Should You Have Coffee Before Zone 2 Training? An Honest Update

TL;DR

An earlier version of this article said that caffeine blunts the AMPK signal that drives mitochondrial adaptation from Zone 2 training. I've gone back to the literature, and that claim doesn't hold up. At the doses you get from coffee, caffeine has no demonstrated effect on the mitochondrial adaptation signal in human muscle — in either direction. There are still three good reasons to be deliberate about it: caffeine distorts the heart-rate data you use to track progress, a recent tendon study raises a question worth respecting if you're rehabbing an Achilles, and — the one with the strongest human evidence — caffeine's long half-life means a post-ride coffee can still be cutting into deep sleep that night, and sleep is where adaptation is actually expressed. Here's the corrected picture, with references, and a protocol that reflects what the evidence actually supports.

A correction, up front

In April I published a version of this article built on a single mechanistic claim: that caffeine interferes with AMPK, the cellular energy sensor that helps trigger mitochondrial biogenesis, and that pre-ride coffee therefore quietly dampens the adaptation you're chasing from easy aerobic work.

I stated that as fact without a citation. When I went looking for the primary evidence, it wasn't there. What I found instead was more interesting than the original claim, and it changes the advice. Rather than quietly editing the old post, I'd prefer to show my working.

What Zone 2 is actually doing

Zone 2 sits at roughly 60–70% of heart-rate reserve — comfortable, conversational, and deceptively productive. Its signature adaptation is mitochondrial biogenesis: more and better mitochondria in the slow-twitch fibres that carry you through long efforts, alongside capillary growth and a greater capacity to use fat as fuel.

The molecular trigger is well mapped. Contraction raises intracellular calcium and shifts the cell's energy state. Those signals converge on a handful of kinases — CaMK, AMPK, p38 MAPK — that switch on PGC-1α, the transcriptional co-activator that coordinates the build (Egan & Zierath, 2013; Hood, 2019). AMPK is one input among several, not the master switch.

Where the AMPK claim came from, and why it fails

Caffeine does interact with AMPK — but not the way the original article said, and not at doses you can drink.

In isolated rat muscle and cultured muscle cells, caffeine activates AMPK, not inhibits it, by triggering calcium release from the sarcoplasmic reticulum (Jensen et al., 2007; Egawa et al., 2009, 2011; Tsuda et al., 2015). So the direction of the original claim was wrong.

But the more important point is dose. Those experiments used caffeine at around 3 millimoles per litre. A 3–6 mg/kg dose in humans — a strong coffee or two — produces plasma concentrations of roughly 20–80 micromoles per litre, about 40 to 150 times lower (Graham & Spriet, 1995). A 2026 review in Frontiers in Physiology that examined this question specifically concludes that at human doses, adenosine receptor antagonism is the only well-supported action, and any AMPK effect "is more likely to be indirect" and unproven in exercising human muscle (Liu et al., 2026).

The honest summary: caffeine at coffee doses has no demonstrated effect on the Zone 2 mitochondrial signal, positive or negative. The original claim was an inversion of a rodent finding that doesn't translate to a morning flat white in any case.

What the human evidence does show

The same review separates three questions that are easy to blur together.

Acute performance. Solid. Caffeine reliably improves endurance output by around 2–3% (Southward et al., 2018), mostly by reducing perceived effort through adenosine receptor blockade in the central nervous system.

Fuel selection. Caffeine raises circulating free fatty acids, and meta-analyses show a small increase in whole-body fat oxidation during exercise (Collado-Mateo et al., 2020; Conger et al., 2022). Inside working human muscle, however, caffeine does not measurably change carbohydrate or fat use (Graham et al., 2000). So if your goal is fat oxidation, caffeine is at worst neutral. It certainly isn't inhibitory.

Chronic adaptation. Unresolved. No human training study has measured mitochondrial content or respiration with and without pre-session caffeine. The available trials are short, mostly resistance-based, and lack muscle biopsies. The best-fitting model is that caffeine improves the quality of individual sessions without directly changing how muscle remodels between them (Liu et al., 2026).

The finding that actually earns caution: tendon

Here's what the original article missed, and it matters for anyone reading this in the context of Achilles rehab.

In late 2025, Keith Baar's laboratory at UC Davis published a study showing that caffeine reduced protein synthesis in cultured muscle and tendon cells, cut collagen content and tensile strength in engineered ligaments by 30–45%, and blunted exercise-induced muscle-mass gains in mice given continuous caffeine at a dose roughly equivalent to 5.7 mg/kg in humans (Steffen et al., 2025).

Three caveats. It's a single study. The cell concentrations were high. And the mice drank caffeine continuously, all day, not as a single pre-session dose — an exposure pattern the Frontiers review flags as not interchangeable with pre-workout use. The review classifies this as hypothesis-generating and calls for replication before it drives practice.

But tendon adapts slowly, collagen synthesis is the whole point of a loading program, and this is the one line of evidence that points toward attenuation in a tissue you care about during rehab. When the potential downside is real and the cost of moderating intake is small, that tilts the decision — even before replication.

The reason that was right all along: clean data

If you track heart-rate drift during Zone 2 — the rise in heart rate at fixed power over a session, a useful marker of aerobic fitness and recovery — caffeine is a significant confound. It elevates heart rate at any given workload, in a naive user by as much as 5–10 beats per minute, and it steepens the drift curve. In my own training log during a reconditioning block, sessions with caffeine averaged a drift rate of 3.4 bpm per 10 minutes against 1.7 without it. That's the difference between "adapting well" and "not adapting" on the same legs.

That confound doesn't affect the adaptation itself. It affects your ability to see it. During a reconditioning block, when you're using session-to-session heart-rate comparisons to decide whether to progress, that visibility has real value.

The route with the best human evidence: sleep

Adaptation happens during recovery, and the deepest part of recovery is slow-wave sleep. Growth hormone pulses, tissue protein synthesis, and glycogen restoration are all concentrated there. The Frontiers review identifies sleep-mediated recovery impairment as the most plausible indirect route by which repeated caffeine could reduce training adaptation — better supported in humans than any direct muscle mechanism (Liu et al., 2026).

The pharmacology here is unambiguous. Caffeine's half-life is typically 4–6 hours, but it ranges from about 2 to over 9 hours depending on liver CYP1A2 activity, so a "slow metaboliser" still has half a morning dose on board at bedtime (Graham & Spriet, 1995). Even in average metabolisers, 400 mg taken six hours before bed cut objectively measured total sleep by more than an hour, and participants didn't notice — their subjective sleep ratings barely moved (Drake et al., 2013). A 2023 meta-analysis put numbers on it: caffeine reduced total sleep time by about 45 minutes, sleep efficiency by 7%, and specifically suppressed slow-wave sleep, with a recommended cut-off of roughly nine hours before bed for a standard coffee and around thirteen hours for a pre-workout dose (Gardiner et al., 2023).

That reframes "coffee after the ride." If the ride finishes at 7 a.m., a post-ride coffee is well clear. If it finishes at 3 p.m., that coffee is landing inside the window that Drake's participants couldn't feel but their sleep recordings could.

One data point, and it's only one: since I stopped caffeine entirely earlier this year, my overnight deep-sleep readings improved markedly and stayed there. That's a single person with a wrist sensor, not evidence. But it's consistent with the controlled data, and it's the kind of change you can test on yourself in a fortnight.

The updated protocol

If you track heart-rate drift or power-at-heart-rate: no caffeine before those sessions. This isn't about adaptation; it's about keeping your measurements comparable. Decaf is fine.

If you're in a tendon rehab block: keep total daily caffeine moderate and avoid all-day continuous intake. The tendon signal comes from sustained high exposure, not a single morning coffee. A pre-session dose of 1–3 mg/kg on harder days is unlikely to be the problem; four coffees spread across the day, every day, is closer to what the mouse study modelled.

If you're doing Zone 2 for its own sake and not measuring drift: have the coffee if you want it. There is no evidence it costs you mitochondria.

Post-session coffee with carbohydrate is a good habit if the timing works. Caffeine co-ingested with carbohydrate after exercise enhances glycogen resynthesis (Pedersen et al., 2008), so an espresso with breakfast after a morning ride is a mild recovery aid. But the glycogen benefit is small and the sleep cost is not. Set your cut-off by bedtime, not by the ride: for most people that means no caffeine after roughly midday, and none at all after an afternoon session. If you use a wearable, watch deep-sleep minutes for two weeks with and without the afternoon coffee and let your own data decide.

For hard sessions and races: the 2–3% performance benefit is real. Rehearse the dose in training before relying on it on the day.

The bigger lesson

Mechanistic stories are seductive. "Caffeine blunts AMPK" sounds precise and is easy to build a protocol around. But a mechanism observed in a dish at a hundred times the achievable concentration isn't a mechanism that applies to you, and a claim about a pathway isn't a claim about outcomes. The question that matters is always: shown in what model, at what dose, and did it change what we actually care about?

The original protocol — water before Zone 2, coffee after — was a reasonable practice built on a wrong explanation. It's still a reasonable practice. It just deserves the right reasons.

Zone 2 work is one part of a well-structured Achilles rehab program, not the whole thing. The tendon itself needs progressive mechanical load to remodel. The Achilles Load Restoration™ program provides a 12-week loading framework across three phases. To find your Zone 2 range, use the heart rate zone calculator.

References

Collado-Mateo D, Lavín-Pérez AM, Merellano-Navarro E, Del Coso J. Effect of acute caffeine intake on the fat oxidation rate during exercise: a systematic review and meta-analysis. Nutrients. 2020;12(12):3603.

Conger SA, Tuthill LM, Millard-Stafford ML. Does caffeine increase fat metabolism? A systematic review and meta-analysis. Int J Sport Nutr Exerc Metab. 2023;33(2):112–120.

Drake C, Roehrs T, Shambroom J, Roth T. Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. J Clin Sleep Med. 2013;9(11):1195–1200.

Gardiner C, Weakley J, Burke LM, et al. The effect of caffeine on subsequent sleep: a systematic review and meta-analysis. Sleep Med Rev. 2023;69:101764.

Egan B, Zierath JR. Exercise metabolism and the molecular regulation of skeletal muscle adaptation. Cell Metab. 2013;17(2):162–184.

Egawa T, Hamada T, Kameda N, et al. Caffeine acutely activates 5′AMP-activated protein kinase and increases insulin-independent glucose transport in rat skeletal muscles. Metabolism. 2009;58(11):1609–1617.

Egawa T, Hamada T, Ma X, et al. Caffeine activates preferentially α1-isoform of 5′AMP-activated protein kinase in rat skeletal muscle. Acta Physiol. 2011;201(2):227–238.

Graham TE, Spriet LL. Metabolic, catecholamine, and exercise performance responses to various doses of caffeine. J Appl Physiol. 1995;78(3):867–874.

Graham TE, Helge JW, MacLean DA, Kiens B, Richter EA. Caffeine ingestion does not alter carbohydrate or fat metabolism in human skeletal muscle during exercise. J Physiol. 2000;529(3):837–847.

Hood DA, Memme JM, Oliveira AN, Triolo M. Maintenance of skeletal muscle mitochondria in health, exercise, and aging. Annu Rev Physiol. 2019;81:19–41.

Jensen TE, Rose AJ, Hellsten Y, Wojtaszewski JFP, Richter EA. Caffeine-induced Ca²⁺ release increases AMPK-dependent glucose uptake in rodent soleus muscle. Am J Physiol Endocrinol Metab. 2007;293(1):E286–E292.

Liu Z, Zheng H, Deng H, et al. Caffeine, skeletal muscle signalling, and exercise adaptation: a narrative review separating acute ergogenic effects from chronic remodelling. Front Physiol. 2026;17:1875283.

Pedersen DJ, Lessard SJ, Coffey VG, et al. High rates of muscle glycogen resynthesis after exhaustive exercise when carbohydrate is coingested with caffeine. J Appl Physiol. 2008;105(1):7–13.

Southward K, Rutherfurd-Markwick KJ, Ali A. The effect of acute caffeine ingestion on endurance performance: a systematic review and meta-analysis. Sports Med. 2018;48(8):1913–1928.

Steffen D, Paulussen KJM, Crone R, Tucker B, Pathak S, Baar K. Caffeine decreases muscle and tendon protein synthesis and engineered ligament strength in vitro and attenuates adaptation to exercise in mice. J Appl Physiol. 2025;139(6):1569–1580.

Tsuda S, Egawa T, Kitani K, Oshima R, Ma X, Hayashi T. Caffeine and contraction synergistically stimulate 5′-AMP-activated protein kinase and insulin-independent glucose transport in rat skeletal muscle. Physiol Rep. 2015;3(10):e12592.

Greg Dea is a Sports Physiotherapist and Strength & Conditioning Coach based on the Mornington Peninsula, Melbourne. He works with athletes across rehabilitation, return to sport, and performance.