Abstract
A prospective study was carried out to investigate any correlation between electrophysiological and sonographic findings in patients with a clinical diagnosis of carpal tunnel syndrome. A total of 113 patients (113 wrists) in 90 women and 23 men, with a mean age of 60 years, underwent sonographic and electrophysiological examination. Fifty-five patients (48%) had mild, 43 (38%) moderate and 12 (11%) had severe conduction disturbances and three patients had normal conduction. Sonographic measurements showed a cross-sectional area of the median nerve of 9.9 mm2 at the forearm and 17.8 mm2 at the tunnel inlet. The mean anteroposterior diameter (height) of the nerve at the tunnel inlet was 2.7 mm, and the lowest height inside the tunnel was 1.8 mm. No correlation was found between sonographic and electrophysiological parameters.
Introduction
The diagnosis of carpal tunnel syndrome (CTS) is based on the presence of typical clinical symptoms and signs and often confirmed by nerve conduction studies. Ultrasonography has recently attracted increased popularity in diagnosing CTS and gained a good reputation as a valuable tool for confirming the diagnosis (Klauser et al., 2009; Lange J, 2012; Seror, 2008).
Numerous studies have shown that the most accurate sonographic parameter reflecting the compression of the median nerve in the carpal tunnel is the cross-sectional area (CSA) of the nerve at the inlet level (Klauser et al., 2009; Pinilla et al., 2008; Seror, 2008). However, sonographic data show great variability in mean median nerve CSA, both in controls (4.8–9.7 mm2) and in patients with CTS (9–16.8 mm2). This means that there is extensive overlap of the distributions of the CSA values considered to be normal or pathological. Likewise, the CSA threshold (cut-off value) for establishing the diagnosis is not determined and varies between studies from 9 to 13 mm2. The sensitivity and specificity of ultrasonography in diagnosing CTS, calculated for different cut-off values, vary from 60% to 98% (El Miedany et al., 2004; Moran et al., 2009; Pinilla et al., 2008; Yazdchi et al., 2012).
Nerve conduction studies (NCS) are still believed to be a ‘gold standard’ in diagnosing CTS and are frequently used as a reference standard for ultrasonography (Moran et al., 2009; Pastare et al., 2009; Seror, 2008). In contrast with sonographic data, electrodiagnostic tests are characterized by less variation. However, both types of examination are additional tools to confirm or exclude the diagnosis of CTS in combination with the clinical signs or history.
The results of some studies show good correlation between electrophysiological and sonographic parameters in CTS: as conduction disturbances become more severe, there is increasing sonographic abnormality (Pastare et al., 2009; Visser et al., 2008; Wong et al., 2002). However, some have questioned this relationship (Koyuncuoglu et al., 2005; Moran et al., 2009; Pinilla et al. 2008). In view of this, we carried out a prospective study to investigate the possible correlation between sonographic and electrophysiological parameters of the median nerve in patients diagnosed with CTS.
Patients and methods
Between 2009 and 2011, 349 patients (279 women and 60 men) were admitted to the authors’ institution for carpal tunnel release. Of this number, 113 patients, 90 women (80%) and 23 men (20%), with a mean age of 60 years (range 37–88), with the condition lasting a mean of 63 months (range 4–260), were examined both sonographically and electrophysiologically. In 33 patients (29%), the disease was unilateral, involving the right wrist in 23 and the left in 11 patients. In 80 patients (71%), there was bilateral involvement. The diagnosis of CTS was made on clinical grounds from the clinical history, symptoms and signs.
Electrodiagnostic studies were carried out in the Laboratory for Neurophysiology with an Alpine Biomed (Copenhagen, Denmark) device. Sensory nerve conduction and sensory nerve action potential were examined after antidromic stimulation of the median nerve with a surface electrode on the distal forearm. Receiving ring electrodes were placed on the index and middle fingers, and the signals were recorded separately from each finger. Motor nerve conduction was examined after orthodromic stimulation at two sites, the cubital fossa and distal forearm. A receiving, bipolar surface electrode was placed over the abductor pollicis brevis muscle. Distal motor latency and compound muscle potentials were recorded. The results of needle electromyography were also analysed for signs of neurogenic changes or muscle denervation. According to the severity of conduction abnormalities, the results were graded as mild, moderate or severe (Table 1).
Classification of the severity of neurophysiologic findings in the study.
APB, abductor pollicis brevis muscle; CMAP, compound muscle action potential; DML, distal motor latency; EMG, electro- myography; NCS, nerve conduction studies; SNAP, sensory nerve action potential.
Each patient was examined one day before the operation, and the assessment included ultrasonography scanning of the median nerve and clinical examination. In patients with bilateral involvement, only the wrist indicated for surgery was examined. The sonographic examinations were done by two hand surgeons, both trained in musculoskeletal sonography, using an 18 MHz linear array transducer (Esaote, Maastricht, The Netherlands). The examiner was blinded to the results of clinical and electrophysiological examinations. The patients were seated in a comfortable position, facing the examiner, with their hands resting on a firm surface and fingers semi-flexed. Particular attention was paid to maintaining correct transducer position, perpendicular to the median nerve. The full course of the nerve in the carpal tunnel was assessed in the longitudinal and transverse planes. For the study purpose, median nerve images were obtained and recorded at three levels:
at the forearm, at the proximal margin of the pronator quadratus muscle;
at the carpal tunnel inlet, at the level of the pisiform bone and scaphoid tubercle; and
at the narrowest site of the nerve in the carpal tunnel.
At levels (a) and (b), the CSA of the median nerve was measured in the coronal plane (Figure 1). At levels (b) and (c), the height of the median nerve was measured in the sagittal plane (Figure 2). The CSAs at the tunnel inlet and at the distal forearm were calculated directly with the area measurement software of the ultrasonography machine.
(a) Coronal section of the median nerve (outlined) at the level of the pronator quadratus. The CSA of the nerve is 13 mm2. The proximal edge of the muscle is marked with an arrow. (b) Coronal section of the median nerve (outlined) at the carpal tunnel inlet. The CSA of the nerve is 20 mm2. The scaphoid tubercle is marked with an arrow.
(a) Sagittal section of the median nerve (marked with +) at the carpal tunnel inlet. The height of the nerve is 2.3 mm. (b) Sagittal section of the median nerve (marked with +1) at the narrowest site in the carpal tunnel. The height of the nerve is 1.8 mm. The sagittal section of the flexor retinaculum is marked with +2 (height 4.8 mm).
Clinical examinations included completion of the Levine questionnaire (Levine et al., 1993). Quantitative data were expressed as the mean, median, range and standard deviation. The correlation between variables was analysed with the Spearman rank test and a coefficient of p<0.05 was considered to indicate a significant difference.
Results
Fifty-five patients (48%) had mild, 43 (38%) moderate and 12 (11%) had severe disturbances of conduction in the median nerve. In three patients (3%), the electrodiagnostic studies were normal. The results of sonographic and clinical measurements are shown in Table 2. The CSA of the median nerve at the tunnel inlet differed significantly from the CSA at the forearm. There was, however, a wide variation of these variables, reflected in the SD values. The height of the median nerve at the tunnel inlet also differed significantly from the narrowest site in the carpal tunnel. The ranges of these variables also were wide, but the SDs were much lower in comparison with those of the CSA measurements.
Results of sonographic and clinical measurements in 113 patients with clinically diagnosed CTS.
CSA, cross-sectional area; NCS, nerve conduction studies.
Using the Spearman test, we calculated the correlation between the sonographic and electrophysiological parameters indicating the severity of CTS. We noted a statistically significant but weak (r=0.22) correlation only between the Levine symptom score and the severity of conduction disturbances (Table 3).
Correlation between the sonographic, clinical and neurophysiological variables and the severity of grade of CTS. Statistically significant correlations are in bold.
CSA, cross-sectional area; NCS, nerve conduction studies.
Discussion
The relatively wide range of variability of sonographic findings contributes to the variety of opinions about its utility in diagnosing CTS. The mean CSA values in our study (17.6 mm2) were generally greater than reported by others (mostly within the range 10–15 mm2). Only Klauser et al. (2009) have reported similar results, with a mean CSA at the tunnel inlet of 16.8 mm2 in patients and 9.0 mm2 in controls. The best diagnostic discrimination, with sensitivity of 94% and specificity of 95%, was noted by using a CSA threshold of 12 mm2.
The correlation between electrophysiological and sonographic findings in CTS has been investigated in several studies, and their results are equivocal. A statistically significant correlation between electrophysiological and sonographic findings was demonstrated by El Miedany et al. (2004). Patients with normal median NCS had a mean CSA of 11.6 mm2, and those with mild, moderate and severe conduction disturbances had values of 11.7 mm2, 16.7 mm2 and 20.7 mm2, respectively. These authors found the CSA cut-off value of 10 mm2 to be the upper limit for the normal median nerve and noted a sensitivity and specificity of ultrasonography >96% for diagnosing CTS. Wiesler et al. (2006) found the average CSA of the median nerve at the distal wrist crease to be 9 mm2 in asymptomatic volunteers, and 14 mm2 in CTS patients. They demonstrated also that higher values of ultrasound measurements were associated with greater abnormalities in NCS findings. A cut-off point of 11 mm2 or more for ultrasound measurement yielded a sensitivity of 91% and a specificity of 84%. The CSA values in patients had a large standard deviation, similar to the findings in our study. Ziswiler et al. (2005) reported that a CSA cut-off of 10 mm2 yielded a sensitivity of 82% and a specificity of 87% in detecting CTS. A cut-off value of <8 mm2 had satisfactory power to rule out CTS, but the fitted-negative likehood ratio was only 0.13. A cut-off of >12 mm2 had excellent power to confirm CTS, with a fitted-positive likehood ratio of 19.9. They noted a high concordance between sonography and nerve conduction studies. Pastare et al. (2009) reported a significant correlation between electrophysiological and sonographic findings, assuming a CSA cut-off value of 9 mm2 as diagnostic for CTS. Similar findings were reported by Karadag et al. (2010). Patients with normal median nerve conduction studies had a CSA of 8.5 mm2, and those with mild, moderate and severe conduction disturbances had values of 11.7 mm2, 14 mm2 and 16.3 mm2, respectively. These values, however, differ considerably from those reported earlier by El Miedany et al. (2004). Yazdchi et al. (2012) reported a statistically significant correlation between the electrophysiological severity of the syndrome and the CSA at the carpal tunnel inlet, but failed to show significant differences in CSA between mild and moderate CTS and controls.
In contrast, Kele et al. (2003) reported that a CSA of 11 mm2 was considered as diagnostic for CTS, but that sonography did not show any correlation with electrophysiological or clinical findings. Koyuncuoglu et al. (2005) reported the results of patients who had clinical symptoms of CTS, but whose nerve conduction studies were normal. An increased CSA (>10.5 mm2) of the nerve was found in 30% of these wrists, but no correlation between sensory conduction velocity and the CSA was observed. Pinilla et al. (2008) examined the median nerve CSA and antero-posterior (AP) diameter (height) in patients with CTS and controls. The patients also underwent electrophysiological evaluation. Statistically significant differences in measured variables between the patients and controls were noted, but the ultrasound failed to correlate with NCS findings. Moran et al. (2009) analysed the results of sonographic and electrodiagnostic studies in patients with CTS and found a CSA greater than 13 mm2 to be diagnostic for CTS. They concluded that sonographic findings could not grade the severity of the syndrome as well as nerve conduction studies. Kwon et al. (2008) compared the usefulness of sonography and nerve conduction studies in the diagnosis of CTS. They reported the best cut-off value of CSA at the tunnel inlet to be 10.7 mm2, which had a sensitivity of 66% and a specificity of 63%. NCS showed a sensitivity of 78% and a specificity of 83%. They concluded that sonography was not accurate enough to replace NCS for the diagnosis of CTS. Our findings are largely consistent with the results reported in these four articles.
It is difficult to explain the discrepancies between these studies. They may result from the use of different cut-off values of the CSA as being diagnostic for CTS. The range of this parameter was wide (9–13 mm2) in the various studies, and this might affect the results. Sonographic assessment is highly operator and equipment dependent, which may introduce an element of bias in the assessments. There are also inherent differences in median nerve size according to handedness, sex, age, race or professional occupation (Pinilla et al., 2008). All these factors result in difficulties in the determination of uniform, normative sonographic parameters for the median nerve. Electrophysiological studies are characterized by less variation, and there are commonly accepted criteria for a diagnosis of CTS based on conduction disturbances (Jablecki et al., 1993; Makanji et al., 2012). However there is some liberality in grading the severity of the electrophysiological findings, which may result in different results for the correlation with sonographic (and clinical) data. The lack of correlation between sonographic and electrophysiological findings in our study might be also explained by the fact that most of our patients (89%) had mild and moderate conduction disturbances, and only 11% had severe disturbances. This tendency has been reported by other authors: sonography is better at discriminating between patients with severe CTS and those with mild/moderate CTS than between those with mild/moderate CTS and controls (Kele et al., 2003; Koyuncuoglu et al., 2005; Pinilla et al., 2008).
It has been suggested that the wrist-to-forearm ratio (WFR) may be a more accurate index than the CSA at the tunnel inlet for diagnosing CTS. Some authors have reported statistically significant correlations between the WFR and electrophysiological findings, and these were significantly related to the severity of the condition. Moreover, the cut-off value for each severity showed higher sensitivity and specificity (Kang et al., 2012). We believe however, that WFR and ΔCSA indices present similar discriminative strengths, because they are both derived from the same basic variables (the CSA at the wrist and at the tunnel inlet). In our study we failed to find the relationship between ΔCSA and electrophysiological parameters.
It is obvious that sonography and electrophysiological tests measure different aspects of median nerve pathology. The former shows only enlargement (swelling) of the median nerve at the tunnel inlet, which is likely to be caused by nerve compression inside the tunnel. It provides no information about the nerve function, mechanism and severity of the lesion or any other possible site of entrapment, as does electrodiagnostic examination. Nerve conduction studies provide much more information than ultrasonography. However, how much of this information is necessary for the surgeon in making a decision about treatment? The most important information that is needed is the exact cause of the patients’ complaints within the hand. If enlargement of the median nerve at the tunnel inlet precisely indicated the cause and site of the problem (entrapment in the carpal tunnel), nothing more would be needed. The problem is that both nerve conduction studies and ultrasonography show no superiority to the clinical findings in the diagnosis of typical cases of CTS (Zyluk and Szlosser, 2013). Therefore, we believe that routine use of sonography (and likewise nerve conduction studies) in clinically typical cases is not necessary. In patients with atypical CTS symptoms or an unclear history, both these additional examinations may help to confirm or exclude the diagnosis of CTS. However, in these doubtful cases, electrodiagnostic tests provide significantly more information than ultrasonography about the condition and function of the nerve. Ultrasonography may be helpful in particular situations, such as the presence of an intracarpal mass (ganglion cyst, muscle, tumour) or anatomical abnormalities (median artery, bifid nerve), but not in routine practice.
Our study has some limitations. The sonography was carried out by surgeons, rather than radiologists, although both were trained in musculoskeletal sonography. Another drawback may be the method of expression of electrophysiological findings by grades, but not by individual variables such as velocity, latency and amplitude. One might expect these variables to correlate better with the sonographic findings, as has been shown it some studies (Pastare et al., 2009). However, we believe that expressing the severity of conduction disturbances in grades is more practical, as it reflects all the variables considered and not just individual ones.
Footnotes
Conflict of interests
None declared.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.