1. Has Mortality from Acute Respiratory Distress Syndrome Decreased over Time? A Systematic Review Jason Phua1,2, Joan R. Badia1,3, Neill K. J. Adhikari1,4, Jan O. Friedrich1,5, Robert A. Fowler1,4, Jeff M. Singh1,6, Damon C. Scales1,4, David R. Stather7, Amanda Li8, Andrew Jones9, David J. Gattas10, David Hallett1, George Tomlinson1, Thomas E. Stewart1,6, and Niall D. Ferguson1,6 1Interdepartmental Division of Critical Care, University of Toronto, Ontario, Canada; 2Division of Respiratory and Critical Care Medicine, Singapore; 3Hospital Clinic of Barcelona, Spain; 4Department of Critical Care Medicine, Sunnybrook Health Science Centre, Toronto, Ontario, Canada; 5Critical Care and Medicine Departments, and The Keenan Research Centre, Canada; 6Department of Medicine, Canada; 7University of Calgary, Alberta, Canada; 8University of Ottawa, Ontario, Canada; 9Department of Critical Care, London, United Kingdom; and 10Royal Prince Alfred Hospital, Sydney, Australia Am J Respir Crit Care Med Vol 179. pp 220–227, 2009 1

2. Introduction Definition of ALI/ARDS –By an American-European Consensus Conference (AECC) –1. The acute onset of arterial hypoxemia (PaO 2 /fraction of inspired oxygen [FiO 2 ] ratio < 300 in ALI /<200 in ARDS) –2. A pulmonary artery wedge pressure < 18 mm Hg or no clinical evidence of left atrial hypertension –3. Bilateral infiltrates consistent with pulmonary edema on chest radiograph Mortality from ALI/ARDS –Initial : 45%~50% –Overall mortality in recent studies : 30% 2

3. Introduction Decreased mort ality rates for pat ients with ALI/A RDS. No decline in ov erall ALI/ARDS mortality 3 ?

4. Introduction 4 Results Seventy-two studies were included in the analysis. There was a wide variation in mortality rates among the studies, from 15-72%. The overall pooled mortality rate for all studies was 43% (95% CI 40-46%). Meta-regression analysis suggested a significant decrease in overall mortality rates of approximately 1.1% per year over the period analyzed (1994-2006). The mortality reduction was also observed for hospital but not for ICU or 28-day mortality. Zambon M, Vincent JL. Chest 2008;133:1120-1127 Zambon M, Vincent JL. Chest 2008;133:1120-1127 Variation in overall mortality rates over time in the ALI/ARDS studies

5. Introduction The limitations of the existing literatures. –Data from RCTs Underestimation of ‘‘real-world’’ mortality –Performed in specialized centers –Selected criteria –Designed to improve safety –Maximize potential treatment effects –Data from observational studies From single centers Including many patients with traumatic injuries, who have a better prognosis –Data from the meta-analysis Uncomprehensible search strategy Some studies conducted before 1994 were labeled as post- 1994 studies. 5

6. Proposal To evaluate…. –Whether mortality has changed over time. –Which patient and study factors are independently associated with mortality. 6

7. Search Strategy –MEDLINE, EMBASE, CINAHL, and Cochrane CENTRAL –A sensitive strategy without language restrictions –Reviewing personal files and references of included studies Study Selection –Full-text copies : distributed to six pairs of independent reviewers –Prospective observational studies and RCTs –Enrolling 50 or more adults with ALI/ARDS(using any definition) Data extraction –The primary outcome : ICU mortality, if unavailable 28/30 days – hospital – 45days – 60days – 90days mortality. 7 Methods

8. Results 8 Study selection

9. Results 9 Study characteristics

10. Results 10 Mortality in Observational studies

11. Results 11 Mortality in Observational studies Mortality in Observational studies : 48.2%, heterogeneity(+)

12. Results 12 Mortality in RCTs

13. Results 13 Mortality in RCTs Mortality in RCTs : 37.5%, lower than observational studies

14. Results 14 Overview of Mortality Figure. The annual pooled weighted mortality by year of study conduct

15. Results 15 Meta-regression Analysis (1)

16. Results 16 Meta-regression Analysis (2) None of variables was significantly assosicated with mortality

17. Results 17 Meta-regression Analysis (3)

18. Discussion The mortality due to ALI/ARDS –44.0% for observational studies and 36.2% for RCTs The current benchmark for mortality of ALI/ARDS in clinical practice should be 25~30% ?? –Decreased mortality over time in observational studies conducted from 1984 to 1993 –Higher mortality in observational studies than in RCTs Due to the characteristics of RCTs –Specialized centers, Ventilation protocols shown to improve outcomes, exclusion of patients with poor prognosis –Age and study variability : influence to mortality 18

19. Discussion In Observational studies of ARDS mortality –A decrease in mortality over time in five studies. –Limitations In single center Included many trauma patients –Multicenter observational studies ARDS mortality : From 39% to as high as 50 to 60%. In Meta-analysis about mortality trends in ARDS over time 1. An older meta-analysis by Krafft, et al. –No relation between 1967 and 1994. –Not directly comparable with this journal Before 1984, Included many retrospective studies with very small sample sizes 2. Zambon M, Vincent JL. Mortality rates for patients with acute lung injury/ARDS have decreased over time. Chest 2008;133:1120–1127 19

20. Discussion 20 Abstract Background Over the last decade, several studies have suggested that survival rates for patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) may have improved. We performed a systematic analysis of the ALI/ARDS literature to document possible trends in mortality between 1994 and 2006. Methods We used the Medline database to select studies with the keywords “acute lung injury”, “acute respiratory distress syndrome”, “acute respiratory failure”, “mechanical ventilation”. All studies that reported mortality for patients with ALI/ARDS defined according to the criteria of the American European Consensus Conference were selected. We excluded studies with less than 30 patients and studies limited to specific subgroups of ARDS patients, such as sepsis, trauma, burns or transfusion-related ARDS. Results Seventy-two studies were included in the analysis. There was a wide variation in mortality rates among the studies, from 15-72%. The overall pooled mortality rate for all studies was 43% (95% CI 40-46%). Meta-regression analysis suggested a significant decrease in overall mortality rates of approximately 1.1% per year over the period analyzed (1994-2006). The mortality reduction was also observed for hospital but not for ICU or 28-day mortality. Conclusions In this literature review, the data are consistent with a reduction in mortality rates in general populations of patients with ALI/ARDS over the last 10 years. Abstract Background Over the last decade, several studies have suggested that survival rates for patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) may have improved. We performed a systematic analysis of the ALI/ARDS literature to document possible trends in mortality between 1994 and 2006. Methods We used the Medline database to select studies with the keywords “acute lung injury”, “acute respiratory distress syndrome”, “acute respiratory failure”, “mechanical ventilation”. All studies that reported mortality for patients with ALI/ARDS defined according to the criteria of the American European Consensus Conference were selected. We excluded studies with less than 30 patients and studies limited to specific subgroups of ARDS patients, such as sepsis, trauma, burns or transfusion-related ARDS. Results Seventy-two studies were included in the analysis. There was a wide variation in mortality rates among the studies, from 15-72%. The overall pooled mortality rate for all studies was 43% (95% CI 40-46%). Meta-regression analysis suggested a significant decrease in overall mortality rates of approximately 1.1% per year over the period analyzed (1994-2006). The mortality reduction was also observed for hospital but not for ICU or 28-day mortality. Conclusions In this literature review, the data are consistent with a reduction in mortality rates in general populations of patients with ALI/ARDS over the last 10 years. Mortality rates for patients with ALI/ARDS have decreased over time. Chest 2008;133:1120-1127

21. Discussion Methodologic differences in 2 studies –Our search strategy Reviewers to work in duplicate More search terms and multiple databases Allowed only prospective studies Excluded small studies with fewer than 50 patients with ALI/ARDS, and was not restricted to the English language. –Other reserch Considered only the last cohort in studies with multiple historical cohorts and only the control group in RCTs. Included several reports, to contain overlapping patient with low mortality : Some high-mortality studies before 1994 + Low-mortality post-1994 RCTs 21

22. Discussion Why mortality for patients with ARDS has not improved? –The effective therapeutic interventions for ARDS continues to pale in comparison to the failed interventions. –Because of ALI/ARDS as a syndrome with multiple pathophysiologic mechanisms –“The optimization of general supportive care” Limitations in this review article –1. The differences in primary outcome –2. The differences in case-mix and/or availability of ICU beds in various countries. –3. Using the median year of conduct for each study –4. Exclusion of unpublished data and abstract only reports. –5. Mixture of ARDS and ALI patients, difficulty in assessing severity of disease 22

23. Mortality due to ARDS has remained relatively unchanged since 1994, coincident with the publication of the current syndrome definition. Higher mortality –in observational study design and increased patient age Expection of a baseline mortality risk from ARDS –For observational studies : 40 to 45% –For RCTs : 35 to 40% Most importantly, our results highlight the need for future effective therapeutic interventions for this highly lethal syndrome. 23 Conclusion