Very low HDL cholesterol in infectious mononucleosis with hepatitis: a real-world evidence study

Background Infectious mononucleosis (IM) with hepatitis suppresses HDL cholesterol (HDL-C), but its magnitude, specificity, recovery, and cardiovascular implications are uncharacterised. Objective To determine the prevalence, specificity, recovery, and long-term cardiovascular significance of very low HDL-C in IM hepatitis. Methods In the TriNetX Global Collaborative Network (∼190 million patients), we studied 1,944 adults with IM and hepatitis. Presenting HDL-C distributions were compared against 14 propensity-score-matched (PSM) comparator cohorts and characterised by Gaussian mixture modelling; longitudinal trajectory was assessed across six post-index windows. Long-term major adverse cardiovascular events (MACE) were analysed in PSM cohorts with very low HDL (≤20 mg/dL, n=979/arm) versus without, over up to 20 years, with independent COVID-19 (n=83,888/arm) and pharyngitis (n=10,618/arm) cohorts as external validation. Robustness was tested with E-values. All cross-sectional and survival analyses were repeated in sex-specific PSM cohorts, and pre-illness HDL-C was retrieved for patients with a lipid measurement recorded at least three months before the index illness. Results Mean presenting HDL was 36.7±22.6 mg/dL versus 49.7 pre-illness; 23.9% had HDL ≤20 mg/dL. Suppression matched CMV hepatitis but exceeded pharyngitis and IM without hepatitis, indicating a hepatitis-driven mechanism. Mixture modelling identified a discrete suppressed subpopulation (mean 16 mg/dL, 41%) absent in non-hepatitis controls. Recovery was rapid in most patients but prolonged when HDL was ≤ 20 mg/dL. Across repeated cross-sectional sampling, HDL returned to pre-illness values by 16–30 days in the cohort overall and by 31–90 days in patients presenting ≤20 mg/dL. Patients who suppressed most deeply had pre-illness values ∼10 mg/dL below the rest of the cohort, and referred to their own baseline had recovered fully by 1–3 years. Very low acute HDL predicted higher long-term MACE (HR 1.92–2.47), reproduced in COVID-19 (HR 2.04–2.70) and pharyngitis (HR 1.43–1.69) ; E-values (1.88 – 4.38) indicated robustness to unmeasured confounding . Women presented with higher HDL-C than men (36.8 vs 31.3 mg/dL; p < 0.001), but the proportion ≤20 mg/dL did not differ by sex (28.3% vs 31.2%; RR 0.91, 95% CI 0.77–1.07), and hazard ratios were similar in women and men for every outcome. Conclusions Very low HDL-C is a prevalent, hepatitis-driven finding in IM that identifies a subgroup at elevated long-term cardiovascular risk. However, the acute HDL-C dip is likely a marker of lipid phenotype rather than an independent risk factor, warranting prospective evaluation of cardiovascular follow-up.