REF-1389 · Nasser et al., 2021 · International Journal of Sports Physical Therapy · Paper · PubMed · Free full text

What helps sit-bone (proximal hamstring) tendon pain?

Nothing is proven: 12 small studies, very low certainty for every treatment. The authors suggest first trying what works for other tendinopathies.

Main caveat: Two small trials and ten case series. No pooling across studies was possible.
moderate evidenceMeasured on recreational to elite athletes

Nasser and colleagues searched MEDLINE, CINAHL, EMBASE, SPORTDiscus and PubMed to January 2019 for any study of a surgical or non-surgical intervention in proximal hamstring tendinopathy — persistent buttock pain with running, sitting and lunging, diagnosed by a health professional — and found 12 that qualified: 2 randomised trials and 10 case series of at least 10 people, 424 participants in all (229 men, 195 women; mean ages 24 to 51).

In the larger trial, 40 professional athletes given four weekly sessions of shockwave did far better than those given a 6-week multimodal programme (a week of rest with NSAIDs, two weeks of manual therapy and therapeutic ultrasound, three weeks of exercise with strengthening and stretching): large effects on pain and function at 1 week, 6 months and 12 months, and 16 of 20 back at their pre-injury level of sport at a mean 9 weeks against none of 20 at a year.

In the smaller trial (15 people, 17 tendons), platelet-rich plasma and autologous whole blood did not differ on function or quality of life. The case series — platelet-rich plasma, corticosteroid injection and surgery — all improved over time, but without a comparison group; surgery had the most complications.

Every outcome was very low certainty, and the authors conclude that there is "insufficient evidence to recommend any one intervention over another", suggesting "a pragmatic approach would be to initially trial approaches proven successful in other tendinopathies".

They also point out that the multimodal comparator gave only three weeks of exercise, shorter than successful programmes for other lower-limb tendons, with exercises (lunge, hamstring stretch, exercise bike) likely to compress the tendon at the sit bone early on.

Our reading: the honest answer to "what fixes my hamstring tendon?" is that nothing has been shown to; tendon-loading principles are borrowed from other sites, and the one trial favouring shockwave compared it with a weak control.

Protocol

Searches and selection: five databases to January 2019, citation tracking and reference checking; two independent screeners; RCTs, prospective cohorts, case-control studies and case series with 10 or more participants; English only; 1924 records, 34 full texts, 12 included. One protocol amendment before the search: case series under 10 participants excluded (three dropped).
Participants: n = 424 (229 males, 195 females) in 12 studies; mean ages 24 to 51 years; diagnosed by a health professional, with MRI or ultrasound in 92% of studies; clinical criteria varied between studies; acute or complete tears and avulsions excluded.
Interventions: platelet-rich plasma (the most common), surgery (4 series: tenotomy, debridement, sciatic nerve release), corticosteroid injection (2), autologous whole blood (1), shockwave (1), multimodal care (NSAIDs, manual therapy, exercise and stretching; 2).
Outcomes and time points: symptoms, physical function (including return to sport), quality of life, ratings of success; short (12 weeks or less), medium (to 1 year) and long term (1 year or more); adverse events minor or major.
Analysis and grading: standardised mean differences between groups (RevMan) and standardised paired differences within groups (R, metafor, r = 0.50 assumed when unreported); 0.2, 0.5 and 0.8 = small, medium, large. Risk of bias: Cochrane tool (RCTs), Joanna Briggs Institute checklist (case series); reporting completeness: TIDieR.GRADE: RCTs start high, case series low; single studies under 300 participants were treated as inconsistent and imprecise, so no single study could rise above low.
Funding and interests: funding statement "none"; "The authors have no conflicts of interest to declare." Affiliations as printed include La Trobe University, the University of Technology Sydney, the University of Queensland, Northern Health and the Gold Coast Suns Football Club. Two authors (Vicenzino, Grimaldi) are also authors of the gluteal tendinopathy trial REF-1390.

Key Results

Intervention (studies, design, n)OutcomeResultCertainty (review's GRADE)
Shockwave (4 weekly sessions, 2500 impulses) vs 6-week multimodal programme (Cacchio 2011, RCT, 40 professional athletes)Pain (VAS)SMD −1.84 (95% CI −2.59 to −1.09) at 1 week; −3.23 (−4.28 to −2.19) at 6 months; −3.22 (−4.28 to −2.16) at 12 months, favouring shockwavevery low
SameFunction (Nirschl phase rating scale)SMD −3.09 (−4.04 to −2.15); −2.90 (−3.88 to −1.92); −2.42 (−3.33 to −1.50)very low
SameReturn to pre-injury level of sport16/20 (80%) at a mean 9 weeks (range 6-15) vs 0/20 at one yearvery low
Platelet-rich plasma vs autologous whole blood (Davenport 2015, RCT, 15 people, 17 tendons)Function, quality of lifeNo difference: e.g. HOS sport at 6 months 0.17 (−0.86 to 1.21); MHHS 0.24 (−0.76 to 1.24); iHOT-33 −0.04 (−1.05 to 0.97); one sciatic-nerve irritation after PRPvery low
Platelet-rich plasma alone (3 case series)Before-after changePain SPD −0.92 (−1.54 to −0.29) at 12 weeks; function −0.44 (−0.82 to −0.06) to −0.90 (−1.52 to −0.28); 3/29 (10%) back to pre-injury sport at 8 weeksvery low
Corticosteroid injection (2 case series)SymptomsShort-term improvement in 14 (78%) of 18, but 56% without improvement beyond three months; in the second series 56% still had symptoms at long-term follow-up (as the text reports)very low
Surgery (4 case series)Before-after changePain SPD −1.89 (−2.36 to −1.41) to −6.02 (−8.10 to −3.94); function 4.08 (−5.53 to −2.63); return to pre-injury level 77-100% (mean time 4.4 to 5 months in the two series that gave it); complications in 21 (10%), 2 (1%) major (wound abscess, deep vein thrombosis)very low
Multimodal care, then PRP for those who failed (Wetzel 2012, retrospective series, 15 people)Return to sport2/11 (18%) after multimodal care; 9/9 (100%) after PRP plus more multimodal carevery low
No high or moderate certainty evidence for any intervention (symptoms, function or quality of life); intervention adherence was reported in no study, and no study described post-surgical rehabilitation adequately.
Not tested at all: a progressive tendon-loading or heavy slow resistance programme, load management advice, PRP against placebo, shockwave against sham — named by the authors as the research still needed.

Mechanisms

Measured: only clinical outcomes (pain, function, return to sport, adverse events); no tissue or imaging outcome was synthesised.
Authors' interpretation: the comparator in the shockwave trial may have been weakened by a short exercise block (three of six weeks) and by early exercises — lunges, hamstring stretching, an exercise bike — that likely put the tendon insertion "under high levels of compression"; avoiding excessive compressive load at the enthesis has been proposed for early to mid-stage exercise selection.Improvements in case series may be natural history, regression to the mean or rehabilitation, not treatment.
Inference: positions of deep hip flexion (sitting, lunges, hamstring stretches) press the proximal hamstring tendon against the sit bone; that this is why they provoke pain is the authors' proposal plus our reading, not a mechanism measured in this review.

Practical Application

What helps sit-bone (proximal hamstring) tendon pain? Nothing is proven: 12 small studies, very low certainty for every treatment. The authors suggest first trying what works for other tendinopathies.
Say it plainly: no intervention — exercise, injection, shockwave or surgery — has been shown to work better than another for proximal hamstring tendinopathy (B+, systematic review, all outcomes very low certainty).
Borrowed, and labelled as borrowed: progressive tendon loading built on the principles tested at the Achilles (REF-1380 guideline; REF-628 pain monitoring; REF-626 heavy slow resistance) is our reading of the authors' general suggestion ("approaches proven successful in other tendinopathies"; the Achilles examples are ours), untested at this tendon.Not from this paper; the base's own practice: the pain-monitoring rule at any site other than the Achilles is transfer (SYNTH-73).
Compression is the authors' caution: early exercise choices that press the tendon against the sit bone (deep hip flexion such as lunges and hamstring stretches) were suggested as a reason the multimodal programme did poorly — a proposal, not a tested rule.
Do not promise shockwave: the one positive trial compared it with a short, possibly provocative programme in 40 professional athletes, at high risk of bias.
Surgery is the last step: improvements after surgery come from uncontrolled series with the most complications; the authors carry over a recommendation from other tendons that surgery wait until 12 months of an evidence-based loading programme has been tried.
What it is good for as a source: stating that the evidence gap is real and the size of it. What it does not show: that any treatment works, how long recovery takes, or anything specific to runners (sport was not reported).

Limitations

Two small trials and ten case series. No pooling across studies was possible; single-study effect sizes are large but very imprecise, and before-after effect sizes from case series are not treatment effects (the authors' caution).
Heterogeneous diagnosis and outcomes. Clinical criteria, outcome scales and follow-up times (one week to six years) varied; interventions were poorly described, adherence never reported.
Search date. Studies after January 2019 are not included.
Population. Mostly athletes by level, sport not given; few women in the shockwave trial (13 of 40); the PRP-versus-blood trial was almost all women (13 of 15).
Numbers that do not add up: the second corticosteroid series is said to leave 56% with symptoms at long-term follow-up, while its Table 2 row lists complete resolution in 11 (29%) of 38 rated patients, which would leave 71% with some symptoms (our arithmetic) — the figures do not reconcile; the abstract gives the surgical function effect as −4.08 and the Results text as 4.08 with the same confidence interval (−5.53 to −2.63).
Funding and interests: none stated; no conflicts declared. No device maker, platform or federation is named.
How this entry was checked and changed
Why this level

a systematic review held below A- because there is no meta-analysis and the evidence it gathers is two small RCTs (40 and 15 participants; both downgraded for high risk of bias) and ten uncontrolled case series, with every outcome rated very low certainty; it carries the weight of its best comparative trial, not of a review of trials. Its main finding — that no intervention is proven — it documents well. Read from the full text (Europe PMC, CC BY-NC-SA 4.0)

Full text (checked 2026-10-28)
Twelve studies (2 RCTs and 10 case series) were included (n=424; males 229) (full text)
One-hundred and ninety-nine (47%) participants were described as professional, competitive or high-level athletes, 75 (18%) were recreational athletes and five participants were non-athletes; activity levels in 145 (34%) not reported (full text)
Shockwave vs multi-modal intervention, self-reported symptoms long-term: -3.22 SMD; 95% CI -4.28, -2.16 (full text)
Sixteen athletes (80%) returned to pre-injury level of sport, at a mean time of nine weeks (range; 6-15 weeks) post shockwave; no participants in the multi-modal group returned to sport at one year (full text)
Platelet-rich plasma vs autologous whole blood, medium-term physical function: (0.17 SMD; 95% CI -0.86, 1.21) to (0.24 SMD; 95% CI -0.76, 1.24) (full text)
Surgery complications occurred in 21 (10%) cases; two (1%) patients experienced major complications (full text)
There is insufficient evidence to recommend any one intervention over another; a pragmatic approach would be to initially trial approaches proven successful in other tendinopathies (full text)
Funding statement: none; the authors have no conflicts of interest to declare (full text)