ObjectiveTo determine the published incidence of adult respiratory distress syndrome (ARDS) as well as the clinical evidence supporting a causal association between ARDS and its major risk factors.Data SourcesThe National Library of Medicine MEDLINE database and the bibliographies of selected articles.Study SelectionClinical studies were selected from the English literature, if they pertained to either the incidence of ARDS or its association with one or more commonly identified risk factors.Data ExtractionAll relevant studies identified by the search were evaluated for strength of design, and risk factors were scored according to established criteria for the strength of causation.Data SynthesisA total of 83 articles were considered relevant: six on incidence and 77 on risk factors. Only 49% of the 83 articles provided a definition of ARDS; a definition of risk factors was given in 64%, and 23% had no definition for either ARDS or risk factors. The published, population-based incidence of ARDS ranges from 1.5 to 5.3/105population/yr. The strongest clinical evidence supporting a cause-effect relationship was identified for sepsis, aspiration, trauma, and multiple transfusions. The weakest clinical evidence was identified for disseminated intravascular coagulation. The following study types were represented by the 77 articles on risk factors: observational case-series (56%); cohorts (23%); case-controls (12%); nonrandomized clinical trials (5%); and randomized clinical trials (3%). Only a single study reported an odds ratio.ConclusionsThe significant variation in the incidence of ARDS is attributed to differences in the type and strength of study designs, as well as definitions of ARDS. While a substantial body of evidence exists concerning a causal role for ARDS risk factors, such as sepsis, aspiration, and trauma, more than 60% of clinical studies employed weak designs. The lack of reproducible definitions for ARDS or its potential risk factors in 49% of studies raises concerns about the validity of the conclusions of these studies regarding the association between ARDS and the supposed risk factors.(Crit Care Med 1996; 24:687-695)