1 TRANSFUSION AND SHOCK
2 FUNDAMENTALS OF MANAGEMENT IN SURGICAL PATIENTS
Page 1 of 119
Transfusion practices and shock are the topics chosen for this issue of Selected Readings
in General Surgery (SRGS). In the section dealing with transfusion practices we focus on risks
of transfusion, methods for salvaging blood and reducing transfusion volume, and man-
agement of the patient who refuses blood and/or blood products. We then move on to a
discussion of shock. Classifications of shock have evolved in the last five decades. Currently
shock states are classified as hemorrhagic/traumatic, neurogenic (spinal cord injury), cardi-
ogenic, obstructive (pulmonary embolus, pericardial tamponade), and distributive (septic,
inflammatory). In this issue of SRGS, a discussion of shock from hemorrhage and/or injury,
and septic shock are offered. Other forms of shock have been discussed in previous issues
of SRGS dealing with trauma and venous disease.
Transfusion in surgical patients—risks and current practices
Blood and blood products are life-saving substances and critically important tools used
by surgeons to manage critically ill and injured patients. Risks associated with blood product
administration include hemolytic transfusion reactions (which usually result from errors in
processing and administration), disease transmission, and transfusion-related lung injury.
Stored leucocytes may cause febrile acute transfusion reactions. Leucocytes also mediate
the process of alloimmunization to platelets. Furthermore, transfusions are associated with
immune modulation related to the duration of storage. Transfusion of older units of banked
red blood cells in patients with few risk factors for infection has been associated with an in-
creased risk of infection and mortality. Measures to reduce the number of leucocytes in
banked blood have not been universally effective in preventing presumed leucocyte-
mediated changes in the immune system.
Page 2 of 119
Because blood products are, for the most part, administered after acquisition from un-
related donors and stored for varying periods before administration to patients, functions
of transfused red blood cells, platelets, and coagulation proteins are not normal and may
not immediately correct abnormalities in oxygen delivery or clot formation. The pathophys-
iologic effects of continued bleeding, combined with the need for multiple transfusions,
lead to a coagulopathic state. In this and following sections of the overview, we discuss the-
The first article reviewed is by Goodnough1 in Critical Care Medicine, 2003. The article is
entitled “Risks of blood transfusion.” The author begins by noting that the rarity of serious
transfusion reactions and transfusion-related disease transmission has made estimation of
the risks of transfusion difficult. Mathematical models have been developed to predict
transfusion risk and these models assume that disease transmission is most likely to occur
when blood from infected donors is collected in the “window period” when the donor is in-
fectious but donor-screening tests are not positive. This assumption may lead to underesti-
mation of disease transmission risk. Another potentially hazardous assumption is related to
the fact that patients who receive blood have 1- and 10-year mortality rates ranging from
24%-52%. The patients may not survive long enough for the diagnosis of disease transmis-
sion to be made. Despite these limitations, Goodnough emphasizes that risks of transmis-
sion of viral diseases (mainly hepatitis B and C and human immunodeficiency virus) are low-
er than at any time in the history of transfusion. He presents a table indicating a risk of hep-
atitis transmission of 1/220,000 transfusions to 1/800,000 transfusions, and a risk of human
immunodeficiency virus transmission of 1/1.4 million transfusions. Available data show that
Page 3 of 119
risk continues to decline with improved practices of donor screening, particularly the re-
cently implemented practice of nucleic acid testing.
The author moves on to discuss specific aspects of viral disease transmission. He notes
that the first documented instance of transmission of human immunodeficiency virus in
banked blood occurred in 1982. Before screening blood tests were available, blood banks
began screening donors for specific characteristics and allowing donors to self-exclude their
blood after donation. This practice resulted in an immediate decline in instances of trans-
mission. With the implementation of antibody testing in March 1985, the rate of reported
transmissions of human immunodeficiency virus decreased from more than 700/year to
5/year. Additional small increments of progress occurred after the introduction of screening
for p24 antigen. Hepatitis B virus transmission declined markedly after the introduction of
third-generation tests for the hepatitis B surface antigen. Hepatitis B currently accounts for
only 10% of the instances of hepatitis transmission in transfused blood products. Hepatitis C
is an important viral disease that can be transmitted via transfusion. Transfusion-related
hepatitis C leads to chronic disease in more than 80% of cases. Twenty percent of infections
progress to cirrhosis; one to five percent of infections lead to hepatocellular carcinoma. The
mortality risk for hepatitis C over 25-30 years after transfusion-related infection is nearly
Screening for human immunodeficiency and hepatitis viral disease is very effective.
Since 1999, worldwide, only two cases of human immune deficiency virus infection and one
Page 4 of 119
case of hepatitis C have been reported where all screening tests performed in the donors
Cytomegalovirus infection related to transfusion of blood products is an important
cause of mortality and morbidity in patients immunocompromised from treatment for ma-
lignant disease. The largest risk is observed in patients undergoing stem cell or bone mar-
row transplantation. The prevalence of cytomegalovirus infection in these patient groups is
60% and viral disease develops in up to one-half of the infected patients. Viral disease can
be reduced but not eliminated by using transfusions of blood cells from donors with nega-
tive serum tests of cytomegalovirus infection. The author cites data to show that granulo-
cyte transfusions from serum-negative donors resulted in seroconversion in 1%-4% of pa-
tients. Other data indicate that the risk of infection is not different with cells originating
from donors who are positive for cytomegalovirus infection compared with negative, but
the risk of clinically overt viral disease seems to be less when negative donor cells are used.
Cytomegalovirus infection is much less common in patients who undergo autologous stem
cell or bone marrow transplantation. This protection is observed despite equivalent rates of
serum assay conversion and urinary excretion of viral products in patients undergoing al-
logeneic or autologous bone marrow transplantation. Goodnough notes that 10%-50% of
blood donors test negative for cytomegalovirus infection. Leuco-reduction seems to reduce
the risk of cytomegalovirus infection also. Current recommendations state that patients re-
ceiving allogeneic or autologous bone marrow transplants should receive blood products
donated by serum-negative patients; these products should undergo leuco-reduction.
Page 5 of 119
Bacterial infection of banked blood products can cause rare but often fatal sepsis in pa-
tients who receive transfusion. Yersinia enterocolitica is the most common pathogen infect-
ing red blood cell units. The contamination rate is less than 1/million donated units but clin-
ical infection may become manifest during the transfusion and the mortality rate for clinical
infection is 60%. Platelets carry the greatest risk for bacterial infection because these units
are stored, for up to five days, at temperatures of 20-24°C. This means that stored platelets
are an excellent culture medium for bacteria. Goodnough cites surveillance data indicating
that bacterial contamination occurs in one of every 1000-2000 units. Because four million
platelet units are transmitted annually in the United States (25% of these are apheresis
platelets), estimates are that 300-1000 cases of bacterial sepsis may be expected each year.
Pooled platelets carry a higher risk than apheresis platelet transfusions. Liquid medium cul-
turing of platelets has been implemented in an effort to reduce the risk of infection.
Other infectious diseases that can potentially be transmitted by blood products include
leukemias caused by human lymphotrophic virus 1 or 2. Only a single case had been report-
ed at the time of publication of Goodnough’s article . Epstein-Barr virus disease,
leishmaniasis, babesiosis, toxoplasmosis, brucellosis, malaria, Chagas disease, West Nile vi-
rus, and prion disease can be transmitted potentially by blood products, but such instances
are extremely rare.
Leucocytes in stored blood have been implicated in the production of febrile transfusion
reactions, postoperative infection, transfusion-related lung injury (discussed in SRGS, Vol.
35, No. 8), and increased recurrence risk for colon cancer. In an article by Bernard and coau-
Page 6 of 119
thors2 (reviewed in SRGS, Vol. 35, No. 7), an increased risk for mortality and surgical site in-
fection was noted using an analysis of data from the NSQIP database. Currently universal
leuco-reduction programs have not resulted in significant protection against these compli-
cations. Another report by Weinberg and coauthors3 presents data on injured patients in a
retrospective study. These authors had previously observed an increased risk for mortality
and infectious morbidity in injured patients who received large volumes of transfused
blood. This association was particularly pronounced in patients who received blood with
storage age > 14 days. In this study, the authors focused on patients less severely injured
(injury severity score of less than 25) who did not receive early or large volume transfusion.
They documented an increased risk for mortality, renal failure, and pneumonia when these
less severely injured patients received any transfusion of blood with a storage age of > 14
days. The blood administered to these patients was leuco-reduced. These authors note that
transfusion carries a risk of mortality and infection related to the age of stored blood, and
this risk is not eliminated by leuco-reduction. Their data confirm a number of other analyses
of transfusion risk in injured patients. Goodnough cites additional data that indicate no re-
duction in the risk of infection after cardiac operations when leuco-reduced blood was used.
Another study noted no change in the risk of recurrence of colorectal cancer when a leuco-
reduction program was implemented.
Additional evidence of immune system alterations occurring in patients transfused with
blood and blood products is found in an article by Dunne and coauthors4 in the Journal of
Trauma, 2008. These authors note that micro-chimerism, or the persistence of donor cells
in the circulation of the recipient, represents additional evidence of immune modulation
Page 7 of 119
occurring after transfusion. These authors investigated the role of fresh versus banked
blood in producing micro-chimerisms in combat casualties. They studied 26 severely injured
soldiers who had received fresh whole blood, banked blood, or blood plus platelets. Blood
samples from these patients were compared with blood samples from patients who had re-
ceived no transfusions. There was no evidence of micro-chimerism in the patients who had
not been transfused. Half of the patients who had received fresh whole blood had evidence
of micro-chimerism. A similar proportion of patients who had received banked blood or
blood plus platelets showed evidence of micro-chimerism. The authors cite data from other
studies that confirm the significant occurrence of micro-chimerism. The clinical significance
of this finding is unknown currently but micro-chimerism is known to be associated with
graft-versus-host disease in patients receiving bone marrow transplants.
Most of the morbidity from blood transfusion results from transfusion reactions associ-
ated with errors of cross matching or blood product administration. Goodnough1 notes that
errors in the blood bank or misidentification of recipients occur at a rate of 1/14,000-18,000
transfusions in the United States and Great Britain. Approximately two-thirds of these
events are the result of patient misidentification or failure to recognize a transfusion reac-
tion. One-third of the errors occur during crossmatching and blood labelling processes in
the blood bank. Approximately 10% of hemolytic transfusion reactions are fatal. The mortal-
ity rate for error-related transfusion complications is roughly twice that for all transfusion-
related mortalities from infection.
Page 8 of 119
Transfusion of blood and blood products is not without risk. To date, measures to estab-
lish the specific mechanism of transfusion-related immune modulation have not clearly
identified the factors responsible for these changes. Knowledge of these alterations has led
to the development of conservative transfusion protocols. It is now recognized that transfu-
sion can be safely withheld in stable patients with hemoglobin levels ≥ 7gm/dL. Higher he-
moglobin targets may be required in the critically ill and unstable, or in patients with severe
coronary artery disease. Blood and blood products are best used for specific indications;
protocols that identify specific goals of blood product use and document evidence of
achievement of carefully chosen outcomes will provide surgeons the best means to achieve
effective use of blood and blood products.
Current transfusion practices
Practice guidelines for the use of blood and blood products in the perioperative period
were developed and promulgated by the American Society of Anesthesiologists. A summary
of the practice guidelines is found in a report published in Anesthesiology in 2006.5 The
guidelines begin with a review of practical measures for detection of patients who might
require blood or blood products. The initial evaluation includes, at a minimum, review of
medical records, detailed history focusing on historical evidence of congenital disorders of
coagulation, hemoglobinopathy, and other blood disorders. Additional findings in the histo-
ry will provide information on the risk of ischemic cardiac disease, any history of use of pre-
scription drugs, over-the -counter drugs, or herbal remedies that alter coagulation, and his-
tory of prior transfusion. Evidence of prior exposure to drugs (aprotinin) that could precipi-
tate allergic reaction on repeat administration is sought. The patient and family are ques-
Page 9 of 119
tioned to document the willingness, or lack of willingness, to accept blood transfusion. Rec-
ommended laboratory assessments include hemoglobin and hematocrit determinations and
a coagulation profile. Management of patients who refuse blood products will be discussed
in a later section of the overview.
The guidelines note that specific evidence to support use of drugs such as aprotinin, ep-
silon-amino caproic acid, and tranexamic acid is lacking except for certain cardiac or ortho-
paedic procedures associated with significant risk of bleeding (for example, reoperation for
open cardiac surgery and reoperation for joint replacement). They emphasize that anti-
fibrinolytic drugs have been implicated in vascular graft thrombosis and severe allergic reac-
tions have resulted from aprotinin use.
Evidence to document the value of preoperative erythropoietin treatment of anemia is
not available although certain patient groups such as patients with renal disease, anemia of
chronic disease, and patients who refuse transfusion may benefit. Several weeks of treat-
ment are required for correction of anemia using erythropoietin. The guidelines note that
preoperative collection of autologous blood has been associated with reduced transfusion
risks. They emphasize, however, that severe transfusion reactions (associated with misiden-
tification of patients) and bacterial infections have been reported from programs using pre-
operative collection of autologous blood.
The guidelines recommend cessation of drugs such as warfarin, aspirin, and clopidogrel
before elective operations. Intervals of one to three weeks are required for reversal of the
affects of clopidogrel and aspirin on platelets. The method chosen for preoperative restora-
Page 10 of 119
tion of clotting function depends on a careful assessment of the type, urgency, and risk of
the surgical procedure chosen. The risk of thromboembolic complications that might occur
soon after the chronic warfarin dose is decreased or eliminated is also assessed. For exam-
ple, patients with cancer are at increased risk for thromboembolic complications. By con-
trast, recent data indicate that stroke risk reduction from warfarin treatment in patients
with atrial fibrillation is overcome by the risk of bleeding when patients are under 70 years
of age and have minimal stroke risk.6 Warfarin cessation can result in restoration of clotting
capability as evidenced by an international normalized ratio of 1.5 within several days to
one week. Operation is usually safe at this level of INR and re-starting of warfarin when the
patient is stable and free of significant bleeding risk usually results in rapid and dependable
re-anticoagulation. If reversal of warfarin affects is needed within 48-72 hours, intravenous
vitamin K is indicated. More rapid reversal than this will require use of fresh frozen plasma
or, occasionally, recombinant activated factor VII. Use of recombinant factor VII is discussed
in more detail later in the overview. Reversal of warfarin with vitamin K or fresh frozen
plasma is associated with more difficult re-anticoagulation postoperatively.
The decision to transfuse before, during, and after an operation is based on clinical as-
sessments of the volume of blood loss and the physiologic effects of blood loss. Monitoring
of ongoing blood loss requires clear and frequent communication between the surgical
team and the anesthesiology team. Use of assessments of blood pressure, heart rate, arte-
rial oxygen saturation, central venous pressure, pulmonary artery occlusion pressure, cen-
tral venous (or mixed venous) oxygen saturation, and lactate levels depends on type of op-
eration and patient risk. For patients with a history of cardiac dysfunction, intraoperative
Page 11 of 119
echocardiography may be indicated. The practice guidelines note that assessments of blood
pressure, heart rate, and arterial oxygen saturation are routine during surgical procedures,
but the literature is unclear about specific contributions these assessments make with re-
gard to transfusion. A synthesis of various information points will usually be required to
support clinical decisions. Intraoperative measurement of hemoglobin and hematocrit is
common but these values may not reflect end-organ perfusion status. The guidelines note
that transfusion is not needed in stable patients who do not have cardiac disease until the
hemoglobin level is below 6 gm/dL. Transfusion is definitely not needed when the hemo-
globin level is above 10 gm/dL.
The decision to transfuse patients whose hemoglobin levels are between 6 gm/dL and
10 gm/dL is made based on assessment of each patient’s cardiovascular reserve, the risk of
ongoing blood loss, and the status of end organ perfusion. The safety of lower hemoglobin
levels in stable, critically ill patients was documented in a 1999 report by Hebert and coau-
thors.7 An appreciation of the risks of transfusion and an understanding of the limitations of
the blood supply has stimulated caution and overall conservatism about use of banked
blood and blood products. Benefits of a conservative approach to transfusion are most ob-
vious in patients under the age of 55 who have no clinical evidence of cardiovascular dis-
ease. Assessment and monitoring of patients with blood loss and shock are discussed in lat-
er sections of the overview.
Coagulopathy from massive transfusion is discussed in a later section. The practice
guidelines support the use of fresh frozen plasma for this condition as well as cryoprecipi-
Page 12 of 119
tate for the support of fibrinogen levels. Recombinant activated factor VII may be useful as
“rescue therapy” for patients with microvascular bleeding. The guidelines note that
measures such as normovolemic hemodilution and deliberately lowered venous pressures
may be helpful in selected patients undergoing cardiac, hepatic, and orthopedic procedures.
Techniques of perioperative blood salvage
Blood recovery and reinfusion during and/or after an operation is potentially a means of
reducing the use of blood and blood products in surgical patients. Collection and reinfusion
of blood after tube thoracostomy for traumatic hemothorax has been shown to be safe. Be-
cause liver surgeons have recognized the value of low central venous pressure as a means
of reducing blood loss during liver parenchymal transection, there is interest in collecting
blood before liver transaction, and re-infusing the blood once liver hemostasis has been ob-
tained. This topic is addressed in an article by Hashimoto and coauthors 8 in Annals of Sur-
gery, 2007. These authors conducted a randomized trial comparing one group of living liver
donors who had removal of a volume of blood equal to 0.7% of body weight with a group of
patients treated with conventional management of liver resection. The endpoints of inter-
est were the amount of blood loss during liver parenchymal transection and the volume of
banked blood infused after parenchymal transection. The patients who had removal and re-
infusion of blood had significantly less blood loss and used less banked blood than the pa-
tients did in the comparison group. The trend toward smaller blood product use in patients
who had intraoperative blood salvage was identified, but this difference was not statistically
significant. Intraoperative blood salvage is used frequently in cardiac surgical procedures.
The value of this approach in abdominal operations is debated. Limitations to the use of in-
Page 13 of 119
traoperative blood salvage in abdominal operations are concerns over depletion of clotting
factors in salvaged blood and the potential for bacterial contamination during abdominal
Recovery of postoperative blood drainage from the operative site is possible using
closed drainage systems that collect blood through a filtration system into a blood admin-
istration bag. This approach is used in Europe and is especially well suited for patients un-
dergoing orthopaedic procedures that, predictably, have significant postoperative blood
drainage. An article describing the potential value of postoperative blood salvage by Mirza
and coauthors9 appeared in the Annals of the Royal College of Surgeons, 2007. The authors
report results in 109 patients undergoing total hip replacement. The amount of periopera-
tive blood loss (as indicated by the change in hemoglobin level) in these patients was com-
pared with a group of historical control patients. Patients who had blood salvage had a
smaller perioperative hemoglobin drop. Nine percent of the salvage patients required peri-
operative transfusion of banked blood compared with 30 percent of the historical control
group. The authors conclude that this approach may reduce use of banked blood in patients
undergoing operations associated with predictable postoperative drainage of uncontami-
Management of patients who refuse blood and blood products
Surgeons inevitably will encounter patients who decline all or some parts of proposed
strategies for care of surgical conditions. Use of blood and blood products is unacceptable
to some patients. Typically, this patient group includes, but is not limited to, Jehovah’s Wit-
nesses. A significant body of knowledge has accumulated that has permitted safe surgical
Page 14 of 119
care for Jehovah’s Witnesses and this knowledge can be used to optimize the care of any
patient who declines the use of blood and blood products. In this section of the overview,
we will discuss two articles that focus on the elements of care for Jehovah’s Witnesses. The
first article discussed is from Hughes and coauthors10 in the Journal of Trauma, 2008, enti-
tled “The contemporary approach to the care of Jehovah’s Witnesses.” This article is sup-
plied as a full-text reprint accompanying this issue of SRGS. The authors begin by providing
background on the origins and development of the Jehovah’s Witnesses sect. They note
that the sect evolved from a Bible study group formed by Charles Taze Russell in Alleghany,
Pennsylvania, in 1869. Based on extensive study, the group identified what they perceived
as fundamental errors of interpretation of biblical text. The interpretation of the group led
to the perception that there were philosophical errors in the practice of Christian doctrine.
The literal interpretations of scripture by the group evolved into commands. Obedience to
these commands, in the eyes of the adherents to this doctrine, was required for any chance
at life after death. In 1879, the group was known as “Bible Students” and the group pro-
duced a religious magazine later called The Watchtower. In 1931, the group adopted the
name “Jehovah’s Witnesses.” Currently, this is the fastest growing religious sect in the
western world with nearly 7 million members. In the United States, more than 1 million
members are enrolled. The sect is based on person-to-person ministry and the group is not-
ed for their work in literacy and disaster relief. Each member must commit a certain amount
of time each month to ministry. Fundamental to their beliefs is the recognition that God’s
Kingdom is the only legitimate government. Thus, members of the sect do not join service
organizations or serve in the military. They remain neutral politically.
Page 15 of 119
The rejection of blood transfusion was codified in the belief system of Jehovah’s Wit-
nesses in 1945. The belief is based on the literal interpretation of several Old Testament
passages that prohibit the eating of food that contains blood. In 1951, a Watchtower article
cited the use of intravenous fluids in hospitals as “intravenous feeding” and declared that
the intravenous administration of blood was equivalent to “feeding” of blood and was,
therefore, unacceptable based on the interpretation of Old Testament texts. The authors
note that the prohibition has been interpreted to extend to the use of banked blood, blood
cells, plasma, and platelets. Preoperative autologous blood donation is prohibited because
the blood is separated from the body. This interpretation permits the use of intraoperative
blood salvage as long as the collection circuit is continuous with the administration device
so that separation of the blood from the patient’s body does not occur. In 2000, the Watch-
tower Society issued a statement saying that the Jehovah’s Witness sect would no longer
excommunicate members who accepted transfusion. Instead, the statement indicated that
the member would be judged to have voluntarily given up membership in the sect based on
having accepted blood transfusion. It is important to note, as Hughes and colleagues em-
phasize, that there is variability among Jehovah’s Witnesses about the acceptability of blood
derivatives. Some adherents will accept albumin, immunoglobulins, and factor concen-
trates. Recombinant erythropoietin may also be accepted. Some Jehovah’s Witnesses do
not believe that they have committed an act requiring that they leave the sect if the agree-
ment to accept a transfusion is made privately with the treating physician or surgeon. In
such circumstances, patient confidentiality is obviously critical. Against this approach is the
Page 16 of 119
frequent presence of a member of the sect physically in the care area to make certain that
the treated member does not receive a transfusion or any unacceptable blood product.
It is important that surgeons decide, ahead of time, whether they will be able to accept
the fact that the patient may not accept transfusion. The surgeon will need to reconcile the
patient’s wishes with the possibility that a “preventable” death may occur if transfusion is
withheld. Acceptable arrangements for transfer of the care of the patient to an alternate
surgeon may become necessary. Hospital resources are available for patients and treating
surgeons. Ethics committees, risk management group, and transfusion medicine specialists
are potentially valuable. In many cities, Jehovah’s Witness groups offer consultative advice
for patients and treating surgeons.
Legal aspects of the care of Jehovah’s Witnesses are grounded in the 1914 legal decision
that established the right of patients to refuse treatment. The decision to exercise this right
by a patient, particularly a Jehovah’s Witness, should be confirmed in writing by the patient.
Many Jehovah’s Witnesses carry an advance directive document with them at all times.
Hughes and associates stress the importance that this knowledge is made known to all
caregivers. Thus, the medical record should display, predominantly, the wishes of the pa-
It is worth noting that many court decisions have determined that parents do not have
the right to refuse transfusion for their children. The ability of adolescents, or mature mi-
nors, to refuse transfusion is less clear and rulings have varied depending on the jurisdic-
Page 17 of 119
tion. Surgeons will need to consult hospital risk management and counsel for assistance in
dealing with these issues.
Where time permits full discussion of the risks and benefits of operation, formal in-
formed consent will be possible. What the surgeon should do to proceed with treatment
when the informed consent discussion has not occurred is challenging. The presence of an
advance directive card does not preclude the necessities of obtaining informed consent or
proceeding with needed treatment in emergencies where informed consent cannot be ob-
tained. This means that the decision to transfuse, or withhold transfusion for a Jehovah’s
Witness patient, in an emergency may be viewed, in retrospect, as a basis for legal action
against the surgeon and/or hospital. To facilitate management of Jehovah’s Witness pa-
tients in emergencies, early knowledge of the religious beliefs of the patient is necessary.
This information should be actively sought and made available to all caregivers. Hughes and
colleagues cite data from a Level I trauma center that indicated that the necessary
knowledge was frequently not obtained. In this analysis, 5% of the Jehovah’s Witness pa-
tients received transfusion but in only one case was this reasonably preventable. A discus-
sion of some of the measures available for minimizing blood loss as well as examples of ad-
vance directive and consent forms are found in an article authored by Gohel and
coauthors11 in the Annals of the Royal College of Surgeons of England, 2005.
It is important to note that data about outcomes of injured Jehovah’s Witness patients
have not shown an increased risk of death from injury in this group. Earlier decision for op-
eration has been consistently noted in studies of the management of injured Jehovah’s
Page 18 of 119
Witnesses patients. In elective surgery patients, measures to prevent bleeding and im-
proved operative approaches have resulted in improved outcomes for Jehovah’s Witnesses
as noted in data cited by these authors. Additional data have shown that postoperative
anemia (hemoglobin level of 6 gm/dL) did not predict mortality or morbidity. Data indicat-
ing an increased risk of mortality have been reported for patients with chronic severe ane-
mia but anemia, in this setting, has been interpreted as evidence of increased overall risk
for this patient group.
Hughes and colleagues note that management of Jehovah’s Witnesses should include
measures to minimize iatrogenic blood loss (phlebotomy), minimize intraoperative blood
loss (hemodilution, intraoperative blood salvage), enhance red blood cell production, en-
sure hemostasis, and maintain blood volume with electrolyte or colloid solutions. Protocols
for elective management of Jehovah’s Witness patients have avoided transfusion without
increasing operative risk. A report on the effectiveness of a transfusion-free program for the
conduct of orthotopic liver transplantation in Jehovah’s Witnesses is the subject of an arti-
cle by Jabbour and coauthors12 in Archives of Surgery, 2006. These authors reported on a
comparison of patients treated by a formal protocol approach to transfusion avoidance
compared with a group of historic controls. The protocol stressed the maintenance of low
venous pressure, use of acute normovolemic hemodilution, and intraoperative blood sal-
vage. These authors noted that their protocol group was sicker as judged by MELD scores
compared with the non-protocol historic controls. Despite this, operative mortality in the
transfusion-free patients was not increased. These data indicate that avoidance of transfu-
sion in complex elective surgery is safe when a protocol approach is used.
Page 19 of 119
From the perspective of the editor, it seems clear that transfusion avoidance is not as-
sociated with increased operative risk in elective operations and that reductions in blood
product use are associated with decreased risk for transfusion-related complications. Pre-
planned protocols designed to identify Jehovah’s Witness patients (and, in fact, any pa-
tients) who are unwilling to accept blood and blood products are helpful. Free and open
discussion of risks and benefits with the patient is critically important. Use of institutional
and community resources to accomplish these goals is helpful. In an emergency when fully
informed consent is not obtainable, or when dealing with children of Jehovah’s Witnesses
parents, therapeutic choices should be directed toward obtaining the optimum safe out-
come for the patient.
Pathophysiology of shock
In this section of the overview, derangements in homeostasis that are unique to various
forms of shock will be reviewed. Along with this discussion, we will emphasize the physio-
logic alterations that are common to most or all forms of shock.
Shock due to hemorrhage and trauma
Traditionally, shock from hemorrhage and trauma has been thought to proceed sequen-
tially through stages of compensation that include a phase of catecholamine mediated vas-
oconstriction and redistribution of blood volume toward the brain and heart and away from
skeletal muscle, kidneys, and splanchnic circulation. During this phase, movement of pro-
teins, water, and ions from the extracellular fluid space into the intravascular space pro-
vides partial refilling of the volume lost to bleeding. This response, coupled with catechola-
mine-mediated vasoconstriction, serves to support blood pressure. A subsequent phase of
Page 20 of 119
partial compensation where the patient is mildly to moderately hypotensive occurs; this is
followed, within a variable interval, by a final phase characterized by systemic vasodilation,
bradycardia, and death. The duration of this phase is determined mainly by the rate and
volume of continuing bleeding. The phase of partial compensation has been envisioned as a
phase wherein autoregulation of vascular beds fails and tissue blood flow becomes pressure
dependent. During this phase, cellular hypoxia is thought to lead to disordered mitochon-
drial function and a shift to anaerobic metabolism. Accumulation of acid metabolites and
lactate are evidences of these phenomena. As partial compensation proceeds to
decompensation, cell membrane function fails with movement of water and ions from the
extracellular space into the intracellular space.
The physiologic construct described above has driven the development of treatment
protocols for shock that were based on early identification of bleeding sites and control of
hemorrhage, replacement of red blood cells, and replacement of volumes of electrolyte so-
lution calculated based on expected losses from the extracellular fluid. Among the major
features of understanding of shock and resuscitation that have undergone modification in
the past two decades are, first, an understanding that patients with compensated and par-
tially compensated shock have expected losses of extracellular fluid smaller than those cal-
culated on the basis of experimental evidence from models of severe, decompensated
shock. In fact, the most common clinical settings in which large volume electrolyte fluid re-
suscitation is required are encountered in patients with massive blood loss from penetrat-
ing torso trauma and in combat injuries. Further, it is now clear that restoration of effective
tissue perfusion pressure and oxygen delivery are the most important resuscitation targets,
Page 21 of 119
and restoration of arterial pressure to “normal” is not necessary. In fact, overzealous resto-
ration of blood pressure may lead to recurrent bleeding and decompensation. Finally, cur-
rent understandings hold that splanchnic ischemia leads to movement of substances into in-
testinal lymph that contribute to creation of a proinflammatory state and progression to
multiple systems organ failure in patients who are genetically susceptible to the develop-
ment of dysfunctional systemic inflammatory responses. In this section of the overview, we
review some of the research and clinical data that have contributed to our current under-
standings of shock from hemorrhage and trauma.
Experimental evidence demonstrating that tissue blood flow was reduced by
microvascular occlusion because of cell swelling and depolarization during shock stemmed
from work completed by several investigators. One example of this research is found in an
older article by Borchelt and coauthors13 in the Journal of Trauma, 1995. These authors ear-
lier demonstrated that cellular dysfunction in shock was mediated by a protein they called
“circulating shock protein.” This substance was shown to cause cellular depolarization that
could lead to sequestration of extracellular fluid in the cells. In this study, they hypothesized
that this substance could produce migration of water and ions into cells. They incubated
cells with the protein and used a double isotope indicator technique to measure cell water.
They were able to demonstrate movement of water into cells in response to exposure of
cells to the protein; cells used in this experiment were skeletal muscle cells. This paper was
presented during the plenary session of the annual meeting of the American Association for
the Surgery of Trauma, September 1994. The discussion that followed the presentation ac-
companies the article. The discussants agree that cell swelling is an important phenomenon
Page 22 of 119
in shock but limitations of isotope techniques have made determination of the magnitude
of cell swelling and fluid migration into cells difficult. Debate has continued about the loca-
tion of cell swelling, the types of cells that swell, and magnitude of the contribution of the
fluid migration into cells.
Defining one location of cell swelling leading to fluid sequestration, cellular hypoxia, and
elaboration of proinflammatory mediators is the topic of an article by Zakaria and coau-
thors14 in Surgery, 2007. These authors produced severe hemorrhagic shock in a rodent
model and used intravital videomicroscopy and Doppler flow measurements to determine
microvascular diameters and flow rates. An index of functional capillary density was calcu-
lated to estimate the effective cross-sectional area of perfused microvessels. All measure-
ments were made in an isolated segment of rat ileum. The authors investigated the follow-
ing hypotheses: 1) that endothelial swelling induced by activation of the sodium/hydrogen
ion exchanger during shock would reduce tissue blood flow; 2) that conventional resuscita-
tion with blood and electrolyte solution would not prevent cell swelling; and 3) that direct
inhibition of the sodium/hydrogen ion exchanger with topical amiloride would prevent cell
swelling and preserve microcirculatory flow. The authors were able to document cell swell-
ing that reduced microvascular flow and functional capillary density. Conventional resusci-
tation did not restore microcirculatory function even though central hemodynamics was re-
turned to baseline levels. When cell swelling was prevented by blocking the sodi-
um/hydrogen ion exchanger, cell swelling was prevented and microvascular flow was re-
stored by conventional resuscitation. As a final component of this experiment, the authors
prevented cell swelling by bathing the ileum tissue in peritoneal dialysis fluid and noted that
Page 23 of 119
direct peritoneal resuscitation using peritoneal dialysis fluid prevented endothelial cell
swelling and preserved microcirculatory capillary density. In this setting as well, conven-
tional resuscitation restored microvascular flow and function to near baseline levels. The
authors note that activation of the sodium/hydrogen ion exchanger permits the cell to
transport hydrogen ions into the extracellular space. Chloride ion is also transported into
the extracellular space. Bicarbonate and sodium are transported into the cell along with wa-
ter. The flux of water, bicarbonate, and sodium into the cell produce cell swelling. The esti-
mated size of the endothelial surface in intestine and skeletal muscle would explain the
large fluid shifts observed in experimental preparations of severe shock as well as in injured
patients sustaining massive blood loss. Persistence of intracellular acidosis could also ex-
plain the observed disorders of cellular energy metabolism and the resultant accumulation
of acid metabolites and lactate. Reversing cellular edema with direct peritoneal resuscita-
tion maintained microvascular flow in the intestinal bed and could reduce the elaboration
of proinflammatory cytokines into intestinal lymph. This group of investigators examined
the effects of direct peritoneal resuscitation on the flow and composition of mesenteric
lymph is a study of rodent hemorrhagic shock authored by Matheson and coauthors15 in Ar-
chives of Surgery, 2009. Earlier work by these authors (one report described above) demon-
strated endothelial swelling, leading to reductions in intestinal microvascular flow. Conven-
tional resuscitation did not prevent this phenomenon but reductions of cell swelling with di-
rect peritoneal resuscitation, combined with conventional resuscitation, preserved intesti-
nal flow. In this experiment, the authors hypothesized that preservation of intestinal
microvascular flow would alter intestinal lymph flow and composition. They were able to
Page 24 of 119
demonstrate increased mesenteric lymph flow with shock and resuscitation. Increased lev-
els of lymph hyaluronic acid and proinflammatory cytokines were noted with shock treated
by conventional resuscitation. Direct peritoneal resuscitation added to conventional resusci-
tation preserved intestinal microvascular flow and function. Mesenteric lymph flow was re-
stored to normal, and cytokine levels were returned to control levels by treating rodent
hemorrhagic shock with the combination of conventional and direct peritoneal resuscita-
tion. These authors have documented a logical mechanism for cellular dysfunction and re-
duced intestinal blood flow. They have explained the mechanisms of cellular swelling, re-
duced tissue blood flow, and elaboration of proinflammatory cytokines from the gut. In
other experiments, they showed that skeletal muscle cell swelling and fluid sequestration
occurs. These experiments confirm the existence and the magnitude of fluid shifts in pro-
found shock and the authors propose a plausible link between mesenteric ischemia and
Clinical evidence of mesenteric ischemia is associated with patient mortality and
postshock multiple organ failure. This topic is reviewed by Malinoski and coauthors 16 in the
Journal of Trauma, 2009. The authors note that animal studies have documented the
movement of pancreatic enzymes from the lumen of the intestine into mesenteric lymph
and subsequently into the general circulation in animal models of shock and multiple organ
failure. The observed elevations of amylase and lipase occurred without tissue evidence of
pancreatic inflammation and seemed to result from gut ischemia. These authors hypothe-
sized that patients with shock would have elevated levels of pancreatic enzymes and that
the elevations would be predictive of mortality and organ failure. They analyzed data from
Page 25 of 119
an institutional trauma registry and identified 481 patients with elevations of amylase and
lipase more than twice as high as normal levels. Patients in this group were twice as likely to
die, to develop organ failure, and require massive transfusion, as were patients without ele-
vations of these enzymes. In the discussion section of their paper, the authors note that ex-
perimental preparations have shown that ischemia to the intestinal mucosa allows move-
ment of pancreatic enzymes into the submucosa where these substances can activate in-
flammatory mediators. The enzymes and mediators can translocate from the submucosa to
the mesenteric lymph channels and enter the general circulation by this route.
The gut-lymph hypothesis of the origin of multiple organ failure has been extensively in-
vestigated in the laboratory. One example of this research is found in an article by Senthil
and coauthors17 in the Journal of Trauma, 2006. In a porcine model of trauma and shock,
the authors demonstrated that injury and shock are necessary for production of lymph that
would activate neutrophils. Neutrophil activation was used as a marker for the presence of
proinflammatory cytokines. The authors further documented that the lymph draining from
the gut in animals sustaining shock and trauma was cytoxic to endothelial cells. In the dis-
cussion section of the article, the authors note that several studies have shown that mesen-
teric lymph from animals sustaining shock and trauma has properties that can produce cel-
lular effects similar to those seen in multiple system organ failure. They note that the lymph
does not contain bacteria or endotoxin, leading to the conclusion that mediators contained
in the lymph, such as tumor necrosis factor alpha and interleukin-6, could lead to cellular
changes that might induce multiple organ failure syndrome in susceptible patients.
Page 26 of 119
The question of whether lymph draining from other tissues rendered ischemic during
shock would also cause cellular changes consistent with multiple organ failure is the topic of
two articles by Diebel and coauthors.18-19 These authors showed that lymph gathered from a
skeletal muscle bed two hours after shock produced in a canine model led to acute lung cell
injury and neutrophil priming. Taken together with the research described earlier, it seems
likely that lymph from tissue rendered ischemic during shock can produce changes con-
sistent with multiple organ failure. Whether these mechanisms are active in human shock
requires further clarification.
The role of proinflammatory cytokines in the production of multiple organ failure syn-
drome has been emphasized in multiple studies. An article documenting the relationship of
elevated cytokine levels and multiple organ failure by Jastrow and coauthors 20 appeared in
the Journal of the American College of Surgeons, 2009. These authors measured sequential
levels of cytokines in severely injured patients treated in a single Level I trauma center. Cy-
tokine levels were measured during the first 24 hours after admission. There were no signif-
icant clinical differences observed when the group of patients who developed multiple sys-
tem organ failure was compared to the group that did not develop multiple system organ
failure. The authors were able to identify, however, four cytokines present at significantly
higher levels in patients who developed organ failure. These included interleukin-10 and in-
terleukin-6 that have been associated with organ failure development in a number of exper-
imental and clinical studies. Interestingly, there were some patients who did not develop
organ failure but who had elevated cytokine levels. This observation suggests that other fac-
Page 27 of 119
tors, perhaps including genetic predisposition, might work in conjunction with the cytokine
expression to produce a cellular environment leading to the development of organ failure.
One potential factor influencing susceptibility to postshock mortality and the risk of or-
gan failure is gender. Males are more likely to be injured and are more likely to have elevat-
ed levels of circulating cytokines after injury. Mortality and the frequency of multiple organ
failure are more common in males as documented in several experimental and clinical se-
ries. This issue is the topic of a report by Sperry and coauthors 21 in the Journal of Trauma,
2008. These authors documented increased levels of interleukin-6 in severely injured male
patients. These levels persisted over time and increased serum levels of interleukin-6 were
associated with an 11-fold increase in the risk of multiple systems organ failure. These au-
thors note that a causal relationship between increased cytokine levels cannot be inferred
from these data but increased cytokine levels could be a marker of increased organ failure
risk. This same group of investigators examined the role of female hormone levels in pro-
ducing the apparent protection against mortality and multiple organ failure in injured wom-
en compared with injured men.22 These authors examined hormone levels in pre- and
postmenopausal women who had sustained severe injury and noted no difference in the
risk for death or multiple organ failure in postmenopausal women compared with premen-
opausal women. The risk for mortality and multiple organ failure in both groups was lower
than that for men after adjustment for multiple risk factors. The authors note that these
findings contrast sharply with experimental animal data that have linked levels of hormones
to the protective effect of female gender after shock and injury. They conclude based on
Page 28 of 119
the data, that there are factors other than hormone levels that contribute to the protective
effect of female gender.
Microvascular flow abnormalities associated with shock and injury may be produced by
alterations in circulating blood cells as well as abnormalities in the small vessels of the mi-
crocirculation. This topic is addressed in an article by Machiedo and coauthors 23 in the Jour-
nal of Trauma, 2009. The authors evaluated the affects of infusion of blood from animals
subjected to trauma and shock on organ blood flow in a rodent model. Organ blood flow
was measured using radioactive microspheres. The authors were able to document in-
creased sequestration of red blood cells from animals subjected to trauma and shock and
subsequently infused into normal animals in organs such as the liver and spleen. Organ
blood flows were reduced. They concluded that red blood cell changes induced by trauma
and shock predispose these cells to lodge in organ microcirculations producing obstruction
of blood flow. These data and the studies discussed earlier suggest a diverse group of
changes in organ microcirculation resulting from trauma and shock. Detailed studies of the
human microcirculation are needed to provide clinical evidence of measures to improve
Protection against cellular death is another avenue that could improve outcomes in se-
verely injured patients with shock. One measure leading to improved cell survival is inhibi-
tion of deacetylation of histones. Histone deacetylation is one means of altering cellular
DNA leading to cell death. Inhibition of this process is the topic of a report authored by
Gonzales and coauthors24 in the Journal of Trauma, 2008. The authors note that hemor-
Page 29 of 119
rhagic shock-induced changes in the liver produce aberrations in gene-regulatory programs;
one important pathway for these changes is recruitment of histone deacetylases in liver tis-
sue. Valproic acid, a regularly prescribed antiepileptic drug, was recently shown to produce
neuroprotective effects through inhibition of histone deacetylation. Pretreatment with
valproic acid in a rodent model of severe hemorrhagic shock produced favorable changes
leading to liver cell survival and these changes were directly related to hyper-acetylation of
The cell protective activity of valproic acid was further evaluated in a porcine model of
severe hemorrhage, trauma, and shock and reported in an article authored by Alam and co-
authors25 in Surgery, 2009. The authors of this report treated animals with valproic acid af-
ter hemorrhage and liver injury in contrast to the pretreatment model examined in the arti-
cle authored by Gonzales.24 Alam and coauthors examined outcomes in animals untreated,
animals treated with fresh whole blood, and animals treated with intravenous valproic acid.
Survival was 100% and 86% in animals treated with fresh whole blood and valproic acid re-
spectively compared with 25% in the control group. The authors examined liver tissue and
found that survival was associated with preservation of the Akt cell survival pathway. Nor-
mal function of this pathway depends on prevention of histone deacetylation. These two
studies indicate that interventions at the cellular level hold promise as therapeutic path-
ways for the management of trauma and shock.
Restoration of perfusion after a period of ischemia is another potential pathway leading to
tissue damage. Restoration of blood flow to tissue is essential for survival from shock but reper-
Page 30 of 119
fusion injury is a potential pathway for production of cellular injury and postshock complica-
tions through production of reactive oxygen species as well as other substances. Reperfusion
injury after hemorrhage and injury is reviewed in an article by Rushing and Britt 26 in Annals of
Surgery, 2008. This article is included as a full-text reprint accompanying this issue of SRGS.
These authors note that tissue level evidence of reperfusion injury is noted, initially, by neutro-
phil adhesion to venular endothelium. They stress that adhesion molecules are expressed on
endothelial cells, platelets, and neutrophils are all active in the adhesion process. Blockade of
cell adhesion is appealing, although trials of adhesion molecule blockade have not shown bene-
fit. The authors note that mitochondrial injury is a constant factor in reperfusion injury and cite
evidence suggesting that cytochrome C oxidase is a mediator of reperfusion damage to mito-
chondria. In addition, enzymatic mediators such as xanthine oxidase are important in the pro-
duction of reperfusion injury. Reperfusion injury produces alterations of intracellular signaling
pathways such as the NF-kappa B pathway that augment production of pro-inflammatory cyto-
kines such as tumor necrosis factor alpha. Immune cells, such as B-lymphocytes, are important
locations of these alterations. Changes within these pathways can lead to cell death via apop-
totic pathways. The events initiating and perpetuating reperfusion injury are important foci of
research efforts as pathways to improved patient outcomes are sought.
Sepsis and septic shock
A review of the pathophysiology of sepsis and septic shock is authored by Hotchkiss and
Karl27 in the New England Journal of Medicine, 2003. The authors begin by noting that sepsis
and septic shock are the leading causes of death among critically ill patients in the United
States; 750,000 cases of sepsis are diagnosed annually and 25%-40% of these patients die
Page 31 of 119
depending on risk factors (age, comorbidities), the etiology of sepsis, and the mode of ther-
apy. They go on to stress that the major clinical manifestations of sepsis are not infection-
specific but represent an inherent pattern of response to infection regardless of the infect-
ing organism. Although clinical sepsis is thought to result from dysfunction of the inflamma-
tory response, proinflammatory cytokines are variably recovered from septic patients and
blockade of cytokines has not shown benefit as a means of improving sepsis outcomes. Re-
cent data have described targeted alterations in the innate immune system that may in-
crease susceptibility to sepsis. As an example, the authors cite data that alterations in the
toll-like receptor 4 gene have been demonstrated in humans who show increased suscepti-
bility to sepsis. The authors go on to discuss the shift from a proinflammatory state to a
state of immune suppression which occurs during the septic event and this immune sup-
pression may predispose to nosocomial infections in patients recovering from sepsis. As
immunosuppression progresses, anergy may occur due to apoptotic cell death in popula-
tions of immune cells such as lymphocytes and gastrointestinal mucosal cells. Animal data
have shown that anti-apoptotic drugs reduce the impact of this process and in animals,
mortality from sepsis is reduced28. Data from surgical patients suggest that restoration of
the proinflammatory state that characterizes the early septic response is associated with
restoration of immune function and improved survival.
Genetic alterations also may act to perpetuate the immunosuppressed state. The au-
thors cite data that alterations in the genes that produce tumor necrosis factors alpha and
beta result in low levels of these cytokines in patients who die of infection during the im-
mune suppressed phase of sepsis.
Page 32 of 119
Neutrophils have traditionally been thought essential to the successful control of infec-
tion. Recent experimental studies have suggested that neutrophils may mediate tissue inju-
ry (see discussion of Rushing’s review, above). Blockade of neutrophil-mediated injury has
not improved sepsis outcomes, however. In contrast, some suggestion of benefit was noted
in patients with pneumonia who were treated with granulocyte colony stimulating factor.
These patients had neutrophil counts in excess of 70,000 cells/mm3 but there was no sug-
gestion of neutrophil-mediated tissue injury.
Observations from autopsy data, cited by Hotchkiss and Karl, suggest that the popula-
tions of cells dying in patients with sepsis are lymphocytes and gastrointestinal epithelial
cells. This cell death is thought to occur because of sepsis-induced acceleration of apoptosis.
No specific therapy is available to forestall lymphocyte and gastrointestinal epithelial cell
death. These cells are popular foci of research into the basic mechanisms of sepsis and sep-
Hotchkiss and Karl note that septic shock, though frequent in septic patients, is an un-
common cause of death in this patient group. Septic shock is a form of distributive shock
characterized by preservation of cardiac output and global indices of oxygenation occurring
in the setting of hypotension. Recent studies have shown that microcirculation dysfunction
can be observed in the sublingual mucosa, which is anatomically and functionally similar to
the intestinal mucosa. Failure to recover microcirculatory function has been associated with
mortality from sepsis and septic shock although shock itself was less often a cause of death
than the subsequent development of organ failure and/or nosocomial infection. Therapy for
Page 33 of 119
septic shock has been directed toward preserving perfusion and oxygen delivery in an at-
tempt to prevent sepsis-associated organ failure that is a more common cause of death in
septic patients than septic shock. Therapy for sepsis and septic shock is discussed later.
Arterial hypotension (systolic blood pressure < 90 mmHg) combined with signs of end-
organ hypoperfusion (acidosis, elevated serum creatinine, abnormal sensorium), depressed
cardiac index, and signs of elevated left ventricular pressure comprise the clinical features
of cardiogenic shock. Cardiogenic shock occurs in 7% of patients after myocardial infarction
and this condition is the most common cause of early death after myocardial infarction.
Other conditions that complicate myocardial infarction such as papillary muscle rupture
with acute mitral insufficiency and ventricular wall rupture with pericardial tamponade are
surgical emergencies that also cause cardiogenic shock.
The epidemiology of cardiogenic shock is discussed in a review by Topalian and coau-
thors29 in Critical Care Medicine, 2008. This article is supplied as a full-text reprint with this
issue of SRGS. These authors cite three reports indicating that cardiogenic shock complicat-
ing myocardial infarction (defined as ST-segment elevation and/or new left bundle branch
block on electrocardiogram) occurred in 7.1%-8.6% of patients after myocardial infarction.
Once cardiogenic shock was diagnosed, subsequent mortality rates reported were 60-80%.
Mortality rates were higher in patients older than 75 years. Topalian and colleagues stress
recent data showing that aggressive attempts to provide complete revascularization of criti-
cal coronary stenoses and occlusions in patients with cardiogenic shock has resulted in low-
er mortality rates approaching 48%. Patients at increased risk for cardiogenic shock after
Page 34 of 119
myocardial infarction include older patients and patients with a history of hypertension,
dyslipidemia, and previous coronary angioplasty. Cardiogenic shock has an average time of
onset, after the appearance of clinical evidence of myocardial infarction, of 7 hours. This re-
view notes that cardiogenic shock can occur after myocardial infarctions that result in ST-
segment elevation and in those patients without ST-segment elevation accompanying the
infarction. The etiology of infarction and cardiogenic shock is usually multivessel coronary
atherosclerosis. In one report cited by the authors, more than half the patients with cardio-
genic shock after myocardial infarction had triple vessel disease, and more than 15% of pa-
tients had significant left main coronary artery stenosis. Anterior infarction is common but
multiple sites of infarction are diagnosed in more than half the patients who develop cardi-
ogenic shock. Nearly 80% of patients who develop cardiogenic shock have left ventricular
failure as the main pathogenesis of the shock event. Acute mitral insufficiency, septal rup-
ture, right ventricular infarction, and ventricular wall rupture with pericardial tamponade
account for the remaining patients. Ventricular wall rupture is the least frequent etiology of
cardiogenic shock, accounting for less than 1.5% of instances. Acute mitral insufficiency ac-
counts for 7% of patients developing cardiogenic shock.
The pathophysiology of cardiogenic shock is the subject of a review by Aymong and co-
authors30 in Medical Clinics of North America, 2007. These authors emphasize the critical
importance of reduced coronary blood flow that results in a state of myocardial oxygen
deprivation and disordered myocardial energy metabolism. They note that heart muscle is
very reliant on aerobic energy metabolism. Oxygen utilization in contracting heart muscle
ranges from 8-15 mL of oxygen/100 gm/tissue/min. The noncontracting muscle uses only
Page 35 of 119
1.5 mL/100 gm/tissue/min. The heart, even in its resting state, extracts 2-3 times more oxy-
gen than any other single organ. Myocardial oxygen demand is determined by ventricular
tension, heart rate, and contractility. Oxygen delivery to the myocardium via the coronary
arteries is primarily a diastolic phenomenon. Coronary perfusion pressure can be estimated
from the difference in aortic diastolic pressure and mean left ventricular pressure. Coronary
vascular resistance interacts with coronary perfusion pressure to determine net coronary
flow. Coronary vascular resistance is determined by complicated interactions among various
local and systemic mediators. Endothelial dysfunction occurs because of coronary athero-
sclerosis and produces abnormal coronary responses to endogenous mediators of vasocon-
striction and vasodilation. The net result is disordered autoregulation, with coronary flow
becoming progressively more dependent on coronary perfusion pressure. Diastolic coronary
perfusion can be reduced during conditions where cardiac output is decreased and when
the diastolic interval is reduced (tachycardia). Myocardial infarction resulting from coronary
thrombosis directly reduces coronary flow because of obstruction. With extensive involve-
ment of the ventricular wall, cardiac output is reduced. Tachycardia results from the cate-
cholamine response to reduced cardiac output and these factors combine to produce a self-
perpetuating cycle of myocardial ischemia and infarction. Topalian and associates29 empha-
size that “ischemia begets ischemia” during this cycle.
Ventricular dysfunction leads to elevated end-diastolic pressures and pulmonary con-
gestion. Elevated ventricular filling pressures increase myocardial wall stress. Restricted left
ventricular filling was noted in more than 60% of patients evaluated by echocardiography in
a series cited by Topalian and coauthors. This pathophysiologic process is complicated by
Page 36 of 119
the elaboration of inflammatory mediators. Clinical evidence of systemic inflammatory re-
sponse syndrome was noted in 20% of patients in a report cited by these authors. Aymong
and colleagues30 cite data indicating that elevated levels of interleukin-6 have been docu-
mented in patients with cardiogenic shock and that these levels are equivalent to the levels
recorded in patients with sepsis and septic shock. Topalian and associates and Aymong and
colleagues further note that the proinflammatory state observed in some patients with car-
diogenic shock is associated with stimulation of inducible nitric oxide. Unfortunately, block-
ade of inducible nitric oxide synthesis did not produce a benefit in patients with cardiogenic
Aymong and coauthors note that myocardial ischemia can result in myocyte necrosis.
Other cells die because of apoptosis (programmed cell death). In addition, myocardial cells
may be “stunned” from ischemia. Stunned cells do not recover full contractile function with
early restoration of perfusion from revascularization procedures. Thus, normal contractile
function might not be restored for several weeks after coronary artery thrombosis followed
The extensive multivessel coronary atherosclerosis noted in autopsy reports of patients
dying of myocardial infarction and cardiogenic shock usually produces multiple myocardial
infarctions of varying age rather than a single massive infarction. Low coronary flow and in-
creased coagulation activity because of the presence of unstable coronary plaque can pro-
duce progressive coronary thromboses that extend to involve branches of the artery where
the thrombosis originated. These authors go on to note that declining myocyte contractility
Page 37 of 119
can occur in areas remote from the infarction and this may be a consequence of production
of hypoxia-inducible-factor-1 which is stimulated when cardiac cell mitochondria sense hy-
poxia. Reduced cell metabolism and contractility are produced by this factor. The resulting
low metabolism state produced in areas remote from the infarction is termed “myocardial
hibernation.” One of the benefits of treatment of cardiogenic shock with intra-aortic bal-
loon counterpulsation may be restoration of coronary perfusion and myocyte oxygenation
that serves to reverse myocyte hibernation and recover contractility.
Monitoring the patient in shock
The clinical hallmark of shock in the critically ill surgery patient is arterial hypotension.
The Advanced Trauma Life Support Course offered by the Committee on Trauma of the
American College of Surgeons defines shock as an arterial blood pressure < 90 mmHg con-
firmed before patient arrival at the hospital and/or present during initial assessment. Addi-
tional bits of clinical information available during the initial assessment are used to supple-
ment the clinical suspicion raised by the finding of hypotension and provide guidance to the
clinician regarding hypoperfusion of critical vascular beds (oliguria, depressed level of con-
sciousness, metabolic acidosis) and prognosis (alternating tachycardia and bradycardia, re-
sponse to fluid challenge). Clinical experience has shown that significant reductions in cardi-
ac output because of hemorrhage, sepsis, cardiac dysfunction, pericardial tamponade, and
massive pulmonary embolus can be present without arterial hypotension. In fact, hypoten-
sion is not a sign of shock but a sign of decompensated shock indicating that the intrinsic
defenses against reductions of cardiac output have failed and, therefore, the hypotensive
patient is literally minutes away from cardiac arrest and death.
Page 38 of 119
Hemodynamic abnormalities resulting from reductions in cardiac output change along a
dynamic continuum. The patient who is hypotensive during transport from a crash scene to
the trauma center may have a normal arterial pressure in the resuscitation area of the hos-
pital because of ongoing adjustments of the blood volume and systemic arteriolar con-
striction induced by catecholamine secretion. Prehospital hypotension can be a marker for
patients who have bled massively or who are continuing to bleed. This topic was addressed
in an article by Lipsky and coauthors31 in the Journal of Trauma, 2006. These authors re-
viewed records on more than 1000 injured patients transported to a single trauma center
over the course of a 12-month interval. Seven percent of patients were hypotensive during
transport as defined by a systolic blood pressure of 90 mmHg or less. Of this group, more
than one-third required an operative procedure for hemorrhage control within the first six
hours at the trauma center. Underscoring the hazards of clinical classifications based on
blood pressure readings was the finding that 11% of patients who were normotensive dur-
ing the prehospital interval and who were normotensive during resuscitation also required
an operative procedure to control bleeding within the first six hours after arrival at the
trauma center. It is clear that clinical decisions about patients at risk for shock will be based
on combinations of data and that all of the variables assessed suffer from a lack of sensitivi-
ty and specificity.
In this section of the overview, a discussion of the capabilities and limitations of moni-
toring the patient in shock will be undertaken.
Page 39 of 119
Basic features of hemodynamic monitoring include intermittent or continuous meas-
urements of arterial pressure, heart rate, and arterial oxygen saturation. Arterial pressure
can be measured with blood pressure cuffs (automatic or manual) and with indwelling arte-
rial catheters attached to continuous monitoring transducers. Automatic blood pressure
cuffs produce blood pressure values consistently higher than actual pressures. This feature
is particularly noticeable in lower pressure ranges. Manual blood pressure cuffs require per-
sonnel, and the measurement process is cumbersome in critically ill patients. Placement of
arterial catheters in peripheral arteries may be challenging in vasoconstricted patients. Ar-
terial pressure measurements can be supplemented with serial measurements of central
venous pressure. In the intensive care unit or the operating room additional variables such
as central venous oxygen saturation, right heart pressures, mixed venous oxygen saturation,
and cardiac output can be measured. Transthoracic or transesophageal echocardiography
can be used to determine the presence and degree of cardiac dysfunction. Controversy con-
tinues about the additional benefit of these assessments. Isolated values derived from he-
modynamic measurements are of limited value. Trends are therefore monitored and re-
sponses to interventions such as endotracheal intubation and ventilation, decompression of
hemo- or pneumothoraces, and intravenous fluid boluses, are recorded.
A critical aspect of early resuscitation from shock involves decisions to undertake an in-
tervention to correct the underlying cause of the shock state. This might be a procedure to
stop bleeding or control the source of sepsis. Interventions may take the form of thrombo-
lysis of pulmonary embolus, revascularization of occluded coronary arteries, or insertion of
Page 40 of 119
an intra-aortic balloon pump. The clinician establishes goals of early resuscitation to help
guide these decisions. These are termed “endpoints of resuscitation.” The first article dis-
cussed in this section addresses this topic. It is by McCunn and Dutton 32 and appeared in
Current Opinion in Anesthesiology, 2000. The authors begin the discussion by confirming the
hazards of depending on specific values for monitoring variables. They stress the consistent
observation that more than one-third of injured patients in shock are not tachycardic. Ex-
perimental studies of shock from blood loss disclose that pure blood loss most often results
in slowing of the heart rate.
They go on to emphasize the importance of separating resuscitation from shock into
phases. The initial phase takes place before a needed intervention to reverse the underlying
cause of shock and the second phase occurs after the intervention has been completed.
During the initial phase, goals of resuscitation include identification of the underlying etiol-
ogy of shock and expansion of the intravascular space with intravenous electrolyte solution.
McCunn and Dutton stress the importance of careful adjustment of the arterial pressure.
The initial goal is to raise mean arterial pressure to a level that will support cell survival. At-
tempts to normalize the arterial pressure will lead to recurrence or acceleration of bleeding
in the acutely injured patient. In addition, McCunn and Dutton note that infusion of a vol-
ume of electrolyte solution sufficient to raise arterial pressure to normal will lead to exces-
sive volume infusion and complications such as abdominal compartment syndrome. They
emphasize the importance of warming intravenous fluids to support core patient tempera-
ture. Arterial pressure goals before controlling the cause of shock should be set at mean ar-
terial pressure of 60-70 mmHg with adjustments upward for elderly patients, patients with
Page 41 of 119
traumatic brain injury, and patients with known cardiac dysfunction. Supplemental infor-
mation can be gleaned from monitoring trends in central venous pressure. In the early
phase of resuscitation, risks associated with central venous catheterization include arterial
injury and pneumothorax. These risks may be reduced, somewhat, by using ultrasound
guidance for central venous catheter insertion. Central venous pressures of 8-12 cm H20 are
adequate to support cardiac output in healthy patients. Pressure goals will need to be ad-
justed upward in patients who are requiring positive pressure ventilation. As in all other
global hemodynamic measurements, trends are more important than isolated values.
A review of hemodynamic monitoring in the intensive care unit appeared in Chest in
2007, by Pinsky.33 The author begins by describing clinical objectives of hemodynamic moni-
toring. Clinicians use hemodynamic monitoring to establish cause of shock and to estimate
the direction and degree of response to therapy. The first steps are to determine how much
cardiac output is likely to increase in response to a first line intervention (usually an intrave-
nous fluid bolus) and then to determine whether the systemic vasomotor tone is increased
or decreased. Distributive shock (most commonly septic shock) is the most common shock
state encountered by surgeons where vasomotor tone is decreased. Finally, an assessment
of the ability of the heart to sustain function when arterial pressure is restored completes
the initial characterization of the shock state. Pinsky emphasizes that the response to intra-
vascular volume expansion (preload responsiveness) cannot be predicted from estimates of
ventricular end-diastolic volume. This observation calls into question the clinical value of da-
ta that can be obtained from placement of a pulmonary artery catheter.
Page 42 of 119
Whether data obtained from pulmonary artery catheter measurements alters outcomes
in patients with shock is controversial. A consensus conference produced recommendations
published in 1997.34 The recommendations pertinent to surgical patients suggested that
preoperative optimization of hemodynamic status using data derived from the pulmonary
artery catheter would probably improve outcomes in high-risk patients undergoing vascular
operations. The consensus noted that clinical decisions made using pulmonary artery cathe-
ter measurements of filling pressures and cardiac output could be useful in the manage-
ment of injured patients when standard hemodynamic measurements were deemed inade-
quate. This recommendation was based on expert opinion because no high-quality data
were available. The recommendation specifically focused on patients with brain injury and
postinjury acute respiratory distress syndrome. Overall, the data available do not support
the widespread use of the pulmonary artery catheter as a means of improving patient out-
comes in shock. In an editorial published in Critical Care Medicine in 2005, Pinsky and Vin-
cent35 addressed the lack of data supporting the use of the pulmonary artery catheter and
noted that the available studies are insufficient to document lack of efficacy of the pulmo-
nary artery catheter because the studies only compared outcomes in patients treated with
and without the device. A valid evaluation would have compared treatment protocols using
data derived from the pulmonary artery catheter compared with treatment protocols using
other hemodynamic data to support treatment decisions. These authors acknowledge the
availability of preload assessments guided by echocardiographic measurements as de-
scribed by Voga.36 Pinsky and Vincent reiterate the well-known lack of echocardiographic
capability at all times in most intensive care units.
Page 43 of 119
Currently, use of the pulmonary artery catheter to guide therapy in critically ill surgery
patients should be selective. Indirect indices of adequacy of blood volume and oxygen
transport are available. Trends of variables that are related to ventricular end-diastolic vol-
ume such as right atrial pressure and pulmonary artery occlusion pressure do not predict
accurately, changes in cardiac output in response to fluid boluses. The author goes on to
note that preload responsiveness can be predicted from alterations in arterial pulse pres-
sure in response to cycles of positive pressure ventilation.
Preload responsiveness can also be predicted from pulse pressure variation measured
using echocardiography. Obvious limitations of these assessments relate to the fact that the
patient must be on a ventilator and echocardiographic assessment must be readily availa-
ble. These assessments also lose their predictive ability when heart rate is irregular as in
atrial fibrillation. Measurement of the response of aortic flow velocity to a passive leg raise
is an assessment that can be performed in patients with any cardiac rhythm. In addition, the
patient does not have to be on the ventilator for this assessment to occur. Since flow veloci-
ty is measured by trans-esophageal Doppler techniques, special equipment is required. In
some critically injured patients, passive leg raising may not be feasible.
Because of the limitations of pressure measurements and the lack of universal availabil-
ity of other equipment, indirect assessments of the adequacy of oxygen transport have
been used and these are discussed in the next section.
Monitoring oxygenation and perfusion
As noted above, continuous measurement of arterial oxygen saturation is a component
of the basic hemodynamic assessments performed on all critically ill patients. Arterial oxy-
Page 44 of 119
gen saturation does not provide information relating to oxygen delivery, oxygen consump-
tion, or the existence of oxygen debt. Mixed venous oxygen saturation can be measured but
this requires placement of a pulmonary artery catheter. Fiberoptic central venous catheters
are now available that can provide continuous assessments of central venous oxygen satu-
ration. The absolute value for central venous oxygen saturation varies compared with abso-
lute values for mixed venous oxygen saturation (obtained using a pulmonary artery cathe-
ter). Trends of these measures in response to interventions change, in parallel, however,
suggest that trends in central venous oxygen saturation are potentially valuable clinically.
Data discussed in later sections of the overview document that a central venous oxygen
saturation value of < 70% is a dependable indicator of inadequate global oxygen transport.
Currently, this assessment is probably the most useful index of adequacy of oxygen delivery.
One obvious limitation of this measurement is in sepsis when values above 70% may be
present in the face of inadequate tissue perfusion. As a general statement, combining arte-
rial pressure and heart rate with central venous pressure measurements supplemented by
continuous measurements of central venous oxygen saturation provide adequate estimates
of hemodynamic function to guide the clinician dealing with most forms of shock encoun-
tered in surgical patients.
A fundamental concept of shock is that inadequacy of the circulation leads to a cellular
shift from aerobic to anaerobic metabolism. Anaerobic metabolism leads to the accumula-
tion of acid metabolites in tissue. Traditionally, assessments of acid base status have been
routine in the management of shock. Assessment of global acid-base status has usually
been achieved by calculating base deficit from values for pH, bicarbonate and PC02 obtained
Page 45 of 119
from arterial blood gas analysis or from assessments of bicarbonate levels obtained from
venous blood samples. Lactic acid is a metabolite that accumulates in cells that are metabo-
lizing energy substrates anaerobically. Serum lactate levels reflect lactic acid accumulation
and are relatively easy to measure. Absolute lactate levels and rates of lactate clearance
have been used as a means of predicting mortality in patients in shock. For example, the
rate of lactate clearance has been shown to be predictive of outcomes in injured patients.
Whether lactate levels can be used to guide therapeutic decisions is less clear.
We discuss several articles dealing with global assessments of acidosis, lactate metabo-
lism, and the roles measurements of these variables may play in managing patients in
The relationship between serum bicarbonate (which can be measured from a venous
sample) and base deficit (which requires an arterial blood sample) is the topic of an article
by Martin and coauthors37 in Archives of Surgery, 2005. These authors reviewed records
from intensive care unit stays for 2291 patients. The patients were critically ill surgery pa-
tients; injured patients were excluded. Pairs of samples compared serum bicarbonate levels
and calculated base deficit. The authors found that trends of the two measurements
changed in parallel. Admission serum bicarbonate predicted mortality and length of stay.
Defining significant acidosis as a base deficit > 5, the authors determined that serum bicar-
bonate was an accurate measure of acid-base status. The authors concluded that serum bi-
carbonate was as accurate as base deficit diagnosing acidosis and that low serum bicar-
bonate was a dependable trigger for treatment of acidosis. This same group of investigators
Page 46 of 119
extended the comparison of serum bicarbonate and base deficit to a group of 3102 injured
patients admitted to a trauma intensive care unit. These data are found in a report by
Fitzsullivan and coauthors38 in the American Journal of Surgery, 2005. The authors note that
base deficit is a calculated value derived from data gathered from analysis of an arterial
blood sample. The value represents the number of milliequivalents of additional base buffer
that would have to be added to bring the pH of the blood sample to normal. Serum bicar-
bonate is measured from a venous sample. Potential advantages of using serum bicar-
bonate are ease of collection and processing as well as lowered cost. The analysis disclosed
that serum bicarbonate was an accurate predictor of acid base status and had equivalent
clinical value compared with base deficit. Most, but not all, of the variability of base deficit
was paralleled by serum bicarbonate values. The authors concluded that serum bicarbonate
was as accurate as base deficit in predicting clinically significant acidosis.
In a final analysis by this group of investigators, a comparison of base deficit to serum
lactate was carried out to determine whether base deficit was as accurate as lactate levels
for prediction of outcomes in critically ill surgical patients. They confirmed that serum lac-
tate was a superior predictor of outcomes compared with base deficit. They concluded from
their analysis that admission base deficit and admission lactate predict mortality but that a
normal base deficit does not always predict a normal lactate and that an abnormal lactate
level was predictive of mortality and length of stay even when the base deficit was normal.
Additional discussion of lactate as a clinical indicator of occult anaerobic metabolism is dis-
cussed in the next segment.
Page 47 of 119
A review of lactate metabolism and an assessment of the clinical utility of lactate level
measurements are found in an article by Jansen and coauthors39 in Critical Care Medicine,
2009. The article is entitled “Blood lactate monitoring in critically ill patients: a systematic
health technology assessment.” This article is supplied as a full-text reprint accompanying
this issue of SRGS. The objective of this report is to attempt to describe the process of
measuring blood lactate and to establish whether these measurements have value in the
management of the critically ill patient. The authors note that there is wide variability
among clinicians and hospitals in the use of blood lactate measurements. Although elevated
blood lactate is thought to be evidence of anaerobic metabolism that can result from shock
and other conditions, there is evidence that elevations of lactate can occur in the absence
of anaerobic metabolism. The authors note that the measurement of blood lactate has
been standardized and the test is dependable. They review experimental and clinical studies
that have evaluated lactate. Experimental preparations that have reduced oxygen delivery
until oxygen consumption was delivery-dependent have consistently shown elevations of
blood lactate in the experimental subjects. Studies of patients with septic and cardiogenic
shock have shown increased lactate/pyruvate ratios consistent with anaerobic metabolism.
Jansen and coauthors cite a study by Rivers40 (discussed later in the overview) that doc-
umented a consistent relationship between hyperlactatemia and low central venous oxygen
saturation indicating inadequate oxygen delivery and suggesting that a shift to anaerobic
metabolism had occurred. They also cite data demonstrating that elevations of lactate con-
sistently occurred in patients with low mixed venous oxygen saturation, low cardiac output,
and clinical signs of shock. The authors stress that there is wide variability in regional oxy-
Page 48 of 119
gen delivery and demand. This fact means there will be no specific value for oxygen delivery
or central venous oxygen saturation associated with lactate accumulation. They describe re-
search conducted in patients with septic shock that demonstrate lowering of lactate levels
when capillary perfusion is improved, even though systemic hemodynamic variables did not
change. This observation suggests that local microcirculatory changes may alter the accu-
mulation of lactate.
The authors go on to review instances where lactate levels may change in the absence
of anaerobic metabolism. Conditions that increase aerobic glycolysis to a point where the
ability of pyruvate dehydrogenase is exceeded by over production of pyruvate will result in
accumulation of lactate. Mitochondrial dysfunction and liver dysfunction can alter lactate
production and clearance with increased lactate levels in the absence of anaerobic metabo-
lism. The lung can also be an important source of lactate when pulmonary production of
lactate is stimulated by inflammatory mediators, even though hypoxia is not present. Final-
ly, alkalosis can increase efflux of lactate from cells and several drugs can stimulate lactate
The article next reviews the prognostic value of lactate and the authors conclude that
lactate elevations accurately predict mortality in acutely injured patients in the emergency
department and critically ill patients in the intensive care unit. The authors reviewed several
studies that confirmed an inconsistent relationship between lactate and other indices of ac-
idosis. The bulk of the evidence suggests lactate levels cannot be predicted from other indi-
ces of acidosis such as base deficit and anion gap. Further, the evidence suggests that ele-
Page 49 of 119
vated lactate, along with evidence of acidosis, strongly suggests a shift to anaerobic metab-
olism in a large proportion of circulatory beds. The authors conclude that the available evi-
dence suggests that lactate changes can be used to indicate the effectiveness of therapeutic
efforts to improve oxygen delivery and tissue blood flow.
For surgeons managing critically ill patients, lactate is most useful when the elevated
lactate level occurs in the setting of metabolic acidosis; acidosis is independently confirmed
by measurements of serum bicarbonate and/or base deficit. The presence of elevated lac-
tate with metabolic acidosis is predictive of mortality and resolution of hyperlactatemia
suggests successful reversal of anaerobic metabolism.
The final article reviewed in this section deals with the potential value of laboratory
tests in the management of shock by Friese and coauthors41 appeared in the Journal of
Trauma, 2007. These authors hypothesized that the transition from adequate resuscitation
of shock to over resuscitation would be associated with increased myocardial wall stretch
and serum levels of B-type natriuretic peptide might be valuable indicators of this transi-
tion. Serum levels of B-type natriuretic peptide were followed in 134 injured patients. Chest
radiographs obtained 24 hours after admission were graded for the presence of pulmonary
edema, which the authors used as an indicator of over resuscitation. Twenty patients who
had increased pulmonary edema scores had mean serum levels of B-type natriuretic pep-
tide that were significantly higher than the mean levels observed in patients without pul-
monary edema. This paper was presented during the plenary session of the annual meeting
of the American Association for the Surgery of Trauma and the discussion that followed the
Page 50 of 119
presentation is included with the article. The discussants note the non-specific nature of
single chest radiograph assessments of pulmonary edema. There was no association be-
tween serum levels of B-type natriuretic peptide and any clinically meaningful endpoint. It
seems safe to conclude that the search for additional laboratory indicators of resuscitation
endpoints should continue.
Monitoring at the tissue level
Global markers of inadequate oxygen delivery to tissue such as central venous oxygen
saturation, base deficit, and lactate levels represent the resultant vector of all vascular
beds. There has been wide recognition that some vascular beds (especially the splanchnic
bed) are persistently underperfused and hypoxic even though global indices of oxygen sup-
ply and consumption have been corrected to normal levels. Studies in humans and experi-
mental animals have demonstrated significant variability in blood flow and recovery of
normal cellular metabolism in resuscitated hemorrhagic shock. Accessible microvascular
beds evaluated in clinical studies include skin, skeletal muscle, the sublingual bed, and the
gastric mucosa. Mucosal pH and PC02, tissue oxygen saturation, and tissue oxygen tension
have been studied as possible indices of successful resuscitation. Data from these evalua-
tions are discussed in this segment of the overview.
Values for pH and PCO2 can be measured in the gastric and sublingual mucosa using sa-
line-filled latex chambers applied to the mucosal surface. Because the chamber is permea-
ble to C02, the carbon dioxide tension of the mucosa is equivalent to the carbon dioxide
tension of the saline within the chamber after equilibration, and this can be measured in a
standard blood gas analyzer. Simultaneous measures of blood bicarbonate permit calcula-
Page 51 of 119
tion of the mucosal pH. An important assumption included in these measurements is that
mucosal and blood bicarbonate levels are equivalent. This assumption is controversial.
The first article reviewed for this discussion is a classic clinical study by Ivatury and coau-
thors42 in the Journal of the American College of Surgeons, 1996, who conducted a random-
ized prospective trial in which 57 severely injured patients were resuscitated from shock us-
ing global indicators of oxygen delivery and consumption. Consistent with the practice at
the time, resuscitation to “supranormal” levels of oxygen delivery and consumption was
performed. Gastric mucosal pH was assessed using the tonometry technique described
above. Elevation of the gastric mucosal pH to a level > 7.3 was defined as adequate tissue-
level resuscitation. Forty-four patients reached target levels of gastric mucosal pH and three
of these developed multiple organ failure. Fifty-four percent of the patients who did not
reach target levels for gastric mucosal pH developed multiple organ failure. The authors
noted that low levels of gastric mucosal pH were observed in the hours before a serious
complication such as abdominal compartment syndrome was diagnosed.
These authors concluded that patients who reach target levels of gastric mucosal pH are
less likely to develop organ failure and the development of low gastric mucosal pH may be a
warning signal of an impending serious complication. Unfortunately, this study did not show
that resuscitation to a target level of gastric mucosal pH was superior to resuscitation using
global indicators of oxygen delivery and consumption. In fact, the frequency of complica-
tions such as organ failure and abdominal compartment syndrome were equivalent in the
Page 52 of 119
group resuscitated to a target mucosal pH compared to the group resuscitated using global
The evidence for and against the use of gastric mucosal pH as a resuscitation target is
reviewed in an article by Hameed and Cohn43 in Chest, 2003. These authors reviewed the
technique of gastric tonometry and concluded that the technique was standardized and val-
id but the controversy concerning equivalency of blood and mucosal bicarbonate had not
been resolved. They made note that newer tonometry probes had eliminated the need for
equilibration with saline and that this technology had simplified the assessment technique.
They then reviewed the evidence supporting the use of gastric tonometry in critically ill pa-
tients and concluded that the evidence was conclusive; persistent gastric mucosal pH levels
< 7.32 predict mortality and complications. In a review of available evidence from trials
evaluating gastric mucosal pH as a resuscitation target, Hameed and Cohn concluded that
there was no evidence that resuscitation to a target gastric mucosal pH improved outcomes
in critically injured patients or patients with septic shock. Because a clinically definable ben-
efit has not been established, the use of gastric tonometry has declined.
The sublingual mucosa has a microcirculatory anatomy and physiology similar to the
gastric mucosa. An available probe will rapidly measure sublingual mucosal carbon dioxide
tension and this may be helpful in establishing prognosis and in defining adequate resusci-
tation. A review of the technical aspects of sublingual capnometry is the topic in an article
by Boswell and Scalea44 in AACN Clinical Issues, 2003. A clinical evaluation of this technique
was published in the Journal of Trauma, 2007 in a report by Baron and coauthors.45 Baron
Page 53 of 119
and colleagues obtained measurements of sublingual mucosal pC02, base deficit, and lac-
tate serially during the first 48 hours after admission of 86 severely injured patients to two
Level I trauma centers. Their data indicated that the ability of the three tests to predict mor-
tality was equivalent. All three tests demonstrated sensitivity > 60% and specificity > 87%
for prediction of mortality. The authors concluded that sublingual capnometry was feasible
from a technical standpoint but was not superior to conventional global assessments of ox-
ygen delivery and consumption as means of predicting outcomes.
It is possible to assess tissue oxygen saturation using near infrared spectrophotometry.
A study assessing this technology as a method for detecting patients with shock who will
require massive transfusion is authored by Moore and coauthors46 in the Journal of Trauma,
2008. This study was a multicenter analysis involving seven Level I trauma centers. One
hundred fourteen patients were categorized as massive transfusion patients. All these pa-
tients were monitored with standard measures and subcutaneous near infrared spectrome-
try measurements of oxygen saturation. The authors found that persistently low tissue oxy-
gen saturation (< 70%) was predictive of outcomes. They found that low admission systolic
blood pressure and base deficit > 11 accurately predicted the need for massive transfusion.
They concluded from their analysis that monitoring subcutaneous oxygen saturation was
potentially helpful for establishing prognosis. Shock associated coagulopathy and massive
transfusion are discussed in a later section of this overview.
Additional data confirming the predictive value of subcutaneous oxygen saturation lev-
els is reported by Leone and coauthors47 in Anesthesiology, 2009. These authors measured
Page 54 of 119
muscle tissue oxygen saturation using near infrared technology in a group of patients who
fulfilled diagnostic criteria for septic shock. Initial measurements were obtained after global
indices of oxygen delivery and consumption were optimized. Tissue oxygen saturations
were lower (mean 73%) in nonsurvivors compared with surviving patients (mean 84%). The
authors concluded that these measurements were useful for establishing prognosis. To
date, there are not sufficient data to support monitoring of tissue oxygen saturation as a
method for gauging adequacy of resuscitation.
An experimental study evaluating measurements of muscle tissue oxygen tension and
muscle blood flow as indices of successful resuscitation is reported in an article by Wan and
coauthors48 in the Journal of Trauma, 2009. This report provides data from a series of ex-
periments in a porcine model. Separate groups of animals were instrumented for monitor-
ing of muscle and brain blood flow as well as muscle and brain oxygen tension. Tissue blood
flow was measured by thermal dilution and tissue oxygen tension by polarographic probes.
Oxygen challenges were administered during each experimental phase by increasing in-
spired oxygen tension to 500 mmHg. Experimental phases included baseline, administration
of phenylephrine to raise mean arterial pressure to 150 mmHg, and hemorrhage to an initial
mean arterial pressure of 60 mmHg and, finally, further hemorrhage to a mean arterial
pressure of 40 mmHg. No resuscitation was carried out. The authors found that muscle oxy-
gen tension increased with oxygen challenge at baseline and muscle blood flow did not
change. With vasopressor administration, muscle blood flow decreased significantly while
brain blood flow increased. Muscle oxygen tension trended downward but the decrease
was not statistically significant. With hemorrhage, brain and muscle blood flow and brain
Page 55 of 119
and muscle oxygen declined in parallel. The authors concluded that muscle oxygen tension
and brain oxygen tension are governed mainly by blood flow and to a lesser extent by oxy-
gen tension in arterial blood. These findings support additional research efforts directed
toward finding ways to improve microcirculatory flow in all vascular beds.
The authors further concluded that muscle oxygen tension has potential as a continuous
measure of microcirculatory flow. In the discussion section of their paper, they point out
the divergent changes in tissue flow and oxygen levels in brain and muscle with vasopressor
infusion. They note that current efforts to increase brain blood flow with vasopressor
agents may have negative effects on other body tissues. This work reconfirms multiple ob-
servations confirming the heterogeneity of microvascular responses in animals and humans
who sustain shock from hemorrhage, sepsis, or other etiologies.
Additional investigation should clarify the complex interrelationships among vascular
beds and permit development of more precise approaches to resuscitation in the future.
Routine use of tissue oxygen measurements, though theoretically attractive, is not support-
ed by sufficient data at this time. The difficulty in using tissue tonometry as a continuous
monitoring methodology has limited its use. Global indices of oxygen delivery and con-
sumption such as central venous oxygen saturation, base deficit, and lactate levels are valu-
able indices when used as supplements to sequential clinical assessments during resuscita-
Page 56 of 119
Newer monitoring technologies
Recently, technology has become available to measure cardiac output noninvasively.
The device uses assessments of aortic blood flow derived from Doppler ultrasound analysis.
The probe is placed in the suprasternal notch and spectral arrays of flow patterns are ana-
lyzed with computer assistance. A single-center, prospective observational study comparing
noninvasive cardiac output measurements with thermodilution cardiac output measure-
ments is by van Lelyveld-Haas and coauthors49 in the European Journal of Anesthesiology,
2008. These authors compared 263 measurements obtained sequentially with the noninva-
sive device with an identical number of thermodilution cardiac output measurements. The
authors found that noninvasive measurements were consistently 12% lower than
thermodilution measurements. The random error rate was 17% and consisted of a 3% inter-
observer variability; an 11% rate of inter-patient variability. The learning curve required to
master the technique of noninvasive cardiac output measurement was significant. Despite
the variability, the authors concluded that the technique was acceptable for detection of
shock in critically ill patients.
An informative comparison of currently available cardiac output devices comes from an
article by Compton and Schäfer in Seminars in Cardiothoracic and Vascular Anesthesia,
2009. These authors concluded that the only available cardiac output devices with accepta-
ble performance were not noninvasive, in the strictest sense of the term, but minimally in-
vasive because each of the two acceptable devices require an indwelling arterial catheter.
Lithium indicator dilution and pulse wave analysis are the methods used by these two de-
vices. Interested readers are encouraged to review this article.
Page 57 of 119
The objective of resuscitation from shock is to restore tissue blood flow and oxygen de-
livery by manipulating systemic arterial pressure, cardiac preload, intravascular volume, and
blood oxygen carrying capacity. Traditionally, resuscitation from shock because of injury,
and hemorrhage, septic shock, and cardiogenic shock have shared several common fea-
tures, namely, an initial fluid bolus using electrolyte solutions followed by red blood cell
transfusion to improve oxygen carrying capacity. If cardiac dysfunction is present, vasopres-
sor and/or inotropic drug therapy are added. It is clear that factors other than fluid and
blood therapy make important contributions to the success of resuscitation. For shock from
hemorrhage, identification of bleeding sites and rapid control of bleeding is the central fac-
tor leading to successful resuscitation. Septic shock patients require identification of the
source of infection and control of that source, if possible, to ensure successful resuscitation.
Patients with cardiogenic shock may benefit from coronary revascularization and/or intra-
aortic balloon pump support. Patients with “obstructive” shock (pulmonary embolus, ten-
sion hemothorax or pneumothorax, pericardial tamponade) require relief of the obstruction
to circulation in order for resuscitation to be successful.
In this section of the overview, we discuss resuscitation strategies. The discussion re-
views fundamental understandings of electrolyte solution resuscitation, colloid resuscita-
tion, hypertonic resuscitation, and the use of nonblood oxygen carrying solutions. Blood
and blood product use are discussed in separate sections of the overview.
Page 58 of 119
Resuscitation with electrolyte solutions
Electrolyte resuscitation from shock has been conceptually based on experimental and
clinical studies of hemorrhagic and burn shock begun by Carl Moyer in the early 1960s. Mo-
yer’s hypothesis suggested that shock produced a deficit of sodium and water in the inter-
stitial space and if this deficit remained uncorrected, resuscitation would not be successful.
Later, Shires and numerous colleagues conducted experiments demonstrating a shift of wa-
ter and sodium into the intravascular space as the body responded to decreased blood vol-
ume by moving fluid along pressure gradients from the interstitium into the intravascular
space. These studies also identified a second mechanism of interstitial fluid depletion, a
movement of water and salt into cells as the shock insult overcame all compensatory de-
fenses and entered the decompensated phase. Clinical recommendations evolved from the-
se data and replacement of blood volume supplemented by a volume of electrolyte solution
equal to 20% of the calculated interstitial volume became the norm for managing patients
in shock. Protocols for the management of traumatic shock emphasized the use of electro-
lyte solution as an initial means of restoring blood pressure to normal while typed and
crossmatched blood was prepared. For patients undergoing major operative procedures
formulas were suggested that led to hourly electrolyte infusions of 2-3 liters.
Unfortunately, independent investigators could not consistently duplicate Shires’s ex-
perimental results, and questions arose about the presence and the magnitude of the inter-
stitial deficit. The emergence of acute respiratory distress syndrome as a major complica-
tion of trauma and shock led to concern that overzealous infusion of electrolyte solution
would lead to pulmonary damage. Previous discussion in the overview has stressed that the
Page 59 of 119
volume of electrolyte solution necessary to raise arterial blood pressure to normal will lead
to significant tissue edema and, most importantly, to recurrent bleeding if control of hem-
orrhage is not achieved.
The importance of targeting lower arterial pressures during the early phases of traumat-
ic shock resuscitation in order to prevent recurrent hemorrhage until definitive control of
bleeding can be achieved was amply demonstrated in the clinical trial reported by Bickell
and coauthors50 in the New England Journal of Medicine, 1994. Recommendations for blood
pressure targets in the interval before hemorrhage is controlled have been for volume ex-
pansion to achieve mean arterial pressures of 60-70 mmHg. Higher pressures may be neces-
sary in elderly patients, patients with cardiac disease, and patients with traumatic brain in-
jury where maintenance of cerebral perfusion pressure is important.
The foregoing discussion of the pathophysiology of shock has documented that a shift of
ions and water into cells (endothelial cells and skeletal muscle cells) does occur but formula-
based replacement of ion and water deficits is not as effective as resuscitation based on
achieving hemodynamic, oxygenation, and blood flow goals.
The first article discussed, by Lucas,51 was presented as the Scudder Oration in Trauma
at the annual Clinical Congress of the American College of Surgeons in 2000. The article is
entitled “The water of life: a century of confusion.” This article is supplied as a full-text re-
print accompanying this issue of SRGS. Lucas reviews his perspective on the dynamics of
protein and fluid flux during shock and resuscitation. He notes that albumin and fluid enter
and leave the intravascular space and the interstitial space based upon physiologic rules
Page 60 of 119
such as Starling’s law of the capillary. He notes that there is sound evidence for a three-
phase understanding of shock as it is observed in the severely injured patient. Phase one
consists of the interval of shock and active hemorrhage, phase two is a phase of obligatory
sequestration of extracellular fluid that begins when bleeding is controlled. The volumes of
blood and electrolyte fluid necessary to maintain life while bleeding is controlled produces
the interstitial edema that characterizes this phase. Phase three consists of mobilization of
fluid and diuresis. Lucas stresses the importance of lymphatic transport of albumin from the
interstitial space to the intravascular space during phase three of trauma and shock. This
explains the rise in serum albumin levels frequently observed during this phase. Lucas em-
phasizes the critical importance of adequate replacement of blood volume and electrolyte
as critical features of successful resuscitation from shock. He notes the lack of evidence of
benefit of colloid based resuscitation strategies and closes by restating the critical im-
portance of early hemorrhage control.
Lucas next reviews data on colloid-rich resuscitation. He notes that studies of albumin
resuscitation in hemorrhagic shock have consistently demonstrated higher mortality rates
compared with resuscitation strategies using blood and electrolyte solutions. Newer col-
loids, such as hydroxyethyl starch and dextran, may adversely affect the coagulation mech-
anism. and these agents have not proven superior to resuscitation strategies based on
blood and electrolyte fluid replacement.
Evidence that serves to guide the choice of electrolyte solution for shock resuscitation is
reported by Todd and coauthors52 in the Journal of Trauma, 2007. These authors compared
Page 61 of 119
infusions of lactated Ringer’s solution to normal saline in a porcine model of hemorrhagic
shock. They chose to resuscitate the animals to their baseline blood pressure. As mentioned
above, this would not be desirable in patients before definitive control of bleeding. They
noted that animals resuscitated with normal saline required significantly larger infusion vol-
umes, developed hyperchloremic metabolic acidosis, and had significantly lower levels of
coagulation factors at the end of resuscitation.
Several recent reports and reviews have emphasized possible complications of aggres-
sive electrolyte solution resuscitation including higher postoperative complication rates,
impaired neutrophil function, augmented inflammatory mediator production, and acute
lung injury.53-54 Lactated Ringer’s solution is a racemic mixture containing equal amounts of
d- and l-lactate. Only l-lactate can be metabolized in the liver. Some have speculated that
the d-lactate isomer may be the cause of the effects on immune function associated with
experimental studies of electrolyte solution resuscitation. In the future, it is possible that a
single isomer solution will be available. It is clear that electrolyte resuscitation is most bene-
ficial if used in targeted fashion to restore microcirculatory flow and oxygen transport. For-
mula-based fluid volume prescriptions clearly increase the risk of complications in patients
not in severe shock and in patients where rapid control of bleeding is achieved. Despite ob-
servations from animal experiments that potentially harmful changes in immune function
result from the use of lactated Ringer’s solution, there currently is not a cost-effective sub-
stitute fluid that has a similar proven safety record.
Page 62 of 119
Resuscitation with hypertonic solutions
The most common hypertonic solution used for the management of patients in shock is
7.5% saline combined with 6% dextran 70. A comprehensive review of the scientific basis
for hypertonic resuscitation of hemorrhage is detailed by Kramer55 in the Journal of Trauma,
2003. This author provides an informative review of the animal experiments that served to
support early clinical trials. Hypertonic saline/dextran rapidly raises arterial pressure by
moving water from erythrocytes and, possibly, endothelial cells into the intravascular space.
Microcirculatory vasodilation occurs, possibly because of movement of water from endo-
thelial cells. With fixed dose hypertonic resuscitation, the degree of blood pressure rise is
not predictable and a risk of recurrent bleeding exists if blood pressure is raised to the point
that clot adherent to injured vessels is displaced. Most studies of hypertonic resuscitation
have shown that, over time, an amount of isotonic electrolyte solution is required that is
equivalent to the amount that would have been administered had resuscitation been car-
ried out with isotonic electrolyte solutions. Hypertonic saline dextran does function well as
a means of increasing cerebral perfusion pressure and restoring cerebral blood flow in pa-
tients (especially children) with traumatic brain injury and this setting is where the current
use of hypertonic resuscitation solutions is common.
Design and use of resuscitation strategies to improve cerebral perfusion pressure is the
focus of a report by Meybohm and coauthors56 in the Journal of Trauma, 2007. These au-
thors used small volume hydroxyethyl starch for resuscitation of porcine hemorrhagic shock
and compared the use of vasopressin and norepinephrine to maintain and improve cerebral
perfusion pressure. Brain energy metabolism was assessed with microdialysis techniques
Page 63 of 119
assaying for glucose, glycerol, lactate and the lactate/pyruvate ratio in brain tissue assessed.
These authors showed that cerebral perfusion pressure was maintained with either vaso-
pressor agent and that cerebral energy metabolism was protected in the groups treated
with either agent.
There is a focus on early aggressive maintenance of cerebral perfusion pressure in pa-
tients with traumatic brain injury. This fact, combined with concern over possible adverse
consequences of aggressive electrolyte fluid resuscitation, has led to increasing use of vaso-
pressor drugs in early shock resuscitation. This approach is the topic of a report authored by
Sperry and coauthors57 in the Journal of Trauma, 2008. The authors reviewed data from a
large multicenter trial of treatment for injured patients in hemorrhagic shock. They per-
formed a Cox proportional hazards analysis to determine the influence on patient out-
comes, of the use of norepinephrine, phenylephrine, dopamine, and/or vasopressin within
the first 24 hours after injury. Their analysis disclosed an almost 80% increase in the risk of
death for patients who were administered these drugs. After adjustment for several con-
founding factors, the authors showed that early vasopressor use was an independent driver
of the risk of mortality. Interestingly, they showed that aggressive electrolyte solution infu-
sion (more than 16 liters in the first 24 hours after injury) was an independent predictor of
patient survival. This latter finding suggests that the “adverse” affects of large volumes of
electrolyte solutions demonstrated in animal experiments and in patients undergoing elec-
tive operation (who were not in shock) are absent in patients sustaining injury and shock.
Page 64 of 119
Hypertonic resuscitation may have salutary effects on the proinflammatory state that is
documented after traumatic shock in some patients. This is the topic of a randomized pro-
spective trial of hypertonic resuscitation carried out in 27 patients in a single trauma center.
This report, by Rizoli and coauthors58 appeared in Annals of Surgery, 2006. In this study, pa-
tients sustaining blunt trauma who had a history of at least one blood pressure reading of
90 mmHg or less were randomized to receive a single 250 mL infusion of hypertonic sa-
line/dextran or normal saline. A diverse array of immune variables was assessed during the
initial hospitalization. The authors noted that patients who received hypertonic sa-
line/dextran had lower levels of proinflammatory cytokine release and a more “balanced”
immune response compared with patients receiving normal saline.
Experimental studies have shown that hypertonic resuscitation, combined with a
vasodilating agent, may attenuate ischemic injury to the gut resulting from shock. This topic
is dealt with in a report by Deree and coauthors59 from the Journal of Trauma, 2007. These
authors showed, in previous studies, that addition of pentoxifylline, a nonspecific
phosphodiesterase inhibitor, to hypertonic saline resuscitation of hemorrhagic shock atten-
uated histologic evidence of injury in the terminal ileum. In this analysis, they explored the
hypothesis that this form of resuscitation attenuates gut damage by down regulating the
activity of inducible nitric oxide synthase. They were able to show, in this experimental
model, that inducible nitric oxide synthase was, in fact down regulated.
One effect of gut ischemia and hypoperfusion seen in experimental simulations of trau-
ma and shock is altered function of neutrophils. Abnormal neutrophil function has also been
Page 65 of 119
linked to resuscitation with isotonic lactated Ringer’s solution. This same group of investiga-
tors60 subjected blood drawn from normal, healthy volunteers to stimulation with endotox-
in and examined the actions of two hypertonic solutions (3% and 7.5% NaCl) and
pentoxifylline on activation of neutrophil by assessing CD14 and CD11b expression and tu-
mor necrosis factor alpha production. They found that hypertonic saline combined with
pentoxifylline prevented neutrophil activation and tumor necrosis factor production more
effectively than hypertonic saline alone. The authors note that these data lend support to
the use of pentoxifylline as an adjunct to hypertonic saline in experimental models of trau-
The authors speculate that the mechanism of reduced tumor necrosis factor secretion is
from the action of pentoxifylline on intracellular pathways such as the cyclic adenosine
monophosphate response element binding (CREB) protein. They further speculate that the
action of pentoxifylline leading to reduction of adhesion molecule expression occurs via the
reduction of intracellular kinase such as P38 MAPK. These demonstrated antiinflammatory
effects, combined with the ability of pentoxifylline to improve microcirculatory function,
imply that this combination could be potentially beneficial in clinical models of traumatic
shock. These findings, taken together with the research studies of Zakaria and Matheson,14
discussed earlier, lend support to the search for resuscitation strategies that target normali-
zation of distal gut microcirculatory function.
Resuscitation with non-blood oxygen carrying solutions
Recognition of the facts that banked and fresh units of red cells and blood for transfu-
sion are scarce resources especially in remote locations and in combat conditions has stimu-
Page 66 of 119
lated a search for nonblood oxygen carrying solutions that could be used for resuscitation of
injured patients and to substitute for blood transfusion in elective surgery patients. In this
segment of the overview, a review of the current experimental and clinical research relating
to the ability of oxygen-carrying solutions to reverse the microcirculatory changes leading to
the oxygen supply/consumption imbalance associated with various shock states is dis-
cussed. The potential usefulness of these solutions as a means of avoiding transfusion in
elective and emergency surgical conditions also is reviewed. The outlook for reducing the
risks of allogeneic transfusion and reducing the demand for donated blood will also be de-
The first article reviewed in this section of the overview by Creteur and Vincent 61 ap-
peared in Critical Care Clinics, 2009. This review is entitled “Potential uses of hemoglobin-
based oxygen carriers in critical care medicine.” This article is included as a full-text reprint
accompanying this issue of SRGS. The authors open the review by pointing out the main
drivers of the effort to develop nonblood oxygen carrying solutions. Concern about poten-
tial risks of transfusion, particularly transmission of disease and the possibility that a global
shortage of donated blood might occur, have been the main influences stimulating this ef-
fort. The characteristics of an acceptable oxygen carrying solution would include a better
shelf life than banked red blood cells, universal compatibility, physiologically useful intra-
vascular half-life, and absence of infection risk.
Of the available solutions, shelf life is one to two years and intravascular half-life for
most solutions is in the 12-24 hour range. None requires compatibility testing. Preparation
Page 67 of 119
processes have included measures to prevent infection transmission of bacteria, viruses,
and prions. The currently available solutions are derived from human, bovine, or recombi-
nant hemoglobins. The hemodynamic effects of the solutions derive from their oncotic ef-
fects and nitric oxide scavenging actions. The currently available hemoglobin solutions all
improve tissue oxygen delivery through a combination of oxygen delivery and blood flow
improving actions. The oncotic effects of hemoglobin solutions serve to increase blood vol-
ume. Keeping solution viscosity near that of blood serves to improve microvascular flow.
The vasopressor effects that derive from nitric oxide scavenging can be offset by encasing
the hemoglobin molecules in liposomes. Vasoconstriction can also occur when the cellular
oxygen tension in the resistance arterioles increases as a result of off-loading of oxygen
from the hemoglobin molecules. This feature of hemoglobin solutions was confirmed in an
experimental study reported by Koehler and coauthors62 who showed that brain arteriolar
vasoconstriction correlated with off-loading of oxygen to arteriolar cells.
Creteur and Vincent stress that modification of hemoglobin solutions that serve to re-
duce the vasoconstrictive affects of the solutions result in more efficient oxygen delivery.
Solutions that have high viscosity, high oxygen affinity, and high colloid osmotic pressure re-
sult in better tissue oxygenation because both oxygen delivery and microcirculatory flow
Creteur and Vincent note that avoidance of red blood cell transfusion has become the
main index of efficacy in clinical studies of hemoglobin solutions. They cite clinical studies of
hemoglobin solution resuscitation of injured patients in shock that have shown reduced red
Page 68 of 119
blood cell transfusion volumes and equivalent mortality in groups treated with hemoglobin
solutions compared with conventional red blood cell based resuscitation methods. In stud-
ies of elective surgical procedures, most patients treated with hemoglobin solutions have
been able to avoid allogeneic blood transfusion.
Despite these beneficial effects, there remains concern that the vasopressor and
vasoconstrictive effects of hemoglobin solutions lead to an increased risk of stroke and my-
ocardial infarction. This concern has been emphasized in a meta-analysis authored by
Natanson and coauthors63 in the Journal of the American Medical Association, 2008. These
authors analyzed five clinical trials involving nearly 4000 patients; they documented a statis-
tically significant increase in the risk of myocardial infarction and stroke in patients treated
with hemoglobin solutions. This meta-analysis has been criticized because of the variability
of experimental situations and hemoglobin solutions studied. A review of the clinical usage
of hemoglobin solutions in clinical situations, by Napolitano,64 appeared in Critical Care Clin-
ics, 2009. In addition to noting the potential shortcomings in the meta-analysis mentioned
previously, she notes that the safety record of hemoglobin solution use in South Africa
seems to be satisfactory. Nonetheless, this meta-analysis and other safety concerns have
prevented approval of hemoglobin solutions for clinical use in the United States.
The most detailed analysis of hemoglobin solution use in trauma patients comes from
Moore and coauthors65 in the Journal of the American College of Surgeons, 2008. This re-
port presents data from a randomized trial of hemoglobin solution use in seriously injured
patients who sustained, for the most part, blunt force injuries. The authors noted a signifi-
Page 69 of 119
cant reduction in allogeneic transfusion in the patients randomized to receive the hemoglo-
bin solution. Mortality was equivalent in both treatment groups. Adverse events were,
however, more frequent in patients receiving hemoglobin solutions. These authors con-
cluded that hemoglobin solution use would be justified in situations where blood might not
be available. It is clear that nearly 50 million Americans live more than one hour from a des-
ignated or verified trauma center. These patients are at risk for being injured in a setting
where blood will not be readily available. Similarly, combat injuries are likely to occur in set-
tings where blood is not readily available. Hemoglobin solution risk would probably be less
than the documented benefit in both situations.
Data in support of this benefit is found in a study from Dudkiewicz and coauthors.66 This
study evaluated hemoglobin solution resuscitation in an experimental model of hemor-
rhage, thoracic injury, and brain injury. In the initial interval, designed to simulate the
prehospital phase of care, the animals received either saline or hemoglobin solution. There-
after, the animals received care designed to simulate protocols of care recommended for
patients with severe traumatic brain injury. Treatment in this phase, designed to simulate
the emergency department and in-hospital phases of care, focused on maintaining cerebral
perfusion pressure and oxygenation. The authors noted that the saline resuscitated animals
had maintenance of systemic oxygenation but did not demonstrate restoration of cerebral
oxygenation until the 120-minute time point when mannitol was given. The animals receiv-
ing hemoglobin solution had restoration of cerebral blood flow and oxygenation within
minutes of receiving the solution and mortality was lower in animals receiving hemoglobin
Page 70 of 119
solution. The authors concluded that hemoglobin solution administration could be benefi-
cial in severely injured patients with traumatic brain injury.
Shock and resuscitation-related coagulopathy
In SRGS, Vol. 35, No. 6, the general principles and effective application of “damage con-
trol” techniques for management of critically injured patients were discussed. The devel-
opment of these techniques was prompted by the recognition of the lethal triad of acidosis,
hypothermia, and coagulopathy leading to death in patients with prolonged bleeding and
hypovolemic shock. The critically important contribution of “damage control” leading to
improved survival in this patient group was to document methods for rapid control of hem-
orrhage followed by restoration of tissue perfusion, body temperature, and coagulation sys-
tem function in the controlled environment of the intensive care unit.
Experience with combat casualties in Iraq and Afghanistan has shown that persistent
microvascular hemorrhage from coagulopathy is a component of severe injury that may
have its onset early in the management of the trauma victim. These observations have led
to the development of “massive transfusion” protocols that stress the replacement of clot-
ting factors along with blood. Combat resuscitation protocols using fresh whole blood are
an outgrowth of the understanding that clotting factor replacement along with restoration
of volume and oxygen carrying capacity ideally should occur simultaneously with hemor-
rhage control in the severely injured patient. In this section of the overview, the coagulopa-
thy of severe injury will be discussed.
Page 71 of 119
Mechanisms of coagulopathy
The first article reviewed in this section of the overview is by Hess and coauthors 67 from
the Journal of Trauma, 2008. The article is furnished as a full-text reprint accompanying this
issue of SRGS. The article is entitled “The coagulopathy of trauma: a review of mecha-
nisms.” The authors begin by emphasizing the central role of hemorrhage as a determinant
of outcomes in injured patients. They note that 40% of all trauma deaths are the result of
hemorrhage. Lethal bleeding can occur in the immediate postinjury interval leading to
death before the injury victim reaches the hospital. The focus of this report, however, is on
that group of patients at risk for lethal bleeding because of persistent microvascular hemor-
rhage that arises when control of bleeding is delayed or incomplete, leading to massive in-
fusions of red blood cells and electrolyte solutions. The authors stress that the adverse af-
fects of activation of the coagulation mechanism are not confined to persistent
Activation of coagulation pathways leads to elaboration of pro-inflammatory mediators,
exposing the patient to the development of systemic inflammatory response syndrome and
multiple system organ failure. Coagulopathy increases the risk of additional intracranial
bleeding and additional secondary tissue injury in patients who sustain traumatic brain inju-
ry. They stress that traumatic brain injury can, by itself, lead to coagulopathy due to release
of brain-specific thromboplastins into the circulation. Acute coagulopathy evidenced by an
abnormal international normalized ratio was observed in 25% of severely injured patients
according to data cited by these authors; in this patient group, there was a four-fold in-
crease in mortality risk.
Page 72 of 119
The authors discuss the combined effects of bleeding and tissue injury on activation of
coagulation pathways. They point out that tissue injury leads to the exposure of type II col-
lagen and tissue factor that is the mechanism for initiation of platelet binding and activation
of the coagulation protein cascade. Massive soft tissue injury, in the absence of bleeding
and shock, rarely, if ever, results in demonstrable coagulopathy. They note that coagulopa-
thy has been thought to occur in patients with multiple long bone fractures by bleeding be-
cause of coagulopathy in patients with long bone fractures essentially never occurs without
accompanying hemorrhage and hypovolemia. This consistent clinical observation under-
scores the importance of the combined effects of tissue trauma and hemorrhagic shock.
The shock-tissue injury combination leads to a hyperfibrinolytic state because of the re-
lease of tissue plasminogen activator from injured tissue and from ischemic endothelial
cells. Inhibitors of plasminogen activator are also down-regulated with hemorrhage and
trauma. As clotting factor levels decrease, abnormal polymerization of fibrin monomers re-
sults in a polymer vulnerable to cleavage by plasmin.
The authors then review the effects of shock, hemodilution, hypothermia, and acidemia
on coagulopathy. End organ hypoperfusion, as evidenced by a base deficit > 6 is associated
with coagulopathy (defined as an abnormal international normalized ratio and prolonged
partial thromboplastin time) in 25% of injured patients. Platelet counts are preserved in this
setting. A consistent observation is that shock is the driver of coagulopathy and this concept
is bolstered by a large number of reports confirming coagulopathy in the presence of shock
with minimal tissue injury (vascular injury from penetrating mechanism) and the absence of
Page 73 of 119
coagulopathy despite extensive tissue injury when hemorrhagic shock is not present. The
mechanism of shock-induced coagulopathy remains unclear. It is known that acidemia will
interfere with coagulation but coagulopathy is present with lesser degrees of acidemia than
necessary to produce this interference. Shock produces changes in coagulation that lead to
a relatively anticoagulant and hyperfibrinolytic state. The authors cite one study that has in-
criminated increased thrombomodulin activity leading to combined activation of protein C
Traditionally, trauma-related coagulopathy has been interpreted to be a “dilutional co-
agulopathy” produced by consumption of coagulation factors at the site of injury and dilu-
tion of the remaining factors due to shifts of fluid from interstitial and intracellular spaces
and infusion of large volumes of blood and fluid deficient in coagulation factors. Electrolyte
fluids and some colloid fluids may alter coagulation according to in vitro studies. Apprecia-
tion of the effects of dilution has provided additional support for efforts to restore coagula-
tion proteins early in resuscitation. Hypothermia is commonly observed in injured patients,
especially those demonstrating abnormal microvascular bleeding.
Studies68 from the early 1990s have documented the effects of hypothermia on platelet
function and coagulation proteins. A more recent analysis by Wolberg and coauthors 69 from
the Journal of Trauma, 2004, analyzed coagulation enzyme function and platelet function at
37°C and 33°C. They noted no change in platelet activation and only mild changes in coagu-
lation enzyme function with reduced temperature. Platelet aggregation and adhesion were
significantly impaired at 33°C and below this level; both platelet function and coagulation
Page 74 of 119
enzyme activity were impaired. These observations lead to the conclusion that hypothermia
plays a role in producing trauma-related coagulopathy when temperature is reduced below
33°C, but does not function as an independent driver of the coagulopathy at body tempera-
tures observed in most injured patients (34-36°C). Acidemia can alter coagulation functions.
Data cited by Hess and colleagues documents significant impairment of activity of coagula-
tion factor complexes on cell surfaces at pH levels < 7.0. Fibrinogen degradation is enhanced
in acidic environments. Noteworthy is the fact that restoration of pH with buffer does not
ameliorate changes in coagulation indicating that combinations of influences are driving the
development and severity of coagulopathy.
These authors note that the proinflammatory state commonly observed in many trauma
patients who survive the initial injury and resuscitation leads to altered coagulation. During
shock and resuscitation, injured patients are in an anticoagulant-hyperfibrinolytic state that
evolves into a hypercoagulable state following resuscitation. The characteristics of this
hypercoagulable state are similar to those observed in patients with sepsis and septic shock
with depletion of protein C.
Hess and associates conclude by noting that trauma-related coagulopathy differs from
disseminated intravascular coagulation seen in some forms of sepsis; the coagulopathy has
an early onset with abnormalities of prothrombin time and partial thromboplastin time and
preserved platelet numbers and normal fibrinogen levels.
The view that trauma-related coagulopathy is distinct from classic disseminated intra-
vascular coagulation is not held by all surgeons. A contrasting view is offered by Gando 70 in
Page 75 of 119
an editorial comment appearing in the Journal of Trauma, 2009. Gando objects to the cate-
gorization of the acute coagulopathy of trauma and shock as a unique form of coagulation
abnormality and emphasizes the classification of disseminated intravascular coagulation in-
to two phenotypes, fibrinolytic (or hemorrhagic) and antifibrinolytic (or thrombotic). In his
view, these two phenotypes correctly characterize the early hemorrhagic and later
prothrombotic phases of trauma-related coagulopathy. Gando offers data in support of the
presence of the hemorrhagic phenotype of disseminated intravascular coagulation early af-
ter injury. He further notes his agreement that hemodilution and hypothermia play permis-
sive but not primary roles in trauma-related coagulopathy. He concludes by stating that dis-
seminated intravascular coagulation noted in injured patients, septic patients, and patients
with medical diseases is a condition characterized by disordered coagulation that may be
hemorrhagic or prothrombotic depending on the stimulus and the pattern of disordered an-
From the perspective of the editor, the authors of the two interpretations of trauma-
related coagulopathy expressed above appear to be in “violent agreement.” Gando empha-
sizes the value of defining coagulopathy according to the criteria promulgated by the Japa-
nese Association for Acute Medicine and the International Society on Thrombosis and He-
mostasis. He agrees with the mechanisms and clinical descriptions offered by Hess and co-
authors. The main point of contention is not clinical but one of terminology.
Page 76 of 119
Diagnosis of trauma-related coagulopathy
Early trauma-related coagulopathy has typically been identified by documenting abnor-
malities of prothrombin time (expressed as the international normalized ratio) and activat-
ed partial thromboplastin time. This practice has been driven mainly by the availability of
these two assays and the fact that results can generally be obtained during the initial as-
sessment and resuscitation phase of trauma management. The facts that these two tests do
not provide sufficient characterization of the coagulopathy to predict, accurately, the risk of
persistent microvascular bleeding or to guide therapy have been widely recognized. The
search for more accurate measures has proceeded and thromboelastography has been one
test investigated. The use of this measure as a means of characterizing, accurately, the ex-
tent of coagulopathy is the subject of an experimental analysis reported by Martini and co-
authors71 in the Journal of Trauma, 2008. The authors created an experimental situation de-
signed to evaluate the effects of shock, hypothermia, and resuscitation on clotting tests cur-
rently used to detect the presence of trauma-related coagulopathy. They then compared
the commonly used clotting tests to thromboelastography, a test that assesses, separately,
initial clot formation time, rapidity of clot formation, time to maximum clotting, and clot
strength. In this set of experiments using porcine subjects, shock was produced by bleeding
the animal a volume equivalent to 35% of estimated blood volume. Mean arterial pressure
was maintained above 40 mm Hg. Resuscitation was carried out with a volume of lactated
Ringer’s solution equivalent to three times the hemorrhage volume. Hypothermia, applied
with or without shock and resuscitation, was begun at the time of initiation of bleeding and
continued until the core temperature was maintained at 32°C.
Page 77 of 119
The authors observed little change in activated partial thromboplastin time or
prothrombin time in any group except the shock-resuscitation-hypothermia group. Platelet
counts were reduced with shock-resuscitation, hypothermia, and combined groups, and fi-
brinogen level was diminished significantly in the combined shock-resuscitation-
hypothermia group. Thromboelastography analyses disclosed significant reductions in clot
strength in the shock-resuscitation and combined groups. Hypothermia caused slowing of
clot formation time, time to maximum clotting, and clotting rapidity. Shock-resuscitation
and combined shock-resuscitation and hypothermia displayed similar findings. The authors
concluded that thromboelastography was a more sensitive measure of disorders of coagula-
tion that accompany hypothermia, shock, and resuscitation than conventionally used tests.
This group of investigators conducted a clinical study to ascertain whether
thromboelastography could detect a hypercoagulable state in injured patients and to de-
termine whether the test could identify patients who developed venous thromboembolism.
They conducted serial assessments of the coagulation mechanism in a group of 33 injured
patients and 25 burn patients. All studies began within 24 hours of admission and continued
for seven days. The control group consisted of 20 healthy volunteers. Thromboelastography
detected a hypercoagulable state that was confirmed by documenting depressed levels of
protein C and elevations of fibrinogen. Activated partial thromboplastin times and
prothrombin times were prolonged in all patients, suggesting an anticoagulant state. No
test detected the three patients who developed pulmonary emboli. The authors concluded
that thromboelastography was able to detect, reliably, a hypercoagulable state in this group
Page 78 of 119
of injured patients. The data from this study however, do not permit evaluation of
thromboelastography as a means of predicting thromboembolic complications.
It is apparent that thromboelastography is a sensitive test for coagulation system dys-
function. The test can be a “point of care” assessment carried out at the bedside or in the
operating room. Data confirming the utility of the test as an instantaneous guide for thera-
peutic decisions has not been forthcoming and this fact probably contributes to lack of
widespread adoption of this assessment in trauma centers. The subjects of diagnosis and
management of trauma-related coagulopathy are dealt with in more detail in the next sec-
Management of coagulopathy
The first article reviewed in this section is by Hoyt and coauthors72 in the Journal of
Trauma, 2008. This article reports results of an international survey of trauma surgeons.
The survey sought to assess practices pertinent to the management of severely injured pa-
tients defined as an injury severity score ≥ 16. Most respondents practiced in trauma cen-
ters and the majority of the respondents had more than 20 years of clinical experience. The
results indicate significant geographic variability. For example, trauma surgeons were more
likely to manage patients in the American region while anesthesiologists and other special-
ists were active in the management of injured patients in the Europe-Mediterranean region
and the Asian-Middle East region. Pharmacologic agents such as aprotinin and
desmopressin were more likely to be used in areas outside of the American region. Most
practitioners used general clinical assessment along with body temperature, activated par-
Page 79 of 119
tial thromboplastin time, and prothrombin time to detect coagulopathy. Most respondents
replaced red blood cells, platelets, and clotting proteins (mostly with fresh frozen plasma).
Calcium, vitamin K, and recombinant activated factor VII were the most common adjunctive
measures used. Fewer than half the respondents followed a prescribed massive transfusion
protocol. Many respondents cited barriers to protocol use including delays in obtaining la-
boratory values and blood products. Although this study is limited because less than one-
third of the potential respondents actually responded, the available results emphasize the
variability and barriers to effective management of coagulopathy in injured patients.
The next series of articles reviewed present evidence of the value of early identification
of patients at risk for coagulopathy. The principle diagnostic criterion for such patients has
been the need for transfusion of more than 10 units of packed red blood cells within the
first 12-24 hours after injury. Once suspicion of trauma-related coagulopathy is present, ag-
gressive efforts to stop bleeding along with institution of replacement of blood, clotting
proteins (fresh frozen plasma), and platelets are indicated. Since ongoing bleeding and
shock are the main conditions leading to coagulopathy, successful identification of bleeding
points and control of these are critical. The exact configurations of protocols for administra-
tion of fresh frozen plasma and platelets continue to evolve.
The first article dealing with the appropriate ratio of blood to plasma and platelets is by
Holcomb and coauthors73 in Annals of Surgery, 2008. This article is supplied as a full-text re-
print accompanying this issue of SRGS. The authors begin by stressing the lethality of early
massive bleeding. They note that death from bleeding usually occurs within the first 2-6
Page 80 of 119
hours after injury and that most of these deaths are contributed to, at least partially, the
development of coagulopathy. They note that recommended civilian injury plasma and
platelet administration protocols focus on a ratio of one unit of plasma to three units of red
blood cells with platelets administered when bleeding persists and/or when platelet counts
are less than 100,000/cm3. Military experience during the conflicts in Iraq and Afghanistan
has emphasized the value of early administration of red blood cells, plasma, and platelets in
ratios of 1:1:1. Holcomb and associates note that the association of resuscitation with large
volumes of electrolyte solution is associated with the diagnosis of abdominal compartment
syndrome and they suggest that reductions of electrolyte solution volumes are desirable,
although clinical evidence to confirm this assertion is scarce. This report involved data from
more than 1500 patients admitted to several Level I trauma centers. They were able to doc-
ument an improvement in overall survival from 41% to 71% with increasing ratios of plasma
and platelets to red blood cells.
Additional data supporting the use of higher plasma to red blood cell ratios comes from
a report by Sperry and coauthors74 in the Journal of Trauma, 2008. These authors analyzed
a prospective cohort of injured patients who were admitted with clinical signs of shock.
Four hundred fifteen of these patients received eight or more units of red blood cells in the
first 24 hours after injury and these patients were the focus of this analysis. These authors
documented a decreased early mortality but an increased frequency of acute respiratory
distress syndrome in patients receiving plasma to red blood cell ratio of ≥ 1.5 to 1. Although
there was a higher frequency of acute respiratory distress syndrome in the patients who re-
ceived larger volumes of plasma, the risk-adjusted frequencies of nosocomial infection and
Page 81 of 119
multiple organ failure were not increased in this patient group. Additional retrospective
analyses75-76 have also suggested that mortality benefit is observed with increasing use of
plasma in patients who have massive hemorrhage. Although these latter studies are limited
by retrospective design and cannot account for the influence of improved overall recogni-
tion and management of massive bleeding, an association remains between improved
short-term outcomes and increasing use of plasma.
Other reports do not uniformly support a reduction in mortality when aggressive plasma
resuscitation is used. It is useful to examine the data to determine which patients will bene-
fit from increasing the ratio of plasma to red blood cell transfusion. The first article re-
viewed is by Scalea and coauthors77 from the Annals of Surgery, 2008. This article is sup-
plied as a full-text reprint accompanying this issue of SRGS. This report analyzed prospec-
tively gathered data from patients admitted to the intensive care unit of a single trauma
center. The authors report that a massive transfusion protocol was in place during the in-
terval covered by this study. This protocol called for aggressive plasma replacement after
five units of packed red blood cells were administered. More than 800 patients were availa-
ble for analysis and 250 of these received plasma and red blood cells during the first 24
hours after injury. Eighty-one patients fulfilled criteria for “massive transfusion” as de-
scribed in the articles reviewed previously. Using logistic regression analysis, these authors
could not identify a mortality benefit for aggressive plasma resuscitation even in the mas-
sively transfused subgroup. The authors note, in the discussion section of their report, that
their study includes brain-injured patients. This patient group is excluded in most of the re-
ports where benefit from increased plasma to red blood cell ratios has been observed.
Page 82 of 119
The authors note that their overall mortality was lower than the mortalities reported
from combat injury data and from other reports supporting aggressive plasma resuscitation.
Finally, the authors note that defining “massive transfusion” in terms of number of units of
red blood cells infused over the first 24 hours after injury actually selects a heterogeneous
group of patients. Some of these patients present with rapid blood loss (10 or more units
within 6-8 hours) while others have slower, more persistent blood loss (10 units spread over
the entire 24-hour interval). The risk for coagulopathy is likely different in these two groups
of patients. The patients who will likely benefit from aggressive replacement of fresh frozen
plasma and platelets (or the use of fresh whole blood) are most likely found in the group
with early evidence of rapid blood loss.
A report describing the potential increased risk of morbidity associated with aggressive
plasma resuscitation appeared in the Journal of Trauma, 2009, by Watson and coauthors.78
This report analyzes data gathered prospectively for a separate study of patients sustaining
blunt injury and admitted with a diagnosis of shock. Patients sustaining traumatic brain inju-
ry were excluded from the analysis. More than one thousand patients were included and
nearly two-thirds of the patients received fresh frozen plasma. This report could not identify
a relationship between aggressive plasma resuscitation and mortality except when early
deaths (within 48 hours) were included. When this was done, a modest reduction in mortal-
ity was associated with increasing use of plasma. The authors confirmed an increased risk
for acute respiratory distress syndrome and multiple system organ failure with increasing
use of fresh frozen plasma. The authors note that a clear determination of an increased risk
of dying from acute respiratory distress syndrome and multiple organ failure could not be
Page 83 of 119
identified in their analysis. This report was presented during the plenary session of the an-
nual meeting of the American Association for the Surgery of Trauma. In the discussion that
followed the presentation (included with the article), several discussants emphasize the
possibility that there was increased risk for morbidity because a higher proportion of pa-
tients survived the early bleeding event.
From the available data, reviewed above, it is logical to accept that a group of patients
exists who will benefit from aggressive measures to control hemorrhage combined with a
resuscitation approach that emphasizes increased ratios of plasma to red blood cells. This
group will most likely have severe torso trauma and major vascular injury. Exsanguination
within the first 6-8 hours after injury will be the predominant risk. An extension of this
thought process includes consideration for implementation of institutional protocols stress-
ing rapid identification of patients at risk, aggressive efforts to control bleeding, and imme-
diate availability of red blood cells and plasma. This topic is reviewed in the next two articles
discussed. The first is by Cotton and coauthors79 from the Journal of Trauma, 2009. The arti-
cle is entitled “Predefined massive transfusion protocols are associated with a reduction in
organ failure and postinjury complications.” The protocol evaluated in this report included a
telephone alert to the blood bank from the attending trauma surgeon immediately on
recognition of the potential for massive transfusion. A blood sample for type and screen of
blood products was delivered to the blood bank simultaneously with the alert telephone
call. The blood bank immediately prepared and delivered 10 units of un-crossmatched red
blood cells, four units of plasma, and two units of platelets and prepared to maintain sup-
plies of these products in the same ratio. One hundred twenty-five patients treated under
Page 84 of 119
the protocol were compared with matched historical control patients. The authors noted a
significant reduction in the frequency of multiple organ failure and abdominal compartment
syndrome compared to historical controls. Even with massive transfusion protocols in place,
obtaining completely crossmatched blood and thawed plasma may require an hour. Many
trauma centers make use of un-crossmatched type O blood for early red blood cell resusci-
tation. Type-specific, un-crossmatched blood is not as desirable as type O blood because of
the risk of clerical error during a hurried typing process. Type O positive blood can be used
for male injured patients; type O negative blood is used for women during the childbearing
A report supporting the safety of un-crossmatched blood by Inaba and coauthors80 ap-
peared in the Journal of Trauma, 2008. These authors reviewed transfusion usage in more
than 25,000 injured patients admitted to a single trauma center over a six-year interval. The
authors noted that use of un-crossmatched blood was closely associated with need for mas-
sive transfusion. They documented, as have others, that massive transfusion is required
most often in young male patients with penetrating torso trauma. They documented, in this
analysis, that the use of un-crossmatched blood independently predicted mortality. They
stress the importance of early recognition of the patient requiring massive transfusion and
they suggest that the order for un-crossmatched blood may be a valuable clue that could be
used as a “trigger” for implementation of an institution’s massive transfusion protocol.
Although it is tempting to attribute improved results to implementation of a single sys-
tem-wide change, there is always a risk that other factors are influencing outcomes. For ex-
Page 85 of 119
ample, a report from by Riskin and coauthors81 in the Journal of the American College of
Surgeons, 2009, examined the influence of implementation of a massive transfusion proto-
col on outcomes. This study reports data from 77 patients treated over a four-year period.
Forty patients were treated before protocol implementation and 37 after implementation.
The authors noted a significant reduction in mortality in the group treated after protocol
implementation. They also documented significantly decreased intervals from admission to
blood and plasma availability. Interestingly, there was no difference in ratios of plasma to
red blood cells administered to these selected, massively bleeding patients in the first 24
hours of treatment in both intervals; ratios were 1:1.8 in both intervals. The authors con-
cluded that factors separate from, or in addition to, the implementation of the massive
transfusion protocol were responsible for the observed reductions in mortality.
The data reviewed documents that a significant proportion of injured patients admitted
to trauma centers (up to 25% of admissions) fulfill diagnostic criteria of hemorrhag-
ic/traumatic shock. Of these, a small proportion will bleed massively and these patients will
probably benefit from resuscitation with increased ratios of plasma to red blood cells. Red
blood cells and fresh frozen plasma are the main agents used in the early care of these pa-
tients. Platelets and cryoprecipitate are used more selectively, based on the time interval
required for control of bleeding to occur, platelet counts, and fibrinogen levels. Blood and
blood products are in short supply and are expensive. The intent of massive transfusion pro-
tocols is to identify patients who would benefit from resuscitation with increased plasma to
red blood cell ratios early and to make the necessary products available rapidly so that im-
Page 86 of 119
proved survival can occur. It must be emphasized that this approach is likely to benefit rela-
tively few patients but it will probably be life-saving in this group. Coupling the availability
of blood and blood products with aggressive efforts to identify bleeding sites and control
bleeding is critical. It is clear from the articles reviewed that essentially all of the patients
who prompt activation of massive transfusion protocols are in transit to the operating room
for hemorrhage control. These patients often have sustained penetrating torso trauma
and/or major vascular injury. The absence of clear benefit of plasma-rich resuscitation in
unselected groups of patients injured primarily by blunt mechanisms emphasizes that mas-
sive transfusion protocols cannot be applied to the majority of patients admitted to trauma
centers. The cost of improved survival in massively transfused patients is an increase in the
risk of acute respiratory distress syndrome. The available data does not examine healthcare
resource consumption associated with massive transfusion. It is legitimate to hope that
massive transfusion protocols would save resources by identifying patients who would ben-
efit and applying resources to this patient group specifically.
Use of specific hemostatic agents to control microvascular bleeding
Although a number of agents are available to assist in hemorrhage control, few of these
are effective in the management of life-threatening diffuse microvascular bleeding. Agents
such as fibrin glue and recombinant thrombin saturated gelatin sponges are helpful with
small areas of bleeding. Desmopressin and aprotinin are limited; these agents are most suc-
cessful when given before the bleeding event. Recent reports have emphasized the poten-
tial benefit of recombinant factor VIIa as an adjunctive measure for control of persistent,
diffuse microvascular bleeding.
Page 87 of 119
The first article reviewed is from Dutton and coauthors82 in the Journal of Trauma, 2004.
The authors begin by pointing out that this agent was originally developed for use in the
management of hemophilia patients who had developed neutralizing antibodies to factor
VIII. Recombinant factor VIIa acts by binding tissue factor at the sight of vascular disruption.
Coagulation is initiated and a “thrombin burst” is produced on the surface of platelets. Dut-
ton and colleagues note that prospective trials have documented the usefulness of this
agent in reducing blood loss during major elective surgery and in the management of coag-
ulopathy resulting from warfarin. They go on to note that no standard dose exists for use of
this agent in injured patients. They chose to use a dose of 100 mcg/kg in injured patients
with bleeding; a dose of 50 mcg/kg was chosen for patients who were coagulopathic sec-
ondary to warfarin use. Repeat dosing was used as needed with up to three total doses ap-
plied in their patient group. The data regarding efficacy are reported separately for patients
with acute traumatic injury, congenital or pharmacologic coagulopathy, and traumatic brain
injury. Patients with bleeding because of acute injuries were more likely to be treated with-
in the first 24 hours of injury. The authors noted reductions in bleeding in most instances.
Thirteen of the 46 patients with acute traumatic bleeding died within the first 24 hours after
admission and all of these received relatively large doses of recombinant factor VIIa, raising
the possibility that these patients represented failures of treatment. Overall, the authors
deemed the treatment successful in assisting to achieve hemostasis in 75% of patients
treated. Most of the treatment failures were deemed unsalvageable because of
nonsurvivable injuries or severe brain damage.
Page 88 of 119
Dutton and coauthors note the main advantage of recombinant factor VIIa is the imme-
diate nature of the drug effect. The authors stress gaps in our knowledge regarding safety.
For example, it is not clear whether the agent will accelerate microvascular thrombosis in
injured tissue such as the brain. Whether use of this agent enhances the hypercoagulable
state seen in patients surviving shock is also not known. Additional data from case series of
combat injuries have supported the use of early recombinant factor VIIa in patients with dif-
fuse microvascular bleeding from injury and shock.83-84 Patients identified as “massive trans-
fusion” patients were treated with doses of recombinant factor VIIa equivalent to120
mcg/kg with the dose being repeated, usually up to three times, as needed. These series
have noted reductions in overall mortality with no notable increase in episodes of venous
thromboembolism. Readers should note, however, that a causal relationship between use
of recombinant factor VIIa and improved survival cannot be established based on data pro-
vided. Similarly, with regard to thromboembolic complications, the approach to thrombo-
embolism detection and protocols for management of thromboembolism prophylaxis were
not provided in these articles. Despite these limitations, the evidence suggests that recom-
binant factor VIIa is useful as an adjunctive means of hemorrhage control and that the drug
has a satisfactory safety profile.
Data about patterns of use of recombinant factor VIIa among trauma surgeons are pre-
sented in an article by Horton and coauthors85 in The American Surgeon, 2008. These au-
thors surveyed surgeons, trauma program managers, and pharmacists who serve in Level I
or Level II trauma centers. The survey had a 36% response rate. The responses disclosed
that recombinant factor VIIa was used by nearly 90% of Level I trauma centers and just over
Page 89 of 119
half of the Level II centers. There was an association between patient volume and usage
with increasing usage in high volume centers. Wide variability of dosing was noted. Few
centers used the 200 mcg/kg dose that was reported from the only prospective evaluation
of the drug in trauma patients reported by Boffard and coauthors.86 Most centers used dos-
es at or near the 90 mcg/kg dose recommended for hemophiliac patients.
Thrombotic complications are associated with the use of recombinant factor VIIa. This
topic is addressed in an article by Rhys Thomas and coauthors87 from the Journal of Trauma,
2007. These authors reviewed data from 285 patients treated with this agent during a five-
year interval. The authors begin by reviewing data on thrombotic complications associated
with the use of recombinant factor VIIa in patients without traumatic bleeding. For exam-
ple, in a reported trial consisting of patients with acute hemorrhagic stroke, 7% of the treat-
ed patients had documented thrombotic complications compared with a frequency of 2% in
patients receiving placebo. In reported trials and retrospective series dealing with injured
patients, no increase in thrombotic complications has been observed. These authors ob-
served an overall incidence of thrombotic complications in 9.4% of the 285 patients ana-
lyzed. The most serious complications recorded included eight instances of mesenteric in-
farction, three instances of myocardial infarction, and five instances of cerebral infarction.
The authors noted that the severity of injury in this patient group prevented definite attrib-
ution of the thrombotic complications to the use of recombinant factor VIIa but the associa-
tion of these events to use of the drug suggests that additional surveillance to document
the actual risk is warranted. When planning for use of recombinant factor VIIa in a patient
Page 90 of 119
with persistent microvascular bleeding, several clinical considerations are appropriate. The-
se are listed in the accompanying table (see Table 1).
The use of recombinant factor VIIa as a means of assisting hemostasis in major elective
surgical procedures is the topic of a systematic review by Ranucci and coauthors 88 in Ar-
chives of Surgery, 2008. The authors report data gleaned from seven randomized controlled
trials. A consistent finding in all studies was a reduction in the number of allogeneic red cell
transfusions in groups where recombinant factor VIIa was used as an adjunct to achieve
hemostasis. The authors report apparent variability of effectiveness depending on the dose
of the agent used in reported trials. They stress the need for additional research so that the
appropriate dose for varying indications can be determined; cost of the drug is an important
factor. Where cost/benefit analyses have been done, cost effectiveness is obtained in pa-
tients requiring massive transfusion only (more than 10 units of red blood cells in 24 hours).
In an editorial comment by Dr. Margaret Knudson accompanying the article, data is cit-
ed supporting the use of recombinant factor VIIa in situations where major blood loss can
be anticipated. She notes that maximum benefit is obtained when the drug is used relative-
ly early and is not relied on as a last resort. Another important point emphasized in the edi-
torial is the need for additional assessments of the risk of thrombotic complications. She
notes that most reports have documented thrombotic events in 2%-4% of patients. This
number is, however, smaller than the frequency reported by Rhys Thomas and coauthors 87
Page 91 of 119
Critical care of the patient with sepsis and septic shock
In the earlier discussion of the pathophysiology of sepsis and septic shock, emphasis was
placed upon the diversity of physiologic responses to inflammation and infection. The septic
state is usually characterized by generalized vasodilation and a hyperdynamic circulation.
Arterial pressure may be within the low normal range and cardiac index may be increased.
Despite these observations, signs of end organ hypoperfusion manifest by elevated blood
lactate levels may be present. In patients with severe sepsis, evidence of end organ
hypoperfusion is clear-cut. Septic shock is present when patients exhibit hypotension re-
sistant to intravascular and extracellular volume expansion. Faced with these combinations
of clinical manifestations, surgeons caring for septic patients need to be familiar with avail-
able means of assessing circulatory status (see the earlier section on monitoring patients in
shock) and to be familiar with the use of volume expanders, vasopressors, inotropic agents,
and adjunctive drugs such as corticosteroids and recombinant activated protein C. These
topics are discussed next.
Goal-directed early comprehensive management of suspected sepsis and/or septic shock
Because sepsis and septic shock are associated with infection-induced inflammation, flu-
id exudation frequently occurs into the sites of inflammation. In postoperative patients who
become septic, recent or ongoing blood loss may also contribute to hypovolemia. Clear evi-
dence of myocardial dysfunction will not become manifest until volume status is optimized.
Fluid therapy that is too aggressive contributes to sequestration of edema fluid in inflamed
tissues. In the lung, edema may contribute to alveolar instability, ventilation/perfusion
mismatching, and hypoxemia. For these reasons, early, goal-directed, fluid therapy of sepsis
and septic shock has become a critical component of all care protocols.
Page 92 of 119
Recognition of the importance of shortening the duration between injury and definitive
care has stimulated the development of prehospital care systems and trauma centers for
the management of acutely injured patients. Similarly, observations of septic patients dis-
close increasing mortality and morbidity with increases in the interval between recognition
of the septic state and implementation of treatment protocols. A classic article dealing with
this topic by Rivers and coauthors40 in the New England Journal of Medicine, 2001, empha-
sized that mortality for sepsis and septic shock has remained stable despite advances in the
effectiveness of therapeutic agents. They note that most reports evaluating these interven-
tions have enrolled patients up to 72 hours after symptom onset. The lack of progress sug-
gests that delay in instituting therapy may contribute to suboptimal outcomes.
The treatment protocol described in this article required placement of a central venous
catheter capable of monitoring central venous oxygen saturation; an arterial catheter was
also inserted in each patient. Specimens were obtained for culture and broad-spectrum an-
tibiotic therapy was begun. After these interventions, the authors randomly assigned pa-
tients to receive conventional intensive care unit admission and treatment and compared
this group with patients assigned to receive protocol-directed treatment for a six-hour in-
terval in the emergency department, followed by intensive care unit admission. Patients re-
ceived electrolyte fluid boluses in volumes of 500 mL/30 minutes until central venous pres-
sure was between 8 and 12 cm H20. If, at this point, arterial pressure was ≤ 65 mmHg, vaso-
pressors were given to maintain mean arterial pressure at or above 65 mmHg. If mean arte-
rial pressure was above 90 mmHg, vasodilator therapy was begun to maintain arterial pres-
sure at or near 90 mmHg. After optimization of vascular pressures, central venous oxygen
Page 93 of 119
saturation was assessed. If this variable was less than 70%, red blood cell transfusion was
given to raise the hematocrit to 30%. If this intervention did not lead to central venous oxy-
gen saturation above 70%, inotropic therapy was begun with dobutamine at 2.5
mcg/kg/min and the dose was increased in 2.5 mcg increments until central venous oxygen
saturation was above 70% or a maximum dose of 20 mcg/kg/min had been reached.
Dobutamine dosage was reduced if heart rate exceeded 120 bpm or if mean arterial pres-
sure was less than 65 mmHg. If optimum oxygen saturation could not be obtained, oxygen
consumption was reduced using adjuvant ventilator therapy and sedation. The authors not-
ed that a similar number of patients were excluded from each group. A few patients in each
group were excluded because of the need for urgent operation. There was no difference in
clinical characteristics of excluded patients when the patients excluded from each group
were compared. The authors documented a significantly reduced mortality at 30, 60, and 90
days in patients who had early optimization of end organ perfusion. The authors note that
beginning assessment of septic patients in the emergency department identified patients
with lower central venous oxygen saturations than is typical for septic patients in the inten-
sive care unit. They speculate that this early identification of insufficient oxygen delivery
may have contributed to the improvement in mortality that was observed.
Despite ongoing debate on use of electrolyte solutions or colloid solutions for early he-
modynamic resuscitation of septic patients, available data confirm that the two types of flu-
id are equivalent in terms of efficacy and safety. Recommendations for fluid therapy are
found in a review from Hollenberg and coauthors89 in Critical Care Medicine, 2004. These
authors note that significant intravascular and interstitial volume deficits are common in
Page 94 of 119
septic patients and, because of this, volume expansion is an important first step in resusci-
tation of this patient group. Hollenberg and associates note that fluid therapy should be
guided by clinical assessment of filling pressures and by monitoring of global indices of tis-
sue oxygen supply such as central venous or mixed venous oxygen saturation (see discus-
sion in previous sections of the overview). They stress that venous oxygen saturation may
be elevated in septic patients and that this variable needs to be considered in the context of
the total hemodynamic picture presented by each individual patient. The report also em-
phasizes that a liter of electrolyte solution (normal saline or Ringer’s lactate) will provide
approximately 200 mL of persistent intravascular volume expansion. Expected infusion vol-
umes within the first 24 hours of management of sepsis may approach 6-10 liters of electro-
lyte solution or 2-4 liters of colloid. Albumin has been the most common colloid evaluated
for use in the critically ill septic patient.
A report comparing albumin resuscitation to saline resuscitation for critically ill patients
is found in a 2004 report by Finfer and coauthors90 from The New England Journal of Medi-
cine. Nearly 7000 patients were included in this randomized prospective study; the vast ma-
jority of both treatment groups consisted of acutely injured patients or patients with sepsis.
Although the patients treated with saline had net positive fluid balances significantly higher
than observed in the patients receiving albumin during the first three days of treatment,
complication rates were equivalent in the two groups. Adequacy of resuscitation was not
different in the two groups. When the subgroup of injured patients was analyzed separate-
ly, there was a suggestion that saline resuscitation was superior. The authors concluded
Page 95 of 119
that hemodynamic resuscitation in the intensive care unit could be safely conducted with
either saline or albumin.
Additional data confirming the importance of early recognition and treatment of sepsis
are found in a report by Kumar and coauthors91 in Critical Care Medicine, 2006. In this anal-
ysis, medical records of nearly 3000 medical and surgical intensive care patients were re-
viewed. For patients with the suspected diagnosis of sepsis or septic shock, groups receiving
antimicrobial therapy appropriate for the infecting organism within six hours of the onset of
hypotension were compared with a group of patients who received antimicrobial therapy
only after hypotension recurred. A significant survival advantage was noted in patients who
received early therapy and there was a linear association between the interval of treatment
delay and mortality. The highest survival rate (79.9%) was observed in patients who had ap-
propriate antimicrobial therapy begun within one hour of symptom recognition. The au-
thors stress, in the discussion section of the report, that data from the report authored by
Rivers and coauthors40 (discussed above) have demonstrated the importance of early he-
modynamic resuscitation as a means of achieving improved survival in septic patients.
The data reported in this article support the addition of appropriate antimicrobial ther-
apy to treatment protocols designed for management of the septic patient. In an editorial
accompanying this article, Dr. Peter Gross emphasizes the strengths of this analysis, which
are its large sample size, and the consistent results obtained in multiple intensive care units
serving a diverse assortment of patients. The strength of the evidence from this study and
that of Rivers supports the addition of early hemodynamic resuscitation and early antimi-
Page 96 of 119
crobial therapy as performance measures for the management of patients with suspected
sepsis or septic shock. Gross notes that these performance measures are being incorpo-
rated into sepsis management protocols (sepsis bundles) that provide evidence-based guid-
ance for the management of these patients.
Early support for a comprehensive management protocol for sepsis and septic shock is
found in a description of the guidelines for management of sepsis promulgated by the Sur-
viving Sepsis Campaign. These guidelines are found in an article by Dellinger and
coauthors92 in Critical Care Medicine, 2004. The guidelines were updated with strengthened
evidence in 2008.92 In these two articles, a group of interventions including early fluid ther-
apy, antimicrobial therapy, source control, glycemic control, and selective use of cortico-
steroids are included in a “bundled” approach to the septic patient. Each measure is sup-
ported by evidence with sufficient strength to justify inclusion of the intervention as a per-
formance measure (supported by evidence stronger than expert opinion). The recommend-
ed measures are listed in the accompanying table (see Table 2).
The effectiveness of “bundling” multiple interventions into a comprehensive manage-
ment protocol for sepsis is examined in an article by Kortgen and coauthors 93 in Critical Care
Medicine, 2006. The report describes a “before and after” study of patients in a single in-
tensive care unit. Thirty patients fulfilling diagnostic criteria for septic shock were treated
with a protocol that stressed early goal-directed resuscitation, appropriate antimicrobial
therapy, stress dose corticosteroids, glycemic control, and selective use of recombinant ac-
tivated protein C. Results in these 30 patients were compared with a similar-sized cohort of
Page 97 of 119
historical control patients. The authors noted that goals of treatment were achieved earlier
in the patients treated with the sepsis protocol. The use of insulin, hydrocortisone, and re-
combinant activated protein C all increased following implementation of the protocol. The
authors noted a reduction in mortality from 53% to 27% when the two groups of patients
The authors conclude that implementation of an evidence-based protocol for sepsis
care is feasible and that uniformity of utilization of the recommended interventions was
achieved in their unit. In an editorial by Rivers that accompanied this report, emphasis is
placed on the wide variability noted in the United States when rates of use of evidence-
based interventions for sepsis are assessed. Rivers asserts that implementation of evidence-
based protocols is appropriate and necessary if improvements in sepsis outcomes are to be
Additional evidence that protocol implementation is feasible as a means of improving
the early care of patients with sepsis and septic shock is found in another article in the same
issue of Critical Care Medicine, 2006.94 In this report, the authors describe some of the edu-
cational and process change measures that eased the protocol implementation process. The
objectives of the protocol were to enhance early recognition of septic patients, to institute
early evidence-based therapy in the emergency department, and to improve outcomes for
septic patients. The authors used the regularly scheduled medical and nursing staffs of a
university hospital. They conducted education and provided real-time feedback. Care pro-
cesses were made easier by providing “sepsis carts” containing equipment and medications,
Page 98 of 119
and written versions of the protocol elements. The 116 protocol patients were compared
with historic controls. The results disclose that early usage of resuscitation fluids, antimi-
crobials, corticosteroids, and recombinant activated protein C all increased after implemen-
tation of the protocol. There was a reduction in overall hospital mortality from 29% to 21%
but this reduction was not statistically significant. Readers should note, however, that the
historic sepsis mortality in this institution was lower than that reported in other series. In
fact, it was already near the 20-25% mortality reported in many of the protocol-treated
groups reported in evaluations of early management of sepsis. An additional statistically
significant reduction was, therefore, unlikely.
Use of vasopressors
A significant proportion of septic patients have mild arterial hypotension. By definition,
patients with septic shock are hypotensive. Because of the varying efficiency of
autoregulation in peripheral vascular beds and the wide distribution of tissue oxygen con-
sumptions, choosing an arterial pressure that will uniformly correct disorders of tissue oxy-
gen dynamics is challenging. Most recommendations for arterial pressure targets for septic
patients have selected a mean arterial pressure range of 60-70 mmHg because, for most
vascular beds, autoregulation is lost and tissue blood flow becomes pressure dependent at
mean pressures below 60 mmHg. Although at least one clinical report has noted no im-
provement in tissue oxygen dynamics when a mean arterial pressure target of 85 mmHg
was chosen, other studies have shown increases in urine output with stepwise increases in
mean arterial pressure.
Page 99 of 119
A study by Jhanji and coauthors95 in Critical Care Medicine reported increases in cutane-
ous oxygen tension and microvascular flow with increasing norepinephrine doses in 16 pa-
tients with septic shock. These authors used transcutaneous oxygen measurements and
Doppler laser flowmetry to assess microvascular oxygen availability and blood flow. Sublin-
gual microvascular flow was assessed with sidestream darkfield imaging. Dosing of norepi-
nephrine was adjusted to achieve mean arterial pressures of 60, 70, 80, and 90 mmHg se-
quentially. Global oxygen delivery as well as cutaneous oxygen tension and cutaneous
microvascular flow increased in keeping with each increase in mean arterial pressure. Sub-
lingual microcirculatory status did not change. Outcome assessments were not obtained in
these patients. The authors concluded that additional research is desirable to refine defini-
tions of optimum endpoints for vasopressor therapy in patients with septic shock. These
observations serve to confirm data from articles discussed in previous sections of the over-
view confirming the heterogeneity of oxygen delivery, blood flow, and autoregulation in
various vascular beds. Studies of capnometry44 have noted the similarity between the sub-
lingual microcirculation and the intestinal microcirculation. The failure to correct flow and
oxygen delivery in the sublingual area suggests that sepsis-induced reductions in flow to the
vital visceral vascular bed may remain impaired despite increasing doses of norepinephrine.
When patients remain hypotensive despite correction of intravascular volume, vaso-
pressor therapy is indicated. This topic is discussed in a review by Hollenberg 96 in Chest,
2007. The author notes that several vasoactive agents are available for arterial pressure en-
hancement in septic patients. These include isoproterenol, dopexamine (may not be widely
available in the United States), dobutamine, dopamine, epinephrine, norepinephrine, and
Page 100 of 119
phenylephrine. Isoproterenol, dopamine and dobutamine are predominantly beta-receptor
agonists. Alpha-receptor stimulatory effects are progressively more pronounced for epi-
nephrine, norepinephrine, and phenylephrine. Dopamine has been recommended as the
first line vasoactive agent for the management of sepsis. At low doses (5 mcg/kg/min), do-
pamine stimulates dopaminergic receptors and vasodilation results. This property was the
basis for using dopamine as a means of improving renal blood flow and preventing renal
failure but this approach failed. In the dosage range from 5-10 mcg/kg/min, dopamine is
primarily a beta-receptor stimulant and improvements in myocardial contractility increase
cardiac output because of an increase in stroke volume. A secondary effect of dopamine is
to increase heart rate that also aids in increasing cardiac output. At doses above 10
mcg/kg/min, dopamine is an alpha-receptor agonist causing vasoconstriction and increases
in arterial pressure result from this action.
Hollenberg cautions that there is substantial overlap in the thresholds for these hemo-
dynamic actions. He goes on to note that dopamine regularly increases global oxygen deliv-
ery but consistent improvements in tissue oxygen delivery have not been observed. Dopa-
mine may predispose to cardiac arrhythmias. Because the drug increases heart rate, myo-
cardial oxygen demand may also increase and this effect may be dangerous in patients with
coronary artery disease. It is worth noting that one article, by Sakr and coauthors97, com-
pared 375 patients who were treated for shock with dopamine with 683 patients who were
in shock but not treated with dopamine. Mortality was significantly higher in patients who
received dopamine. In a multivariate analysis, dopamine treatment was independently as-
sociated with mortality.
Page 101 of 119
Hollenberg notes that norepinephrine is primarily an alpha-receptor agonist and im-
proves blood pressure by producing peripheral vasoconstriction. Norepinephrine may pro-
duce modest increases in cardiac output. The dose of norepinephrine required to produce
needed increases in mean arterial pressure may vary because of alpha-receptor down regu-
lation that occurs in sepsis. The author cites a randomized prospective trial comparing
norepinephrine to dopamine in hypotensive septic patients. Norepinephrine at doses of 1.5-
3 mcg/kg/min raised arterial pressure to the target range in more than 90% of patients
compared with 31% satisfactory response to dopamine. Furthermore, 90% of the patients
who failed dopamine therapy responded to norepinephrine. Available evidence cited by
Hollenberg indicates that norepinephrine increases blood pressure without impairing cardi-
ac function and there are data suggesting improved mortality for septic patients treated
with norepinephrine. Because of the evidence cited above, norepinephrine is recommended
as the first choice drug for managing hypotension in septic patients who remain hypoten-
sive following optimization of volume status.
Hollenberg notes that phenylephrine is generally used as a second line drug when
tachyarrhythmias complicate therapy with other vasopressor drugs. Epinephrine raises
blood pressure but the drug increases lactate levels, produces hypoglycemia, and predis-
poses to tachyarrhythmias.
Arginine vasopressin has been suggested as an adjunct to low-dose norepinephrine for
the management of hypotension in septic patients. Hollenberg96 notes that this hormone is
synthesized in the hypothalamus and stored in the pituitary gland. The hormone is released
Page 102 of 119
in response to decreases in intravascular volume and is an arterial vasoconstrictor, acting
via V1 receptors on arterial smooth muscle cells. There is evidence that V1 receptors are
upregulated in sepsis and septic shock and that, with time, there is depletion of vasopressin
stores. These factors suggest that replacement of vasopressin to achieve physiologic plasma
levels would augment the needed vasoconstrictor response in sepsis. Hollenberg cites data
suggesting that a constant infusion of vasopressin (0.03 U/min) improves the blood pres-
sure response to low-dose norepinephrine.
A recent randomized, prospective, double blind trial compared norepinephrine treat-
ment alone with norepinephrine combined with arginine vasopressin infusion at the dosage
level cited above.98 The outcomes of interest were 28 day and 90 day mortality. The authors
did not confirm a mortality benefit in the group treated with norepinephrine plus vasopres-
sin. In the subset of patients with less severe septic shock, there was a trend toward low-
ered mortality in the group treated with vasopressin but the study was not designed to ana-
lyze patients based on shock severity. The authors of the study note that overall mortality
rates for both groups were low compared with prior experience and this may reflect an
overall improvement in outcomes for patients with sepsis treated with modern protocol
A recent report suggesting a synergistic effect of vasopressin and corticosteroid therapy
in patients with septic shock is from Russell and coauthors99 in Critical Care Medicine, 2009.
Nearly 800 patients were diagnosed with septic shock and treated with norepinephrine
and/or vasopressin. The vasopressin dose ranged from 0.01 U/min–0.03 U/min. Hydrocorti-
Page 103 of 119
sone in “stress” doses (150-200 mg/day) was given according to the judgment of the prima-
ry caregiver. The authors noted a significant mortality reduction in patients who received
hydrocortisone and vasopressin compared with patients who received norepinephrine and
hydrocortisone. The authors speculate that interaction between vasopressin and hydrocor-
tisone at the V1b receptor may serve to increase serum cortisol levels and produce a syner-
gistic effect on arterial tone.
Because vasopressin, in the low doses recommended, has few, if any, adverse side ef-
fects, combining norepinephrine, cortisone, and vasopressin in selected patients may be
beneficial. Additional data regarding this association is needed. Review of the use of “stress-
dose” hydrocortisone therapy in septic shock will occur in a later section of the overview.
Use of inotropic drugs
Traditional approaches to the management of septic patients have included the use of
inotropic agents such as dobutamine to assist in support of cardiac output in patients with
persistent low cardiac output despite volume loading and vasopressors. The use of inotropic
agents in patients with sepsis-associated myocardial depression producing decreased cardi-
ac output is a topic reviewed by Leone and Martin100 in Advances in Sepsis, 2006. These au-
thors note that some septic patients have myocardial depression evidenced by low central
venous oxygen saturation and echocardiographic evidence of myocardial dysfunction. Such
patients could possibly benefit from dobutamine therapy. Leone and Martin note that one
study reported increased cardiac output when dobutamine was added to norepinephrine
therapy in septic shock patients with myocardial dysfunction. Leone and Martin caution,
however, that beta-1 stimulation has been shown to lead to cardiac myocyte apoptosis.
Page 104 of 119
Fortunately, the proportion of septic patients with documentable myocardial dysfunction is
Use of “stress-dose” corticosteroids
High-dose corticosteroid therapy (30 mg/kg methylprednisolone) was suggested as a
means of reducing mortality in patients with septic shock in the 1970s. This approach be-
came controversial and a meta-analysis authored by Cronin and coauthors101 in Critical Care
Medicine, 1995, concluded that “current evidence provides no support for the use of corti-
costeroids in septic shock and suggests that their use may be harmful.” In subsequent years,
the discovery that certain proinflammatory states are associated with relative adrenal insuf-
ficiency and/or inflammation induced glucocorticoid receptor resistance led to trials that fo-
cused on diagnosis of relative adrenal insufficiency with therapy targeted toward patients
with adrenal insufficiency.
The first article reviewed in this segment of the overview is from Annane and
coauthors102 in the Journal of the American Medical Association, 2002. The article is entitled
“Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in
patients with septic shock.” The authors open their report by citing data suggesting that rel-
ative adrenal insufficiency can be detected in critically ill patients using a short corticotropin
test. This assessment is performed as follows: a blood sample for measurement of baseline
blood cortisol is obtained and this is followed by a single 250 mcg intravenous dose of
tetracosactin. Additional samples are drawn for blood cortisol levels 30 and 60 minutes af-
ter the injection. An elevation of blood cortisol over baseline of 9 mcg/dL or less is consid-
ered evidence of relative adrenal insufficiency. All patients in this trial fulfilled usual diag-
Page 105 of 119
nostic criteria for septic shock. All underwent testing for adrenal insufficiency, and all were
randomized to receive hydrocortisone (50 mg four times daily for seven days) and oral (or
via feeding tube) fludrocortisone 50 mcg daily. The endpoint of interest was overall mortali-
Mortality in patients diagnosed with relative adrenal insufficiency was 63% in the place-
bo group and 53% in the treatment group. This difference was statistically significant. Mor-
tality reductions were observed for ICU, hospital, and one-year mortality time points. An
additional trial appeared in the New England Journal of Medicine in 2008 by Sprung and co-
authors.103 In this trial, patients were enrolled if they had infection and hypotension requir-
ing volume and vasopressor therapy for at least one hour. Randomization could occur with-
in three days of identification of septic shock. A standard corticotropin test was adminis-
tered to all enrolled patients. This trial could not identify a mortality benefit for stress dose
steroids in patients who fulfilled diagnostic criteria for relative adrenal insufficiency com-
pared with those who did not.
Several important differences exist between this trial and the trial reported by
Annanne102 discussed above. First, the patients entered into the 2008 trial were, as a group,
not as severely septic or as dependent on vasopressors as the patients in the Annane trial.
The observed mortality in the 2008 trial was less than 30% compared with a mortality rate
of 63% in placebo patients in the earlier trial. This would suggest that the groups are not
comparable and that there have been overall improvements in the management of septic
shock that may serve to reduce the impact of cortisone replacement therapy. The results of
Page 106 of 119
these trials have led to recommendations for the diagnosis and management of corticoster-
oid insufficiency in critically ill adult patients.
These recommendations can be found in an article by Marik and coauthors104 in Critical
Care Medicine, 2008. This article is supplied as a full-text reprint accompanying this issue of
SRGS. The recommendations state that diagnosis of relative adrenal insufficiency be estab-
lished using the response to a standard corticotropin test or a random blood cortisol of < 10
mcg/dL. The recommendations further state that cortisone replacement therapy be admin-
istered in a dose of 50 mg four times daily or a continuous infusion of 240 mg/day. This
therapy should be offered to patients with severe, vasopressor dependent septic shock.
Higher dose therapy (methylprednisolone 1 mg/kg/day for 14 days) is appropriate for pa-
tients with early, severe, acute respiratory distress syndrome.
Use of recombinant activated protein C
Sepsis and septic shock are known to trigger an inflammatory and procoagulant cascade
that predisposes to microvascular thrombosis and this may be the etiology of organ failure
leading to death in septic patients. A number of interventions designed to modulate these
cascades have been evaluated in an effort to target the underlying pathophysiology of sep-
sis. Success has been limited in these trials. An agent targeting the coagulation system is re-
combinant activated protein C. This agent has been shown to improve microcirculation in
sepsis according to data cited in a review by Fuller and Dellinger.105 Additional data cited by
these authors indicate that early improvement (within 12-24 hours of beginning treatment)
in tissue oxygen was predictive of survival from severe septic shock.
Page 107 of 119
An early trial reported by Bernard and coauthors106 in the New England Journal of Medi-
cine, 2001, documented a 6% absolute mortality reduction in patients with severe sepsis.
This survival advantage was observed in patients who were severely ill only (APACHE score >
25). The main adverse affects of treatment with recombinant activated protein C is bleed-
ing; fatal intracranial hemorrhage has been reported in patients where the drug has been
used. Use of this agent in patients with increased bleeding risk (abnormal coagulation pa-
rameters, peptic ulcer disease, vascular injury, or solid organ injury) is contraindicated.
Management of cardiogenic shock
Management of patients with suspected cardiogenic shock begins with a rapid assess-
ment to determine adequacy of the airway, oxygenation, and circulation. Patients suspect-
ed to be hypovolemic will benefit from a fluid challenge and/or blood transfusion (similar to
fluid challenges used in the management of hemorrhagic and septic shock, discussed
above). Physical examination assisted by plain chest radiograph can disclose clinical evi-
dence of pulmonary congestion. Electrocardiography and echocardiography are useful, ear-
ly in patient management, to quantify the extent of ischemia and left ventricular dysfunc-
tion. Similarly, pulmonary artery catheter placement can supply sequential data on pulmo-
nary artery pressures and cardiac index. Topalian and coauthors29 supply an algorithm sug-
gesting therapeutic approaches adjusted for the presence of hypotension and/or signs of
end-organ hypoperfusion. They stress that signs of hypoperfusion can occur without pro-
found hypotension. For patients with arterial pressures consistently above 100 mmHg, in-
travenous nitroglycerine in doses of 10-20 mcg/min will produce coronary vasodilation and
improve coronary flow. Dobutamine is used for hypotensive patients without signs of end-
Page 108 of 119
organ hypoperfusion to elevate arterial pressure. Doses of 2-20 mcg/kg/min are indicated
initially. If signs of hypoperfusion are present, the initial dose can be increased to 5-15
mcg/kg/min. Failure to achieve improved blood pressure is an indication for using intrave-
nous norepinephrine. Norepinephrine doses are titrated in the 0.5-30 mcg/min as a contin-
uous intravenous infusion.
Early revascularization using percutaneous coronary angioplasty or coronary artery by-
pass grafting has reduced overall long-term mortality for patients with myocardial infarction
complicated by cardiogenic shock. This result is particularly noticeable in patients < 75 years
old although, as noted by Topalian and associates, mortality rates for older patients are bet-
ter in patients who are revascularized than in patients not revascularized. Thus, older pa-
tients with good functional status are candidates for early revascularization. These authors
note that despite the improved overall outcomes with early revascularization, such inter-
ventions are still underutilized. They cite a report that was published in 2004 indicating that
only half of eligible patients had revascularization.
Overall results of systemic administration of fibrinolytic agents have been inferior to di-
rect revascularization procedures. This lack of efficacy probably results from poor penetra-
tion of the thrombus by fibrinolytic drugs and early reocclusion of coronary arteries that
have been opened by fibrinolysis.
Topalian and colleagues go on to discuss the use of intra-aortic balloon
counterpulsation. They note that these devices, which can now be introduced
percutaneously, can improve coronary flow and reduce left ventricular afterload. Currently,
Page 109 of 119
these devices are largely used as “bridges” to revascularization interventions. Nearly one-
third of the insertions of intra-aortic balloon devices, in a report cited by these authors,
were for cardiogenic shock. Intra-aortic balloon counterpulsation is useful, as well, in pa-
tients with cardiogenic shock because of acute mitral insufficiency and septal rupture. Pa-
tients with right ventricular infarction refractory to fluid therapy may also benefit. Compli-
cations of the devices include arterial injury, bleeding, and hemolysis. Intra-aortic balloon
pumps may be contraindicated in patients with aortic insufficiency, aortic aneurysm, and
severe peripheral arterial atherosclerosis.
Ventricular assist devices, which can be placed in the cardiac catheterization laboratory,
are useful as a means to “buy time” before cardiac transplantation or placement of an im-
planted ventricular assist device. Topalian and coauthors conclude by noting that death
from cardiogenic shock occurs more often in patients who experience a delay in diagnosis of
myocardial infarction. They encourage public education efforts to inform patients of the
importance of early medical care at the first sign of developing acute myocardial ischemia.
Page 110 of 119
Page 111 of 119
Clinical considerations for use of recombinant factor VIIa.
1. Is patient salvage realistically possible? (best results are obtained when
recombinant factor VIIa is used before the onset of terminal acidosis,
5 hypothermia, and coagulopathy)
6 2. Has accessible, surgically controllable bleeding been addressed?
7 3. Has patient been resuscitated to pH > 7.25? (Recombinant factor VIIa is
less effective in acidic environments)
4. Have clotting protein levels been restored with fresh frozen plasma?
5. Is patient platelet count > 50,000? (Recombinant factor VIIa requires
10 adequate numbers of functioning platelets?
11 6. Is potential benefit from treatment with recombinant factor VIIa suffi-
cient to justify the risk? (serious thrombotic complications occur in ap-
proximately 10% of patients who have received recombinant factor
Table 1: Clinical Considerations when use of recombinant factor VIIa is contemplated
Page 112 of 119
Recommended interventions for the initial management of sepsis and
5 septic shock.
6 1. Early (within 6 hrs or less) goal-directed hemodynamic resuscita-
tion guided by central venous pressure (8-12 cmH20 or higher if
7 patient is on positive pressure ventilation), mean arterial pressure
(titrate to > 65 mmHg), and central venous oxygen saturation (ti-
trate to > 70%) using electrolyte or colloid fluids.
2. Blood and other cultures prior to beginning antimicrobial drugs.
3. Imaging studies to confirm origin of sepsis.
4. Administration of broad-spectrum antimicrobial drugs within first
hour of clinical suspicion of sepsis. Continue antimicrobial therapy
for 7-10 days. Adjust antimicrobial therapy depending upon cul-
5. Vasopressor therapy using norepinephrine or dopamine to raise
mean arterial pressure if fluid therapy is unsuccessful.
6. Consider dobutamine therapy if fluids and vasopressors do not re-
store cardiac output.
7. Source control using least invasive procedure that has a high pos-
sibility of success. (see discussion in SRGS Vol. 35 No. 7
8. Stress dose corticosteroid therapy for patients with septic shock
that is poorly responsive the therapy.
9. Use of lung protective ventilation and weaning protocols including
intermittent cessation of sedation. (see discussion in SRGS Vol. 35
10. Use of preventive measures for ventilator associated pneumonia.
11. Control of hyperglycemia.
12. Accept hemoglobin level of 7 gm/dL if patient is stable and free of
clinically evident coronary artery disease.
13. Venous thromboembolism prophylaxis.
Page 113 of 119
14. Stress ulcer prophylaxis.
15. Veno-venous hemofiltration for renal insufficiency. (see discussion
5 1. Goodnough LT. Risks of blood transfusion. Crit Care Med 2003; 31(12 Suppl):S678-86.
6 2. Bernard AC, Davenport DL, Chang PK, et al. Intraoperative transfusion of 1 U to 2 U packed red
7 blood cells is associated with increased 30-day mortality, surgical-site infection, pneumonia, and
8 sepsis in general surgery patients. J Am Coll Surg 2009; 208(5):931-7, 937 e1-2; discussion 938-9.
9 3. Weinberg JA, McGwin G, Jr., Marques MB, et al. Transfusions in the less severely injured: does
10 age of transfused blood affect outcomes? J Trauma 2008; 65(4):794-8.
11 4. Dunne JR, Lee TH, Burns C, et al. Transfusion-associated microchimerism in combat casualties. J
12 Trauma 2008; 64(2 Suppl):S92-7; discussion S97-8.
13 5. Practice guidelines for perioperative blood transfusion and adjuvant therapies: an updated
14 report by the American Society of Anesthesiologists Task Force on Perioperative Blood
15 Transfusion and Adjuvant Therapies. Anesthesiology 2006; 105(1):198-208.
16 6. Singer DE, Chang Y, Fang MC, et al. The net clinical benefit of warfarin anticoagulation in atrial
17 fibrillation. Ann Intern Med 2009; 151(5):297-305.
18 7. Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of
19 transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators,
20 Canadian Critical Care Trials Group. N Engl J Med 1999; 340(6):409-17.
21 8. Hashimoto T, Kokudo N, Orii R, et al. Intraoperative blood salvage during liver resection: a
22 randomized controlled trial. Ann Surg 2007; 245(5):686-91.
23 9. Mirza SB, Campion J, Dixon JH, et al. Efficacy and economics of postoperative blood salvage in
24 patients undergoing elective total hip replacement. Ann R Coll Surg Engl 2007; 89(8):777-84.
25 10. Hughes DB, Ullery BW, Barie PS. The contemporary approach to the care of Jehovah's witnesses.
26 J Trauma 2008; 65(1):237-47.
27 11. Gohel MS, Bulbulia RA, Slim FJ, et al. How to approach major surgery where patients refuse
28 blood transfusion (including Jehovah's Witnesses). Ann R Coll Surg Engl 2005; 87(1):3-14.
29 12. Jabbour N, Gagandeep S, Shah H, et al. Impact of a transfusion-free program on non-Jehovah's
30 Witness patients undergoing liver transplantation. Arch Surg 2006; 141(9):913-7.
31 13. Borchelt BD, Wright PA, Evans JA, et al. Cell swelling and depolarization in hemorrhagic shock. J
32 Trauma 1995; 39(2):187-92; discussion 192-4.
33 14. Zakaria el R, Li N, Matheson PJ, et al. Cellular edema regulates tissue capillary perfusion after
34 hemorrhage resuscitation. Surgery 2007; 142(4):487-96; discussion 496 e1-2.
35 15. Matheson PJ, Mays CJ, Hurt RT, et al. Modulation of mesenteric lymph flow and composition by
36 direct peritoneal resuscitation from hemorrhagic shock. Arch Surg 2009; 144(7):625-34.
37 16. Malinoski DJ, Hadjizacharia P, Salim A, et al. Elevated serum pancreatic enzyme levels after
38 hemorrhagic shock predict organ failure and death. J Trauma 2009; 67(3):445-9.
39 17. Senthil M, Brown M, Xu DZ, et al. Gut-lymph hypothesis of systemic inflammatory response
40 syndrome/multiple-organ dysfunction syndrome: validating studies in a porcine model. J
41 Trauma 2006; 60(5):958-65; discussion 965-7.
42 18. Diebel LN, Liberati DM, Ledgerwood AM, et al. Systemic not just mesenteric lymph causes acute
43 lung injury following hemorrhagic shock. Surgery 2008; 144(4):686-93; discussion 693-4.
Page 114 of 119
1 19. Diebel LN, Liberati DM, Lucas CE, et al. Systemic not just mesenteric lymph causes neutrophil
2 priming after hemorrhagic shock. J Trauma 2009; 66(6):1625-31.
3 20. Jastrow KM, 3rd, Gonzalez EA, McGuire MF, et al. Early cytokine production risk stratifies trauma
4 patients for multiple organ failure. J Am Coll Surg 2009; 209(3):320-31.
5 21. Sperry JL, Friese RS, Frankel HL, et al. Male gender is associated with excessive IL-6 expression
6 following severe injury. J Trauma 2008; 64(3):572-8; discussion 578-9.
7 22. Sperry JL, Nathens AB, Frankel HL, et al. Characterization of the gender dimorphism after injury
8 and hemorrhagic shock: are hormonal differences responsible? Crit Care Med 2008; 36(6):1838-
10 23. Machiedo GW, Zaets SB, Berezina TL, et al. Trauma-hemorrhagic shock-induced red blood cell
11 damage leads to decreased microcirculatory blood flow. Crit Care Med 2009; 37(3):1000-10.
12 24. Gonzales ER, Chen H, Munuve RM, et al. Hepatoprotection and lethality rescue by histone
13 deacetylase inhibitor valproic acid in fatal hemorrhagic shock. J Trauma 2008; 65(3):554-65.
14 25. Alam HB, Shuja F, Butt MU, et al. Surviving blood loss without blood transfusion in a swine poly-
15 trauma model. Surgery 2009; 146(2):325-33.
16 26. Rushing GD, Britt LD. Reperfusion injury after hemorrhage: a collective review. Ann Surg 2008;
18 27. Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med 2003;
20 28. Ayala A, Perl M, Venet F, et al. Apoptosis in sepsis: mechanisms, clinical impact and potential
21 therapeutic targets. Curr Pharm Des 2008; 14(19):1853-9.
22 29. Topalian S, Ginsberg F, Parrillo JE. Cardiogenic shock. Crit Care Med 2008; 36(1 Suppl):S66-74.
23 30. Aymong ED, Ramanathan K, Buller CE. Pathophysiology of cardiogenic shock complicating acute
24 myocardial infarction. Med Clin North Am 2007; 91(4):701-12; xii.
25 31. Lipsky AM, Gausche-Hill M, Henneman PL, et al. Prehospital hypotension is a predictor of the
26 need for an emergent, therapeutic operation in trauma patients with normal systolic blood
27 pressure in the emergency department. J Trauma 2006; 61(5):1228-33.
28 32. McCunn M, Dutton R. End-points of resuscitation how much is enough? Curr Opin Anaesthesiol
29 2000; 13(2):147-53.
30 33. Pinsky MR. Hemodynamic evaluation and monitoring in the ICU. Chest 2007; 132(6):2020-9.
31 34. Taylor RW, Jr., Calvin JE, Matuschak GM. Pulmonary Artery Catheter Consensus Conference: the
32 first step. Crit Care Med 1997; 25(12):2064-5.
33 35. Pinsky MR, Vincent JL. Let us use the pulmonary artery catheter correctly and only when we
34 need it. Crit Care Med 2005; 33(5):1119-22.
35 36. Voga G. Pulmonary artery occlusion pressure estimation by transesophageal echocardiography:
36 is simpler better? Crit Care 2008; 12(2):127.
37 37. Martin MJ, FitzSullivan E, Salim A, et al. Use of serum bicarbonate measurement in place of
38 arterial base deficit in the surgical intensive care unit. Arch Surg 2005; 140(8):745-51.
39 38. FitzSullivan E, Salim A, Demetriades D, et al. Serum bicarbonate may replace the arterial base
40 deficit in the trauma intensive care unit. Am J Surg 2005; 190(6):941-6.
41 39. Jansen TC, van Bommel J, Bakker J. Blood lactate monitoring in critically ill patients: A systematic
42 health technology assessment*. Crit Care Med 2009.
43 40. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe
44 sepsis and septic shock. N Engl J Med 2001; 345(19):1368-77.
45 41. Friese RS, Dineen S, Jennings A, et al. Serum B-type natriuretic peptide: a marker of fluid
46 resuscitation after injury? J Trauma 2007; 62(6):1346-50; discussion 1350-1.
Page 115 of 119
1 42. Ivatury RR, Simon RJ, Islam S, et al. A prospective randomized study of end points of
2 resuscitation after major trauma: global oxygen transport indices versus organ-specific gastric
3 mucosal pH. J Am Coll Surg 1996; 183(2):145-54.
4 43. Hameed SM, Cohn SM. Gastric tonometry: the role of mucosal pH measurement in the
5 management of trauma. Chest 2003; 123(5 Suppl):475S-81S.
6 44. Boswell SA, Scalea TM. Sublingual capnometry: an alternative to gastric tonometry for the
7 management of shock resuscitation. AACN Clin Issues 2003; 14(2):176-84.
8 45. Baron BJ, Dutton RP, Zehtabchi S, et al. Sublingual capnometry for rapid determination of the
9 severity of hemorrhagic shock. J Trauma 2007; 62(1):120-4.
10 46. Moore FA, Nelson T, McKinley BA, et al. Massive transfusion in trauma patients: tissue
11 hemoglobin oxygen saturation predicts poor outcome. J Trauma 2008; 64(4):1010-23.
12 47. Leone M, Blidi S, Antonini F, et al. Oxygen tissue saturation is lower in nonsurvivors than in
13 survivors after early resuscitation of septic shock. Anesthesiology 2009; 111(2):366-71.
14 48. Wan JJ, Cohen MJ, Rosenthal G, et al. Refining resuscitation strategies using tissue oxygen and
15 perfusion monitoring in critical organ beds. J Trauma 2009; 66(2):353-7.
16 49. van Lelyveld-Haas LE, van Zanten AR, Borm GF, et al. Clinical validation of the non-invasive
17 cardiac output monitor USCOM-1A in critically ill patients. Eur J Anaesthesiol 2008; 25(11):917-
19 50. Bickell WH, Wall MJ, Jr., Pepe PE, et al. Immediate versus delayed fluid resuscitation for
20 hypotensive patients with penetrating torso injuries. N Engl J Med 1994; 331(17):1105-9.
21 51. Lucas CE. The water of life: a century of confusion. J Am Coll Surg 2001; 192(1):86-93.
22 52. Todd SR, Malinoski D, Muller PJ, et al. Lactated Ringer's is superior to normal saline in the
23 resuscitation of uncontrolled hemorrhagic shock. J Trauma 2007; 62(3):636-9.
24 53. Cotton BA, Guy JS, Morris JA, Jr., et al. The cellular, metabolic, and systemic consequences of
25 aggressive fluid resuscitation strategies. Shock 2006; 26(2):115-21.
26 54. Koustova E, Stanton K, Gushchin V, et al. Effects of lactated Ringer's solutions on human
27 leukocytes. J Trauma 2002; 52(5):872-8.
28 55. Kramer GC. Hypertonic resuscitation: physiologic mechanisms and recommendations for trauma
29 care. J Trauma 2003; 54(5 Suppl):S89-99.
30 56. Meybohm P, Cavus E, Bein B, et al. Small volume resuscitation: a randomized controlled trial
31 with either norepinephrine or vasopressin during severe hemorrhage. J Trauma 2007; 62(3):640-
33 57. Sperry JL, Minei JP, Frankel HL, et al. Early use of vasopressors after injury: caution before
34 constriction. J Trauma 2008; 64(1):9-14.
35 58. Rizoli SB, Rhind SG, Shek PN, et al. The immunomodulatory effects of hypertonic saline
36 resuscitation in patients sustaining traumatic hemorrhagic shock: a randomized, controlled,
37 double-blinded trial. Ann Surg 2006; 243(1):47-57.
38 59. Deree J, de Campos T, Shenvi E, et al. Hypertonic saline and pentoxifylline attenuates gut injury
39 after hemorrhagic shock: the kinder, gentler resuscitation. J Trauma 2007; 62(4):818-27;
40 discussion 827-8.
41 60. Coimbra R, Loomis W, Melbostad H, et al. Role of hypertonic saline and pentoxifylline on
42 neutrophil activation and tumor necrosis factor-alpha synthesis: a novel resuscitation strategy. J
43 Trauma 2005; 59(2):257-64; discussion 264-5.
44 61. Creteur J, Vincent JL. Potential uses of hemoglobin-based oxygen carriers in critical care
45 medicine. Crit Care Clin 2009; 25(2):311-24, Table of Contents.
46 62. Koehler RC, Fronticelli C, Bucci E. Insensitivity of cerebral oxygen transport to oxygen affinity of
47 hemoglobin-based oxygen carriers. Biochim Biophys Acta 2008; 1784(10):1387-94.
Page 116 of 119
1 63. Natanson C, Kern SJ, Lurie P, et al. Cell-free hemoglobin-based blood substitutes and risk of
2 myocardial infarction and death: a meta-analysis. JAMA 2008; 299(19):2304-12.
3 64. Napolitano LM. Hemoglobin-based oxygen carriers: first, second or third generation? Human or
4 bovine? Where are we now? Crit Care Clin 2009; 25(2):279-301, Table of Contents.
5 65. Moore EE, Moore FA, Fabian TC, et al. Human polymerized hemoglobin for the treatment of
6 hemorrhagic shock when blood is unavailable: the USA multicenter trial. J Am Coll Surg 2009;
8 66. Dudkiewicz M, Harpaul TA, Proctor KG. Hemoglobin-based oxygen carrying compound-201 as
9 salvage therapy for severe neuro- and polytrauma (Injury Severity Score = 27-41). Crit Care Med
10 2008; 36(10):2838-48.
11 67. Hess JR, Brohi K, Dutton RP, et al. The coagulopathy of trauma: a review of mechanisms. J
12 Trauma 2008; 65(4):748-54.
13 68. Reed RL, 2nd, Bracey AW, Jr., Hudson JD, et al. Hypothermia and blood coagulation: dissociation
14 between enzyme activity and clotting factor levels. Circ Shock 1990; 32(2):141-52.
15 69. Wolberg AS, Meng ZH, Monroe DM, 3rd, et al. A systematic evaluation of the effect of
16 temperature on coagulation enzyme activity and platelet function. J Trauma 2004; 56(6):1221-8.
17 70. Gando S. Acute coagulopathy of trauma shock and coagulopathy of trauma: a rebuttal. You are
18 now going down the wrong path. J Trauma 2009; 67(2):381-3.
19 71. Martini WZ, Cortez DS, Dubick MA, et al. Thrombelastography is better than PT, aPTT, and
20 activated clotting time in detecting clinically relevant clotting abnormalities after hypothermia,
21 hemorrhagic shock and resuscitation in pigs. J Trauma 2008; 65(3):535-43.
22 72. Hoyt DB, Dutton RP, Hauser CJ, et al. Management of coagulopathy in the patients with multiple
23 injuries: results from an international survey of clinical practice. J Trauma 2008; 65(4):755-64;
24 discussion 764-5.
25 73. Holcomb JB, Wade CE, Michalek JE, et al. Increased plasma and platelet to red blood cell ratios
26 improves outcome in 466 massively transfused civilian trauma patients. Ann Surg 2008;
28 74. Sperry JL, Ochoa JB, Gunn SR, et al. An FFP:PRBC transfusion ratio >/=1:1.5 is associated with a
29 lower risk of mortality after massive transfusion. J Trauma 2008; 65(5):986-93.
30 75. Teixeira PG, Inaba K, Shulman I, et al. Impact of plasma transfusion in massively transfused
31 trauma patients. J Trauma 2009; 66(3):693-7.
32 76. Duchesne JC, Islam TM, Stuke L, et al. Hemostatic resuscitation during surgery improves survival
33 in patients with traumatic-induced coagulopathy. J Trauma 2009; 67(1):33-7; discussion 37-9.
34 77. Scalea TM, Bochicchio KM, Lumpkins K, et al. Early aggressive use of fresh frozen plasma does
35 not improve outcome in critically injured trauma patients. Ann Surg 2008; 248(4):578-84.
36 78. Watson GA, Sperry JL, Rosengart MR, et al. Fresh frozen plasma is independently associated with
37 a higher risk of multiple organ failure and acute respiratory distress syndrome. J Trauma 2009;
38 67(2):221-7; discussion 228-30.
39 79. Cotton BA, Au BK, Nunez TC, et al. Predefined massive transfusion protocols are associated with
40 a reduction in organ failure and postinjury complications. J Trauma 2009; 66(1):41-8; discussion
42 80. Inaba K, Teixeira PG, Shulman I, et al. The impact of uncross-matched blood transfusion on the
43 need for massive transfusion and mortality: analysis of 5,166 uncross-matched units. J Trauma
44 2008; 65(6):1222-6.
45 81. Riskin DJ, Tsai TC, Riskin L, et al. Massive transfusion protocols: the role of aggressive
46 resuscitation versus product ratio in mortality reduction. J Am Coll Surg 2009; 209(2):198-205.
47 82. Dutton RP, McCunn M, Hyder M, et al. Factor VIIa for correction of traumatic coagulopathy. J
48 Trauma 2004; 57(4):709-18; discussion 718-9.
Page 117 of 119
1 83. Spinella PC, Perkins JG, McLaughlin DF, et al. The effect of recombinant activated factor VII on
2 mortality in combat-related casualties with severe trauma and massive transfusion. J Trauma
3 2008; 64(2):286-93; discussion 293-4.
4 84. Perkins JG, Schreiber MA, Wade CE, et al. Early versus late recombinant factor VIIa in combat
5 trauma patients requiring massive transfusion. J Trauma 2007; 62(5):1095-9; discussion 1099-
7 85. Horton JD, DeZee KJ, Wagner M. Use of rFVIIa in the trauma setting--practice patterns in United
8 States trauma centers. Am Surg 2008; 74(5):413-7.
9 86. Boffard KD, Riou B, Warren B, et al. Recombinant factor VIIa as adjunctive therapy for bleeding
10 control in severely injured trauma patients: two parallel randomized, placebo-controlled,
11 double-blind clinical trials. J Trauma 2005; 59(1):8-15; discussion 15-8.
12 87. Thomas GO, Dutton RP, Hemlock B, et al. Thromboembolic complications associated with factor
13 VIIa administration. J Trauma 2007; 62(3):564-9.
14 88. Ranucci M, Isgro G, Soro G, et al. Efficacy and safety of recombinant activated factor vii in major
15 surgical procedures: systematic review and meta-analysis of randomized clinical trials. Arch Surg
16 2008; 143(3):296-304; discussion 304.
17 89. Hollenberg SM, Ahrens TS, Annane D, et al. Practice parameters for hemodynamic support of
18 sepsis in adult patients: 2004 update. Crit Care Med 2004; 32(9):1928-48.
19 90. Finfer S, Bellomo R, Boyce N, et al. A comparison of albumin and saline for fluid resuscitation in
20 the intensive care unit. N Engl J Med 2004; 350(22):2247-56.
21 91. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective
22 antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med
23 2006; 34(6):1589-96.
24 92. Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Campaign: international guidelines for
25 management of severe sepsis and septic shock: 2008. Intensive Care Med 2008; 34(1):17-60.
26 93. Kortgen A, Niederprum P, Bauer M. Implementation of an evidence-based "standard operating
27 procedure" and outcome in septic shock. Crit Care Med 2006; 34(4):943-9.
28 94. Shapiro NI, Howell MD, Talmor D, et al. Implementation and outcomes of the Multiple Urgent
29 Sepsis Therapies (MUST) protocol. Crit Care Med 2006; 34(4):1025-32.
30 95. Jhanji S, Stirling S, Patel N, et al. The effect of increasing doses of norepinephrine on tissue
31 oxygenation and microvascular flow in patients with septic shock. Crit Care Med 2009;
33 96. Hollenberg SM. Vasopressor support in septic shock. Chest 2007; 132(5):1678-87.
34 97. Sakr Y, Reinhart K, Vincent JL, et al. Does dopamine administration in shock influence outcome?
35 Results of the Sepsis Occurrence in Acutely Ill Patients (SOAP) Study. Crit Care Med 2006;
37 98. Russell JA, Walley KR, Singer J, et al. Vasopressin versus norepinephrine infusion in patients with
38 septic shock. N Engl J Med 2008; 358(9):877-87.
39 99. Russell JA, Walley KR, Gordon AC, et al. Interaction of vasopressin infusion, corticosteroid
40 treatment, and mortality of septic shock. Crit Care Med 2009; 37(3):811-8.
41 100. Leone MM, C. How to use adrenergic support in sepsis. Advances in Sepsis 2006; 5(3):78-87.
42 101. Cronin L, Cook DJ, Carlet J, et al. Corticosteroid treatment for sepsis: a critical appraisal and
43 meta-analysis of the literature. Crit Care Med 1995; 23(8):1430-9.
44 102. Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone
45 and fludrocortisone on mortality in patients with septic shock. JAMA 2002; 288(7):862-71.
46 103. Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock. N Engl
47 J Med 2008; 358(2):111-24.
Page 118 of 119
1 104. Marik PE, Pastores SM, Annane D, et al. Recommendations for the diagnosis and management
2 of corticosteroid insufficiency in critically ill adult patients: consensus statements from an
3 international task force by the American College of Critical Care Medicine. Crit Care Med 2008;
5 105. Fuller BM, Dellinger RP. Hemodynamic resuscitation in septic shock: cardiovascular support and
6 adjunctive therapy. Curr Infect Dis Rep 2009; 11(5):357-64.
7 106. Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and safety of recombinant human activated
8 protein C for severe sepsis. N Engl J Med 2001; 344(10):699-709.
Page 119 of 119