Association of C-reactive protein with future development of diabetes: a population-based 7-year cohort study among Norwegian adults aged 30 and older in the Tromsø Study 2007–2016

OBJECTIVES: The extent to which observed associations between high-sensitivity C-reactive protein (hs-CRP) and incident diabetes are explained by obesity and hypertension remains unclear. This study aimed to investigate the association of hs-CRP with developing diabetes in a Norwegian general population sample. DESIGN: A cohort study using two population-based surveys of the Tromsø Study: the sixth survey Tromsø6 (2007–2008) as baseline and the seventh survey Tromsø7 (2015–2016) at follow-up. SETTING: Tromsø municipality of Norway, a country with increasing proportion of older adults and a high prevalence of overweight, obesity and hypertension. PARTICIPANTS: 8067 women and men without diabetes, aged 30–87 years, at baseline Tromsø6 who subsequently also participated in Tromsø7. OUTCOME MEASURES: Diabetes defined by self-reported diabetes, diabetes medication use and/or HbA1c≥6.5% (≥48 mmol/mol) was modelled by logistic regression for the association with baseline hs-CRP, either stratified into three quantiles or as continuous variable, adjusted for demographic factors, behavioural and cardiovascular risk factors, lipid-lowering medication use, and hypertension. Interactions by sex, body mass index (BMI), hypertension or abdominal obesity were assessed by adding interaction terms in the fully adjusted model. RESULTS: There were 320 (4.0%) diabetes cases after 7 years. After multivariable adjustment including obesity and hypertension, individuals in the highest hs-CRP tertile 3 had 73% higher odds of developing diabetes (OR 1.73; p=0.004; 95% CI 1.20 to 2.49) when compared with the lowest tertile or 28% higher odds of incidence per one-log of hs-CRP increment (OR 1.28; p=0.003; 95% CI 1.09 to 1.50). There was no evidence for interaction between hs-CRP and sex, hypertension, BMI or abdominal obesity. CONCLUSIONS: Raised hs-CRP was associated with future diabetes development in a Norwegian adult population sample. The CRP-diabetes association could not be fully explained by obesity or hypertension.