The effect of daily caffeine use on cerebral blood flow: How much caffeine can we tolerate?

Abstract

Caffeine is a commonly used neurostimulant that also produces cerebral vasoconstriction by antagonizing adenosine receptors. Chronic caffeine use results in an adaptation of the vascular adenosine receptor system presumably to compensate for the vasoconstrictive effects of caffeine. We investigated the effects of caffeine on cerebral blood flow (CBF) in increasing levels of chronic caffeine use. Low (mean = 45 mg/day), moderate (mean = 405 mg/day), and high (mean = 950 mg/day) caffeine users underwent quantitative perfusion magnetic resonance imaging on four separate occasions: twice in a caffeine abstinent state (abstained state) and twice in a caffeinated state following their normal caffeine use (native state). In each state, there were two drug conditions: participants received either caffeine (250 mg) or placebo. Gray matter CBF was tested with repeated-measures analysis of variance using caffeine use as a between-subjects factor, and correlational analyses were conducted between CBF and caffeine use. Caffeine reduced CBF by an average of 27% across both caffeine states. In the abstained placebo condition, moderate and high users had similarly greater CBF than low users; but in the native placebo condition, the high users had a trend towards less CBF than the low and moderate users. Our results suggest a limited ability of the cerebrovascular adenosine system to compensate for high amounts of daily caffeine use.

Figure 1

Gray matter CBF means for placebo and caffeine conditions in the abstained and native caffeinated states among all subjects. Caffeine, relative to placebo, reduced CBF in both states within each caffeine use group. High users had greater CBF than low users in the abstained placebo condition (p < 0.05), and there was greater CBF in the abstained placebo condition than in the native placebo condition for the moderate and high (p < 0.05), but not the low users. The repeated‐measures ANOVA interaction between drug condition, caffeine state, and caffeine use was significant (p < 0.005). Error bars are SEM.

Figure 2

Shown are the differences in CBF between drug conditions in each caffeine state to explore the repeated‐measures ANOVA interaction between caffeine use and caffeine state. In the abstained state (AP‐AC), caffeine reduced CBF equally among the caffeine use groups. In the native state (NP‐NC), caffeine reduced CBF more among the low users than the high users (p < 0.005). Reductions in CBF following caffeine were smaller among the moderate and high users in the native state compared to the abstained state (p < 0.001). Error bars are SEM.

Figure 3

Average CBF maps (at MNI coordinates z = 14) of the low, moderate, and high caffeine users in the native placebo and native caffeine conditions. The difference in CBF following caffeine was greatest among the low users, followed by the moderate users, and the smallest difference was among the high users.

Figure 4

Correlation between the reduction in CBF following caffeine in the native state (NP‐NC) and caffeine use. Caffeine produced smaller reductions in CBF as daily caffeine use increased (r = −0.48, p < 0.005).

Figure 5

Shown are the differences in CBF between caffeine states in each drug condition to explore the repeated‐measures ANOVA interaction between caffeine use and caffeine state. Differences in CBF between the abstained and native state are apparent in the placebo condition (AP‐NP). The difference in CBF in the placebo condition was greater among the high users than the low users (p < 0.001), and the difference in the placebo condition was greater than the difference in the caffeine condition for the moderate and high users (p < 0.001). Error bars are SEM.

Figure 6

Average CBF maps (at MNI coordinates z = 14) of the low, moderate, and high caffeine users in the abstained placebo and native placebo conditions. Although there is virtually no difference in CBF between the two states among the low users, increased CBF following abstention is apparent among the moderate and high users.

Figure 7

Correlation between the reduction in CBF from the abstained state to the native state (AP‐NP) and caffeine use. The difference in CBF increased as daily caffeine use increased (r = 0.62, p < 0.001).