Coffee consumption and health: we need randomised controlled trials

The umbrella review of coffee consumption and health by Poole and colleagues includes a gentle plea to consider a randomised controlled trial of coffee for chronic liver disease as well as the outcomes of a focus group, which indicate interest in this particular usage.1 I hope that such a trial does result from this publication.

I was able to find one small (n=40) randomised controlled trial of coffee—a 30 day crossover study in patients with hepatitis C, which confirmed its beneficial effects on telomere length, 8-hydroxydeoxyguanosine, aspartate transaminase, alanine transaminase, and collagen, despite high viral load in coffee drinkers at baseline.2

Poole and colleagues do not include a meta-analysis of the many studies showing an association between coffee drinking and lower aspartate transaminase, alanine transaminase, and γ glutamyltransferase.3

The accompanying editorial says that coffee is often consumed with sugar and unhealthy fats, without defining the latter.4 But saturated fats, such as beef tallow and cocoa butter, are strongly hepatoprotective in animal models of alcoholic liver disease, whereas polyunsaturated fats, such as corn oil and soy oil, induce disease progression.5 A trial of this effect in humans would also be very welcome, although it may be extremely difficult to carry out. In the randomised controlled trials cited by Poole and colleagues, coffee caused increased low density lipoprotein cholesterol concentrations, yet it was associated with reduced rates of all cardiovascular outcomes in the prospective cohort studies.1 This is a similar pattern to the relation between saturated fat, low density lipoprotein cholesterol, and cardiovascular disease in the PURE study, and shows that low density lipoprotein cholesterol does not reliably inform dietary effects on health in the same way that it informs statin efficacy.6

So having cream in one’s coffee may not be a bad habit after all, but we should probably hold the cake.