Sex differences in the association between diabetes and cancer: a systematic review and meta-analysis of 121 cohorts including 20 million individuals and one million events

Abstract

Aims/hypothesis: Diabetes has been shown to be a risk factor for some cancers. Whether diabetes confers the same excess risk of cancer, overall and by site, in women and men is unknown.

Methods: A systematic search was performed in PubMed for cohort studies published up to December 2016. Selected studies reported sex-specific relative risk (RR) estimates for the association between diabetes and cancer adjusted at least for age in both sexes. Random-effects meta-analyses with inverse-variance weighting were used to obtain pooled sex-specific RRs and women-to-men ratios of RRs (RRRs) for all-site and site-specific cancers.

Results: Data on all-site cancer events (incident or fatal only) were available from 121 cohorts (19,239,302 individuals; 1,082,592 events). The pooled adjusted RR for all-site cancer associated with diabetes was 1.27 (95% CI 1.21, 1.32) in women and 1.19 (1.13, 1.25) in men. Women with diabetes had ~6% greater risk compared with men with diabetes (the pooled RRR was 1.06, 95% CI 1.03, 1.09). Corresponding pooled RRRs were 1.10 (1.07, 1.13) for all-site cancer incidence and 1.03 (0.99, 1.06) for all-site cancer mortality. Diabetes also conferred a significantly greater RR in women than men for oral, stomach and kidney cancer, and for leukaemia, but a lower RR for liver cancer.

Conclusions/interpretation: Diabetes is a risk factor for all-site cancer for both women and men, but the excess risk of cancer associated with diabetes is slightly greater for women than men. The direction and magnitude of sex differences varies by location of the cancer.

Keywords: Cancer; Diabetes; Meta-analysis; Sex differences; Systematic review.

Fig. 1

Flow chart of study selection

Fig. 2

Maximum-available-adjusted RR for all-site cancer, comparing individuals with diabetes with those without diabetes by sex: (a) women; and (b) men. ANZ, Australia and New Zealand; ARIC, Atherosclerosis Risk in Communities; BRFSS, Behavioral Risk Factor Surveillance System; CLUE II, Give Us a Clue to Cancer and Heart Disease; CPS II, Cancer Prevention Study II; DECODE, Diabetes Epidemiology: Collaborative analysis of Diagnostic criteria in Europe; DERI, Diabetes Epidemiology Research International; DRT, Diabetes Registry Tyrol; 2001 ENTRED study, 2001–2006 National representative sample of people with diabetes study; HSE, Health Survey for England; MHS, Maccabi Healthcare Services; NDSS, National Diabetes Services Scheme; NIH-AARP, National Institutes of Health-American Association of Retired Persons; NHIS-NSC, Korean National Health Insurance Service-National Sample Cohort; SHeS, Scottish Health Survey; VHM&PP, The Vorarlberg Health Monitoring and Promotion Programme

Fig. 3

Maximum-available-adjusted women-to-men RRR for all-site cancer, comparing individuals with diabetes with those without diabetes. For definition of study acronyms, please refer to Fig. 2 legend. ^aThe BRFSS did not report the total number of cancer events

Fig. 4

Subgroup analyses of women-to-men RRR for all-site cancer, comparing individuals with diabetes with those without diabetes. ^aSix studies were excluded because the baseline year bridged over 1985 (i.e. included both pre-1985 and 1986 onwards). ^bResults using multiple adjustment were used when available and age-adjusted otherwise, as in Fig. 3. ^cTen studies were excluded because absolute risks for men and women were unavailable

Fig. 5

Maximum-available-adjusted pooled women-to-men RRR for cancer at each site, comparing individuals with diabetes with those without diabetes