ASSOCIATION BETWEEN DIABETES MELLITUS AND SELF-REPORTED HISTORY OF CANCER AMONG ADULTS: A CROSS-SECTIONAL ANALYSIS OF NHANES DATA

Abstract

Background: Diabetes mellitus and cancer may coexist, but their observed relationship may reflect shared metabolic and behavioural risk factors. This study examined the association between diabetes mellitus and self-reported history of any cancer among U.S. adults.

Methods: This cross-sectional analysis used NHANES 2017–2018 data. Adults aged ≥20 years with complete primary-analysis data on self-reported diabetes, self-reported cancer history, body mass index, demographic variables, smoking, and mobile examination center weights were included (n = 4,347). The outcome was report that a health professional had ever diagnosed cancer or malignancy. Survey-weighted logistic regression estimated crude, demographic-adjusted, and fully adjusted odds ratios. A sensitivity analysis used diagnosed diabetes or glycated hemoglobin (HbA1c) ≥6.5% (n = 4,214).
Results: Weighted cancer-history prevalence was 20.1% (95% CI, 15.2%–26.1%) among adults with diabetes and 9.4% (95% CI, 7.6%–11.6%) among adults without diabetes. Diabetes was associated with higher unadjusted odds of cancer history (OR, 2.41; 95% CI, 1.57–3.70; p < .001). The estimate attenuated after adjustment for age, sex, and race/ethnicity (OR, 1.24; 95% CI, 0.76–2.02) and remained non-significant after full adjustment (OR, 1.31; 95% CI, 0.80–2.15; p = .262). The sensitivity model was also non-significant (OR, 1.34; 95% CI, 0.85–2.09; p = .189).

Conclusions: Diabetes was associated with greater unadjusted co-occurrence of self-reported cancer history, but no statistically significant independent cross-sectional association remained after measured confounding was addressed. Findings do not establish temporality, causation, cancer-site-specific associations, or incident cancer risk.