Introduction

Retro-trochanteric pain has been attributed to as the “piriformis syndrome” and can often be unrecognized or misdiagnosed. In 1928, Yeoman described a syndrome where he proposed that arthritic changes in the sacroiliac joint may cause sciatic pain due to secondary inflammation of the piriformis muscle [21]. The concept “piriformis syndrome” was later introduced by Robinson in 1947 [16].

The syndrome is characterized by buttock and lower extremity pain caused by a hypertrophic or inflamed muscle, which increases the pressure from the small external rotators of the hip towards the sciatic nerve [2, 3, 6, 13, 15]. The aetiology of the syndrome is not clearly known, although it has been argued that the pain may occur secondary to blunt trauma to the buttock or gluteal region [3, 12]. Anatomical abnormalities of the piriformis muscle or the sciatic nerve as bipartite piriformis muscle and piriformis muscle lying anterior to the nerve have been described as irritating the sciatic nerve [3, 4, 6, 9, 17]. It has also been described that the cause could be a recurrent problem after spinal surgery [4]. Benson et al. [3] found adhesions between the piriformis muscle and the sciatic nerve. Cox et al. [7] argued that the gemelli-internal obturator muscles and the associated bursa should be considered as possible sources of retro-trochanterically located sciatic-like pain. Overuse of the piriformis muscle was suggested to contribute to the “piriformis syndrome” by Mayrand et al. [12]. The “piriformis syndrome” has also been reported following hip replacement [19]. There are no laboratory or radiographical methods to establish a diagnosis of the syndrome [1, 10]. Windisch et al. [20] argued that the variations of the piriformis muscle were not mentioned in radiological text books, which are of importance for the interpretation of ultrasonographic and magnetic resonance tomography images. In an anatomical cadaver study, the authors found that 48 of 112 cadaveric specimens had a fusion of the piriformis and the internal obturator muscles [20].

The role of the internal obturator in retro-trochanteric pain syndrome is not especially well described in the literature and suggested treatment options for this type of pathology are seldom seen. A palpation test with tenderness over the insertion of the external rotators behind the trochanter major appears to be the main diagnostic sign [10, 13, 18]. Pain and weakness on forced passive internal rotation of the extended hip (Freiberg’s sign) and pain and weakness on resisted abduction and external rotation of the thigh in sitting position (Pace’s sign) are also important findings. A number of methods exist for the treatment of the “piriformis syndrome” in the hip region. These include physical therapy [7, 11, 12], injection of anaesthetic agents with or without steroids [3, 14], and surgical release of the tendon [6, 13, 18]. Dezawa et al. [8] described an arthroscopic technique for the release of the piriformis tendon.

In 2003, Meknes et al. [13] reported significant short-term pain reduction both at 6 and 12 weeks after surgery in patients with radiating pain in the buttock after sectioning the internal obturator muscle. However, at 6 months, the reduction in pain was no longer statistically significant. The aim of the present study was to perform a long-term clinical and radiographic follow-up of the patients in the randomized study by Meknas et al. [13]. The hypothesis of the study was that surgical release of the internal obturator muscle would render a long-term decrease in pain in patients with retro-trocanterically located pain in the buttock.

Patients and methods

Twelve patients, three males and nine females, mean age 47 (25–66) years with retro-trochanteric pain in the buttock, radiating distal to the knee and intolerance to sit more than 40 min, were included in a prospective, randomized trial. At inclusion the patients were randomly allocated by sealed envelopes to either operative treatment or a control group. The median duration of symptoms was 7.5 (2–20) years (Table 1), and all patients had undergone various conservative treatments with either physiotherapy or injections with local anaesthetic agents combined with steroids before inclusion in the study.

Table 1 The demographics of the patients at inclusion
Full size table

At inclusion, 6 weeks, 12 weeks, 6 months and at 8 years the patients were tested for pain and tenderness when examined with deep palpation over the small external rotators dorsal to the trochanter major and over the sciatic nerve in the same region. The clinical examinations also included tests for pain and weakness on resisted abduction and external rotation of the thigh in a sitting position (Pace’s sign) (Fig. 1), pain and weakness on forced passive internal rotation of the extended thigh (Freiberg’s sign) (Fig. 2) and buttock and leg pain during passive straight leg raising (Lasegue’s sign). Furthermore, an evaluation of limping and of problems at walking and sitting was performed. All the above tests were categorically classified by the patient as positive (pain/problems) or negative (no pain/no problems). The 6- and 12-week results have been reported in a previous publication by Meknas et al. [13].

Fig. 1
figure 1

The Pace’s sign. The patient is in the sitting position. During resisted abduction and external rotation of the thigh, the small rotators of the hip are stretched. The test is classified as positive if pain is registered (illustration photograph)

Full size image
Fig. 2
figure 2

The Freiberg’s sign. The patient is in the supine position with the thigh extended. The leg and thigh are passively internally rotated by the examiner. The test is classified as positive if pain is registered (illustration photograph)

Full size image

Patients with obvious spinal pathology, cancer or severe organic diseases, patients older than 70 years and patients who suffered from mental illness preventing them from following simple rehabilitation instructions were excluded from the study. At inclusion and at 8 years, the patients underwent standard antero-posterior radiographs of the pelvis and hips, and lateral view of the hips (bilaterally) and either CT using a Siemens Somatom Sensation (Siemens AG, Erlangen Germany) or MRI using a Philips Intera 1.5 Tesla (Royal Philips, Electronics Amsterdam Netherlands) of the lumbar spine (Table 2).

Table 2 The radiographic findings before the treatment and 8 years after treatment
Full size table

The clinical examinations at 8 years as well as all radiographic assessments were performed by independent observers who were not involved in the treatment of the patients.

In the present study, pain was registered using the visual analogue scale (VAS) preoperatively, at 6 months and at 8 years. The use of analgesic and anti-inflammatory drugs and the level of work for each patient at inclusion, 6 months and at 8 years were registered and classified according to Tables 3 and 6. All patients provided written consent, and the study protocol was approved by the National Committee for Research Ethics in Norway.

Table 3 The classification of pain medication
Full size table

Surgical treatment and regimen in the control group

An explorative operation was done using a postero-lateral approach in the hip region. The fascia lata was split, the external rotators and the sciatic nerve were identified (Fig. 3a, b). An examination of the anatomy as well as the relationships between structures during passive flexion, internal rotation, and the Lasegue’s test were performed during the operation. The sciatic nerve was found to pass anterior to the piriformis muscle in all cases in the surgical group, with no part of the nerve passing through this muscle. As for the superior gemellus and internal obturator muscles, the sciatic nerve passed behind these in all cases in the surgical group. The internal obturator muscle was found tense, hyperaemic and in close contact with the sciatic nerve (Fig 3a). The nerve was flattened and slightly hyperaemic. When performing the Lasegue manoeuvre on the operating table, the internal obturator, and not the piriformis muscle, impinged on the sciatic nerve at an early stage during hip flexion. To relieve the tension towards the sciatic nerve from the internal obturator muscle, sectioning of the tendon was performed at its insertion to the greater trochanter. An immediate release of the tension towards the scatic nerve during the Lasegue’s manoeuvre was observed after sectioning of the tendon (Fig 3b). Prophylaxis against infection was administered intravenously using 2 g of Cefalotin (ACS Dobfar Generics Luxembourg) just before the operation. Weight bearing supported by crutches was allowed immediately after surgery. A gradual increase in activity as tolerated by the patients was allowed. The patients underwent no formal sessions of physiotherapy. No new approaches for treatment were initiated in the control group during the observation period.

Fig. 3
figure 3

a The sciatic nerve and the internal obturator tendon as found during operation for retro-trochanteric pain syndrome. The internal obturator tendon is tense and hypertrophic, lying in close contact with the sciatic nerve, which turns sharply over the tendon. b After sectioning of the internal obturator tendon the sciatic nerve is released from the tendon. Used and modified with permission from IASP® (International Association for the Study of Pain®), originally published by Meknas et al. [13]

Full size image

Statistical methods

The power analyses was performed before the start of the long-term follow-up using the knowledge from the short term study by Meknas et al. [13]. The decrease in pain in the treatment group as measured with the VAS was the primary variable. It was hypothesized that there would be a mean long-term decrease in the pain score of 3 on the VAS with a standard deviation of 1.5 compared with the preoperative values. With the alpha value set at 0.05 and the power at 80%, the required sample size would be four patients in the treatment group. Based on these calculations, it was decided to proceed with the long-term follow-up.

Mean (SD) values are reported for the VAS and median (range) values for the other variables. The repeated measures ANOVA test and the Fisher’s post hoc test were used to analyse the change over time in terms of the VAS for pain within the study groups. For all other ordered variables within group comparisons were made using the Wilcoxon test. Dichotomous variables were analysed using the Fisher’s exact test.

Results

The demographics of the patients are presented in Table 1. At inclusion, there were no significant differences between the study groups in terms of gender, age and duration of symptoms. At inclusion, all patients underwent both clinical examinations and radiographic examinations of the hips and either CT or MRI of the lumbar spine (Tables 2, 4). At 6 months, all patients underwent clinical examinations (Table 4). At 8 years, one patient in the surgical group had died and one patient in each group was lost to follow-up. The remaining four patients in the surgical group and five patients in the control group underwent clinical (Table 4) and radiographic (Table 2) examinations at 8 years. No per or postoperative complications or re-operations were registered during the period of the study.

Table 4 The clinical examination tests at inclusion, after 6 months and after 8 years
Full size table

At 8 years, a significant decrease in pain was found in the surgical group but not in the control group (Table 5).

Table 5 The VAS at inclusion, 6 months and 8 year
Full size table

All clinical examination tests at inclusion, 6 months and 8 years are presented in Table 4. At 8 years, the Lasegue’s test was significantly better compared with the findings at inclusion in the control group.

All individual radiographic findings at inclusion and at 8 years are presented in Table 2. At inclusion and at 8 years, no pathology was found in the hip joints in neither group. Minor degenerative changes were found at the L3–L5 level in 4/12 patients at inclusion. At 8 years, the corresponding changes were found in 7/9 patients (P = 0.025).

The level of pain medication decreased significantly in the whole study cohort at 6 months (P = 0.03) and at 8 years (P = 0.02) compared with the levels at inclusion. If the study groups were analysed separately, the decrease was only significant in the surgical group at 6 months (P = 0.04) (Table 6).

Table 6 Levels of pain medication and work at inclusion, 6 months and 8 years
Full size table

Two patients in the surgical group were working full and half time, respectively, at 6 month follow-up; both patients still worked half time at 8 years. Before the start of the study the patients had been out of work for 3 and 10 years, respectively (Table 6).

Discussion

Twelve patients with diffuse retro-trochanterically located radiating pain in the hip and thigh region were enrolled in this prospective randomized study.

The principal finding of the present study was that a significant decrease in pain was found in the surgical group but not in the control group, 8 years after sectioning of the internal obturator tendon. Meknas et al. [13] have previously reported that patients with symptoms and findings similar to the “piriformis syndrome” had a remarkable and significant decrease in pain during the first months after surgical release of the internal obturator muscle tendon. A corresponding finding was not seen in the control group. However, after 6 months, the decrease in pain was no longer significant. The reason for this is unknown. One explanation could be that the operative procedure included sectioning of the internal obturator tendon, and also exploration of the sciatic nerve, which might have caused secondary scar formation.

The radiographic findings are also interesting. A significant increase in minor degenerative changes in the lumbal columna was found in the whole study cohort. This finding suggests that retro-trochanteric pain syndrome might be associated with early degenerative changes in the lumbal columna, and for diagnostic reasons, changes in the lumbal columna must be taken into consideration.

A number of methods exist for the treatment of the “piriformis syndrome” with variable results; however, no particular treatment has resulted in long-term improvement. Cox et al. [7] suggested that distraction and manual stretching of the gemmeli-internal obturator and piriformis muscles are successful for treating “retro-trochanteric pain syndrome”. Keskula et al. [11] described the importance of stretching exercises for the “piriformis syndrome” and Mayrand et al. [12] considered chiropractic care and muscle stretching beneficial. Benzon et al. [4] recommended an injection technique with special placement of the needle to avoid damage to the sciatic nerve and Mullin et al. [14] reported significant pain relief after injection of cortico-steroids and local anaesthetics in 12 patients with a follow-up period of 9–24 months.

In two patients, Solheim et al. [18] found complete pain relief immediately after surgical release of the piriformis muscle with 11 and 10 months of follow-up, respectively. In a technical note, Dezawa et al. [8] reported good results in six patients operated with an arthroscopic release of the piriformis tendon. To our knowledge, there are no other randomized studies on surgical treatment of retro-trochanteric pain syndrome found in the literature except the study by Meknas et al. [13].

Anatomically, the internal obturator muscle is deep to both the piriformis muscle and the sciatic nerve and it runs parallel to the piriformis in its attachment to the trochanter major. Because of its proximity, similar pathway and similar function, most treatments for patients with “piriformis syndrome” would affect the internal obturator muscle as well [5].

In an anatomical cadaver study, Windisch et al. [20] found that 48 of 112 cadaveric specimens had a fusion of the piriformis and the internal obturator muscles. Their finding is in line with the present observation that the internal obturator muscle can cause symptoms similar to those of the “piriformis syndrome”. There is some confusion in the literature since retro-trocanterically located pain has been given different names such as “piriformis syndrome”, “internal obturator syndrome”, “small external rotator syndrome”, or other similar names. We suggest that the syndrome consisting of symptoms and findings as described in the present study should be named “retro-trochanteric pain syndrome” in the future.

Larger and controlled studies should be set up to thoroughly evaluate the treatment of this type of pathology. The present study was a randomized trial which is of course a strength of the study. One more strength of this study is that independent examiners performed the long-term clinical examinations and all radiographic assessments.

A weakness of the study is the small number of patients involved and that the study was not designed with enough power to compare the surgical treatment group with the control group. In addition, the patients did not undergo Lauenstein radiographs or MRI evaluation of the hip. Although it is unlikely, there is a possibility that the patients could have suffered from femuro-acetablular impingement or labral tears in their hip joint.

Conclusion

Surgical release of the internal obturator muscle resulted in long-term decrease in pain in patients with retro trochanteric pain syndrome. Thus, the hypothesis of the present study could be verified.