Can I rehab a hamstring strain through some pain?
In 43 men, rehab allowing up to 4/10 pain cleared them about as fast as pain-free rehab (median 17 vs 15 days) and kept more strength. Clearance still needed zero pain.
Main caveat: Small, men only. 43 men; powered to detect a large effect (effect size 1.2), so a moderate difference in time to clearance cannot be excluded.
Hickey and colleagues tested the long-standing advice to rehabilitate an acute hamstring strain only in the absence of pain.
Between February 2016 and May 2017 in Melbourne, 43 men (mean age 24.9-27.4 by group; 32 Australian footballers and 30 with an earlier hamstring strain, our sums of Table 4) with a clinically confirmed acute strain began rehabilitation about 3 days after injury and were randomised to perform and progress the same protocol either only at 0 on a 0-10 pain scale (pain-free, n = 22) or with pain up to 4 (pain-threshold, n = 21).
The protocol, twice a week and fully supervised, combined hamstring bridges, 45-degree hip extension, eccentric sliders and the Nordic hamstring exercise, progressed by exercise-specific criteria, with a 9-stage progressive running drill over 50 m; return-to-play clearance required no pain on palpation, on stretch tests, in maximal isometric contractions and in a full sprint.
Median time to clearance was 15 days (95% CI 13-17) pain-free and 17 days (11-24) pain-threshold (hazard ratio for clearance 0.75, 0.40-1.40; P = .37): no evidence of a difference, with the pain-threshold median two days longer.
Strength recovery at 90 degrees of hip and knee flexion was 15% greater in the pain-threshold group at clearance (1-28%) and at 2 months (1-29%), and biceps femoris long-head fascicle length was 0.91 cm (0.34-1.48) better maintained at 2 months; fear of movement fell similarly in both. Each group had 2 reinjuries in the 6 months after clearance (HR 1.05, 0.14-7.47).
Our reading: one small, well-run trial showing that mild pain during hamstring-strain rehabilitation exercise was not harmful and may help strength; it is the base's second tested pain rule after the Achilles trial (REF-628), and it concerns rehabilitation exercise after a muscle strain, not running through pain in training.
Protocol
Design: single-centre "efficacy, double-blind" randomised controlled trial at the Australian Catholic University, Melbourne, CONSORT; ethics approval 2015-307H; registered ACTRN12616000307404. Participants were told only their own group's pain limit and were blinded to the existence of the other arm; all outcome measures were collected by investigators blinded to allocation. The lead investigator, who supervised every session, was not blinded.
Eligibility (Table 1): men and women aged 18-40 with acute-onset posterior thigh pain from a clear mechanism (for example high-speed running or kicking) that stopped activity, seen within 7 days, with pain on palpation and pain at the injury site during isometric knee flexion.Excluded: other causes of posterior thigh pain (hamstring tendinopathy, referred low back pain) and suspected complete rupture needing a surgical opinion. Eligibility verified by an independent physiotherapist; no MRI and no categorical grading.
Participants (Table 4): 52 assessed, 9 excluded, n = 43 randomised (pain-free 22, pain-threshold 21); all men, as no women presented. Age 27.4 ± 5.2 vs 24.9 ± 5.3 y; sport 3 ± 1 days a week in both; Australian football 18 vs 14; earlier hamstring strain 16 vs 14; injured in competition 14 vs 15; lateral injury site 18 vs 15; first assessment 3 ± 2 vs 3 ± 1 days after injury; isometric knee-flexor strength at 90/90 degrees 60.1% of the uninjured leg in both.
Randomisation: stratified by previous hamstring strain and sex, sealed envelopes in blocks of 4 per stratum.
The pain rule: pain at the injury site rated on a 0-10 numeric rating scale during each exercise and each running stage. Pain-free group: perform and progress only with "a complete absence of pain (0 on a 0-to-10 numeric rating scale [NRS])". Pain-threshold group: perform and progress with "a pain rating of 4 or less on the NRS during exercise".The authors chose 4/10 as a slightly more conservative version of the 5/10 pain-monitoring model used in patellofemoral pain and Achilles tendinopathy (REF-628 among its sources).
Rehabilitation (Figure 1, Tables 2-3): twice a week, every session supervised; no ice, medication or topical pain relief in sessions (outside sessions not controlled); no extra rehabilitation at home.Exercises: bilateral then unilateral hamstring bridge, 45-degree hip extension and eccentric slider, plus the Nordic hamstring exercise; at most 3 sets per exercise; progress when the repetition range was done through full range within the pain limit; then add resistance in 5-kg steps.All bilateral variations were attempted in the first session (on average 3 ± 2 days after injury).Running: 9 stages of 50 m repetitions with shrinking acceleration and deceleration phases and a growing hold phase, from walk-jog-walk to run-sprint-run (jog under 50%, run under 70%, sprint above 90% of perceived maximal speed); started once walking was normal within the pain limit; next stage after 3 repetitions within the limit; at most 9 repetitions per session.Participants could gradually rejoin team training but were told to keep any running below the intensity reached in supervised running.
Return-to-play clearance (Table 3, the same for both groups): no pain on palpation; no pain during the active knee extension or passive straight-leg raise, with range at least 90% of the uninjured leg; no pain during maximal isometric knee flexion at 0/0 and 90/90 degrees of hip/knee flexion; no pain or apprehension sprinting at 100% of perceived maximal intensity.After clearance, at least 2 full training sessions before competition were recommended; continuing at least one hip-extension and one eccentric knee-flexion exercise a week was encouraged but not monitored.
Outcomes: primary — days from injury to clearance; secondary — biceps femoris long-head fascicle length (ultrasound), isometric knee-flexor strength (injured leg as a percentage of the uninjured leg at the first assessment), Tampa Scale of Kinesiophobia (17 items, 0-68) at first assessment, clearance and 2 months after clearance, and reinjuries over 6 months (confirmed by a blinded physiotherapist).
Analysis: intention to treat; Cox proportional hazards for time to clearance and to reinjury (Kaplan-Meier curves); linear mixed models for the secondary outcomes; R 3.4.3. Sample size: 29 for 80% power with 20% dropout, from an effect size of 1.2 in earlier lengthening-exercise trials.
Funding and interests: "Dr Jack Hickey and Dr Maniar were recipients of research support funding through the Australian Government Research Training Program Scholarship for the duration of this study. The authors certify that they have no affiliations with or financial involvement in any organization or entity with a direct financial interest in the subject matter or materials discussed in the article." No device maker, federation or data platform is named.
Key Results
Pain-free (0/10) vs pain-threshold (4/10 or less) rehabilitation (Results; Figures 3-7):
| Outcome | Pain-free, n = 22 | Pain-threshold, n = 21 | Between groups |
|---|---|---|---|
| Reached return-to-play clearance | 21 of 22 (one stopped rehabilitation at day 24) | 21 of 21 | — |
| Days from injury to clearance, median (95% CI) | 15 (13, 17) | 17 (11, 24) | HR for clearance 0.75 (0.40, 1.40), P = .37 |
| Isometric knee-flexor strength at 90/90 degrees, first assessment to clearance | +35% (26%, 44%) | +49% (36%, 63%) | 15% (1%, 28%) greater with pain-threshold |
| Same, first assessment to 2 months after clearance | — | — | 15% (1%, 29%) greater |
| Isometric knee-flexor strength at 0/0 degrees, to clearance | +32% (22%, 41%) | +39% (26%, 52%) | no evidence of a difference (−6%, 20%); at 2 months (−6%, 22%) |
| Fascicle length (biceps femoris long head), to clearance | +1.70 cm (1.33, 2.08) | +1.95 cm (1.41, 2.48) | no evidence of a difference (−0.29, 0.78) |
| Fascicle length, first assessment to 2 months | +0.56 cm (0.16, 0.97) | +1.47 cm (0.90, 2.04) | 0.91 cm (0.34, 1.48) greater |
| Fear of movement (Tampa Scale), to clearance | −7 points (−5, −9) | −8 points (−5, −11) | −1 (−4, 2); at 2 months −4 (−6, 0), not significant |
| Reinjuries in the 6 months after clearance | 2 (50 and 67 days after clearance) | 2 (8 and 17 days after clearance) | HR 1.05 (0.14, 7.47), P = 1.0 |
Strength is the injured leg as a percentage of the uninjured leg at the first assessment, so "15% greater" means 15 points of that relative measure (our reading of the method). Bold values are those named in the abstract.
Safety: no adverse events (reinjuries) before clearance. One pain-threshold session was stopped as a precaution when a participant reported 7/10 pain while sprinting; reinjury criteria were not met and he continued.
Early clearance and reinjury: the two pain-threshold reinjuries happened in men cleared at 6 and 11 days after injury, 8 and 17 days after clearance; the pain-free ones were cleared at 13 and 26 days and reinjured 50 and 67 days later.Overall 4 of the 37 men followed for 6 months (11%) reinjured; 3 of the 4 within 2 months of clearance, all 3 having been cleared within 2 weeks of the first injury. The authors read this as a possible link between accelerated clearance and reinjury and as a sign that the clearance criteria may need to account for tissue healing time.
Follow-up: 6-month data from 17 of 22 pain-free and 20 of 21 pain-threshold participants; 2-month data missing for 4 and 3.
Exercise tolerance: the authors observed that participants could often do the Nordic exercise and the unilateral eccentric slider without pain while still having pain or more than a 10% deficit on isometric testing, which conventional guidance would use to delay those exercises.
Numbers that do not add up: the Results say "All 51 potential participants screened for eligibility were men" and, in the next sentence, "Of these 52 potential participants"; the CONSORT diagram gives 52 assessed. Table 2's footnote defines a sprint as above 90% of perceived maximal speed, while the text gives 100% as the sprint upper limit and the clearance test is at 100%.
Mechanisms
Measured: both groups gained isometric strength and fascicle length quickly; the pain-threshold group's strength at 90/90 degrees recovered more and its fascicle-length gain was better kept at 2 months, although all participants were pain-free at clearance and at 2 months when those measures were taken.
Authors' interpretation: participants allowed some pain may have been more willing to contract maximally once pain was less of a barrier; early eccentric and long-length loading, progressed by exercise-specific criteria rather than waiting for pain or strength deficits to resolve, explains the large gains in both groups; the specific threshold is "somewhat of an arbitrary task", but "it is unnecessary to completely avoid pain during HSI rehabilitation".
Inference: the trial cannot separate the pain allowance from what it permitted — earlier and heavier progression — so it supports the reading that mild pain during rehabilitation exercise is acceptable more than the reading that pain itself is useful. The 4/10 ceiling applied within supervised sessions with fixed progression rules; the same number without those rules, or in unsupervised running, is untested.
Practical Application
Can I rehab a hamstring strain through some pain? In 43 men, rehab allowing up to 4/10 pain cleared them about as fast as pain-free rehab (median 17 vs 15 days) and kept more strength. Clearance still needed zero pain.
Start early, within a ceiling: rehabilitation here began about 3 days after injury and included eccentric work from the first session (bilateral eccentric sliders; the Nordic and unilateral variations once exercise-specific criteria were met); mild pain during the exercise (up to 4/10, not worse) was allowed (A-, one RCT in men). This is a clinician-led protocol, supervised every session; the assistant's role is to encourage seeing one, not to run the protocol.
The return itself is pain-free: clearance meant no pain on palpation, on stretch, in a maximal contraction and in a 100% sprint, then two full training sessions before competition. The pain allowance applied to rehabilitation, not to the return or to racing.
A fast clearance is not a finished one: both pain-threshold reinjuries came in men cleared within 6 and 11 days of injury, and 3 of 4 reinjuries came within 2 months of clearance; the authors suggest clearance criteria may need to account for tissue healing time. Too few events to make a rule.
Where it changes the base: SYNTH-73 says "At any site other than the Achilles the rule is transfer". For hamstring-strain rehabilitation exercise there is now a tested rule — one RCT of 43 men, a 4/10 ceiling during supervised exercise and graded run-throughs.It does not test running through pain in training, tendon pain at the sitting bone (REF-1389) or bone pain, where the load is meant to produce no symptoms during, after or the day after loading (REF-1383).Not from this paper; the base's own practice: "The ceiling the assistant gives is ≤3/10 — the base's own, more conservative practice" (SYNTH-73), which sits below the 4/10 tested here.
What it is good for / what it does not show: good for telling an athlete with a diagnosed hamstring strain that mild pain during rehabilitation exercise was not harmful in a randomised trial and did not slow the return.It does not show that a faster return follows (the pain-threshold median was 2 days longer, no evidence of a difference either way), that reinjury risk is unchanged (2 vs 2, far too few), that it holds for women, for MRI-graded or severe strains, or for unsupervised rehabilitation, or that running sessions can be trained through pain.
Limitations
Small, men only. 43 men; powered to detect a large effect (effect size 1.2), so a moderate difference in time to clearance cannot be excluded; no women presented despite being eligible.
Clinical diagnosis only. No MRI (the authors' limitation); some participants may have had MRI-negative injuries, which return faster. The authors argue this matches what many sports clinicians can do.
Clearance, not return to full sport. The primary outcome is meeting clearance criteria; time back to full competition was not reported, so complete recovery time is unclear (the authors' limitation).
Reinjury data too sparse: 4 reinjuries in total, and more pain-free participants were lost to the 6-month follow-up (5 of 22 vs 1 of 21), a possible bias the authors name.
Unblinded supervisor; pain relief outside sessions uncontrolled. The lead investigator ran every session and knew the allocation; analgesics outside sessions were not controlled.
Population. Club-level team-sport men with acute strains, most of whom had a previous hamstring strain; transfer to distance runners and to slow-onset hamstring pain is inference.
Funding and interests: two authors held Australian Government Research Training Program Scholarships; the authors certify no affiliation or financial involvement with any organisation with a direct financial interest. No device maker, federation or data platform funded it.
How this entry was checked and changed
Why this level
a parallel RCT with at least 40 participants (rubric A-), read from the full text (the journal's typeset PDF held in the Maynooth University repository). Not higher: 43 men, powered only for a large effect (29 needed for an effect size of 1.2), diagnosis by clinical assessment without MRI, every session supervised by the unblinded lead investigator, and 2 vs 2 reinjuries are too few to compare
Full text (checked 2026-10-28)
Forty-three men with acute HSIs were randomly allocated to a pain-free (n = 22) or pain-threshold (n = 21) rehabilitation group (full text)
Pain-free group: only with a complete absence of pain (0 on a 0-to-10 numeric rating scale); pain-threshold group: with a pain rating of 4 or less on the NRS during exercise (full text)
Median time from HSI to RTP clearance was 15 days (95% CI: 13, 17) in the pain-free group and 17 days (95% CI: 11, 24) in the pain-threshold group; hazard ratio 0.75 (95% CI: 0.40, 1.40) (full text)
Isometric knee flexor strength recovery at 90°/90° was greater in the pain-threshold group by 15% (95% CI: 1%, 28%) at RTP clearance and by 15% (95% CI: 1%, 29%) at 2-month follow-up (full text)
Improvement in BFLH fascicle length from baseline was 0.91 cm (95% CI: 0.34, 1.48) greater at 2-month follow-up in the pain-threshold group (full text)
Two reinjuries occurred in both groups between RTP clearance and the 6-month follow-up; hazard ratio for reinjury 1.05 (95% CI: 0.14, 7.47) (full text)
The pain threshold of 4/10 or less was selected as a slightly more conservative version of the pain-monitoring model of 5/10 or less (full text)
Hickey and Maniar received research support funding through the Australian Government Research Training Program Scholarship; the authors certify no affiliations with or financial involvement in any organization with a direct financial interest (full text)