Introduction

The gluteus medius and minimus muscles have been referred to as the “rotator cuff of the hip” [1–4]. While pain over the greater trochanter has classically been associated with bursitis, gluteus medius and minimus tendinopathy may be a true underlying cause for greater trochanteric pain syndrome [5–12]. Rotator cuff tears of the shoulder have been described to show a progression of pathology from tendinosis to partial thickness tears to full thickness tears with subsequent muscle atrophy and secondary osteoarthritis [13–18]. Rotator cuff tears of the shoulder have been treated aggressively in the early stages in order to avoid muscle atrophy and mechanical instability [19–23]. Similar to the shoulder, it has been anecdotally noted from our experience that degeneration and tendon tear precede atrophy of the gluteus medius and minimus muscles and that these muscles commonly have disproportionate atrophy compared to the other muscles surrounding the hip. However, unlike the rotator cuff of the shoulder, the gluteus medius and minimus tendons rarely tear completely with muscle retraction based on our anecdotal experience. Nonetheless, while rotator cuff pathology of the shoulder may interfere with daily activities, atrophy of the gluteus muscles contributes to gait disturbance and increases the risk of fall-related fractures [24–29]. While current treatment is primarily directed toward pain reduction, some surgeons have started to treat gluteus tendon tears aggressively with repair [30–34] with recent literature showing successful outcomes of repair [35].

We sought to investigate the prevalence of gluteus medius and minimus tendinopathy and muscle atrophy with age using MRI. We also sought to describe the pattern of involvement of the different insertion sites of the gluteus medius tendon and the distribution of fatty atrophy of both the gluteus medius and minimus muscles.

Materials and methods

Institutional review board approval was obtained prior to the performance of this study. A retrospective review of records and images without informed consent was performed in accordance with Health Insurance Portability and Accountability Act.

Study design and patient selection

We performed a database search of our institution for all pelvic MRI reports between January 2010 and December 2011. The inclusion criterion was age ≥50 years. Exclusion criteria included hip surgery, fracture, infection or tumor, or inadequate image quality. A non-interpreting researcher stratified the available patients by decade group and randomly selected the study group to meet requirements of power analysis. A group sample size of 20 was calculated to achieve 96 % power to detect a difference between group proportions of 0.60 with an alpha target of 0.05. Patient history and physical examination related to pain over the greater trochanter were not available.

MRI protocol

MR imaging of the pelvis was performed using the standard protocol on a 1.5-T scanner using a body coil. Large field of view (36–40 cm), coronal T1-weighted spin echo (TE 10–20, TR 400–800), coronal short tau inversion recovery (TE 20–40, TR >2000), axial T1-weighted spin echo (TE 10–20, TR 400–800), axial T2-weighted fast spin-echo fat-saturated (TE 30–40, TR 3000), and sagittal T2-weighted fast spin echo fat-saturated (12–16 cm FOV, TE 40–50, TR >2000) sequences were acquired. A routine bony pelvis MRI included the area from the anterosuperior iliac spine to the ischia including both hips. Images were obtained using a slice thickness of 4–5 mm with a 1-mm gap.

Evaluation of MRI examinations

Two musculoskeletal radiologists with 24 and 28 years of experience were blinded and reviewed the MR examinations in consensus on a PACS (picture archiving and communication system) workstation. Both hips were evaluated for each individual.

Greater trochanteric bursitis was graded as none, mild (slip of fluid), moderate (distended bursa with round margins), or severe (displacement of adjacent structures). Gluteus medius and gluteus minimus tendon pathology was graded as normal, tendinosis (intermediate signal, not fluid), low-grade partial thickness tear (<50 % tendon fluid signal), high-grade partial thickness tear (≥50 % tendon fluid signal), or full thickness tear (complete fluid signal). The gluteus medius and gluteus minimus were graded for fatty muscle atrophy using the classification system established by Goutallier et al. [36–38], which compares the amount of fat to the amount of muscle for rotator cuff muscle atrophy but has been applied elsewhere in the body [39]. Atrophy grading score 0 was assigned for normal muscle (no fat), score 1 for minimal atrophy (some fatty streaks), score 2 for mild atrophy (fat infiltration less than muscle), score 3 for moderate atrophy (fat infiltration equal to muscle), and score 4 for marked atrophy (fat infiltration greater than muscle). The site of insertion of tendon pathology and anterior or posterior location of muscle atrophy was also analyzed. The tensor fascia lata and iliopsoas were also scored in each patient as internal controls using the same tendon and muscle atrophy scales as above.

Descriptive and statistical analysis

Statistical analyses were performed by a statistician using SAS software. Subjects were stratified by age into decade groups. Given the number of patients, the last three decade groups were combined. Mean values and standard deviations for greater trochanteric bursitis and for tendon pathology and muscle atrophy for the gluteus medius, gluteus minimus, and iliopsoas were calculated. Spearman correlation coefficients were calculated for age and tendinopathy, age and atrophy, bursitis and tendinopathy, and tendinopathy and atrophy. Statistical significance was considered with Bonferroni correction for p-values <0.00625.

Percentages of tendinosis, low-grade partial tear, high-grade partial tear, and full thickness tear for the gluteus medius and minimus tendons were calculated. For the gluteus medius, the percentage of involvement of the different insertions of the gluteus medius tendon was calculated. The percentages of anterior or posterior distribution of the gluteus medius and minimus muscles were also calculated. Fisher's exact tests were calculated between the insertion sites of the gluteus medius tendons and location of atrophy in the stratified age groups. Statistical significance was considered with Bonferroni correction for p-values <0.025.

Results

Our initial database search yielded 2697 MRI reports of the pelvis performed at our institution between January 2010 and December 2011. Exclusion criteria were met in 213 MRI reports (18 hip surgery, 77 hip prosthesis, 92 fracture, 1 infection, 5 tumor, or 16 inadequate image quality). Of the remaining 2484 MRI reports, a non-interpreting researcher randomly selected 185 subjects 50 years old or greater (102 males, 83 females; age range, 50.3–95.7 years; mean age, 65.7 years) to serve as the study population.

Mean values (± standard deviation) for greater trochanteric bursitis, gluteus tendon pathology, and gluteus muscle atrophy increased with advancing age (Table 1, Fig. 1). There was a statistically significant moderate to strong positive correlation between age and tendinopathy and age and muscle atrophy of the gluteus medius and minimus. There was a statistically significant weak positive correlation between age and muscle atrophy for the tensor fascia lata. There was no statistically significant correlation between age and tendinopathy or between age and muscle atrophy for the iliopsoas. There was a statistically significant positive correlation between bursitis and gluteus medius and minimus tendon pathology as well as a statistically significant positive correlation between gluteus medius and minimus tendinopathy and muscle atrophy (Table 2).

Table 1 Mean values for bursitis, tendinopathy, and muscle atrophy
Full size table
Fig. 1
figure 1

a Axial T2-weighted fast spin echo fat-saturated MR image (top) obtained in a 59-year-old male shows gluteus minimus tendinosis (white arrows), and coronal T1-weighted spin echo image (bottom) shows no gluteus minimus muscle atrophy (white arrowheads). b Axial T2-weighted fast spin echo fat-saturated MR image (top) obtained in a 66-year-old female shows a low-grade partial gluteus minimus tendon tear (white arrows), and coronal T1-weighted spin echo image (bottom) shows grade 3 (fat = muscle) gluteus minimus muscle atrophy (white arrowheads). c Axial T2-weighted fast spin echo fat-saturated MR image (top) obtained in a 75-year-old male shows high-grade partial gluteus minimus tendon tear (white arrows) and coronal T1-weighted spin echo (bottom) shows grade 4 (fat > muscle)  gluteus minimus muscle atrophy (white arrowheads)

Full size image
Table 2 Quantitative analysis of age, tendinopathy, atrophy, and greater trochanteric bursitis
Full size table

There was a progression of tendon pathology with higher percentages of low- and high-grade partial tears of the gluteus medius and minimus tendons with advancing age (Table 3). There are two insertion sites of the gluteus medius tendon on the superoposterior and lateral facet of the greater trochanter of the femur routinely identified by MRI [1]. There was greater isolated involvement of the superoposterior insertion of the gluteus medius tendon in the 50–59 year group with greater involvement of both the superoposterior and lateral facet insertions in subjects greater than 60 years old. There is a statistically significant association between the gluteus medius insertion sites and location of muscle atrophy (Table 4). The majority of cases shows no muscle atrophy if there was no insertion site involved (79.5 %), isolated superoposterior insertion involvement (61.8 %), or isolated lateral insertion involvement (68.4 %) of the gluteus medius. If both gluteus medius insertions were involved, both anterior and posterior muscle atrophy was seen in 46.9 % of cases. If the superoposterior facet insertion was involved in isolation, there was greater atrophy of the anterior muscle or both anterior and posterior muscle bulk. If the lateral facet insertion was involved in isolation, there was greater atrophy in the posterior muscle or both anterior and posterior muscle bulk.

Table 3 Percentages of gluteus medius and minimus tendon pathology, insertion site involvement, and atrophy location
Full size table
Table 4 Gluteus medius insertion involvement and muscle atrophy location
Full size table

There is one insertion site of the gluteus minimus tendon on the anterior facet of the greater trochanter routinely identified by MRI [1]. There was greater isolated atrophy of the anterior gluteus minimus in the 50–59- and 60–69-year groups and greater involvement of both the anterior and posterior gluteus minimus muscle in the 70–99-year group (Table 3).

Discussion

The progression of tendon pathology from tendinosis to tears and subsequent muscle atrophy has been demonstrated in several muscle groups, including the rotator cuff of the shoulder [13–18]. While the gluteus medius and minimus have been referred to as the “rotator cuff of the hip” [1–4], the pathoetiology and natural progression of tendon pathology and subsequent muscle atrophy in these muscles are less well understood. Some tendons are prone to complete tears while others degenerate with thickening and subsequent fraying, partial thickness tearing, and thinning with associated muscle atrophy and dysfunction [40]. As the gluteus medius and minimus muscles are known to play an important role in normal gait, dysfunction of these muscles may play a significant role in increased risk of falls and subsequent fractures [24–29]. Our study aims to characterize the prevalence of tendon pathology and muscle atrophy with advancing age as well to investigate the distribution of the involvement of the tendon insertions and location of muscle atrophy.

Gluteus medius and minimus tendon pathology and muscle atrophy increase with age above 50 years. There appears to be progression of the gluteus medius and minimus pathology from tendinosis to tendon tears with advancing age with associated progression in muscle atrophy. There is a positive association between greater trochanteric bursitis and tendon pathology with advancing age, suggesting that tendon pathology may often be the true underlying cause of greater trochanteric pain syndrome. Of note, no full thickness tears of either the gluteus medius or minimus tendons were observed in our study, indicating that these tendons may undergo a gradual process of degeneration with repetitive injury and repair leading to dysfunction and atrophy rather than rupture. The tensor fascia lata was selected as an internal control for nerve-related atrophy given that the superior gluteal nerve innervates the gluteus medius, gluteus minimus, and tensor fascia lata. The iliopsoas was selected as an internal control for generalized muscle atrophy as it rarely atrophies in isolation and is innervated by a separate nerve supply from the gluteus medius and minimus. Disproportionate tendon pathology and atrophy of the gluteus medius and minimus muscles about the hip suggest that atrophy of the gluteus medius and minimus is more strongly associated with tendon pathology than generalized atrophy or denervation/nerve impingement. Anatomic compression of the tendons as they drape over the greater trochanter may cause gradual wear of the tendons with age and explain the disproportionate atrophy of these muscles about the hip in addition to age-related changes related to generalized neuropathy, denervation, or nerve impingement.

Of the abductors, tendinopathy and atrophy of the gluteus medius may have more clinical significance in symptomatic patients [40]. In our study, isolated involvement of the superoposterior facet insertion of the gluteus medius was demonstrated in a higher proportion in younger patients with involvement of both insertions in the older patients, suggesting that the superoposterior insertion is the first to degenerate. For isolated involvement of the superoposterior facet insertion, there was greater atrophy in the anterior muscle bulk or in both anterior and posterior muscle bulk compared to isolated posterior muscle bulk atrophy. However, the majority of cases with isolated superoposterior or lateral facet tendon involvement demonstrated no muscle atrophy, suggesting that one tendon attachment may be sufficient to prevent atrophy of the overall muscle bulk. Degeneration of the both insertion sites demonstrated both anterior and posterior muscle atrophy in the highest proportion of cases. For the gluteus minimus, atrophy of the posterior portion of the gluteus minimus may have more clinical significance in symptomatic patients [41]. In our study, the atrophy of the anterior portion of the gluteus minimus muscle was more frequent in patients less than 70 years old with greater involvement of both the anterior and posterior minimus in patients greater than 70 years old. The etiology of the distribution of gluteus minimus atrophy is unclear as only one insertion of the gluteus minimus is visible by MRI.

There are certain limitations to our study methodology. Our study was a retrospective review of gluteus tendon pathology and muscle atrophy, limiting our ability to demonstrate direct causality between tendon tears and muscle atrophy. An attempt was made to perform a subgroup analysis in patients with multiple studies over time to evaluate the progression of tendinopathy and atrophy in individuals over time; however, there was not an adequate number of patients with repeat studies in the timeframe of our approved IRB to provide useful analysis. Our study was also limited by the number of patients in the older groups above age 70 years. No correlation could be made with the grading of tendon and muscle pathology to histological or pathological specimens as these data were not available. Interobserver variability in the assessment of tendon pathology and muscle atrophy was not included in our study; previous literature has demonstrated the reliability of the Goutallier staging system by MRI [37–39]. Lastly, MR imaging in this study did not include 3.0-T systems as this was not a routine protocol at our institution.

Conclusion

In conclusion, gluteus medius and minimus tendon pathology appears to progress from tendinosis to low-grade partial tears to high-grade partial tears with corresponding progression in muscle atrophy with advancing age. Given the important function of the gluteus medius and minimus in normal gait and the association of muscle atrophy with fall-related hip fractures, more aggressive therapy of tendon tears to prevent atrophy and subsequent falls could be beneficial.