| Sign In to gain access to subscriptions and/or personal tools. |
DOI: 10.1054/JHSB.2002.0869 © 2003 The British Society for Surgery of the Hand
Carpal Tunnel Syndrome – Is High-Frequency Ultrasound a Useful Diagnostic Tool?From the Hartlepool District General Hospital, Holdforth Road, Hartlepool, UK and the Department of Orthopaedic Surgery, Royal United Hospital, Bath, UK Correspondence: Mrs L Leonard, Orthopaedic Registrar, Foxhill, The Old Rope Walk, TETBURY, Glos GL8 8XQ, UK. Tel.: +44 1225 428 331 bleep 7049, E-mail: lisa.leonard{at}virgin.net
This study assessed the clinical use of three ultrasound measurements; median nerve cross-sectional area, median nerve flattening ratio and palmar displacement of the flexor retinaculum, for the diagnosis of carpal tunnel syndrome. The measurements were made in 20 carpal tunnel sufferers and 20 controls. The sensitivity, specificity and predictive values of each were calculated in various clinical settings. Values for each of the three variables were significantly different in the patient and control populations. The differences we recorded were smaller than those found in previous studies. The tests had a sensitivity of 72% and a specificity of 90%. Alterations in the morphology of the carpal tunnel in patients with carpal tunnel syndrome can be measured in the district general hospital setting. The measurements described maybe a useful non-invasive confirmatory test in patients in whom there is a strong clinical suspicion of carpal tunnel syndrome. However, they would be of no benefit in epidemiological surveys of populations with a low incidence of carpal tunnel syndrome.
The use of high-frequency ultrasound to image the carpal tunnel contents in patients with suspected carpal tunnel syndrome has been described previously (Buchberger et al., 1991, 1992, 1997; Chen et al., 1997). Three measurements, cross-sectional area of the median nerve at the level of the pisiform, the flattening ratio of the median nerve at the level of the hook of the hamate and the palmar displacement of the flexor retinaculum from a line drawn between the trapezium and the hamate, are influenced by the presence of carpal tunnel syndrome (Buchberger et al., 1992). These measurements can also be made using magnetic resonance imaging (Allmann et al., 1997; Mesgarzadeh et al., 1989; Middleton et al., 1987). One study has shown that successful carpal tunnel decompression results in resolution of the abnormal magnetic resonance imaging findings in a high percentage of cases (Allmann et al., 1997). This prospective study was designed to assess the reproducibility of these measurements using ultrasound. We also evaluated the use of these measurements for clinical settings by calculating their sensitivity and specificity and determining their positive and negative predictive values for populations with different prevalences of carpal tunnel syndrome. We also determined whether ultrasound was better or worse than other diagnostic tools.
The study was carried out in a district general hospital between March and July 1998 with full ethical committee approval. Patients attending outpatients with a clinical diagnosis of carpal tunnel syndrome were asked to participate. We recruited and included 20 patients and 20 controls. The patients had to have a characteristic history of paraesthesiae in the median nerve distribution with a positive Phalen’s test and no other co-existent neurological disease. No other diagnostic tests were relied upon, but patients had to have experienced relief of their symptoms after surgery for inclusion. This group were taken to represent true positives for the sensitivity and specificity analyses. Patients were excluded if they did not wish to participate. They were also excluded if an alternative pathology was diagnosed during the ultrasound examination or if they did not have relief of their symptoms following surgery. Consenting patients underwent examination with high-frequency ultrasound (Diasonics System Five, 10 MHz probe with a linear array) prior to their surgery and measurements were made of the three variables described by Buchberger et al. (1992). Axial scans were obtained at each level. To calculate the cross-sectional area of the the median nerve, the ulnoradial and anteroposterior diameters of the median nerve were measured at the level of the pisiform. The ratio of the same diameters, but at the level of the hamate, was used to calculate the flattening ratio of the median nerve. After drawing a line between the trapezium and hamate the maximum distance from this line to the palmar apex of the transverse carpal ligament was measured (palmar displacement). This took approximately 10 min in total for each patient. Each patient had an open carpal tunnel decompression under either local or general anaesthesia a short time after the scan (range, 1–85 days; median, 14 days). A control group of patients with no symptoms of carpal tunnel syndrome and no confounding variables were recruited from clinics and hospital staff. They were matched for sex and side studied. They were matched as closely as possible for age (15 to within 1 year, 3 to within 10 years and two with 13 years difference). These volunteers also underwent high-frequency ultrasound examination and similar data sets were recorded. Several ultrasonographers were involved in the study but one of the authors (GD) was present in each case to provide continuity. Those making the measurements were blinded as far as possible as to each individual’s group and hospital staff recruited to the control group did not attend wearing their uniform or identity badge. On occasions, some participants in the patient group were so keen to know the results of the test that it became obvious to the ultrasonographer that they were carpal tunnel sufferers. We therefore recognize that the blinding was not successful in some cases.
There were 18 women and two men in each group. The mean age was 46 (range, 34–66) years for the patients and 47 (range, 34–66) years for the controls. Only four of the patients elected to have general anaesthesia for decompression, the other 16 had surgery under local anaesthesia. Eighteen of the patients were right-hand dominant, one was left handed and dominance was not recorded for the other. Ten of the patients had a right-sided decompression, and ten had a left-sided decompression. All of the patients included in the final analysis had relief of their symptoms following surgery, which was taken as confirmation of the carpal tunnel diagnosis. One patient was found to have a ganglion within the carpal tunnel: this was decompressed percutaneously in the ultrasound department by one of us (GD) and resulted in complete relief of symptoms. This patient was not included in the analysis. The mean median nerve cross-sectional area was 11.6 (95% CI, 9.4–13.8) mm2 for the patients and 7.8 (95% CI, 7.1–8.5) mm2 for the control group (P=0.004). For the flattening ratio, the mean value was 3.5 (95% CI, 3.1–3.9) for the patients and 2.8 (95% CI, 2.5–3.1) for the control group (P=0.013). For the palmar displacement, the mean value was 2.8 (95% CI, 2.0–3.6) mm for the patients and 1.8 (95% CI, 1.2–2.4) mm for the control group (P=0.52): two patients did not have palmar displacement recorded. Multiple logistic regression was used to combine the three measurements and provide an estimate of the probability that each individual was suffering with carpal tunnel syndrome. All individuals with a calculated probability of greater than 0.5 were deemed to have had a positive set of tests and vice versa. Within the patient group, 13 of the 18 sets of results were deemed positive by this method and five were deemed negative (the two patients who did not have a PD recording were excluded from the analysis). In the control group, two sets of results were positive and 18 were negative. This test thus had a sensitivity of 72% and a specificity of 90%. The three variables studied were also evaluated separately and the cross-sectional area measurements were found to contribute most strongly to the sensitivity and specificity of the tests. The positive and negative predictive values for this test are dependent on the prevalence of carpal tunnel syndrome in the population being sampled. In a group consisting solely of patients who are thought to have carpal tunnel syndrome on the basis of their clinical history and examination, the prevalence of this condition is approximately 94% (Gunnarsson et al., 1997) and the positive predictive value of the test is 99%. Similarly, the negative predictive value is 17%. If our ultrasound test was used to screen for carpal tunnel syndrome in the general population, with an estimated prevalence of 0.6% for men and 6.8% for women (de Krom et al., 1992), its positive predictive value would be 0% for men and 33% for women. The negative predictive values would be 100% for men and 97% for women.
Table 1 shows the mean values obtained for the three ultrasound measurements made during this and two other studies. The values obtained by all three groups for asymptomatic controls are very similar, but the values for carpal tunnel syndrome patients varied greatly. The differences between carpal tunnel patients and controls were larger in previous ultrasound studies (Buchberger et al., 1991) than in our study and larger still in previous magnetic resonance imaging studies (Allmann et al., 1997). There are a number of possible explanations for this. One is that the patients in the previous studies had more severe compression of their median nerve compared to our group. Buchberger et al. (1991) had 13 out of 20 patients with severe carpal tunnel syndrome using electrodiagnostic test criteria. However Allmann et al. (1997) had patients with less marked changes on electrodiagnostic testing so this is unlikely to be the sole source of the discrepancy. Another possible explanation is that the other groups were making more accurate measurements, either because of higher resolution machinery or greater user experience. This seems less plausible given the similarity of the control values in all three groups. It seems more likely that the differences are explained by our use of a more general population and clinical findings to confirm the diagnosis.
The results obtained in this study suggest that ultrasound examination is a relatively poor predictor of true carpal tunnel syndrome with a sensitivity of only 72%. A clinical history and examination by an experienced hand surgeon has been estimated to have a sensitivity of 94% (Gunnarsson et al., 1997). In the same paper, electrodiagnostic studies gave a sensitivity of only 85%, but this can be enhanced by more detailed testing. Techniques such as palmar studies and comparisons of median to ulnar nerve conduction velocities can yield sensitivity values of up to 95% (Jablecki et al., 1993) However, in some population groups, ultrasound tests may still prove useful. For instance the positive predictive value of these tests in a group of patients clinically suspected to have carpal tunnel syndrome by an experienced hand surgeon would be 100% though the negative predictive value would be very low so that a negative test would not effectively exclude the diagnosis. In the general population, the reverse situation applies. Prevalence rates are low and the negative predictive value of the tests is then high, between 97% and 100%. In this group, a negative test would effectively exclude the diagnosis of carpal tunnel syndrome, but the positive predictive value is also low so that a positive test would not confirm the diagnosis: more weight would have to be placed on the clinical history and examination.
Received for publication October 12, 2001. Accepted for publication September 9, 2002.
  CiteULike  Connotea  Del.icio.us  Digg  Reddit  Technorati What's this?
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
Abstract
INTRODUCTION