Caffeine is a very popular ergogenic aid. It’s been a staple in multi-ingredient pre-workout formulas since they hit the market, and it’s been shown to improve many different types of exercise performance, including endurance, strength, and power (2). However, caffeine is also very popular for non-exercise applications, with some research suggesting that up to 85% of the US population consumes at least one caffeinated beverage per day (3). Any regular caffeine user can tell you that some of caffeine’s effects diminish over time as we become habituated to regular consumption, which leads to important questions: do performance improvements fade away over time? Do we need to cycle on and off caffeine in order to resensitize our receptors and restore caffeine’s ergogenic effects? The answers to these questions are surprisingly unclear, despite the large body of performance-related caffeine research, and they are questions we’ve repeatedly revisited in our research review MASS (one, two, three, four).
A recent meta-analysis (1) sought to determine if habitual caffeine consumption influences the acute ergogenic effect of caffeine. The researchers systematically searched the literature for studies that met the following criteria:
- Included healthy males or females, with no restrictions on age or training status, within a randomized single-blind or double-blind study design.
- Quantified habitual caffeine consumption of participants in mg/kg/day, or provided sufficient data for the meta-analysts to calculate habitual intakes in these units.
- Assessed the acute ergogenic effect of caffeine (at any dose and in any form) consumed before an exercise task, with a direct comparison to a placebo group or placebo condition.
Once the literature search was complete and the researchers excluded all of the studies that failed to meet their criteria, they were left with 60 studies. Collectively, these studies included 1,137 total participants; 958 were males and 179 were females; 718 were “trained,” 400 were “untrained,” and 19 were “elite.” At that point, they went through to extract a ton of data for comparison purposes. For example, they went through each study and noted the age, weight, sex, training status, and habitual caffeine consumption of participants, the characteristics of the exercise task, the supplementation protocol, and the primary exercise outcome tested.
The researchers used meta-regression to determine if habitual caffeine intake was predictive of caffeine’s effect size for performance outcomes. The primary model indicated that caffeine had a statistically significant ergogenic effect (effect size [ES], expressed as Hedges’ g, = 0.25)
Figure 1 shows the results of their comparisons. In short, the results indicated that caffeine had a small but positive effect on endurance, power, and strength outcomes. However, habitual caffeine intake (in mg/kg/day) did not significantly influence the observed effect size (p = 0.59). The researchers dug deeper to explore a number of other comparisons, which are presented in Figure 1. In short, caffeine was ergogenic across different exercise types (endurance, strength, and power), training statuses (trained or untrained), and sexes (male or female), with fairly similar effect sizes across these categories, and no significant influence of habitual caffeine intake within these comparisons. Acute caffeine was effective whether the experimental dose was higher or lower than the habitual intake of study participants, and the ergogenic effect was not meaningfully influenced by the duration of the caffeine withdrawal periods prior to the caffeine intervention. The only thing that seemed to threaten the ergogenic effect of caffeine was the dose ingested; doses below 3mg/kg and between 3-6 mg/kg were similarly effective, but doses above 6mg/kg led to a smaller pooled effect size (which was not statistically significant) and a wider confidence interval. This lines up pretty well with experimental evidence showing that higher caffeine doses (>6mg/kg) are more likely to induce uncomfortable side effects which may impair average performance at the group level and increase the inter-individual variability of performance responses (4).
Since we’ve visited this topic several times in MASS Research Review issues (one, two, three, four), I’ll focus on the highlights and conclusions from this literature instead of retreading the study-by-study trajectory of this research topic. I think this is a really useful meta-analysis, because it confirms something that’s been observed many times in isolated studies: when you evaluate the acute ergogenic effect of caffeine, it tends to work for self-reported habitual caffeine consumers and non-consumers alike, and the effect sizes tend to be pretty similar when compared directly. This meta-analysis also provides several additional comparisons that have a ton of practical utility – it’s very helpful to see head-to-head comparisons of effect sizes for different sexes, exercise types, training statuses, dosing ranges, and so on. At this point, I think we can make a few conclusions with a reasonable degree of confidence: caffeine still works for habitual caffeine users, caffeine works across a broad range of exercises, caffeine works across different sexes and training statuses, and your “best bet” for dosing is probably in the range of 3-6mg/kg.
Having said that, I do want to highlight a few considerations to prevent overconfidence in some of these conclusions. A meta-analysis can only enable inferences based on the experimental data available, and there are some shortcomings in the caffeine literature. First, the literature skews heavily toward male subjects (in this meta-analysis, 84% of the data came from male participants), which isn’t rare for supplement research. All signs point to similar effects between males and females, but we should remember that we’re working with a limited set of female data. Speaking of limited data, the presently reviewed meta-analysis noted that only 24% of the studies in their search reported the mean habitual caffeine consumption of their participants. I still feel pretty confident that 60 studies give us plenty of data to draw from, but we should acknowledge that these findings come from less than a quarter of the relevant caffeine research to date, and may not be perfectly representative of the literature as a whole. Finally, and most importantly (by far): the randomized controlled trial is a very powerful tool, and we should use it whenever possible.
Most of our knowledge about this caffeine habituation question comes from research that is observational in nature. Sure, they are employing randomized, placebo-controlled methods when determining the acute effect of caffeine intake, but the determination of habitual caffeine intake is based on self-reported habits rather than experimental manipulation of caffeine habituation status. The presently reviewed meta-analysis provides our current “best guess” about the impact of habituation on the ergogenic effect of caffeine, but there are two major limitations of the underlying literature. First, we’re looking at habitual caffeine intake from each study at the group level, not the individual level. This meta-analysis is not strictly comparing the effects of caffeine among individuals who habitually consume large amounts versus small amounts of caffeine; rather, it’s comparing average, group-level effects of caffeine among groups of people with high versus low average daily intakes of caffeine. This approach should generally point us in the right direction, but there’s absolutely no question that there are individuals who consume very minimal caffeine who have been grouped into samples that are categorized, on average, as high caffeine consumers. Second, the time course of caffeine habituation (and re-sensitization) isn’t fully understood with a tremendous amount of detail, but appears to occur over periods of days rather than months. Let’s say you’re participating in a caffeine study and the research team asks how much caffeine you typically consume. You typically consume about 5 cups of coffee per day, but you haven’t been consuming much over the last 10-14 days due to changes in your schedule (because of work, social events, family obligations, and so on). So, how does a researcher classify your habitual intake? You’re generally consuming a pretty high level of daily caffeine intake, but you might not be particularly habituated at the time of testing. Individual studies don’t typically go into a lot of detail about this consideration, but it’s a pretty big deal when it comes to nuanced interpretation.
The reason I focus so much on the observational nature of categorizing habitual caffeine consumption is because Greg reviewed one of the only studies (to my knowledge) that actually addressed the question of caffeine habituation and exercise performance from a truly experimental perspective. In the study (5), Lara and colleagues studied the ergogenic effect of pre-exercise caffeine supplementation (3mg/kg) over a 20-day period. The results generally suggested that the magnitude of caffeine’s ergogenic effect decreased with repeated use. However, the relative degree of effect size reduction varied among the different performance outcomes measured, and an effect size getting smaller is not the same as an effect size disappearing entirely. It’s also unclear if the effect sizes were on a trajectory involving continuous decreases over time, or if the effect size reductions had effectively plateaued at a lower (but still non-zero) magnitude of performance enhancement. Looking at the broader literature, I have no hesitation when stating that habitual caffeine users can still enjoy an ergogenic effect from acute caffeine use. It also seems that effect sizes are pretty similar when comparing self-reported heavy caffeine users to self-reported light caffeine users. However, the limited experimental evidence available causes me to hesitate a little bit when suggesting that there is absolutely no attenuation of caffeine’s ergogenic effects when it is consumed habitually. I wish there were more randomized controlled trials to sort out the discrepancy, but I find it difficult to unequivocally accept the conclusions from observational findings while entirely ignoring the small amount of experimental data available. We never want to place a disproportionate amount of confidence in a single study, but we also have to reconcile experimental findings with observational findings when both are available, and experimental data generally tend to warrant a heightened level of consideration when compared to observational data.
In summary, I’m quite confident that habitual caffeine consumers are still able to attain ergogenic benefits from acute caffeine consumption. However, in order to develop a more nuanced understanding of the time course by which caffeine habituation develops and reverses, and how performance varies across these processes, we probably need to lean on the good old-fashioned experiment. I’d love to see more randomized controlled trials that explore this question directly, and they are feasible studies to complete. Caffeine has an excellent safety profile within ergogenic dosing ranges and is an extremely affordable study ingredient. The biggest logistical challenge would be finding participants who are willing to come into the lab for very frequent performance testing, and selecting a performance outcome that is simultaneously sensitive enough to capture small changes in neuromuscular performance, easy enough to avoid soreness or performance decrements when completed frequently, but familiarized well enough to avoid time-related performance improvements over the course of the study. I think there’s a very intuitive series of studies on caffeine tolerance, habituation, and withdrawal time courses that would make for a great PhD dissertation or thesis project, and if I were entering a PhD program today (a daunting prospect that I’d prefer not to entertain, even hypothetically), it’s definitely the route I would pursue. Until more randomized controlled trials with experimental manipulation of caffeine habituation status become available, the current evidence suggests that it probably isn’t a huge deal for performance outcomes, and that caffeine is still ergogenic (to a fairly similar degree) when comparing heavy caffeine users to people who consume minimal caffeine.
Note: This article was published in partnership with MASS Research Review. Full versions of Research Spotlight breakdowns are originally published in MASS Research Review. Subscribe to MASS to get a monthly publication with breakdowns of recent exercise and nutrition studies.
References
- Carvalho A, Marticorena FM, Grecco BH, Barreto G, Saunders B. Can I Have My Coffee and Drink It? A Systematic Review and Meta-analysis to Determine Whether Habitual Caffeine Consumption Affects the Ergogenic Effect of Caffeine. Sports Med. 2022 May 10; ePub ahead of print.
- Grgic J, Grgic I, Pickering C, Schoenfeld BJ, Bishop DJ, Pedisic Z. Wake Up And Smell The Coffee: Caffeine Supplementation And Exercise Performance-An Umbrella Review Of 21 Published Meta-Analyses. Br J Sports Med. 2020 Jun;54(11):681-688.
- Mitchell DC, Knight CA, Hockenberry J, Teplansky R, Hartman TJ. Beverage Caffeine Intakes In The U.S. Food Chem Toxicol. 2014 Jan;63:136–42.
- Guest NS, VanDusseldorp TA, Nelson MT, Grgic J, Schoenfeld BJ, Jenkins NDM, et al. International Society Of Sports Nutrition Position Stand: Caffeine And Exercise Performance. J Int Soc Sports Nutr. 2021 Jan 2;18(1):1.
- Lara B, Ruiz-Moreno C, Salinero JJ, Del Coso J. Time Course Of Tolerance To The Performance Benefits Of Caffeine. PloS One. 2019;14(1):e0210275.
