Arterial lactate and enteral feeding intolerance within the vasopressor gray zone: a retrospective cohort study

Qinqin Shu, Min Tang, Qi Si, J F Xu, Yan Li

Background Enteral feeding tolerance during vasopressor therapy remains uncertain, particularly within the moderate-dose gray zone. This preregistered analysis tested whether pre-feeding arterial lactate modifies the dose–response relationship between norepinephrine-equivalent dose and enteral feeding intolerance (vomiting). Methods Using the same MIMIC-IV cohort ( n = 8,055), we examined whether baseline arterial lactate (most recent value within 24 h before enteral nutrition initiation) modifies the norepinephrine-equivalent dose–vomiting association. Effect modification was assessed via multiplicative interaction in multivariable logistic regression and stratified restricted cubic spline analyses, with lactate dichotomized at 2 mmol/L. A joint risk surface was constructed using generalized additive modeling. Results Among 5,542 patients with available lactate (68.8% of the parent cohort), 278 (5.0%) developed vomiting. The global test for multiplicative interaction was not significant ( p = 0.27). Exploratory stratified analysis showed a localized signal at doses of 0.3–0.5 μg/kg/min: vomiting was 9.8% among patients with lactate >2 mmol/L versus 4.9% among those with lactate ≤2 mmol/L. Separately, in the dichotomized joint-exposure model, norepinephrine-equivalent dose ≥0.3 μg/kg/min plus lactate >2 mmol/L was associated with an adjusted OR of 2.02 (95% CI 1.33–3.07; p = 0.001) versus dose <0.3 μg/kg/min plus lactate ≤2 mmol/L. Dose–response curves showed steeper risk escalation in the elevated-lactate stratum, though confidence bands overlapped throughout; this divergence was directionally consistent across alternative lactate thresholds (1.5 and 4.0 mmol/L) and exposure windows (6-h, 12-h). Conclusion At moderate vasopressor doses, elevated arterial lactate may identify a subgroup with a higher observed vomiting rate—an exploratory signal not fully captured by vasopressor dose alone. If confirmed prospectively, this pattern could inform a lactate-stratified approach to enteral feeding advancement: at norepinephrine-equivalent 0.3–0.5 μg/kg/min, lactate >2 mmol/L might support slower advancement with closer monitoring rather than standard escalation. Given that feeding intolerance is a key driver of enteral nutrition interruption, these findings have potential downstream implications for nutritional adequacy and clinical outcomes. Because the global interaction was not statistically significant, these findings are hypothesis-generating and require prospective validation.