REF-1654 · Han et al., 2024 · Nutrition Reviews, · Paper · PubMed

Drinking coffee daily for 2-24 weeks did not change resting heart rate: +0.40 beats per minute (95% CI −0.78 to 1.57; P = 0.506; I² = 0%) across 6 randomised trials with 11 arms and 485 people, at 3-6 cups a day

Main caveat: Six trials of generally healthy people, some with hypertension, hypercholesterolaemia or overweight; their sex, age and fitness are not in the abstract.
moderate evidenceMeasured on a mixed groupRead from the abstract only

Han and colleagues asked whether regular coffee changes resting heart rate, which matters because resting heart rate is an indicator of disease and mortality, for individual coffee drinkers and for public health.

They searched PubMed, Web of Science and the Cochrane Library for randomised trials published before March 2023, followed PRISMA, assessed risk of bias with Cochrane methods and tested for publication bias.

Six RCTs with 11 intervention arms and 485 participants qualified; participants were generally healthy, though some had hypertension, hypercholesterolaemia or overweight, and trials lasted 2 to 24 weeks.

Coffee raised resting heart rate by a negligible, non-significant 0.40 beats per minute (95% CI −0.78 to 1.57; P = 0.506), with no heterogeneity (I² = 0.0%), and every planned subgroup agreed. The authors conclude that 3 to 6 cups of coffee a day for 2 to 24 weeks has no significant effect on resting heart rate.

Participants

485 people in 6 RCTs (11 arms); generally healthy, some with hypertension, hypercholesterolaemia or overweight.
Sex, age and training status: not in the abstract.

Methods

Search: PubMed, Web of Science, Cochrane Library; RCTs of coffee and resting heart rate published before March 2023.
Extraction and risk of bias: Cochrane guidelines; PRISMA reporting.
Pooling: random- or fixed-effects models; heterogeneity by Cochran's Q; publication bias by Begg's test, Egger's test and funnel plots.
Not in the abstract: the risk-of-bias judgements, the publication-bias results, the subgroup categories, the control conditions, and when resting heart rate was measured relative to the last cup.

Key Results

OutcomeCoffee vs control95% CIPHeterogeneity
Resting heart rate+0.40 beats/min−0.78 to 1.570.506I² = 0.0% (heterogeneity P = 0.756)
Trials: 6 RCTs, 11 arms, 485 people, 2 to 24 weeks, 3 to 6 cups a day.
All subgroup analyses agreed with the overall result.

Practical Application

A daily coffee habit does not explain a raised resting heart rate. Weeks of 3-6 cups a day left resting heart rate unchanged (+0.40 beats/min, confidence interval spanning zero); when morning heart rate is up on a usual coffee day, look first at other causes, such as heat, illness, short sleep or recent load (PHYS-12).
The interval is narrow. Its upper limit was 1.57 beats/min. Inference, not tested here: an effect that small sits inside ordinary day-to-day variation, so a habitual coffee drinker need not adjust a resting heart-rate baseline for it.
Inference, not tested here: an unusual dose, or coffee taken shortly before the reading, is a different question from a stable habit, and this review does not answer it.
Not settled by this design / not from this paper: athletes; non-habitual drinkers; heart-rate variability; heart rate during exercise (REF-1652); when the readings were taken relative to the last cup.

Limitations

Six trials of generally healthy people, some with hypertension, hypercholesterolaemia or overweight; their sex, age and fitness are not in the abstract.
The risk-of-bias judgements, the subgroup categories and the timing of the heart-rate measurement are not in the abstract.
Chronic intake only; the acute effect of a cup in the hour before a reading was not isolated.
Written from the abstract and the PubMed record; full text not read.
How this entry was checked and changed
Why this level

risk of bias not read (abstract only: the abstract says risk of bias was assessed by Cochrane methods but does not report the judgements); 6 trials with no heterogeneity (I² = 0%); resting heart rate is the subject of this entry, so no marker step

Content correction 2026-11-07
2026-11-07: added from the abstract and the PubMed record; level B+ (meta-analysis of RCTs, risk of bias not read; abstract only).