Sex-Specific Differences in Survival After Out-of-Hospital Cardiac Arrest

A Nationwide, Population-Based Observational Study

Yoshikazu Goto; Akira Funada; Tetsuo Maeda; Hirofumi Okada; Yumiko Goto


Crit Care. 2019;23(263) 

In This Article


Hierarchical propensity score matching did not indicate any differences in the 1-month survival and neurologically intact survival between men and women in any of the age groups analyzed in this nationwide, population-based observational study. This is the first large cohort study to clearly demonstrate the absence of differences in clinically meaningful survival between men and women after OHCA.

Hormonal effect on survival outcomes after cardiac arrest remains controversial: estrogen mediates hormonal responses to ischemia–reperfusion injury in women of childbearing age,[33–37] whereas androgen mediates negative effects on the cardiovascular system in young men.[38–40] Previous studies have revealed that apart from reproductive roles, sex hormones possess cardioprotective and neuroprotective functions.[41,42] Estrogen has previously been reported to have cardioprotective and neuroprotective effects after cardiac arrest.[35–37] Although the cardioprotective effects of estrogen are widely appreciated, little is known about the effects of progesterone.[42] Testosterone is associated with various adverse events in young men, including coronary plaque formation, platelet aggregation, inflammation, heart failure, and coronary artery disease.[38–40] Increased testosterone levels in men have been associated with decreased sudden cardiac arrest events; further, increased estradiol levels in both sexes have been associated with increased sudden cardiac arrest events.[43] Nevertheless, the impact of testosterone on the cardiovascular system remains controversial.[42] However, in the present study, there were no significant sex-specific differences in the 1-month survival outcomes in each age group after adjustment for prehospital confounders. These results possibly reflect estrogen-related effects on women and testosterone-related effects on men. Regarding the neuroprotective effects of sex hormones, estrogen reportedly slows down the progression of brain injury and diminishes the extent of cell death by suppressing apoptotic pathways.[44] In agreement with that report, another study demonstrated that brain edema after intracerebral hemorrhage-induced injury was reportedly less severe in female than in male rats.[45] The neuroprotective properties of progesterone following cerebral ischemia, which are mediated by a reduction in edema, are most likely related to the suppression of interleukin-1β production.[46] Testosterone also possesses neuroprotective properties that result from the activation of androgen pathways as well as antioxidant and anti-apoptotic activities. The neuroprotective effects of sex hormones may thus partially explain the lack of significant differences in the neurologically intact survival outcomes between women and men in this study.[41] Other factors that may have contributed to the sex-specific differences in survival outcomes after OHCA include sex-specific anatomical differences of chest wall compliance,[47] coronary risk factors,[48] genomic variation,[49] and care process.[50] The statistical analysis did not adjust for those factors as confounding variables because of the lack of data in the registry.

A recent analysis by Benjamin et al. proposed changing the significance level from P < 0.05 to P < 0.005 to avoid a high rate of false-positive results, even in the absence of other reporting problems.[32] Thus, the lack of reproducibility of sex differences and outcomes after OHCA in previous studies might be related to the statistical analysis. In this study, P < 0.005 was considered statistically significant. Our findings that female patients with OHCA were older, were more likely to present with nonshockable rhythms, had more unwitnessed arrests, and received bystander CPR more frequently than male patients are consistent with the findings of some[6,12,14,20] but not all previous studies.[8–11,16] Our findings showed that women of childbearing age had no survival advantage over age-matched men after OHCA, which differs from previously reported results in Japan.[6,7,14] For example, Kitamura et al. reported that from 1998 to 2007 in Osaka, female patients aged > 13 years had higher 1-month survival rates and that those aged 13–49 years had higher 1-month neurologically intact survival rates after OHCA compared with male patients of the same age.[6] Using the 2005–2007 Utstein data, Akahane et al. reported that female patients aged 30–79 years with OHCA with an initial shockable rhythm had improved 1-month survival and that those aged 40–59 years had better 1-month neurologically intact survival than male patients of the same age.[7] Taken together, these findings indicate that the survival rates were not solely attributable to the effects of estrogen. In addition, we used hierarchical propensity score matching analysis within each age group to adjust for confounding factors; this was not followed by previous studies.[6,7] Differences in the results reported in the above studies may also reflect improvements in the management OHCA after the 2010 and 2015 updates of the international CPR guidelines. Previous studies from Korea, Australia, and New Zealand reported no sex-specific differences in survival outcomes after OHCA, which is consistent with our results.[16,17,51] Importantly, our data further showed the relationship between sex and neurologically intact survival after cardiac arrest in which sex was not associated with clinically meaningful survival.

The present study has some limitations. First, the study analyzed data collected from a large national population by standard procedures; however, because of the retrospective observational design, we could not exclude uncontrolled confounders. For example, patient estrogen, progesterone, and testosterone levels were not available. Sex difference was a surrogate for both sex-related (sex hormone levels, autonomic modulation, and electrophysiological properties) and gender-specific (socioeconomic, environmental, educational, and community) factors that could not be adjusted for. A prospective study of patients with OHCA including sex hormone assays would be needed to clarify the relationship between sex hormone levels and outcomes. The present study also lacked data on pre-existing comorbidities, the location of arrest, the quality of bystander- and EMS-initiated CPR, and the in-hospital treatments. Even in propensity score matching analysis, we cannot exclude numerous unknown confounding factors that may mislead the sex-specific differences in the outcomes after OHCA. Other limitations are common to epidemiological studies, including ascertainment bias and lack of data integrity and validity. The relevance of our results to other communities with different emergency care systems and protocols is not known; similar studies in other countries would help validate our results.