Abstract / Summary
Background and Objectives: Composite indices from routine inflammatory, haematological and nutritional measurements are widely reported as prognostic markers after acute coronary syndromes, but whether they add information beyond their components is uncertain. Materials and Methods: In 30-day survivors of non-ST-segment elevation myocardial infarction (NSTEMI) undergoing coronary angiography at a single centre, thirteen composite indices and six prespecified components were evaluated for five-year all-cause mortality using ROC analysis, Cox regression and the change in Harrell’s C-index (ΔC) over a study clinical model. Each index was also compared with a freely weighted model of its own constituents and with a reference model adding CRP and albumin. Results: Among 561 patients (median follow-up 5.48 years), 94 died, 86 within five years. Composite-index AUCs ranged from 0.616 to 0.689; none exceeded CRP alone after multiplicity correction. Eight indices were nominally associated with mortality after adjustment (hazard ratios 1.26–1.38 per 1 SD), but every bootstrap confidence interval for ΔC over the clinical model included zero. Fixed indices and freely weighted constituents performed similarly (cross-validated ΔC −0.0089 to +0.0053), with one small exploratory difference favouring the fixed MLR ratio. Adding CRP and albumin raised the C-index from 0.7720 to 0.7826; over this reference, no estimable composite showed evidence of association (all p ≥ 0.19). Conclusions: In 30-day survivors of NSTEMI, composite indices showed no evidence of improved long-term mortality discrimination over a study-specific clinical model and contained no information beyond their measured constituents. Their independent associations may largely reflect information already present in the components. Small gains cannot be excluded; the findings do not establish equivalence.