Lifting for endurance: what it does and what it does not

The fear, and what the reviews found

The old objection to lifting was that it makes you heavier, and an endurance athlete carrying more mass has more to move. The review of strength training for running and cycling does not find that trade: total body mass does not go up, even though the trained muscles themselves grow by roughly three to six per cent ¹.

What improves is economy and the performance measures that sit on it — velocity and power at VO2max, and time-trial performance in trained endurance athletes ². Economy is the mechanism worth understanding, because it is where the benefit shows up. Two riders with the same VO2max and the same threshold can need different amounts of oxygen to hold the same pace. Strength work lowers that cost, and a lower cost means either a faster pace at the same effort or the same pace at less.

The explanation the review settles on is adaptation within the trained muscle rather than one single mechanism: later recruitment of the less efficient fast-twitch fibres, better neuromuscular efficiency, a shift of type IIX fibres toward the more fatigue-resistant IIA, and stiffer tendons ¹. Two of those are neural, two are structural. The muscle changes; the scale does not.

Heavy, few, and moved fast

The protocols behind these findings are heavy loads for few repetitions, performed alongside normal endurance training rather than instead of it: four to ten repetition maximum, two to three sets, two to three minutes of rest ¹.

The part most often dropped is intent. The review is specific that the concentric phase should be performed with maximal intended velocity — the bar moves slowly under a heavy load, but you try to move it fast. It also reports that explosive, lighter work adds to heavy work rather than substituting for it, so the two are not rivals.

One scope limit is worth stating plainly, because it decides what this app will and will not claim. The running-and-cycling review is built on lower-body work — squats, leg press, plyometrics — so it cannot speak to a pressing session. The broader review does include upper-body training, but for cross-country skiers, whose event is driven by poling ². A cyclist benching is outside both.

The injury case, and exactly what was tested

A meta-analysis of randomised trials found strength-training interventions reduced sports injuries to under a third of the control rate, and balance training to a little over half, while stretching showed no effect at all ³. That is a large effect for something most endurance athletes skip.

It is worth knowing what sat behind the strength figure, because it is not what the word suggests. The four trials pooled there were eccentric hamstring work on a flywheel, the Nordic hamstring exercise, army-recruit calisthenics, and a neuromuscular warm-up — not barbell programmes. So the finding supports bodyweight and eccentric work rather than arguing against it, which is the opposite of what a reader might assume from the label.

Two further limits. The overuse-injury figure in that paper pools every kind of exercise intervention, not the strength ones specifically, so it should not be read as "strength training halves overuse injuries". And the cohorts are team-sport athletes and army recruits — football, basketball, handball. There is no endurance population in it, so applying the number to a cyclist is an extrapolation, not a citation.

Where the tier badge comes from, and what it is not

The strength tab places each of five barbell lifts on a five-step scale from untrained to elite, for the athlete's sex, bodyweight and age band. Those thresholds come from accumulated competitive-lifting classification data ⁴.

That source is a classification of lifters, not a survey of a population, and its own authors say so: the tables are not regression-derived and are not strength norms. The distinction changes what the badge can honestly say. It answers "where does this lift sit among people who lift", not "you are stronger than N per cent of people your age". A percentile needs a distribution behind it — a real sample with a real spread — and a table of category boundaries is not one, however tempting it is to interpolate between the steps.

The number beside it is usually an estimate, not a measurement. A logged single is taken as it stands; anything heavier-rep is converted, and those conversions carry a real, direction-consistent bias. Of the seven equations tested against measured maxima, every one underestimated the deadlift — by nine to fourteen per cent on average — and none of them predicted all three lifts well ⁵. Treat the tier as a band you are in, not a score you hold.

Fitting it around the training that matters

Strength work is a supplement to endurance training in every study above, never a replacement for it. The sessions are added to an existing programme, which means they add load — and load is load whether it came from a bike or a barbell.

Start light. The review recommends easing into the concurrent load over the first two to three weeks rather than beginning at full prescription ¹.
Two sessions a week is the usual prescription in these protocols, not five ¹.
Expect early soreness to blunt some endurance sessions while the pattern is new. That is the adaptation cost, not a sign it is wrong.

How to sequence lifting against hard interval days is a real question and one this evidence does not answer — the reviews above report weekly frequency, not session order. Treat the usual advice to separate them as coaching convention rather than a finding.

References

Numbered in order of first appearance. Every link goes to the original work.