Early adjunctive vasopressin was not associated with a lower 7-day risk of renal replacement therapy in norepinephrine-treated patients with cirrhosis: a target-trial-informed study with external replication

Abstract

Background Early adjunctive vasopressin can reduce catecholamine exposure in vasodilatory shock, but whether it prevents severe kidney failure in norepinephrine-treated patients with cirrhosis is uncertain. Methods We conducted a target-trial-informed landmark study using MIMIC-IV as the primary cohort and eICU-CRD as an external replication cohort. Adults with a first intensive care unit admission, cirrhosis or advanced chronic liver disease, and norepinephrine initiation within 24 h of admission were eligible. Treatment was classified at a 6-h landmark as vasopressin initiated within 6 h of norepinephrine or no vasopressin during that window; patients in the latter group could receive vasopressin later. The primary outcome was new renal replacement therapy (RRT) between the landmark and day 7. Propensity-score overlap weighting was prespecified as the primary analysis. Results The MIMIC-IV cohort included 1,086 patients (early vasopressin, n = 295; no early vasopressin, n = 791), and the eICU-CRD cohort included 356 patients (n = 91 and n = 265, respectively). Before weighting, vasopressin recipients had greater shock and liver-disease severity. Early vasopressin was not associated with lower odds of new RRT in MIMIC-IV (overlap-weighted odds ratio [OR] 1.37, 95% confidence interval [CI] 0.83–2.26; P = 0.214), eICU-CRD (OR 1.10, 95% CI 0.45–2.69; P = 0.833), or the pooled analysis (OR 1.30, 95% CI 0.84–2.01; P = 0.235). Mortality was higher among early-vasopressin recipients, but this secondary association remained highly vulnerable to confounding by indication. Weighted absolute-risk estimates, cumulative RRT incidence, subgroup analyses, and sensitivity analyses provided no supporting renal-benefit signal; creatinine trajectories were evaluated descriptively because follow-up measurements were susceptible to survivor selection. Conclusion Among norepinephrine-treated patients with cirrhosis or advanced chronic liver disease who were alive and RRT-free 6 h after norepinephrine initiation, early adjunctive vasopressin was not associated with less RRT by day 7. The findings do not challenge vasopressin use for hemodynamic support when clinically indicated, but do not support early initiation solely as a renal-protective strategy.