U-shaped association between hemoglobin levels and cardiometabolic multimorbidity in older adults: a large-scale cross-sectional study.

BACKGROUND

Multimorbidity is common in older adults and associated with poor outcomes. Evidence for a nonlinear relationship between hemoglobin (Hb) levels and multimorbidity risk remains limited, particularly in large-scale studies. This study examines the dose-response relationship between Hb and cardiometabolic multimorbidity (CMM) to inform early risk assessment.

METHODS

This cross-sectional study analyzed data from the National Primary Public Health Service System (NPPHSS) in Hainan Province, China, including 466,751 adults aged ≥ 65 years. Key variables included Hb levels, measured using the sodium lauryl sulfate method, and CMM (defined as the coexistence of ≥ 2 conditions among hypertension, diabetes, and hyperlipidemia). We used logistic regression and restricted cubic spline (RCS) models to assess the nonlinear dose-response relationship between Hb and CMM risk. Subgroup and sensitivity analyses were performed to assess the robustness of the findings.

RESULTS

The mean age was 74.6 ± 6.9 years; 56.4% were female. CMM prevalence was 32.1%. Mean Hb was 132.9 ± 15.4 g/L. RCS analysis revealed a significant nonlinear relationship (P for non-linear < 0.001). Odds ratios (ORs) formed a U-shaped curve, decreasing to a nadir at 112 g/L (OR = 0.885, 95% CI: 0.868 - 0.902) and rising with higher Hb levels, exceeding 1.0 above approximately 132 g/L. Subgroup analyses by sex, age, BMI, and multimorbidity patterns confirmed consistent U-shaped associations. Especially, among participants with all three chronic conditions, Hb ≥ 132 g/L or ≤ 83 g/L was positively associated with multimorbidity, and a negative association was noted in the Hb range of 104 - 132 g/L.

CONCLUSION

Our findings demonstrate a significant U-shaped association between Hb levels and CMM in Chinese older adults. Maintaining Hb within a reasonable range may be a consideration for primary prevention strategies.