VIEWS: 76 PAGES: 9 POSTED ON: 8/14/2011
Clin Chest Med 27 (2006) 549–557 Clinical Epidemiology of Acute Lung Injury and Acute Respiratory Distress Syndrome: Incidence, Diagnosis, and Outcomes Jaime F. Avecillas, MDa,b, Amado X. Freire, MD, MPHa, Alejandro C. Arroliga, MDc,* a Division of Pulmonary, Critical Care, and Sleep Medicine, University of Tennessee Health Science Center at Memphis, 956 Court Avenue, Memphis, TN 38163, USA b Memphis VA Medical Center, 1030 Jeﬀerson Avenue, Room G401H, Memphis, TN 38104, USA c Division of Pulmonary and Critical Care Medicine, Scott & White/Texas A&M University System, 2401 South 31st Street, Temple, TX 76508, USA Acute respiratory distress syndrome (ARDS) by the development of newer therapeutic ap- and acute lung injury (ALI) are devastating proaches that have resulted in decreased mortality clinical syndromes that are characterized by the and increased number of days without ventilator sudden onset of severe hypoxemia and diﬀuse use in patients who have ARDS [5,6]. Many ques- bilateral pulmonary inﬁltrates in the absence of tions remain regarding the incidence, diagnosis, left atrial hypertension . The severity of the and outcomes of these entities, however. hypoxemia diﬀerentiates these two entities, with This article reviews the currently used deﬁni- a ratio of partial pressure of oxygen in arterial tions of ARDS and ALI and the diﬀerent studies blood (PaO2) to inspired fraction of oxygen that have advanced the understanding of the (FIO2) of less than 200 mm Hg for ARDS and of epidemiology and outcomes of these entities. It less than 300 mm Hg for ALI. Both ALI and also discusses several diagnostic issues that are ARDS can occur from a wide diversity of causes important for both clinicians and researchers. or associated conditions, such as sepsis, pneumo- nia, aspiration, trauma, pancreatitis, massive blood transfusions, and smoke or toxic gas inhala- Incidence tion [2,3]. Although ARDS initially was reported in 1967 Although signiﬁcant progress has been made in by Ashbaugh and colleagues , it was not until the understanding of the natural history and 1992 that the American-European Consensus pathogenesis of ARDS, its incidence has remained Conference (AECC) on ARDS developed a stan- uncertain. One of the ﬁrst estimates of the in- dardized deﬁnition for ARDS and ALI . Dur- cidence of ARDS in the United States is from ing the last 4 decades signiﬁcant advances have a National Heart, Lung and Blood Institute been made in the understanding of the etiology, workshop organized in 1976 to provide a better pathophysiology, and epidemiology of ALI and understanding of the mechanisms of ALI . The ARDS. These advances have been accompanied expert panel of this conference estimated that there are 150,000 cases of ARDS per year in the United States, an incidence of about 75 cases per 100,000 persons. Since this initial estimate * Corresponding author. , diﬀerent estimates of the incidence of ARDS E-mail address: email@example.com and ALI have been reported, varying from 1.5 (A.C. Arroliga). to 78 cases per 100,000 persons depending on 0272-5231/06/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ccm.2006.06.001 chestmed.theclinics.com 550 AVECILLAS et al the population studied and the deﬁnition of about 3.5 million persons. Another population- ARDS and ALI used (Table 1) [8–17]. based study of patients who had ARDS performed From 1983 to 1995 most studies showed a rel- in the United States by Thomsen and colleagues atively low incidence for ARDS, ranging from 1.5  estimated the incidence of ARDS as between to 8.3 cases per 100,000 persons [8–11,18]. Between 4.8 to 8.3 cases per 100,000 person-years. This 1983 and 1985 Villar and colleagues  took ad- study, however, included only 6 of the 40 general vantage of the unique characteristics of the Canary acute-care hospitals in the state of Utah. All these Islands (an island with a relatively large popula- studies were designed before the AECC deﬁnition tion) to determine the incidence of ARDS. The of ARDS and ALI became available in 1994. overall incidence of ARDS in that study was Therefore, no standard criterion for the diagnosis reported to be as low as 1.5 to 3.5 cases per of ARDS was used. 100,000 persons, depending on the diagnostic crite- Recent data have shown higher incidences for ria used. In 1988 Webster and colleagues  re- ALI and ARDS [14,16,17]. In a prospective co- ported a similar incidence for ARDS (4.5 cases hort study with extrapolation to the United States per 100,000 persons). This study was conducted population, Goss and colleagues  estimated in a British health district with a population of the incidence of ALI in the United States at 64.2 Table 1 Comparison of selected incidence studies for acute lung injury and acute respiratory distress syndrome Geographic area, length, Authors and year of study Deﬁnition of ARDS/ALI Incidence Villar et al  Canary Islands; 36 months; PaO2 % 55 mm Hg with 1.5–3.5 cases per 105 1983–1985 FIO2 O 0.5, bilateral person-years for ARDS inﬁltrates, no evidence of left atrial hypertension Webster et al  UK; 12 months; 1985 Not speciﬁed 4.5 cases per 105 person-years for ARDS Thomsen et al  Utah; 12 months; PaO2/FIO2 %200 mm Hg, 4.8–8.3 cases per 105 1989–1990 bilateral inﬁltrates, no person-years for ARDS evidence of left atrial hypertension Lewandowski et al  Berlin; 8 weeks 1991 LIS R 2.5 3 cases per 105 person-years for ARDS Luhr et al  Sweden, Denmark, Iceland; AECC 17.9 cases per 105 8 weeks; 1997 person-years for ALI 13.5 cases per 105 person-years for ARDS Arroliga et al  Cleveland, OH; 35 months; AECC 15.3 cases per 105 1996–1999 person-years for ARDS Goss et al  US; 1996–1999/ARDS AECC 22.4–64.2 cases per 105 network data person-years for ALI Hughes et al  Scotland; 8 months; 1999 AECC 16 cases per 105 person-years for ARDS Bersten et al  Australia; 8 weeks; 1999 AECC 34 cases per 105 person-years for ALI 28 cases per 105 person-years for ARDS Rubenfeld et al  King County, WA; AECC 78.9 cases per 105 12 months; 1999–2000 person-years for ALI 58.7 cases per 105 person-years for ARDS Abbreviations: AECC, American European Consensus Conference deﬁnition of acute lung injury and acute respira- tory distress syndrome; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, inspired fraction of oxygen; LIS, lung injury score; OH, Ohio; PaO2, partial pressure of oxygen in arterial blood; US, United States; UK, United Kingdom; WA, Washington. CLINICAL EPIDEMIOLOGY OF ALI AND ARDS 551 cases per 100,000 person-years. To obtain this es- 3. Diﬀerent lengths of the study period timate, they used prospective clinical trial screen- 4. Diﬀerent study designs and assumptions ing data (ARDS network) and conservative (sometimes unveriﬁed) to estimate incidence assumptions about where patients who have ALI 5. Heterogeneity of diseases underlying ALI and receive care (they assumed that ALI cases oc- ARDS curred only at hospitals with 20 or more ICU 6. Diﬀerences in the structure of the health care beds). Moreover, a recent large prospective, systems within the countries where studies population-based cohort study by Rubenfeld have been performed (ie, diﬀerent use of and colleagues  showed that the crude inci- ICU resources) dence of ALI in King County, Washington was 7. Seasonal variation 78.9 per 100,000 person-years, and the incidence 8. Diﬃculties ﬁnding all cases in a given geo- of ARDS was 58.7 cases per 100,000 person-years. graphical area This study included all 18 hospitals in King It is important to determine the real incidence County, and patients were identiﬁed according of ALI and ARDS for the eﬃcient distribution to the deﬁnition of the AECC on ALI and and allocation of clinical resources. Although it is ARDS. This last study, however, was conducted diﬃcult to reconcile the diﬀerent incidences re- in a geographic area with a population that diﬀers ported by the diﬀerent authors, the authors signiﬁcantly from the rest of the United States, be- believe that the recent estimate by Rubenfeld ing wealthier, younger, with more Asians, and and colleagues  may be more accurate than with fewer African-Americans. previous estimates, because this study included Other factors (eg, age and the underlying risk a large number of patients from multiple ICUs factors associated with the syndrome) may aﬀect who were followed for a long period of time the reported incidence of ALI and ARDS. The (1 year). This study and other recent studies incidence of ALI and ARDS seems to increase [14,16] clearly show that ALI and ARDS are rel- with age [2,17]. Rubenfeld and colleagues  atively common conditions that have the potential showed that the incidence of this entity increases to become a public health burden. with age, from 16 per 100,000 person-years for those 15 through 19 years of age to 306 per 100,000 person-years for those 75 through 84 years of age. Few studies actually have mea- Diagnosis sured the incidence of ARDS prospectively in The diagnosis of ALI and ARDS (Table 2) is selected groups of patients thought to be at in- based on clinical and radiologic information, ab- creased risk for the syndrome [2,18]. Fowler normalities of oxygenation, and occasionally on and colleagues  found that the incidence of lung biopsies. The accuracy of these criteria may ARDS varies by predisposed groups, being low- be limited by several factors, some of which are est in most categories associated with nonpulmo- discussed in this section. nary injury and the highest after witnessed The pathologic diagnosis of diﬀuse alveolar pulmonary aspiration. On the other hand, in damage (DAD) is considered the standard for the a similar study Hudson and colleagues  diagnosis of ALI and ARDS . After reviewing showed that the highest incidence of ARDS oc- 420 cases of DAD, however, Katzenstein and col- curred in patients who had sepsis syndrome and leagues  concluded that DAD is not a diagnosis multiple transfusions, whereas the lowest inci- but a nonspeciﬁc reaction of the lung to a multitude dence was seen in trauma patients and patients of injurious agents. The pathologic diagnosis of who had drug overdose. DAD is rarely available in clinical practice because The medical literature has reported a wide of the low rates of open-lung biopsy seen in this range of values for the incidence of ALI and group of patients. Nonetheless, Patel and col- ARDS, as already shown. Variations among leagues , in a study involving 57 patients who studies might be ascribed to the following factors: had ARDS and who underwent open-lung biopsy, 1. The use of diﬀerent deﬁnitions of ALI and showed that in selected patients who have clinical ARDS ARDS, open-lung biopsy can be performed safely, 2. Cultural, demographic, and socioeconomic often reveals an unsuspected diagnosis, and fre- diﬀerences in the diﬀerent populations quently leads to alterations in therapy. This study studied revealed that a pathologic diagnosis other than 552 AVECILLAS et al Table 2 American European Consensus Conference deﬁnition of acute lung injury and acute respiratory distress syndrome Acute lung injury Acute respiratory distress syndrome Acute onset Acute onset PaO2/FIO2 of less than 300 mm Hg PaO2/FIO2 of less than 200 mm Hg Bilateral alveolar inﬁltrates on chest radiograph Bilateral alveolar inﬁltrates on chest radiograph PAWP less than 18 mm Hg or no clinical evidence PAWP less than 18 mm Hg or no clinical evidence of left atrial hypertension of left atrial hypertension Abbreviations: PaO2/FIO2, ratio of partial pressure of oxygen in arterial blood to inspired fraction of oxygen; PAWP, pulmonary artery wedge pressure. (Data from Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Deﬁnitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149 (3 Pt 1):819.) DAD or ﬁbroproliferation was found in up to 60% speciﬁcity of the AECC deﬁnition of ARDS may of patients, biopsy results led to a change in man- be diﬀerent in clinical practice. agement in most patients, and major complica- The AECC deﬁnition of ALI and ARDS tions occurred in only 7% of cases . requires the presence of bilateral alveolar inﬁl- In clinical practice the AECC deﬁnition of ALI trates on chest radiograph. The identiﬁcation of and ARDS is used to diagnose these two entities, individuals who have ARDS based on character- even though the deﬁnitions were devised mainly to istic radiographic ﬁndings can be challenging, assist with clinical and epidemiologic research. The however, and diﬀerent studies have shown that AECC deﬁnition of ALI and ARDS (see Table 2) the interpretation of chest radiographs in these requires the acute onset of bilateral alveolar inﬁl- patients is not as reliable as one would expect trates on chest radiograph, a PaO2/FIO2 ratio of [23,24]. Rubenfeld and colleagues  asked 21 less than 200 mm Hg for ARDS and of less than experts to review 28 randomly selected chest ra- 300 mm Hg for ALI, and a pulmonary artery wedge diographs from critically ill hypoxemic patients pressure (PAWP) less than 18 mm Hg or no clinical and decide if the chest radiograph fulﬁlled the evidence of left atrial hypertension . Although AECC deﬁnition for ALI and ARDS. The inter- diﬀerent authors have pointed out some of the lim- observer agreement in applying the AECC deﬁni- itations of this deﬁnition [12,19,22], it is used widely tion was moderate (k ¼ 0.55), and there was full in research and in clinical practice. agreement on fewer than half of the chest radio- The deﬁnition suggested by the AECC has graphs (43%). Meade and colleagues  also limitations. In a recent series of autopsy patients found that intensivists are able to achieve only Esteban and colleagues  compared the clinical moderate levels of agreement in the radiologic di- criteria for ARDS with autopsy ﬁndings and agnosis of ARDS. The level of agreement, how- showed that the accuracy of the AECC deﬁnition ever, could increase substantially after consensus of ARDS was only moderate, with a sensitivity of training (some physicians participated in pilot 75% (95% conﬁdence interval, 66%–82%) and training as part of a randomized trial of a pressure- a speciﬁcity of 84% (95% conﬁdence interval, and volume-limited ventilation strategy in patients 79%–88%). In this study the deﬁnition of at high risk for ARDS). Finally, in clinical prac- ARDS was more accurate for patients who had tice it is common to encounter patients who extrapulmonary risk factors (sepsis syndrome, have ALI or ARDS but who initially present multiple trauma, several blood transfusions, with an atypical chest radiograph or with bilateral shock, or pancreatitis) than for patients who had inﬁltrates caused by chronic conditions such as id- pulmonary risk factors (pneumonia, aspiration, iopathic pulmonary ﬁbrosis, asbestosis, sarcoido- near-drowning, inhalational injury, or lung contu- sis, or lymphangitic carcinoma. In these cases it sion). Nonetheless, this study included only pa- is important for the clinician to do a general as- tients who died and underwent autopsy; sessment of oxygenation and overall clinical con- therefore, these results may not apply to less se- dition to be able to make the right diagnostic vere cases of ARDS. Another limitation of this and therapeutics decisions in a timely fashion. study is that the diagnosis of ARDS was made us- Another limitation of the AECC deﬁnition is ing retrospective analysis of clinical and radio- that, besides using clinical information including graphic ﬁndings. Therefore, the sensitivity and physical examination and radiologic evaluation, it CLINICAL EPIDEMIOLOGY OF ALI AND ARDS 553 may consider the measurement of PAWP to deﬁnition is developed (see the article by Rice evaluate the presence of left atrial hypertension. and Bernard in this issue). The accuracy of this evaluation can vary depending on the method used. Al-Kharrat and colleagues  conducted a study to determine Outcomes inter- and intraobserver variability in measure- Mortality ment of pulmonary artery occlusion pressure (Ppao). These authors compared values recorded Diﬀerent studies report a wide range of by critical care nurses and physician specialists mortality rates for patients who have ALI and (cardiologists and critical care physicians) and ARDS, varying between 23% and 71% (Table 3) found signiﬁcant diﬀerences in the measurement [5,6,12,15–18,28–36]. Diﬀerences between reported of Ppao tracings, not only between physicians fatality rates may be explained by diﬀerent factors: and critical care nurses but also among physi- 1. The use of diﬀerent proportions of patients cians. The ﬁndings of this study are extremely im- who have ARDS caused by pulmonary or ex- portant, because critical care nurses perform the trapulmonary causes, because signiﬁcant dif- measurements of Ppao in most hospitals, and sev- ferences in mortality have been reported in eral aspects of patient care depend these measure- these two patient populations [2,18,32,33]. ments. A more recent study by Rizvi and (For instance, a variable proportion of colleagues  showed that the addition of an air- ARDS cases caused by trauma or by Pneumo- way pressure signal (Paw) to pressure tracings of cystis jiroveci in patients who have AIDS Ppao would improve interobserver agreement by might aﬀect the mortality rate, because these facilitating identiﬁcation of end-expiration. In subsets of patients have been associated with this study, agreement increased from 71% without a lower and a higher mortality rate, respec- Paw to 83% with Paw. The use of a PAWP of less tively [15,28].) than 18 mm Hg as the cut-oﬀ point for the diag- 2. Diﬀerent severity of underlying illness in the nosis of ALI and ARDS also has been challenged populations studied and the use of diﬀerent because mild-to-moderate levels of elevated left scoring systems for disease severity. atrial pressure and ALI may coexist . In 3. The use of diﬀerent deﬁnitions of the main a post hoc analysis of a randomized, controlled outcome measure (ie, hospital, ICU, or trial of pressure- and volume-limited ventilation 28-day mortality). involving 120 patients who had ARDS but who 4. The type of ventilatory strategy used [5,6]. had no risk factors for congestive heart failure; Ferguson and colleagues  found that Failure to control for these factors has resulted a PAWP higher than 18 mm Hg is seen commonly in results that vary widely from study to study, in these patients, even though these patients had making it more diﬃcult to translate the reported relatively high cardiac indices. Possible explana- mortality rates to individual ICU populations. tions for an increased PAWP in ARDS patients Early reports of the mortality of ARDS include aggressive intravenous ﬂuid resuscitation, showed a high mortality rate (64%–70%) [8,18]. increased pleural pressures, or the use of positive Over the last 20 years, however, the fatality rate end-expiratory pressure during mechanical venti- for patients who have ARDS has shown a trend lation . toward improvement, being between 29% and Finally, the AECC deﬁnition of ALI and 42% in more recent studies [17,37–40]. A retro- ARDS does not specify the exact timing for the spective study by Stapleton and colleagues  evaluation of the chest radiograph and the PaO2/ showed that overall fatality has decreased at their FIO2 ratio, and that timing could have a signiﬁcant institution during the last 2 decades (from 68% in impact at the time of patient selection for the dif- 1981–1982 to 29% in 1996). In this study the in- ferent studies. All the diﬀerent factors discussed creased survival was accounted for entirely by pa- here can potentially alter the sensitivity and the tients who presented with trauma, but survival for speciﬁcity of the AECC deﬁnition of ALI and patients whose risk factor was sepsis did not ARDS, making comparisons among the diﬀerent change. Milberg and colleagues  also observed studies more diﬃcult to achieve. Therefore, all a signiﬁcant decline in fatality rates (ranging from these limitations of the AECC deﬁnition of ALI 53%–68% between 1983 and 1987 to a low 36% and ARDS must be considered at the time in 1993). In this study, however, the decline in of study design, at least until a more suitable mortality occurred largely in patients who had 554 AVECILLAS et al Table 3 Diﬀerent reported mortality rates for acute lung injury and acute respiratory distress syndrome Geographic area of Authors study (year published) Deﬁnition of ARDS/ALI Mortality rate (%) Baumman et al  US (1986) PaO2/FIO2 % 200 mm Hg, 64 for ARDS bilateral inﬁltrates, no evidence of left atrial hypertension Hickling et al  New Zealand (1994) PaO2/FIO2 % 150 mm Hg, 26.6 for ARDS bilateral inﬁltrates, no evidence of left atrial hypertension Doyle et al  US (1995) AECC 58 for ALI Lewandowski et al  Germany (1995) LIS R 2.5 58.8 for ARDS Zilberberg et al  US (1998) AECC 58 for ALI 58 for ARDS Amato et al  Brazil (1998) LIS R 2.5, no evidence of 38–71 for ARDS left atrial hypertension depending on the ventilation modality used Roupie et al  France (1999) AECC 60 for ARDS Luhr et al  Sweden, Denmark, AECC 42 for ALI Iceland (1999) 41 for ARDS ARDS Network  US (2000) PaO2/FIO2 % 300 mm Hg, 31–39.8 for ARDS bilateral inﬁltrates, no depending on the evidence of left atrial ventilation modality hypertension used Suntharalingam et al  UK (2001) AECC 27.9–47.4 for ARDS depending on the risk factor Bersten et al  Australia (2002) AECC 32 for ALI 34 for ARDS Estenssoro et al  Argentine (2002) AECC 58 for ARDS Arroliga et al  US (2002) AECC 48.5 for ARDS Hughes et al  Scotland (2003) AECC 60.9 for ARDS Rubenfeld et al  US (2005) AECC 38.5 for ALI 41.1 for ARDS Abbreviations: AECC, American European Consensus Conference deﬁnition of acute lung injury and acute respira- tory distress syndrome; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, inspired fraction of oxygen; LIS, lung injury score; PaO2, partial pressure of oxygen in arterial blood; US, United States; UK, United Kingdom. sepsis syndrome as their risk factor for ARDS and mortality of patients who had ARDS remained in those younger than 60 years of age. Rubenfeld constant (about 50%) throughout the period and colleagues  recently showed that the in- they studied (1967–1994), and other studies in hospital mortality was 38.5% for patients who the ARDS literature still report mortality rates had ALI and 41.4% for those who had ARDS. as high as 60% [15,34], keeping open the debate This improvement in mortality may be explained about ARDS mortality. Therefore, intensivists by a general improvement in supportive care for who take care of ARDS patients in their daily the critically ill and probably by the use of new practice should be familiar with the most recently strategies for mechanical ventilation in these pa- reported data regarding fatality rates for this dis- tients. Most of these studies are from academic ease, because this information makes possible bet- tertiary care centers, and the patient population ter counseling of patients and their families. and management of ARDS in these centers may ALI and ARDS do not occur in isolation, and diﬀer from that in community hospitals. mortality will vary according to the risk factor On the other hand, Kraﬀt and colleagues  implicated. Rubenfeld and colleagues  found in a meta-analysis of older series found that the that mortality varied from 24.1% among patients CLINICAL EPIDEMIOLOGY OF ALI AND ARDS 555 who had severe trauma, to 40.6% among patients during the admission to the ICU, the presence of who had severe sepsis with a suspected pulmonary illness acquired during the ICU stay, and the source, to 43.6% among patients in whom aspira- rate of resolution of the lung injury and multior- tion was witnessed. Sepsis syndrome with multiple gan dysfunction during the ICU stay (as reﬂected organ failure (see the article by Vincent in this is- by the slope of the Lung Injury Score and the sue), rather than isolated respiratory insuﬃciency, Multiple Organ Dysfunction Score, respectively). is the most common predisposing risk factor for It seems that a signiﬁcant proportion of ARDS ALI and ARDS [1,17, 30,34] and is the most com- survivors also have neurocognitive impairments, mon cause of death in these patients [13,15,39]. including impaired memory, attention, concentra- Martin and colleagues  analyzed the occur- tion, mental processing speed, and global intellec- rence of sepsis from 1979 trough 2000 using a na- tual decline [46,47]. The neurocognitive and tionally representative sample of all nonfederal emotional morbidity and decreased quality of acute care hospitals in the United States and life persist at least 2 years in these patients, with found an 8.7% annualized increase in the moderate to severe depression and anxiety in ap- incidence of sepsis. Future research in ARDS proximately 23% of patients . In addition, and ALI also should target the morbidity of sep- Schelling and colleagues  evaluated a group sis and the systemic inﬂammatory response of long-term ARDS survivors for posttraumatic syndrome. stress disorder and found that 27.5% had this condition . Their data showed that all these patients who had posttraumatic stress disorder Morbidity were deeply sedated (Ramsay sedation score 5 to Early studies that focused on aspects of re- 6) during the critical phase of their illness, suggest- covery from ALI and ARDS paid more attention ing that the majority of the traumatic events oc- to respiratory symptoms and lung function curred during weaning, when the sedation is [43,44]. In fact, most survivors present moderate- usually decreased. A signiﬁcant proportion of to-severe impairment with low forced vital ARDS survivors undergo rehabilitation, but capacity (FVC), forced expiratory volume in 1 sec- most of these patients are not evaluated for cogni- ond (FEV1), FEV1/FVC ratio, total lung capacity tive impairment . Every rehabilitation center (TLC), and carbon monoxide diﬀusing capacity taking care of ARDS survivors should be able (DLCO) [43,44]. McHugh and colleagues  in to provide screening and therapy for neurocogni- a prospective analysis of a cohort of 82 ARDS tive deﬁcits in a timely fashion. survivors determined the rate of improvement in Diﬀerent studies have shown that a signiﬁcant pulmonary function abnormalities and self- number of ARDS survivors will not resume perceived health during the 12 months after the normal activities or return to work within 1 episode of ARDS. At extubation the majority of year of the acute event [43,47]. Some of these patients had abnormally low FVC, TLC, and patients may need prolonged rehabilitation and DLCO. Pulmonary function tests improved sub- disability, representing a substantial health care stantially between extubation and month 3, with burden. If, in fact, the mortality rate after additional slight improvement at 6 months. ARDS is declining, more survivors will be at More recent studies have shown that patients risk for ARDS-related morbidity, which seems who survive ARDS also are at risk for other types to be signiﬁcant. Most of the available informa- of physical, neuropsychologic, and emotional tion regarding outcomes in patients who have impairment and diminished quality of life [45– survived ARDS comes from studies that have 48]. Herridge and colleagues  followed 109 not included a comparison group of ICU survi- survivors of ARDS 3, 6, and 12 months after vors who did not have ARDS. Therefore, it is discharge from the ICU. This study showed that necessary to use caution when interpreting these these patients have persistent functional limita- data, because all these sequelae may be related tions, largely as a result of muscle wasting and to morbidities from critical illness in general weakness and, to a lesser extent, caused by entrap- and not be unique to ARDS. More studies ment neuropathy, heterotopic ossiﬁcation, and de- about ARDS demography and associated out- creased lung function . In this cohort of comes are needed to provide more accurate in- patients the most important determinants of the formation to public health entities so they can inability to exercise during the ﬁrst year of allocate resources to help in reintegrate these pa- follow-up were use of systemic corticosteroids tients to society. 556 AVECILLAS et al Summary failure in Berlin, Germany. Am J Respir Crit Care Med 1995;151(4):1121–5. A good understanding of the epidemiology of  Luhr OR, Antonsen K, Karlsson M, et al. Incidence ARDS is important for the allocation of health and mortality after acute respiratory failure and care resources and research funds and also for the acute respiratory distress syndrome in Sweden, evaluation of new therapeutic interventions and Denmark, and Iceland. The ARF Study Group. health care delivery. Unfortunately, the only way Am J Respir Crit Care Med 1999;159(6):1849–61. such data can be obtained is through large pro-  Arroliga AC, Ghamra ZW, Perez Trepichio A, et al. Incidence of ARDS in an adult population of north- spective population-based studies that could be east Ohio. Chest 2002;121(6):1972–6. impractical and diﬃcult to perform. Nonetheless,  Goss CH, Brower RG, Hudson LD, et al. Incidence the authors believe that there is enough informa- of acute lung injury in the United States. Crit Care tion in the medical literature to show that ARDS Med 2003;31(6):1607–11. is a relatively common entity with signiﬁcant  Hughes M, MacKirdy FN, Ross J, et al. Acute respi- morbidity and mortality, and although diﬀerent ratory distress syndrome: an audit of incidence and studies suggest that the outcomes of ARDS may outcome in Scottish intensive care units. Anaesthesia be improving, this entity still has the potential to 2003;58(9):838–45. have a signiﬁcant burden on public health.  Bersten AD, Edibam C, Hunt T, et al. Incidence and mortality of acute lung injury and the acute respira- tory distress syndrome in three Australian States. References Am J Respir Crit Care Med 2002;165(4):443–8.  Rubenfeld GD, Caldwell E, Peabody E, et al. Inci-  Bernard GR, Artigas A, Brigham KL, et al. The dence and outcomes of acute lung injury. N Engl American-European Consensus Conference on J Med 2005;353(16):1685–93. ARDS. Deﬁnitions, mechanisms, relevant out-  Fowler AA, Hamman RF, Good JT, et al. Adult re- comes, and clinical trial coordination. Am J Respir spiratory distress syndrome: risk with common pre- Crit Care Med 1994;149(3 Pt 1):818–24. dispositions. Ann Intern Med 1983;98(5 Pt 1):593–7.  Hudson LD, Milberg JA, Anardi D, et al. Clinical  Esteban A, Fernandez-Segoviano P, Frutos-Vivar F, risks for development of the acute respiratory dis- et al. Comparison of clinical criteria for the acute re- tress syndrome. Am J Respir Crit Care Med 1995; spiratory distress syndrome with autopsy ﬁndings. 151(2 Pt 1):293–301. Ann Intern Med 2004;141(6):440–5.  Ware LB, Matthay MA. The acute respiratory dis-  Katzenstein AL, Bloor CM, Leibow AA. Diﬀuse tress syndrome. N Engl J Med 2000;342(18):1334–49. alveolar damage–the role of oxygen, shock, and  Ashbaugh DG, Bigelow DB, Petty TL, et al. Acute related factors. A review. Am J Pathol 1976;85(1): respiratory distress in adults. Lancet 1967;2(7511): 209–28. 319–23.  Patel SR, Karmpaliotis D, Ayas NT, et al. The role  The Acute Respiratory Distress Syndrome Network. of open-lung biopsy in ARDS. Chest 2004;125(1): Ventilation with lower tidal volumes as compared 197–202. with traditional tidal volumes for acute lung injury  Ferguson ND, Meade MO, Hallett DC, et al. High and the acute respiratory distress syndrome. values of the pulmonary artery wedge pressure in pa- N Engl J Med 2000;342(18):1301–8. tients with acute lung injury and acute respiratory  Amato MB, Barbas CS, Medeiros DM, et al. Eﬀect distress syndrome. Intensive Care Med 2002;28(8): of a protective-ventilation strategy on mortality in 1073–7. the acute respiratory distress syndrome. N Engl  Rubenfeld GD, Caldwell E, Granton J, et al. Inter- J Med 1998;338(6):347–54. observer variability in applying a radiographic deﬁ-  Conference report: mechanisms of acute respiratory nition for ARDS. Chest 1999;116(5):1347–53. failure. Am Rev Respir Dis 1977;115(6):1071–8.  Meade MO, Cook RJ, Guyatt GH, et al. Interob-  Villar J, Slutsky AS. The incidence of the adult respi- server variation in interpreting chest radiographs ratory distress syndrome. Am Rev Respir Dis 1989; for the diagnosis of acute respiratory distress syn- 140(3):814–6. drome. Am J Respir Crit Care Med 2000;161(1):  Webster NR, Cohen AT, Nunn JF. Adult respira- 85–90. tory distress syndromedhow many cases in the  Al-Kharrat T, Zarich S, Amoateng-Adjepong Y, UK? Anaesthesia 1988;43(11):923–6. et al. Analysis of observer variability in measurement  Thomsen GE, Morris AH. Incidence of the adult re- of pulmonary artery occlusion pressures. Am spiratory distress syndrome in the state of Utah. Am J Respir Crit Care Med 1999;160(2):415–20. J Respir Crit Care Med 1995;152(3):965–71.  Rizvi K, Deboisblanc BP, Truwit JD, et al. Eﬀect of  Lewandowski K, Metz J, Deutschmann C, et al. In- airway pressure display on interobserver agreement cidence, severity, and mortality of acute respiratory in the assessment of vascular pressures in patients CLINICAL EPIDEMIOLOGY OF ALI AND ARDS 557 with acute lung injury and acute respiratory distress distress syndrome (ARDS): 1983–1993. JAMA syndrome. Crit Care Med 2005;33(1):98–103 [discus- 1995;273(4):306–9. sion: 243–4].  Jardin F, Fellahi JL, Beauchet A, et al. Improved  O’Quin R, Marini JJ. Pulmonary artery occlusion prognosis of acute respiratory distress syndrome pressure: clinical physiology, measurement, and in- 15 years on. Intensive Care Med 1999;25(9):936–41. terpretation. Am Rev Respir Dis 1983;128(2):  Stapleton RD, Wang BM, Hudson LD, et al. Causes 319–26. and timing of death in patients with ARDS. Chest  Baumann WR, Jung RC, Koss M, et al. Incidence 2005;128(2):525–32. and mortality of adult respiratory distress syn-  Valta P, Uusaro A, Nunes S, et al. Acute respiratory drome: a prospective analysis from a large metropol- distress syndrome: frequency, clinical course, and itan hospital. Crit Care Med 1986;14(1):1–4. costs of care. Crit Care Med 1999;27(11):2367–74.  Hickling KG, Walsh J, Henderson S, et al. Low mor-  Kraﬀt P, Fridrich P, Pernerstorfer T, et al. The tality rate in adult respiratory distress syndrome us- acute respiratory distress syndrome: deﬁnitions, se- ing low-volume, pressure-limited ventilation with verity and clinical outcome. An analysis of 101 clin- permissive hypercapnia: a prospective study. Crit ical investigations. Intensive Care Med 1996;22(6): Care Med 1994;22(10):1568–78. 519–29.  Doyle RL, Szaﬂarski N, Modin GW, et al. Identiﬁ-  Martin GS, Mannino DM, Eaton S, et al. The cation of patients with acute lung injury. Predictors epidemiology of sepsis in the United States from of mortality. Am J Respir Crit Care Med 1995; 1979 through 2000. N Engl J Med 2003;348(16): 152(6 Pt 1):1818–24. 1546–54.  Zilberberg MD, Epstein SK. Acute lung injury in the  McHugh LG, Milberg JA, Whitcomb ME, et al. Re- medical ICU: comorbid conditions, age, etiology, covery of function in survivors of the acute respira- and hospital outcome. Am J Respir Crit Care Med tory distress syndrome. Am J Respir Crit Care 1998;157(4 Pt 1):1159–64. Med 1994;150(1):90–4.  Roupie E, Lepage E, Wysocki M, et al. Prevalence,  Ghio AJ, Elliott CG, Crapo RO, et al. Impairment etiologies and outcome of the acute respiratory dis- after adult respiratory distress syndrome. An evalu- tress syndrome among hypoxemic ventilated pa- ation based on American Thoracic Society recom- tients. SRLF Collaborative Group on Mechanical mendations. Am Rev Respir Dis 1989;139(5): Ventilation. Societe de Reanimation de Langue 1158–62. Francaise. Intensive Care Med 1999;25(9):920–9.  Herridge MS, Cheung AM, Tansey CM, et al. One-  Suntharalingam G, Regan K, Keogh BF, et al. Inﬂu- year outcomes in survivors of the acute respiratory ence of direct and indirect etiology on acute outcome distress syndrome. N Engl J Med 2003;348(8): and 6-month functional recovery in acute respira- 683–93. tory distress syndrome. Crit Care Med 2001;29(3):  Hopkins RO, Weaver LK, Collingridge D, et al. 562–6. Two-year cognitive, emotional, and quality-of-life  Estenssoro E, Dubin A, Laﬀaire E, et al. Incidence, outcomes in acute respiratory distress syndrome. clinical course, and outcome in 217 patients with Am J Respir Crit Care Med 2005;171(4):340–7. acute respiratory distress syndrome. Crit Care Med  Hopkins RO, Weaver LK, Pope D, et al. Neuropsy- 2002;30(11):2450–6. chological sequelae and impaired health status in sur-  Abel SJ, Finney SJ, Brett SJ, et al. Reduced mortal- vivors of severe acute respiratory distress syndrome. ity in association with the acute respiratory distress Am J Respir Crit Care Med 1999;160(1):50–6. syndrome (ARDS). Thorax 1998;53(4):292–4.  Weinert CR, Gross CR, Kangas JR, et al. Health-  Esteban A, Anzueto A, Frutos F, et al. Characteris- related quality of life after acute lung injury. Am J tics and outcomes in adult patients receiving me- Respir Crit Care Med 1997;156(4 Pt 1):1120–8. chanical ventilation: a 28-day international study.  Schelling G, Stoll C, Haller M, et al. Health-related JAMA 2002;287(3):345–55. quality of life and posttraumatic stress disorder in  Milberg JA, Davis DR, Steinberg KP, et al. Im- survivors of the acute respiratory distress syndrome. proved survival of patients with acute respiratory Crit Care Med 1998;26(4):651–9.