Disease-specific C-reactive protein heterogeneity across common infections is associated with a shared systemic inflammatory laboratory profile

Astratt

Background C-reactive protein (CRP) is widely used to characterize systemic inflammation in infectious diseases, yet its absolute concentration may be influenced by infection type, anatomic extent, host characteristics, disease severity, and treatment exposure. We examined whether CRP distributions and the laboratory profile accompanying CRP elevation show reproducible cross-disease directions when evaluated within a harmonized analytical framework. Methods This study was a baseline cross-sectional analysis nested within a single-center retrospective multicohort study of community-acquired pneumonia (CAP), Mycoplasma pneumoniae pneumonia (MPP), influenza, active tuberculosis, sepsis, acute pyelonephritis (APN), diagnosis-record-defined candidiasis, and diagnosis-record-defined acute COVID-19. The first eligible hospitalization per patient within each disease was retained. Baseline laboratory values were the first reported results from 6 h before to 48 h after admission or the disease index time. Cross-disease CRP differences were quantified using nonparametric tests and effect sizes; disease-specific associations were assessed using Spearman correlations with Benjamini-Hochberg false-discovery-rate correction, bootstrap 95% confidence intervals, and CRP-quartile analyses. CRP-by-disease interaction models evaluated whether CRP-associated laboratory slopes differed across diseases. Age- and sex-adjusted cross-disease models and a narrower -6 h to +24 h sensitivity analysis were also performed. Results The analysis included 2,717 disease-specific hospital records from 2,714 unique patients. CRP distributions differed across diseases (Kruskal-Wallis chi-square=245.64, P<0.001; epsilon-squared=0.088). Median CRP was highest in sepsis (92.3 mg/L), intermediate and overlapping in CAP (38.0), influenza (33.9), APN (29.7), and MPP (25.2), and lower in active tuberculosis (8.6), candidiasis (11.1), and COVID-19 (14.1). Across all eight cohorts, CRP correlated positively with NLR and neutrophil count and inversely with albumin; lymphocyte count was inversely associated in seven cohorts. However, CRP-by-disease interactions were significant for NLR (P = 0.020), neutrophils (P = 0.037), and albumin (P = 0.008), but not clearly for lymphocytes (P = 0.059), indicating shared directions with disease-dependent association strength rather than identical slopes. Cross-disease heterogeneity persisted after age and sex adjustment and in the narrower -6 h to +24 h window (n=2,003; P<0.001). Conclusions Common infectious-disease cohorts show marked heterogeneity in absolute CRP concentration while higher CRP is accompanied by recurrent directional changes in routine inflammatory laboratory measures. These cross-sectional findings support disease-context-aware interpretation of CRP as a measure of systemic inflammatory burden. Diagnostic accuracy and prognostic performance were not directly evaluated, and the observed overlap should not be interpreted as establishing a universal etiologic, severity, or prognostic CRP threshold.