Radiological Anatomy of the Obturator Nerve and Its Articular Branches: Basis to Develop a Method of Radiofrequency Denervation for Hip Joint Pain

Transcription

1 Blackwell Publishing IncMalden, USAPMEPain Medicine American Academy of Pain Medicine? Original ArticleRadiofrequency Denervation of Hip JointLocher et al. PAIN MEDICINE Volume 9 Number Radiological Anatomy of the Obturator Nerve and Its Articular Branches: Basis to Develop a Method of Radiofrequency Denervation for Hip Joint Pain Stephan Locher, MD,* Helge Burmeister, MD, Thomas Böhlen, MD,* Urs Eichenberger, MD,* Christophoros Stoupis, MD, Bernhard Moriggl, MD, Prof, Klaus Siebenrock, MD, Prof, and Michele Curatolo, MD, Prof* *Department of Anesthesiology, Division of Pain Therapy, University Hospital of Bern; Department of Orthopedics, University Hospital of Bern; Department of Diagnostic Radiology, University Hospital of Bern, Inselspital, Switzerland; Department of Anatomy, Histology and Embryology, Innsbruck Medical University, Innsbruck, Austria ABSTRACT ABSTRACT Objective. A previous study of radiofrequency neurotomy of the articular branches of the obturator nerve for hip joint pain produced modest results. Based on an anatomical and radiological study, we sought to define a potentially more effective radiofrequency method. Design. Ten cadavers were studied, four of them bilaterally. The obturator nerve and its articular branches were marked by wires. Their radiological relationship to the bone structures on fluoroscopy was imaged and analyzed. A magnetic resonance imaging (MRI) study was undertaken on 20 patients to determine the structures that would be encountered by the radiofrequency electrode during different possible percutaneous approaches. Results. The articular branches of the obturator nerve vary in location over a wide area. The previously described method of denervating the hip joint did not take this variation into account. Moreover, it approached the nerves perpendicularly. Because optimal coagulation requires electrodes to lie parallel to the nerves, a perpendicular approach probably produced only a minimal lesion. In addition, MRI demonstrated that a perpendicular approach is likely to puncture femoral vessels. Vessel puncture can be avoided if an oblique pass is used. Such an approach minimizes the angle between the target nerves and the electrode, and increases the likelihood of the nerve being captured by the lesion made. Multiple lesions need to be made in order to accommodate the variability in location of the articular nerves. Conclusions. The method that we described has the potential to produce complete and reliable nerve coagulation. Moreover, it minimizes the risk of penetrating the great vessels. The efficacy of this approach should be tested in clinical trials. Key Words. Radiofrequency Denervation; Hip Joint Pain; Anatomy; Obturator Nerve Introduction H ip joint pain is a common condition, with an estimated prevalence of 7% in men and 10% in women, in a population sample aged over Reprint requests to: Michele Curatolo, MD, PhD, Department of Anesthesiology, Division of Pain Therapy, Inselspital, 3010 Bern, Switzerland. Tel: ; Fax: ; michele.curatolo@insel.ch. 45 years [1]. Hip pain has been shown to have a considerable impact on health status, with impairment of physical functioning and need for medical treatment already before the first consultation [2]. Patients who fail to respond to conservative treatment usually undergo surgery, i.e., hip replacement. However, not all patients can be operated on, mostly because of poor health status or the patient s unwillingness to undergo major surgery. American Academy of Pain Medicine /08/$15.00/ doi: /j x

2 292 Locher et al. Radiofrequency destruction of the obturator nerve for hip joint pain has been described [3], but the description of clinical outcomes in that study does not allow an evaluation of the effectiveness of the method. Moreover, a loss of sensation was observed in nearly all patients. Radiofrequency denervation of the articular branches of the obturator and femoral nerve has been proposed as an alternative treatment for hip joint pain, and appeared effective in a retrospective uncontrolled study [4]. This report was relevant, because it opened a new perspective for the treatment of hip joint pain. Selectively targeting articular nerves and sparing the parent trunk avoids unnecessary sensory and motor loss. However, it is unclear whether the described surgical method was based on an accurate investigation of the anatomy of the target nerves. Radiofrequency electrodes produce lesions circumferentially around the shaft of the active tip. The radius of the lesion is approximately equal to one electrode-width and no lesion is produced distal to the electrode tip [5]. It is therefore mandatory that the needle be placed as close as possible to the target nerve and, whenever feasible, parallel and not perpendicular to it [6]. Because of possible anatomical variations and because nerves cannot be seen under fluoroscopy, multiple lesions may be necessary to increase the chance of coagulating the nerve. The fact that only one lesion was produced in the aforementioned study on hip pain [4] may explain the finding that complete pain relief was not achieved in all patients. The aim of the present study was to define a rational surgical approach for radiofrequency denervation of the articular branches of the obturator nerve that could be tested in clinical trials. For this purpose, the anatomy of the obturator nerve and its branches was studied by cadaver dissection. Because the denervation procedure relies on the use of fluoroscopy, the nerves were marked by wires and the radiological relationship of the nerves to the bone structures on fluoroscopy was analyzed. Furthermore, a magnetic resonance imaging (MRI) study was undertaken to describe the anatomical structures that would be encountered by the radiofrequency electrode during different possible percutaneous approaches. The anatomy of the articular branches of the femoral nerve, which also supplies the hip joint and can be a target for radiofrequency treatment [4], were not investigated. Methods Anatomical Dissection Ten embalmed cadavers (3 men and 7 women) were studied at the Institute of Anatomy of the University of Bern, Switzerland. Four of them were dissected on both sides, six only unilaterally because they had been dissected previously by medical students. All dissections were performed by an anesthesiologist with experience in radiofrequency denervation of the hip joint and an orthopedic surgeon experienced in hip surgery. A physician from the anatomical institute was available for support. An anesthesiologist experienced in interventional pain management was in charge for the fluoroscopy imaging. A combination of a Smith-Petersen and a Pfannenstiel approach was chosen in order to dissect the pathway of the obturator nerve and its articular branches to the hip joint [7]. The sartorius muscle and the rectus femoris muscle were dissected from their origin at the anterior superior and inferior iliac spines, and were mobilized distally. The iliopsoas muscle and the external iliac vessels were cut at the level of the linea arcuata and were mobilized distally. The iliocapsularis muscle [8] was carefully dissected from the capsule of the hip joint. The pectineus muscle was dissected from its origin at the pectineal line of the superior ramus of the pubis. The external obturator muscle was then dissected from the membrane of the obturator foramen. After the rectus abdominal muscle was detached from the pubic crest, an osteotomy of the superior ramus of the pubis was performed, in order to better expose the obturator nerve. A segment of approximately 1.5 cm in size was removed to visualize the obturator nerve from its intrapelvic course through the obturator canal. Then the nerve and its branches were carefully prepared distally using magnifying glasses with a 2.3 factor. Metal wires were placed at the obturator nerve and at the upper and lower edge of its articular branches. The width of the band of articular branches after their origin from the obturator nerve (medial end of the band) and before penetrating the articular capsule (lateral end of the band), as well as the length of the branches from their medial to their lateral end were measured. Fluoroscopy After the metal wires were placed at the nerves, anteroposterior and lateral fluoroscopic imaging were recorded using an OEC workstation (GE

3 Radiofrequency Denervation of Hip Joint 293 OEC Medical Systems, Inc., Salt Lake City, UT). Then, the fluoroscopy beam was directed with the same orientation as the articular branches of the obturator nerve, so that the wires lying on these nerves appeared as small as possible on the fluoroscopy image. This was carried out in order to define the site of puncture for the radiofrequency electrode if it were to be placed parallel to the targeted nerves. The angle of the fluoroscopy beam to the frontal, transverse, and sagittal plane were recorded using the goniometer of the C-arm. This was carried out in order to provide information concerning the electrode direction for a radiofrequency procedure. As has been shown for coagulation of the nerves that supply the cervical zygapophysial joints [9], a matrix of several lesions is required in order to cover the whole region where the nerves are expected to be, taking into account the anatomical variability of the nerve location. To define the bony landmarks for the multiple electrode placement, a drawing of the bony structures on which the articular branches lie was made and the position of the wires placed at the obturator nerve and its articular branches were drawn on it. MRI Magnetic resonance images were obtained from 20 consecutive diagnostic pelvic procedures on patients using the archives of the Department of Radiology of the University of Bern, Switzerland. Ideal electrode positioning was simulated on the images using two approaches: 1) the anteroposterior approach described by Kawaguchi et al. [4], by which the electrode is introduced on an anteroposterior line, perpendicular to the frontal plane; and 2) an oblique approach lateral to the femoral vessels and the femoral nerve. The anatomical structures that would be penetrated by the electrode with each approach were identified by plotting a line that corresponded to the electrode path on the MRI scans. The percent of scans in which the line intersected the femoral artery or vein using the anteroposterior approach was calculated. The distance between the primary target of a radiofrequency procedure (see Results of MRI analysis for its definition) and the posterior aspect of the bundle of femoral vessels and nerve was measured by electronically generated distance points in the archive system. Data Analysis All data were analyzed by descriptive statistics and presented as median, 25th and 75th percentiles, minimum, and maximum values. The analyses were made on all the 14 dissections performed on the 10 analyzed cadavers. Results On both sides of one cadaver, no articular branches were detected, probably because of errors in dissection. In all other 12 dissections, articular branches from the obturator nerve to the hip joint numbered between two and seven. They radiated across the front of the hip joint across band-like areas whose disposition varied from specimen to specimen (Figure 1). Each band had a stem located lateral to the obturator foramen, just below the acetabulum. On radiographs, the stem was located immediately below the teardrop silhouette formed by the anterior inferior of the acetabulum (Figure 1). The width of the band ranged from 9 to 19 mm at its stem, and from 7 to 38 mm at its lateral end (Table 1). On MRI scans, the lower end of the anterior lip of the acetabulum corresponds with the upper corner of the medial end of the band of distribution of articular nerves. We refer to this point as the primary target of the radiofrequency procedure. The femoral vessels lie directly anterior to this point (Figure 2). Therefore, there is a high likelihood of puncturing one of these vessels if a radiofrequency electrode was inserted along an anteroposterior line to these nerves, as proposed Figure 1 Articular branches of the obturator nerve and target region for radiofrequency denervation. Covering the right hip are tracings of the anteroposterior projections of the metal wires used to mark the location of the articular branches in cadavers. In each cadaver, articular branches were spread across band-like areas. The bold lines represent the upper boundary of each area, and the dotted lines represent the lower boundary. The stem of each band was located below the teardrop shape of the inferior end of the acetabulum. Over the left hip, the matrix of lesions required to coagulate the articular branches is illustrated. Its medial margin lies opposite and below the teardrop silhouette of the acetabulum. For reference, tracings of wires covering the obturator nerve have been depicted.

4 294 Locher et al. Table 1 Dimensions of the articular branches of the obturator nerve and their bands of distribution, as found in 11 cadavers, eight left side and three right side Minimum 25th Percentile Median 75th Percentile Maximum Width of nerve band, lateral end (mm) Width of nerve band, medial end (mm) Length of articular branches (mm) Figure 2 Magnetic resonance imaging (MRI) of the right pelvic area. The primary target for electrode positioning is the bony structure corresponding to the teardrop on an anteroposterior fluoroscopic view. To reach this target and avoid the femoral nerve-vessel bundle, the electrode should be introduced using an oblique pass. Based on our 20 MRI analyses, we recommend that the electrode be introduced at an angle of 70 with the sagittal plane and 20 with the transversal plane in order to always avoid the vessels and minimize the angle between the electrode and the target nerves. A = acetabulum; FA = femoral artery; FN = femoral nerve; FV = femoral vein; G Min = gluteal minor muscle; G Medical = gluteal medial muscle; Head = head of femur; IP = iliopsoas muscle; P = pectineal muscle; RF = rectus femoris muscle of the quadriceps muscle; S = sartorius muscle; α = recommended angle of 70 with the sagittal plane to introduce the electrode. by Kawaguchi et al. [4]. In the 20 MRI scans analyzed, a hypothetical electrode inserted in this manner incurred the femoral vein in 11 cases, the femoral artery in five, and both vessels in three cases. In only one case did the electrode avoid a vessel. However, puncture of vessels can be avoided if the electrode is introduced lateral to the femoral vessels. In the 20 MRI scans analyzed, a line drawn from a point just lateral to the femoral vessels to the target point formed an angle with the sagittal plane whose median, minimum, and maximum value was 39, 2.2, and 57.6, respectively. This means that the trajectory of the electrode should form an angle of at least 57.6 with the sagittal plane in order to avoid puncture of the vessels in the cases that we analyzed. For our proposed technique we suggest that an angle of 70 be chosen in order to account for possible further anatomical variabilities concerning vessel location and to minimize the angle between the electrode and the target nerves (see Discussion) (Figure 2). The median distance between target point and the posterior aspect of the great vessels in the MRI scans were 11.4 mm, with a range of 7 21 mm. This suggests that there is no danger of injuring the vessels during the coagulation procedure. Based on these observations, an oblique approach, designed to deliver an electrode parallel to the plane of the articular nerves and to avoid vessel puncture, was developed. Proposed Technique The target for the sites of radiofrequency lesions are illustrated in Figure 1. Data from specimen 2 and 8 (three dissections) are not included in this analysis because of failure to identify nerves, and technical radiological failure, respectively. Because the depth of the target nerves is variable, the active tip of the electrode should be long, at best 10 mm, in order to maximize the chance of coagulating them. We propose that first a spinal needle (e.g., 21 G) be introduced under anteroposterior fluoroscopic view in a parasagittal plane (Figures 3 and 4). The skin point of entry is the projection of the caudal-lateral edge of the teardrop in an anteroposterior view (Figure 3). The teardrop figure appears on an anteroposterior fluoroscopic view (Figure 3) and is formed by a lateral, medial, and caudal line corresponding to the wall of the acetabulum, the wall of the lesser pelvis, and the acetabular notch, respectively. The femoral artery is palpated and, if on the trajectory of the needle, a more medial site of puncture is chosen. This needle is used to anesthetize the nerves before heating them and as a reference for the target of the electrode positioning. A possible puncture of a femoral vessel with such a thin needle is not expected to lead to clinically significant hematoma in the absence of coagulation disorders. Then the beam of the fluoroscopy device is turned from an anteroposterior view in lateral direction until forming an angle of 70 with the

5 Radiofrequency Denervation of Hip Joint 295 Figure 3 This and the following images show the steps of the proposed radiofrequency technique. Here a 21 G needle was introduced under an anteroposterior view to anesthetize the articular branches of the obturator nerve. The needle tip also serves as reference to guide the radiofrequency electrode to the target nerves. sagittal plane. According to our MRI analysis, this angle should be wide enough to avoid the risk of puncturing femoral vessels and nerve. Then the fluoroscopy device is turned in a cranial direction to form an angle of 20 with the transverse plane. In that way, the fluoroscopy beam will be almost parallel to the course of the articular branches (Figure 1). In order to introduce the electrode in the direction of the fluoroscopy beam, the site of puncture is the projection of the tip of the 21 G needle on the skin using this 70 oblique and 20 craniocaudal view. After local anesthesia of the skin, the electrode is introduced in a tunnel view (i.e., in the direction of the fluoroscopy beam) toward the tip of the 21 G needle, i.e., the caudallateral edge of the teardrop, until bone is contacted. Figure 5 shows the electrode in place in the tunnel fluoroscopy view. Then, the fluoroscopy device is turned back to the anteroposterior view. The electrode is slipped caudal to the contacted bone and 5 mm deeper. Care should be taken that the tip of the electrode remains lateral to the teardrop in an anteroposterior view, in order not to Figure 4 Lateral view of the anesthetic block of the articular branches of the obturator nerve. N C Figure 5 The fluoroscopy beam was turned from an anteroposterior view in lateral direction until forming an angle of 70 with the sagittal plane, then in a cranial direction to form an angle of 20 with the transverse plane. In that way, the fluoroscopy beam is almost parallel to the course of the articular branches (Figure 1). N: spinal needle for the local anesthesia of the articular branches. E: radiofrequency electrode inserted into the target in a tunnel view, i.e., in the direction of the fluoroscopy beam. C: cable of the radiofrequency electrode. The site of puncture for the electrode insertion is the projection of the tip of the 21 G needle (N) on the skin using this view. After local anesthesia of the skin, the electrode (E) is introduced in the tunnel view toward the tip of the 21 G needle (N). E

6 296 Locher et al. Figure 6 The fluoroscopy device has been turned back to the anteroposterior view. The electrode is positioned for the first lesion. Before the first coagulation is performed, the spinal needle must be withdrawn in order to avoid dispersion of the heat along the needle. The technique for the further electrode positions (Figure 1) is described in the discussion section. coagulate the obturator nerve (Figure 1). Previous local anesthetic block may prevent detection of coagulation of that nerve. This is the first site of lesion (Figure 6). The electrode is then withdrawn by 10 mm and redirected more posteriorly until contact with the joint capsule is felt. Here, a second lesion is performed. In this way, a 20-mmlong lesion is produced, possibly adjacent to the joint capsule where the articular branches are expected to be. A third more lateral lesion can be considered to increase the chance to coagulate a longer portion of the nerves and delay nerve regeneration, thereby prolonging the duration of pain relief. In order to coagulate all the articular branches and considering the anatomical variations of the nerves location, further lesions caudal to the aforementioned ones are mandatory (Figure 1). The depth of the electrode at the first two lesions on the lateral view is recorded, e.g., by printing the images. This will serve as reference for the further caudal lesions. The electrode is then positioned 2 3 mm caudal to the previous location, depending on the diameter of the electrode. Because the size of the lesion around the active tip of the electrode is approximately equal to one electrode-width [5], the wider the diameter, the wider the size of the lesion, thereby allowing a higher distance between two adjacent lesions. At these places, two or three lesions are made at the same depths as for the first two to three lesions, the position being controlled on the lateral view, having as reference the documented electrode positions of the first two to three lesions. The procedure is repeated by further caudal placements of the electrode. The most caudal position should be at a horizontal line crossing the junction between the middle and caudal third of the obturator foramen in the anteroposterior view (Figure 1). As shown in Figure 1, the site of the lesions from the cranial to the caudal ones is progressively moved laterally. This is carried out to minimize the risk of coagulating the obturator nerve, which has a slightly mediolateral course. The number of parallel lesions from the cranial to the caudal end of the target region depends on the diameter of the electrode: the bigger the diameter, the higher the distance between two adjacent lesions, the lesser the number of lesions that needs to be produced. Discussion General Aspects We found that the width of the bundle of the articular branches is up to 1.9 and 3.8 cm at their medial and lateral end, respectively (Table 1). Furthermore, there is substantial anatomical variation in the location of the branches (Figure 1). As a result, it is unlikely that a single lesion as proposed by Kawaguchi et al. [4] would reliably coagulate all the branches. We strongly recommend that a matrix of lesions be produced to cover all possible nerve locations. Theoretically, the ideal technique of electrode placement would bring the active tip parallel to the articular branches of the obturator nerve. Because of the course of these branches, the technique would imply a skin point of entry of the electrode on the lateral aspect of the thigh. The electrode would then be introduced medially, passing through the gluteus medius and minimus muscles, the rectus femoris of quadriceps femoris muscle, the iliopsoas muscle, and finally posterior to the great femoral vessels. Because the anatomical variations of the location of the articular branches were high, we had to abandon this method. While on the anteroposterior view the bony landmarks were clear (Figure 1), this was not the case for the lateral view, where no constant reference for the location of the nerves could be identified. As a result, in order to account for the anatomical variations, a matrix of lesions in a three-dimensional space would have to be made. This would render the procedure exceedingly long.

7 Radiofrequency Denervation of Hip Joint 297 A possible approach is to reach the articular branches from the anterior aspect of the thigh, introducing the electrode in a parasagittal plane, as performed by Kawaguchi et al. [4]. As mentioned above, however, one of the great femoral vessels or the femoral nerve would be punctured in almost every case (Figure 2). While the femoral artery can be palpated and therefore avoided, the femoral vein and the femoral nerve cannot. Although no complications such as hematomas or nerve damage are described by Kawaguchi et al. [4], the number of cases reported in that study is too small to state that the method is not associated with increased risk of clinically significant complications. Furthermore, while the method described by Kawaguchi et al. involves a single lesion [4], we recommend that multiple lesions be performed. The resulting multiple electrode positioning would increase the risk of hematoma or nerve damage. We therefore suggest that an oblique pass be adopted for radiofrequency denervation of the articular branches of the obturator nerve. This approach avoids puncture of the great femoral vessels and of the femoral nerve, would position the electrode almost parallel to the target nerves, and would be minimally exposed to the risk of missing the nerves because of anatomical variations. Limitations The principles of the described approach of multiple lesions are similar to those underlying the validated method for radiofrequency neurotomy of the cervical zygapophysial joints [10]. However, with our method, the electrode is not strictly parallel, but forms a small angle to the nerves. As mentioned above, this approach is not ideal because the nerves may be coagulated on a short length, possibly reducing the duration of pain relief due to nerve regeneration. This limitation is partially overcome by performing two to three lesions on the same line, as previously described. The hip joint is supplied not only by the obturator nerve, but also by branches from the femoral nerve, the accessory femoral and accessory obturator nerves when present, the nerve supplying the quadratus femoris muscle, the superior gluteal nerve, and direct branches from the sciatic nerve [11,12]. Radiofrequency denervation of all these nerves is unfeasible, mostly because of the strong anatomical variability and difficult access to the nerves. Kawaguchi et al. [4] also proposed to coagulate the articular branches of the femoral nerve that supply the anterior part of the hip joint. They inserted the electrode caudal and lateral to the anterior superior iliac spine and advanced it to the anterolateral aspect of the hip, where a lesion was made. However, the innervation of the hip joint by branches arising from the femoral nerve is highly variable. The articular branches may arise either from the main trunk of the nerve or from its branches to the rectus femoris, the pectineus, or the vastus lateralis muscle [11,12]. The course of these branches themselves is again variable. In order to cover all these anatomical variations by radiofrequency lesioning, a space of several square centimeters has to be coagulated. Because of the limited size of the lesion around the active tip of the electrode [5], we anticipate that approximately 50 lesions or more would need to be done reliably to denervate the portion of the hip joint supplied by the femoral nerve. This is not feasible in clinical practice. We therefore did not extend our study to the femoral branches, and believe that a reliable coagulation of these nerves using the currently available radiofrequency technology cannot be accomplished clinically. Despite the above limitations, the obturator nerve still supplies an important portion of the hip joint and denervation of its articular branches has the potential to offer substantial pain relief with a relatively simple and minimally invasive procedure. A caution concerning patient selection applies. The proposed method can only be successful in those patients in which a significant proportion of the nociceptive signal is transmitted by the obturator nerve. In order to select them, the paradigm that has been validated for the diagnosis and treatment of zygapophysial joint pain can be adopted [13,14]. The procedure implies the use of prognostic local anesthetic blocks that are performed in a controlled randomized double-blind fashion [13 15]. In the case of zygapophysial joint pain, a positive test predicts at least a 70% success rate of a subsequent radiofrequency denervation [6,16]. Diagnostic intra-articular blocks and/or blocks of the articular branches of the obturator nerve have been performed by Kawaguchi et al. to select patients for radiofrequency denervation [4]. Because of the multiple innervation of the hip joint, intra-articular blocks may have a lower predictive value than the block of the nerves that are to be coagulated. Single diagnostic blocks for cervical and lumbar zygapophysial joint pain are associated with a false positive rate of approximately one-third, probably as a result of a placebo effect or chance; the false positive rate is minimized by

8 298 Locher et al. performing the block in a controlled fashion [17,18]. Because this likely applies to other types of diagnostic blocks, we believe that controlled rather than single blocks of the articular branches of the obturator nerve should be performed for prognostic purposes. Conclusions Radiofrequency denervation of the articular branches of the obturator nerve should be accomplished by multiple lesions to account for the location of the target nerves and their variability. The method that we describe has the potential to produce complete and reliable interruption of nociceptive impulses arising from the hip joint that are conveyed by the obturator nerve. Moreover, the method may be safer than previous approaches because the great vessels are not punctured. The efficacy of this technique needs to be tested in clinical trials. Acknowledgments We thank Dr. Gudrun Hermann, staff member at the Institute of Anatomy of the University of Bern, for her support during the anatomical dissections. We thank the scientific committee of the Department of Anesthesiology of the University of Bern for providing financial support. References 1 Birrell F, Lunt M, Macfarlane G, Silman A. Association between pain in the hip region and radiographic changes of osteoarthritis: Results from a population-based study. Rheumatology (Oxford) 2005;44: Birrell F, Croft P, Cooper C, Hosie G, Macfarlane G, Silman A. Health impact of pain in the hip region with and without radiographic evidence of osteoarthritis: A study of new attenders to primary care. The PCR Hip Study Group. Ann Rheum Dis 2000;59: Akatov OV, Dreval ON. Percutaneous radiofrequency destruction of the obturator nerve for treatment of pain caused by coxarthrosis. Stereotact Funct Neurosurg 1997;69: Kawaguchi M, Hashizume K, Iwata T, Furuya H. Percutaneous radiofrequency lesioning of sensory branches of the obturator and femoral nerves for the treatment of hip joint pain. Reg Anesth Pain Med 2001;26: Bogduk N, Macintosh J, Marsland A. Technical limitations to the efficacy of radiofrequency neurotomy for spinal pain. Neurosurgery 1987;20: Lord SM, Barnsley L, Wallis BJ, McDonald GJ, Bogduk N. Percutaneous radio-frequency neurotomy for chronic cervical zygapophyseal-joint pain. N Engl J Med 1996;335: Smith-Petersen MN. Approach to and exposure of the hip for mold arthroplasty. J Bone Joint Surg 1968;50:334. 8Ward WT, Fleisch ID, Ganz R. Anatomy of the iliocapsularis muscle. Relevance to surgery of the hip. Clin Orthop Relat Res 2000;374: Lord SM, Barnsley L, Bogduk N. Percutaneous radiofrequency neurotomy in the treatment of cervical zygapophysial joint pain: A caution. Neurosurgery 1995;36: Lord SM, McDonald GJ, Bogduk N. Percutaneous radiofrequency neurotomy of the cervical medial branches: A validated treatment for cervical zygapophysial joint pain. Neurosurg Q 1998;8: Birnbaum K, Prescher A, Heßler S, Heller KD. The sensory innervation of the hip joint An anatomical study. Surg Radiol Anat 1997;19: Gardner E. The innervation of the hip joint. Anat Rec 1948;101: Barnsley L, Bogduk N. Medial branch blocks are specific for the diagnosis of cervical zygapophyseal joint pain. Reg Anesth 1993;18: Barnsley L, Lord S, Bogduk N. Comparative local anaesthetic blocks in the diagnosis of cervical zygapophysial joint pain. Pain 1993;55: Lord SM, Barnsley L, Bogduk N. The utility of comparative local anesthetic blocks versus placebo-controlled blocks for the diagnosis of cervical zygapophysial joint pain. Clin J Pain 1995;11: McDonald GJ, Lord SM, Bogduk N. Long-term follow-up of patients treated with cervical radiofrequency neurotomy for chronic neck pain. Neurosurgery 1999;45: Schwarzer AC, Aprill CN, Derby R, et al. The falsepositive rate of uncontrolled diagnostic blocks of the lumbar zygapophysial joints. Pain 1994;58: Barnsley L, Lord S, Wallis B, Bogduk N. Falsepositive rates of cervical zygapophysial joint blocks. Clin J Pain 1993;9: