Dynamic lactate trajectories improve risk prediction beyond baseline severity assessment in cardiogenic shock

Background Early risk stratification in cardiogenic shock (CS) is essential for guiding therapeutic decision-making. The Society for Cardiovascular Angiography and Interventions (SCAI) shock classification is widely used and provides structured baseline risk assessment. However, CS represents a dynamic condition in which evolving metabolic responses may convey additional prognostic information. Serial lactate measurements may therefore refine outcome prediction beyond admission severity assessment, as recently highlighted by the SCAI Door-to-Lactate Clearance Initiative.¹ Purpose To investigate whether early lactate trajectories during intensive care treatment improve prediction of 30-day mortality beyond baseline clinical risk assessment in patients with cardiogenic shock. Methods Data were obtained from an ongoing prospective multicentre cohort study including consecutive patients with cardiogenic shock of all causes. Patients with uncertain shock aetiology or pulmonary embolism were excluded. Lactate trajectories at 4, 8, 16, and 24 hours after ICU admission were categorised as clearance (≥50% reduction or lactate <2 mmol/L), partial clearance (10–50% reduction), or no clearance (<10% reduction). Associations with 30-day mortality were analysed using Cox proportional hazards models adjusted for age, SCAI stage at admission, and baseline lactate. Model discrimination was assessed using Harrell’s concordance index (C-index). Incremental prognostic value of lactate trajectories was evaluated by changes in C-index (ΔC) and likelihood ratio testing compared with baseline models. Results A total of 571 patients were included in the analysis. Impaired lactate trajectories were consistently associated with increased mortality across all timepoints. Trajectory-based models demonstrated moderate discrimination, with an adjusted C-index of 0.72 as early as 8 hours after ICU admission (Fig. 1). Addition of lactate trajectories to baseline risk assessment significantly improved model discrimination, with incremental prognostic value detectable from 8 hours onwards and increasing over time (ΔC +0.06 at 24 hours, Figure 2). The incremental benefit was more pronounced in clinically relevant subgroups. In patients with non-AMI cardiogenic shock, inclusion of lactate trajectories markedly improved prediction from 8 hours onward (ΔC +0.13 at 24 h, LRT p <0.001). Similarly, enhanced discrimination was observed in women (ΔC +0.06 at 8 h, LRT p <0.001). Conclusion Lactate trajectories provide early prognostic information in cardiogenic shock beyond baseline SCAI severity classification. Incremental improvement in risk discrimination is detectable as early as 8 hours after ICU admission. While gains in the overall population are modest, substantial improvement is observed in non-AMI cardiogenic shock and in female patients, suggesting particular value for risk stratification in clinically heterogeneous and traditionally underrepresented phenotypes.