Background A better understanding of the epidemiological impact of suicidal ideation after stroke is required to identify subjects needing personalised interventions.

Objective The aim of this meta-analysis was to estimate rates and correlates of suicidal ideation among stroke survivors.

Methods We searched via Ovid, Medline, Embase and PsycInfo from database inception until August 2016. Predefined outcomes were (1) rates of suicidal ideation based on random-effects pooled proportion and (2) relevant sociodemographic and clinical correlates, using random-effects odds ratio (OR) or standardised mean difference (SMD) for categorical and continuous variables, respectively.

Results Fifteen studies and 13 independent samples, accounting for 10 400 subjects, were included in meta-analyses. The pooled proportion of suicidal ideation among stroke survivors was 11.8% (7.4% to 16.2%), with high heterogeneity across studies (I²=97.3%). Current (OR=11.50; p<0.001) and past (OR=6.96; p<0.001) depression, recurrent stroke (OR=1.77; p<0.001), disability (SMD=0.58; p=0.01) and cognitive impairment (SMD=−0.22; p=0.03) were all associated with suicidal ideation. Moreover, suicidal ideation was less likely in stroke survivors who were married (OR=0.63; p<0.001), employed (OR=0.57; p=0.02) and had higher education levels (OR=0.55; p=0.002).

Conclusion Despite some limitations, this meta-analysis shows that about one out of eight stroke survivors has suicidal ideation. Thus, there is enough evidence to support the use of routine screening and early interventions to prevent and treat suicidal ideation after stroke, especially among subjects carrying specific correlates.