Bifid Median Nerve in Patients With Carpal Tunnel Syndrome

Transcription

1 Article Bifid Median Nerve in Patients With Carpal Tunnel Syndrome Ilkay Koray Bayrak, MD, Ayse Oytun Bayrak, MD, Melike Kale, MD, Hande Turker, MD, Barıs Diren, MD Objective. The aim of this study was to investigate the frequency of the anatomic variation of a bifid median nerve in patients with carpal tunnel syndrome (CTS) and to determine the size criteria for sonography. Methods. On axial sonograms of 320 hands of 170 patients with CTS and 240 hands of 120 unaffected individuals, the median nerve was evaluated morphologically for a bifid median nerve, and the cross-sectional area was measured at 3 levels (radial-ulnar junction, pisiform, and hook of the hamate). Electrophysiologic studies were performed in addition to clinical and sonographic evaluations in all patients, controls with a bifid median nerve, and controls with a cross-sectional area of greater than 0.09 cm 2. Results. A bifid median nerve was seen in 32 (19%) of 170 patients and 11 (9%) of 120 controls. It occurred relatively frequently in patients with CTS (P <.01). The cross-sectional area of the bifid median nerve was relatively higher than that of the nonbifid median nerve in controls at 2 of the 3 levels (P <.001; P =.226; P <.01). The cutoff values for the cross-sectional area at the level of the pisiform were 0.11 cm 2 (sensitivity, 90%; specificity, 99%; P <.001) for patients with a bifid median nerve and 0.10 cm 2 (sensitivity, 98%; specificity, 81%; P <.001) for patients with a nonbifid median nerve. Conclusions. A bifid median nerve occurs relatively frequently in patients with CTS. It may facilitate compression of the median nerve in the carpal tunnel because of its relatively higher crosssectional area compared with a nonbifid median nerve. The size criterion for CTS in patients with a bifid median nerve is slightly higher than in those with a nonbifid median nerve. Key words: bifid median nerve; carpal tunnel syndrome; electrophysiology; high division of median nerve; sonography. Abbreviations CTS, carpal tunnel syndrome; ROC, receiver operating characteristic Received December 17, 2007, from the Departments of Neurology (I.K.B., M.K., B.D.) and Radiology (A.O.B., H.T.), Ondokuz Mayıs University, School of Medicine, Samsun, Turkey. Revision requested January 4, Revised manuscript accepted for publication May 7, Address correspondence to Ilkay Koray Bayrak, MD, Department of Radiology, Ondokuz Mayıs University Medical Faculty, Kurupelit, Samsun, Turkey. ilkaykoray@hotmail.com Carpal tunnel syndrome (CTS) is a common neuropathy that is caused by entrapment of the median nerve in the wrist. 1 A high division of the median nerve proximal to the carpal tunnel, also known as a bifid median nerve, is an anatomic variation that may be associated with CTS. 2 The occurrence of a bifid median nerve has been widely reported in the surgical literature, and in most cases it coexists with a persistent median artery. 2 8 The diagnosis of CTS is based on the patient s clinical history, physical examination, and electrophysiologic studies. Recently, sonography at high frequencies has been proven a useful diagnostic tool in patients with CTS Sonography can also depict the nerve along its course and can effectively show median nerve variations in cases of atypical clinical and electrophysiologic abnormalities A bifid median nerve that would be associated with CTS can also be depicted by sonography. To our knowledge, no previous study has determined the fre by the American Institute of Ultrasound in Medicine J Ultrasound Med 2008; 27: /08/$3.50

2 Bifid Median Nerve in Carpal Tunnel Syndrome quency of a bifid median nerve on sonography in patients with CTS and compared it with that in controls. Cross-sectional area cutoff values that can be used to diagnose CTS in bifid median nerves have not been reported before. The aim of this study was to investigate the frequency of bifid median nerves in patients with CTS and to determine the size criteria for CTS in bifid median nerve cases. Materials and Methods Study Participants Three hundred twenty hands of 170 patients with a diagnosis of CTS both clinically and electrophysiologically who were referred to our neurophysiology laboratory between December 2006 and May 2007 were evaluated sonographically. For the clinical diagnosis of CTS, patients were questioned about their primary symptoms (paresthesias, a needling sensation, clumsiness, and nocturnal symptoms) in the area of median nerve distribution. On the basis of 2 or more of these primary symptoms, CTS was diagnosed clinically. If only 1 of these symptoms was present, we looked for at least 2 of the secondary symptoms (swelling, sensations of burning and cold, tightness, and pain/discomfort). 16 Patients were excluded if they had any symptoms or signs of polyneuropathy on neurologic examinations or a history of surgery for CTS. Two hundred forty hands of 120 controls were selected from patients who underwent sonographic evaluations for other reasons. They were excluded if they described 1 of the CTS symptoms mentioned above or had systemic disease that may have been associated with CTS. Patients with diseases and conditions such as diabetes mellitus, hypothyroidism, rheumatoid arthritis, amyloidosis, chronic renal failure treated by hemodialysis, and pregnancy were excluded. Electrophysiologic evaluations were performed in all patients and controls with a bifid median nerve. Electrophysiologic evaluations were performed to exclude CTS when a cross-sectional area of greater than 0.09 cm 2 at the level of pisiform was measured by sonography in controls without a bifid median nerve. All patients and controls were asked to sign an informed consent form. Electrophysiologic Evaluations Routine nerve conduction studies for CTS were conducted with a Neuropack-8 electromyography system (Nihon Kohden, Tokyo, Japan). The skin temperature was kept between 31 C and 32 C. Motor nerve conduction studies of both median and ulnar nerves were performed using supramaximal stimuli at the wrist and recording from the abductor pollicis brevis and the abductor digiti minimi muscles, respectively. Sensory conduction studies were done antidromically, stimulating at the wrist and recording from digits 1 and 3 for the median nerve and from digit 5 for the ulnar nerve. We compared the findings with the reference values used in our laboratory and looked for median nerve neuropathy related to a median nerve action potential latency abnormality at the palm-wrist segment. When routine nerve conduction study results were normal (standard negative hand results), a medianulnar sensory latency difference ( 0.4) was required. 17 In accordance with the results provided in electrophysiologic studies, all of the hands were categorized into 5 stages 18 : (1) minimal CTS, characterized by standard negative hand results and abnormal comparative test results; (2) mild CTS, characterized by slowing of the median digit-wrist segment and normal distal motor latency; (3) moderate CTS, characterized by slowing of the median digit-wrist segment and abnormal distal motor latency; (4) severe CTS, characterized by the absence of a median sensory response and abnormal distal motor latency; and (5) extreme CTS, characterized by the absence of a thenar motor (and sensory) response. Sonographic Evaluations Sonography was performed in real time with an Aplio SSA-770A system (Toshiba Medical Systems Co, Ltd, Tokyo, Japan) and a 12-MHz linear probe. Each hand involved in the study was placed on a hard surface in a neutral position. The full course of the median nerve in the tunnel was evaluated first in the sagittal plane. The median nerve was located just beneath the flexor retinaculum in the axial section. The nerve was accepted as bifid if and when it branched out before the level of the distal radial-ulnar junction. Three internal anatomic landmarks were 1130 J Ultrasound Med 2008; 27:

3 Bayrak et al used for the axial images. Axial images of the nerve were obtained at the radial-ulnar junction (just proximal to the flexor retinaculum) and at the level of the pisiform (proximal tunnel) and the hook of the hamate (mid tunnel); the crosssectional area at each level was measured. A manual tracing method was used for cross-sectional area measurements (Figure 1A). The area measurements were done for lateral and medial branches of a bifid median nerve separately, and 2 measurements were added to obtain a total crosssectional area value at each level. Bifid median nerves were also grouped according to the thickness of the medial and lateral branches (equal branches, thicker lateral branch, or thicker medial branch; Figure 1). Palmar displacement of the flexor retinaculum was not assessed, but the median nerve was traced until it reached the deep antecubital fossa for possible nerve or surrounding structure abnormalities. If needed, Doppler sonography was used to differentiate vessels between the 2 branches. All sonographic evaluations were performed by the same radiologist. Statistical Analysis A statistical analysis was performed with SPSS version 11 software for Windows (SPSS Inc, Chicago, IL). To determine optimal cutoff points for cross-sectional area measurements at the 3 different levels of the median nerve in patients with CTS, a receiver operating characteristic (ROC) curve analysis was done. The student t test was used to compare cross-sectional area measurements at the 3 levels of the bifid and nonbifid median nerves of patients and controls. Before the t test was performed, data were tested for normality by a Wilks-Shapiro test. The χ 2 test was used to determine the frequency of a bifid median nerve in patients with CTS compared to those without CTS. The Fisher exact test was used to compare patients and controls with a bifid median nerve according to findings of branch dominance (thicker lateral, thicker medial, or equal) and the presence or absence of a persistent median artery. Results Study Participants Among 170 patients with CTS, 32 (19%) had a bifid median nerve. There were no significant differences in age (P >.05) or sex (P >.05) between the patients with CTS with and without a bifid median nerve. The bifid median nerve was unilateral in 22 patients (13%, 14 [64%] in nondominant and 8 [36%] in dominant hands) and bilateral in 10 (6%). We determined a high division of the median nerve in 12 hands of 11 (9%) of 120 patients. A bifid median nerve was relatively frequent in patients with CTS (P <.01; χ 2 = 9.48). Patient and control characteristics are summarized in Table 1. Electrophysiologic Findings Electrophysiologic studies were done in all patients with CTS and patients and controls with a bifid median nerve. Carpal tunnel syndrome was excluded by electrophysiologic studies in 18 controls with a cross-sectional area of greater than 0.09 cm 2 at the level of the pisiform and a Figure 1. A, Bifid median nerve with almost equal branches (branches A and B). B, Bifid median nerve with a thicker lateral (L) than medial (M) branch. Persistent median arteries (arrowheads) are shown along both bifid median nerves. A B J Ultrasound Med 2008; 27:

4 Bifid Median Nerve in Carpal Tunnel Syndrome Table 1. Characteristics of Patients and Controls Patients Without Patients With Controls Without Controls With Characteristic BMN (n = 138) BMN (n = 32 [19%]) BMN (n = 109) BMN (n = 11 [9%]) Mean age, y Male Female Total Sex, n (%) Male 16 (12) 4 (12.5) 20 (18) 2 (18) Female 122 (88) 28 (87.5) 89 (82) 9 (82) CTS, n/bmn, n (%) Bilateral (31) 0 1 (9) Unilateral (69) 0 10 (81) BMN indicates bifid median nerve. nonbifid median nerve. Similar to the remaining 102 participants, they did not have any of the primary or secondary symptoms of CTS. In 2 participants with no clinical symptom of CTS but a large cross-sectional area on sonography, CTS was confirmed by electrophysiologic studies, and these participants were excluded. Electrophysiologic findings of the patients with a bifid median nerve are summarized in Table 2. A unilateral bifid median nerve was in the nondominant hand of 14 patients. Twelve of them had CTS in both hands; the electrophysiologic stages of the 2 sides were comparable (difference not more than 1 stage) in 10 patients, whereas the electrophysiologic stage was more advanced in the nondominant hand in 2 patients. We did not observe any atypical electrophysiologic evidence in the patients with a bifid median nerve. Table 2. Carpal Tunnel Syndrome Stage Distribution in Patients With a Bifid Median Nerve CTS Stage Dominant Nondominant Unilateral BMN (n = 22) Minimal 1 1 Mild 4 7 Moderate 2 2 Severe 1 4 Extreme 0 0 Bilateral BMN (n = 10) Minimal 4 0 Mild 1 3 Moderate 4 4 Severe 1 3 Extreme 0 0 BMN indicates bifid median nerve. Sonographic Findings A bifid median nerve was observed in 42 wrists of 32 patients with CTS. In the control group, a bifid median nerve was seen in 12 wrists of 11 individuals. No significant difference was noted between patients and controls with a bifid median nerve according to the thickness of the branches (P =.078) and the presence of a persistent median artery (P =.32; Table 3). The 2 (lateral and medial) branches were measured individually, and a total cross-sectional area value was obtained for each level. Mean total cross-sectional area measurements of bifid and nonbifid median nerves in both patients and controls are summarized in Table 4. Between patients and controls with or without a bifid median nerve, a significant difference was noted in cross-sectional area measurements at all measured levels of the median nerve (P <.001; Table 4). Between bifid and nonbifid median nerves in patients, no significant difference was noted in cross-sectional areas at any of the measured levels (radial-ulnar junction, P =.961; level of pisiform, P =.584; level of hamate, P =.067; Table 4). The cross-sectional area of the bifid median nerve was significantly higher than that of the nonbifid median nerve in controls at the levels of the radial-ulnar junction (P <.001) and hamate (P <.01). At the level of pisiform, however, the difference was not statistically significant (P =.226; Table 4). On the ROC curve analysis, cutoff values for crosssectional area measurements for CTS in patients with a bifid median nerve were 0.11 cm 2 at the level of the radial-ulnar junction (sensitivity, 88%; specificity, 99%; P <.001), 0.11 cm 2 at the level of the pisi J Ultrasound Med 2008; 27:

5 Bayrak et al Table 3. Sonographic Characteristics of a Bifid Median Nerve Characteristic BMN With CTS (n = 42) BMN Without CTS (n = 12) P Thicker lateral branch ( 2-fold), n (%) 22 (52) 8 (67) Thicker medial branch ( 2-fold), n (%) 4 (10) 3 (25).078 Almost equal branches, n (%) 16 (38) 1 (8) Persistent median artery 19 (45) 3 (25).32 BMN indicates bifid median nerve. form (sensitivity, 90%; specificity, 99%; P <.001), and 0.12 cm 2 at the level of the hook of the hamate (sensitivity, 95%; specificity, 99%; P <.001). These cutoff values for the cross-sectional area of the nonbifid median nerve were 0.09 cm 2 (sensitivity, 96%; specificity, 81%; P <.001), 0.1 cm 2 (sensitivity, 98%; specificity, 81%; P <.001), and 0.1 cm 2 (sensitivity, 96%; specificity, 87%; P <.001), respectively. The ROC curves of crosssectional area measurements for detection of CTS in patients with and without a bifid median nerve are shown in Figure 2. No space-occupying lesion (dislocated bone, mass, or accessory muscle) was noted around the median nerves. It was not easy to differentiate isolated compression of the branches of the nerve because the 2 branches were so close to each other (not more than 2 mm in all participants). Although there are cases of a bifid median nerve with branches having a marked difference in cross-sectional areas, we observed no apparent difference in echogenicity. Discussion Carpal tunnel syndrome, compression of the median nerve by the flexor retinaculum at the wrist, is the most common nerve entrapment syndrome. It may be idiopathic or associated with various conditions such as space-occupying lesions, Colles fractures, and systemic diseases such as diabetes mellitus, hypothyroidism, and rheumatoid arthritis. 1 Recently, Tanzer 19 suggested that an anatomic variation may be a predisposing factor in the etiology of CTS. A bifid median nerve and persistent median artery are anatomic variations that have been reported extensively in the surgical literature. 2 8 The bifid median nerve anomaly has been reported to have an incidence of 0.8% to 2.8% in patients with CTS, and in most cases it has been reported with a concomitant persistent median artery. 2 4,6 Table 4. Cross-sectional Area Measurements of All Nerves in Patients and Controls With and Without a Bifid Median Nerve CTS Controls Measurement Mean ± SD Range Mean ± SD Range P t BMN, CTS (n = 42) vs controls (n = 12) Radial-ulnar junction, cm ± ± < Level of pisiform, cm ± ± < Level of hamate, cm ± ± < Non-BMN, CTS (n = 278) vs controls (n = 228) Radial-ulnar junction, cm ± ± < Level of pisiform, cm ± ± < Level of hamate, cm ± ± < P t P t BMN vs non-bmn Radial-ulnar junction < Level of pisiform Level of hamate < BMN indicates bifid median nerve. J Ultrasound Med 2008; 27:

6 Bifid Median Nerve in Carpal Tunnel Syndrome A B Figure 2. Receiver operating characteristic curves of cross-sectional area measurements at the levels of the radial-ulnar articulation (RUA), pisiform (LOP), and hook of the hamate (HOH) for detection of CTS in patients with (A) and without (B) a bifid median nerve. Imaging of bifid median nerves has been described in very few studies, whereas to the best of our knowledge, electrophysiologic and sonographic findings had not been fully described in the literature. Iannicelli et al 13 compared sonographic and magnetic resonance imaging findings in 6 patients with a bifid median nerve selected from a population of 294 patients with CTS. In another study, the authors reported 3 cases of a bifid median nerve, 1 of which was in a patient with CTS, whereas the remaining 2 were found in cadaveric specimens. 12 Both studies concluded that sonography can allow effective diagnosis and delineation of a bifid median nerve. In a recent study, Gassner et al 14 described Doppler sonographic findings in 2 patients with CTS associated with a persistent median artery, and they reported 16 hands with a persistent median artery among 50 asymptomatic volunteers. In 10 of 16 hands, the persistent median artery was associated with a bifid median nerve. To our knowledge, this was the largest study in the literature that provided a discussion for the imaging and electrophysiologic findings concerning bifid median nerves. In our study of 170 patients with CTS, we found a bifid median nerve in 32 cases, unilaterally in 22 and bilaterally in 10. It is not clear why our data revealed a relatively high percentage of bifid median nerves (19%). Nineteen hands were associated with a persistent median artery. Thus, the incidence of bifid median nerves in our study was higher than the previous rates, and in nearly half (45%) of our cases, a concomitant persistent median artery was found. Radiologic criteria for diagnosing CTS on magnetic resonance imaging or sonography include swelling of the median nerve proximal to the carpal tunnel and bowing of the flexor retinaculum. 11,12,20 There are many studies acknowledging that sonography is an effective choice in patients with CTS, relying on enlargement of the median nerve with an area of greater than 9 or 10 mm 2 at the level of the pisiform. 21,22 Our data showed a cutoff value of 10 mm 2 for CTS in nonbifid median nerves at the level of the pisiform. Interestingly, Propeck et al 12 indicated that the sonographic size criteria for diagnosing CTS in nonbifid median nerves may not be accurate in evaluating bifid median nerves. To our knowledge, no study attempted to establish the size criteria for CTS in bifid median nerves. In patients with a bifid median nerve, our data revealed a cutoff value of 11 mm 2 for CTS. This study determined a relatively higher cutoff value for CTS in bifid median nerves than in nonbifid ones. Singer and Ashworth 23 reviewed the surgical findings of 147 hands that underwent carpal tunnel release and reported 47 variations in 60 hands. They concluded that patient age and hand dominance were important factors in predicting the presence of an anatomic variation J Ultrasound Med 2008; 27:

7 Bayrak et al For patients younger than 40 years, the odds of observing an anatomic variation were higher than for patients older than 40 years. The odds of observing an anatomic variation in the dominant hand were also higher than in the nondominant hand. In our study, we investigated only bifid median nerves and noted no significant age difference. The occurrence of a bifid median nerve in the nondominant hand was relatively higher in patients with a unilateral bifid median nerve in our study group. It is not clear why our data showed a relatively high frequency of bifid median nerves in the nondominant hand. The major weakness of our study was the lack of interobserver and intraobserver variability measurements. In addition, although the controls did not have any primary or secondary symptoms, electrophysiologic studies were not done in all of them. In contrast to sonography, no atypical clinical or electrophysiologic findings were noted in the hands of participants with a bifid median nerve. We also did not find a significant difference in electrophysiologic stages between the 2 groups. Only 2 patients with a bifid median nerve in the nondominant hand had unilateral and severe CTS, and 3 patients with a bifid median nerve in the nondominant hand had more severe CTS than in the dominant hand. It is known that in idiopathic cases, the dominant hand is almost always the affected hand. If symptoms are bilateral, then the dominant hand is usually more severely affected than the contralateral hand. Unilateral CTS or bilateral CTS that is considerably worse in the nondominant hand should raise suspicion of the presence of a specific underlying cause. 1 We think that it would be important to investigate the bifid median nerve and other variations or abnormalities when unilateral CTS in the nondominant hand is determined. Our data also showed a higher cross-sectional area in controls with a bifid median nerve than in those with a nonbifid median nerve at the levels of the radial-ulnar junction and hook of the hamate, but the difference at the level of the pisiform was not significant. We think that these findings and the higher incidence of bifid median nerves in patients with CTS support the idea of the bifid median nerve as the cause of CTS. In conclusion, a bifid median nerve occurs relatively frequently in patients with CTS. It may facilitate compression of the median nerve in the carpal tunnel because of its relatively higher cross-sectional area than that of a nonbifid median nerve. Although it has no electrophysiologic or clinical differential diagnosis, in cases with unilateral or severe CTS especially in the nondominant hand, physicians may consider the presence of a median nerve variation. Because sonography is a noninvasive and effective method for assessing median nerve abnormalities such as a bifid median nerve, we recommend its use in evaluations of the median nerve. In addition, the sonographic size criterion for CTS in patients with a bifid median nerve is higher than in those with a nonbifid median nerve. References 1. Preston DC, Shapiro BE. Median neuropathy at the wrist. In: Electromyography and Neuromuscular Disorders. 2nd ed. Philadelphia, PA: Elsevier Butterworth-Heinemann; 2005: Lanz U. Anatomical variations of the median nerve in the carpal tunnel. J Hand Surg [Am] 1977; 2: Eiken O, Carstam N, Eddeland A. Anomalous distal branching of the median nerve: case reports. Scand J Plast Reconstr Surg 1971; 5: Tountas CP, Bihrle DM, MacDonald CJ, Bergman RA. Variations of the median nerve in the carpal canal. J Hand Surg [Am] 1987; 12: Stancić MF, Eskinja N, Stosić A. Anatomical variations of the median nerve in the carpal tunnel. Int Orthop 1995; 19: Ahn DS, Yoon ES, Koo SH, Park SH. A prospective study of the anatomic variations of the median nerve in the carpal tunnel in Asians. Ann Plast Surg 2000; 44: Lindley SG, Kleinert JM. Prevalence of anatomic variations encountered in elective carpal tunnel release. J Hand Surg [Am] 2003; 28: Barbe M, Bradfield J, Donathan M, Elmaleh J. Coexistence of multiple anomalies in the carpal tunnel. Clin Anat 2005; 18: Buchberger W. Radiologic imaging of the carpal tunnel. Eur J Radiol 1997; 25: Buchberger W, Schön G, Strasser K, Jungwirth W. Highresolution ultrasonography of the carpal tunnel. J Ultrasound Med 1991; 10: Bayrak IK, Bayrak AO, Tilki HE, Nural MS, Sunter T. Ultrasonography in carpal tunnel syndrome: comparison with electrophysiological stage and motor unit number estimate. Muscle Nerve 2007; 35: J Ultrasound Med 2008; 27:

8 Bifid Median Nerve in Carpal Tunnel Syndrome 12. Propeck T, Quinn TJ, Jacobson JA, Paulino AF, Habra G, Darian VB. Sonography and MR imaging of bifid median nerve with anatomic and histologic correlation. AJR Am J Roentgenol 2000; 175: Iannicelli E, Chianta GA, Salvini V, Almberger M, Monacelli G, Passariello R. Evaluation of bifid median nerve with sonography and MR imaging. J Ultrasound Med 2000; 19: Gassner EM, Schocke M, Peer S, Schwabegger A, Jaschke W, Bodner G. Persistent median artery in the carpal tunnel: color Doppler ultrasonographic findings. J Ultrasound Med 2002; 21: Padua L, Pazzaglia C, Insola A, et al. Schwannoma of the median nerve (even outside the wrist) may mimic carpal tunnel syndrome. Neurol Sci 2006; 26: Nathan PA, Meadows KD, Doyle LS. Sensory segmental latency values of the median nerve for a population of normal individuals. Arch Phys Med Rehabil 1988; 69: Jablecki CK, Andary MT, Floeter MK, et al; American Association of Electrodiagnostic Medicine, American Academy of Neurology, American Academy of Physical Medicine and Rehabilitation. Practice parameter: electrodiagnostic studies in carpal tunnel syndrome. Report of the American Association of Electrodiagnostic Medicine, American Academy of Neurology, and the American Academy of Physical Medicine and Rehabilitation. Neurology 2002; 58: Padua L, LoMonaco M, Gregori B, Valente EM, Padua R, Tonali P. Neurophysiological classification and sensitivity in 500 carpal tunnel syndrome hands. Acta Neurol Scand 1997; 96: Tanzer RC. The carpal-tunnel syndrome: a clinical and anatomical study. J Bone Joint Surg Am 1959; 41: Monagle K, Dai G, Chu A, Burnham RS, Snyder RE. Quantitative MR imaging of carpal tunnel syndrome. AJR Am J Roentgenol 1999; 172: Duncan I, Sullivan P, Lomas F. Sonography in the diagnosis of carpal tunnel syndrome. AJR Am J Roentgenol 1999; 173: Chen P, Maklad N, Redwine M, Zelitt D. Dynamic highresolution sonography of the carpal tunnel. AJR Am J Roentgenol 1997; 168: Singer G, Ashworth CR. Anatomic variations and carpal tunnel syndrome: 10-year clinical experience. Clin Orthop Relat Res 2001; 392: J Ultrasound Med 2008; 27: