Losing weight as an endurance athlete
Screen for low energy availability first, then lose it slowly, in the off-season, with the key sessions fuelled, protein high and the lifting kept; a deficit inside a build block puts the training it serves at risk
The question it answers
"I want to lose X kg — how fast, how, and when?" and "Will it make me faster?" — asked often, with no card to answer from until now. It decides whether a deficit fits at all, its pace, what protects training and muscle, where in the season it goes, and which signs end it.
Default (what to prescribe first)
Screen before advising. Ask about missed or lengthening periods, libido and morning erections in men, bone stress injuries, falling performance, sleep and mood; ask also about persistent fatigue (not on the IOC list; the base's own addition) and hunger (an athlete-reported symptom in the consensus, not a scored indicator; expected in any deficit, so context, not a stop sign on its own).
Any of the listed indicators → not a weight-loss conversation. This screen is the REDs indicator list of the IOC consensus (REF-341, A), applied as a coaching filter — not from a study; the base's own practice.
Ask also how the athlete feels about their body and about eating: in 105 female athletes, negative body image was associated with REDs risk on validated questionnaires (r = 0.61), with disordered eating a statistical mediator (REF-956, B, cross-sectional — a correlation, not a cause; women only, a risk score rather than a diagnosis).
A worrying answer is part of the screen, not a diagnosis.
Slow rate: about 0.5-0.7% of body mass a week. In 24 elite athletes (sports unnamed, lifting 4×/week), a 0.7%/week target with 19% less energy intake gained lean mass (+2.1%) while a faster cut (1.4% target, 1.0% achieved, 30% less energy) did not (−0.2%), for the same total loss (5.6% vs 5.5%) (REF-760, B+). Transfer: not endurance athletes at endurance volumes; the 0.5 lower bound is not from a study; the base's own practice.
Keep heavy strength work in the week (REF-760 had all athletes lifting; REF-345, B: in 15 young adults a 5-day deficit cut resting muscle protein synthesis 27% and a resistance bout restored it). Transfer: short lab studies, not endurance blocks.
Across RCTs of ≥3 weeks, lifting in a deficit kept strength gains comparable to lifting without one but impaired lean-mass gains (ES −0.57), and a deficit of about 500 kcal/day prevented them (REF-1087, A, meta-analysis and meta-regression; populations not described in the abstract) Inference, not tested there (no non-lifting arm): in a cut the lifting is there to protect muscle; expecting it to grow muscle beyond a slow cut like REF-760 is optimistic.
Protein toward the top of the endurance band: ~1.8 g/kg/day advocated, possibly above 2.0 during carbohydrate-restricted training and on rest days per its abstract, with Table 1 targets of 1.95 g/kg/day on low-carbohydrate training days and 2.0 on recovery days (corrected 2026-10-12) (REF-151, A, narrative review); UCI: 1.6-2.1 g/kg/day, the upper end in energy-restricted periods (REF-748, A).
In a steep 2-week cut (60% of habitual energy), ~2.3 g/kg kept lean mass that ~1.0 g/kg lost (−0.3 vs −1.6 kg), with the same fat loss and no performance difference (REF-880, B+, 20 resistance-trained young men — a transfer).
Fuel every key session in full and take the deficit elsewhere (see Decision). UCI: train-low sessions must not compromise key sessions or habitual energy availability (REF-748, A).
Do it outside the build and race blocks — not from a study; the base's own practice, consistent with the REF-341 entry's advice never to combine caloric restriction with hard training in an athlete aiming at performance gains.
Decision (who → what)
| Situation | What to do | Source and how sure |
|---|---|---|
| Any REDs indicator present (below) | No deficit. Name the concern, restore fuelling, refer to a sports physician/dietitian | REF-341 A: restoring energy availability is the core of treatment; the entry says not to assign a severity level yourself |
| Healthy, clear fat to lose, off-season or early base | Deficit at ~0.5-0.7% BM/week, lifting kept, protein high | REF-760 B+ (elite, non-endurance); window = practice |
| Wants it faster ("1 kg a week") | Explain the trade: in REF-760 the faster cut reached the target sooner (5.3 vs 8.5 wk) but gave up the lean-mass gain | REF-760 B+; rate and duration were confounded |
| In a build block or within weeks of an A-race | Hold weight; eat for the training | REF-341 entry advice; REF-545 B: in 12 trained-to-elite men, 14 days at ~22 kcal/kg FFM cut explosive power and haemoglobin and worsened well-being |
| Recreational, higher body fat, lifting | A moderate deficit may still allow fat-free-mass gain | REF-738 B: n=30, recreationally trained, BMI ~26, protein 2.5 g/kg/day: FFM +1.04 kg (deficit) vs +0.97 kg (maintenance); the fat-loss difference was not significant |
| Already lean, asks for "race weight" | Default answer is maintenance; the lean-mass result above does not transfer | REF-738's own limitation (BMI 26 is the condition); inference for lean athletes |
| Masters (40+) | Slower end of the rate, lifting non-negotiable, ~0.4 g/kg protein per meal | REF-348 B (healthy, untrained men): MPS plateaued at ~0.40 g/kg per meal at ~71 y vs ~0.24 at ~22 y (p = .055 per kg body mass; no one aged 40-60 was studied, so the 40+ row is a transfer); REF-579 A: the protein effect on FFM shrank with age (−0.01 kg per year); slower rate = inference |
| Postmenopausal woman | Expect strength gains, not easy muscle gains; volume higher than 6-8 sets/muscle/week may be needed | REF-523 B+ (healthy women, 52 y): 1RM rose in all groups; FFM rose only premenopause |
| Wants time-restricted eating | Possible short tool, not a default; watch the hormone signal | REF-520 B: 16 elite U23 male cyclists, 8-h window at 100% of estimated needs, 4 wk: −2% weight, FFM unchanged, performance no different, free testosterone and IGF-1 fell. Pooled over 16 small trials in 245 athletes: small reductions in body mass (SMD −0.27) and body fat% (SMD −0.39), lean mass not significantly changed in the main model (SMD −0.16, 95% CI −0.42 to 0.09; a small loss in a sensitivity analysis), aerobic capacity, strength and power not significantly changed, energy intake not significantly lower (SMD −0.35, 95% CI −0.77 to 0.07); certainty very low to moderate, no trial at low risk of bias, intake uncontrolled in several, and no hormone or REDs indicator pooled (REF-1007, B+) — so it does not clear the hormone signal, and the screen above still comes first |
Fuel and protein during the deficit
Where the deficit sits in the week — not from a study; the base's own practice: keep the carbohydrate before, during and after quality sessions and long rides at the day's UCI row (light endurance 5-7 g/kg, moderate <3 h/day 6-10 g/kg, REF-748, A), and take the modest cut from low-demand days and from fat and discretionary food, never from protein. Do not make rest days automatic "eat less" days — the REF-342 entry carries the same rule as base practice.
Protein, stated with its source each time:
Train-low is not a weight-loss method. Train-low augmented signalling in 73% of 11 studies but improved performance in 37% of 11 (REF-543, B+, framework review), and its entry frames it as timing, with total energy still covering the training. Using fasted rides to deepen a deficit stacks two stresses — inference.
How to progress / periodise
When NOT to use it / red flags
Stop the deficit, restore fuelling and refer when any of these appear — indicators from the IOC REDs CAT2 table (REF-341, A):
What the assistant must never do — not from a study; the base's own practice built on REF-341 and REF-748:
What the evidence does and does not show
Evidence gaps
How this entry was checked and changed
(rate rests on one B+ RCT in elite athletes from unnamed sports; protein on A/B+ entries; the REDs line on the A-level IOC consensus; season timing and where the deficit sits in the week are the base's own practice)