Novel predictors of response to therapy with terlipressin and albumin in hepatorenal syndrome–acute kidney injury

Introduction

Hepatorenal syndrome (HRS) is defined as renal failure developing in patients with advanced cirrhosis of the liver, in the absence of significant structural abnormalities in the kidneys, due to vasoconstriction of the renal circulation [1]. There are 2 types of HRS [2]: (i) HRS with acute kidney injury (AKI; earlier called Type 1 HRS), characterized by a rapidly progressive impairment in renal function, associated with high mortality but with a potential for reversibility. The most common precipitants for HRS-AKI are bacterial infections [3], followed by drugs, gastrointestinal bleeding and severe alcoholic hepatitis; and (ii) HRS-non-AKI (earlier called Type 2 HRS), characterized by a moderate renal impairment with a progressive course that steadily evolves over weeks to months. HRS-non-AKI is further subclassified as HRS-acute kidney disease (renal dysfunction lasting for less than 3 months but not meeting the AKI criteria) and HRS-chronic kidney disease (estimated glomerular filtration rate <60 mL/min/1.73 m² for >3 months).

The pathogenesis of HRS involves marked splanchnic vasodilatation due to portal hypertension and activation of systemic vasoconstrictors, both of which result in a significant reduction in the effective circulating blood volume and lead to marked impairment of renal blood flow [4].

Untreated HRS-AKI has an extremely dismal prognosis, with a median survival of weeks to months [5]. The transplant-free survival can, however, be improved by the identification and removal of the precipitating factor of AKI and appropriate pharmacologic treatment. Various studies have shown that the best available pharmacologic therapy for HRS is a combination of intravenous terlipressin (splanchnic vasoconstrictor) and intravenous albumin [6,7]. Terlipressin counteracts the splanchnic arterial vasodilatation, while albumin is a volume expander that further improves effective circulating volume and cardiac contractility. The safety and efficacy of terlipressin in patients with HRS-AKI have previously been evaluated in randomized trials, with response rates varying from 40-60% [8-10]. A meta-analysis of clinical trials [11] found that the pooled rate of patients with a complete response to terlipressin and albumin was 52%. In view of the modest rates of response to pharmacologic therapy, it is crucial to identify those patients with a low likelihood of treatment response early in the course, so that they can be offered other modalities of treatment, and workup for liver transplantation can be expedited.

The majority of studies assessing the response rates to vasoconstrictor therapy were performed in patients in whom Type 1 HRS was defined according to the old International Club of Ascites (ICA) criteria [1]. However, the definition of AKI in cirrhosis was recently modified [12] to align with Kidney Disease: Improving Global Outcomes (KDIGO) criteria [13]. Very few studies have evaluated the impact of the new AKI definition on treatment response rates. This study aimed to assess response rates to terlipressin based on the new definition of HRS-AKI and to identify novel predictive factors of improved renal function and survival.

Patients and methods

Results

During the study period, we identified 90 patients diagnosed with HRS-AKI and started on treatment with terlipressin and albumin. Six were excluded as they had adverse responses to terlipressin requiring discontinuation of therapy.

Predictors of treatment response

The clinical and demographic variables of patients with a complete response to treatment were compared with non-responders to identify predictors of successful treatment (Table 1). The factors associated with complete treatment response to treatment on univariate analysis were the presence of SIRS at baseline, serum creatinine, rise in mean arterial pressure by day 3 of treatment, stage of AKI, and a reduction in the renal resistive index (ΔRRI) by day 3 of treatment (P<0.001 for all). Forty-five patients had features of SIRS (≥2 SIRS components) at baseline, and in 76% of these patients (n=34) there was a complete response to treatment. In patients without SIRS, HRS reversal was seen only in 31% (12/39). Among the various stages of AKI (as defined by KDIGO), the rates of complete response to treatment were 91.3% (21/23), 71.4% (20/28), and 23.8% (5/21) for AKI stages 1, 2 and 3, respectively.

Table 1 Comparison of sociodemographic and clinical characteristics of complete responders and non-responders

According to the multivariate analysis, the independent predictors of treatment response were the presence of SIRS at baseline (P=0.022; odds ratio [OR] 15.749, 95% confidence interval [CI] 1.478-167.82) and ΔRRI by day 3 (P=0.048; OR 6.67, 95%CI 1.021-43.625) (Table 2). The cutoff level for ΔRRI that best predicted response to treatment was ≥5% by day 3 (area under the curve, 0.803; P<0.001; sensitivity 83%; specificity 69%). Fig. 1 shows the ROC curve for the correlation of ΔRRI with treatment response. In those with ΔRRI ≥5% by day 3, there was a complete response to treatment in 85.4% (41/48), while in those with a ΔRRI <5% by day 3, only 20.8% (5/24) showed a complete treatment response.

Table 2 Binary logistic regression analysis to determine variables with independent predictive value for treatment response

Figure 1 Correlation of ΔRRI with treatment response (AUROC 0.803)

ΔRRI, change in the renal resistive index; CI, confidence interval; AUROC, area under the receiver operating characteristic curve

Survival

Mortality was higher in non-responders compared to responders at 1 month (27% vs. 9%) and 6 months (73% vs. 45%). Response to treatment was significantly associated with transplant-free survival at 6 months (log-rank P<0.001) (Fig. 2). Mean transplant-free survival was significantly longer in the complete responders than in non-responders (148 days vs. 90 days). Response to terlipressin treatment was an independent predictor of transplant-free survival at 6 months (hazard ratio [HR] 0.527, 95%CI 0.342-0.810; P=0.004, while a higher MELD-sodium score was independently associated with 6-month mortality (HR 1.071, 95%CI 1.012-1.135; P=0.018) (Table 3). The cutoff level of MELD-sodium that best predicted 6-month mortality was ≥23 (area under the curve, 0.664; P=0.019; sensitivity, 84%; specificity, 43%).

Figure 2 Kaplan-Meier survival analysis of the patient cohort

df, degrees of freedom

Table 3 Cox proportional hazards regression analysis and independent predictors of 6-month mortality in HRS-AKI

Discussion

HRS is one of the most challenging complications of advanced liver disease and is associated with extremely high mortality. Successful treatment of HRS-AKI with pharmacologic therapy has the potential to improve short-term transplant-free survival and reduce the need for RRT. A combination of terlipressin and albumin is the first-line pharmacologic therapy to counteract splanchnic arterial vasodilation and improve renal perfusion in HRS-AKI.

Various studies have evaluated the efficacy of pharmacologic therapy in HRS-AKI. However, most studies involved patients in whom type 1 HRS was defined according to the old ICA criteria [1]. This study used the new AKI definition to define response status and compared it with existing studies to assess whether the change in the definition of AKI in cirrhosis has an impact on the response rates to therapy.

Among the 84 patients with HRS-AKI included in this study, a response to treatment was observed in 69.1% (complete response in 54.8% and partial response in 14.3%.). The overall response rates were better with the new definition of HRS compared to existing studies, where response rates varied between 40-60% [8-10]. However, the rate of complete response to treatment observed in this study (54.8%) was similar to that observed in previous studies [10,11,19,20]. Despite being the first-line pharmacologic treatment of HRS, it was observed that less than 60% of patients have a complete response to therapy with terlipressin. Therefore, it is crucial to identify non-responders early so they can be fast-tracked for liver transplantation. We compared the clinical and sociodemographic variables between complete responders and non-responders to identify predictors of successful treatment of HRS.

In this study, the factors associated with a complete response to treatment on univariate analysis were serum creatinine, the presence of SIRS on day 1, a rise in mean arterial pressure by day 3, the stage of AKI, and a reduction in RRI by day 3. Pre-treatment serum creatinine and a rise in mean arterial pressure as predictors of HRS reversal were also noted in various studies [21-23]. Other predictive factors of treatment response reported in previous studies [8,20,24-26] included age, alcohol etiology of cirrhosis, leucocyte count, serum bilirubin, serum albumin, Child-Pugh and MELD scores. However, these were not found to be statistically significant in our study. The independent predictors of treatment response on multivariate analysis were the presence of SIRS on day 1 and a reduction in the RRI (ΔRRI) ≥5% by day 3.

The fact that the presence of SIRS improved the renal response to terlipressin in patients with HRS-AKI is an interesting observation. Even in the absence of a clinically apparent infection, SIRS occurs frequently in patients with decompensated cirrhosis because of an increase in gut bacterial translocation, leading to a persistent inflammatory state that can be exacerbated during acute decompensations [27-29]. Terlipressin, by reducing the portal pressure, may lead to decreased gut bacterial translocation, which in turn reduces endotoxemia and the production of proinflammatory cytokines [30,31]. Terlipressin is also thought to have a direct anti-inflammatory action, independent of its effect in reducing portal hypertension [32]. Thus, in the presence of SIRS there is an exaggerated response to terlipressin, which leads to an enhanced renal response and resolution of HRS-AKI. Therefore, in patients with HRS-AKI, the presence of SIRS helps to identify a subset of patients that may not respond to volume expansion with albumin alone and could have a better response to terlipressin. In our study, the presence of SIRS on day 1 was found to have an independent predictive value for complete treatment response in cirrhosis patients with HRS-AKI. Very few studies have evaluated the impact of SIRS as a determinant of treatment response to terlipressin in HRS-AKI. Wong et al noted that HRS reversal with terlipressin was observed in 42.9% of subjects with SIRS, as compared to 24.6% in patients without SIRS [33].

The major pathophysiology of HRS is splanchnic arterial vasodilatation, leading to pooling of blood in the splanchnic vascular bed with associated renal artery vasoconstriction and hypoperfusion of the kidneys. This intrarenal arterial vasoconstriction causes an increase in the resistive index of the renal arterial system, which can be assessed by Doppler ultrasound [14-16]. Terlipressin causes selective vasoconstriction of splanchnic arterial vessels, which tends to reverse alterations in renal hemodynamics. RRI is a useful index for quantifying renovascular resistance in cirrhotic patients before HRS develops [34], but its utility in predicting treatment response in HRS has not been evaluated. We made serial measurements of RRI in patients with HRS-AKI who were started on terlipressin. The baseline RRI did not differ significantly between responders and non-responders. However, a change in RRI in the interlobar artery of ≥5% by day 3 of treatment predicted a complete treatment response. This was a novel finding in our study, and it will require validation in larger multicenter studies to assess its reproducibility across different settings (Fig. 3).

Figure 3 The postulated relationship between Renal Resistive Index, SIRS and the renal response to terlipressin in HRS-AKI

HRS, hepatorenal syndrome; AKI, acute kidney injury; ΔRRI, change in the renal resistive index; SIRS, systemic inflammatory response syndrome

The results of the current study confirm data from previous reports indicating that patients with type 1 HRS who respond to treatment with terlipressin and albumin have longer survival compared with non-responders Both the long-term and short-term mortality were significantly higher in non-responders as compared to responders at 1 month (27% vs. 9%) and 6 months (74% vs. 45%). The mean transplant-free survival at 6 months was significantly longer in the responder group compared with non-responders to terlipressin therapy (148 vs. 90 days). Non-response to pharmacologic treatment and model for end-stage liver disease-sodium score ≥23 were found to be independent predictors of 6-month mortality in HRS-AKI. This highlights the fact that non-responders should be identified early and fast-tracked for liver transplantation.

The independent predictors of transplant-free survival at 6 months by Cox regression analysis were a response to treatment and MELD-sodium, which is consistent with the findings of Nguyen-Tat et al [9] and Heidemann et al [10].

Our study had certain limitations. The study was conducted on subjects from a single center, and the duration of follow up was 6 months. In the future, multi-center studies will be needed with larger sample sizes and a longer duration of follow up to assess long-term survival. Another limitation of our study is that in the absence of a placebo group, we cannot causally attribute survival advantage to the terlipressin response alone.

In conclusion, the rates of response to terlipressin and albumin in HRS-AKI are limited, even though this is the first-line pharmacologic therapy. Response to treatment is an independent predictor of transplant-free survival at 6 months. Non-responders have higher mortality and should be identified early to expedite liver transplantation. SIRS and ΔRRI are simple parameters to predict treatment response in HRS-AKI. In HRS-AKI, the RRI is a measure of the intrarenal vasoconstriction, and SIRS reflects exaggerated systemic inflammation, both of which are reversed by terlipressin, leading to an enhanced treatment response.