Dehydroepiandrosterone sulfate and dehydroepiandrosterone sulfate/cortisol ratio in cirrhotic patients with septic shock: another sign of hepatoadrenal syndrome?

Septic shock is accompanied by activation of the hypothalamic–pituitary–adrenal (HPA) axis, which is highlighted by increased serum corticotropin and glucocorticoid [1‐3]. Parallel to glucocorticoid secretion, HPA activation leads to the release of dehydroepiandrosterone (DHEA) and its sulfate (DHEAS). Both of these are adrenal androgens and practically all is secreted by the adrenal glands [4]. The serum concentration of DHEAS is 300–500 times higher than that of DHEA and can be considered a circulating reservoir [4]. In contrast to serum cortisol or DHEA concentrations, serum DHEAS levels do not exhibit a diurnal variation, as a consequence of a longer half-life [4]. Therefore, serum DHEAS levels have been proposed to serve as a potential biomarker of adrenal function in different clinical settings [5, 6]. In fact, investigators have combined adrenocorticotropic hormone (ACTH) stimulation tests or insulin tolerance tests with measurements of DHEAS to define normality of the HPA axis [7, 8]. Importantly, similar to that of glucocorticoid, the secretion of adrenal androgen is almost exclusively under the trophic effect of ACTH [4]. Normally, adrenal androgen is secreted synchronously with glucocorticoid from the adrenal cortex [9]. Although the specific physiological function of DHEA is still unclear, it has been shown that DHEA modulates the function of the immune system. DHEA plays an important role in the interaction between the endocrine and immune systems [10, 11]. The immune-modulatory properties of DHEA include induction of T-cell activation and interleukin-2 production [12, 13], and enhancing cytotoxic functions of monocytes [14]. Because of these immune-enhancing effects, adrenal androgen has been considered a functional antagonist of glucocorticoids, which can be immunosuppressive [4]. In fact, a recent study indicated that DHEA and glucocorticoids regulate the same immune-modulating genes in opposite directions [15]. Considering lines of evidence for activities of adrenal androgen that offset those of glucocorticoid, investigators have proposed quantification of their levels as a ratio, to serve as an alternative index of adrenal activity in different clinical settings [16, 17]. Interestingly, the responses of glucocorticoid and androgen demonstrate a significant discrepancy during critical illness, with adrenal androgen levels decreasing and glucocorticoid levels increasing [18]. This phenomenon is more pronounced in most severely ill patients and nonsurvivors [18‐20], suggesting that functional adaptation of adrenal steroidogenesis may exhaust the counterregulatory mechanisms between adrenal androgen and glucocorticoid, thus negatively impacting the prognosis of critical illness. Indeed, an increased cortisol to DHEAS ratio was shown in nonsurvivors [19, 20], indicating that an exhausted adrenal reserve can serve as a prognostic factor.

Liver cirrhosis is significantly associated with an increased risk of sepsis and sepsis-related mortality [21, 22]. Indeed, upregulated inflammatory cytokines, which mediate inflammation and organ dysfunction in the setting of liver cirrhosis and sepsis [22, 23], are major components facilitating interaction between immune and neuroendocrine systems [24]. Critical illness-related corticosteroid insufficiency (CIRCI) or relative adrenal insufficiency is common in cirrhotic patients with severe sepsis and septic shock, and is associated with increased mortality [25‐28]. However, relevant data about adrenal androgen in critically ill cirrhotic patients are still lacking despite growing interest in the role of adrenal androgen replacement in critical illness [29]. In fact, decreased levels of adrenal androgen are reported in noncritical patients with liver cirrhosis [30, 31]. In a sense, overshooting inflammation in septic shock may overwhelm the adaptive strategy of adrenal steroidogenesis in those patients who succumb. We hypothesized that defective DHEAS production and adrenal exhaustion are associated with inflammation and poor prognosis in cirrhotic patients with septic shock. In this prospective observational investigation, we used the short corticotropin stimulation test (SST) to evaluate adrenal function and studied whether low concentrations of DHEAS and a decreased DHEAS/cortisol ratio are associated with poor outcome in patients with liver cirrhosis and septic shock.

The major findings of this study are as follows. First, the ratio of DHEAS (an immunostimulant) to cortisol (an immunosuppressant) is significantly lower in patients with liver cirrhosis and septic shock, compared to their noncirrhotic counterparts. Second, upon admission to the ICU there is a dissociation between baseline DHEAS (reduced) and cortisol (increased) in nonsurvivors of the cirrhotic group, suggesting a functional adaptation of adrenal steroidogenesis in this subgroup. Finally, upon admission to the ICU a low DHEAS/cortisol ratio is associated with CIRCI and hospital mortality. The DHEAS/cortisol ratio can serve as a prognostic marker in this clinical setting.

Considering the synchronized synthesis of cortisol and DHEAS and their opposing effects to each other, there has been growing interest in measuring these two hormones as a ratio rather than as separate values [16, 17]. In this regard, our study is the first to measure the ratio of DHEAS to cortisol in liver cirrhosis. Interestingly, we found that a low DHEAS/cortisol ratio is associated with inflammation, disease severity, CIRCI, and hospital mortality (Table 3). Our results suggest that the balance between these two mutually dependent hormones may influence the pathophysiological process and ultimately outcomes. Indeed, this index has been used as an indicator in different diseases [16, 17, 19, 20]. This approach also helps to conceptualize the preferential synthesis of glucocorticoid over adrenal androgen in liver cirrhosis and septic shock. Previous studies have shown that there is a discrepancy between DHEAS (decreased) and cortisol (increased) in critical illness [18]. In accordance with this notion, we showed that nonsurvivors of the cirrhotic group had significantly lower baseline DHEAS and significantly higher baseline cortisol, and thus a lower DHEAS/cortisol ratio. The combination of low DHEAS levels with high cortisol levels suggested a shift in pregnenolone metabolism away from adrenal androgen pathways toward the glucocorticoid pathway in this subset of patients. It is conceivable that this functional adaptation could serve to minimize the synthesis of adrenal androgen and maximize the secretion of glucocorticoid which is acutely necessary to antagonize inflammation. Other intriguing findings are that baseline cortisol levels were negatively correlated to cortisol increments, whereas baseline DHEAS levels were positively correlated to DHEAS increments. Furthermore, both DHEAS and cortisol increment were negatively correlated to disease severity scores and associated with hospital mortality. Taken together, preferential adaptation of adrenal steroidogenesis favoring glucocorticoid pathway may deplete adrenal reserve and exhaust the counterbalance between adrenal androgen and glucocorticoid, thus negatively impacting the prognosis of liver cirrhosis with septic shock.

Adrenal androgens are pleiotropic hormones with multiple biological functions. In addition to the immune-enhancing effects mentioned previously, adrenal androgen has been shown to modulate cardiovascular functions. DHEA can exert its protective effects on the cardiac function and integrity of microcirculation of different vasculatures through activating Akt–eNOS signaling pathways [39‐42]. Indeed, in experimental animals subjected to sepsis, administration of DHEA can attenuate organ dysfunction and improve survival. This effect was paralleled by decreased inflammatory cytokine levels and an improved activity of T-cell-mediated immunity [10, 43]. Moreover, salutary effects on hepatic perfusion were also demonstrated in experimental animals treated with androstenediol, a metabolite of DHEA, and subjected to trauma-hemorrhagic shock [44]. Interestingly, these beneficial effects of androstenediol on intrahepatic hemodynamics are due to induction of eNOS-mediated NO and a decrease in endothelin-1 [44], which may also represent a potential remedy for endothelial dysfunction in the microcirculation of cirrhotic liver [45]. It is unknown whether adrenal androgen can provide beneficial effects on immunity and microcirculation and thus improve outcomes in liver cirrhosis with septic shock.

Recently, CIRCI or relative adrenal insufficiency has been used to describe a suboptimal adrenal response to ACTH in critical illness, in which the glucocorticoid levels, although high in terms of absolute value, are inadequate to control inflammation [1, 37]. In most studies, only glucocorticoid levels were measured although the adrenal cortex secretes androgen as well as glucocorticoid upon ACTH stimulation [4]. In this study, we showed that CIRCI is associated with significantly lower DHEAS/cortisol ratio and significantly higher inflammatory cytokines and hospital mortality (Table 4). Additionally, the DHEAS/cortisol ratio had excellent capacity to predict hospital survival in cirrhotic patients with septic shock (AUROC = 0.925), suggesting that these two tightly coordinated hormones should be assessed in concert to better reflect adrenal dysfunction. Although administration of low doses of glucocorticoid in septic patients can restore vascular hyporeactivity [46] and reverse the shock status [26, 47‐50], its impact on survival remains controversial [26, 48‐50]. The reasons for the discrepancy are unclear. Indeed, there are more episodes of superinfection, and higher rates of severe hyperglycemia associated with glucocorticoid administration [26, 49, 50]. In this regard, DHEA has been shown to enhance immunity and help overcome the catabolic effects of glucocorticoid [29]. It is plausible to speculate that glucocorticoid treatment alone may be incomplete unless the imbalance of steroidogenesis is offset by adrenal androgen in addition to glucocorticoid. Further studies are needed to clarify this issue.

Finally, another finding deserving discussion is that cirrhotic patients with septic shock had significantly lower baseline DHEAS, lower baseline DHEAS/cortisol ratio, and decreased cortisol and DHEAS increments upon ACTH challenge, when compared to the noncirrhotic group. However, the baseline cortisol levels were comparable between both groups. This phenomenon can be interpreted as a shift of steroidogenesis toward glucocorticoid biosynthesis at the expense of adrenal androgen in liver cirrhosis. We speculate that liver cirrhosis per se represents a risk factor for altered adrenal steroidogenesis, which may portend a poor prognosis for septic shock if upregulation of glucocorticoid becomes unopposed by adrenal androgen but is still insufficient to control inflammation. In this regard, glucocorticoid resistance may contribute to adrenal dysfunction in our patients. Finally, hepatoadrenal syndrome [27, 28, 51] may need to be redefined and addressed in the context of coordination between separate adrenal hormones.

There are limitations in our study. First, the number of patients is small. Second, the association does not indicate causality because of the observational design. Further studies with a larger cohort and therapeutic intervention are needed in the future.

We demonstrated a divergent biosynthesis between cortisol and DHEAS in cirrhotic patients with septic shock. This phenomenon, as evidenced by an altered DHEAS/cortisol ratio, is associated with more severe diseases and higher levels of inflammatory cytokines and can serve as a prognostic marker. More investigations are needed to evaluate the role of adrenal androgen in this clinical setting.