Posttraumatic stress disorder in critical illness survivors: a metaanalysis

Abstract

Objective: To conduct a systematic review and metaanalysis of the prevalence, risk factors, and prevention/treatment strategies for posttraumatic stress disorder symptoms in critical illness survivors.

Data sources: PubMed, Embase, CINAHL, PsycINFO, and Cochrane Library from inception through March 5, 2014.

Study selection: Eligible studies met the following criteria: 1) adult general/nonspecialty ICU, 2) validated posttraumatic stress disorder instrument greater than or equal to 1 month post-ICU, and 3) sample size greater than or equal to 10 patients.

Data extraction: Duplicate independent review and data abstraction from all eligible titles/abstracts/full-text articles.

Data synthesis: The search identified 2,817 titles/abstracts, with 40 eligible articles on 36 unique cohorts (n = 4,260 patients). The Impact of Event Scale was the most common posttraumatic stress disorder instrument. Between 1 and 6 months post-ICU (six studies; n = 456), the pooled mean (95% CI) Impact of Event Scale score was 20 (17-24), and the pooled prevalences of clinically important posttraumatic stress disorder symptoms (95% CI) were 25% (18-34%) and 44% (36-52%) using Impact of Event Scale thresholds greater than or equal to 35 and greater than or equal to 20, respectively. Between 7 and 12 months post-ICU (five studies; n = 698), the pooled mean Impact of Event Scale score was 17 (9-24), and pooled prevalences of posttraumatic stress disorder symptoms were 17% (10-26%) and 34% (22-50%), respectively. ICU risk factors for posttraumatic stress disorder symptoms included benzodiazepine administration and post-ICU memories of frightening ICU experiences. Posttraumatic stress disorder symptoms were associated with worse quality of life. In European-based studies: 1) an ICU diary was associated with a significant reduction in posttraumatic stress disorder symptoms, 2) a self-help rehabilitation manual was associated with significant posttraumatic stress disorder symptom reduction at 2 months, but not 6 months; and 3) a nurse-led ICU follow-up clinic did not reduce posttraumatic stress disorder symptoms.

Conclusions: Clinically important posttraumatic stress disorder symptoms occurred in one fifth of critical illness survivors at 1-year follow-up, with higher prevalence in those who had comorbid psychopathology, received benzodiazepines, and had early memories of frightening ICU experiences. In European studies, ICU diaries reduced posttraumatic stress disorder symptoms.