Does Caffeine Still Work If You Drink Coffee Every Day? (2025 Meta-Analysis)

Short answer: yes, but not nearly as much. A 2025 meta-analysis of 12 studies and 230 trained lifters found caffeine improved mean bar velocity by an SMD of 0.42. Split by how much caffeine people already drank daily, the gap was brutal: under 3 mg/kg/day scored 0.87, above that scored 0.21. Same supplement, roughly a quarter of the payoff.

Key findings at a glance

• Caffeine improved mean bar velocity (SMD 0.42) and mean power output (SMD 0.21) during resistance training.
• Lifters with habitual intake under 3 mg/kg/day responded 4x better than heavier daily users (0.87 vs 0.21, p < 0.01).
• Men responded more than women (0.56 vs 0.22, p < 0.01).
• Higher doses did not reliably produce bigger effects. Dose subgroup differences were not statistically significant.
• 3 mg/kg/day is roughly 245 mg for a 180 lb lifter, about two cups of coffee.

Source: Xiao et al., Frontiers in Nutrition, October 2025 (PMID 41127092). Reviewed and published September 2026 by Danny Lehr, CrossFit Level 3 Trainer and USA Weightlifting coach.

What did the study actually measure?

Xiao and colleagues pooled 12 studies covering 230 participants, published in Frontiers in Nutrition in October 2025. They looked at movement velocity and power output during resistance training with a fixed number of reps, then ran subgroup analyses on sex, caffeine dose, habitual caffeine intake, muscle group, and load.

Two headline numbers:

• Mean velocity: SMD 0.42 (95% CI 0.19 to 0.65)
• Mean power output: SMD 0.21 (95% CI 0.12 to 0.30)

Both were statistically significant. Caffeine moves the bar faster. That part isn't in question.

Why does habitual caffeine intake matter so much?

This is the finding worth your attention. Of everything the researchers tested, habitual intake produced one of only two splits that reached statistical significance.

Subgroup	Mean velocity SMD	Significant difference?
Habitual intake under 3 mg/kg/day	0.87	Yes, p < 0.01
Habitual intake 3 mg/kg/day or more	0.21
Males	0.56	Yes, p < 0.01
Females	0.22
Across caffeine doses	0.31 to 0.78	No
Across muscle groups	0.32 to 0.54	No
Across loads	0.37 to 0.49	No

The authors put it this way in their own conclusion: These benefits were more pronounced in males, at higher caffeine doses, among low habitual caffeine consumers, and during lower-body exercises. Note that only two of those four actually cleared significance in the results section.

For a 180 lb lifter, 3 mg/kg/day is about 245 mg, or roughly two decent cups of coffee. Plenty of people reading this are well past that before they get to the gym.

Should I quit caffeine to get the effect back?

No, and that's not what the data says. The high-habitual group still showed a positive effect. It was smaller, not absent. Nothing in this analysis tested what happens when a daily drinker quits and comes back, because no such trial was included. That's a hypothesis people like to sell you, not a finding.

The defensible read is simpler: if you already drink coffee daily, your pre-workout caffeine is doing less than the label promises. Plan around that instead of pretending otherwise.

Does taking more caffeine fix it?

The study can't say it does. Effect sizes across dose buckets ranged from 0.31 to 0.78, but the differences between dose subgroups were not statistically significant. That's a real limitation and it's the part most write-ups skip. Anyone pointing at the top of that range as proof that megadosing works is reading a spread as a result.

Same story for muscle group and load. Lower body scored higher than upper body, heavy scored close to light, and none of those gaps cleared significance.

What should a lifter actually do with this?

• Know your baseline. Add up what you actually drink in a day. Most people underestimate it.
• Time it, don't just stack it. A cup you drink at 6am isn't a pre-workout at 5pm.
• Judge by the bar, not the buzz. Velocity is measurable. Feeling wired isn't the same as moving faster.
• Don't chase the dose. The evidence for more being better didn't hold up statistically here.

If you want coffee built for the session rather than the habit, PR Blend runs about 50% more caffeine than a standard cup. The rest of the lineup is in our coffee collection, and the Coffee of the Month plan sends a different roast every month.

Frequently asked questions

How much is 3 mg/kg of caffeine?
Around 200 mg for a 150 lb person and about 245 mg for a 180 lb person. A typical 8 oz cup of drip coffee runs 95 to 165 mg.

Does caffeine help heavy lifts more than light ones?
The effect showed up at every load tested, 0.37 to 0.49, with no significant difference between them.

Why did men respond more than women?
The analysis found the split (0.56 versus 0.22) but didn't establish a mechanism. Sample sizes for women were smaller across the included trials.

Is a 0.42 effect size big?
It's conventionally read as small to moderate. Real, worth having, not transformative.

Should I cycle off caffeine before a competition?
This study can't answer that. It never tested withdrawal or reintroduction, so anyone citing it for a caffeine cycling protocol is going beyond the evidence.

What this does not prove

Subgroup comparisons pool across different studies with different designs. They are not randomized head to head tests, so the 0.87 versus 0.21 split is a strong signal rather than a proven cause. Heterogeneity on the velocity outcome was high (I² = 85%), meaning the individual studies disagreed with each other a fair amount. Total sample was 230 people. And nothing here tested deloading, cycling off, or tolerance reversal.

Sources

1. Xiao Y, Ding L, Xu Z, Liu J, Guo L, Barnes MJ, Cao Y, Girard O. Effects of acute caffeine intake on muscular power during resistance exercise: a systematic review and meta-analysis. Front Nutr. 2025 Oct 7;12:1686283. PMID 41127092
2. Raya-Gonzalez J, et al. Acute Effects of Caffeine Supplementation on Movement Velocity in Resistance Exercise: A Systematic Review and Meta-analysis. Sports Med. 2020;50(4):717-729. PMID 31643020