LVEF-defined septic cardiomyopathy modifies the association between the prognostic nutritional index and mortality in adults with sepsis: a 72-h landmark analysis

Abstract

The prognostic nutritional index (PNI) combines albumin and lymphocyte count and may capture nutrition-related immune-inflammatory status in sepsis. Whether its association with mortality differs by left ventricular ejection fraction (LVEF)-defined septic cardiomyopathy (SCM) is unclear. We conducted a retrospective 72-h landmark study of MIMIC-IV adults with Sepsis-3 who survived to the landmark, had clinically indicated pre-landmark echocardiographic LVEF assessment, and had calculable early PNI. The prespecified primary analysis tested the lower-PNI-by-LVEF-defined-SCM interaction for 28-day mortality in an adjusted Cox model. Ninety-day and sensitivity analyses were supportive or exploratory. A single-center cohort was examined separately, not as formal validation. Among 1,976 patients, 240 had LVEF-defined SCM; 506 and 633 died by 28 and 90 days, respectively. Each 5-point lower PNI was associated with higher 28-day mortality without SCM (adjusted hazard ratio [HR], 1.18; 95% CI, 1.11–1.24) but not with LVEF-defined SCM (HR, 0.96; 95% CI, 0.87–1.05); the ratio of HRs was 0.82 (95% CI, 0.73–0.91; p  < 0.001). The secondary 90-day ratio was 0.84 (95% CI, 0.76–0.94; nominal p  = 0.001). Absolute-risk and restricted-mean-survival-time analyses supported attenuation, whereas same-calendar-day PNI pairing and some alternative LVEF definitions attenuated the interaction estimates toward the null. The external estimate was imprecise and statistically inconclusive (ratio of odds ratios, 0.89; 95% CI, 0.51–1.55). Among selected 72-h survivors with clinically indicated pre-landmark echocardiography, the association between lower PNI and 28-day mortality was attenuated with LVEF-defined SCM. PNI should not be used alone for nutritional or clinical decision-making.