@article{3174943,
    title = "Alcoholic beverage preference and diabetes incidence across Europe: the
Consortium on Health and Ageing Network of Cohorts in Europe and the
United States (CHANCES) project",
    author = "Sluik, D. and Jankovic, N. and Hughes, M. and O'Doherty, M. G. and and Schoettker, B. and Drygas, W. and Rolandsson, O. and Maennistoe, S. and and Ordonez-Mena, J. M. and Ferrieres, J. and Bamia, C. and De Gaetano, G. and and Kiefte-De Jong, J. C. and Franco, O. H. and Sluijs, I. and and Spijkerman, A. M. W. and Sans, S. and Eriksson, S. and Kromhout, D. and and Trichopoulou, A. and Wilsgaard, T. and Brenner, H. and Kuulasmaa, K. and and Laatikainen, T. and Soederberg, S. and Iacoviello, L. and Boffetta, P. and and Kee, F. and Feskens, E. J. M.",
    journal = "European Journal of Clinical Nutrition",
    year = "2017",
    volume = "71",
    number = "5",
    pages = "659-668",
    publisher = "Nature Publishing Group",
    issn = "0954-3007, 1476-5640",
    doi = "10.1038/ejcn.2017.4",
    abstract = "BACKGROUND/OBJECTIVES: It is unknown if wine, beer and spirit intake
lead to a similar association with diabetes. We studied the association
between alcoholic beverage preference and type 2 diabetes incidence in
persons who reported to consume alcohol.
SUBJECTS/METHODS: Ten European cohort studies from the Consortium on
Health and Ageing: Network of Cohorts in Europe and the United States
were included, comprising participant data of 62 458 adults who reported
alcohol consumption at baseline. Diabetes incidence was based on
documented and/or self-reported diagnosis during follow-up. Preference
was defined when. >= 70% of total alcohol consumed was either beer,
wine or spirits. Adjusted hazard ratios (HRs) were computed using Cox
proportional hazard regression. Single-cohort HRs were pooled by
random-effects meta-analysis.
RESULTS: Beer, wine or spirit preference was not related to diabetes
risk compared with having no preference. The pooled HRs were HR 1.06
(95% confidence interval (CI) 0.93, 1.20) for beer, HR 0.99 (95% CI
0.88, 1.11) for wine, and HR 1.19 (95% CI 0.97, 1.46) for spirit
preference. Absolute wine intake, adjusted for total alcohol, was
associated with a lower diabetes risk: pooled HR per 6 g/day was 0.96
(95% CI 0.93, 0.99). A spirit preference was related to a higher
diabetes risk in those with a higher body mass index, in men and women
separately, but not after excluding persons with prevalent diseases.
CONCLUSIONS: This large individual-level meta-analysis among persons who
reported alcohol consumption revealed that the preference for beer,
wine, and spirits was similarly associated with diabetes incidence
compared with having no preference."
}