Coffee: friend or foe for diabetes risk?


Coffee is one of the most popular drinks in the world and can have a mild stimulating effect because of its caffeine content. Worldwide coffee consumption is highest in Finland, with an average per-person consumption of 4.7 cups per day, which is equivalent to 12 kg/year; Norway, Iceland, Denmark and the Netherlands are the next highest with 3.9–3.3 cups per day. One of the earliest records of coffee drinking comes from Ethiopia. Myth has it that Kaldi, a 9th century goat herder noticed his flock eating a bright red berry, after which they became very excitable and started to jump everywhere. The goat herder took some of these berries to the local monk, whereupon they were thrown into the fire with disapproval. The monk then noted the ensuing lovely aroma; he retrieved the beans and used them for one of the first cup of coffees.

​Recent meta-analyses have indicated a modest benefit of coffee consumption on both Type 2 diabetes and cardiovascular risk. For cardiovascular (both coronary heart disease and stroke) risk this is a non-linear relationship, with lowest risk associated with 3–5 cups per day and no adverse effect of heavy coffee consumption. A recent meta-analysis in Type 2 diabetes has compared caffeinated and decaffeinated coffee and concluded that both are associated with reduced risk. A dose response was also found compared with no or rare coffee consumption; a relative risk for 1 cup of coffee per day of 0.92 (0.90–0.94) reducing to 0.67 (0.61–0.74) for 6 cups per day. Controversy exists over what in coffee may be leading to a protective effect on Type 2 diabetes. Furthermore, whilst meta-analyses are commonly used to support evidence of association, they may be complicated by publication bias, leading to false positives, and they certainly do not prove causality.

​The situation is further complicated in that one of the first reports of coffee consumption and diabetes was in fact a positive association with Type 1 diabetes reported in the British Medical Journal in 1990, but subsequently there were few follow-up studies. It is with this background that Lofvenborg and colleagues report in the current issue of Diabetic Medicine (page 799) that coffee consumption might affect the risk of latent autoimmune diabetes in adults (LADA). The investigators used a strong study design studying incident diabetes in the Epidemiology Study of Risk Factors for LADA and Type 2 diabetes (ESTRID)—a population-based study from Sweden. As expected, an inverse relationship of coffee consumption and Type 2 diabetes (759 cases) was confirmed [odds ratio 0.92 (0.87–0.98)]. With regard to LADA (254 cases), the findings were not significant [odds ratio 1.04 (0.96– 1.13)], although clearly limited by the comparatively lower number of incident cases. The investigators did find a positive elationship between coffee consumption and autoantibody [glutamic acid decarboxylase (GAD)] titre with a dose effect. Whilst an intriguing relationship, this study is likely to be underpowered, limiting conclusions with regard to LADA. Nonetheless, I have chosen this article as editor’s choice for this month as it clearly demands to be studied further and, unlike Type 1 diabetes, not left another 24 years.

​G. A. Hitman, Editor-in-Chief Diabetic Medicine