Sex‐dependent disparities in critical illness: methodological implications for critical care research

Despite persistent efforts by government and research funding agencies to include sex and gender in human health-related research (Institute of Medicine, 2001) and acknowledgement that sex and gender are major determinants of many health-related outcomes (Vlassof, 2007), exploration of these issues in critical illness remains sparse. Historically, sex differences referred to biological dissimilarities between males and females, whereas, gender differences suggested the effect of psychosocially conditioned factors and sociopolitical environments (Holdcroft, 2007). More recently, epigenetic studies have clearly demonstrated interactions between social determinants (like gender) and past experiences and cellular responses to stress (Miller et al., 2011; Stankiewicz et al., 2013; Babenko et al., 2014), suggesting a blurring of these independent definitions. We will use the more inclusive term gender in this review, unless summarizing what authors have stated are sex-specific physiological processes. Although the gender gap in life expectancy, favouring women by 5 years on average (Rochelle et al., 2014), is still unexplained, women exhibit specific health-related susceptibilities (e.g. autoimmune, mood and anxiety disorders) compared to men (Verma et al., 2011). Moreover, in such disease states as coronary artery disease, health-related outcomes are less favourable in women. While some differences are likely attributed to differences in treatments and insidious gender bias (Alspach, 2012), in some conditions women may still have worse outcomes, despite identical care (Anderson and Pepine, 2007). In spite of a long-held belief that women have a better chance to survive critical illness (Kristensen et al., 2014), a closer look at the evidence reveals a complex interactive picture. This editorial aims to briefly synthesize current evidence on the influence of sex and gender on adult critically ill patients' outcomes and to discuss specific methodological implications for critical care research. Methodological limitations account for equivocal data and challenges to make cross-study comparisons (Guidet and Maury, 2013). Most studies involving intensive care unit (ICU) patients do not consider gender in their sampling strategy, often resulting in under-representation of women, and they may or may not address gender in subgroup analyses. Moreover, when gender is addressed, the degree to which it may modify the observed associations is not explored. It was only recently that awareness on the importance of gender–age interactions was raised in critical care research literature. Another significant limitation is that sex and gender are often used interchangeably and data that can examine gender differences are not collected. Most studies addressing gender-specific outcomes in general ICU populations are retrospective, usually involving large pre-existing databases with limited control over confounding variables. In general, crude mortality rates are higher in women, but results vary after adjustment for confounders. Although in a large retrospective study in the USA (Mahmood et al., 2012), adjusted odds ratios (ORs) of death were lower in younger women compared to men, and no significant differences were observed for older women [≥50 years of age (yoa)], in a Canadian study, older women (≥50 yoa) exhibited greater adjusted OR for death (Fowler et al., 2007). Likewise, in a smaller study in Belgium, a higher mortality than men was observed among older women (Romo et al., 2004). This gender gap was present only in the first days after admission and decreased over time. Being male may be a risk factor at least for the development of sepsis and multiple organ dysfunction syndrome (MODS) (Nachtigall et al., 2011; Sakr et al., 2013; Angele et al., 2014); however, effects on mortality are not clear. In a large retrospective trauma study in Germany, OR for MODS and sepsis were higher in men (ORs were 1.55 and 1.19 higher in men for sepsis and MODS, respectively), predominantly in patients of reproductive age (Trentzsch et al., 2014a). Despite these differences in incidence, mortality rates did not differ. However, in a smaller American study the odds for lung dysfunction in women were higher (OR: 1.6) (Heffernan et al., 2011). Despite observations of lower incidence of sepsis in females, effects of gender on sepsis/MODS-related mortality and outcomes vary. Overall, sepsis mortality in the general ICU population appears to be higher in women (Nachtigall et al., 2011; Sakr et al., 2013). Nonetheless, when older women are considered (≥50 yoa), the effect of gender may be reversed with women exhibiting lower adjusted OR for death (Adrie et al., 2007). Therefore, varying results on gender-related differences in sepsis mortality may be attributed to varying but generally low percentages (<10%) of younger (<50 yoa) women in the whole ICU population. It is likely that observed discrepant data across studies arise from different approaches to controlling for severity of injury and other related variables, including differences in the most prevalent types of injuries between the two genders (Schoeneberg et al., 2013). For example, in two retrospective studies, when controlling for severity of injury, mortality appears to be higher in female trauma victims, despite lower risk for sepsis and MODS (Schoeneberg et al., 2013; Fröhlich et al., 2014). On the contrary, in a prospective cohort (Guidry et al., 2014), and a matched-pair analysis of ICU patients (Trentzsch et al., 2014a), female trauma victims exhibited lower OR for death after adjustment for other variables. Importantly, critical illness trajectories after trauma may differ in women compared to men. Although they may tolerate shock better (Trentzsch et al., 2014a), women appear to be more vulnerable in the first days after trauma. When non-survivors are analysed separately, men appear to survive approximately twice as long as women (Schoeneberg et al., 2013). In burn trauma, the effect of gender may be accentuated with male and female survival curves diverging early after injury and women exhibiting higher mortality (Summers et al., 2014). The most prominent theory attempting to interpret the impact of sex on critical illness prognosis focuses on the role of gonadal steroid hormones; however, clinical evidence challenges this hypothesis. In animal models, estradiol may protect against sepsis by favouring anti-inflammatory responses and suppressing pro-inflammatory responses (Chen et al., 2014). Nonetheless, in an experimental human study with administration of endotoxin, females exhibited increased pro-inflammatory response compared to males (van Eijk et al., 2007). Indeed, although many animal studies suggest that female hormonal phenotype favours better outcomes (Angele et al., 2006), the matter remains controversial since low testosterone/high estradiol profile has been associated with organ dysfunction (Dossett et al., 2008; Heffernan et al., 2011). Hence, in order to draw useful practice implications from experimental and clinical studies, a crucial issue to resolve is whether female steroid hormones are protective, or, conversely, male sex hormones may accentuate the pathophysiology of sepsis, or whether both scenarios are simultaneously in effect. Moreover, an alternative, most possibly concomitant, mechanism of sex dimorphism in sepsis is that female sex-related ‘protection’ is linked to genetic traits inherited on the X chromosome (Torres et al., 2005; Spolarics, 2014). In traumatic injury and hemorrhagic shock, pathophysiologic responses may be differentiated by sex as well. Experimental evidence suggests that sex hormones may shape host responses to trauma, with estrogens producing salutary effects and male steroids contributing to suppression of cardiac function and immunity (Choudhry et al., 2006). These effects may be related to the presence, on immune cells and diverse cell types, of oestrogen receptors, signalling through which may down-regulate pro-inflammatory cytokine production (Sperry and Minei, 2008). Critically ill patients' survival and outcomes largely depend on: (a) neuro-endocrine responses to stress and metabolism, (b) immune responses and inflammation, and (c) endothelial responses which are key in regulating coagulation, immune cell activation and intravascular volume. Research evidence suggests the existence of dimorphism in all three processes. Overall, neural activation in stress appears to be sex-specific (Wang et al., 2007). Men and women exhibit differential hypothalamic–pituitary–adrenal (HPA) axis responses, possibly mediated by gonadal steroids (Verma et al., 2011), as well by differences in the oxytocin and arginine vasopressin systems (Seng, 2013; Bisagno and Cadet, 2014; Buisman-Pijlman et al., 2014; Steinman et al., 2014). Presumably, HPA responses may have a role in differential regulation of cell-mediated immunity in critical illness (Butts and Sternberg, 2008), and possibly in the higher prevalence of depressive symptoms during the post-ICU period in women (Schandl et al., 2012). The role of past traumatic experiences has not been considered in critical care studies, yet there is strong evidence that HPA axis parameters can be persistently altered to by such experiences (Babenko et al., 2014). Thus, it is challenging to establish ‘baseline’ HPA axis functioning prior to the critical illness event. With regard to immune responses, not only sex steroids influence the responsiveness and activity of immune cells, but also genetic differences attributed to the presence of the X chromosome, as well sex-specific regulation of autosomes may modulate immune functions (Oertelt-Prigione, 2012). Overall, women appear to have more vigorous cellular and humoral immunity and different cytokine profile (Darnall and Suarez, 2009), but this is still an open area of study. Endothelial dysfunction is central to the development of sepsis/MODS (Paulus et al., 2011). Endothelial cells bear oestrogen receptors, and estradiol regulates endothelial nitric oxide synthase (eNOS) and endothelium-derived hyperpolarizing factor (EDHF) (Chow et al., 2010), which have a major role in vasodilation and coagulation activation (Kublickiene and Luksha, 2008). Evidence-based and patient-centred health care aims to fine-tune treatments based on relative efficacy and safety of treatment approaches. To this end, randomized controlled trials, as well as large-scale correlational and case–control studies provide most of the evidence for evidence-based clinical guidelines. However, the great majority of guidelines for the critically ill do not contain gender-specific recommendations. Moreover, gender does not feature in many severity scoring systems, which may also be applied as mortality predictive tools (e.g. SAPS II, SOFA). Additionally, a recent review has revealed a marked male dominance in clinical trials in critical care (Kristensen et al., 2014). The CONSORT guidelines for reporting parallel group randomized trials (Moher et al., 2012; Lee et al., 2014) briefly mention sex as a potential stratification or sub-group analysis factor; however, they do not provide any specific recommendation for incorporating sex and gender in research design and analysis strategy. The effect of gender on critical illness is complex, and the underlying mechanisms are not clearly understood. From both a practice and research point of view, it is important to acknowledge that even if gender differences are not currently evident, they may still exist. It may be useful to never assume that male–female responses are similar, and to also keep in mind that women are more vulnerable in the first days of critical illness. In fact, this short review of evidence shows that there may be two very distinct paradigms of critical illness – one male, one female – both largely unexplored due to the universal confounding effect of gender in the current status of research. Although a bit exaggerated for the sake of argument, it may be possible that cross-gender transfer of data in critical illness is even more challenging than cross-species extrapolation of research results. This is not to say that men and women are two different species, but rather that they are evolutionary counterparts. Critical illness is the ultimate test of survival. It is only natural that the very same gender-dependent disparities that secured survival of our species for over 50 000 years become prominent when we are critically ill. It is therefore time to map the twists and turns of female and male critical illness pathways, both in the ICU and post ICU discharge. This will bring unprecedented insight into the nature of critical illness itself and the constituents of survival.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.