Abstract
Purpose
Complete proximal hamstring avulsion injuries are anatomically complex because of their proximity to the sciatic nerve. This study characterises neurological abnormalities following injury and surgical repair. Although denervation has been described, its severity, pattern and diagnostic thresholds in relation to magnetic resonance imaging (MRI) findings, anatomical innervation and tendon retraction distance remain poorly defined.
Methods
In this prospective longitudinal cohort study, 18 patients undergoing surgical repair of MRI‐confirmed complete proximal hamstring avulsion were evaluated using serial electromyography (EMG) and MRI performed preoperatively and postoperatively over 12 months. Tendon retraction distance, muscle and sciatic nerve MRI characteristics and EMG evidence of denervation were recorded. EMG findings were used to distinguish traumatic from postoperative nerve injury. Receiver operating characteristic analysis and Youden's J‐statistic were applied to determine a tendon retraction threshold associated with neurotrauma.
Results
Among the 18 patients, 5 (28%) had preoperative nerve injuries. Of the 13 patients without denervation preoperatively, 3 (23%) experienced iatrogenic injuries postsurgery. SHORE scores and MRI did not differ significantly between normal and abnormal EMG cohorts, although neurological symptoms were numerically more frequent in the abnormal group. Increased tendon retraction was significantly associated with more severe EMG abnormalities in the hamstring muscles, with a 5‐cm threshold demonstrating good discrimination.
Conclusion
Proximal hamstring avulsion injuries exhibit varying degrees of neuropathology and recovery. MRI‐measured retraction, not MRI signal changes, may predict neurotrauma. EMG is required to confirm denervation. Retraction distance over > 5 cm (nerve at risk distance, NARD) is associated with a substantially increased risk of neurotrauma, but the long‐term clinical consequences of underlying nerve injury, especially in the athlete, require further investigation.
Level of Evidence
Level II, prospective longitudinal cohort study.
Keywords: athlete proximal hamstring, proximal hamstring avulsion, proximal hamstring nerve injury, proximal hamstring repair, proximal hamstring sciatic nerve
MRI‐measured tendon retraction distance predicts EMG‐confirmed neurotrauma in proximal hamstring avulsion, identifying a nerve‐at‐risk distance beyond which neurological injury is more likely.
Abbreviations
- EMG
electromyography
- FPR
false positive rate
- MRI
magnetic resonance imaging
- MRN
magnetic resonance neurography
- NARD
nerve at risk distance
- NCS
nerve conduction studies
- ROC AUC
receiver operating characteristics area under the curve
- SHORE
Sydney hamstring origin rupture evaluation
- STIR
short tau inversion recovery
- TPR
true positive rate
INTRODUCTION
Proximal hamstring avulsion injuries represent a significant athletic impairment, most commonly caused by forceful eccentric contraction of the hamstring muscles during activities such as water skiing and hurdle jumping [2, 14].
These injuries, in which the hamstring detaches from its origin at the ischium, are predominantly observed in middle‐aged individuals and demonstrate a notable gender disparity: women are more likely to sustain the injury during daily activities, whereas men sustain it predominantly in sports [14]. Accompanying sciatic nerve‐related symptoms are common and range from motor weakness to sensory disturbances and neuropathic pain [26, 27]. Nonoperative management of complete avulsions may result in persistent deficits in strength and function, although some patients manage to return to their previous levels of sporting activity [1, 11]. Surgical intervention is advocated for complete avulsions, particularly for athletes, to expedite a return to sport and enhance functional outcomes, and in sciatic nerve involvement [24, 27]. The described surgical techniques for reattaching the proximal hamstring avulsion generally yield satisfactory postoperative results in terms of strength and endurance relative to the uninjured side [28]. However, some complete avulsions can result in more substantial injury to the proximal hamstring innervation due to the close neuroanatomical relationship between the ischial tuberosity origin and sciatic nerve and its branches [19, 20]. Delayed surgical repair may increase the risk of further sciatic nerve compromise from traction and tethering, often necessitating extensive intraoperative neurolysis [15, 28].
Despite these risks, patients with sciatic nerve‐related symptoms following proximal hamstring avulsion often demonstrate improvement in motor deficits, sensory symptoms and neuropathic pain following surgical repair [27].
Current evidence does not allow reliable differentiation between traumatic and iatrogenic sciatic nerve injury, complicating prognostication and patient counselling. While magnetic resonance imaging (MRI) is the gold standard imaging modality for assessment of proximal hamstring injury itself, the optimal adjunct for evaluation of neuropathology remains undefined [12, 29].
The aims of this study were to determine whether nerve injury associated with proximal hamstring avulsion is primarily traumatic or iatrogenic, to characterise the pattern and temporal evolution of motor denervation using serial electromyography (EMG), and to assess whether MRI‐derived measures of injury severity are associated with EMG‐confirmed neural injury. It was hypothesised that more extensive avulsion injuries on MRI, including greater tendon retraction, would be associated with a higher prevalence and greater extent of denervation on EMG.
METHODS
This is a prospective longitudinal cohort study.
Participants
Participants, recruited by a single orthopaedic surgeon (D.W.), had MRI‐confirmed proximal hamstring avulsions and chose surgical repair, agreeing to complete up to four EMG and MRI procedures over 12 months. Ethics approval was granted by St Vincent's Hospital Human Research Ethics (2019/ETH10683). Written informed consent was provided by all participants prior to data collection.
Variables of interest
At the preoperative consultation, the following data was collected from each participant: age, sex, side of injury, date of injury, mechanism of injury, previous treatment, injury classification [28], tendon retraction distance, sciatic nerve involvement from MR appearances, presence of preoperative neurological symptoms, injury severity, preinjury activity levels, date of surgical repair and the Sydney Hamstring Origin Rupture Evaluation (SHORE) score [9]. Any complication experienced within 12 months postoperatively was recorded for each participant.
Timeline
Following injury, all participants underwent preoperative MRI and EMG assessment prior to surgical repair. Surgery was performed after completion of baseline imaging and neurophysiological testing. Postoperative MRI and EMG assessments were performed at 3 months, with additional assessments at 6 and 12 months in participants with abnormal findings, to distinguish traumatic from postoperative nerve injury and to evaluate recovery over time.
Operational definitions
The mechanism of injury was classified by velocity; low (walking speed or less), medium (between walking and running speeds) or high (above running speed). All injuries were confirmed by MRI and classified according to a method previously described [28]. Tendon retraction distance was assessed on MRI. A single, trained radiologist measured tendon retraction distance, and oedema and denervation of hamstring muscle based on intensity in T1 and T2 signals. Preoperative neurological symptoms were considered present if participants reported tingling or altered sensation. Injury presentation was dichotomised into acute (< 6 weeks) or chronic. Participants' preinjury activity level was categorically classified as: recreational, competitive and elite.
SHORE score
Functional outcomes were assessed using the SHORE score, a patient‐reported outcome measure evaluating pain, walking/running ability, activity limitation, sitting/driving tolerance, stair negotiation and symptoms of ache and tightness, assessed both preinjury and at follow‐up.
Participants' pain and functional outcome was assessed at their preoperative, 6‐ and 12‐month appointments. The patient‐reported outcome measure was devised and validated by the senior author (D.W.) [9]. (Table 1, SHORE score).
Table 1.
SHORE score.
| Before injury/current | SHORE—Sydney hamstring origin rupture evaluation | ||
|---|---|---|---|
| Name: | Side: | Left | Right |
| Injury date: | Operation date: | ||
| Pain (tick one box in each column) | ||
|---|---|---|
| Before Injury | Current | |
| None | ||
| Mild—no limit to activity | ||
| Moderate—limit of strenuous exercise | ||
| Fairly severe—some limitation of day‐to‐day activity | ||
| Very severe—severe limitation | ||
| Totally disabled | ||
| Running/walking (circle one in each column) | |
|---|---|
| Before injury | Current |
| Sprint/Run/Jog/Brisk Walk/Walk/Unable to walk | Sprint/Run/Jog/Brisk Walk/Walk/Unable to walk |
| Activity (tick one box in each column) | ||
|---|---|---|
| Before injury | Current | |
| Unrestricted | ||
| Mildly restricted | ||
| Unable to do strenuous exercise | ||
| Unable to do moderate exercise | ||
| Unable to do gentle exercise | ||
| Unable to do activities of daily living | ||
| Sitting/driving (tick one box in each column) | ||
|---|---|---|
| Before injury | Current | |
| Any type of seat as long as you like | ||
| Comfortable seat for as long as you like | ||
| Pain stops me driving/sitting for more than an hour | ||
| Pain stops me driving/sitting for more than 30 min | ||
| Pain stops me driving/sitting for more than 10 min | ||
| Unable to sit/drive | ||
| Stairs/inclines | ||
|---|---|---|
| Before injury | Current | |
| Easily | ||
| Little difficult | ||
| Moderate difficulty (stop frequently) | ||
| Extreme difficulty (use of aids intermittently) | ||
| Only with support of aids (every step) | ||
| Unable to do | ||
| Aches and tightness (tick one box) | ||||
|---|---|---|---|---|
| Before injury ache | Current ache | Before injury tightness | Current tightness | |
| Never | ||||
| Only with exercise | ||||
| Rare (once a month) | ||||
| Occasional (1 week) | ||||
| Regularly (everyday) | ||||
| Always (constant) | ||||
Note: The SHORE score is a patient‐reported outcome measure assessing pain, function and activity limitation before injury and at current follow‐up in patients with proximal hamstring injury.
MRI
Analysis of MRI imaging was performed by a single musculoskeletal fellowship‐trained radiologist (S.F.). Assessment of muscle changes was based on qualitative assessment of signal intensity of the hamstring muscles, both in terms of the degree of signal intensity on fat‐saturated images as well as the percentage of cross‐sectional area involvement of the muscle on axial scans. Oedema and muscle atrophy were assessed and graded according to intensity (none, mild = 1%–33%, moderate = 34%–66%, severe > 67%–100%), consistent with established MRI descriptions of denervated skeletal muscle. Acute or subacute denervation‐related changes were identified by increased T2/short tau inversion recovery (STIR) signal intensity, while chronic changes were characterised by muscle atrophy and fatty infiltration, as described in prior experimental and clinical studies of peripheral nerve injury [12, 29].
Oedema was considered reticular or linear increased soft tissue signal, scar tissue was considered linear or band‐like low signal in the soft tissues, both which were determined to involve the epineurium of the sciatic nerve if there was no clear fat plane between it and the surface of the nerve. Haematoma was considered to be fluid signal on T2 imaging or high signal on T1 imaging in the acute setting. This grading was performed for all four muscles of the proximal hamstring (semimembranosus, semitendinosus, short and long head of biceps femoris), thus giving each a numerical grading between 0 and 9, ascending with pathological change. The sum of all gradings was deemed the MRI total score.
Although MRI‐based classification systems for proximal hamstring tears and avulsions have been described, no validated MRI scoring system currently exists for quantifying denervation‐related muscle changes or perineural soft‐tissue involvement in correlation with EMG‐confirmed neurotrauma [6]. Accordingly, this ordinal grading approach was defined a priori to provide a clinically interpretable summary of conventional MRI findings relevant to neural risk. Formal intra or interobserver reliability analysis was not performed.
Neurophysiology
Nerve conduction studies (K.N.) consisted of the measurement of sensory and motor amplitudes with conduction velocities from sural and superficial peroneal nerves, and from extensor digitorum brevis and abductor hallucis muscles. Long latency F‐wave recordings were also recorded from motor studies. Needle EMG was routinely performed in the following muscles: tibialis anterior, medial gastrocnemius and the four hamstring muscles of semitendinosus, semimembranosus, long and short head of biceps. The presence of acute electrophysiological denervation changes was noted, and a binary outcome of normal or abnormal was assigned if there were any features of denervation. This could consist of a high‐firing motor unit pattern on needle EMG, especially if intervention was within a week, ± the presence of spontaneous activity if the injury was at least a week old.
An abnormal EMG was defined as the presence of either spontaneous activity on needle EMG, which was more commonly observed or neurogenic motor unit changes, which were less common in the early phase. Spontaneous activity included fibrillation potentials and positive sharp waves, while neurogenic changes were characterised by larger‐amplitude, prolonged‐duration or polyphasic motor unit potentials. A binary outcome of normal or abnormal EMG was assigned if any of these features were present.
Muscles were sampled at a site distal to any MRI changes, to minimise confounding traumatic changes for spontaneous activity. In this way, sciatic trunk injuries could be assessed from nerve conduction and needle EMG of the muscles below the knee, and branch injuries after leaving the trunk proper by needle EMG of the hamstring muscles, enabling early detection of denervation or conduction block from injury.
Surgical procedure
All surgical procedures were performed by the senior author (D.W.). Surgery was performed in a standardised, previously published technique [7]. In all cases, three suture anchors were utilised for repair and fixation (Q‐Fix All‐Suture; Smith & Nephew), and routine sciatic nerve identification and neurolysis was performed.
Postoperative study protocol
All participants were followed up at three months postoperatively. Follow‐up included a routine postoperative clinical consultation with the senior surgeon (D.W.), and an nerve conduction studies (NCS)/EMG and MRI at 3, 6 and 12 months (K.N.). Participants with abnormal EMG findings preoperatively and/or 3 months postoperatively were followed up again at 6 months postoperatively. A final follow‐up was conducted at 12 months postoperatively for those participants with abnormal EMG findings at 6 months. Two consecutive normal NCS/EMG results signifying either no injury at any point, or sufficient recovery after an abnormal result, warranted termination of further neurophysiological testing. Participants wishing to withdraw at any point were recorded and not contacted for further follow‐up.
Statistical analysis
All statistical analyses were carried out using SPSS (Version 29, SPSS Inc.). Continuous variables were calculated as means and categorical variables as frequency counts. All continuous data were checked for the assumptions of linearity, normality, homoscedasticity, independence of observations and multicollinearity.
Between‐group differences for baseline demographic, clinical and surgical data were analysed using independent t‐tests for continuous data or chi‐square test for independent dichotomous data. The ability of MRI‐measured tendon retraction distance to discriminate with normal or abnormal EMG findings was assessed using receiver operating characteristics (ROC) curve analysis. Tendon retraction distance was treated as a continuous predictor and EMG findings were dichotomised as normal or abnormal. Discriminatory performance was quantified using the ROC curve (area under the curve [AUC]) with corresponding confidence intervals. The optimal threshold was determined using Youden's J‐statistic (sensitivity + specificity – 1), with sensitivity and specificity reported for the selected cut‐off. An AUC of 0.5 indicated no discrimination. Figures 1 and 2; Statistics 1 and 2).
Figure 1.
Statistic 1. Youden's J‐statistic plotted against increasing magnetic resonance imaging (MRI) tendon retraction distance thresholds (cm) for predicting abnormal electromyography (EMG) findings following proximal hamstring avulsion. The optimal threshold corresponds to the maximal Youden's J‐value.
Figure 2.
Receiver operating characteristic (ROC) curve demonstrating the diagnostic performance of magnetic resonance imaging (MRI)–measured proximal hamstring tendon retraction distance for predicting abnormal electromyography (EMG) findings following proximal hamstring avulsion. Sensitivity is plotted against 1–specificity across increasing retraction distance thresholds. The area under the curve (AUC) was 0.88, indicating good discriminatory ability. The optimal cutoff value was determined using Youden's J statistic, with a threshold of 5.0 cm corresponding to a sensitivity of 0.75 and specificity of 0.90. n = 18.
No a priori power or sample size calculation was performed, as this study was exploratory and hypothesis‐generating.
RESULTS
Etiology and trajectory of nerve injury
Of 28 eligible participants who underwent surgical repair of their proximal hamstring avulsion, 24 met inclusion criteria and agreed to participation between September 2020 and July 2021. Six participants were excluded due to incomplete preoperative EMG data. (Figure 3; Consort Flow Diagram). Eighteen participants were included in the final analysis (15 males, 3 females), with a mean age of 54 years (SD 11.3). Injury laterality, mechanism of injury, injury classification and preinjury activity level are summarised in Table 2. Preoperatively, 5 participants had abnormal EMG studies. By 3 months, 2 abnormal EMG studies normalised, while 3 additional participants developed abnormalities. At 12 months, 2 additional abnormal EMGs normalised, resulting in 14 normal EMG results, 3 abnormal and 1 lost to follow‐up.
Figure 3.
CONSORT flow diagram illustrating patient identification, eligibility assessment, exclusions and final cohort inclusion for analysis of proximal hamstring avulsion injuries, including allocation to normal and abnormal electromyography (EMG) groups for subsequent radiological and clinical analyses.
Table 2.
Comparison of normal and abnormal EMG cohorts of patients with proximal hamstring avulsion.
| Variable | Total cohort (n = 18) | EMG classification subgroup | p‐value | |
|---|---|---|---|---|
| Normal (n = 10) | Abnormal (n = 8) | |||
| Mean (SD) | Mean (SD) | Mean (SD) | ||
| Age (years) | 54 (11.2) | 54 (10.9) | 53 (11.5) | 0.83 |
| Time from injury to surgery (days) | 26 (15.5) | 27.9 (19.7) | 23.6 (6.7) | 0.59 |
| Retraction distance (cm) | 4.3 (2.4) | 2.9 (1.5) | 5.9 (2.2) | 0.005* |
| SHORE scores | ||||
| Preinjury | 32.94 (1.72) | 33.4 (1.3) | 32.4 (2.2) | 0.23 |
| Preoperative | 7.17 (4.67) | 6.3 (3.8) | 8.3 (5.9) | 0.41 |
| 6‐months | 29.39 (2.69) | 29.6 (2.3) | 29.1 (3.4) | 0.73 |
| 12‐months | 31.61 (1.57) | 31.8 (1.0) | 31.4 (2.2) | 0.60 |
| n (%) | n (%) | n (%) | ||
| Sex | 0.81 | |||
| Male | 13 (72.2) | 7 (70) | 6 (75) | |
| Female | 5 (31.8) | 3 (30) | 2 (25) | |
| Side of injury | 0.91 | |||
| Right | 11 (61.1) | 6 (60) | 5 (62.5) | |
| Left | 7 (42.3) | 4 (40) | 3 (37.5) | |
| Injury mechanism | 0.87 | |||
| Stretch | 16 (88.8) | 9 (90) | 7 (87.5) | |
| High speed running/sprint | 2 (11.12) | 1 (10) | 1 (12.5) | |
| Injury mechanism velocity | 0.53 | |||
| Low | 6 (33.3) | 3 (30) | 3 (37.5) | |
| Medium | 7 (38.9) | 5 (50) | 2 (25) | |
| High | 5 (30.0) | 2 (20) | 3 (37.5) | |
| Previous treatment | 0.25 | |||
| Yes | 1 | 0 | 1 | |
| No | 17 | 10 | 7 (12.5) | |
| Injury classification | 0.50 | |||
| V | 6 (33.3) | 4 (40) | 2 (25) | |
| VI | 12 (66.7) | 6 (60) | 6 (75) | |
| Preoperative sciatic nerve appearance (MRI) | 0.91 | |||
| Normal | 7 (38.9) | 4 (40) | 3 (37.5) | |
| Abnormal | 11 (61.1) | 6 (60) | 5 (62.5) | |
| Presence of preoperative neurological symptoms | 0.07 | |||
| Yes | 7 (38.9) | 2 (20) | 5 (62.5) | |
| No | 11 (61.1) | 8 (80) | 3 (37.5) | |
| Presentation of injury | 0.18 | |||
| Acute | 16 (0.9) | 8 (80) | 8 (100) | |
| Chronic | 2 (0.1) | 2 (20) | 0 | |
| Previous activity level | 0.63 | |||
| Elite | 1 | 1 (10) | 0 (0) | |
| Competitive | 7 | 4 (40) | 3 (27.5) | |
| Recreational | 10 | 5 (50) | 5 (62.5) | |
| Postoperative complication | 0.20 | |||
| Yes | 6 | 4 | 1 | |
| DVT | 1 | 0 | ||
| Hamstring muscle atrophy | 1 | 0 | ||
| Reduced ROM | 1 | 0 | ||
| Sitting pain | 1 | 0 | ||
| New numbness | 1 | 1 | ||
| No | 11 | 6 | 7 | |
Note: Demographic, clinical, imaging and functional variables are compared between patients with normal and abnormal EMG findings. Continuous variables are reported as mean (± standard deviation) and categorical variables as number (percentage). p‐Values refer to between‐group comparisons. Asterisk (*) Statistical significance was defined as p < 0.05.
Abbreviations: DVT, deep vein thrombosis; EMG, electromyography; MRI, magnetic resonance imaging; ROM, range of motion; SD, standard deviation.
Injury outcomes
SHORE scores did not significantly differ between normal and abnormal EMG cohorts at any time‐point (Table 1). There were no other significant differences between those with normal and abnormal EMG, in various characteristics, including the mechanism of injury, such as the velocity, the injury classification or the previous activity level of the subjects.
Overall, there were no statistically significant differences between total MRI Scores and EMG results (Tables 2 and 3). There were some differences in the MRI appearance of the muscles in denervated versus normal muscles defined by EMG at some time points (e.g., preoperatively for the semimembranosus muscle), but overall, there was not a strong correlation. No difference in MRI appearances of the sciatic nerve between these groups dichotomised by neurophysiology.
Table 3.
Comparison of muscle MR composite scores by normal and abnormal EMG result.
| Preoperative | p‐ value | 3‐month post‐op | p‐ value | 6‐month post‐op | p‐ value | 12‐months post‐op | p‐ value | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Normal | Abnormal | Normal | Abnormal | Normal | Abnormal | Normal | Abnormal | |||||
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |||||
| SH | 0.5 (0.89) | 0.5 (0.71) | 1.00 | 0.0 (0.00) | 0.0 (0.00) | ‐ | 0.00 (0.00) | 0.00 (0.00) | ‐ | 0.25 (0.50) | 2.00 (0.00) | 0.00* |
| n = 16 | n = 2 | n = 15 | n = 3 | n = 4 | n = 4 | n = 4 | n = 3 | |||||
| LH | 1.38 (1.26) | 1.50 (0.71) | 0.89 | 3.50 (2.03) | 5.50 (2.65) | 0.12 | 5.25 (2.06) | 5.00 (2.45) | 0.88 | 5.20 (2.17) | 8.00 (1.41) | 0.16 |
| n = 16 | n = 2 | n = 14 | n = 4 | n = 4 | n = 4 | n = 5 | ||||||
| SM | 1.44 (0.89) | 3.50 (3.54) | 0.04* | 2.85 (2.27) | 5.00 (2.65) | 0.10 | 4.00 (2.83) | 4.33 (2.42) | 0.86 | 4.25 (2.06) | 7.33 (3.06) | 0.17 |
| n = 16 | n = 2 | n = 13 | n = 5 | n = 2 | n = 6 | n = 4 | n = 3 | |||||
| ST | 0.93 (0.96) | 1.33 (0.58) | 0.50 | 2.67 (1.88) | 5.00 (2.00) | 0.07 | 2.75 (1.50) | 4.75 (1.26) | 0.09 | 2.75 (1.50) | 6.33 (1.52) | 0.03* |
| n = 15 | n = 3 | n = 15 | n = 3 | n = 4 | n = 4 | n = 4 | n = 3 | |||||
Note: Mean (± standard deviation) MRI composite scores are shown for individual hamstring muscles preoperatively and at 3, 6 and 12 months postoperatively. Values are reported separately for patients with normal and abnormal EMG findings at each time point. n denotes the number of patients available for analysis. p‐Values refer to between‐group comparisons. Asterisk (*) Statistical significance was defined as p < 0.05. A hyphen (–) indicates that comparative analysis was not performed.
Abbreviations: EMG, electromyography; MR, magnetic resonance; SD, standard deviation; SH, short head of biceps femoris; SM, semimembranosus; ST, semitendinosus.
Preoperative tendon retraction distance and nerve injury
When the abnormal EMG cohort was compared with the normal EMG cohort, a significant difference between the cohorts was only observed in preoperative tendon retraction distance (Table 2). While more participants in the abnormal EMG cohort reported neurological sensory symptoms preoperatively, the difference was not statistically significant between cohorts (p = 0.07; Table 3).
Tendon retraction distance versus EMG
There was a significant difference in preoperative tendon retraction distances between the normal and abnormal EMG cohorts (5.9 cm vs. 2.9 cm; p = 0.005). Calculated Cohen's d (1.59) CI (0.657, 2.516), These findings indicate a significant large effect size, suggesting a substantial difference in tendon retraction distance between the normal and abnormal EMG groups. The ROC AUC was 0.88, indicating excellent discrimination ability. (Figure 1, Youdens J‐Statistic 1). Sensitivity of 90% (95% CI: 66.7%–100%), and specificity of 75% (95% CI: 40%–100%). Cross‐validation confirmed a mean AUC of 0.80 (SD: 0.19).
Optimal threshold analysis based on Youden's J‐statistic was 5 cm. (Figure 2, ROC Tendon Retraction Distance, Statistic 2.).
DISCUSSION
This study demonstrates that proximal hamstring avulsion injuries are frequently associated with occult neurotrauma, which may be traumatic or iatrogenic in origin. EMG‐confirmed denervation was present preoperatively in a substantial proportion of patients and developed postoperatively in a further subset. Greater MRI‐measured tendon retraction distance was strongly associated with the presence and severity of neurotrauma, whereas conventional MRI muscle signal changes did not reliably correlate with functional denervation.
Surgical repair of acute complete proximal hamstring ruptures is generally associated with high patient satisfaction and favourable return‐to‐sport outcomes [4, 5, 17]. This is supported by systematic reviews of operative management [5, 10].
Nevertheless, delayed recovery and persistent hamstring weakness may occur in a subset of patients, suggesting a neuropathological contribution [3, 27]. Sciatic nerve–related symptoms have been reported in approximately one quarter of patients, although motor and sensory deficits are less common and recovery appears more favourable following operative treatment [8, 21, 30].
Anatomy and innervation
The anatomical proximity of the sciatic nerve to the ischial tuberosity renders it vulnerable to injury both at the time of avulsion and during surgical repair [13, 19, 21, 26]. Seidel et al. described distinct zones of motor branch distribution, highlighting a complex branching pattern rather than dominance by a single motor branch. (Figure 4, innervation of the proximal hamstring.) In the present study, greater tendon retraction was associated with a higher likelihood of EMG abnormalities both pre and postoperatively. A nerve‐at‐risk distance (NARD) of 5 cm lies within Seidel's zone 1, which contains early motor branches to the hamstrings [22]. Prior anatomical work has demonstrated that recurrent motor branches may lie as close as 1.5 cm from the ischial tuberosity, supporting the concept that increasing retraction distance represents an anatomical risk factor for neurotrauma [20, 22]. Tendon retraction distance should therefore be interpreted as a marker of neural risk rather than a direct causal determinant, as injury severity, mechanism and anatomical variability limit causal inference.
Figure 4.
Innervation proximal hamstring. Schematic illustration of proximal hamstring innervation showing the branching patterns of the tibial division of the sciatic nerve to the long head of biceps femoris, semitendinosus and semimembranosus muscles. Coloured areas indicate variability in motor branch anatomy.
Longitudinal analysis demonstrated that all hamstring motor branches were susceptible to traumatic injury, with the branch to the semimembranosus muscle appearing most susceptible to iatrogenic injury. Delayed postoperative denervation was observed between 3 and 6 months in all muscles except the long head of biceps femoris. Among affected muscles, semimembranosus demonstrated the highest rate of recovery, with denervation persisting at 12 months in only half of cases identified at 6 months. This recovery pattern may reflect anatomical or surgical factors, although the present data are insufficient to confirm a mechanism.
In addition to acute or iatrogenic injury, chronic tethering or compression of the sciatic nerve by scarred proximal hamstring tissue has been described as a cause of delayed neurologic symptoms, sometimes requiring surgical neurolysis [3, 19, 21]. Such presentations include proximal hamstring syndrome and extraspinal sciatica, which may manifest months or years after the index injury [18, 30]. Magnetic resonance neurography has demonstrated perineural scarring and nerve signal change in cases of chronic sciatic neuropathy related to remote proximal hamstring injury [6].
MRI/EMG
In this cohort, MRI muscle signal abnormalities did not significantly correlate with EMG‐confirmed denervation, indicating that structural MRI alone is unreliable for predicting sciatic nerve injury. Acute muscle denervation has been shown to correlate with T2 hyperintensity and STIR signal changes in experimental and selected clinical settings [12, 25, 29]. In proximal hamstring avulsion, soft‐tissue trauma, extensive hematoma and inflammation may obscure denervation‐related changes on conventional sequences [2, 12].
Furthermore, clinically relevant nerve injury—particularly partial or early axonal injury—may occur without detectable MRI abnormalities on conventional sequences. MRI findings should therefore be interpreted cautiously and in conjunction with clinical assessment and EMG, which remains necessary to confirm functional denervation.
Advanced magnetic resonance neurography has shown promise in characterising sciatic neuropathy by directly depicting nerve signal change and perineural scarring, but remains less widely available and was not used in this cohort. Accordingly, the present findings apply specifically to conventional MRI protocols [6, 16].
Limitations
This study was limited by a small sample size and a single‐surgeon cohort, introducing potential selection bias and limiting power to detect more subtle associations. The absence of a nonoperative comparison group precludes assessment of whether surgical repair reduces delayed noniatrogenic neural injury. MRI assessments were performed by a single musculoskeletal radiologist and EMG studies were interpreted by two neurologists without formal interobserver reliability analysis, which may limit reproducibility.
Implications for treatment and management
MRI‐measured tendon retraction may aid prognostic stratification, while EMG remains essential for confirming functional denervation. A retraction threshold of 5 cm may identify patients at increased neurological risk. Future studies should evaluate surgical timing, technique and adjunct imaging modalities, including neuromuscular ultrasound, in prospective cohorts [23].
CONCLUSION
Proximal hamstring avulsion injuries are associated with variable neuropathology and recovery. MRI‐measured tendon retraction is useful for defining nerve‐at‐risk distance but does not reliably identify denervation. EMG remains essential for confirming functional neurotrauma and characterising its pattern. A nerve‐at‐risk distance greater than 5 cm is associated with a substantially increased risk of neurotrauma, although the long‐term clinical significance of underlying nerve injury, particularly in athletes, requires further investigation.
AUTHOR CONTRIBUTIONS
David Wood: Conceptualisation; methodology; surgical data acquisition; study design; data curation; investigation; writing—review and editing. Milos Spasojevic: Formal analysis; writing—original draft; writing—review and editing; project administration. Sofie French: Study design; data curation; investigation; formal analysis. Ran Wei: Data curation; data management; visualisation. Sebastian Fung: Radiological analysis; methodology; data acquisition. Karl Ng: Conceptualisation; electrophysiology expertise; methodology; validation; data interpretation; writing—review and editing; supervision; writing—review and editing.
CONFLICT OF INTEREST STATEMENT
The authors confirm there are no conflicts of interest to disclose. The authors, their immediate family, and any research foundation with which they are affiliated did not receive any financial payments or other benefits from any commercial entity related to the subject of this article.
ETHICS STATEMENT
Ethics approval was granted by St Vincent's Hospital Human Research Ethics (2019/ETH10683).
ACKNOWLEDGEMENTS
Funding received from The Friends of the Mater.
Wood D, Spasojevic M, French S, Wei R, Fung S, Ng K. MRI‐measured tendon retraction distance is associated with EMG‐confirmed neurotrauma in proximal hamstring avulsion. J Exp Orthop. 2026;13:e70672. 10.1002/jeo2.70672
David Wood and Milos Spasojevic contributed equally to this study.
DATA AVAILABILITY STATEMENT
Data available on request due to privacy/ethical restrictions. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
- 1. Aldridge SE, Heilpern GNA, Carmichael JR, Sprowson AP, Wood DG. Incomplete avulsion of the proximal insertion of the hamstring. J Bone Jt Surg Br. 2012;94 B(5):660–662. [DOI] [PubMed] [Google Scholar]
- 2. Beltran L, Ghazikhanian V, Padron M, Beltran J. The proximal hamstring muscle‐tendon‐bone unit: a review of the normal anatomy, biomechanics, and pathophysiology. Eur J Radiol. 2012;81(12):3772–3779. [DOI] [PubMed] [Google Scholar]
- 3. Bertiche P, Mohtadi N, Chan D, Hölmich P. Proximal hamstring tendon avulsion: state of the art. J ISAKOS. 2021;6:237–246. [DOI] [PubMed] [Google Scholar]
- 4. Birmingham P, Muller M, Wickiewicz T, Cavanaugh J, Rodeo S, Warren R. Functional outcome after repair of proximal hamstring avulsions. J Bone Jt Surg. 2011;93(19):1819–1826. [DOI] [PubMed] [Google Scholar]
- 5. Bodendorfer BM, Curley AJ, Kotler JA, Ryan JM, Jejurikar NS, Kumar A, et al. Outcomes after operative and nonoperative reatment of proximal hamstring avulsions: a systematic review and meta‐analysis. Am J Sports Med. 2018;46(11):2798–2808. [DOI] [PubMed] [Google Scholar]
- 6. Bucknor MD, Steinbach LS, Saloner D, Chin CT. Magnetic resonance neurography evaluation of chronic extraspinal sciatica after remote proximal hamstring injury: a preliminary retrospective analysis. J Neurosurg. 2014;121(2):408–414. [DOI] [PubMed] [Google Scholar]
- 7. Carmichael J, Packham I, Trikha SP, Wood DG. Avulsion of the proximal hamstring origin. Surgical technique. J Bone Jt Surg. 2009;91(Suppl 2):249–256. [DOI] [PubMed] [Google Scholar]
- 8. Cohen S, Bradley J. Acute proximal hamstring rupture. J Am Acad Orthop Surg. 2007;15(6):350–355. [DOI] [PubMed] [Google Scholar]
- 9. French SR, Kaila R, Munir S, Wood DG. Validation of the Sydney Hamstring Origin Rupture Evaluation (SHORE): a clinical hamstring assessment tool. Bone Jt J. 2020;102 B(3):388–393. [DOI] [PubMed] [Google Scholar]
- 10. Harris JD, Griesser MJ, Best TM, Ellis TJ. Treatment of proximal hamstring ruptures—a systematic review. Int J Sports Med. 2011;32(7):490–495. [DOI] [PubMed] [Google Scholar]
- 11. Hofmann KJ, Paggi A, Connors D, Miller SL. Complete avulsion of the proximal hamstring insertion: functional outcomes after nonsurgical treatment. J Bone Jt Surg. 2014;96(12):1022–1025. [DOI] [PubMed] [Google Scholar]
- 12. Holl N, Echaniz‐Laguna A, Bierry G, Mohr M, Loeffler JP, Moser T, et al. Diffusion‐weighted MRI of denervated muscle: a clinical and experimental study. Skeletal Radiol. 2008;37(12):1111–1117. [DOI] [PubMed] [Google Scholar]
- 13. Imam N, Grace ZT, Cherian C, Alberta F, Bassora R. Pudendal neurapraxia after proximal hamstring repair: a case report. JBJS Case Connect. 2022;12(4):e22.00238. [DOI] [PubMed] [Google Scholar]
- 14. Irger M, Willinger L, Lacheta L, Pogorzelski J, Imhoff AB, Feucht MJ. Proximal hamstring tendon avulsion injuries occur predominately in middle‐aged patients with distinct gender differences: epidemiologic analysis of 263 surgically treated cases. Knee Surg Sports Traumatol Arthrosc. 2019;28:1221–1229. [DOI] [PubMed] [Google Scholar]
- 15. Koc I, Shafiyev J, Ozturk B, Odabasi Z. Proximal sciatic neuropathy secondary to hamstring tendon avulsion. Ideggyogy Sz. 2022;75(11–12):429–432. [DOI] [PubMed] [Google Scholar]
- 16. Ku V, Cox C, Mikeska A, MacKay B. Magnetic resonance neurography for evaluation of peripheral nerves. J Brachial Plex Peripher Nerve Inj. 2021;16(1):e17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. van der Made AD, Reurink G, Gouttebarge V, Tol JL, Kerkhoffs GM. Outcome after surgical repair of proximal hamstring avulsions. Am J Sports Med. 2015;43:2841–2851. [DOI] [PubMed] [Google Scholar]
- 18. Matsuda DK. Editorial commentary: proximal hamstring syndrome: another pain in the Buttock. Arthroscopy. 2018;34(1):122–125. [DOI] [PubMed] [Google Scholar]
- 19. Miller SL, Gill J, Webb GR. The proximal origin of the hamstrings and surrounding anatomy encountered during repair. A cadaveric study. J Bone Jt Surg. 2007;89(1):44–48. [DOI] [PubMed] [Google Scholar]
- 20. Rebuffa N, Strauther S, Holton C, Hackney R. The recurrent motor nerve, a possible cause of denervation following hamstring injury and repair: a cadaveric case CASE REPORT. J Anat. 2020;24(4):281–284. [Google Scholar]
- 21. Saikku K, Vasenius J, Saar P. Entrapment of the proximal sciatic nerve by the hamstring tendons. Acta Orthop Belg. 2010;76(3):321–324. [PubMed] [Google Scholar]
- 22. Seidel PMP, Seidel GK, Gans BM, Dijkers M. Precise localization of the motor nerve branches to the hamstring muscles: an aid to the conduct of neurolytic procedures. Arch Phys Med Rehabil. 1996;77(11):1157–1160. [DOI] [PubMed] [Google Scholar]
- 23. Solankee J, Srinivasan S, Payne R, Cai H. Neuromuscular ultrasound in delayed sciatic neuropathy following hamstring tear. PM & R. 2025;17(5):582–584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Sonnery‐Cottet B, Archbold P, Thaunat M, Fayard JM, Canuto SMG, Cucurulo T. Proximal hamstring avulsion in a professional soccer player. Orthop Traumatol Surg Res. 2012;98(8):928–931. [DOI] [PubMed] [Google Scholar]
- 25. Viddeleer AR, Sijens PE, van Ooijen PMA, Kuypers PDL, Hovius SER, De Deyn PP, et al. Quantitative STIR of muscle for monitoring nerve regeneration. J Magn Reson Imaging. 2016;44(2):401–410. [DOI] [PubMed] [Google Scholar]
- 26. Wilson T, Spinner R, Krych A. Surgical approach impacts posterior femoral cutaneous nerve outcomes after proximal hamstring repair. Clinical Journal of Sport Medicine. 2017;29(4):281–284. [DOI] [PubMed] [Google Scholar]
- 27. Wilson T, Spinner R, Mohan R, Gibbs CM, Krych A. Sciatic nerve injury after proximal hamstring avulsion and repair. Orthop J Sports Med. 2017;5(7):2325967117713685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Wood DG, Packham I, Trikha SP, Linklater J. Avulsion of the proximal hamstring origin. J Bone Jt Surg Am Vol. 2008;90(11):2365–2374. [DOI] [PubMed] [Google Scholar]
- 29. Yamabe E, Nakamura T, Oshio K, Kikuchi Y, Ikegami H, Toyama Y. Peripheral nerve injury: diagnosis with MR imaging of denervated skeletal muscle—experimental study in rats. Radiology. 2008;247(2):409–417. [DOI] [PubMed] [Google Scholar]
- 30. Young IJ, van Riet RP, Bell SN. Surgical release for proximal hamstring syndrome. Am J Sports Med. 2008;36:2372–2378. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data available on request due to privacy/ethical restrictions. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.