Document Sample
       Cardiac and respiratory complications are the two most frequent and most lethal
   groups of complications that occur after general surgery operations. Using modern
   understandings of cardiac and pulmonary pathophysiology, surgeons can now prevent
   or manage these events with frequent patient salvage and full recovery. This issue of
   Selected Readings in General Surgery (SRGS) reviews current information pertinent to
   the successful management of cardiac and respiratory diseases and complications in
   general surgery patients.

Perioperative cardiac complications
       Evidence of atherosclerotic cardiovascular disease is found at autopsy on nearly all
   patients dying after the age of 40 years. Symptoms of atherosclerotic cardiovascular
   disease have become increasingly common as the population of the United States ages
   and cardiovascular disease is the leading cause of death among older adults in North
   America. Increasingly, older patients with moderate-to-severe comorbid cardiovascular
   diseases are presenting for surgical care. Current data estimate that 60%-80% of
   postoperative deaths after elective operation are traceable to cardiovascular
   complications of surgical procedures. In the first section of the overview for this issue of
   SRGS, we review pertinent data on the topic of perioperative cardiac complications.
   Important issues relevant to risk recognition, risk modification, and prevention are
   discussed. Data pertinent to the diagnosis and management of myocardial infarction,
   cardiac failure, arrhythmias, and cardiac arrest will be reviewed. Fundamental aspects
   of the diagnosis and management of cardiac conduction system disorders and
   management of pacemakers and implantable defibrillators are included.

Risk factors for postoperative cardiac complications
       Effective prevention of perioperative cardiac complications is possible only if
   patients at risk can be identified. Identification of high-risk patients can lead to
   development and use of preventive strategies. These approaches obviously will be most
   useful for patients who are scheduled to undergo elective operations. In this patient
   group, there is time for a detailed history and physical examination, laboratory studies,
   electrocardiogram, and imaging. The articles reviewed in this section of the overview
   detail the fundamental features of perioperative cardiac risk assessment and risk

      The first article reviewed is by Davenport and coauthors1 entitled, “Multivariable
   predictors of postoperative cardiac adverse events after general and vascular surgery:
   results from the patient safety in surgery study.” This article is supplied as a full-text
   reprint with this issue of SRGS. The authors begin noting that cardiac complications
   occur after 1%-5% of surgical procedures. Given an annual number of operations
   currently exceeding 30 million, this estimate would result in as many as 1.5 million
adverse cardiac events annually. The estimated mortality for adverse cardiac events
exceeds 50%. Thus, as many as 750,000 deaths could be expected annually.

    The authors cite data that have identified age > 75years, diabetes mellitus,
hypertension, and baseline electrocardiographic findings indicative of ischemia as risk
factors for perioperative adverse cardiac events. The current means of estimating the
risk of perioperative cardiac events are summarized in three available scoring systems
focusing on factors pertinent to the operation (elective versus emergency; simple
versus complex), and on cardiac-specific risk factors such as a history of hypertension,
symptomatic ischemic heart disease, diabetes, and cardiac failure. The authors stress
that popular risk scoring systems, introduced in the late 1980s, award one point for
each of several risk factors; clinical reviews of these systems have noted increased risk
of adverse cardiac events with increasing risk scores. Nonetheless, there remains
controversy over the ability of the available scoring systems to identify accurately
patients for whom the procedure should be delayed in order to conduct further
evaluation. Furthermore, assigning a specific risk to an individual patient is difficult
using existing scoring systems.

    In an attempt to clarify and improve cardiac risk scoring for surgical patients,
Davenport and coauthors used the Patient Safety in Surgery database that contains a
standard dataset for patients from 128 Veterans Administration hospitals and from 14
academic medical centers. This database contains multiple demographic, preoperative,
perioperative, and outcomes variables obtained from medical record reviews
conducted in a standard fashion by experienced nurse reviewers using standardized
definitions. Data on more than 180,000 patients were subjected to multivariate logistic
regression analysis. Adverse cardiac events were defined as cardiac arrest or acute
myocardial infarction within 30 days of operation. Adverse cardiac events were
recorded in 2362 patients and the mortality rate for these events was 60%. The authors
tested a predictive model on a sample of patients drawn from the database after logistic
regression modelling of risk factors from the entire database.

    Prediction of adverse events was accurate using a model that included ASA score,
operation complexity (as reflected in procedure relative value work units), age, and
type of operation. Interestingly, none of the conventional cardiac specific risk factors
such as hypertension, prior history of myocardial infarction, or prior history of a
cardiac surgical procedure was valuable as a predictor of perioperative adverse cardiac
events. When the subgroup of patients from non-VA medical facilities was considered,
the authors found that these patients, as a group, were younger and contained more
women than the VA cohort. The frequency of adverse cardiac events was less in this
subgroup but cardiac-specific risk factors failed to predict outcomes in this subgroup
also. The authors provide a table of risk point assignments for the factors they
   identified as most influential in determining outcomes. A graph in their report indicates
   that significant cardiac risk (>1% risk of adverse event) is recognized beginning with a
   total risk score of 12-15 points.

       Davenport and colleagues emphasize the diminished predictive power of
   conventional cardiac specific risk factors and report that these factors become less
   predictive when considered together with more global risk indicators such as ASA
   score. They also note that, in the current era, patients with known conventional risk
   factors are often treated preoperatively with medications and, occasionally,
   interventions that serve to reduce risk. This would also work to reduce the influence of
   cardiac-specific risk factors.

       They further emphasize the lethality of adverse cardiac events. The mortality risk
   for patients who sustain these events is large, and recognition of this serves as a
   stimulus to improve perioperative management. Use of measures such as avoiding
   emergency operation, preoperative stabilization of cardiac failure and rhythm
   disturbances, optimization of intraoperative monitoring, use of regional anesthesia, and
   use of drugs to control heart rate and stabilize atherosclerotic plaque, are potentially
   useful measures. These are discussed in more detail in the following sections of the
   overview. If emergency operation can be avoided, preoperative approaches to optimize
   coagulation, renal function, and nutrition might assist in minimizing the risk of cardiac
   adverse events. The authors conclude that their approach to outcomes prediction is
   well suited for inclusion in efforts to identify high and low performing hospitals as is
   done in the National Surgical Quality Improvement Program (NSQIP) sponsored by the
   American College of Surgeons. Furthermore, their risk scoring system is designed for
   easy incorporation into electronic medical records. The risk prediction model could be
   made available at the bedside on a hand-held computer.

Risk assessment
       Additional detailed data relevant to cardiac risk assessment is in an article by
   Poldermans and coauthors2 in the Journal of the American College of Cardiology, 2008.
   These authors report European experience with adverse cardiac events in the
   perioperative period. An annual rate of 400,000 perioperative cardiac events has been
   recorded in the European Union. One hundred thirty-three thousand deaths occurred
   because of these complications. They agree with Davenport and colleagues that the type
   of operation is a major driver of risk. They cite data showing that patients older than 40
   years of age have a 2.5% risk of adverse cardiac events after operation. The risk rises to
   more than 6% in patients undergoing vascular surgical procedures. They stress that the
   incidence of perioperative cardiac events varies because of the means used to make the
   diagnosis. When the diagnosis was made with Troponin T or I assays, the frequency of
events rose to 25% in high-risk patients. They further agree that advancing age is a
driver of cardiac risk.

    Postoperative myocardial infarction is the most important adverse cardiac event.
The pathophysiology of this complication is complex. Patients whose preoperative
images show discrete areas of impaired myocardial perfusion are thought to be at
increased risk for perioperative myocardial infarction. It is now clear, however, that
"culprit" coronary lesions are not the predominant cause of perioperative myocardial
infarction. Plaque rupture and thrombosis in coronary arteries at sites of noncritical
coronary artery stenosis are frequent causes of perioperative myocardial infarction.
This understanding helps to explain the lack of benefit of preoperative myocardial
revascularization of culprit lesions. Emphasis has shifted away from identification of
culprit coronary lesions and toward global pharmacologic measures for reducing
cardiac risk. The topic of preoperative interventions is discussed in a later section of the

    Poldermans and associates note that features of the metabolic response to
operation contribute to imbalances in myocardial oxygen demand and availability. The
increased secretion of catecholamines results in tachycardia, which can create
unfavorable myocardial oxygen demand/supply situations. This topic is addressed in
more detail in a report by Sander and coauthors3 in Critical Care Medicine, 2005. These
authors identified 69 patients deemed at high risk for adverse cardiac events. In a
subgroup of 39 patients with sustained (>12 hours) tachycardia (heart rate > 95 bpm),
the risk of a major adverse cardiac event was 49%. In the 30 high-risk patients who did
not have tachycardia, the risk of an adverse cardiac event was 13%. The majority of the
tachycardic rhythms were sinus tachycardia, although there were 16 patients with new-
onset atrial fibrillation.

    The authors cite data that document an association of tachycardia with prolonged
ST-segment depression, a finding known to predict perioperative myocardial infarction.
Poldermans and coauthors2 agree, noting that prolonged ST-segment depression is a
known precursor of perioperative myocardial infarction in patients undergoing
vascular surgical procedures. In Sander’s report, tachycardia occurred during the 24-
hour period in which the myocardial infarction occurred in 90% of patients. Sander and
associates conclude with the observation that the subgroup of their patients where no
tachycardia occurred were more likely to be receiving β-blocking drugs and epidural
analgesia. They suggest that these factors might be protective against tachycardia and
the adverse cardiac events that accompany this change in cardiac rhythm.

   Poldermans and colleagues point out that the inflammatory response that
sometimes follows major surgery procedures creates an environment that contributes
to hypercoagulability because of activation of the coagulation mechanism and reduced
fibrinolytic activity. As noted in the discussion of atherosclerotic vascular disease in a
previous three-issue series of SRGS (Volume 35, Numbers 1-3), inflammatory cytokines
are potent forces that produce plaque instability and rupture. Inflammation might also
contribute to the onset of postoperative tachyarrhythmias. This topic is discussed in an
article by Anselmi and coauthors4 in Annals of Thoracic Surgery, 2009, focusing on
causes of new-onset atrial fibrillation after cardiac operations.

    These authors note that cardiopulmonary bypass is a potent stimulus of the
inflammatory response. Inflammation, as evidenced by elevated levels of C-reactive
protein, is associated with increased risk of atrial fibrillation in surgery and nonsurgery
patients. Lower C-reactive protein levels have also been associated with improved
responses to cardioversion for new-onset atrial fibrillation. They cite one study where
reduction of risk for recurrent atrial fibrillation occurred with specific anti-
inflammatory therapy with the antioxidant Vitamin C. Anselmi and associates point out
that reductions of perioperative inflammation observed with off-pump coronary artery
bypass, use of perioperative corticosteroids, and use of preoperative statin drugs are all
associated with lowered risk of atrial fibrillation. These observations support an
association between perioperative inflammation and perioperative atrial fibrillation.

    Because new-onset atrial fibrillation is associated with perioperative cardiac events,
as noted by Sander and associates3, efforts to control the inflammatory response seem
warranted. Available pharmacologic therapies that reduce inflammation such as β-
blockers and statins are discussed in a subsequent section of the overview.

    Poldermans and coauthors2 go on to note that existing cardiac risk scoring systems
are imprecise. Improved risk assessment would result with the inclusion of global risk
factors such as age and operation characteristics. This assertion is in agreement with
the findings of Davenport and coauthors1, noted above. The difficulty encountered by
clinicians attempting to quantify cardiac risk preoperatively using the available scoring
systems has stimulated researchers to investigate alternative means of assessing risk
for perioperative cardiac events. Two of these approaches are discussed here.

    Normal cardiac function depends, in large measure, on an appropriate balance
between the influences of the sympathetic and parasympathetic nervous systems. This
balance is assessable clinically by use of special ambulatory electrocardiographic
monitoring with assessment of heart rate variability. Heart rate variability is a sign of
healthy cardiac function, while loss of variability signals an imbalance in the relative
influences of the sympathetic and parasympathetic nervous systems on the heart. Loss
of heart rate variability has been associated with an increased risk of sudden cardiac
death, an increased risk of death in multiple trauma patients, and adrenal insufficiency
in the critically ill surgical patient.

    The fundamental physiology of heart rate variability and the potential application of
this assessment to the preoperative evaluation are discussed in an article by Laitio and
coauthors5 in Anesthesia and Analgesia in 2007. These authors begin by noting that
heart rate variability is an indicator of the integrated function of the parasympathetic
nervous system (especially vagus nerve activity), the sympathetic nervous system, and
the baroreceptor system. They describe the various measures used in analyses of heart
rate variability, including time domain analyses that express variability in terms of
instantaneous heart rate and intervals between normal QRS complexes. Frequency
domain analyses commonly express variability in terms of “power-law” spectral
analyses of RR-interval variability.

    These methods of assessing heart rate variability are time-tested and accepted but
they do not adequately describe the complex, fractal system that is heart rate
variability. Because of this, dynamic assessments of heart rate variability have been
developed to analyze correlations of multiple time series of RR intervals. These
analyses can be graphed and patterns typical of normal patients, patients with heart
failure, and patients prone to ischemic events can be displayed. Similar graphic displays
are obtained using Poincare plots. These show typical compact “comet shaped” patterns
in normal patients and patients analyzed after myocardial revascularization. These
graphs show diurnal variation. Heart rate variability changes during ischemic episodes
are characterized by irregular widely spread graphic patterns with loss of diurnal

    The actions of anesthetic drugs to down-regulate vagal activity result in changes in
heart rate variability. Changes in heart rate variability accurately predict hypotensive
episodes after induction of spinal anesthesia especially when the block reaches the
thoracic spinal levels. Loss of heart rate variability in elderly patients and in diabetic
patients with dysfunction of the autonomic nervous system accurately predicts
episodes of hemodynamic instability. The authors cite several studies that have related
loss of heart rate variability to short- and long-term operative mortality from
myocardial ischemia and prolonged ICU stays.

    The presence of heart rate variability abnormalities improves the predictive
capability of the available cardiac risk scales, especially for predicting long-term cardiac
mortality. In the cited studies, the combination of abnormal heart rate variability and
high-risk scores accurately predicted perioperative cardiac morbidity for patients
undergoing cardiac and noncardiac procedures. Laitio and colleagues speculate that
loss of heart rate variability indicates unopposed sympathetic influence on the heart
   that might increase myocardial oxygen demand by augmenting ventricular contractility.
   This situation favors the development of ischemia in susceptible patients. The need for
   24-hour electrocardiographic monitoring and manual assessment of the tracings are
   significant disincentives that have reduced the utility of heart rate variability
   measurements for preoperative patients. With improved, computer-assisted
   methodologies, these disadvantages might be overcome.

      Additional data on the use of heart rate variability assessments as a means of
   predicting perioperative cardiac morbidity risk are found in an article by Hanss and
   coauthors6 in Anaesthesia, 2008. These authors report an initial analysis of 50 patients
   who underwent heart rate variability analysis preoperatively and, during the
   postoperative period, had 24-hour electrocardiographic monitoring and sequential
   measurements of creatine kinase MB band in blood samples. Cardiac events were
   detected by a combination of electrocardiographic changes and elevations of the CPK-
   MB level. Seventeen of the initial patients had cardiac events and the authors
   established that a heart rate variability power value <400 ms2 Hz-1 was a useful cut-off
   value for prediction of cardiac events. This cut-off value was then assessed
   prospectively in 50 additional patients. Cardiac events and hospital length of stay were
   both increased in the 26 patients with low power scores in the prospective group. The
   authors conclude that heart rate variability power analysis is a useful predictor of
   postoperative cardiac events and the additional information might improve the
   predictive power of cardiac risk scoring systems.

       Additional predictive power can be obtained using serum markers that reflect
   vulnerability of the myocardium to ischemia. Two of these, B-type natriuretic peptide
   (BNP) and N-terminal pro-brain natriuretic peptide (NT-proBNP), are discussed. The
   first of these is by Cuthbertson and coauthors7 in the British Journal of Anaesthesia,
   2007, who analyzed outcomes data on 204 patients undergoing major noncardiac
   surgery procedures. Preoperative BNP levels were obtained in each patient.
   Perioperative cardiac events were defined as an elevation of the troponin level or death
   within three days of operation. The authors found that a preoperative elevation of BNP
   >40 pg/mL was predictive of perioperative cardiac events. They performed rigorous
   multivariate statistical analysis and found that the preoperative BNP level was more
   accurate for risk prediction than findings on history, physical examination, or
   electrocardiogram. BNP was more predictive than the revised cardiac risk score.
   Nonetheless, five patients with significant postoperative cardiac events were not
   identified by the preoperative BNP elevation. This observation suggests that BNP
   cannot be used alone to establish risk for perioperative adverse cardiac events.

An article analyzing the potential usefulness of NT-proBNP levels assessed preoperatively
and with a single postoperative sample by Mahla and coauthors8 appeared in
Anesthesiology, 2007. These authors analyzed results in 218 patients who underwent major
vascular reconstructive procedures. Patients were followed for 30 months postoperatively.
Twenty percent sustained a significant cardiac event during the followup interval. The
authors found that a preoperative value for NT-proBNP equal to 280 pg/mL coupled with a
postoperative increase to 557 pg/mL accurately predicted short- and long-term cardiac
morbidity. The authors state that NT-proBNP is released from cardiac myocytes in
response to ischemia and stretch and, therefore, this hormone might be a good candidate
for prediction of perioperative myocardial ischemia. The accuracy of the preoperative and
postoperative levels combined was good, with an area under the receiver-operating-
characteristic curve of 0.8 that is as good as or better than all available risk-scoring
systems. The authors conclude that this test might offer improved prediction of cardiac
events in high-risk patients undergoing major vascular operations.

Preoperative evaluation for coronary artery disease and preoperative interventions
       Once a high-risk patient is identified, the next decision concerns the need for
   additional preoperative cardiac testing. Exercise testing, dobutamine stress
   echocardiography, and myocardial scintigraphy with vasodilator stress are tests
   commonly contemplated. Recommendations by the American Heart Association state
   that testing is indicated in patients who have symptoms suggesting unstable cardiac
   syndromes (decompensated cardiac failure or unstable angina), patients who have poor
   functional capacity where a high-risk operation is contemplated (major vascular
   reconstruction), and patients with known valvular heart disease. Poldermans and
   colleagues stress the lack of a positive contribution of preoperative cardiac testing in
   cardiac stable patients, especially those who are already using β-blocking drugs and
   statins with good control of heart rate. They further emphasize that coronary artery
   bypass in cardiac stable patients has not resulted in improved surgical outcomes and
   the intervention delays the planned noncardiac procedure. Percutaneous coronary
   interventions similarly delay the planned procedure because the risk of stent
   thrombosis is substantial in the first weeks following stent placement when multiple
   drug antiplatelet therapy is used. With drug-eluting stents, this interval might be as long
   as one year.

       Additional data on the use of extensive preoperative cardiac testing and
   preoperative cardiac interventions to prevent perioperative adverse cardiac events are
   reported by Jaroszewski and coauthors9 in the Journal of Thoracic and Cardiovascular
   Surgery, 2008, who performed a retrospective review of 294 patients who underwent
   thoracotomy for a noncardiac operation in a single institution. One hundred eighty-four
   patients underwent extensive preoperative assessment including, in addition to history
   and physical examination with 12-lead electrocardiogram, stress testing, stress
   echocardiography, and/or myocardial scintigraphy. Based on preoperative test
   findings, 40 patients were selected to undergo coronary angiography and four of these
   had preoperative coronary revascularization by either operation or stenting. There was
   no difference in the frequency of perioperative myocardial infarction in patients who
   had testing and intervention compared with those who did not have testing. In fact, of
   the four patients who underwent revascularization, two had perioperative myocardial
   infarction. One of these was from perioperative coronary stent thrombosis. These
   authors concluded that there was no benefit to testing and intervention in cardiac
   stable patients.

Medications for reducing cardiac risk
       Poldermans and coauthors2 stress the importance of general medical approaches to
   cardiac risk modification in patients undergoing major elective operations. These
   approaches have generally consisted of careful risk stratification, and use of
   pharmacologic approaches designed to alter, favorably, myocardial oxygen
   consumption/oxygen demand relationships as well as stabilize plaque through control
   of perioperative inflammatory responses. The authors note that the high catecholamine
   release states created by the stress of operation alter both myocardial energetics and
   inflammation. Initial approaches to balancing myocardial energetics included the use of
   β-blocking drugs. Initially, drugs used were combined β-1 and β-2 agents, such as

       As additional human trial data have become available, important lessons have been
   learned. Poldermans and associates describe these progressive steps. For example, they
   stress the observation that trials of beta blockade have generally shown reductions in
   the frequency of perioperative cardiac events but some trials have disclosed risks of
   hypotension and stroke, especially in older patients not judged at high risk for cardiac
   events. Two prospective, randomized trials cited by these authors (references 40 and
   45 in their bibliography) showed effective reduction in cardiac events but at the cost of
   a significantly increased risk of stroke and overall mortality. These trials disclosed the
   potential danger of pharmacologically lowering blood pressure to 100 mmHg or lower
   in elderly patients. The type of beta-blocking drug used, timing of drug therapy, and
   dosing are also important features of approaches that achieve maximum success. Drugs
   that are β-1 selective agents (such as bisoprolol) are more effective than drugs that
   target both the β-1 and β-2 receptors. Blockade of both receptor types results in a state
   of predominant α receptor stimulation that results in hypertension and increased
   myocardial stress.

      Prospective, randomized trials have shown no impact of beta blockade on the risk of
   perioperative cardiac events if the drug is started on the day before or the day of the
   operation. This finding implies that maximum stabilization of cardiac energetics and
   plaque requires time. In fact, the DECREASE trial (reference 37 in Poldermans’
   bibliography) started treatment, on average, 37 days before operation and with
incremental adjustment of the dose upward based on blood pressure and heart rate.
This study disclosed a 10-fold reduction in the risk of perioperative cardiac events and

   All of the data cited by Poldermans and associates support the use of β-1 selective
drugs with a long half-life. The drug should be started at least one month before
operation and adjusted to obtain optimum heart rate (current American Heart
Association recommendations are resting heart rates in the 60-66 bpm range) without
episodes of hypotension. Patients at low risk for cardiac events should not take beta-
blocking drugs unless they are using these drugs chronically. Moderate risk patients
undergoing major vascular operations are acceptable candidates for beta-blocking
therapy and high-risk patients undergoing any type of major operation are excellent
candidates for this approach. Drugs should not be withdrawn in the perioperative
period because benefit has been shown for protection against both short-term and
long-term cardiac morbidity. Downsides of beta-blocking drug usage include a range of
contraindications (asthma, for example) and consistent observations that up to 25% of
patients have episodes of tachycardia in the perioperative period despite seemingly
adequate beta blockade.

    Further analysis of the use of beta blockade for prevention of perioperative cardiac
events comes from a review by Chopra and coauthors10 entitled “Perioperative beta-
blockers for major noncardiac surgery: Primum non nocere.” This article appeared in
The American Journal of Medicine in 2009 and a full-text reprint of the article is
provided with this issue of SRGS. These authors review the actions of beta-blocking
drugs. They note that there are three subtypes of beta-receptors and these receptors
are presented on the cell surface of many types of human tissue. Beta one receptors are
found in the myocardium, the kidney, and the eye. Beta two receptors are found in
adipose tissue, liver, pancreas, smooth muscle, and skeletal muscle. Beta three
receptors are primarily involved with metabolic regulation and lipolytic pathways. The
receptors are G-coupled proteins that activate intracellular adenyl cyclase and produce
intracellular effects via adenosine monophosphate production and opening of
excitatory channels. Chopra and associates confirm the observations of Poldermans and
coauthors2 and suggest that beta blockade use be targeted toward patients at high risk
for perioperative cardiac events. Beta-blocking drug therapy should be started at least
one month before operation, and tight heart rate control should be sought with
maximum protection against hypotensive events. Drug therapy should be continued
during the postoperative period, and that use of statins and/or aspirin should be

   Supporting data on the value of tight heart rate control is in a meta-analysis
authored by Beattie and coauthors11 in Anesthesia and Analgesia, 2008. These authors
provide an analysis that specifically focuses on reasons for inadequate heart rate
control in trials of beta-blocking drugs. They emphasize that early implementation of
beta blockade with progressive upward adjustment of dose to obtain consistent resting
heart rates in the 60-65 bpm range is associated with maximum reductions of risk of
perioperative cardiac events. Several recent trials have shown that fixed dose
approaches do not produce maximum protection against perioperative cardiac events.
They stress that the available trial data strongly indicate that variation in achieving
optimum heart rate control accounts for 60% of the variability in trial results. They also
document that the type of drug used and the concomitant use of calcium channel
blockers might alter the heart rate response.

    Based on their analysis they recommend that beta-blocking drugs other than
metoprolol be chosen, especially in patients concomitantly using calcium channel
blockers. Most of the available trials disclose failure to achieve heart rate goals in 20%-
35% of patients. Suboptimum heart rate control might be the result, according to
Beattie and colleagues, of the presence of the AA variant of the beta-receptor resistant
to the beta-blocking drugs. In the setting where optimum heart rate is not achieved,
combination therapy with calcium channel blockers might be necessary.

    Beattie and colleagues point out that available data does not adequately address the
risk of exacerbation of congestive heart failure from tight heart rate control with beta-
blockers. Nor does the data adequately evaluate the use of other approaches to
stabilization of myocardial energetics such as the use of regional anesthesia/analgesia
and α-2 receptor agonists, which have both been shown to reduce the frequency of
perioperative cardiac events. The authors recommend therapy to obtain optimum heart
rate control but caution that the best approach to achieve this goal might not be
available yet.

    An alternative approach to achieving optimum balance between control of heart
rate and maintenance of cardiac output is described in an article by Suttner and
coauthors12 in the British Journal of Anesthesia, 2009. These authors note that concerns
about the effect of beta-blocking drugs on blood pressure and cardiac output have led to
reluctance on the part of some clinicians to use beta blockade for high-risk patients,
especially those who might need urgent or emergent intervention. In the current study,
an analysis is presented of results in 75 high-risk (as defined by three or more risk
factors) vascular surgery patients randomized to receive continuous perioperative beta
blockade with intravenous esmolol alone, esmolol plus enoximone (a
phosphodiesterase type III inhibitor), and standard therapy.

   Perioperative cardiac events were documented by elevations of troponin or BNP.
The authors explain that phosphodiesterase inhibitors such as enoximone and
milrinone have the potential to maintain cardiac contractile function when
catecholamine pathways are pharmacologically blocked. In this study, they noted no
abnormalities of troponin in either group receiving esmolol. BNP was lowest in the
esmolol + enoximone group and this group had the best maintenance of cardiac index.
They suggest that enoximone support of cardiac index occurred because of its action
that promotes influx of calcium into myocytes, thereby favoring increased contractility.
In vascular smooth muscle, enoximone promotes calcium efflux, favoring vasodilation.

    These authors rightly caution that dosing of esmolol and enoximone should be
managed carefully because higher doses might predispose to new-onset arrhythmias.
While these salutary effects were obtained in high-risk vascular surgery patients, the
numbers of patients are small. Furthermore, even though these patients were said to be
at high risk for perioperative cardiac events, fewer than 20% of each group received
preoperative beta-blocker and/or statin therapy. The results of this study suggest, but
do not prove, that an approach such as described might have protective effects in high-
risk patients who are not using beta-blocking drugs and who require urgent or
emergent operation.

    Additional plaque stability effects accrue to patients from the use of drugs such as
statins and aspirin. Several trials have shown protection against perioperative cardiac
events, and both short- and long-term mortality with the use of statins. Sustained
release preparations are preferred because intravenous statin preparations are not
available. As with beta-blocking drugs, therapy should not be withdrawn
postoperatively, and patients who are chronically using statins should have these
continued in the postoperative period. Low-dose aspirin has also been shown to be
protective against both short- and long-term cardiac events and death.

    Because of data suggesting benefit in terms of reduction of perioperative cardiac
events for high-risk patients from statins and from beta-blocking drugs, it is useful to
determine whether using the drugs in combination would be helpful. This issue is
addressed in a study by Dunkelgrun and coauthors13 in Annals of Surgery, 2009. This
article concerns a randomized prospective trial of beta blockade using bisoprolol
compared with the use of a statin drug (fluvastatin) alone, a combination of the two
drugs, or neither drug. The patients were deemed intermediate risk (cardiac event risk
of up to 6%). The authors noted that cardiac event rates were significantly reduced in
patients receiving beta blockade with or without the statin drug. A lesser reduction
(nonsignificant) was seen with the statin drug alone. Although this study does not
support the addition of statin drugs to beta blockade as a means of gaining additional
control of cardiac event risk, the study is limited because of the small number of
enrolled patients.
      From the perspective of the editor, there is convincing evidence to support careful
  preoperative cardiac risk assessment. Furthermore, it is expected that an increasing
  number of patients will present for operation already taking beta-blocking drugs,
  statins, or both. In this case, drug therapy with both drugs should be continued during
  and after the perioperative period with dose and type of drug adjusted to make certain
  that full effects of both drugs are maintained. For intermediate-risk patients undergoing
  high-risk operations (abdominal or thoracic vascular procedures) and for high-risk
  patients, beta-blocking drug therapy, at least, should be implemented and dosage
  adjusted progressively during the preoperative interval to obtain a resting heart rate in
  the 55-65 bpm range. Therapy should continue into the postoperative recovery period.
  Other adjuncts, such as regional anesthesia/analgesia, aspirin therapy, and statin
  therapy might be useful.

Perioperative myocardial infarction
      In the foregoing discussion, emphasis was placed on the vulnerability of coronary
  artery plaque and the hazard of plaque rupture with thrombosis of the coronary artery
  as the proximate cause of perioperative myocardial infarction. The significant, and
  increasing, prevalence of coronary artery disease in surgery patients is a reminder to
  surgeons that this problem is a continuing challenge. Increased resource consumption
  from postoperative myocardial infarction is significant. In a 2006 report by Mackey and
  coauthors,14 results from a prospective analysis of 236 patients deemed at high risk
  showed significant incremental increases in both hospital and ICU lengths of stay when
  vascular surgery patients developed a perioperative myocardial infarction.
  Perioperative myocardial infarction was a marker for long-term use of healthcare
  resources as well.

      Nearly one-quarter of the study patients discharged alive returned to the emergency
  department for care during the year after discharge. Frequently, postoperative
  myocardial infarction occurs without chest pain. Nonspecific signs such as hypotension,
  dyspnea, arrhythmia, onset of new cardiac murmur, and alterations in the level of
  consciousness might be the only clinical symptoms. Electrocardiographic diagnosis and
  laboratory diagnosis using serum markers such as troponin might yield nonspecific
  results. The typical electrocardiographic findings of spontaneous myocardial infarction
  include the appearance of Q waves, ST-segment elevation, and T-wave inversion. In
  contrast, postoperative myocardial infarction is associated with intervals, occasionally
  prolonged, of ST-segment depression indicating subendocardial ischemia. Increasingly,
  echocardiographic cardiac imaging is used to obtain diagnostic information. Features of
  the pathophysiology, diagnosis, and management of perioperative myocardial
  infarction will be discussed.
Pathophysiology of myocardial infarction
       The first article discussed is by Burke and Virmani15 entitled “Pathophysiology of
   myocardial infarction.” The review appeared in Medical Clinics of North America in
   2007. The authors begin by noting that 80% of spontaneous myocardial infarctions are
   caused by thrombosis of coronary arteries critically narrowed by atherosclerotic
   plaque. Unusual causes of myocardial infarction are coronary embolization, coronary
   spasm, and thromboses of nondiseased coronary arteries. Concentric subendocardial
   necrosis that might result from prolonged global ischemia from cardiac arrest can also
   lead to coronary artery thrombosis. Myocardial ischemia results in acute pallor of the
   myocardium, visible grossly within 12 hours of the onset of ischemia. Tetrazolium salt
   staining of the myocardium can detect myocardial necrosis within 2-3 hours of the
   onset of ischemia. After 5-7 days, the infarcted area is soft with a hyperemic border. If
   reperfusion occurs, the infarcted area might be reddened from trapped red blood cells.
   Healing of a myocardial infarction takes from 4 weeks to 3 months and the lesion
   evolves into a white scar, which might be the source of rhythm disturbances. Histologic
   findings begin with the development of tissue eosinophilia followed by typical
   inflammatory changes, followed by fibrosis and scarring. Infarctions that involve more
   than 50% of the myocardial wall thickness are termed transmural and these produce Q-
   wave changes in the electrocardiogram.

       In humans, reperfusion of ischemic myocardium within 4-6 hours of the onset of
   ischemia results in myocardial salvage. In this circumstance, the ischemic area remains
   subendocardial and transmural extension does not occur.

      Myocardial energy metabolism depends upon the oxidation of free fatty acids to
   produce ATP. Ischemia causes an immediate shift to anaerobic glycolysis. Exhaustion of
   ATP supply leads to inhibition of Na/K ATPase with breakdown of cell membrane
   defenses and influx of sodium and chloride into the myocardial cell. Increases in
   cytosolic calcium and cellular acidosis lead to myocyte contractile dysfunction. Cell
   death can result from necrosis, oncosis, apoptosis, or autophagy. Because apoptosis is
   an energy consuming function, this occurs in perfused myocardium surrounding the
   necrotic area. Autophagic cell death also requires energy and occurs in a manner that is
   independent of the caspase-mediated pathway leading to apoptosis.

       Infarct size is determined by the extent and efficiency of coronary collateral
   circulation. Well-developed coronary collaterals are present in approximately 40% of
   adult men and these individuals are resistant to the development of transmural
   infarctions. Rather, coronary atherosclerosis in these patients produces anginal pain. In
   patients with well-developed collateral circulation, another means of myocardial
   protection is ischemic preconditioning. Ischemic preconditioning is the term applied to
   the phenomenon of preservation of myocyte energy-producing capability after an
ischemic event preceded by a short interval (10 minutes) of ischemia followed by
reperfusion. Potassium-ATP channels play a central role in ischemic preconditioning.
Blockage of these channels prevents the protective effect of ischemic preconditioning.
Interestingly, cardiac myocyte protection can be induced by ischemic events in distant
tissue sites. This phenomenon is known as remote ischemic preconditioning.

    A review of the potential for remote ischemic preconditioning to produce cardiac
myocyte protection comes from an article by Walsh and coauthors16 in the Journal of
Vascular Surgery, 2009. These authors report that myocyte protection has been
produced after the production of ischemia to kidney, intestine, and skeletal muscle.
Preoperative tourniquet ischemia of an upper extremity was associated with reduced
risk for postoperative cardiac events in patients undergoing coronary artery bypass
grafting. These authors report results of ischemic preconditioning in randomized
analysis involving 82 patients undergoing open abdominal aortic aneurysm repair. Ten
minutes of ischemia to each leg was produced by clamping the iliac arteries
individually. Thirteen of the 42 control patients developed clinically significant
perioperative myocardial ischemia. Only two of the 40 patients who had ischemic
preconditioning developed myocardial ischemic events. Because this study was
conducted in patients who had undergone maximum preoperative preparation with
beta-blocking drugs, the results suggest there might be incremental protection because
of ischemic preconditioning; this technique should be further evaluated.

    Burke and Virmani15 assert that plaque instability universally preceded coronary
thrombosis. Seventy-five percent of coronary thromboses are the result of plaque
rupture and the remaining 25% result from plaque erosion. The left anterior
descending coronary artery is the most frequent site of thrombosis, followed by the
right coronary artery and the left circumflex coronary artery. Arrhythmias and
contractile dysfunction in myocardium distal to a thrombosis might be aggravated by
post-thrombosis microembolization. Complications of myocardial infarction include
cardiac rupture, ventricular aneurysm, mural thrombus with embolization, mitral valve
insufficiency from papillary muscle rupture, and pericardial effusion.

   Additional information on complications of myocardial infarction is in a review by
Wilansky and coauthors17 in Critical Care Medicine in 2007. These authors provide short
descriptions of clinical characteristics of the most important complications of
myocardial infarction. Left ventricular free wall rupture, a frequently lethal
complication of myocardial infarction, traditionally has afflicted up to 6% of patients
sustaining myocardial infarction. With the onset of rapid reperfusion protocols and
angioplasty, the frequency of this complication has dropped to 1%. Nonetheless, up to
17% of the deaths from myocardial infarction result from ventricular free wall rupture.
This complication occurs within the first week after infarction with nearly half
occurring during the first 24 hours. Older age, male gender, first infarction, single vessel
disease, lack of ventricular hypertrophy, transmural infarction and anterior location of
the infarction are all risk factors for left ventricular free wall rupture. This condition can
result in acute hemopericardium and pericardial tamponade.

    Approximately one-third of patients with free wall rupture present with a more
subacute clinical picture characterized by persistent chest pain, right heart failure, and
hemodynamic deterioration. Electrocardiogram findings are nonspecific.
Echocardiography might show pericardial effusion. As noted in the article by Burke and
Vimani,15 approximately 25% of myocardial infarction patients will have nonspecific
pericardial effusion, and this will make diagnosis of cardiac rupture difficult. Doppler
imaging or contrast echocardiography might be needed to show pericardial blood clot
or the rupture site. Surgical repair of the rupture will be required. Some patients might
be amenable to stabilization with fluids, pressors, and/or intraaortic balloon pump.

    A variant of cardiac rupture is ventricular septal rupture. Older age, female gender,
hypertension, absence of a smoking history, and anterior infarction location are risk
factors for septal rupture. Clinically, this complication presents with hemodynamic
collapse in the presence of a new systolic murmur. Diagnosis is established with
echocardiography. Surgical revascularization and septal repair are therapies of choice.

    Left ventricular outflow tract obstruction from severe systolic anterior motion of
the mitral valve is an unusual complication of myocardial infarction. The clinical
presentation is one of a new systolic murmur and refractory hypotension in the setting
of an apical infarction. Echocardiography can confirm the diagnosis. Therapy includes
volume expansion, beta-blocking drugs to reduce hyperdynamic contraction of the
heart, and alpha agonists to support blood pressure.

    Mitral regurgitation might complicate myocardial infarction because of ischemia of
the valve or from papillary muscle rupture. Ischemic mitral regurgitation might be
clinically silent and evidenced only by the presence of a cardiac murmur.
Transesophageal echocardiography is the mainstay of diagnosis. Management varies
according to the clinical status of the patient and the hemodynamic effects of the
valvular dysfunction. Papillary muscle rupture is a critical care emergency with acute
pulmonary edema and cardiogenic shock commonly present. A loud systolic murmur is
present. Immediate management includes support of cardiac function with afterload
reduction and the use of an intraaortic balloon pump. Transesophageal
echocardiography provides accurate delineation of the valvular anatomy and the extent
of dysfunction. Surgical management of the mitral regurgitation and critical coronary
stenoses is associated with significant operative mortality (25%-40%), but survivors
have good quality of life in long-term followup.
Diagnosis and management of postoperative myocardial infarction
       Traditionally, the diagnosis of myocardial infarction is made based on the presence
   of typical chest pain, electrocardiographic evidence of ischemia (ST-segment elevation,
   presence of Q waves), and elevation of biomarkers such as troponin. As noted
   previously, perioperative myocardial infarction might be clinically silent. Chest pain
   might be absent because patients are receiving analgesics, are sedated for mechanical
   ventilation, or are emerging from general anesthesia. Troponin levels might be elevated
   in surgery patients in the absence of myocardial infarction, but persistent elevations of
   troponin > 3, especially combined with ST segment depression intervals of > 60 min on
   electrocardiographic monitoring, predict an increased risk of myocardial infarction and

       An article evaluating diagnostic accuracy of the electrocardiogram in critically ill
   patients by Lim and coauthors appeared 18 in Critical Care Medicine, 2006. These
   authors determined intra- and inter-rater reliability for electrocardiogram
   interpretation in patients at high risk for myocardial infarction in a single ICU. The
   authors reaffirm the difficulties in detecting clinical symptoms of myocardial ischemia.
   Interpreting troponin levels in patients recovering from noncardiac operations and in
   patients who are critically ill is also challenging. Lim and colleagues state that the lack
   of reliability of troponin measurements has led to increased emphasis on
   electrocardiographic changes as a means of confirming the diagnosis of myocardial
   infarction. This study was an analysis by two observers of all electrocardiograms
   obtained on patients at risk for myocardial infarction in a single ICU during two months.

       The changes sought as evidence of myocardial infarction were those recommended
   by the European Society of Cardiology/American College of Cardiology diagnostic
   criteria. The findings included pathologic Q waves, ST-segment elevation in at least two
   contiguous leads, ST-segment depression in at least two contiguous leads, symmetric
   inversion of T-waves (> 1mm) in at least two contiguous leads, T-wave flattening, and
   new onset left bundle branch block. The last criterion was chosen because left bundle
   branch block could obscure ST-segment elevation. The analysis of rater performance
   indicated that intra-rater and inter-rater reliability was poor when the raters had no
   knowledge of the serum troponin level. The raters were more likely to diagnose
   accurately electrocardiographic signs of myocardial infarction if they knew that there
   was a significant elevation of the serum troponin level. Electrocardiographic
   abnormalities most often identified accurately were T-wave inversion, Q-waves, and left
   bundle branch block. These authors conclude that accurate diagnosis of myocardial
   infarction in critically ill patients (who have physiologic similarities to postoperative
   patients) are facilitated using a synthesis of clinical information that includes the
   electrocardiogram, troponin levels, and, possibly, echocardiographic imaging.
       In an editorial by Engel19 that accompanies Lim’s article, the difficulty in arriving at
   an accurate diagnosis of myocardial infarction is reemphasized. Engel agrees that use of
   the electrocardiogram as the principle means of diagnosis of myocardial infarction in
   postoperative or critically ill patients is hazardous. Furthermore, choosing
   interventional therapy in this patient subgroup is challenging because thrombolysis,
   coronary angiography, and percutaneous coronary interventions requiring antiplatelet
   therapy might not be safe.

Management of the patient who has had a perioperative myocardial infarction is based on
providing support for the patient’s heart function while planning for appropriate means of
revascularization. Support from cardiologists and cardiothoracic surgeons will be needed
to facilitate these decisions. Cardiogenic shock is the most common lethal complication of
perioperative myocardial infarction. Management of this condition is discussed in detail in
the next issue of SRGS.

Perioperative cardiac arrhythmia
       In an earlier portion of this overview, we noted the association of postoperative
   inflammation, postoperative tachycardia, and postoperative atrial fibrillation with
   cardiac morbidity. In patients at high risk for perioperative cardiac events, control of
   heart rate and rapid diagnosis and therapy for treatable tachycardias are important for
   prevention of cardiac complications. The most common treatable tachycardias
   encountered in postoperative patients are supraventricular tachycardias and atrial
   flutter/fibrillation. In this section of the overview, we review pertinent features of the
   diagnosis and management of these cardiac rhythm disorders.

Management of supraventricular tachycardia and atrial fibrillation
        Supraventricular tachycardia is the subject of a review by Fox and coauthors20 in
   Mayo Clinic Proceedings, 2008. A full-text reprint of this article is included with this
   issue of SRGS. The authors provide a working definition of supraventricular tachycardia
   that includes all tachycardias arising cephalad to the bifurcation of the His bundle and
   all tachycardias dependent on the His bundle for impulse transmission. These
   tachycardias usually have rates exceeding 100 bpm (unless atrioventricular conduction
   block is present), and QRS morphology is usually normal. In the presence of bundle
   branch block, however, QRS complexes might be widened or otherwise abnormal in
   shape. Data from long-term ambulatory electrocardiographic monitoring have
   permitted estimates of the incidence of supraventricular tachycardia. The authors cite
   data that disclose an incidence of 76% in a group of elderly patients with a 20%
   incidence of symptomatic coronary artery disease. In studies of asymptomatic healthy
   patients aged 18-65 years, the incidence ranged from 12-18%.
    Supraventricular tachycardia is usually of sudden onset and might spontaneously
terminate. The patient might complain of chest pain, and syncope occasionally occurs
(usually in very rapid tachycardias associated with reductions in cardiac output).
Although no clear association between chest pain during a tachycardia episode and
coronary artery disease has been established, the diagnosis might be suspected in
elderly patients with tachycardia and chest pain. Patients usually complain of
palpitations; patients with chronic heart failure might not sense the palpitations but,
instead, present with cardiac decompensation. The catecholamine response stimulated
by tachycardia and hypotension serves to perpetuate the rhythm disturbance.

    These authors note that atrioventricular nodal re-entry, atypical atrioventricular
nodal re-entry, or atrial tachycardias are the usual mechanisms of these rhythm
disturbances. Atrioventricular node dependent tachycardias are usually terminated by
inducing atrioventricular nodal block with a vagal stimulating maneuver (Valsalva,
carotid sinus massage, or immersion of the face in cold water), or pharmacologically.
Atrioventricular node independent rhythms include atrial flutter and atrial fibrillation.

    Diagnosis of tachycardia is usually possible using a 12 lead electrocardiogram,
which is preferred over a rhythm strip. QRS morphology is usually normal with QRS
duration of 90 milliseconds or less. QRS complexes might be abnormal if there is
intermittent or permanent bundle branch block. Other factors to be considered in
interpreting the electrocardiogram are the heart rate, mode of onset and termination of
the tachycardia, relative position of the P-wave within the RR interval, and morphology
of the P wave. The tachycardia rate is usually higher than 100 bpm and can be variable.
A steady rate of 150 bpm suggests atrial flutter with a 2:1 atrioventricular block,
according to Fox and associates.

    Another means of determining the type of tachycardia is by examining the
relationship of the P wave to the preceding and subsequent R wave. When the distance
between the R wave and the next P wave is longer than the subsequent PR interval, the
tachycardia is a “long RP” rhythm. If the distance between the R wave and the
subsequent P wave is shorter than the subsequent PR interval, the rhythm is termed
“short RP.” Long RP tachycardias are atrial and might progress to flutter or fibrillation.
Supraventricular tachycardias, according to these authors, are mainly short RP
rhythms. At very rapid heart rates, RP and PR intervals become very short and might be
difficult to interpret.

    Management of supraventricular tachycardia is usually straightforward because the
patients are usually hemodynamically stable. If there is instability, the patient is
managed according to the typical ABC approach emphasizing airway, breathing, and
circulation. Vagal maneuvers such as carotid sinus massage might terminate the rhythm
promptly and these maneuvers are ineffective in atrial flutter/fibrillation. Carotid sinus
massage should not be done if there is a carotid bruit present. Pharmacologic
management of supraventricular tachycardia is accomplished using adenosine, calcium
channel blockers, or β-blocking drugs. Adenosine is the first-line drug and is given in 6
mg or 12 mg boluses. Smaller doses are used in patients taking dipyridamole.

     Broadening of the QRS complex might occur in supraventricular tachycardia if there
is bundle branch block. Fox and colleagues caution that if the patient is older than 70
years or there is a history of symptomatic coronary artery disease, a broad QRS
tachycardia should be considered a ventricular tachycardia until proved otherwise. An
article discussing the use of response to adenosine bolus therapy as a means of
differentiating supraventricular from ventricular tachycardia when wide QRS
tachycardia comes from Critical Care Medicine, 2009, by Marill and coauthors.21 The
authors note that differentiation of atrial from ventricular tachycardia when the heart
rate is steady and the QRS complex is widened is important, but current algorithms are
neither sensitive nor specific in identifying the type of rhythm present. Drug therapy
using procainamide or amiodarone might effectively treat the rhythm but side effects
such as hypotension limit the usefulness of these agents. Electrical cardioversion is
effective but is painful, does not protect against recurrence of the rhythm, and offers
little diagnostic information.

    These authors hypothesize that adenosine will safely terminate most
supraventricular tachycardias, will slow heart rate enough to allow detection of atrial
flutter or fibrillation, and will be not predictably alter ventricular tachycardia. In a 15-
year interval, these authors treated 197 patients with steady-rate wide QRS complex
tachycardia with a 12 mg bolus of adenosine. Patients determined to have ventricular
tachycardia were older, more often had a history of myocardial infarction and prior
episodes of ventricular tachycardia. Two of 81 patients with ventricular tachycardia
responded to adenosine while 104 of 116 patients with nonventricular tachycardia
responded to adenosine. There were no serious adverse events (defined as emergent
drug therapy or electrical shock) observed in either subgroup. These authors concluded
that nonresponse to adenosine was the only factor that diagnosed ventricular
tachycardia with a high sensitivity and specificity.

    The nondihydropyridine group of calcium channel blockers (verapamil and
diltiazem) are alternative drugs used to terminate supraventricular tachycardia. A
summary of the data supporting these drugs in comparison to adenosine, by Anugwom
and coauthors22 appeared in American Family Physician, 2007. These authors reviewed
data from eight studies involving nearly 600 patients. The data disclose that adenosine
and calcium channel blockers are equivalently effective in terminating paroxysmal
supraventricular tachycardias. Transient, minor side effects such as flushing, nausea,
and headache are common with adenosine. Severe side effects (cardiac arrest and
hypotension) were observed only in patients treated with calcium channel blockers.
These authors note that the American Heart Association guidelines recommend
adenosine as first-line therapy for paroxysmal supraventricular tachycardia because of
the low risk of severe side effects, the rapid onset of action, and the short half-life of the
drug. The advanced cardiac life support course also recommends adenosine for the
management of supraventricular tachycardia.

    Of the atrial arrhythmias, atrial fibrillation is the most commonly encountered. A
discussion of the management of acute atrial fibrillation is in an article by Siu and
coauthors23 in Critical Care Medicine, 2009. These authors report a randomized,
nonblinded trial comparing the effectiveness of diltiazem, digoxin, and amiodarone for
rate control and symptom improvement in patients presenting acutely with
symptomatic, new-onset atrial fibrillation. The authors note that atrial fibrillation is a
common arrhythmia and the frequency of this condition is increasing. Traditionally,
two approaches have been used to manage atrial fibrillation, rhythm control and rate
control. Rhythm control approaches use direct current cardioversion; this modality
might not be available on a 24/7 basis.

    Guidelines published from the American Heart Association recommend emergency
direct current cardioversion only for patients with acute atrial fibrillation who are
hemodynamically unstable. Direct current cardioversion might require that the patient
be anticoagulated, especially if there is atrial enlargement. This fact limits application of
this modality to postoperative patients. The authors analyzed results in 166 patients.
Patients were excluded from the study if they were unstable, had evidence of
symptomatic coronary artery disease, were hypotensive, had an implanted defibrillator,
had a history of recent myocardial infarction, had a history of heart failure, or had
angina pectoris. Drug therapies used were diltiazem, digoxin, and amiodarone. The
endpoints examined were control of heart rate (heart rate < 90 bpm, sustained, at 24
hours after initiation of therapy) and improvement of symptoms. In this study, rate
control and symptom improvement was best achieved with diltiazem. There was only
one adverse event recorded, an episode of phlebitis at the injection site, in one of the
patients receiving amiodarone.

   In an editorial by Karth24 that accompanies Siu’s article, the editorialist stresses that
these data, though valuable and convincing, were obtained in relatively healthy patients
and, because of this, the data might not be directly applicable to typical postoperative
patients since surgical patients are increasingly presenting with significant comorbid
conditions. Nonetheless, there is sufficient reason, based on the data reported by Siu, to
consider diltiazem as initial therapy in patients with acute, new-onset atrial fibrillation
when rhythm control strategies are not appropriate.
       In surgery patients, prevention of postoperative atrial fibrillation would be
   desirable if risk for the development of this arrhythmia could be quantified, and if safe,
   pharmacologic prevention strategies were available. A prevention strategy is discussed
   in a report by Zebis and coauthors25 in Annals of Thoracic Surgery, 2007. These authors
   report a randomized, placebo controlled, double blind trial comparing amiodarone with
   placebo in a group of patients undergoing coronary artery bypass (a known high-risk
   group for the development of postoperative atrial fibrillation). These authors noted a
   14% absolute risk reduction for patients treated prophylactically with amiodarone. Of
   the patients in the placebo group who developed atrial fibrillation, more than 80% were
   symptomatic; just over 40% of the patients in the amiodarone group who developed
   atrial fibrillation were symptomatic. While these data have limited application to typical
   general surgery patients, a preventive strategy might be considered in patients who
   have previously undergone cardioversion for atrial fibrillation if antiarrhythmia drugs
   are not already being used.

Management of surgical patients with disorders of the cardiac conduction system
       A single review article is discussed in this section of the overview by Allen26 from
   Anaesthesia in 2006. The article is entitled “Pacemakers and implantable cardioverter
   defibrillators” and a full-text reprint of this article is provided with this issue of SRGS.
   The author opens the discussion noting that pacemaker implantation is increasing with
   increasing age of the surgery patient population. Likewise, the number of implanted
   cardioverter defibrillators is increasing. Patients who have these devices are elderly
   with histories of significant symptomatic heart disease.

       The author notes that modern pacemakers work by delivering, via an intracardiac
   electrode, a low-voltage impulse to cardiac muscle. Devices in current use are capable of
   detecting the intrinsic electrical signals within the heart so that the devices deliver
   pacing impulses only when they are needed. Improvement in pacing lead design has led
   to “active fixation” leads that ensure optimum contact with the endocardial surface of
   the heart. These leads also are designed to elute steroid medications to minimize
   inflammation at the contact site.

       Battery life has improved so that battery replacement is only necessary once in each
   10-year interval. Furthermore, the titanium casing of modern pacemakers is light and
   protects the device from outside electromagnetic interference so that patients can
   safely use microwave ovens, electric shavers, and mobile telephones. In addition,
   modern devices carry electromagnetic interference detection software that offers
   additional protection. For patients who undergo surgical procedures, the most common
   form of electromagnetic interference comes from use of electrical coagulation devices.
   Bipolar diathermy is preferred when the patient has an implanted cardiac device. If
   monopolar diathermy use is unavoidable, the contact plate should be placed as far away
   from the pacemaker as possible. Advice from the clinician who implanted the
   pacemaker can be sought to reprogram the device if necessary. Reprogramming ideally
   occurs just before beginning the procedure. Ideally, the physician who inserted the
   pacemaker would remain in the area until the procedure is completed.

       While earlier devices paced the ventricle alone, current devices offer dual-channel
   pacing which improves cardiac output by taking advantage of atrial systolic contraction.
   Allen emphasizes data that have documented reductions in risk for mitral and tricuspid
   regurgitation and reductions in frequency of heart failure and chronic atrial fibrillation
   with dual-chamber pacing.

       Allen goes on to provide information on the various pacing modes of current
   pacemaker devices. More than three-quarters of currently used pacemakers are rate-
   sensing so that pacing current is supplied only when heart rates fall below a preselected
   level. More than half of currently implanted devices are dual-chamber pacing devices.
   Currently, rate-sensing pacemakers adjust current output based on surrogates for
   increased physical activity such as body movement and respiratory excursion. Ideally,
   rate-sensing devices would assess catecholamine levels or autonomic activity. Such
   sensors are under development but, as of 2006, were not available. Pacemaker rate
   sensors can sometimes interpret signals from intraoperative monitoring devices (such
   as respiratory rate monitors that determine thoracic impedance) as body movement.
   This results in rapid pacing.

       In patients with chronic heart failure, multiple sites within the cardiac chambers are
   paced; this is termed cardiac resynchronization therapy. In these devices, impulse
   delivery to both ventricles in multiple sites can be timed to maximize cardiac output.
   Implanted cardioverter defibrillators are equipped with complex algorithm software
   that tailors a response to a detected dangerous ventricular rhythm. Rate, beat-to-beat
   variation, atrial activity, and QRS morphology can be detected by the software and
   electrical shocks are delivered based on the rhythm detected. All implantable cardiac
   convertor defibrillators have pacemaker capability. These devices are not generally
   sensitive to external electromagnetic interference, but it will be wise to obtain advice
   from the clinician who implanted the device about any precautions anticipated during
   anesthesia and surgery.

Cardiac failure in the surgical patient
       Cardiac failure is an extremely common medical problem. More than 1 million
   hospitalizations annually in the United States are for cardiac failure; there is a 50%
   likelihood of death or recurrence of cardiac failure during the six months subsequent to
   a hospital admission. Cardiac failure will develop in up to one-third of patients with
   symptomatic ischemic cardiac disease; this condition will develop in 15% of diabetics
    and 10% of patients with hypertension. While it is unlikely that surgeons will be
    involved in the first-line management of patients with acutely decompensated cardiac
    failure, surgeons will be called to assist in the care of patients with heart failure who
    develop conditions requiring elective or urgent surgical conditions. It is important that
    surgeons understand the fundamentals of disordered cardiac function characteristic of
    the various forms of heart failure, and the pharmacology and side effects of the various
    therapies employed in these patients. This set of topics is reviewed in this section of the

Systolic cardiac failure
        The first article reviewed by Chatterjee and Rame27 appears in Critical Care
    Medicine, 2008, entitled “Systolic heart failure: chronic and acute syndromes.” The
    authors define systolic cardiac failure as inadequate function of the heart as a pump
    manifest by reduced ejection fraction. The condition most often emerges in patients
    with diabetes, hypertension, or ischemic heart disease. Systolic heart failure might also
    be encountered in patients with dilated cardiomyopathy from other conditions such as
    myocarditis. Systolic cardiac failure results from a process termed “ventricular
    remodelling.” The ventricles take on a more globular shape and chamber size increases.
    Although ventricular muscle mass increases, chamber size increases results in an
    increased chamber/ventricular wall ratio. The alteration in the chamber/ventricular
    wall ratio results in increased ventricular wall stress; the result of these changes is an
    increase in end diastolic and end systolic chamber volumes, resulting in diminished
    ejection fraction.

        Chatterjee and Rame emphasize the importance or neurohumoral activation as the
    process required for progression of systolic cardiac failure. Adrenergic, renin-
    angiotensin, and aldosterone systems are all activated and the degree of activation is
    linearly related to severity of symptoms and outcome. In the cardiac myocyte, results of
    neurohumoral activation are hypertrophy, apoptosis, necrosis, and fibrosis. There is
    evidence of increased oxidative stress that produces additional cytotoxicity. Increases
    in peripheral vascular resistance, ventricular filling pressures, and arterial stiffness are
    also results of neurohumoral activation, and these features contribute to cardiac failure

        Additional insight into the complex metabolic processes that influence the severity
    and progression of heart failure is from an article by Ashrafian and coauthors28 in
    Circulation, 2007. These authors open their discussion with a description of myocardial
    energy metabolism and the balances necessary for efficient energy use. They point out
    that daily myocardial ATP turnover is much greater than the myocardial ATP pool, and
    normal myocardial energy metabolism extracts only 25% of available substrate.
    Because of these facts, subtle changes in the efficiency of myocardial energy metabolism
have far-reaching implications for cellular energy levels. One of the most important
areas of study has been altered myocardial carbohydrate metabolism and the related
state of myocardial insulin resistance. At the cellular level, as insulin concentrations
vary, an attenuated glucose response results. These authors cite research data that
demonstrate a steadily increasing risk of heart failure with age in diabetic patients.
There is also an increasing risk of heart failure as hemoglobin A1c levels increase.
Persistent hyperglycemia predicts increased risk for the development of heart failure
and for heart-failure-related hospitalizations. They refer to additional evidence
supporting a linkage between myocardial insulin resistance and the subsequent
development of cardiac failure. Neurohumoral disorders characteristic of cardiac failure
also facilitate the development of hyperglycemia. Persistent inflammation,
demonstrable in patients with cardiac failure, contributes to hyperglycemia and
myocardial insulin resistance.

    The aggregate result of the metabolic dysfunctions noted in cardiac failure is a heart
that is energy deficient. Because the heart must produce ATP in amounts many times
the weight of the heart, energy deficiency becomes a major factor in the onset and
progression of heart failure. In addition, heart failure is associated with major
reductions (approximating 70%) in phosphocreatine, the “energy reserve” of the heart.

    Implications for management of cardiac failure emphasize control of the
neurohumoral dysfunction concurrently with optimization of glucose levels as a means
of combating insulin resistance. Ashrafian and associates discuss several new
pharmacologic agents that have the potential to improve myocardial energetics in
cardiac failure patients.

    Patients with Type 2 diabetes and the metabolic syndrome are at increased risk for
the development of cardiac failure. Management of this patient group is challenging
because the two mainstays of diabetic therapy for Type 2 diabetes, the biguanides
(metformin) and the thiazolidinediones (rosiglitazone) are currently contraindicated in
patients with clinical evidence of cardiac failure. Several classes of diabetic drugs are
available as adjuncts to conventional neurohumoral modulating agents in this patient
group. This topic is reviewed in detail in an article by Masoudi and Inzucchi.29
Interested readers are encouraged to review this article.

    Anemia is an additional condition frequently observed in patients with cardiac
failure. A discussion of this topic comes from an article by Mitchell30 in the American
Journal of Cardiology, 2007. The author notes that anemia is present, overall, in 33% of
heart failure patients and the proportion of patients who are judged to be anemic
(hemoglobin level < 12 gm/dL) increases with increasing severity of heart failure. New
York Heart Association Class IV patients have a 76% prevalence of anemia. The causes
of anemia are complex, with contributions from impaired erythropoietin synthesis and
utilization, hemodilution, impaired iron and vitamin B12 absorption, and persistent
gastrointestinal bleeding in patients who take aspirin.

    Anemia, like other disease features, is associated with increased levels of
proinflammatory mediators and oxidative stress factors. Anemia is known to be an
independent driver for increased rates of heart failure hospitalization and death.
Mitchell cites several confirming data sources. Mortality risk is particularly high when
anemia and renal insufficiency coexist. Because ischemic cardiac disease is an
important precursor of cardiac failure, assessment of this patient group for anemia has
been carried out by several investigators cited by Mitchell. Data disclose an association
of anemia with the onset and progression of ischemic cardiac disease. Anemia might
lead to increased cardiac output that contributes to imbalances of myocardial energy
availability/utilization that contribute to progression of cardiac failure.

    Mitchell notes that elevation of hemoglobin levels is associated with improved left
ventricular ejection fraction and improved quality of life indices. Elevation of
hemoglobin levels with erythropoietin analogues and iron is desirable. Red blood cell
transfusion has lowered short-term mortality in a small group of elderly patients but it
is not clear, according to this author, whether the benefit of transfusion outweighs the
risks. Additional information on this topic is in an article by Gerber31 in Critical Care
Medicine, 2008. The focus of this article is the use of transfusion in patients with
ischemic cardiac disease. It is likely, however, that many of the basic findings pertinent
to the ischemic cardiac disease patient will also be appropriately applied to patients
with cardiac failure.

    The author begins by reviewing the complications of transfusion with acute
complications such as transfusion reaction and transfusion-related lung injury (this
topic is discussed in more detail later in the overview), and the medium term
complication of blood-borne disease transmission. Because there are significant risks to
transfusion, the decision to use transfusion must depend on an assessment of the extent
to which oxygen availability to cells will be increased by raising the number of red
blood cells with transfused cells and documentation of improved outcomes in anemic
heart disease patients who receive transfusions. Gerber notes that the average storage
age of transfused red blood cells is 17 days. Currently, stored red cells have lost 2,3
diphosphoglycerate and the p50 of the cells has changed so that cellular affinity for
oxygen is increased and the ability to offload oxygen from transfused red cells to tissue
is reduced. Structural changes in red cells have also occurred and the cells have become
stiff so that passage into and through the microcirculation is impaired. Although
measured oxygen content of blood might increase following transfusion of stored red
cells, increased cellular oxygen availability is by no means assured.
       Data from studies of septic patients and patients in septic shock, cited by Gerber,
    suggest that cellular oxygen delivery is not increased by red blood cell transfusion.
    Gerber then reviews several studies where outcomes have been analyzed in anemic
    heart disease patients who have been transfused. Only one study has shown improved
    outcomes, and the improvement was observed only in elderly patients with admission
    hematocrits < 33%. In all the other studies there was no improvement, with several
    studies suggesting worse outcomes in transfused patients. He concludes by stressing
    that there is no convincing evidence to support the routine use of transfusion to
    improve outcomes in anemic patients with cardiac disease.

Diastolic cardiac failure
        Approximately 50% of patients with acute symptoms of cardiac failure, manifest as
    dyspnea with radiologic signs of pulmonary edema, have preserved left ventricular
    ejection fraction, according to data presented in an article by Kumar and coauthors32 in
    Critical Care Medicine, 2008. Patients with this form of cardiac failure are often elderly,
    female, and less likely to be African American than are patients with other forms of
    cardiac failure. The clinical presentation in many patients consists of signs of acute
    pulmonary edema associated with elevated systolic blood pressure. Because
    echocardiographic imaging that documents maintenance of left ventricular ejection
    fraction is performed, in many patients, after treatment for heart failure has begun, the
    suggestion has been made that ejection fractions were depressed at the time of
    symptom onset and improved with treatment. Kumar and associates cite a report of
    echocardiographic analyses performed in patients acutely and after 24 hours of
    treatment. There was maintenance of left ventricular ejection fraction at both time
    points, suggesting that heart failure occurred in the presence of normal left ventricular
    ejection fraction.

       These authors note that diastolic cardiac failure and pulmonary edema are likely
    caused when venous return to the right ventricle acutely increases and an increased
    volume of blood is delivered to the pulmonary circulation. Left ventricular dysfunction
    creates a situation in which the left ventricle cannot accept the increased blood flow
    without elevating left atrial pressure. In the setting of elevated left ventricular pressure
    (especially with a peak late in systole), left ventricular relaxation is impaired.
    Pulmonary blood volume increases and this overcomes the ability of the pulmonary
    lymphatics to remove fluid from the pulmonary interstitium. Pulmonary edema is the

       Additional data on diastolic cardiac function are found in an article entitled “Left
    ventricular diastolic function” by Hoit33 in Critical Care Medicine, 2007. Hoit notes that
    cardiac diastole is the result of processes whereby the heart loses ability to generate
    contractile force produced by myocyte shortening. The heart returns to a precontractile
   state in preparation for filling and the subsequent systole. Responsible for this series of
   events are myocardial relaxation and the pressure/volume properties of the ventricle.
   Relaxation is an energy-consuming process. Calcium is released from troponin C and
   actin-myosin cross bridges detach. Calcium is sequestered in the sarcoplasmic
   reticulum and, simultaneously, calcium is extruded from myocyte cytoplasm by active
   sodium-calcium exchange. Multiple factors influence the left ventricular end diastolic
   pressure-volume relationship including left ventricular physical properties (stiffness),
   the efficiency of relaxation, and extrinsic factors such as pericardial restraint and
   intrapleural pressure. Echocardiography is a valuable means for quantifying left
   ventricular diastolic function. Ventricular compliance and left atrial volume can be
   assessed with echocardiographic imaging. Using Doppler imaging, flow velocities across
   the mitral valve and in the pulmonary veins can be measured and the relaxation
   dynamics of the left ventricle can be determined.

       Acute echocardiographic imaging is emerging as an important tool permitting
   quantification of cardiac function and intravascular volume status in patients with
   suspected myocardial infarction, hypovolemia, or cardiac failure. Features of acute
   echocardiographic evaluation are reviewed in detail in an article by Glassberg and
   coauthors34 in Critical Care Medicine, 2008. These authors describe the use of Doppler
   echocardiographic imaging to assess preload and afterload. They note that recent data
   disclose an increased rate of cardiac adverse events in patients with acute cardiac
   decompensation monitored using pulmonary artery catheters. They further note that
   Doppler echocardiography has the capability of providing accurate estimates of cardiac
   output, right atrial pressure, pulmonary artery mean, systolic, and diastolic pressures
   as well as left ventricular filling pressure. Echocardiographic imaging might produce
   clinical information that is equivalent to the information gained from the pulmonary
   artery catheter without the risk of central venous catheterization. They conclude that
   acute echocardiographic imaging is an important component of the evaluation of
   patients with acute hemodynamic instability where cardiac failure is an important part
   of the differential diagnosis.

Management of heart failure in the surgical patient
       Where systolic or diastolic cardiac failure is suspected, the history, physical
   examination, and acute echocardiographic imaging are used to establish a diagnosis.
   Laboratory studies, including serum assays of brain natriuretic peptide (BNP) or N-
   terminal pro-brain natriuretic peptide (NT-proBNP) might be helpful in providing
   additional diagnostic information. The use of these serum markers is discussed in an
   article by Omland35 in Critical Care Medicine, 2008. Omland stresses the value of
   diagnostic information that can be gained from serum levels of BNP or NT-proBNP
   obtained in the emergency department or in the ICU when patients present with acute
   dyspnea. Abnormal BNP or NT-proBNP was 84%-90% accurate diagnosing diastolic
cardiac failure as the cause of acute dyspnea in several studies cited by this author.
Omland stresses that BNP levels are frequently normal in patients with chronic heart
failure. Furthermore, BNP and NT-proBNP levels were not consistently useful as means
of assessing progression or improvement of cardiac failure. Data discussed earlier
describe the limitations of serum tests in postoperative patients.

    Therapy for systolic cardiac failure depends on the clinical presentation. The
presence of echocardiographic evidence of increased filling pressures suggests the use
of loop diuretics (furosemide) to improve pulmonary congestion, dyspnea, and hypoxia.
Significant low cardiac output states in patients with systolic cardiac failure can be
treated with afterload reduction using vasodilators. Sublingual nitroglycerin is the first-
line approach in this regard. With very low cardiac output, short duration inotropic
therapy can be considered.

    The use of inotropic drugs for systolic cardiac failure is the topic of an article by
Petersen and Felker36 in Critical Care Medicine, 2008. These authors emphasize data,
which they review in this article, indicating a lack of clinical value of inotropic drugs in
patients without clearly documented end-organ hypoperfusion. They further report the
clinical challenges in documenting end-organ hypoperfusion. Traditionally, this
diagnosis has been made by documenting worsening renal function. Petersen and
Felker note that increases in serum creatinine after the institution of loop diuretic
therapy might indicate presence of cardiorenal syndrome and not end-organ
hypoperfusion. These authors note that some patients with very low cardiac output
states will maintain normal levels of serum creatinine.

    These patients will frequently have nonspecific symptoms such as abdominal pain,
nausea, fatigue, and diminished cognitive function. Documentation of low cardiac
output with echocardiography or pulmonary artery catheter monitoring will likely
provide confirmatory evidence. The authors note that documented low cardiac output
in patients with systolic heart failure is a marker for increased short-term mortality. If
inotropic therapy is contemplated, dobutamine and milrinone are the first-line drugs.
Both drugs produce improvements in cardiac output via augmentation of cellular cyclic
AMP. Milrinone has greater vasodilating function than dobutamine and might have
lower risk of inciting arrhythmias. Devices useful for supporting cardiac function
include the intraaortic balloon pump, left ventricular assist devices, and ultrafiltration
devices. These devices reliably support cardiac function until definitive therapies using
revascularization or transplantation can be organized and implemented. These devices
are discussed in a review by Kale and Fang37 in Critical Care Medicine, 2008.

   According to Kumar and coauthors,32 treatment of acute pulmonary edema, the
main clinical manifestation of diastolic cardiac failure, focuses on improving
   oxygenation and relieving patient symptoms. Noninvasive ventilation with continuous
   positive airway pressure is valuable for reversing hypoxia. Early administration of a
   loop diuretic along with intravenous β-blocking drugs will improve pulmonary
   congestion, lower blood pressure and heart rate, and relieve patient symptoms. These
   authors stress that diuretic-naïve patients might have a very brisk diuresis and,
   therefore, lower diuretic doses initially might provide a greater margin of safety.
   Morphine is helpful for relieving symptoms also. Afterload reduction with sublingual
   nitrate drugs is frequently helpful.

       Perioperative management of diastolic cardiac failure is discussed in a review by
   Pirrachio and coauthors38 in the British Journal of Anesthesia, 2007, who stress that the
   focus of perioperative management is to choose an anesthetic strategy that will not
   decrease left ventricular function. Intravenous agents such as propofol and most muscle
   relaxants do not affect left ventricular function. Volatile anesthetics such as sevoflurane
   and desflurane also do not change left ventricular function. These authors stress the
   importance of aggressively controlling the catecholamine response that accompanies
   operation and they recommend preoperative beta blockade supplemented by
   intravenous short-acting agents such as esmolol for management of hypertension and

Cardiopulmonary resuscitation
       Data on the frequency of out-of-hospital and in-hospital cardiac arrest appear in
   articles by Ramsay and Maxwell,39 Ali and Zafari,40 and Ehlenbach and coauthors.41 The
   articles by Ramsay and Maxwell and Ali and Zarari are supplied as full-text reprints
   with this issue of SRGS. These articles confirm that there are more than 400,000 sudden
   deaths annually ascribed to cardiac disease resulting in cardiac arrest. Ramsay and
   Maxwell cite data indicating that there are 165,000 witnessed episodes of out-of-
   hospital cardiac arrest in the United States each year. In-hospital cardiac arrest occurs
   at a rate of nearly three events/1000 admissions, according to data cited by Ehlenbach
   and coauthors. Cardiac arrest is the cause of 5.6% of all deaths annually in the United
   States, according to data cited in the article by Ali and Zafari. Despite the availability of
   effective methods of cardiopulmonary resuscitation, mortality for witnessed out-of-
   hospital and in-hospital cardiac arrest exceeds 80%. All the authors cited note the
   disappointing statistics indicating that nearly three-quarters of the patients who sustain
   witnessed cardiac arrest have no attempt at resuscitation made. In this section of the
   overview, we review several topics pertinent to effective management of witnessed out-
   of-hospital cardiac arrest and in-hospital cardiac arrest.

History of cardiopulmonary resuscitation
      Ramsay and Maxwell describe a short history of cardiopulmonary resuscitation in
   their article from The American Surgeon in 2009. The authors note that descriptions of
   mouth-to-mouth rescue breathing appear in the Old Testament. In the 14th century,
   rescue breaths were administered using bellows devices placed intranasally or through
   a reed inserted into the trachea via an anterior neck incision. During the 18th and 19th
   centuries, “humane societies” were formed in several European countries to foster the
   use of artificial respiration techniques for drowning victims. In studies on animals, John
   Hunter noted that cessation of breathing led to cardiac standstill and immediate
   resumption of breathing led to restoration of cardiac action.

       The use of electricity for defibrillation was championed by Wiggers who also
   supported the use of open cardiac massage. Open massage was used for resuscitation of
   intraoperative cardiac arrest by Beck of Johns Hopkins Medical School, and this method
   of resuscitation was the focus of his research from 1920–1937. Closed chest massage
   was developed at Johns Hopkins and described in a 1960 publication in the Journal of
   the American Medical Association by Kouwenhoven, Knickerbocker, and surgeon James
   Jude. Training in techniques of cardiopulmonary resuscitation for emergency medical
   services personnel and citizen responders was made simpler and more effective by the
   development of life-like mannequins for intubation and resuscitation by Safar and
   Laerdal. Currently, national standards for citizen, emergency medical services, and in-
   hospital cardiopulmonary resuscitation are promulgated by courses sponsored by the
   American Heart Association (Basic Cardiac Life Support and Advanced Cardiac Life

Current practice and outcomes for cardiopulmonary resuscitation
       Ali and Zafari40 note that sudden cardiac arrest is, in the main, caused by coronary
   artery disease. They cite data from autopsy studies indicating that more than 80% of
   nonsurvivors of cardiac arrest have severe coronary artery disease confirmed by post-
   mortem examination. Other causes of cardiac arrest are aortic stenosis, Wolf-
   Parkinson-White syndrome, cardiomyopathy, and congenital cardiac disease. The
   presence of a “shockable” (ventricular tachycardia or ventricular fibrillation) rhythm is
   associated with better outcomes of cardiopulmonary resuscitation. These authors note
   that these rhythms are being documented less often during cardiopulmonary
   resuscitation events. Fewer than one-third of patients have a shockable rhythm on
   initial electrocardiographic tracing. Asystole and pulseless electrical activity rhythms
   are being recorded with increasing frequency.

       These authors describe a “four phase” classification of a cardiac arrest event. The
   “electrical phase” extends from time 0–4 minutes after arrest. The “circulatory phase”
   extends from 4-10 minutes post arrest. The “metabolic phase” begins at 10 minutes
   post arrest. During the electrical phase, defibrillation is the most effective therapy if a
   shockable rhythm is noted. During the circulatory phase, qood-quality cardiac
   compression is critical. In the metabolic phase, resuscitative efforts focus on reversing
the effects of global ischemia. The importance of defibrillation during the electrical
phase supports the distribution of automatic defibrillators and the use of these devices
by trained citizen rescuers since it is unlikely that trained emergency medical services
personnel will arrive on the scene before the late circulatory or metabolic phase of

    Data cited in Table 2 of the article by Ali and Zafari confirm the value of early
defibrillation if a shockable rhythm is discovered within the first five minutes following
the arrest event.

    Adequate cardiac compressions (optimum rate with optimum excursion) given
before defibrillation shock are associated with improved outcomes, according to data
cited by Ali and Zafari. They also note that optimum cardiac compressions provide
coronary perfusion that serves to minimize the depleting affect of ventricular
fibrillation on cardiomyocyte energy stores. Rescue breaths (two breaths administered
by mouth-to-mouth or mouth-to-airway respiration before instituting cardiac
compression) are currently recommended by the American Heart Association, but this
is controversial and subject to change.

    Ramsay and Maxwell39 note that current recommendations urge rescuers to
perform chest compressions with two hands in adults at a rate of 100
compressions/minute with a compression excursion of 4 cm. For patients who have an
airway placed, the ratio of compressions/breaths is recommended at 30:2. Mouth-to-
mouth and mouth-to-airway “rescue breaths” previously recommended to precede
chest compressions are now eliminated in many regional protocols recognizing that
encouragement to administer mouth-to-mouth breaths is a strong disincentive to
provision of any sort of rescue resuscitation. As noted above, in current studies, no
resuscitation attempt is made in the majority of witnessed out-of-hospital cardiac
arrests. Recent data cited by Ramsay and Maxwell (reference 9 in their bibliography)
indicate that chest compressions without rescue breaths result in improved outcomes
for cardiopulmonary resuscitation in witnessed out-of-hospital cardiac arrest events.

    A more favorable neurologic outcome, more frequent occurrence of shockable
cardiac rhythm on initial electrocardiogram, and improved overall survival when
resuscitation was begun within four minutes of cardiac arrest were all confirmed in the
study cited. These observations have lent support to the primacy of supplying effective
chest compressions. Rescue personnel are no longer encouraged to supply a “stack” of
three electrical defibrillation shocks when a shockable rhythm (ventricular tachycardia
or ventricular fibrillation) is discovered on the initial electrocardiogram. Instead, a
single shock is applied and compressions are resumed. In addition, drug administration
(intravenous or endotracheal instillation) is recommended while compressions
continue rather than a “drug-breath-shock-compression” cycle. Despite dissemination
of this information nationally, data cited by Ramsay and Maxwell indicate an
unsatisfactory level of compliance with these guidelines. In a study of in-hospital
cardiac arrest, compression rates of less than 100/min were noted in more than 90% of
resuscitations (reference 10 in their bibliography). An additional finding of this study
was a disturbing frequency of “no-flow” intervals (intervals during which there are no
compressions). These exceeded 10 seconds/minute of resuscitation events.

   Research confirming the critical importance of high-quality chest compressions is
from a report by Ristagno and coauthors42 in Chest, 2007. These authors performed a
study in pigs; cardiac arrest was produced by ligation of the left anterior descending
coronary artery. Chest compressions were begun 5 minutes after onset of cardiac
arrest. “Adequate” chest compression was defined as compression excursion equal to
25% of the anterior-posterior chest diameter (6 cm excursion) and “conventional” chest
compression as 4 cm excursion. A single defibrillation shock of 150 joules was delivered
before or after compressions began, according to the experimental protocol for each
animal group; each group consisted of six animals. The data presented indicate that
end-tidal PCO2 and coronary perfusion pressures (both variables are related to
adequate myocardial and peripheral perfusion) were lower in animals receiving
conventional compressions. With optimal chest compressions, fewer shocks were
required to restore cardiac rhythm and all animals were resuscitated.

    No animals were resuscitated when defibrillation shock preceded conventional
compressions. Two of six animals were resuscitated successfully when shocks were
administered after a period of conventional chest compression. This animal study
supports current clinical recommendations that stress the importance of adequate
chest compressions. Adequacy of chest compression is important whether a “shock
first” or “shock after compression” protocol is followed. These data lend support to the
urgent need to improve the quality of compressions offered to patients who sustain
cardiac arrest. These authors cite data that indicate adequacy of chest compressions in
less than one-third of cardiopulmonary resuscitation events, and they stress the
importance of aggressive educational efforts to improve the quality of chest
compressions in cardiopulmonary resuscitation events.

    Ramsay and Maxwell go on to discuss methods for ventilating the intubated patient
during a cardiac arrest event. The foregoing discussion has noted the recommendation
of a breath-to-compression ratio of 30:2. Hyperventilation raises intrathoracic
pressures with resultant reduction in venous return and depression of compression
mediated cardiac output. With increased intrathoracic pressure, coronary perfusion is
reduced. If hypocarbia is induced, cerebral vasoconstriction might reduce cerebral
oxygenation. Placement of an impedance threshold device in the ventilation circuit
serves to reduce intrathoracic pressure and improve venous return. This device works
by reducing air entry into the lung during the chest recoil phase that follows a cardiac
compression. Lowered intrathoracic pressures, improved end-tidal PCO2, and improved
coronary perfusion pressures have been documented with the use of this device during
experimental and clinical cardiopulmonary resuscitation.

    A meta-analysis of research results relevant to the use of the impedance threshold
device is reported by Cabrini and coauthors43 in Critical Care Medicine, 2008. These
authors reviewed data from five high-quality studies involving more than 800 patients.
They found that use of the impedance threshold device was associated with significant
improvements in return of effective cardiac rhythm, early survival, and early favorable
neurologic outcomes with no significant effect noted on long-term survival. The authors
note that this study focused on data generated from studies of out-of-hospital cardiac
arrest. They further note that optimum results in terms of improved venous return to
the heart depend on rescuers permitting full recoil of the chest wall by lifting the palms
off the chest after each compression. They stress that data they reviewed showed
improved outcomes even in patients with event-to-compression times of 10 minutes or
more. Improvements are also noted in outcomes of patients with unfavorable rhythms
such as asystole.

    The use of adjunctive drug therapy during cardiopulmonary resuscitation is
reviewed by Ali and Zafari.40 The objective of initial drug therapy is to provide strong α-
receptor agonist capability. Alpha stimulation works to redistribute blood flow to the
brain. Epinephrine 1mg intravenously or instilled into the endotracheal tube has been
the initial drug for a number of years. Recently, additional pressor activity has been
achieved with vasopressin in a dose of 40 international units given intravenously. These
authors cite data from studies comparing epinephrine and vasopressin; no difference in
survival-to-hospital discharge was noted in either group. A subsequent study showed
improved survival in patients receiving both epinephrine and vasopressin. Current
recommendations support use of both drugs in combination.

    Although epinephrine has been used for many years to improve brain blood flow,
research data questions whether this drug results in true increases in brain cellular
oxygen delivery. This issue is the topic of a research study by Ristagno and coauthors44
in Critical Care Medicine, 2009. These authors conducted experiments on pigs and note
that the sine qua non of successful cardiopulmonary resuscitation is successful recovery
of brain function. The alpha-receptor stimulating properties of epinephrine have been
thought to facilitate redistribution of blood flow to the brain during cardiopulmonary
resuscitation. These authors cite evidence for beta-receptor stimulating activity of
norepinephrine during experimental shock. This property resulted in failure of the drug
to restore appropriate nutrient tissue blood flow.
    In these experiments, the authors assessed brain nutrient delivery by measuring
microcirculatory blood flow using orthogonal polarization spectral imaging. Tissue
carbon dioxide and oxygen tensions were also measured. Ventricular fibrillation was
induced and drug therapy administered after three minutes. Animals were divided into
groups that received placebo, epinephrine, and epinephrine in the presence of two
different forms of alpha and beta receptor blockade. The data disclosed reductions in
both brain oxygen tension and microcirculatory blood flow in animals receiving
epinephrine. This was accompanied by increased brain tissue carbon dioxide levels. The
effect was traced to the alpha stimulatory effects of epinephrine.

    Drug therapy has also been suggested to facilitate the return of effective cardiac
rhythm. This topic is also reviewed by Ali and Zafari. Drugs used for this purpose
include lidocaine, atropine, and amiodarone. Studies of atropine use have not
demonstrated statistically improved survival although there might be improved
conversion of slow pulseless electrical activity with atropine use. Lidocaine is currently
not recommended as a means of improving conversion from ventricular tachycardia
and fibrillation to an effective rhythm. Amiodarone is the current drug of choice to
assist with rhythm conversion. Concern that the diluent in which amiodarone is
delivered (polysorbate 80 and benzoyl alcohol) might be a cause of hypotension has
been voiced but data, to date, do not support this as a frequent complication. There is
now an aqueous preparation of amiodarone that can be infused rapidly and this
preparation has not shown increased frequencies of hypotension. The current
recommended dose of amiodarone is 300 mg. Another drug occasionally used in
cardiopulmonary resuscitation is magnesium sulfate. This drug is used in 1-2 gram
doses for one specific rhythm, torsade de pointes. This rhythm is a polymorphic
ventricular tachycardia associated with a prolonged QT interval.

   Recovery of effective cardiac activity after cardiopulmonary arrest is associated
with a pro-inflammatory state characterized by elaboration of inflammatory cytokines,
changes in coagulation, and increased oxidative stress. Because of this observation and
because adrenal insufficiency has also been observed after cardiopulmonary arrest,
anti-inflammatory therapy has been suggested as a means of improving outcomes. This
topic is addressed by Mentzelopoulos and coauthors45 in an article entitled
“Vasopressin, epinephrine, and corticosteroids for in-hospital cardiac arrest.” These
authors begin by citing animal experiment data supporting the use of vasopressin,
epinephrine, and corticosteroids as a means of improving neurologic outcomes of
cardiac arrest. These authors conducted a single-center, randomized, prospective,
double-blind study to assess effects of this drug combination with the use of stress dose
corticosteroid replacement in patients found to have adrenal insufficiency on the
outcomes of cardiopulmonary resuscitation. The authors noted improved recovery of
effective cardiac rhythm and improved survival-to-hospital discharge in patients who
received the corticosteroid drug combination. Two patients in the corticosteroid group
survived to hospital discharge versus none in the control group. Unfortunately, both
survivors had severe neurologic deficits. Although this study is suggestive of a benefit in
a very high-risk group of patients, the small numbers and the questionable clinical
significance of two neurologically disabled survivors versus none indicates that these
results should be interpreted very cautiously.

    Optimization of recovery of effective cardiac rhythm, myocardial perfusion, and
cerebral perfusion are efforts directed toward the electrical and circulatory phases of
cardiac arrest. The main attempt to ameliorate the effects of the metabolic phase of
cardiopulmonary resuscitation has emphasized the use of therapeutic hypothermia.
Ramsay and Maxwell39 note experimental data indicating a reduction in brain oxidative
stress and oxygen demand when post-cardiac arrest hypothermia is used. These
authors review data from the Hypothermia After Cardiac Arrest Group. This group
conducted a randomized trial of hypothermia (bladder temperature 32-34°C for 24
hours) following successful resuscitation in patients with ventricular fibrillation. There
were significantly improved neurologic outcomes (55% good outcomes versus 39% in
controls) when hypothermia was used. Six-month mortality was 41% in the
hypothermia patients versus 55% in controls. These data support use of post-
resuscitation hypothermia in patients with witnessed out-of-hospital cardiac arrest
where successful defibrillation occurs.

   The topic of hypothermia application is discussed in more detail in two additional
reviews and an editorial which accompanied one the reviews.46-48 One of these articles,
by Schneider and coauthors,48 is a detailed, yet readable comprehensive review of
metabolic approaches to cardiac arrest and this article is included as a full-text reprint
with this issue of SRGS. These articles stress the importance of maximizing nutrient
blood flow by optimizing compressions, improving venous return, and protecting
against hyperventilation and lung hyperinflation. They also note the utility of mild post-
arrest hypothermia in patients who have witnessed arrests, prompt resuscitation, and a
shockable rhythm. Hypothermia is designed to protect cerebral tissue metabolism.

   The articles and the editorial stress the cellular causes of cerebral dysfunction,
which result from massive accumulations of calcium in brain cells after cardiac arrest.
Currently, hypothermia is indicated for children remaining comatose following cardiac
arrest and adults with recovery of spontaneous circulation after out-of-hospital
ventricular fibrillation. Hypothermia can potentially assist recovery of patients with
other rhythms, but data confirming this benefit are not available. Data are badly needed
because ventricular fibrillation is declining as the main rhythm for out-of-hospital
cardiac arrest. For in-hospital cardiac arrest, ventricular fibrillation is not often
   discovered on the initial electrocardiogram. The authors of these three articles note
   that the multiple comorbid conditions found in victims of in-hospital cardiac arrest
   makes recovery in these patients less likely. These reports note the potential for agents
   such as growth factors and apoptosis inhibitors; data about these approaches should be
   forthcoming in the future.

Another approach to applying hypothermia and circulatory support for victims of cardiac
arrest is with the use of extracorporeal membrane oxygenation. This topic is the subject of
a report by Thiagarajan and coauthors 49 in Annals of Thoracic Surgery, 2009, of an analysis
of a large extracorporeal membrane oxygenation database. Eleven percent of patients in
the database had the device applied as an adjunct to management of cardiac arrest. The
most frequent diagnosis recorded was “cardiac disease.” The authors documented a 27%
survival in these patients. The proportion of in-hospital versus out-of hospital cardiac
arrests is not provided, and the presence of a shockable rhythm is likewise unknown. These
data suggest potential utility of the extracorporeal membrane oxygenator in patients
sustaining cardiac arrest. Improved outcomes occurred when the device was applied
within two hours of arrest. A diagnosis other than myocarditis was also associated with
improved survival. Renal insufficiency requiring dialysis was associated with increased
mortality risk.

Cardiopulmonary resuscitation for in-hospital cardiac arrest: epidemiology and outcomes
       As has been noted previously, outcomes for in-hospital cardiac arrest are, in general,
   worse than those for out-of-hospital cardiac arrest, primarily due to the multiplicity of
   unfavorable risk factors in patients who sustain in-hospital cardiac arrest. In addition,
   the presenting rhythms for patients with in-hospital cardiac arrest are more often
   pulseless electrical activity or asystole. The first article examining the epidemiology of
   in-hospital cardiac arrest in elderly patients is by Ehlenbach and coauthors 41 in the
   New England Journal of Medicine, 2009. These authors analyzed data from the Medicare
   database over the interval 1992-2005. They examined data from records of 434,000
   patients who underwent in-hospital cardiopulmonary resuscitation. The overall
   survival of these patients was 18%, which is, interestingly, very close to the overall
   survival for out-of-hospital cardiac arrest.

       Survival of nonwhite patients was worse than survival in white patients. The
   authors note that the data for nonwhite patients were drawn, disproportionately, from
   hospitals with lower overall cardiac arrest survival and this might explain the racial
   discrepancy, at least in part. The analysis indicated that results of in-hospital
   resuscitation did not improve during the study interval. In the discussion section of this
   report, the authors make several interesting observations and speculations. They note,
   for example, that between-hospital comparisons suggest that hospitals that make use of
more extensive monitoring seem to have faster responses to cardiac arrest and
improved outcomes.

    One approach to shortening response time for cardiac arrest has been to institute
rapid response teams. This topic is the subject of a report by Chan and coauthors50 in
the Journal of the American Medical Association, 2008. These authors compared the
rates of cardiorespiratory arrest codes and mortality for cardiopulmonary resuscitation
before and after implementation of a rapid response team. Nearly 25,000 patients were
available for analysis in each group. The analysis was performed prospectively in a
single institution. The results of the analysis indicated that there were 376 rapid
response team activations in the 20 months of experience analyzed after team
implementation. The frequency of out-of-ICU codes declined (this is a hoped-for result
of rapid response team implementation). Unfortunately, there was no reduction in the
overall mortality when the two periods were compared.

    The authors note that there is no standard approach to rapid response team
activation and no standard organization of the team. These problems make
comparisons difficult. It is interesting to note recent developments in several hospitals
where rapid response team activation can be accomplished by patients or patients’
families without the need to summon nursing staff. Additional data will be interesting
to document the potential effectiveness of rapid response teams.

   Ehlenbach and coauthors note that their data indicate that the proportion of in-
hospital deaths preceded by formal efforts at cardiopulmonary resuscitation increased
during each year of the study they conducted. They speculate that this increase might
be because nonwhite patients tend to have lower rates of do-not-resuscitate orders
than do white patients. They further speculate that do-not-resuscitate orders might be
ignored in a significant proportion of in-hospital cardiac arrest incidents.

    Prearrest predictors of survival after cardiopulmonary resuscitation have been
sought. This subject is the topic of an analysis authored by Gonzalez and coauthors51 in
Circulation, 2008, who examined outcomes of cardiopulmonary resuscitation in a group
of patients who had undergone echocardiographic assessment of left ventricular
function on average 11 days before the arrest event. The authors found that a left
ventricular ejection fraction of less than 45% predicted a worse outcome. Mortality in
the group with diminished left ventricular function was 92% compared with 81% in
patients with normal left ventricular function.

    The analysis also produced data showing that post-cardiopulmonary resuscitation
of left ventricular function was depressed in all patients by an average of 25%. The
authors concluded that the depression of ventricular function attendant to
cardiopulmonary resuscitation could not be tolerated in patients with depressed
ventricular function before the cardiac arrest event. This analysis is interesting. The
high proportion of echocardiographic analyses before arrest (77%) suggests that these
patients were largely patients hospitalized with cardiac disease and generalization of
these data to other patient groups might not be possible. The authors speculate that the
hyperadrenergic state that follows restoration of cardiac rhythm might be especially
stressful for patients with diminished left ventricular function.

    Ehlenbach and associates also noted differences in outcomes of resuscitation
depending on the day and time of cardiac arrest occurrence. This topic is the focus of a
report by Peberdy and coauthors52 in the Journal of the American Medical Association,
2008. These authors analyzed data from cardiac arrest events occurring from 0700-
2259 versus 2300-0700. They also compared mortalities from cardiac arrests occurring
from 2300 on Friday to 0659 on Monday. The database analyzed was the National
Registry of Cardiopulmonary Resuscitation that contains data from more than 500
participating hospitals. They noted significant reductions in survival-to-discharge and
good neurologic outcomes in patients who sustained cardiac arrest at night or on
weekends. This difference was stable when the data were adjusted for differences in
patient comorbidity and risk. The data suggested a somewhat increased odds of death
for cardiac arrest occurring in the operating room or ICU and a slight (nonsignificant)
reduction of risk for cardiac arrest occurring in the emergency department. Patients
admitted after injuries were, as a group, slightly more likely to survive cardiac arrest.

    Data such as those discussed above suggest the need to improve effective
cardiopulmonary resuscitation in hospitals. Educational activities such as performance
debriefings and feedback might function to improve cardiopulmonary resuscitation
performance and, perhaps, results. The first article that deals with this topic is by
Edelson and coauthors53 and is entitled “Improving in-hospital cardiac arrest process
and outcomes with performance debriefing.” This article appeared in Archives of
Internal Medicine, 2008. A debriefing conference was held for residents who
participated in cardiopulmonary resuscitation events. Data from sensing defibrillators
and observations made by trained observers were used to identify areas for
improvement and training. One hundred twenty-three patient events after
implementation of the debriefing session were compared with 101 historic controls.
These authors found that adherence to national recommendations about compression
rate, excursion, and ventilation improved in the post-debriefing period. There was an
increased rate of return of spontaneous circulation in the post-debriefing patients but
there was no improvement in survival-to-discharge. Another method of performance
assessment and improvement includes the use of real-time audiovisual feedback
combined with debriefing. This approach is analyzed in a study by Dine and coauthors54
in Critical Care Medicine, 2008. These authors compared debriefing alone with
   debriefing with real-time audiovisual feedback in two groups of nurses undergoing
   training using a cardiopulmonary resuscitation simulator. The authors found that
   performance, evidenced by compliance with recommendations for compression rate
   and excursion, improved with debriefing alone. The addition of audiovisual feedback
   provided significant additional improvement. Overall, twice as many participants
   provided optimum compressions after debriefing with audiovisual feedback as with
   debriefing alone.

Editorial comment
       From the perspective of the editor, it seems clear that we have the capability to
   predict perioperative cardiac complications using global risk factors that overcome, at
   least partially, the imprecision associated with the use of risk scoring systems that focus
   on risk factors specific to the cardiovascular system. Once high-risk patients or patients
   at moderate risk who are scheduled to undergo high-risk operations (open vascular
   reconstruction or thoracotomy) are identified, careful preoperative preparation using
   beta-blocking drugs (selective beta-1 receptor blocking drugs) will provide control of
   heart rate and blood pressure in approximately 80% of this patient group. Preparation
   will need to begin at least one month preoperatively. If patients have indications for
   statin drugs based on lipid profiles or risk scores for cardiac disease (such as the
   Framingham score), statin drugs and low-dose aspirin are potentially useful additions.
   Ideally, these measures will have been implemented by the patient’s primary care
   physician. Surgeons need to insure that there is no interruption of drug therapy during
   the perioperative interval.

       Perioperative myocardial infarction is a potentially lethal complication occurring as
   a result of coronary artery plaque instability or rupture with coronary artery
   thrombosis. Prediction of which plaque will rupture is not possible currently. Because
   of this, planning preoperative revascularization interventions based on identification of
   a “culprit” coronary stenosis does not reliably reduce risk of perioperative myocardial
   infarction. Indications for preoperative coronary imaging and revascularization are
   made based on conventional indications, and these are undertaken in patients with
   “unstable” ischemic diseases such as unstable angina, recent myocardial infarction, and
   cardiac failure. The usual diagnostic clues for diagnosis of myocardial infarction (chest
   pain, Q-waves or ST segment elevation on electrocardiogram, and elevated troponin
   levels) lack specificity in the patient who has recently undergone a surgical procedure.
   Clinical signs of perioperative myocardial infarction might be vague and include
   intermittent hypotension, changing mental status, new onset arrhythmia, and ST-
   segment depression on electrocardiographic tracings. Because of these facts, a low
   threshold for use of serial troponin levels, continuous electrocardiographic monitoring,
   and echocardiographic imaging are necessary to make a prompt diagnosis.
       Perioperative tachycardia will occasionally require pharmacologic intervention or
   even electrical cardioversion. Knowledge of the elements of diagnosis and emergency
   treatment of these arrhythmias will be valuable. Even though surgeons will not
   normally be the lead caregiver for patients with cardiac failure, it is useful to
   understand the pathophysiology of this condition so that factors that increase cardiac
   stress can be minimized during the perioperative interval. Echocardiography is the
   most useful modality for quantification of the severity of cardiac failure.

       Finally, the features of successful resuscitation of patients who sustain out-of-
   hospital or in-hospital cardiac arrest are important components of the knowledge base
   of surgeons. Resuscitation maneuvers such as chest compressions and ventilation
   maneuvers are frequently not performed in compliance with recommendations from
   national groups like the American Heart Association. It is important to recall that
   maneuvers to provide effective chest compression and optimum venous return to the
   heart are critical features leading to successful resuscitation.

       Surgeons will be consulted to assist in the management of injuries sustained during
   cardiopulmonary resuscitation. Injuries from cardiopulmonary resuscitation are
   relatively common; clinically significant injuries are discovered in 10%-15% of
   autopsied patients. Injuries might be discovered in a larger proportion of survivors. Rib
   fracture and/or costochondral separation are the most commonly diagnosed injuries.
   Pneumothorax, hemothorax, diaphragm injury, and lacerations of the liver and spleen
   are occasionally encountered.

Perioperative respiratory complications
        Respiratory complications after major surgical procedures might range from minor
   complications like microatelectasis that can be cleared with coughing, deep breathing,
   and early ambulation, to major, life threatening events such as postoperative
   respiratory failure. Risk of major respiratory failure requiring ventilatory intervention
   is increasing as the surgical patient population ages and the frequency of preexisting
   respiratory diseases, such as obstructive sleep apnea, is increasingly recognized. The
   pro-inflammatory state stimulated by anesthesia, operation, and transfusion leads to
   acute lung injury that often progresses to acute respiratory distress syndrome. In order
   to minimize the negative impact of postoperative respiratory complications, surgeons
   require knowledge of the pathophysiology, effective preventive measures, features of
   diagnosis, and therapies available. These topics will be addressed in this section of the

Risk factors for respiratory complications
      Discussion of risk factors for development of postoperative respiratory
   complications and postoperative respiratory failure opens with a review of an article by
Johnson and coauthors55 entitled “Multivariable predictors of postoperative respiratory
failure after general and vascular surgery: results from the patient safety in surgery
study.” This article was published in the Journal of the American College of Surgeons in
2007 and a full-text reprint accompanies this issue of SRGS. This article is one part of a
series of reports detailing the results of the patient safety in surgery study that was the
precursor of NSQIP. Another component of this series was cited previously in the
overview in the discussion of risk factors for postoperative cardiac events.

    These authors analyzed demographic and outcomes data from 128 Veterans
Administration hospitals and 14 academic medical centers. Postoperative respiratory
failure was defined as the need for unplanned postoperative intubation and/or more
than 48 hours of mechanical ventilation assistance postoperatively. In contrast to the
previously discussed analysis of postoperative cardiac complications where global risk
factors markedly reduced the impact of cardiac-specific risk factors, this analysis found
that postoperative respiratory failure was best predicted using a combination of global
and lung-specific risk factors. More than 180,000 patients were analyzed. Respiratory
failure occurred in 3% of this group. Global risk factors such as higher ASA score, the
need for emergency operation, more complex procedures, preoperative sepsis, and
signs of renal insufficiency all predicted postoperative respiratory failure. Patients who
developed postoperative respiratory failure tended to be male and older.

    Smoking, a diagnosis of chronic obstructive pulmonary disease (COPD), and a
history of congestive heart failure were risk factors specific to the cardiopulmonary
system that also predicted postoperative respiratory failure. The authors developed a
“respiratory risk index” by assigning points based on rounding of the calculated odds
ratio for respiratory failure determined from the risk analysis. For example, a calculated
odds ratio of 1.25 would add one point to the respiratory risk index. Patients were then
divided into risk groups of low, medium, and high risk based on the calculated
probability of respiratory failure. The risk scoring system was validated in a separate
sample drawn from the database. The data disclose that respiratory risk index of eight
or lower is associated with a 0.1% risk of respiratory failure. Risk index score of 8-12 is
associated with a 1% incidence of respiratory failure. For scores >12, overall risk is 7%
but the risk steadily increases with increasing risk score; a score of 25 predicts a
frequency of respiratory failure of 40%.

    The authors use the discussion section of the article to confirm that respiratory
failure is associated with more healthcare resource utilization than any other group of
postoperative complications. They further note that other analyses have identified
similar arrays of risk factors, both global and lung specific, that predict postoperative
respiratory failure. Not surprisingly, the risk of respiratory failure increases with a
thoracic incision. After sternotomy or thoracotomy for cardiac procedures, the overall
frequency of respiratory failure was 7% compared with 3% associated with general and
vascular procedures in the analysis presented by Johnson and colleagues.

    Obstructive sleep apnea is being diagnosed with increasing frequency. The presence
of this disorder in obese patients and patients with the “metabolic syndrome” of
obesity, hypertension, and hyperglycemia is firmly established. The diagnosis of
obstructive sleep apnea is predictive of postoperative respiratory complications. The
next article reviewed investigates the possibility that preoperative testing for
obstructive sleep apnea might identify patients at increased risk for episodes of
postoperative hypoxemia. The article is by Gali and coauthors56 and appeared in
Anesthesiology, 2009. The authors cite data indicating a substantial rate of
underdiagnosis of obstructive sleep apnea. They refer to 1993 reports (references 6
and 7 in their bibliography) that estimated that 4% of men and 2% of women in the 30-
60 year age group had obstructive sleep apnea and that this condition was an
independent risk factor for postoperative mortality.

   By the end of the 1990s, there had been a 12-fold increase in the diagnosis of
obstructive sleep apnea. Later estimates have concluded that 82% of men and 93% of
women with obstructive sleep apnea remain undiagnosed. It is likely that many of these
individuals will require surgical care and that this group will be at increased risk for
perioperative respiratory complications. The authors describe the pathophysiology of
respiratory complications in patients with obstructive sleep apnea. The anatomic and
physiologic abnormalities of obstructive sleep apnea can be brought on by the
diminished responses to hypoxia and hypercapnia as well as the diminished pharyngeal
tone produced by anesthetic and analgesic medications.

    In this report, the authors hypothesize that a preoperative risk assessment for
obstructive sleep apnea applied in patients not known to have obstructive sleep apnea,
coupled with post-anesthesia monitoring for hypoxemic events, will identify patients at
risk and prevent complications. The preoperative assessment used consisted of
obtaining a sleep apnea clinical score. This score assigns points based on responses to
questions about the presence of hypertension and a history that patients had been told
by persons sharing their sleeping area that they snore. For this question, 1 point is
assigned for snoring 3-5 times/week or for snoring every night. The patients are also
asked whether they have been told that they gasp, choke, or snort while sleeping. The
point assignments are based on frequency of symptoms.

    The final assessment is a measurement of neck circumference. Points are assigned
for various neck circumferences with hypertension, historic features, or both. A score of
>15 indicates a high likelihood of obstructive sleep apnea. Postoperative monitoring of
patients enrolled in this study included continuous oxygen saturation monitoring, and
monitoring for apnea, bradypnea, and pain level/sedation mismatch. The last
assessment is accomplished when a patient indicates severe pain on a visual-analog
scale but appears too sedated to receive additional analgesia.

    In all, 673 patients were enrolled. Sleep apnea scores of > 15 predicted episodes of
desaturation and recurrent potential hypoxemic events in the post-anesthesia care
area. The combination of a high sleep apnea score and post-anesthesia area hypoxemic
events predicted postoperative respiratory complications. A high sleep apnea clinical
score was recorded in nearly 32% of this patient group. This patient group had higher
ASA scores also. Patients with high clinical scores and recurrent hypoxemic events in
the post-anesthesia care area had a frequency of diagnosed postoperative respiratory
complications of 33%. Patients with low scores and recurrent post-anesthesia events
had a frequency of postoperative respiratory complications of 11%. Patients with low
scores and no events developed postoperative respiratory complications in less than
1% of patients. The authors note that the gold standard for diagnosis of obstructive
sleep apnea is polysomnography.

   Validation of the sleep apnea clinical score with comparisons to other scores and
polysomnography are found in two articles57-58 conducted in 1994 and 2003 confirming
that the sleep apnea clinical score has a positive predictive value for an accurate
diagnosis of obstructive sleep apnea of more than 80% compared with
polysomnography. A limitation of the study authored by Gali and associates is that
polysomnography was not used to validate the findings reported. Nonetheless, the data
suggest that obstructive sleep apnea might be underdiagnosed. Furthermore, sleep
apnea clinical scores of > 15 are, when combined with assessments performed in the
post-anesthesia care area, predictive of postoperative respiratory complications.
Finally, this straightforward assessment can be used to identify patients at risk.

    Because obstructive sleep apnea is a common condition accompanying morbid
obesity, identification of patients with this condition who are scheduled to undergo
bariatric procedures could lead to preventive interventions, reduction of risk for
postoperative pulmonary failure, and lower consumption of healthcare resources. An
article dealing with this topic by Hallowell and coauthors 59 appeared in Surgery, 2007.
This analysis compares the need for ICU admission for respiratory complications in a
group of 318 morbidly obese patients undergoing bariatric procedures.
Polysomnography for diagnosis of obstructive sleep apnea was performed in all the
patients. The frequency of ICU admission in this group was compared with a historic
control group. In the control group, obstructive sleep apnea assessment using
polysomnography was used based on clinical suspicion and/or surgeon preference.
      After implementation of routine polysomnography, ICU admission for respiratory
   complications decreased from 5% to less than 1%. The authors note other changes in
   their practice that occurred simultaneously. They note an increasing use of laparoscopic
   gastric bypass in the second group. They also note that the rate of ICU admission for any
   reason was already declining in their practice. In spite of these limitations, the analysis
   does disclose a significant risk for undiagnosed obstructive sleep apnea in morbidly
   obese patients. Improving the rate of diagnosis of obstructive sleep apnea in this patient
   group could create an opportunity for implementing preventive measures to reduce
   perioperative respiratory complications.

       In the discussion section of the article, Hallowell and colleagues review the changes
   occurring because of increasing body weight that contribute to respiratory
   complications. These include upward displacement of the diaphragm, increased chest
   wall mass leading to decreased chest wall compliance, and increased pulmonary
   vascular resistance resulting from chronically expanded blood volume. These authors
   also stress that obstructive sleep apnea might be occult in this patient group. The use of
   preoperative and postoperative noninvasive ventilation with continuous positive
   airway pressure is one effective preventive measure. The authors note that concerns
   over application of continuous positive airway pressure in a patient with a newly
   constructed gastrojejunostomy seem unfounded. They cite data from clinical reviews
   documenting no increase in anastomotic leak rates in patients who used continuous
   positive airway pressure devices. They conclude by reviewing the costs of the measures
   they used to prevent respiratory complications leading to ICU admission. They note that
   ICU admission for two days would add more than $12,000 of hospital costs in their
   institution. Polysomnography and perioperative continuous positive airway pressure
   add less than $3000 of additional cost.

Editorial comment
       From the perspective of the editor, it seems that easily obtained information from
   the history and physical examination could identify patients for use of
   polysomnography and/or preventive continuous positive airway pressure. Information
   on a history of hypertension, snoring, obesity, and neck circumference > 17 inches
   would provide a basis for further evaluation.

Prevention of respiratory complications
      Patients having the constellation of risk factors discussed above are candidates for
   preventive strategies to reduce risk of respiratory complications. Multivariable risk
   studies have identified systemic risk factors such as elevated ASA score, smoking,
   obesity, older age, and need for complex operation as significant predictors of
   postoperative respiratory complications. Congestive heart failure and COPD are
   patient-specific factors potentially modifiable. It is well recognized that smoking
   cessation and measures to stabilize cardiovascular disease require two months or more
   of preoperative effort for a meaningful impact on complication risk to occur. Lung
   specific interventions such as treatment of lung infection, sputum reduction measures,
   use of bronchodilators, and preoperative respiratory muscle training have potential
   value. These topics will be discussed in the following section of the overview.

Preoperative maneuvers
       The first article reviewed here is by Gore60 in Gerontology, 2007. The article is
   entitled “Preoperative maneuvers to avert postoperative respiratory failure in elderly
   patients.” The author opens the review by emphasizing the clinical importance of
   abnormal postoperative ventilation, hypoxemia, and hypercarbia. These lead to the
   need for intubation and mechanical ventilator support. Dysfunctional ventilation
   leading to intubation greatly increases the risk of ventilator-associated pneumonia
   associated with a mortality risk exceeding 50% (see discussion in SRGS, Vol. 35, No. 7).
   Mortality rates for ventilator-associated pneumonia have not changed in many years.
   This consistent observation supports the importance of preventive strategies to reduce
   the need for intubation in this patient group.

       Particular problems that might predispose elderly patients to perioperative
   respiratory complications include 1) a gradual decline in one-second forced expiratory
   volume (FEV1) with advancing age; 2) increased ventilation/perfusion mismatching
   caused by increased early airway closure in dependent lung units; and 3) depressed
   responses to hypoxemia and hypercarbia. Gore cites data that document decreased
   FEV1 as an accurate predictor of postoperative respiratory complications, especially in
   patients with COPD. Ventilation/perfusion mismatching causes an age-related decline
   in resting arterial oxygen tension. Furthermore, older patients develop blunted
   responses to hypoxemia and hypercarbia and are vulnerable to analgesic and sedative-
   induced respiratory depression. Age-related decreases in mucociliary function reduce
   clearance of bacteria from the airway and contribute to increases in risk for
   perioperative pneumonia. This abnormality is particularly pronounced in smokers.

       Preoperative evaluation can include easily obtained data from the history and
   physical examination that can guide subsequent preventive efforts. In addition to
   obstructive sleep apnea screening (discussed above) a history of smoking, reactive
   airway disease, allergy, cough, and excessive sputum production can be obtained. The
   degree of chronic cough can be ascertained using a “cough test.” The patient is asked to
   cough. If the cough results in production of sputum or repeated coughing, additional
   testing (such as quantification of FEV1) might be helpful. If excessive sputum
   production is documented, a sputum culture is obtained. Recovery of a pathogen such
   as H. influenza, S. Pneumoniae, or MRSA can prompt a short course of preemptive
   antibiotics. Additional interventions that can strengthen cough and reduce sputum
   production include postural drainage, assisted cough, and deep breathing exercises.

       Inhaled bronchodilators are indicated preoperatively in patients with reactive
   airway disease and in patients with chronic bronchitis. Data cited by Gore suggest that
   ipratropium bromide (Atrovent®) is a useful first-line inhalant. Aminophylline has also
   been used for this purpose but data cited by Gore suggest that the association of this
   drug with tachycardia limits its use in elderly surgical patients. In patients with
   documented COPD, preoperative corticosteroid therapy might be useful. Effectiveness
   of steroid therapy is monitored with sequential assessments of FEV1. If this variable
   improves with steroid therapy, preoperative and postoperative therapy is valuable.
   Gore stresses that, overall, fewer than one-third of COPD patients will have significant
   responses to corticosteroids although degrees of improvement in some patients are
   substantial. Gore emphasizes the importance of continuing preoperative therapy into
   the postoperative recovery period.

Postoperative maneuvers
       Time-honored patient care processes for minimizing postoperative pulmonary
   complications include early ambulation, encouraging cough, elevation of the head of the
   bed, and judicious use of analgesics and sedatives. These interventions are valuable for
   preventing atelectasis and maintaining lung inflation. One device used for maintenance
   of lung inflation is the incentive spirometer. Use of this device is the topic of an article
   by Westwood and coauthors61 in Surgeon, 2007. These authors analyzed data from 263
   patients; the study was not randomized. One group of patients had intensive chest
   physiotherapy supported by a physiotherapist visit at least once daily during the
   postoperative hospitalization. The other group had similar physiotherapy with the
   addition of the incentive spirometer. Respiratory complications were defined as new
   fever > 38°C, signs of atelectasis or infiltrate on chest radiograph, and/or institution of
   antimicrobial therapy for suspected pulmonary infection. Both patient groups consisted
   of elderly patients (mean age 68 years); more than half of each group had a history of
   smoking and both groups underwent high-risk abdominal or noncardiac thoracic
   operations. Respiratory complications, according to the authors’ definition, occurred in
   17% of controls and 6% of patients using the incentive spirometer.

       The Westwood article and the article authored by Gore60 review data from other
   studies relevant to the use of the incentive spirometer. Of the available studies, half
   show benefit from use of the device and half show no favorable effect on postoperative
   respiratory complications. The available studies vary in the application of other
   approaches such as intensive chest physiotherapy. The main complication of incentive
   spirometer use is gastric dilatation, which has been reported several times.
    Westwood’s study reports no instances of this complication. It is difficult to ascribe
a consistent clinical benefit to incentive spirometer use as an isolated intervention.
Significant benefit for the device when it is added to dedicated chest physical therapy
might be associated with the fact that a well-trained patient can use the device during
the intervals between physical therapist visits.

    A randomized trial evaluating intensive inspiratory muscle training under the
supervision of a physical therapist as a means of reducing perioperative respiratory
complications by Dronkers and coauthors62 appeared in Clinical Rehabilitation, 2008. In
this study, 20 patients undergoing open abdominal aortic aneurysm repair were
randomized to receive intensive inspiratory muscle training (one physical therapist
supervised session and five unsupervised sessions/week for two weeks prior to
operation). This group was compared with a control group receiving instruction in deep
breathing and incentive spirometer use. The primary endpoint of this study was
detection of atelectasis on chest radiograph. The analysis disclosed a nonsignificant
trend toward less atelectasis in the group that received intensive inspiratory muscle
training. Maximum inspiratory force increased by 10% in the intervention group.
Patient satisfaction with the intervention was high. The authors acknowledge the need
for additional studies involving larger patient groups.

    A meta-analysis of available data from studies evaluating prophylactic respiratory
physical therapy by Pasquina and coauthors63 from Chest, 2006, evaluated 35 trials that
provided data on the possible value of respiratory physical therapy as a means of
preventing perioperative respiratory complications. Significant differences in
postoperative respiratory events were reported in only four studies that included a “no
intervention” control group. In some studies differences occurred in the frequency of
atelectasis (usually defined as a change on chest radiograph). Most studies did not focus
on important complications such as pneumonia, need for intubation, or ventilator

    In the single study analyzing effects of respiratory physical therapy on the frequency
of pneumonia, a significant reduction was recorded but the frequency of pneumonia in
the control group was higher than baseline rates for this complication recorded in other
clinical series. This fact limits the external validity of this study. Unspecified respiratory
complications were reduced in one analysis. These authors concluded that routine use
of physiotherapy is not indicated in low and moderate risk patients undergoing
abdominal operations. There were no reports of adverse events associated with the use
of physical therapy. Readers will recall that we noted earlier reports of gastric
distention associated with the use of the incentive spirometer but few of the studies
supporting use of this device report occurrences of gastric distention. Failure to
   consider potential harm from an intervention weakens data supporting use of the

       Gore60 concludes his article with a discussion of nutritional support and cessation of
   alcohol use as a means of preventing perioperative respiratory complications. He notes
   that use of supplemental parenteral nutrition as a means of improving nutrition has
   been limited by the known complications of this intervention (hyperglycemia, liver
   dysfunction), its cost, and the need for central venous access. Use of this intervention is
   unusual except in patients who have lost more than 5% of ideal body weight or patients
   who were profoundly hypoalbuminemic from nutritional impairment. Use of
   preoperative enteral nutrition is also unusual because of the need for enteral access and
   limited patient tolerance. Enteral access should be acquired in patients undergoing
   high-risk abdominal procedures where preoperative weight loss and/or
   hypoalbuminemia suggest nutritional deficit. Although there are data indicating value
   of human growth hormone therapy in managing burn patients (especially burned
   children), Gore stresses that there are no high-quality data supporting use of this agent
   to reduce perioperative respiratory complications.

       Gore concludes his article by emphasizing the importance of cessation of alcohol use
   by heavy drinkers before a major operation. This will require careful counseling of each
   individual patient. Abstinence from alcohol use, even for short intervals before
   operation might be valuable. Acute cessation of alcohol use might precipitate clinical
   alcohol withdrawal syndrome, which carries its own risk of mortality. This eventuality
   should be avoided to the extent possible.

Editorial comment
       In the section of the overview, we have discussed preoperative and postoperative
   maneuvers that might be helpful for prevention of respiratory complications.
   Interpretation of the available data is challenging because of the variable definitions of
   respiratory complications and the small patient groups that make up most of the
   available studies. Most patients undergoing abdominal or thoracic operations should
   have early ambulation, elevation of the head of the bed, brief training in coughing and
   deep breathing, and careful pain control. Interventions such as preoperative antibiotics,
   inhaled bronchodilators, continuous positive airway pressure breathing, and
   corticosteroids should be applied in carefully selected patients.

       The most consistent marker for postoperative respiratory complications is the need
   for intubation in the postoperative period. Analyses of this event in two recent
   studies64-65 indicated that intubation in the post-anesthesia care unit is closely related
   to the presence of residual neuromuscular blockade. Careful tracking and recording of
   the level of neuromuscular blockade effectively prevented early, unplanned intubations.
   In populations of postoperative general surgery patients, unplanned intubation is an
   event associated mainly with nonmodifiable events such as chronic disorders of
   consciousness, severe cardiovascular and/or pulmonary disease, and the development
   of postoperative sepsis.

       Data indicate that the overall incidence of unplanned postoperative intubation is
   low (less than 3%) but the event carries a mortality risk in excess of 40%. Current
   preventive interventions such as emergency response teams have had limited impact
   on the frequency and mortality risk of this complication. Most unplanned postoperative
   intubations occur in patients who are already in the ICU. Prevention of unplanned
   intubation seems to be most useful for patients in the post-anesthesia care unit.
   Recognition of risk factors such as severe cardiopulmonary disease and chronic
   depressed level of consciousness might help identify patients at increased risk for
   postoperative unplanned intubation.

Noninvasive ventilation for postoperative respiratory complications
       Increasing recognition that the morbidity of postoperative respiratory failure is
   driven, at least in part, by complications of intubation and mechanical ventilation
   (especially ventilator-associated pneumonia) has stimulated efforts to use
   nonintubation interventions for early treatment of this complication. The first article
   reviewed that deals with this topic is by Michelet and coauthors66 from the British
   Journal of Surgery, 2009. This article is supplied as a full-text reprint with this issue of
   SRGS. The authors note that respiratory failure and anastomotic leak are linked
   complications in patients undergoing esophagectomy. Anastomotic leak occurs when
   ischemia or diminished oxygen delivery to the anastomotic area occurs. Thus, hypoxia,
   which might accompany the onset of postoperative respiratory failure, can contribute
   to the risk of anastomotic leak.

       The reported study is a case-control design study in which a group of patients
   treated with postoperative noninvasive ventilation was compared with a group of
   patients who did not receive this intervention. In all other aspects, the patient groups
   were comparable. Thirty-six patients comprised each study group. Acute respiratory
   failure was characterized by dyspnea and use of accessory muscles of ventilation, new
   infiltrates visible on chest radiograph, purulent sputum, fever, and hypoxemia
   (Pa02/FI02 ratio of < 200). Noninvasive ventilation was delivered with a face mask and
   a ventilator. Continuous positive airway pressure and positive end-expiratory pressure
   were used and pressures incrementally increased until tidal volume reached a level of 6
   mL/kg estimated ideal body weight and arterial oxygen saturation exceeded 90%.
   Maximum inspiratory pressure was maintained below 25 cm H20. Episodes of
   noninvasive ventilation were interspersed with 45-60 minute intervals without assisted
    Nine of 36 patients in the noninvasive ventilation group eventually required
endotracheal intubation, but only one of these because of intolerance of the mask. In
comparison, 19/36 patients in the control group required intubation. This very low
incidence of mask intolerance contrasts with data from a report by Conti and
coauthors67 that describes a comparison between a mask and a helmet interface for
delivery of noninvasive ventilation. In this study, the patients treated with the helmet
device were compared with historic control patients treated with mask ventilation.
Diagnostic criteria for postoperative acute respiratory failure used by Conti and
colleagues were similar to the criteria used by Michelet and associates. Twenty percent
of the helmet patients and 48% of the mask patients required intubation and the main
reason was intolerance of the noninvasive intervention.

    Both studies indicate that noninvasive ventilation can effectively improve
oxygenation and prevent the need for intubation in many patients. Overall, 20%-50% of
patients in whom noninvasive ventilation is attempted will fail and failure is often times
because the patient cannot tolerate the device. Careful patient selection and availability
of close bedside supervision of the ventilation might reduce this risk, but at significant
added cost in terms of respiratory therapist and nursing time. In general, patients
selected for noninvasive ventilation should have normal sensorium and not have severe
dyspnea or difficulty managing secretions. If these criteria can be met, this intervention
is often effective as a means of supporting oxygenation and avoiding intubation.

    The use of noninvasive ventilation in patients with more severe forms of respiratory
failure is controversial. The main benefit of this approach would be to avoid intubation
with the attendant reduction in risk of ventilator-induced lung injury and ventilator-
associated pneumonia. This subject is the focus of two articles reviewed at this time.
The first is by Antonelli and coauthors68 in Critical Care Medicine, 2007. The authors cite
data from other studies indicating the possibility that intubation rates for patients with
early acute respiratory distress syndrome might be reduced by as much as 50% with
use of noninvasive ventilation. This report deals with 147 patients admitted to two
ICUs. The authors note that both units have extensive experience in the use of
noninvasive ventilation. The patients were diagnosed with acute respiratory distress
syndrome using standard criteria. Noninvasive ventilation was supplied using a mask
or a helmet device. Continuous positive airway pressure and positive end-expiratory
pressure were gradually increased in increments until exhaled volumes reached 6
mL/kg, respiratory rate was < 25 breaths/min, and oxygen saturation was consistently
>90%. Failure to achieve these goals and/or failure of the patient to tolerate
noninvasive ventilation defined failure of noninvasive ventilation.

    The authors achieved success avoiding intubation in 54% of the patients enrolled in
this study. They noted predictors of failure of noninvasive ventilation as higher severity
  of illness scores, older age, requirement for PEEP > 12 cm H20, and failure to improve
  oxygenation after one hour of noninvasive ventilation. Patients who required intubation
  were more likely to have severe sepsis or septic shock and a mortality rate of 54% was
  recorded in the group requiring intubation. Only 12% of patients could not tolerate
  noninvasive ventilation. It is likely that the extensive experience of these clinicians with
  noninvasive ventilation contributed to this level of success.

      Contrasting data are presented in an article by Rana and coauthors69 in Critical Care,
  2006, entitled “Failure of noninvasive ventilation in patients with acute lung injury:
  observational cohort study.” This report presents data on 54 patients who had
  noninvasive ventilation attempted as the first intervention for respiratory distress
  requiring admission to the ICU. The patients were severely ill with sepsis diagnosed in
  88% of patients. Septic shock was present in 19 of the 54 patients. Data on Pa02/Fi02
  ratios disclose that this patient group met criteria for acute respiratory distress
  syndrome rather than acute lung injury in essentially all patients. This was, thus, a high-
  risk cohort of patients. The authors observed failure of noninvasive ventilation in 70%
  of patients.

      The presence of shock, acidosis, and severe hypoxemia were predictive of failure of
  noninvasive ventilation. The higher failure rate in this series serves to emphasize the
  importance of patient selection. Severely ill patients, especially those with severe
  hypoxemia and hemodynamic instability, are at increased risk for failure of noninvasive
  ventilation. These findings are similar to those of Antonelli and coauthors68 discussed
  above. Because concerns have been raised about possible harmful effects of delaying
  intubation in patients who are at high risk of progressing to severe acute respiratory
  distress syndrome, candidates for noninvasive ventilation should be carefully selected.

Acute lung injury and acute respiratory distress syndrome
       Acute lung injury is a term applied to a complex response of the lung to systemic
  and localized inflammatory stimuli. A multitude of injuring agents, acting singly or in
  combination, can produce the histologic, radiologic, and clinical manifestations of acute
  lung injury. These agents might act by direct injury to the lung tissue (pulmonary
  contusion, pulmonary blast injury) or to the airway (aspiration, inhalation injury). In
  other instances, the inflammatory process begins with a remote stimulus (peritonitis,
  pancreatitis, sepsis, combined traumatic injury, shock, and resuscitation) and the lung
  is injured because of circulating factors that act directly on the lung microcirculation
  and/or lung tissue or through activation of inflammatory mediators within the lung
  microcirculation. Pneumonia can trigger the development of acute lung injury in
  adjacent noninfected lung through propagation of the inflammatory process.
       Patients can recover from acute lung injury or might progress to acute respiratory
   distress syndrome, a clinical entity that is manifest by hypoxemia from
   ventilation/perfusion mismatching, loss of lung compliance because of alveolar flooding
   and consolidation of lung tissue, and increased dead space ventilation resulting from
   pulmonary microvascular occlusion. There is no specific therapy for acute lung injury.
   Support of ventilation and oxygenation using adjuvant ventilation therapies can assist
   the lungs in the effort to maintain oxygen transfer from alveolus to blood but these
   therapies have no positive effect on the severity or clinical course of acute respiratory
   distress syndrome. In fact, as clinicians have learned over the past decade, adjuvant
   ventilator therapy can produce additional injury in the lung through effects of pressure,
   volume, cycling of ventilation, and promotion of the inflammatory process.

       In this section of the overview, we discuss important clinical features of the
   pathophysiology, diagnosis, and management of acute lung injury and acute respiratory
   distress syndrome. Entire volumes have been written on these topics and, because of
   the vastness of the information available, this review will not be comprehensive; the
   review will focus on clinically valid understandings and effective interventions.

       Biomarkers of inflammation, altered coagulation, fibrinolysis, and increased
   oxidative stress are activated, elevated, or suppressed to varying degrees and with
   different trajectories in all inflammation-mediated diseases including acute lung injury.
   Pro-inflammatory cytokines can be recovered from blood and from alveolar fluid in
   animals and patients with acute lung injury and adult respiratory distress syndrome.
   Elevations of some biomarkers such as interleukin-6 (IL-6), interleukin-8 (IL-8), and
   intercellular adhesion molecule-1 (ICAM-1) are associated with poorer clinical
   outcomes for acute lung injury. This is also true for coagulation factors. Lower levels of
   Protein C and elevations of thrombomodulin indicate a pro-coagulant state; this pattern
   is associated with poor outcomes. Finally, impaired fibrinolysis is indicated by
   elevations of plasminogen activator inhibitor-1 (PAI-1) and elevated levels of this
   substance have been associated with poor outcomes of acute lung injury and acute
   respiratory distress syndrome.

       Recent changes in clinical approaches to acute respiratory distress syndrome
   include the use of low tidal volume (6 mL/kg), low mean airway pressures (25 cm H20),
   positive end expiratory pressure (12 cm H20), and permissive hypercapnia as clinicians
   attempt to minimize inflammatory stimuli and reduce the impact of ventilator-
   associated lung injury. This approach is termed “open-lung” ventilation or “lung-
   protective” ventilation. There are data suggesting that “open-lung” ventilation reduces
   inflammatory mediators and the reduction is associated with improved clinical
   outcomes. The first article discussed in this section analyzes the effect of open lung
ventilation on inflammatory mediators and seeks to explore the question whether
patterns of inflammatory mediator levels remain predictive of outcomes in the era of
“open-lung” ventilation. The article by McClintock and coauthors70 appeared in Critical
Care, 2008.

    The authors collected ventilator data and serum levels of biomarkers in 50 patients
with acute respiratory distress syndrome. The causes of acute respiratory distress
syndrome varied and included most of the common causes; no single cause
predominated. The ventilator data confirmed that patients were treated with the “open-
lung” approach. Serum biomarker patterns were significantly different in survivors and
nonsurvivors. After adjustment for various risk factors, elevated levels of ICAM-1 and
IL-8, and depressed levels of Protein C were predictive of mortality. The authors
concluded that patterns of biomarker activation are predictive of outcomes and this
predictive value has not been eliminated by “open-lung” ventilation strategies. In the
discussion section of this report, lower levels of Protein C were predictive of mortality
even when the data were adjusted for the frequency of sepsis as the cause of acute
respiratory distress syndrome. This observation suggests that Protein C administration
might favorably affect recovery in some patients with acute respiratory distress
syndrome. It is disappointing that recent data have not shown a benefit for
administration of Protein C in patients with acute respiratory distress syndrome.

    Burn induced lung injury combined with inhalation injury is a prototypical example
of combined lung parenchyma and airway injury. The features of combined
burn/inhalation injury are discussed in an article by Enkhbaatar and Traber71 in
Clinical Science, 2004. The article is supplied as a full-text reprint accompanying this
issue of SRGS. The authors begin by noting that inflammation induced increases in
microcirculatory permeability characterizes burn injuries that exceed 30% of the body
surface area. This hyperpermeability affects the microcirculation at the burn site and in
tissues remote from the site of the burn injury. The result is a large flux of fluid and
protein from the intravascular to the interstitial space with edema formation in the area
of the burn and in all tissues. Inhalation injury alone and combined burn and inhalation
injury cause increased pulmonary microcirculatory permeability. Pulmonary edema
occurs not only because of flux of protein and fluid from the pulmonary circulation but
also because of enormous increases in blood flow to the tracheal-bronchial tree.
Anatomic connections between the bronchial arteries and the pulmonary circulation
deliver a portion of this increased blood flow to the pulmonary circulation and this
contributes to pulmonary edema formation. In animal experiments these investigators
noted that occlusion of the bronchial-pulmonary connecting channels in a smoke
inhalation model greatly reduces pulmonary edema and improves lung function after
inhalation injury.
   In burn-induced acute lung injury, there is also a strong pro-inflammatory state.
Nitric oxide and metabolites of this substance play major roles in the inflammation-
induced lung injury caused by burn. The authors have shown in experimental
preparations that there is upregulation of nitric oxide production. Stable plasma
metabolites of nitric oxide increase 2-2.5 fold after burn injury.

    Nitric oxide exists in three forms, neuronal nitric oxide (nNOS), endothelial nitric
oxide (eNOS), and inducible nitric oxide (iNOS); nNOS and eNOS are constitutive
isoforms and iNOS is induced by multiple components of the pro-inflammatory state.
These authors note that the pro-inflammatory factors IL-1 and endotoxin activate
nuclear factor κ-B. This factor is a potent stimulus for production of iNOS. This same
activation pathway leads to increased production of superoxide that contributes to the
oxidative stress characteristic of the pro-inflammatory state. Elevated levels of iNOS
also contribute to the oxidative stress by combining with superoxide to produce
peroxynitrite, which can damage the alveolar capillary membrane. When stores of
arginine are depleted, iNOS produces superoxide, which can cause tissue damage.

    The authors note that experimental studies have demonstrated arginine depletion
in combined inhalation/burn injury with pulmonary dysfunction. The vasodilating
properties of iNOS contribute to one of the most important features of acute lung injury,
ventilation/perfusion mismatching. Hypoxic vasoconstriction, the protective reaction
that redistributes blood from underventilated alveoli to ventilated alveoli, cannot
function in a high iNOS environment. In acute lung injury, underventilated alveoli have
sustained perfusion leading to delivery of unoxygenated blood to the pulmonary veins.
The finding of hypoxemia and an abnormal alveolar-arterial gradient in patients with
acute lung injury can be explained by the failure of the hypoxic vasoconstriction. Data
from the authors’ laboratory have shown improved lung function when animals are
pretreated with an iNOS inhibitor.

   Enkhbaatar and Traber caution that iNOS is a component of the complex defense
against inflammation and a contributor to the pathophysiology of acute lung injury.
Because of this, inhibition of iNOS has not produced improved outcomes in clinical
studies of pro-inflammatory states such as septic shock. Specific inhibitors of one or
another of the NOS isoforms might produce better outcomes. The authors note data
from animal experiments showing improved lung function in acute lung injury treated
with a specific nNOS inhibitor or with the anti-inflammatory agent, ketorolac.

    Another factor activated by inflammation with resultant cell damage and death is
poly-(ADP-ribose)-polymerase or PARP. This substance is activated in cells in response
to DNA damage and this factor is active in DNA repair processes. Over activation results
in depletion of cellular energy stores that can lead to necrotic cell death. PARP is an
important contributor to endotoxin-induced lung inflammation and inhibition of PARP
can preserve ATP levels in the lung after acute lung injury according to data cited in the
report by Enkhbaatar and Traber.

    Combined inhalation and burn injury results in small airway obstruction because of
leakage of exudate rich in neutrophils and products of coagulation, such as fibrin, into
the airway lumen. Airway obstruction results in increasing dead space ventilation and
contributing to the pathophysiology of acute lung injury. Experimental data from
studies completed by these authors demonstrate that this process can be reversed,
partially by using nebulized tissue plasminogen activator.

    The complex pathophysiology of acute lung injury is produced by inflammatory
injury to the lung microcirculation, the alveolar capillary interface, and the airways.
Each component is present, to varying degrees, depending on the agent producing the
inflammatory state and the resulting lung injury. Transfusion of banked blood is one
cause of acute lung injury. This topic is reviewed by Swanson and coauthors72 in Lung,
2006. Trauma and critical care surgeons have consistently observed an association
between massive transfusion and post-trauma acute respiratory distress syndrome.
That single transfusions of blood products (except albumin) can cause lung injury and
respiratory distress acutely was first recognized in 1951. Characterization of
transfusion related lung injury as a transfusion reaction resulted from studies
demonstrating anti-leucocyte antibodies in patients with acute lung injury closely
following transfusion. The term transfusion related lung injury (TRALI) was first coined
in a 1983 report cited by Swanson and coauthors.

    The TRALI clinical syndrome affects approximately 0.2% of patients who receive
transfusions. Eight percent of transfusion reactions result in lung injury and this entity
is the cause of 13% of transfusion-related fatalities. Overall mortality for a TRALI
episode is 5%-10%. Respiratory distress usually develops within six hours of
transfusion. The clinical picture is typical of acute lung injury and consists of
hypoxemia, tachypnea, and pulmonary infiltrates on chest radiograph. Supplemental
oxygen is required in all patients and nearly three-quarters of the symptomatic patients
will require intubation and ventilator support. The typical TRALI episode will resolve
with supportive care in 48-96 hours.

   A definitive clinical diagnosis requires exclusion of other causes of acute lung injury.
Most patients who require blood product transfusions are significantly injured or ill and
exclusion of other causes for acute lung injury might not be possible. Recovery of anti-
HLA and/or anti-leucocyte antibodies from donor blood supports a diagnosis of TRALI,
especially if the recipient is shown to have a leucocyte antigen phenotype matching the
antibody recovered from donor blood. Culprit antibodies are found in up to 90% of
   donor blood samples when a TRALI episode occurs. Despite this observation, TRALI
   episodes occur without demonstrable immune reaction. The observation that older
   banked blood is more likely to cause TRALI has led to the development of nonimmune
   models of TRALI. These involve factors that prime neutrophils and are present in
   increased concentrations in older banked blood. The specific factor or factors
   responsible are not known but plasma from recipients who develop TRALI contains
   increased concentrations of neutrophil-priming substances. Histopathology of the lung
   shows pronounced leucocyte sequestration in the lung. It is likely that lung damage is
   caused by oxygen free radicals induced by the sequestered leucocytes.

Therapy for TRALI is mainly focused on support of oxygenation until the process resolves.
Prevention of TRALI is challenging. Rejection of blood donation by donors implicated in a
TRALI episode is one avenue. Probably the most effective approach will be to adopt
conservative transfusion protocols so that banked blood transfusion is reduced, overall.

Epidemiology and outcomes of acute lung injury and acute respiratory distress syndrome
       Two older articles are discussed in this section to provide perspective about
   morbidity and mortality associated with acute lung injury and acute respiratory
   distress syndrome. Prognostic factors will also be discussed. The first article discussed
   is by Luhr and coauthors73 from the American Journal of Respiratory and Critical Care
   Medicine, 1999. These authors begin by citing several reports that document mortality
   rates for acute respiratory failure and acute respiratory distress syndrome ranging
   from 40%-50%. They note, also, that several reports have suggested that mortality for
   acute respiratory distress syndrome might be decreasing. Interpretation of data is
   challenging because of the variability of clinical definitions used in the reported studies.
   The aim is this analysis was to perform a prospective cohort study involving patients
   aged 15 years and older admitted to ICUs in Sweden, Denmark, and Iceland during an
   eight-week interval.

       Acute respiratory failure was defined as endotracheal intubation followed by 24
   hours or more of ventilator support. Acute lung injury and acute respiratory distress
   syndrome were defined according to criteria promulgated by the American-European
   Consensus Conference on ARDS. The consensus conference definition includes the
   following criteria 1) acute symptom onset; 2) Pa02/FI02 ratio of < 300 for acute lung
   injury and < 200 for acute respiratory distress syndrome; 3) bilateral infiltrates on
   chest radiograph; and 4) pulmonary artery occlusion pressure < 18 mmHg or no clinical
   evidence of left atrial hypertension. Each patient was enrolled at the time of the first
   admission to the ICU and this admission was the only incident of respiratory distress
   counted in the study. The study included 1231 patients who fulfilled criteria of acute
   respiratory failure. Of this group 287 patients fulfilled criteria for acute lung injury and
   221 fulfilled criteria for acute respiratory distress syndrome. Mortality for the total
group was 41%. Mortality rates for acute lung injury (42.2%) and acute respiratory
distress syndrome (41.2%) were nearly identical to the total group mortality. The
question of whether these patients died from the respiratory disease or died with the
respiratory disease remains unanswered by these data.

     The authors note that the close agreement of death rates for all three clinical
diagnoses suggests that death from refractory hypoxemia might be less common than
death from a condition associated with or, perhaps, precipitating hypoxemia. Luhr and
associates were able to demonstrate significant mortality prediction from the presence
of liver disease, advanced age, and a nonpulmonary cause of respiratory dysfunction.
These factors also suggest that death might not have occurred solely because of
respiratory insufficiency. Additional perspective on this issue appears in an article by
Rocco and coauthors74 in Annals of Surgery, 2001. These authors conducted a
retrospective, single institution review analyzing mortality and prognostic factors in
980 consecutive patients who were intubated and who received ventilator support in
the ICU. Among these were 111 patients who fulfilled the American-European
Consensus Conference on ARDS criteria for acute respiratory distress syndrome (see
discussion above).

    Lung injury scores were calculated for each patient. The lung injury score assigns
points for assessments of the chest radiograph, degree of hypoxemia, level of PEEP, and
lung compliance. A score of 2.5 or more is indicative of acute respiratory distress
syndrome. Lung injury scores were > 2.5 in all 111 patients. Patients were divided into
subgroups depending on whether the patient was a “surgical” patient or a “trauma”
patient. Patients were also divided into a group of patients admitted between January 1,
1990, and December 31, 1994, and those admitted between January 1, 1995, and
December 31, 1998. The data indicate that surgical patients were older and more likely
to have respiratory failure related to intraabdominal infection. Trauma patients were
more likely to have respiratory failure related to multiple injuries and/or direct
thoracic injury. Mortality rates declined in both periods but the magnitude of decline
was statistically significant for trauma patients only.

    In the second interval, overall mortality for acute respiratory distress syndrome
declined from 72% to 38% and this decline was statistically significant. The authors
emphasize that the decline in mortality occurred even though the patients in the second
interval were older. One characteristic of the more recent group was that emergency
operation frequency declined. A decrease in emergency operations was probably
associated with a lower risk for intraabdominal infection, which was the most common
cause of fatal acute respiratory distress syndrome. The authors also point out that they
used lower tidal volumes and lower mean airway pressures as ventilator strategies in
   the later group. The authors note that advanced age and comorbid illness (particularly
   liver disease) were strongly predictive of mortality.

Editorial comment
       The data discussed above document the correlation between age, increasing illness
   severity, infection, and acute respiratory distress syndrome. The role of inflammation in
   the genesis of acute lung injury and as a driver of progression to acute respiratory
   distress syndrome is confirmed in the clinical series. Younger patients with direct lung
   injury (trauma patients) are more likely, as a group, to survive. Ventilator strategies
   might also play a role in producing the improved outcomes observed in the article by
   Rocco and colleagues.74 The current approaches to ventilator support for patients with
   acute lung injury and acute respiratory distress syndrome is discussed in more detail in
   the next section.

Ventilator strategies
       The approach to ventilator therapy for patients with acute lung injury and acute
   respiratory distress syndrome has changed in one major and several minor ways over
   the past three to five years. Traditionally the approach to ventilation has been designed
   to maintain oxygenation and carbon dioxide removal. An understanding of the basic
   mechanisms of acute respiratory failure (many of these discussed above) has improved
   understanding. Most importantly, surgeons are now aware of the potential patient
   harm that might accompany ventilator therapy. The effects of ventilator pressures and
   volumes on hemodynamics have been well recognized for many years.

       More recently, a group of phenomena has been recognized known as ventilator-
   induced lung injury. Included in this category are barotrauma (mediastinal emphysema,
   pneumothorax), which refers to lung damage from disruption of alveoli resulting from
   excess alveolar pressures. Barotrauma can also result from the combination of rapid
   ventilator rates with PEEP that produces “auto-PEEP,” a phenomenon that produces
   successively increasing airway pressure because the rapid respiratory rate does not
   allow return of airway pressure to the set PEEP level before the next breath is
   delivered. Volutrauma refers to alveolar damage caused by ventilation of high-
   compliance areas of the lung with large inspired volumes. The large volumes are
   delivered to high-compliance areas because consolidated areas of the lung lose
   compliance and the inspired gas is “shunted” to the high-compliance alveoli.
   Atelectrauma is the term for alveolar injury that occurs from successive deflation and
   inflation of alveoli during the ventilator cycle. Unstable alveoli might collapse at end
   expiration and have to be reopened with the next inspiration, which produces injury to
   the alveolus. The “collapse-reopen” cycle also increases the intensity of the lung
   inflammatory response. Production of pro-inflammatory cytokines is stimulated and
   this phenomenon contributes to ventilator-associated lung injury.
    The approach to minimizing ventilator-associated lung injury is based on an
understanding that lung injury results from interactions of the ventilator cycle, mean
airway pressure, and tidal volume. High mean airway pressure required to deliver high
tidal volumes is the main cause of ventilator-associated lung injury. Approaches to lung-
protective ventilation strategies emphasize the need to lower tidal volume and lower
mean airway pressure. In patients with acute respiratory distress syndrome,
microcirculatory obstruction and small airway obstruction might increase dead space
ventilation to the extent that PaC02 rises. This rise can be made tolerable for the patient
so that additional respiratory distress does not occur. The process of allowing PaC02 to
increase is termed “permissive hypercapnia” and this is a component of lung-protective
ventilation strategies.

    Data from randomized trials have demonstrated a significant reduction in mortality
for acute respiratory distress syndrome with the use of lung-protective ventilation. In
this section of the overview, we will review some recent contributions to the medical
literature pertinent to ventilator therapy for patients with acute respiratory distress
syndrome. The scientific basis and clinical effectiveness of “recruitment maneuvers”
designed to reopen and maintain alveoli and alternative strategies for “weaning” the
patient from ventilator therapy are reviewed.

    The first article was published in 2003 in Critical Care Medicine75 and is a
randomized controlled trial of alveolar recruitment maneuvers sponsored by the
ARDSNet group of investigators. A full-text reprint of this article accompanies this issue
of SRGS. Brower and coauthors note that lung-protective ventilation has been shown to
decrease mortality. From the available data, it is not clear that lung protective
ventilation contributed to alveolar recruitment, lowered risk of ventilator-associated
lung injury, or improved outcomes in terms of lung function. This study was performed
to assess the effectiveness of maneuvers designed to reopen and maintain alveolar
inflation. In the experimental group, recruitment maneuvers were administered on the
first and third or second and fourth mornings after enrollment. The recruitment
maneuver consisted of changing the ventilator mode to continuous positive airway
pressure mode and increasing airway pressure in increments up to 35-40 cmH20
depending on body weight. The recruitment maneuver was held for 30 seconds unless
hemodynamic instability, cardiac arrhythmia, decreased oxygen saturation, or
tachycardia (>130 bpm) occurred. The control patient group received sham
recruitment maneuvers. PEEP was adjusted according to the level of inspired oxygen
required by the patient. An upper limit of 24 cmH20 was permitted.

   The authors noted immediate improved oxygenation after the use of recruitment
maneuvers but the effect diminished over time. The authors note that this observation
might have occurred because near maximal alveolar inflation had already been
achieved using PEEP. They note that the single recruitment maneuver might not be as
effective as repeated maneuvers. The topic of multiple recruitment maneuvers is the
focus of the next article reviewed.

    The second article is by Meade and coauthors76 in the Journal of the American
Medical Association, 2008. These authors report a randomized trial of patients whose
acute respiratory distress syndrome was diagnosed by standard criteria. One group of
patients was treated with standard lung-protective ventilation (respiratory rate < 25,
tidal volume of 6 mL/kg, and PEEP of 8-12 cmH20). Another group had lung protective
ventilation combined with recruitment maneuvers (inspiration with breath hold at a
steady pressure of 40 cm H20 for 10-15 seconds during each ventilator disconnection
for suctioning or other reasons up to 4 times/day). In the experimental group, PEEP
was set dependent on the Fi02 required by the patient. PEEP pressures ranged from 5-
10 cmH20 for Fi02 of 0.6 or less and ranged up to 24 cmH20 for patients requiring Fi02 of
1.0. Volume-controlled assisted breaths were used in control patients and pressure-
controlled ventilation was used in the experimental group. “Rescue therapies” (prone
position, high frequency oscillation ventilation, extracorporeal membrane support)
were permitted for patients with refractory hypoxemia and/or acidosis.

    The frequency of barotrauma (mediastinal emphysema, pneumothorax) was
recorded. The data demonstrate that the experimental group had improved
oxygenation with no increase in the risk of barotrauma or rescue interventions. The
mortality for the entire group was 38% and there was no significant difference in
mortality risks for the two groups. Of interest is that attributable mortality for
refractory adult respiratory syndrome was only 6% overall. The experimental group
had a significant reduction in mortality from refractory hypoxemia. The authors
concluded that addition of recruitment maneuvers and incremental increases in PEEP
(based on the required inspired oxygen concentration) resulted in improved
oxygenation that was durable over time. This strategy represents an acceptable
alternative to conventional therapy.

    An abiding question regarding the effectiveness of recruitment maneuvers relates to
whether recruitment maneuvers, combined with optimum PEEP, distend already
inflated alveoli or recruit previously contracted or collapsed alveoli. This issue is the
topic of an article by Schreiter and coauthors77 in Critical Care Medicine, 2004. This
study analyzed helical CT images obtained before and after recruitment maneuvers and
PEEP adjustment in 17 patients with direct lung trauma resulting in acute respiratory
distress syndrome. The authors observed increased lung inflation on CT images that
was obtained by reducing consolidated lung area rather than expanding inflated lung.
They concluded that recruitment maneuvers combined with optimum PEEP reopens
contracted or collapsed alveoli and provides sustained inflation of these lung areas.
    The next article reviewed by Mercat and coauthors78 appeared in the Journal of the
American Medical Association, 2008. This report describes a randomized controlled trial
in patients who had acute respiratory distress syndrome diagnosed by standard
criteria. One group of patients was treated with lung-protective ventilation with PEEP
levels set from 5-9 cmH20 based on the level of oxygenation. The second group of
patients had PEEP adjusted upward to establish a plateau airway pressure of 28-30
cmH20. In this study the experimental group had higher fluid requirements (probably
because of the hemodynamic effects of higher airway pressures), but experienced
better oxygenation, a lower risk of requiring “rescue” interventions, and decreased
ventilator and organ failure days. The authors caution that the patients in this study
who met criteria for acute lung injury versus acute respiratory distress syndrome had
less benefit from the higher PEEP strategy, and this patient group might actually
experience lung injury from high PEEP.

    The data about the durability of alveolar recruitment are variable, as is obvious
from the articles previously reviewed. A systematic review of available data on
recruitment maneuvers by Fan and coauthors79 in the American Journal of Respiratory
and Critical Care Medicine, 2008, analyzed data from studies involving nearly 1200
patients. Available studies all showed improved oxygenation after application of
recruitment maneuvers but most studies also disclosed that improvement was
transient. Adverse events are unusual but arterial hypotension accompanies most
recruitment maneuvers. Hypotension is noted more often in patients with less severe
lung injury. They note that PEEP elevations after a recruitment maneuver improve the
durability of the improvement in oxygenation. The authors conclude with the caution
that the value of transient improved oxygenation noted after recruitment maneuvers is
currently unknown. Significant impact of recruitment maneuvers on global outcomes
measures such as mortality has not been demonstrated. These authors urge that the
decision to employ recruitment maneuvers be based on the severity of respiratory
distress (less severely hypoxemic patients probably do not benefit) and the response of
the individual patient to PEEP and recruitment maneuvers. The least improvement in
oxygenation was observed in patients with low lung compliance. This might indicate
that such patients have limited capacity for alveolar recruitment.

    The fundamental concept supporting the use of PEEP and recruitment maneuvers is
that alveolar collapse is a major driver of ventilation/perfusion mismatching,
hypoxemia, and loss of compliance in acute lung injury. An experimental study
examining this question is by Mertens and coauthors80 appeared in Critical Care
Medicine, 2009. These authors used darkfield intravital microscopy to view the lung
parenchyma visualized through a transthoracic window. Lung injury was induced with
intratracheal hydrochloric acid.
    These authors found that alveolar distention increases with ventilation pressure in
normal lungs with a sigmoid shaped curve demonstrated when the percent increase in
alveolar volume was plotted against inflation pressure. Damage to the lung resulted in
alveolar thickening and reduced alveolar volume but alveolar collapse was not
observed. In an editorial by Hubmightr accompanying Mertens’ article, the editorialist
notes that the elegant observations reported in the work of Mertens and colleagues
shows that in the normal lung alveoli are not recruited but distend and contract. The
lung damage produced in this study did not result in lung edema, and Hubmightr notes
that alveolar damage leading to alveolar collapse can result from fracture of liquid
bridges in the edematous lung with inflation and deflation. Thus, the injury in this study
might not have reproduced the pathophysiology of acute lung injury. He concludes that
additional work is required before a full understanding is achieved of alveolar inflation
in the face of lung injury.

    Another ventilation strategy for patients with severe acute respiratory distress
syndrome is high-frequency oscillatory ventilation. This modality is used frequently in
premature infants with respiratory distress. Ventilation of the lung occurs from rapid
administration of very small tidal volumes (1-3 mL/kg) delivered at high ventilatory
rates that allow mixing of gas within the lung so that oxygenation is preserved and
carbon dioxide is removed. High-frequency oscillating ventilation allows maintenance
of high end-expiratory lung volume without overdistention of alveoli. The topic of high-
frequency oscillatory ventilation and ventilator-induced lung injury is addressed in an
article by Imai and Slutsky81 in Critical Care Medicine, 2005. Consistent data from
studies in neonates indicate the effectiveness of high-frequency oscillatory ventilation
combined with maneuvers designed to maintain lung inflation volumes using
recruitment maneuvers and PEEP adjustments. Minimal frequencies of ventilator-
associated lung injury were observed in these studies.

   The authors note that studies in adult patients with acute respiratory distress
syndrome have not shown consistent benefit from the use of high-frequency oscillatory
ventilation. Ideal strategies for maximizing pressure in the proximal and distal airways
and optimizing lung volumes with this ventilatory strategy have not yet been developed
and this limits application of this modality. Although there are data to suggest transient
improvement in oxygenation, there is no demonstration of improved mortality
outcomes and routine use of this modality is not recommended.

   The main objectives of lung-protective ventilation are to preserve adequate
oxygenation, maintain lung inflation, facilitate re-inflation of contracted or collapsed
alveoli, and minimize the risk of ventilator-associated lung injury. Continuous positive
airway pressure strategies are well suited to these objectives. The limitations of
continuous positive airway approaches include that use of these approaches requires
an alert; cooperative, spontaneously breathing patient and these approaches are
sometimes difficult to apply in intubated patients. In addition, a relatively small
proportion of ICU ventilators can deliver continuous positive airway pressure
efficiently. Thus, continuous positive airway pressure is most useful during the
“liberation” or “weaning” process as the patient is assisted through the transition from
ventilator support to normal breathing. One variant of continuous positive airway
pressure, airway pressure release ventilation, can be used in intubated patients. The
patient must be breathing spontaneously in order for this mode of ventilation to work
properly. In suitable patients, airway pressure release ventilation can maintain lung
inflation and recruit additional alveoli in the dependent areas of the lung during
spontaneous breathing intervals.

    An article describing airway pressure release ventilation is by Habashi82 in Critical
Care Medicine, 2005. This article is supplied as a full-text reprint with this issue of SRGS.
Habashi notes that airway pressure release ventilation was initially described in two
articles authored by Stock and Downs in 1987. This approach to ventilation uses
continuous positive airway pressure (Phigh) to maintain lung inflation for a preselected
interval (Thigh). Elimination of carbon dioxide is facilitated by scheduling periodic
releases of airway pressure that permit airway pressure to fall to a preselected level
(Plow). Low pressure is maintained for a preselected interval (Tlow) and carbon dioxide
is eliminated by this exhalation. Spontaneous, patient-generated breaths assist in
recruiting alveoli by supplying diaphragmatic contractions. The recruited alveoli are in
the dependent lung areas adjacent to the diaphragm. The author notes that the process
of alveolar recruitment proceeds along variable time courses because inflation of one
group of alveoli affects the inflation rate of neighboring alveoli. Recruitment, therefore,
proceeds in a wave or “avalanche” fashion. Maintenance of continuous positive airway
pressure assists in maintaining inflation as additional alveolar units open. Recruitment
occurs because of decreases in pleural pressure rather than increases in airway
pressure and Habashi emphasizes that the intermittent airway pressure releases also
work to prevent lung overdistention. To minimize de-recruitment, low pressure
intervals are kept as short as possible (0.2–0.8 seconds in adults).

   Airway pressure release ventilation contrasts with pure continuous positive airway
pressure breathing in that the work of breathing increases with continuous positive
airway pressure alone because of the need for the patient to expend energy to remove
carbon dioxide. In patients with decreased lung compliance and respiratory muscle
deconditioning, this increased work of breathing might not be tolerated by the patient.
Airway pressure release ventilation effectively addresses this problem. Habashi notes
that alveolar ventilation is intermittent while carbon dioxide delivery to the alveolus is
continuous. The intermittent pressure releases refreshed alveolar gas and re-
establishes the gradient for diffusion of carbon dioxide from blood to alveolar gas.

    Improvement of oxygenation during airway pressure release ventilation occurs
because of maintenance of high mean airway pressure that serves to increase the
number of ventilated, perfused alveoli. Spontaneous breaths during the high-pressure
interval serve to recruit additional alveoli in the dependent, perfused lung areas and
this mechanism assists in supporting oxygenation as well. The author notes that
resistance of the artificial airway during the early phase of the pressure release interval
provides airway resistance that effectively produces PEEP which also assists in
supporting oxygenation. Because of the PEEP that results from airway resistance, the
low pressure setting is preferably zero.

    Initial set-up of airway pressure release ventilation is accomplished depending on
whether the patient is newly intubated or whether this modality is being used to assist
in transition to weaning. Suggested set-up strategies are presented in Table 2 of
Habashi’s article. For example, an adult patient newly intubated would have a desired
plateau airway pressure selected (20-35 cmH20), and this would be the high-pressure
setting. Higher pressures might be required where combined lung and chest wall
compliance are reduced (obese patients). Low pressure would be set at zero. The high
time interval would be set at 4-6 seconds and the low time interval would be 0.2-0.8
seconds. Longer low-time intervals might be required in patients with chronic
obstructive lung disease. These settings would produce 10-12 exhalations/minute. In
patients who are transitioning from conventional ventilator support, the high pressure
is set at the prior ventilation mode plateau pressure.

    Habashi notes that airway pressure release ventilation is a useful ventilation mode
for spontaneously breathing patients who are ventilated in the prone position or in
kinetic beds (see discussion in SRGS, Vol. 35, No. 6). He also notes that addition of
pressure support ventilation to airway pressure release ventilation produces
unfavorable increases in transpulmonary pressure. The author notes that spontaneous
breathing is required for effective use of airway pressure release ventilation and,
therefore, this approach is not indicated in patients who require aggressive
sedation/analgesia or neuromuscular blockade. The modality is associated with less
patient discomfort from the use of adjuvant ventilation compared with conventional
ventilation. This consistent observation suggests that intervals of heavy sedation use or
neuromuscular blockade might be shortened by applying airway pressure release
ventilation. Habashi concludes by noting that weaning from airway pressure release
ventilation is a simple process involving reductions of the high pressure setting and
extension of the high pressure time interval. Decreasing the number of releases as the
pressure changes are made facilitates transition to normal patient breathing. Finally,
Habashi notes that this modality can be applied using noninvasive ventilation

    Comparative clinical data documenting the benefit of airway pressure release
ventilation are found in a review by Siau and Stewart83 in Clinics in Chest Medicine,
2008. These authors note that clinical series evaluating airway pressure release
ventilation have been retrospective observational studies or comparative studies
employing historic controls. These studies suggest a reduction of mortality with the use
of this modality in traumatic lung injury patients. There has been little control of
confounding variables in these analyses and, because of this, a definite reduction in
mortality cannot be assumed. Reductions in ICU lengths of stay and ventilator intervals
have been reported. Direct comparisons of airway pressure release ventilation to lung-
protective ventilation have not been reported. Siau and Stewart conclude that airway
pressure release ventilation is appropriate for carefully selected patients but a
recommendation for widespread use of this approach cannot be made based on current

    One area where airway pressure release ventilation might be valuable is in patients
who need a transition between conventional ventilation and implementation of a
formal “weaning” protocol. Weaning from ventilator support is clinically challenging.
During full ventilator support respiratory muscle deconditioning occurs and, because of
this, muscle weakness might limit ventilatory effort. Lung and chest wall compliance are
decreased by the primary lung disease, body habitus (obesity), pain from incisions,
chest drainage tubes, and rib fractures. Successful weaning requires that the primary
disease causing acute respiratory distress syndrome be under control. In addition, the
patient should be capable of initiating spontaneous breathing efforts. Ideally, use of
sedation and analgesia are reduced to the point that the patient can cough, make deep
breathing efforts, and participate in patient care by changing position in the bed or
moving from bed to chair with assistance. Correction of nutritional deficits should be
underway. Once these conditions are met, transition to an assisted-ventilation strategy
is a first step. Spontaneous breathing trials can be scheduled three or four times daily
under nursing and/or respiratory therapist supervision, and intervals of spontaneous
breathing can be incrementally increased (“wind sprints”). When extubation is possible,
noninvasive interfaces can be used to assist patients with “graduation” to normal
breathing. These approaches are particularly useful in patients with COPD.

   A systematic review of available data on the use of noninvasive ventilation as a
weaning adjunct appears in an article by Burns and coauthors84 in British Medical
Journal, 2009. These authors reviewed 12 trials involving 530 patients. They note that
most patients enrolled in weaning trials using noninvasive ventilation had COPD. COPD
was not necessarily the main contributor to the need for ventilator support in the
   reported trials. The authors note that pooled data suggest a reduction in mortality and
   ICU length of stay for patients with COPD weaned with noninvasive ventilation
   protocols. There was no increased risk of weaning failure, pneumonia, or reintubation
   in the reported trials. These authors conclude that the evidence in support of
   noninvasive ventilation as a means of facilitating weaning is sufficiently strong to
   recommend this modality in patients with COPD.

Editorial comment
       From the perspective of the editor, weaning critically ill surgical patients from
   mechanical ventilation is highly dependent on the success of efforts to control the
   process that led to the need for ventilation in the first place. Deconditioning is an
   especially challenging problem that limits success of weaning in the elderly and in
   patients with severe comorbid conditions. Failure of weaning with deterioration of
   oxygenation and lung compliance leading to reinstitution of ventilation is a high price
   the patient pays for suboptimal timing of weaning. Weaning failure and extubation
   failure resulting in the need to reinstitute mechanical ventilation are, in my view,
   largely avoidable complications. Underestimation of the need for analgesia/sedation,
   underuse of assistive exercise physical therapy programs, inadequate patient
   counseling, delay or nonuse of nutritional support, and suboptimal level of
   consciousness are common modifiable factors contributing to weaning failure.
   Education of nurses, physical therapists, and respiratory therapists in the use of
   weaning protocols permits weaning and extubation when criteria for acceptable
   patient-controlled breathing are fulfilled. Use of these multidisciplinary protocols can
   improve the success rates for weaning in critically ill surgical patients.

Nonventilator adjunctive measures
       Several adjuncts to traditional ventilator therapy might improve outcomes of acute
   lung injury and acute respiratory distress syndrome in carefully selected patient
   groups. These adjuncts can be categorized as 1) measures for reducing the risk of
   additional lung injury (fluid therapy); 2) measures that modulate the inflammatory
   process (corticosteroid therapy); 3) interventions that modify pulmonary hypertension
   (nitric oxide); 4) measures designed to improve distribution of ventilation and
   perfusion (prone positioning, discussed in SRGS, Vol. 35, No. 6); 5) adjuncts that replace
   lung function (extracorporeal membrane oxygenation); and 6) interventions that alter
   the airway (tracheostomy). In this section of the overview, we review these topics.

       The first article reviewed appeared in the New England Journal of Medicine in 200685
   entitled “Comparison of two fluid-management strategies in acute lung injury.” This
   article is supplied as a full-text reprint with this issue of SRGS. The article by Wiedemann
   and coauthors is a report of a randomized, prospective trial comparing two approaches
   to fluid management in patients with acute respiratory distress syndrome. The study
was conducted by the ARDSNet group of investigators. One thousand patients were
enrolled and randomly assigned to a conservative or liberal fluid management group.
The characteristics of fluid management for each enrolled patient were determined in
real time depending on the central venous pressure or pulmonary artery pressure, the
presence of shock (arterial pressure < 60 mmHg), and external signs of inadequate
perfusion, such as skin mottling or oliguria. Thus, a patient with effective circulation
and no oliguria who had a central venous pressure of >13 and was assigned to the
conservative fluid group would receive furosemide and intravenous fluids at a
minimum rate until central venous pressure was in the 9-13 cm H20 range.

    Patients in shock were treated with fluids and vasoactive agents as needed until
appropriate hemodynamic response and venous or pulmonary artery target pressures
were achieved. All patients enrolled met standard criteria for acute respiratory distress
syndrome and all were ventilated with standard ventilator strategies. The investigators
found that there was no statistically significant difference in mortality in the two groups
of patients. There were significant improvements in oxygenation in the conservative
strategy group. Ventilator days and ICU stay were reduced in the conservative group
and there was no increase in the frequency of renal insufficiency or the diagnosis of
shock in the conservative therapy group.

    In the discussion section, data indicate that small increases in pulmonary artery
occlusion pressure above the normal range can be associated with large increases in
extravascular lung water. The authors also cite studies indicating that removal of excess
interstitial space fluid with furosemide is associated with improved oxygenation.
Supplemental albumin infusion given to hypoalbuminemic patients to improve oncotic
pressure did not result in improved oxygenation unless furosemide was given along
with the colloid infusion. These results suggest that there is a “mobilizable” fluid space
within the lung amenable to movement of fluid along pressure gradients but that
protein leakage might limit the effectiveness of efforts to improve oncotic pressure. The
lack of mortality difference in this study most likely relates to the fact that patients who
died were actually dying of an associated illness, and not from acute respiratory

   Recognition of the role of inflammation in the pathophysiology of acute lung injury
has stimulated evaluation of anti-inflammatory strategies for treatment of acute
respiratory distress syndrome. This topic is the subject of a report by Tang and
coauthors86 in Critical Care Medicine, 2009, who report a meta-analysis of randomized
controlled trials and observational studies of low-dose corticosteroids (0.5-2.5
mg/kg/day) in patients with acute respiratory distress syndrome. For cohort studies,
analyses of data drawn from 307 patients were included. Randomized trials included
341 patients. Both types of studies demonstrated improved mortality risk and both
types of studies demonstrated improved oxygenation and decreased ICU length of stay.
Overall, there was a 38% reduction in risk of death for patients treated with low dose
methylprednisolone. There was no increased risk of infection, neuromyopathy, or major
complications in the steroid treated groups.

    The authors note that their study effectively deals with the challenges faced by other
investigators who attempted to determine whether there was a benefit to
corticosteroid treatment with no increase in complications. Earlier analyses dealt with
widely varying dosage ranges, differing types of steroid drugs, and heterogeneous
patient groups. This study dealt with studies using low-dose medication with standard
definitions of acute respiratory distress syndrome and standard reporting of outcomes.
The authors conducted subgroup analysis that indicated efficacy of low-dose
corticosteroids even when treatment was started several days after the onset of acute
respiratory distress syndrome. In addition, treatment effect was independent of the use
of “open lung” ventilation strategies. The analysis also confirmed that the treatment
effect was independent of any affect on the outcomes of sepsis. The authors conclude
that low-dose steroid treatment is effective and safe in patients with acute respiratory
distress syndrome. The data indicate that steroid therapy should be tapered and not
stopped abruptly. Abrupt cessation of steroids can be associated with rebound
inflammation and worsening of lung function. Tang and associates acknowledge that
this study is limited by the fact that they had no knowledge of the presence or absence
of dysfunction of the pituitary-adrenal axis in these studies.

    An approach to the diagnosis of adrenal insufficiency in critically ill patients is the
topic of a consensus report by Marik and coauthors87 in Critical Care Medicine, 2008.
This article is supplied as a full-text reprint with this issue of SRGS These authors
recommend that the diagnosis of adrenal insufficiency in critically ill patients can be
established by documenting an increase of less than 9 μg/dL of total serum cortisol
after a dose of adrenocorticotrophic hormone of 250 μg or a random total serum
cortisol level of < 10 μg/dL. Once the diagnosis is established, treatment with
corticosteroid replacement is valuable especially in patients with septic shock and
inadequate responses to fluids and vasoactive drugs. Steroid therapy is useful in the
treatment of early acute respiratory distress syndrome confirming the observations of
Tang and coauthors, discussed earlier. Readers should note that the treatment benefit
observed by Tang and colleagues was not limited to early acute respiratory distress

    Inhaled nitric oxide has potent pulmonary vasodilating properties and, according to
some data, anti-inflammatory features also. Thus, this agent has been suggested as a
means of improving ventilation/perfusion matching in patients with acute respiratory
distress syndrome. A meta-analysis of data pertinent to this topic appears in an article
by Adhikari and coauthors88 in the British Medical Journal, 2007; analyzed data from 12
trials of acceptable quality that had enrolled more than 1200 patients. The data
disclosed modest, transient, improvements in oxygenation but no affect on mortality
from acute respiratory distress syndrome. In addition, patients receiving nitric oxide
incurred a significant increased risk of renal dysfunction. The authors concluded that
available data do not support a role for inhaled nitric oxide in the treatment of acute
respiratory distress syndrome.

    Another adjunctive therapy applied to patients with acute respiratory distress
syndrome is extracorporeal membrane oxygenator support. Using this support device
involves connecting a membrane oxygenation device and a heat exchanger to the
patient using a venovenous circuit or, in patients with hemodynamic instability, a
venoarterial circuit. Limited anticoagulation is required for the circuit to function.
Blood is drawn from a central venous or arterial source and returned via a second
venous access after oxygenation and warming. This device has been primarily used as
“rescue” therapy for patients who cannot be adequately ventilated. Intuitively, the best
results should be obtained in patients with isolated lung damage caused by
noninfectious etiologies. For these reasons, patients with direct pulmonary injury
would be a patient group where the device would probably achieve the best results.

    A clinical series reporting results of extracorporeal membrane oxygenator usage in
patients with multiple injuries is reported in an article by Cordell-Smith and
coauthors89 in Injury, 2005. These authors report a series of 28 patients who received
extracorporeal membrane oxygenator support for severe acute respiratory distress
syndrome developing after direct pulmonary traumatic injury or after multiple trauma
(mostly pelvic and long bone fractures). Because of the need for anticoagulation, use of
this rescue approach would be limited in patients with central nervous system injuries
or intraabdominal injuries. The authors note that the duration of support in this patient
group was, on average, 141 hours. This support interval is shorter than intervals of
support used in other patient groups. Twenty of the 28 patients survived. Nonsurvival
was noted more often in patients with systemic sepsis or pulmonary infection.

   The report does not detail the characteristics of the ventilator therapy used in these
patients and it is, therefore, not completely clear whether the use of extracorporeal
membrane oxygenation was to “rescue” patients or not. This mortality rate is somewhat
lower than the mortality reported for patient groups containing both surgery and
trauma patients treated with aggressive “open-lung” ventilation strategies, but the
patient numbers in this report are small and the process of selecting the patients for
therapy with the external device is not described in detail. Nonetheless, trauma patients
might represent a favorable group for use of extracorporeal membrane oxygenator
    Long-term results of extracorporeal membrane oxygenator support are important
because these data provide insight into the process of lung healing and offer the
opportunity to assess potential chronic adverse effects of extracorporeal support. An
article providing data relevant to this topic, by Linden and coauthors,90 appeared in
Acta Anaesthesiologica Scandinavica, 2009. These authors report results of high-
resolution lung CT scans, extensive pulmonary function tests, lung scintigraphy, and
lung-specific quality of life questionnaire responses in a group of 21 patients who
survived severe acute respiratory distress syndrome treated with extracorporeal
membrane oxygenator support. During extracorporeal membrane oxygenator support
episodes, patients were maintained on low-level continuous positive airway pressure
ventilation. Clinical assessments were performed at least one year after therapy in all
patients. High-resolution CT images disclosed changes consistent with lung fibrosis in
all patients, but the extent of the changes was limited and the distribution of CT changes
was not the typical anterior distribution of ventilator induced lung injury. Pulmonary
function tests showed abnormal carbon dioxide diffusing capacity in nearly two-thirds
of the patients tested. The abnormality was small, however, and functional impairment
was mild. Overall, pulmonary function tests were within the normal range. Pulmonary
scintigraphy showed residual airway obstructive patterns in all patients characterized
by prolonged washout intervals for the inhaled radioisotope. Exercise testing was
performed; reduced exercise tolerance was seen in one-third of patients, but the
limitation was leg fatigue rather than pulmonary symptoms in all of these patients.

   All the patients responded to the quality of life questionnaire and all stated that
quality of life was reduced after treatment with extracorporeal membrane oxygenator
support. Importantly, none of the patients required supplemental oxygen and all were
employed full-time in the same occupations held before their illness. The authors
concluded that long-term impairment after extracorporeal membrane oxygenator
support is usually mild.

    Tracheostomy is an adjunctive treatment used in patients with acute respiratory
distress syndrome. The objective of tracheostomy use is mainly to improve patient
comfort, permit speech, reduce the risk for laryngeal injury, and optimize airway access.
Available data suggest that tracheostomy facilitates discharge of patients from the ICU,
shortens ventilator intervals, and, possibly, reduces the frequency and severity of
ventilator-associated pneumonia. These potential benefits are accompanied by costs
and complications.

   Tracheostomy, at a minimum, results in disfiguring scarring of the anterior neck.
Airway bleeding, tracheal ring fracture, tracheal stenosis, and esophageal injury might
occur. Tracheal-innominate artery fistula is a complication that has nearly disappeared
with reductions in post-tracheostomy local wound infection and appropriate choice of
the level of tracheostomy tube insertion (third tracheal ring). Bedside percutaneous
tracheostomy techniques have reduced the need to transfer patients to the operating
room for formal surgical procedures. Overall reductions in healthcare resource use
have been reported with the use of bedside percutaneous tracheostomy; bedside
tracheostomy requires partial removal of the endotracheal tube and use of the flexible
fiberoptic bronchoscope to guide tracheostomy tube insertion. Thus, additional costs
are incurred along with the risk for sudden airway loss during the procedure. In this
section of the overview, several articles dealing with the use of tracheostomy as an
adjunct to other treatments for acute respiratory distress syndrome are discussed.

    The first article discussed is by De Leyn and coauthors.91 It appeared in the
European Journal of Cardio-Thoracic Surgery, 2007, and is supplied as a full-text reprint
accompanying this issue of SRGS. It reports practice guidelines for the use of
tracheostomy developed by a joint committee of the Belgian Society of Pneumonology
and the Belgian Association for Cardiothoracic Surgery. The process of guideline
development is described in the article. This process included collection and evaluation
of peer-reviewed articles, discussion in committee meetings, posting of proposed
guidelines online for comment, and final promulgation of the guidelines.

    The authors begin by listing the indications for tracheostomy, which include long-
term ventilation, failure to wean, upper airway obstruction, and copious secretions.
They also list contraindications such as active soft tissue infection in the anterior neck,
and extensive scarring from earlier surgical procedures and/or radiation therapy.
Current approaches to the care of patients with cervical spine injury might include early
open reduction and internal fixation of spinal fractures. The presence of a fresh surgical
incision with implanted devices is a relative contraindication to tracheostomy. The
authors next discuss technique for open tracheostomy. If possible, the patient is
positioned with the neck extended. The conventional approach is to make a transverse
or vertical skin incision 1 cm below the lower border of the cricoid cartilage. Soft
tissues are separated and, if necessary, the thyroid isthmus is divided and retracted.
The anterior tracheal wall is identified and the second and third tracheal rings are
located. The endotracheal tube is withdrawn until the distal orifice is above the
tracheostomy site. The trachea is entered and the opening is dilated. Incision of the
tracheal ring above or below the opening is sometimes necessary to facilitate insertion
of the tracheostomy appliance.

   The authors recommend using a tracheostomy appliance with a low-pressure cuff.
Lubrication of the tracheostomy appliance facilitates insertion because the low-
pressure cuff is redundant and might “hang-up” on the tracheal rings. Advance
preparation should be made to connect the airway circuit to the tracheostomy
appliance immediately on completion of successful insertion. Some surgeons, including
the editor, prefer to perform immediate fiberoptic bronchoscopy to make certain that
blood clots and mucous plugs are cleared from the airway and that there is no residual
airway bleeding.

    De Leyn and colleagues discuss percutaneous dilational tracheostomy. The two most
commonly used devices are the “Blue Rhino®” device and the “Percu-twist®” device.
Each device uses a percutaneous needle for tracheal access, and optimal patient safety
concerns have dictated the use of flexible fiberoptic bronchoscopic control of the
procedure so that the point of needle entry and tracheostomy device placement is
documented. The main difference between the two devices lies in the means of dilation
of the skin, subcutaneous tissue, and tracheal wall channel into the tracheal lumen. The
Blue Rhino device uses a curved, hydrophilic-coated dilator, which needs to be passed
through the channel a minimum of three passages using the previously placed
guidewire. The Percu-twist device uses a hydrophilic-coated screw that is rotated in a
clockwise direction to create the entry channel for the tracheostomy device.

    The authors list and discuss early and late complications of both surgical and
percutaneous dilational tracheostomy. They note that conventional wound care
approaches are necessary to reduce the risk of peri-tracheostomy infection. In addition,
they note that the swallowing dysfunction that accompanies tracheostomy usually
means that patients will not be able to eat normally. Aspiration episodes are frequent
and collection of secretions above the tracheostomy balloon must be anticipated. If
ventilator support can be interrupted, patients might be able to speak if the
tracheostomy tube orifice is covered with a gloved finger. Tracheostomy appliance
balloons tend to increase in volume over time with concomitant increases in balloon
pressure. Pressures in excess of 25 mmHg might interrupt tracheal mucosal blood flow.
Tracheostomy cuff pressure should be monitored to keep pressures in the appropriate

    A comparison of surgical tracheostomy to percutaneous tracheostomy is the topic of
a report by Beltrame and coauthors92 in Minerva Anesthesiologica, 2008. These authors
report an analysis comparing surgical tracheostomy to percutaneous tracheostomy.
Three hundred sixty-seven patients undergoing percutaneous tracheostomy were
compared with 161 historic control patients who had surgical tracheostomy. Procedure
duration was shorter for percutaneous tracheostomy. Complications were low and
equivalent for both techniques. ICU length of stay was shorter for percutaneous
tracheostomy patients; there is no report of concomitant changes in ICU patient care
processes that might have worked to shorten ICU stay in the later group. The authors
cite data showing that patients with tracheostomy might require less analgesia and
sedation compared to patients with endotracheal tubes. The authors note that
analgesia/sedation protocols are not standardized in most reports so it is not possible
to determine whether intubated patients are simply oversedated.

    De Leyn and colleagues note there is controversy over the timing of tracheostomy
and the impact of tracheostomy on outcomes of ventilation for acute respiratory
distress syndrome. Observational studies and one randomized clinical trial
demonstrating reduced mortality, intensive care length of stay, and reduced frequency
of pneumonia are cited by these authors. The numbers of patients in these studies are
small and, because of this, categorical statements of benefit from early tracheostomy
cannot be made. They cite, in addition, a systematic review of early tracheostomy in
trauma patients that did not demonstrate clear evidence of benefit from early
tracheostomy. Much of the debate centers on the definition of “early tracheostomy.”
Studies that do not show benefit usually report tracheostomy performed within the first
10 days of ventilation; studies showing benefit report tracheostomy performed within
the first 48 hours of ventilation. Obviously, both approaches are subject to selection
bias since prediction of outcomes within the first 48 hours is challenging and, in the
reported studies, some patients subjected to early tracheostomy are weaned from the
ventilator within 3-4 days of the tracheostomy. This subgroup of patients could possibly
have been weaned to extubation without the tracheostomy. Patients who have
tracheostomy at the end of the first week of ventilation are probably patients who are
going to require prolonged support regardless of the approach to airway management.

    A comparison of “early” versus “late” tracheostomy in injured patients is the focus of
a report by Arabi and coauthors93 in Critical Care, 2004. These authors queried a
prospectively managed trauma ICU database and compared a group of 29 patients who
had tracheostomy performed earlier to day 7 of ventilation with 107 patients who had
tracheostomy performed after 7 days. Lower ICU lengths of stay were noted in the
patients who had “early” tracheostomy. The authors noted that the patients having
early tracheostomy were more likely to have severe brain injury and this raises the
question whether early tracheostomy was associated with benefit in terms of improved
outcomes for respiratory failure or, rather, the tracheostomy facilitated transfer of
brain-injured patients to other care areas. Overall hospital outcomes was not influenced
by early tracheostomy in this study.

    A report of an analysis of the affects of early tracheostomy versus late tracheostomy
on patient outcomes from a statewide trauma database is the topic of a report by
Schauer and coauthors94 in the Journal of Trauma, 2009. These authors reviewed
patient data on 685 patients who underwent tracheostomy. Early tracheostomy was
defined as tracheostomy performed within the first four days after injury. The authors
calculated survival probability using standard injury severity indices. Low survival
probability was defined as a probability of survival of less than 25%. The authors noted
   there was high early mortality in the patients in the low survival probability group and
   this group of patients did not benefit from early tracheostomy. In patients with survival
   probability of more than 25%, early tracheostomy resulted in shorter ICU lengths of
   stay and shorter hospital lengths of stay.

       With the increased use of percutaneous dilational tracheostomy, patients referred
   to surgeons for formal surgical tracheostomy are often patients with very obese necks,
   prior neck scarring, patients who cannot be optimally positioned, and patients with
   difficult upper airway anatomy and/or history of difficult intubation. These changes in
   patient characteristics mean that surgical tracheostomy usually means a formal
   procedure performed under general anesthesia in the operating room. Patients will
   frequently not have optimum ventilator support during transport and in the operating
   room (see discussion of ventilatory associated pneumonia in SRGS, Vol. 35 No. 6).
   Careful coordination of the surgical and anesthesia teams is necessary because of the
   hazard of airway loss. Surgical exposure is often challenging and technical measures to
   minimize the risk of bleeding include positioning the patient in a slightly head-up
   position (to lower venous pressure in neck veins), ventilation of the patient using low
   airway pressures, and the use of larger incisions.

Editorial comment
       From the perspective of the editor, it seems clear that early tracheostomy facilitates
   the care of certain types of injured patients. This patient group consists mainly of
   patients with moderate-to-severe brain injury who can be expected to survive and,
   perhaps, recover brain function over time. Tracheostomy can facilitate transfer of such
   patients to rehabilitation facilities or to assisted care facilities. The reductions in ICU
   lengths of stay and hospital lengths of stay reported in the articles discussed above
   supports this conclusion. This patient group is also at increased risk for pneumonia and
   tracheostomy might improve the diagnostic process for pneumonia. By reducing
   aspiration of secretions, pneumonia risk might be reduced. The reasons for decreased
   length of ventilator support reported by some authors in patients undergoing
   tracheostomy (especially early tracheostomy) are not clear. Reductions in airway-
   related complications and pneumonia risk could contribute to such reductions.

      The available data, and clinical experience, have shown that provision of consistent
   open-lung ventilation (discussed in the previous section) with minimal interruption of
   ventilation will maximize the likelihood of weaning and recovery of lung function as
   long as the process producing acute respiratory distress syndrome can be adequately
   controlled and ventilator-associated pneumonia can be avoided.

       In the editor’s experience, clinical judgment can usually determine, with acceptable
   accuracy, those patients who will likely require prolonged ventilator support. In this
   patient group, early tracheostomy is likely to have its greatest benefit. In my view,
   tracheostomy facilitates the diagnosis of ventilator-associated pneumonia using
   bronchoalveolar lavage. I have consistently observed less use of supine positioning and
   more frequent suctioning of patients with tracheostomies. Patients with tracheostomies
   are repositioned in bed more frequently and are easier to move from bed to chair
   compared with patients who have endotracheal tubes. These features might contribute
   to lowering of pneumonia risk.

       Percutaneous tracheostomy is the most efficient procedure and this approach is, in
   the experience of the editor, associated with the shortest interval of interruption of
   ventilator therapy. As higher risk patients have increasingly been selected for open
   tracheostomy under general anesthesia in the operating room, maneuvers to safely
   place the tracheostomy when exposure is limited have been sought.

One valuable maneuver, in the experience of the editor, is to use the percutaneous
dilational tracheostomy insertion equipment to assist in placing the tracheostomy
appliance once the anterior tracheal wall is identified. The needle and guidewire are
inserted using bronchoscopic control and the dilator is used to establish entry into the
tracheal lumen. Stay sutures are placed on each side of the trachea entry site so that the
trachea can be elevated. The tracheostomy appliance (sometimes a long appliance is
necessary and advanced planning to make certain such devices are available is helpful) can
then be placed safely.

1.    Davenport DL, Ferraris VA, Hosokawa P, et al. Multivariable predictors of postoperative cardiac
      adverse events after general and vascular surgery: results from the patient safety in surgery
      study. J Am Coll Surg 2007; 204(6):1199-210.
2.    Poldermans D, Hoeks SE, Feringa HH. Preoperative risk assessment and risk reduction before
      surgery. J Am Coll Cardiol 2008; 51(20):1913-24.
3.    Sander O, Welters ID, Foex P, et al. Impact of prolonged elevated heart rate on incidence of
      major cardiac events in critically ill patients with a high risk of cardiac complications. Crit Care
      Med 2005; 33(1):81-8; discussion 241-2.
4.    Anselmi A, Possati G, Gaudino M. Postoperative inflammatory reaction and atrial fibrillation:
      simple correlation or causation? Ann Thorac Surg 2009; 88(1):326-33.
5.    Laitio T, Jalonen J, Kuusela T, et al. The role of heart rate variability in risk stratification for
      adverse postoperative cardiac events. Anesth Analg 2007; 105(6):1548-60.
6.    Hanss R, Block D, Bauer M, et al. Use of heart rate variability analysis to determine the risk of
      cardiac ischaemia in high-risk patients undergoing general anaesthesia. Anaesthesia 2008;
7.    Cuthbertson BH, Amiri AR, Croal BL, et al. Utility of B-type natriuretic peptide in predicting
      perioperative cardiac events in patients undergoing major noncardiac surgery. Br J Anaesth
      2007; 99(2):170-6.
8.    Mahla E, Baumann A, Rehak P, et al. N-terminal pro-brain natriuretic peptide identifies patients
      at high risk for adverse cardiac outcome after vascular surgery. Anesthesiology 2007;
9.    Jaroszewski DE, Huh J, Chu D, et al. Utility of detailed preoperative cardiac testing and incidence
      of post-thoracotomy myocardial infarction. J Thorac Cardiovasc Surg 2008; 135(3):648-55.
10.   Chopra V, Plaisance B, Cavusoglu E, et al. Perioperative beta-blockers for major noncardiac
      surgery: Primum Non Nocere. Am J Med 2009; 122(3):222-9.
11.   Beattie WS, Wijeysundera DN, Karkouti K, et al. Does tight heart rate control improve beta-
      blocker efficacy? An updated analysis of the noncardiac surgical randomized trials. Anesth Analg
      2008; 106(4):1039-48, table of contents.
12.   Suttner S, Boldt J, Mengistu A, et al. Influence of continuous perioperative beta-blockade in
      combination with phosphodiesterase inhibition on haemodynamics and myocardial ischaemia in
      high-risk vascular surgery patients. Br J Anaesth 2009; 102(5):597-607.
13.   Dunkelgrun M, Boersma E, Schouten O, et al. Bisoprolol and fluvastatin for the reduction of
      perioperative cardiac mortality and myocardial infarction in intermediate-risk patients
      undergoing noncardiovascular surgery: a randomized controlled trial (DECREASE-IV). Ann Surg
      2009; 249(6):921-6.
14.   Mackey WC, Fleisher LA, Haider S, et al. Perioperative myocardial ischemic injury in high-risk
      vascular surgery patients: incidence and clinical significance in a prospective clinical trial. J Vasc
      Surg 2006; 43(3):533-8.
15.   Burke AP, Virmani R. Pathophysiology of acute myocardial infarction. Med Clin North Am 2007;
      91(4):553-72; ix.
16.   Walsh SR, Tang TY, Sadat U, et al. Remote ischemic preconditioning in major vascular surgery. J
      Vasc Surg 2009; 49(1):240-3.
17.   Wilansky S, Moreno CA, Lester SJ. Complications of myocardial infarction. Crit Care Med 2007;
      35(8 Suppl):S348-54.
18.   Lim W, Qushmaq I, Cook DJ, et al. Reliability of electrocardiogram interpretation in critically ill
      patients. Crit Care Med 2006; 34(5):1338-43.
19.   Engel TR. Electrocardiographic diagnosis of coronary syndromes in the critical care unit. Crit Care
      Med 2006; 34(5):1546-7.
20.   Fox DJ, Tischenko A, Krahn AD, et al. Supraventricular tachycardia: diagnosis and management.
      Mayo Clin Proc 2008; 83(12):1400-11.
21.   Marill KA, Wolfram S, Desouza IS, et al. Adenosine for wide-complex tachycardia: efficacy and
      safety. Crit Care Med 2009; 37(9):2512-8.
22.   Anugwom C, Sulangi S, Dachs R. Adenosine vs. calcium channel blockers for supraventricular
      tachycardia. Am Fam Physician 2007; 75(11):1653-4.
23.   Siu CW, Lau CP, Lee WL, et al. Intravenous diltiazem is superior to intravenous amiodarone or
      digoxin for achieving ventricular rate control in patients with acute uncomplicated atrial
      fibrillation. Crit Care Med 2009; 37(7):2174-9; quiz 2180.
24.   Karth GD, Heinz G. Atrial fibrillation: how to slow the pace? Crit Care Med 2009; 37(7):2309-10.
25.   Zebis LR, Christensen TD, Thomsen HF, et al. Practical regimen for amiodarone use in preventing
      postoperative atrial fibrillation. Ann Thorac Surg 2007; 83(4):1326-31.
26.   Allen M. Pacemakers and implantable cardioverter defibrillators. Anaesthesia 2006; 61(9):883-
27.   Chatterjee K, Rame JE. Systolic heart failure: chronic and acute syndromes. Crit Care Med 2008;
      36(1 Suppl):S44-51.
28.   Ashrafian H, Frenneaux MP, Opie LH. Metabolic mechanisms in heart failure. Circulation 2007;
29.   Masoudi FA, Inzucchi SE. Diabetes mellitus and heart failure: epidemiology, mechanisms, and
      pharmacotherapy. Am J Cardiol 2007; 99(4A):113B-132B.
30.   Mitchell JE. Emerging role of anemia in heart failure. Am J Cardiol 2007; 99(6B):15D-20D.
31.   Gerber DR. Transfusion of packed red blood cells in patients with ischemic heart disease. Crit
      Care Med 2008; 36(4):1068-74.
32.   Kumar R, Gandhi SK, Little WC. Acute heart failure with preserved systolic function. Crit Care
      Med 2008; 36(1 Suppl):S52-6.
33.   Hoit BD. Left ventricular diastolic function. Crit Care Med 2007; 35(8 Suppl):S340-7.
34.   Glassberg H, Kirkpatrick J, Ferrari VA. Imaging studies in patients with heart failure: current and
      evolving technologies. Crit Care Med 2008; 36(1 Suppl):S28-39.
35.   Omland T. Advances in congestive heart failure management in the intensive care unit: B-type
      natriuretic peptides in evaluation of acute heart failure. Crit Care Med 2008; 36(1 Suppl):S17-27.
36.   Petersen JW, Felker GM. Inotropes in the management of acute heart failure. Crit Care Med
      2008; 36(1 Suppl):S106-11.
37.   Kale P, Fang JC. Devices in acute heart failure. Crit Care Med 2008; 36(1 Suppl):S121-8.
38.   Pirracchio R, Cholley B, De Hert S, et al. Diastolic heart failure in anaesthesia and critical care. Br
      J Anaesth 2007; 98(6):707-21.
39.   Ramsay PT, Maxwell RA. Advancements in cardiopulmonary resuscitation: increasing circulation
      and improving survival. Am Surg 2009; 75(5):359-62.
40.   Ali B, Zafari AM. Narrative review: cardiopulmonary resuscitation and emergency cardiovascular
      care: review of the current guidelines. Ann Intern Med 2007; 147(3):171-9.
41.   Ehlenbach WJ, Barnato AE, Curtis JR, et al. Epidemiologic study of in-hospital cardiopulmonary
      resuscitation in the elderly. N Engl J Med 2009; 361(1):22-31.
42.   Ristagno G, Tang W, Chang YT, et al. The quality of chest compressions during cardiopulmonary
      resuscitation overrides importance of timing of defibrillation. Chest 2007; 132(1):70-5.
43.     Cabrini L, Beccaria P, Landoni G, et al. Impact of impedance threshold devices on
        cardiopulmonary resuscitation: a systematic review and meta-analysis of randomized controlled
        studies. Crit Care Med 2008; 36(5):1625-32.
44.     Ristagno G, Tang W, Huang L, et al. Epinephrine reduces cerebral perfusion during
        cardiopulmonary resuscitation. Crit Care Med 2009; 37(4):1408-15.
45.     Mentzelopoulos SD, Zakynthinos SG, Tzoufi M, et al. Vasopressin, epinephrine, and
        corticosteroids for in-hospital cardiac arrest. Arch Intern Med 2009; 169(1):15-24.
46.     Ewy GA, Kern KB. Recent advances in cardiopulmonary resuscitation: cardiocerebral
        resuscitation. J Am Coll Cardiol 2009; 53(2):149-57.
47.     Davis DP. Cardiocerebral resuscitation: a broader perspective. J Am Coll Cardiol 2009; 53(2):158-
48.     Schneider A, Bottiger BW, Popp E. Cerebral resuscitation after cardiocirculatory arrest. Anesth
        Analg 2009; 108(3):971-9.
49.     Thiagarajan RR, Brogan TV, Scheurer MA, et al. Extracorporeal membrane oxygenation to
        support cardiopulmonary resuscitation in adults. Ann Thorac Surg 2009; 87(3):778-85.
50.     Chan PS, Khalid A, Longmore LS, et al. Hospital-wide code rates and mortality before and after
        implementation of a rapid response team. JAMA 2008; 300(21):2506-13.
51.     Gonzalez MM, Berg RA, Nadkarni VM, et al. Left ventricular systolic function and outcome after
        in-hospital cardiac arrest. Circulation 2008; 117(14):1864-72.
52.     Peberdy MA, Ornato JP, Larkin GL, et al. Survival from in-hospital cardiac arrest during nights
        and weekends. JAMA 2008; 299(7):785-92.
53.     Edelson DP, Litzinger B, Arora V, et al. Improving in-hospital cardiac arrest process and outcomes
        with performance debriefing. Arch Intern Med 2008; 168(10):1063-9.
54.     Dine CJ, Gersh RE, Leary M, et al. Improving cardiopulmonary resuscitation quality and
        resuscitation training by combining audiovisual feedback and debriefing. Crit Care Med 2008;
55.     Johnson RG, Arozullah AM, Neumighter L, et al. Multivariable predictors of postoperative
        respiratory failure after general and vascular surgery: results from the patient safety in surgery
        study. J Am Coll Surg 2007; 204(6):1188-98.
56.     Gali B, Whalen FX, Schroeder DR, et al. Identification of patients at risk for postoperative
        respiratory complications using a preoperative obstructive sleep apnea screening tool and
        postanesthesia care assessment. Anesthesiology 2009; 110(4):869-77.
57.     Tsai WH, Remmers JE, Brant R, et al. A decision rule for diagnostic testing in obstructive sleep
        apnea. Am J Respir Crit Care Med 2003; 167(10):1427-32.
58.     Flemons WW, Whitelaw WA, Brant R, et al. Likelihood ratios for a sleep apnea clinical prediction
        rule. Am J Respir Crit Care Med 1994; 150(5 Pt 1):1279-85.
59.     Hallowell PT, Stellato TA, Petrozzi MC, et al. Eliminating respiratory intensive care unit stay after
        gastric bypass surgery. Surgery 2007; 142(4):608-12; discussion 612 e1.
60.     Gore DC. Preoperative maneuvers to avert postoperative respiratory failure in elderly patients.
        Gerontology 2007; 53(6):438-44.
61.     Westwood K, Griffin M, Roberts K, et al. Incentive spirometry decreases respiratory
        complications following major abdominal surgery. Surgeon 2007; 5(6):339-42.
62.     Dronkers J, Veldman A, Hoberg E, et al. Prevention of pulmonary complications after upper
abdominal surgery by preoperative intensive inspiratory muscle training: a randomized controlled pilot
study. Clin Rehabil 2008; 22(2):134-42.