Coffee consumption and coronary heart disease: paradoxical effects on biological risk factors versus disease incidence.

Coffee consumption has been shown to have adverse effects on various biological markers of coronary heart disease (CHD) risk, including serum cholesterol (1 ), blood pressure (2 ), insulin resistance (3 ), and plasma homocysteine (4 ). In contrast, higher coffee consumption has not been associated with a higher risk of CHD in prospective cohort studies (5–7 ). What could be the explanation for this “coffee paradox”? First, the acute effects of coffee consumption can be different from the effects of long-term habitual consumption. Second, the physiological effects of coffee can depend on the type of coffee consumed and are not necessarily the same as for caffeine in isolation. Third, coffee consumption may have beneficial effects on other biological pathways implicated in the development of CHD that could compensate for any adverse effects. Fourth, risk markers may not causally affect the development of CHD, or their effects may be too modest for any increase caused by coffee consumption to translate into a substantial increase in disease risk. Research supporting these explanations are discussed, with particular focus on the association between coffee consumption and homocysteine concentrations reported by Ulvik et al. in this issue of Clinical Chemistry (8 ). In caffeine-naive individuals, caffeine intake leads to a marked reduction in insulin sensitivity (3 ) and marked increases in postload glucose concentrations (9 ), epinephrine concentrations (10 ), and blood pressure (10 ). Within a week of initiating caffeine intake, however, an individual experiences an attenuation of coffee’s effects on epinephrine release and blood pressure (10 ). For blood pressure, only a partial tolerance develops, and a modest increase remains after several weeks (2 ). Given that caffeine’s stimulation of epinephrine release appears to contribute to insulin resistance (3 ), it is plausible that the effect of caffeine on glucose metabolism would also attenuate after continued caffeine intake. There is limited evidence that the increase in insulin concentration caused by high caffeine intake remains after 1– 4 weeks (11, 12 ). In contrast, the results of prospective cohort studies suggest that long-term caffeinated coffee consumption is not substantially associated with a risk of hypertension (13 ) and may lower the risk of type 2 diabetes (14 ). Thus, partial tolerance to the effects of caffeine can develop, but longer-term trials on the effects of coffee on biological risk factors are needed to bridge the gap in the data between short-term trials and cohort studies. With regard to the type of coffee, the effect of coffee consumption on serum cholesterol depends on the brewing method. In randomized trials, high consumption of boiled coffee led to a substantial increase in LDL cholesterol, whereas high consumption of paper-filtered coffee had little effect (1). The diterpene cafestol has been identified as the cholesterol-raising component of coffee (15). For paper-filtered coffee, the filter prevents substantial amounts of cafestol from getting into the coffee, whereas cafestol remains in boiled, French press, and Turkish/Greek coffee (15). Coffee is a complex mixture of hundreds of plant compounds that may interact in their physiological effects. Most research on coffee components has focused on caffeine, and it is tempting to directly extrapolate the physiological effects that have been found for caffeine in isolation to those of coffee. However, the effects of coffee on exercise performance (16), epinephrine concentrations (16), blood pressure (2), and hyperglycemia (9) all appear to be weaker than the effects of the same amount of caffeine used in isolation. Recently, the acute effects of caffeine, caffeinated coffee, and decaffeinated coffee on responses to an oral glucose tolerance test were evaluated (9). Compared with placebo, caffeine increased postload glucose concentrations, caffeinated coffee showed a nonsignificant increase, and decaffeinated coffee produced a significant decrease in glucose concentrations. Similarly, trials of caffeinated coffee tended to show weaker effects on blood pressure than trials of caffeine in isolation (2), and the coffee phenol chlorogenic acid may reduce blood pressure (17). In line with these results, the consumption of other caffeinecontaining drinks, but not caffeinated coffee, was associated with a higher risk of developing hypertension (13 ). These findings suggest that components in coffee other than caffeine have physiological effects that are opposite to those of caffeine. For example, quinides may counteract the adenosine receptor antagonism caused by caffeine (18 ). Ulvik et al. report an association between higher coffee consumption and higher plasma homocysteine concentrations in a large population-based crossClinical Chemistry 54:9 1418–1420 (2008) Editorial