Palmar bowing of the flexor retinaculum on wrist MRI correlates with subjective reports of pain in carpal tunnel syndrome

Abstract

Purpose

To investigate the utility of palmar bowing of the flexor retinaculum (PBFR) on magnetic resonance imaging (MRI) as an adjunct to presurgical evaluation of carpal tunnel syndrome (CTS).

Materials and Methods

The CTS group comprised 48 women with CTS diagnosed from clinical history, physical examination, and electrophysiological findings. Another 21 healthy women with no symptoms of CTS served as controls. The CTS group was divided into four subgroups based on symptom duration. CTS patients were preoperatively assessed for symptom severity using a Likert scale. In both groups, palmar bowing of the flexor retinaculum (PBFR) was measured on MRI. Furthermore, relationships with disease duration, subjective symptom severity, and electrophysiological findings were assessed in the CTS group.

Results

Although PBFR increased significantly in all CTS groups when compared to the control group, PBFR in patients who complained of severe pain tended to exceed that in patients with mild to moderate pain. Statistical analysis showed a positive correlation between PBFR and pain severity. In contrast, PBFR did not reflect median nerve function.

Conclusion

PBFR as seen on MRI seems to correlate significantly with patients' subjective reports of pain severity. J. Magn. Reson. Imaging 2009;29:1102–1105. © 2009 Wiley-Liss, Inc.

PROPERLY PERFORMED ELECTRODIAGNOSTIC STUDIES are highly sensitive and specific in the diagnosis of carpal tunnel syndrome (CTS) (1). However, prolongation of latency is frequently minimal. Furthermore, recent clinical studies with validated outcome assessments have demonstrated that subjective symptoms of CTS do not correlate well with electrophysiological data (2, 3). In fact, patients with mild to moderate median nerve dysfunction tend to report significantly more pain than those with severe median nerve dysfunction, and over 10% of CTS patients with significant symptoms present with normal electrophysiology (4).

From the viewpoint of pathogenesis, increased pressure within the carpal tunnel due to noninflammatory tenosynovial swelling is generally accepted as the cause of idiopathic CTS (5, 6). Histological studies have demonstrated that drastic connective tissue remodeling takes place in the flexor tenosynovium with progression of CTS, with various inflammatory mediators such as cytokines, prostaglandins, and proteases being involved in the process (7–15). Since the critical pathophysiology of CTS arises in the flexor tenosynovium, with the median nerve affected secondarily as a result of changes in the physical properties of the tenosynovium, magnetic resonance imaging (MRI), with its ability to detect soft-tissue abnormalities, has been suggested as an ideal diagnostic adjunct to evaluate the soft tissues within the carpal tunnel.

The present study investigated the role of MRI in the diagnosis and evaluation of CTS with special emphasis on the relationship between intracarpal tunnel volume change and subjective symptom severity.

MATERIALS AND METHODS

The CTS group comprised 48 women (mean age, 51.9 ± 8.9 years; range, 31–73 years) who underwent surgical treatment for CTS. The clinical diagnosis was based on history, clinical examination, and electrodiagnostic studies. Another 21 healthy female volunteers with comparable demographics and no symptoms of CTS served as controls. Sensory nerve conduction studies were performed using an orthodromic technique; sensory nerve conduction velocity (SCV) was assessed with sensory electrodes stimulating the index finger and recording electrodes at the wrist. Motor conduction studies were performed with bipolar surface stimulating electrodes proximal to the carpal tunnel and recording over the abductor pollicis brevis and compound action potentials (CMAPs) were recorded.

Individuals with a history of diabetes mellitus, inflammatory arthritis, or mass lesions in the carpal tunnel were excluded from this study. Most of the CTS group were homemakers and none received worker's compensation. They were subdivided into four groups based on duration from onset of symptoms to surgery: group A (n = 14; ≤3 months); group B (n = 10; 4–6 months); group C (n = 10; 7–12 months); and group D (n = 14, ≥13 months). Demographics of the patients in each group are shown in Table 1. Most patients in group D reported a long-term presence of mild symptoms, with surgery considered only after sudden exacerbation of symptoms. On admission, patients completed a questionnaire regarding preoperative daytime pain, nocturnal pain, and paresthesia in the median nerve distribution of the hands. Each item comprised five possible responses (“none: 1,” “mild: 2,” “moderate: 3,” “severe: 4,” and “very severe: 5”). MR examinations of the wrists were performed using a 0.5T superconductive system equipped with a shielded gradient coil (Toshiba, Japan). Axial images of the wrist were obtained using T1-weighted (repetition time (TR) = 500 msec, echo time (TE) = 20 msec) and T2-weighted (TR = 2000 ms, TE = 80 msec) spin echo sequences with slice thicknesses of 5 mm. Palmar bowing of the flexor retinaculum (PBFR) was calculated as the bowing ratio as reported by Mesgarzadeh et al (16). First, a straight line (line-TH) was drawn between the attachments of the flexor retinaculum to the tubercle of the trapezium and the hook of the hamate, and the length of this line was measured. Next, line-PD was drawn perpendicularly from line-TH to the palmar apex of the flexor retinaculum and the length was measured. Then the bowing ratio was calculated as line-PD divided by line-TH (Fig. 1a). An MR image from the control group is shown for comparison (Fig. 1b).

Table 1. Demographics of Patients in Each Carpal Tunnel Syndrome Group and in Controls

Group	A	B	C	D	Control
Number	14	10	10	14	21
Affected side (right: left)	6: 8	6: 4	6: 4	7: 7	21: 0
Mean age ± SD	51.1 ± 5.3	50.2 ± 9.0	56.6 ± 6.5	53.9 ± 14.2	38.6 ± 14.9

Figure 1. Wrist MRI findings in a CTS patient (a) and in a control (b). PBFR was derived by dividing the length of line-PD by the length of line-TH. Line-PD is the line from the apex of the flexor retinaculum to line-TH. Line-TH is the line between the insertion of the flexor retinaculum into the trapezium and the hook of the hamate.

For statistical analysis, StatView 5.0 for Windows (SAS Institute, Cary, NC) was used. Statistical differences between the groups were evaluated by Kruskal–Wallis Analysis and Mann–Whitney U-test. Correlations were estimated using Spearman rank correlation coefficients. P-values less than 0.05 were considered to indicate statistical significance.

RESULTS

Figure 2a shows preoperative symptom severity in each of the CTS subgroups (A–D). Daytime and nocturnal pain were significantly more severe in the early phase of the syndrome than in the late phase. The only exception was patients in group D, who complained of significantly more severe pain than those in group C. As mentioned above, patients in group D had experienced mild symptoms for a number of years before a sudden exacerbation of symptoms led them to consider surgery, hence this finding seems to be related to the characteristic clinical history of patients in group D. In contrast, no significant differences in paresthesia were identified between groups.

Figure 2. a: Severity of nocturnal pain, daytime pain, and paresthesia in each CTS subgroup (groups A–D). b: PBFR in each CTS subgroup and the control group (group N). *P < 0.05.

Figure 2b shows PBFR expressed as mean ± SD for each group. All four CTS subgroups showed significant increases in PBFR compared to controls. Furthermore, PBFR was also significantly greater in groups A and D than in groups B and C. This is a similar tendency to that observed for pain severity and indicates that patients with more severe pain exhibited increased PBFR.

PBFR was significantly correlated with subjective reports of both daytime and nocturnal pain severity (Fig. 3a,b). However, it did not show any significant relationship with paresthesia, a symptom that can reflect median nerve pathology (Fig. 3c).

Figure 3. Correlations between symptom severity and PBFR: for nocturnal pain (a), daytime pain (b), and paresthesia (c). *P < 0.05.

Figure 4a,b shows CMAP and SCV in each CTS subgroup. Neither CMAP nor SCV differed significantly among the groups. Furthermore, neither CMAP nor SCV showed any correlation with PBFR (Fig. 4c,d).

Figure 4. Electrophysiological findings in each CTS subgroup and their correlation with PBFR: CMAP (a,c) and SCV (b,d).

DISCUSSION

Clinically, CTS can be subdivided into early, intermediate, and advanced phases. Patients with early CTS report intermittent pain and paresthesia, while in the intermediate phase, paresthesia and numbness become constant. As the disease becomes advanced, patients experience permanent loss of sensory and motor function in addition to obvious thenar muscle atrophy (17). In contrast to functional status, recent clinical studies with outcome assessment have demonstrated that subjective symptoms are worse in earlier phases than in the advanced phase (2, 3) and that patients place more emphasis on subjective symptoms when deciding whether to undergo surgery (18). Thus, it is important to evaluate subjective symptoms when diagnosing and treating CTS. We used a Likert scale to measure symptom severity in this study because several researchers have found that patients find this type of scale easier to understand and use than a VAS (visual analog scale), and because Likert scale and VAS responses are highly correlated even in children or the elderly (19, 20). In the present study, only women were investigated because previous studies have found differences in the incidence of CTS and in carpal tunnel area between genders (21, 22). In addition, an Italian CTS research group demonstrated significant gender differences in CTS symptom duration (23), which we used to stratify CTS patients in this study.

Electrophysiological studies have been widely used for the diagnosis of CTS and have sensitivity and specificity of ≈90% (1). However, it is well known that electrophysiological findings are not correlated with subjective symptoms, as confirmed in the present study. The technique has a false-negative rate in patients with clinically defined CTS, as well as a false-positive rate in normal, asymptomatic individuals (2–4). In addition, electrophysiological studies require experienced personnel.

MRI has provided visual evidence of changes to structures within the wrist that are believed to be associated with CTS. Abnormal swelling of the median nerve and PBFR are currently the only two specific parameters that are reliable in the diagnosis of CTS (24–26). PBFR is considered a useful parameter even in low-field extremity MRI such as used in this study. There must be enlargement of some structures within the carpal tunnel for the flexor retinaculum to be pushed in a volar direction. In most patients in this study, tenosynovial thickening within the carpal tunnel was noted during surgery; however, we did not attempt to correlate surgical findings with MRI findings in the present study.

Furthermore, we and others have correlated histological changes of the flexor tenosynovium with clinical findings and found that early CTS shows edematous thickening while advanced CTS exhibits extensive type VI collagen fiber degeneration and significant type III collagen deposition (7, 10). Collectively, these findings suggest that pain in CTS originates not from the median nerve, but from connective tissue surrounding the nerve. In fact, a biochemical study of the flexor tenosynovium in CTS demonstrated significant expression of interleukin 6 and prostaglandin E2 in the absence of inflammation (12, 14, 27).

A prospective cohort study by Jarvik et al (28) concluded that the reliability of MRI is high but that diagnostic accuracy is only moderate. The utility of MRI in CTS diagnosis has largely been determined by comparing specific parameters measured on MRI to severity of electrophysiological findings.

In conclusion, an important observation made in the present study is that PBFR seen on MRI seemed to correlate significantly with patients' subjective reports of pain severity, although neither correlated well with electrophysiological or surgical findings, in contrast to previous studies.