Association between sex and survival after out‐of‐hospital cardiac arrest: A systematic review and meta‐analysis

Abstract The current literature on sex differences in 30‐day survival following out‐of‐hospital cardiac arrest (OHCA) is conflicting, with 3 recent systematic reviews reporting opposing results. To address these contradictions, this systematic literature review and meta‐analysis aimed to synthesize the literature on sex differences in survival after OHCA by including only population‐based studies and through separate meta‐analyses of crude and adjusted effect estimates. MEDLINE and Embase databases were systematically searched from inception to March 23, 2022 to identify observational studies reporting sex‐specific 30‐day survival or survival until hospital discharge after OHCA. Two meta‐analyses were conducted. The first included unadjusted effect estimates of the association between sex and survival (comparing males vs females), whereas the second included effect estimates adjusted for possible mediating and/or confounding variables. The PROSPERO registration number was CRD42021237887, and the search identified 6712 articles. After the screening, 164 potentially relevant articles were identified, of which 26 were included. The pooled estimate for crude effect estimates (odds ratio [OR], 1.42; 95% confidence interval [CI], 1.22–1.66) indicated that males have a higher chance of survival after OHCA than females. However, the pooled estimate for adjusted effect estimates shows no difference in survival after OHCA between males and females (OR, 0.93; 95% CI, 0.84–1.03). Both meta‐analyses involved high statistical heterogeneity between studies: crude pooled estimate I2 = 95.7%, adjusted pooled estimate I2 = 91.3%. There does not appear to be a difference in survival between males and females when effect estimates are adjusted for possible confounding and/or mediating variables in non‐selected populations.


Background
It is well-known that males are more likely to experience out-ofhospital cardiac arrest (OHCA) than females. 1,2 For example, the age-standardized incidence of OHCA in the Netherlands between 2009 and 2015 was 87.3 per 100,000 person-years for males versus 30.9 per 100,000 person-years for females. 3 This difference is most likely due to males having a higher risk of cardiovascular disease (CVD) than females. 4 However, in contrast to the well-known sex differences in incidence of OHCA, the existing evidence on sex differences in survival after OHCA is inconclusive. 5 8 which subsequently decreases the chance of 30-day survival after OHCA. Females are less likely to have an initial shockable rhythm than males (24% vs 42%, respectively). 8,10 Presenting with an initial shockable rhythm increases the chance of survival after OHCA dramatically. 10 Three recent systematic reviews summarizing sex differences in survival after OHCA have reached contradictory conclusions, reporting lower survival in females, 11 higher survival in females 12 and no difference in survival. 13 Feng et al 12  where patients had an initial shockable rhythm, received targeted temperature management (TTM), and had a witnessed arrest. 13 Malik et al 13 included 30 studies, all of which reported adjusted effect estimates. However, the heterogeneity of included study populations limits the interpretation of summary estimates from these analyses.

Importance
The differing inclusion criteria and heterogeneity in terms of pooling crude and/or adjusted effect estimates may account for the contradictory findings from these previous systematic reviews and meta-analyses. However, it is important to examine why different conclusions have been reached to direct future research to focus on specific research areas and subsequently inform stakeholders to target these areas to improve any sex-related survival gap after OHCA.

Goals of this investigation
Therefore, the aim of this systematic review and meta-analysis is to synthesize the current evidence on sex differences in survival after OHCA in non-selected populations only, with separate pooling of crude and adjusted effect estimates, to reconcile the conflicting conclusions from previous reviews on this topic.

Study design and registration
This systematic review adheres to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. 14 The protocol of this review is registered with PROSPERO (CRD42021237887).

Search strategy
We

Data analysis
The primary analysis consisted of 2 separate meta-analyses looking at the association between sex and survival after OHCA. between-study variance. 21 The I 2 statistic measures whether variation is likely due to chance or more likely due to study heterogeneity, 20 I 2 0%-25%, I 2 26%-50%, and I 2 ≥50% are respectively considered low, moderate, and high heterogeneity in this review. 22 Forest plots were created to visualize the results of the metaanalyses. A prediction interval is reported in the forest plots, this is an estimate of the between-study variance in the random-effects metaanalyses. 23 Subgroup analyses were planned a priori to address possible clinical heterogeneity 24 and performed including only studies that adjusted for initial cardiac rhythm and/or bystander cardiopulmonary resuscitation (bCPR) (these are important mediating/confounding variables 10,25,26 ).
To explore possible sources of high statistical heterogeneity, sensitivity analyses were performed. [27][28][29][30][31] For the sensitivity analyses Baujat plots and forest plots were created (using the "leave-one-out method") to assess that single studies have a disproportionately large effect on either heterogeneity or the pooled effect size. In further sensitivity analyses basic outliers, determined by statistical analysis, were removed. Publication bias was assessed using different methods, namely the Rücker's Limit Meta-Analysis Method, 29

RESULTS
The initial search yielded 6712 articles and from these, 1351 duplicates were removed. After screening the studies by abstract and title, 164 studies remained. After screening full texts, 26 studies were eligible for inclusion in this systematic review ( Figure 1 shows the selection process). Studies were mostly excluded based on having a selected population-based setting.

Description of included studies
Baseline characteristics of included studies are summarized in indicating that studies were of moderate to high quality.

Meta-analyses
The pooled analysis of 10 studies reporting crude odds ratios (ORs) for the association between sex and survival showed that male sex was significantly associated with increased odds of 30-day survival after OHCA (OR, 1.42; 95% confidence interval (CI), 1.22-1.66) (Figure 2;

Sensitivity and subgroup analyses
The pooled effect estimates did not change significantly when conducting the "leave-one-out method." Heterogeneity was somewhat lower but still moderate to high (meta-analysis crude effect estimates, I 2 = 79.0%; meta-analysis adjusted effect estimates, I 2 = 55.3%). The results of the sensitivity and subgroup analyses were also in line with the results of the main meta-analyses ( Table 4). Results of the publication bias analyses indicate that publication bias is unlikely to have had a major impact on the results of the meta-analyses (Table 5).

F I G U R E 2
Forest plot meta-analysis crude odds ratios.

TA B L E 4
Results of meta-analyses of crude-and adjusted effect estimates for the association between sex and OHCA survival.

Limitations
The high statistical heterogeneity between studies included in the 2 meta-analyses is a limitation of this systematic review. High statistical heterogeneity in systematic reviews is a common issue, especially when using observational studies in a meta-analysis. 55 Important sex-based differences in cardiac arrest characteristics that influence survival after OHCA exist. For example, the incidence of small vessel disease is higher in females than in males, 57,58 whereas the incidence of myocardial infarction is higher in males than in females. 59,60 These differences are important as the etiology of OHCA influences survival after OHCA. [61][62][63] Furthermore, females are less likely to present with an initial shockable rhythm 8,53 and less likely to receive bCPR than males. 8,64 The difference in initial shockable rhythm might partially be explained by the fact that females delay calling emergency medical services (EMS) when experiencing symptoms of acute myocardial infarction because the symptoms differ between males and females, although the magnitude of this impact is unknown. 65,66 On average females are likely to be older than males when experiencing OHCA. 67 All these differences have been reproduced and reported multiple times by large studies like OPALS 10 and CARES. 9 Beyond the prehospital resuscitation phase, differences have been shown in the way sex neutral protocols for the management of acute coronary syndromes are applied to males versus females, with males more likely to receive treatment. 68,69 Some of these variables might explain the male survival benefit found in crude effect estimates. When looking at the effect estimates adjusted for these characteristics, the male survival benefit seems to have disappeared and therefore it seems likely that the crude survival difference between the sexes is due to differences in OHCA characteristics. Most of these variables are considered to be mediators and not confounders, therefore they should be handled differently than confounders and not be adjusted for when conducting regression analysis.
A key rationale for the current systematic review was to explore reasons for the conflicting results of 3 fairly recently published systematic literature reviews and provide clarity on this topic for research user groups. [11][12][13] Our results indicate that the contrasting finding that survival is lower in females than males from Lei et al 11   Writing-review and editing. Nynke Halbesma: Conceptualization, methodology, writing-review and editing, and supervision.