Randomized Controlled Trial on Acute Effects of Coffee Consumption

Drinking coffee has been linked to an assortment of health benefits from lower rates of diabetes (QH) to reduced risk of Parkinson’s Disease to living longer (JHU). There have been numerous studies providing evidence for these benefits but nearly all have been observational studies which raises the possibility of confounders that muddy any causal relationships between coffee drinking and health outcomes. Indeed some have argued that reverse causation is plausible with healthier individuals being more likely to drink coffee. 

One way to resolve this issue is to employ the gold standard of clinical trials the randomized controlled trial (RCT) in which one group would drink coffee and one group would not, and then monitor health variables. But how could one run such a trial over an extended period of time forcing some coffee drinkers not to drink coffee for say years while forcing some non-coffee drinkers to drink coffee?

A group of researcher got around this problem by running a very short trial to examine the acute effects of coffee consumption on various heart arrhythmias, physical activity, sleep, and glucose levels. The study enrolled 100 healthy adult participants between the ages of 18 and 65 years. The participants were randomly assigned to consume either caffeinated coffee or avoid caffeine for 14 days while using continuously recording, wearable sensors. 

The study design was a case-crossover RCT meaning that each individual was both case and control (i.e. had coffee some days but not others). Participants were randomly assigned to consume caffeinated coffee or to avoid caffeine during 2-day periods. The sequence of these periods was arranged in such a way that no participant had to consume or abstain from coffee for more than two consecutive days. Such a design is possible when measuring acute (short-term) effects on a daily basis.

The participants were asked to complete a daily survey about their caffeine intake, step counts, sleep minutes, and serum glucose levels. They also wore a continuously recording electrocardiogram device to measure heart rhythm, and a wrist-worn accelerometer to count steps. Adherence to the program was deemed to be high.

The primary outcome of the study was the number of premature atrial contractions per 24-hour period. The secondary outcomes were the daily number of premature ventricular contractions, daily number of episodes of supraventricular tachycardia or ventricular tachycardia, daily step counts, daily minutes of sleep, and mean daily serum glucose levels.

Premature atrial contractions (PACs) are a type of heart arrhythmia but milder than the more well-known atrial fibrillations. As a reminder, atrial fibrillation (AFib) is a type of heart arrhythmia characterized by an irregular and often rapid heart rate. In AFib, the heart's two upper chambers (atria) quiver or fibrillate instead of contracting properly, which can lead to an inefficient pumping of blood through the heart. This irregular rhythm can cause blood to pool in the atria, which increases the risk of blood clots and stroke.

PACs are extra heartbeats that also start in the atria. These extra beats disrupt your regular heart rhythm. The big difference from AFib is that PACs are common and usually harmless; the extra beats do not disrupt heart function in the same way as atrial fibrillation. They can happen to anyone, but are more common in people over age 65. PACs are often caused by stress, caffeine, alcohol, or smoking. 

One secondary outcome examined in the study was premature ventricular contractions (PVCs), which are the ventricular analog of PACs, i.e. extra heartbeats originating in the ventricles. They are also common and harmless.

The other secondary outcomes, and perhaps of greater interest in understanding the long-term impact of coffee on health, were the number of steps, time sleeping, and blood glucose. These are possibly risk factors for disorders like type 2 diabetes that drinking coffee has been shown to provide benefits.

The results of the study were as follows (NEJM):

"The consumption of caffeinated coffee was associated with 58 daily premature atrial contractions as compared with 53 daily events on days when caffeine was avoided (rate ratio, 1.09; 95% confidence interval [CI], 0.98 to 1.20; P=0.10). The consumption of caffeinated coffee as compared with no caffeine consumption was associated with 154 and 102 daily premature ventricular contractions, respectively (rate ratio, 1.51; 95% CI, 1.18 to 1.94); 10,646 and 9665 daily steps (mean difference, 1058; 95% CI, 441 to 1675); 397 and 432 minutes of nightly sleep (mean difference, 36; 95% CI, 25 to 47); and serum glucose levels of 95 mg per deciliter and 96 mg per deciliter (mean difference, −0.41; 95% CI, −5.42 to 4.60)."

Thus for the primary outcome there was a slight increase in daily PACs (rate ratio = 1.09, P=0.10) for coffee consumption days which was not statistically significant (Table 1). There was also a modest increase in PVCs among coffee drinkers (rate ratio = 1.51). Perhaps of greater interest was the increase in the daily steps by about 1000 (10,646 coffee versus 9665 no coffee), and a 36 minute difference in sleep with drinking coffee leading to less sleep. There was no difference in serum glucose, and none of the secondary outcomes were statistically significant. 

The roughly 10% increase in step count hints at a possible positive effect from the stimulant caffeine on activity level, but a larger and longer trial is needed to see if step increase is maintained over time and whether there is a discernible impact on physiologic variables such as blood glucose or weight on a longer time scale, i.e. perhaps serum glucose did not change because 14 days was too short. In addition, the researchers did not regulate the number of cups of coffee on the days in which coffee consumption was permitted, and that is another factor that may need to be controlled in a future trial. In general, more cups of coffee led to the same directional changes in outcomes of coffee versus no coffee.

Table 1. Daily frequency of arrhythmia. Daily mean number of arrhythmia events adjusted for day of the week. Treatment effects are expressed as rate or odds ratios (from Table 2 of Marcus et al. NEJM, 2023).