Is the interference effect real?

It is real but much smaller than folklore claims, and it is conditional. Three recent meta-analyses on concurrent training found no significant compromise to strength or hypertrophy overall. The effect appears reliably only under specific conditions: same-session training, running rather than cycling, and high endurance volume combined with frequent lifting.

How long should I wait between lifting and cardio?

Six hours or more. The measured interference effect for same-session concurrent training is around -0.66; separated by six or more hours or performed on different days it falls to about -0.10, which is indistinguishable from zero. Six hours is the single highest-leverage scheduling decision available to a hybrid athlete.

Does cycling interfere less than running?

Yes, and at statistical significance. Running produces measurable interference with lower-body strength and hypertrophy; cycling does not. The usual mechanistic explanation is eccentric loading — running is thousands of eccentric contractions per session producing muscle damage that has to be repaired, and cycling is almost purely concentric.

Does the AMPK-mTOR explanation hold up?

It is much weaker in humans than the textbook version suggests. The molecular-signalling story — endurance exercise activating AMPK which inhibits mTOR and blunts protein synthesis — was built largely on rodent and acute-signalling data, and it does not predict the human training outcomes well. Muscle damage and accumulated fatigue explain the observed pattern better.

Should I do cardio on rest days or lifting days?

It depends which is easier to keep easy. Putting endurance work on lifting days preserves genuine rest days but requires real separation to avoid the same-session penalty. Putting it on rest days protects the lifting session but risks eliminating true recovery, especially if the rest-day session drifts upward in intensity, which it usually does.

The Interference Effect: What the Research Actually Says

Where the folklore came from

The classic reference is Hickson 1980, which found that adding endurance training to a strength programme blunted strength gains after several weeks. It was a real finding in a real study, and it was also an extreme protocol: high-frequency, high-intensity endurance work stacked onto heavy lifting in untrained subjects, with no separation.

What happened next is a familiar pattern. A finding under extreme conditions became "cardio kills gains" in gym vernacular, and a molecular mechanism was retrofitted to explain it: endurance exercise activates AMPK, AMPK inhibits mTOR, mTOR drives protein synthesis, therefore cardio blocks muscle growth.

That mechanistic story is much weaker in humans than the textbook version suggests. It was built largely on rodent work and on acute post-exercise signalling — the measurable spike in the hours after a session — and acute signalling is a poor predictor of chronic training outcomes. Signalling studies routinely show effects that never materialise as differences in actual muscle after twelve weeks.

The explanation that fits the human data better is duller and more actionable: accumulated fatigue and muscle damage, competing for the same recovery capacity. That framing predicts the moderators correctly, which the signalling story does not.

The four moderators

Modality: running interferes, cycling does not

This is the largest and most consistent moderator. Running produces measurable interference with lower-body strength and hypertrophy at statistical significance. Cycling does not.

The mechanistic account is eccentric loading. Running is thousands of eccentric contractions per session — the landing phase of every stride decelerates your bodyweight — and eccentric contractions are the ones that produce muscle damage. Cycling is almost purely concentric: you push the pedal, nothing lengthens under load.

That damage is repaired out of the same budget your lifting is drawing on. It is not a competing signal, it is a competing bill.

The practical consequence is unusually clean. If your endurance goal is sport-agnostic — general conditioning, health, work capacity — biasing toward the bike is close to a free lunch. If you are training for a running race, it is not an option, and you manage the other three moderators instead.

Timing: six hours

Separation	Effect size	Interpretation
Same session	≈ −0.66	Substantial and reliable
6+ hours apart, or different days	≈ −0.10	Indistinguishable from zero

This is the single highest-leverage scheduling decision in hybrid programming, and it costs nothing but logistics. Six hours turns a real effect into a rounding error.

It also explains a great deal of the disagreement between people who train this way. An athlete who lifts in the morning and runs in the evening genuinely does not experience much interference and will tell you the effect is a myth. An athlete who runs and then lifts twenty minutes later genuinely does experience it and will tell you it is obvious. Both are reporting their conditions accurately.

Volume: a dial, not a switch

Moderate endurance frequency — two to four sessions a week — is safe. High endurance volume combined with frequent lifting raises the risk.

There is no threshold at which interference switches on. What there is, is a total fatigue budget that both disciplines spend from, and the more endurance volume you add the less lifting volume is recoverable. That is not interference in the molecular sense at all; it is arithmetic.

The practical version is a volume ladder rather than a rule:

Endurance load	Fractional sets per muscle per week	Realistic expectation
None	14–20	Full hypertrophy potential
1–3 h/week	12–18	Essentially unaffected
4–6 h/week	12–15	Slower gains, still real
7+ h/week	10–14	Strength maintenance, not growth

Sequence: whichever is second is compromised

If two sessions must share a slot, the second one is worse. This is uncontroversial and unsurprising, and the rule that follows is simply to put the priority first — accepting explicitly that the other session is a B-effort that day rather than pretending both are full quality.

What the meta-analyses disagree about

They are not unanimous, and being straight about that is worth more than a clean summary.

The disagreements cluster around training status and outcome measure. Untrained subjects show less interference — everything works in untrained subjects, and the ceiling is far away. Explosive and rate-of-force-development measures show more interference than maximal strength does, which makes physiological sense: power expression is the most fatigue-sensitive quality there is. And hypertrophy shows less interference than strength, which is the finding most people find counterintuitive and which points again at fatigue rather than at a blocked growth signal — you can build muscle you cannot currently express as force.

The other genuine limitation is that most of these studies run eight to twelve weeks. A hybrid athlete's actual question is what happens over two years, and nobody has run that trial.

What this means for programming

The four moderators translate directly into five constraints, and any workable hybrid week satisfies all of them:

Hard lifting and hard endurance are never in the same session.
If they share a day, six hours apart, priority first.
Endurance volume and lifting volume trade against each other. Adding one means subtracting the other.
Where the endurance modality is a free choice, choose cycling.
Easy sessions are actually easy — because an easy run at tempo pace is a hard session that you have failed to budget for, which is how people who have followed rules one to four still end up experiencing interference.

That last one deserves emphasis. It is not in the research as a moderator because studies prescribe intensity; in the field it is where most of the damage happens. Why your easy runs are too fast covers it directly.

What to do next

If you have been avoiding endurance work because of the interference effect, the evidence does not support that. Add two to three sessions a week, separated from lifting by six hours or on different days, biased toward cycling if the modality is your choice, and hold lifting volume constant for eight weeks before deciding anything. If you are already doing both and something is degrading, the moderator to check first is timing and the second is whether your easy days are easy.

The other half of the problem — the one that causes more failures than interference does — is energy, and it is covered in adaptive maintenance for hybrid athletes.