Customize your JAMA Network experience by selecting one or more topics from the list below.
Tuomainen T, Lagundoye A, Voutilainen S. Coffee Intake and Glucose Homeostasis: Is There a Role for Body Iron? Arch Intern Med. 2010;170(15):1400–1401. doi:10.1001/archinternmed.2010.252
Since the original report by van Dam et al,1 coffee drinking has been associated with a decreased risk of type 2 diabetes mellitus in a number of epidemiological studies. However, body iron has for over a century been known to cause diabetes if in overt excess, manifested as the “bronzed diabetes”–hereditary hemochromatosis. We hypothesize in line with Mascitelli et al2 in their letter to the editor regarding a study by Pereira et al3 that the protective effect that coffee shows toward type 2 diabetes mellitus is perhaps, at least partially, explained by the iron absorption inhibitory effect of coffee. If this were so, subjects who consume much coffee should have both lower body iron stores and better glucose homeostasis compared with people who drink less or no coffee.
We looked at the association of coffee consumption with body iron and glucose homeostasis in 2682 men, aged 42 to 60 years, in the Kuopio Ischaemic Heart Disease Risk Factor Study (KIHD) in eastern Finland. The KIHD study has been approved by the joint research ethics committee of the University of Kuopio and Kuopio University Hospital. Dietary intake of foodstuffs was estimated by a 4-day food record,4 body iron was assessed as serum ferritin concentration, and glucose homeostasis was studied by the updated homeostasis model assessment (HOMA2) insulin resistance (IR) and pancreatic β-cell function (%β). The steady-state nonlinear HOMA2 models the interplay between hepatic glucose output, body glucose uptake, and insulin secretion and produces computational IR and %β parameters.5 The model sets the normal IR to 1.0 and the normal %β to 100. Ferritin, glucose, and insulin measurements were carried out in fasting state samples, as previously described.6
The mean (SD) values for the study subjects was 53.1 (5.1) years for age; 566 (297) mL/d for coffee intake; 168 (152) μg/L for serum ferritin concentration; 1.51 (0.89) for IR; and 112% (39%) for %β.
In an unadjusted linear regression model, IR decreased 2.1% per 100-mL increase in coffee intake and %β decreased 0.8%. Adjustment for ferritin level attenuated the association of coffee intake with IR from −0.021 (P < .001) to −0.009 (P = .03) (difference, −54%) and the association of coffee intake with %β from −0.843 (P = .001) to −0.706 (P = .008) (difference, −16%).
In the multivariate-adjusted linear regression models, serum ferritin adjustment weakened the age- and body mass index (BMI)-adjusted association of coffee intake with IR markedly (Table, model 2). Multivariate analyses of coffee intake and %β were very resistant to adjustments. Furthermore, serum ferritin concentration was not associated with %β in these models, probably because adding BMI to our statisical model made the association between iron and B-cell function disappear (regression coefficient, −0.002 [P = .96] with BMI [Table, model 2] vs 0.200 [P < .001] without BMI [model not shown]). Further adjustment for other dietary factors that relate to iron balance, such as intakes of red meats, vegetables, and milk, and for total energy intake did not confound the observed associations significantly (data not shown).
These results suggest that coffee consumption may be associated with both body iron stores and glucose homeostasis as measured by HOMA2 IR and HOMA2 %β. Furthermore, the results suggest that the association of coffee intake with IR may be partially explained by a decrease in body iron level.
Correspondence: Dr Tuomainen, Institute of Public Health and Clinical Nutrition, University of Eastern Finland, PO Box 1627, 70211 Kuopio, Finland (Tomi-Pekka.Tuomainen@uef.fi).
Author Contributions:Study concept and design: Tuomainen, Lagundoye, and Voutilainen. Acquisition of data: Tuomainen, Lagundoye, and Voutilainen. Analysis and interpretation of data: Tuomainen, Lagundoye, and Voutilainen. Drafting of the manuscript: Tuomainen, Lagundoye, and Voutilainen. Critical revision of the manuscript for important intellectual content: Tuomainen, Lagundoye, and Voutilainen. Statistical analysis: Tuomainen. Obtained funding: Tuomainen. Administrative, technical, and material support: Tuomainen. Study supervision: Tuomainen.
Financial Disclosure: None reported.
Funding/Support: This study was partially funded by a Juho Vainio Foundation grant to Dr Tuomainen.