Table 5.

Multivariable-adjusted HRs (95% CIs) for the association of diabetes with site-specific cancer risk stratified by levels of body mass indexa

BMI (kg/m2) categories
CancerNo. of casesb among diabeticsNo. of cases among diabeticsHR (95% CI)No. of cases among diabeticsHR (95% CI)
Esophagus8261.49 (0.98–2.26)300.91 (0.60–1.37)
Colon13560.93 (0.71–1.22)1331.33 (1.09–1.61)
Rectum9301.24 (0.85–1.81)401.23 (0.86–1.75)
Liver6243.19 (1.99–5.12)402.61 (1.73–3.95)
Kidney6261.30 (0.87–1.95)541.33 (0.95–1.87)
Bladder13271.30 (0.85–2.01)461.06 (0.71–1.59)
Breast premenopausal4822
Breast postmenopausal24570.99 (0.76–1.29)1431.21 (1.01–1.44)
Endometrium1222.28 (1.45–3.59)621.32 (1.00–1.75)
Prostate461880.83 (0.72–0.97)1600.64 (0.54–0.75)
Lung among ever smokers28761.30 (1.02–1.65)951.00 (0.79–1.26)
Lung among never smokers347
Melanoma8360.69 (0.46–1.14)601.07 (0.79–1.44)
Multiple myeloma11114
  • Note: In each category of BMI, the nondiabetic group was used as reference category.

  • Abbreviations: DLBCL, diffuse large B-cell lymphoma; NHL, non-Hodgkin lymphoma.

  • aHRs and 95% CI adjusted for age, sex, education (≥ college), non-white race, smoking status (present, past, current), smoking-years (packs), alcohol intake (never, special occasion, 1–3 drinks/week, 3–4 drinks/week, daily), physical activity (quintiles of MET-hour/week), and BMI (kg/m2; model 2, as described in the text). HRs for cancer risk at specific sites are additionally adjusted as reported in the footnotes of Table 2.

  • bNo diabetes associated HRs were estimated for participants with BMI <25 kg/m2 since few cancer cases occurred among diabetes in that category.