Antipsychotic use and risk of life-threatening medical events: umbrella review of observational studies

Abstract

Objective: To quantify the risk of hip fracture, thromboembolism, stroke, myocardial infarction, pneumonia and sudden cardiac death associated with exposure to antipsychotics.

Methods: Systematic searches were conducted in Medline, Embase and PsycINFO from inception until 30/07/2018 for systematic reviews of observational studies. AMSTAR-2 was used for the quality assessment of systematic reviews, while the strength of associations was measured using GRADE and quantitative umbrella review criteria (URC).

Results: Sixty-eight observational studies from six systematic reviews were included. The association between antipsychotic exposure and pneumonia was the strongest [URC = class I; GRADE = low quality; odds ratio (OR) = 1.84, 95% confidence interval (CI) = 1.62-2.09; participants = 28 726; age = 76.2 ± 12.3 years], followed by the association with hip fracture (URC = class II; GRADE = low quality; OR = 1.57, 95% CI = 1.42-1.74; participants = 5 288 118; age = 55.4 ± 12.5 years), and thromboembolism (URC = class II; GRADE = very low quality; OR = 1.55, 95% CI = 1.31-1.83; participants = 31 417 175; age = 55.5 ± 3.2 years). The association was weak for stroke (URC = class III; GRADE = very low quality; OR = 1.45, 95% CI = 1.24-1.70; participants = 65 700; age = 68.7 ± 13.8 years), sudden cardiac death (URC = class III; GRADE = very low quality; OR = 2.24, 95% CI = 1.45-3.46; participants = 77 488; age = 52.2 ± 6.2 years) and myocardial infarction (URC = class III; GRADE = very low quality; OR = 2.21, 95% CI = 1.41-3.46; participants = 399 868; age = 74.1 ± 9.3 years).

Conclusion: The most robust results were found for the risk of pneumonia, followed by the risk of hip fracture and thromboembolism. For stroke, sudden cardiac death and myocardial infarction, the strength of association was weak. The observational nature of the primary studies may represent a source of bias.

Keywords: antipsychotics; methodology; pharmacoepidemiology; side effects.