Median Nerve and Carpal Tunnel Morphology Before and After Endoscopic Carpal Tunnel Release: A 6-Year Follow-up Study

Abstract

Background:

Our purpose was to describe structural and morphological features of the median nerve and carpal tunnel on magnetic resonance imaging (MRI) studies obtained before, immediately after, 6 weeks after, and 6 years after endoscopic carpal tunnel release (ECTR).

Methods:

In this prospective cohort study, 9 patients with a diagnosis of carpal tunnel syndrome (CTS) underwent ECTR. Standardized MRI studies were obtained before ECTR, immediately after ECTR, and 6 weeks and 6 years after surgery. Structural and morphological features of the median nerve and carpal tunnel were measured and assessed for each study with comparisons made between each time point.

Results:

All 9 patients had complete symptom resolution postoperatively. On the immediate postoperative MRI, there was a discrete gap in the transverse carpal ligament in all patients. There was retinacular regrowth noted at 6 weeks in all cases. The median nerve cross-sectional area and the anterior-posterior dimension of the carpal tunnel at the level of the hamate increased immediately after surgery and these changes were maintained at 6 years.

Conclusions:

We defined structural and morphological changes on MRI for the median nerve and carpal tunnel in patients with continued symptom resolution 6 years after ECTR. Changes in median nerve and carpal tunnel morphology that occur immediately after surgery remain unchanged at mid-term follow-up in asymptomatic patients. Established imaging criteria for CTS may not apply to postoperative patients. Magnetic resonance imaging appears to be of limited clinical utility in the workup of persistent or recurrent CTS.

Introduction

Carpal tunnel syndrome (CTS) remains the most frequently encountered form of peripheral compressive neuropathy in the upper extremity.^1,2 Since its introduction, endoscopic carpal tunnel release (ECTR) has increased in popularity with nearly 40% of American Society for Surgery of the Hand members using this technique for the majority of their carpal tunnel releases (CTRs). ³ While median nerve decompression at the wrist for CTS results in high rates of symptom resolution and patient satisfaction, it is estimated that 3% to 12% of patients will undergo a revision CTR.^4-7 While there have been a number of studies analyzing the long-term clinical outcomes of CTR, there have been a paucity of investigations analyzing structural and morphological changes to the carpal tunnel and median nerve beyond 1 year after release.^8-10

Recurrent or persistent symptoms related to median nerve compression can be multifactorial with technical complications, iatrogenic injuries, initial misdiagnosis, and disease recurrence among the most common etiologies.^7,11 Both diagnosis and management of recurrent CTS remain challenging as electrodiagnostic studies can remain abnormal even with symptomatic clinical improvement. ¹² Diagnostic injection, magnetic resonance imaging (MRI), and ultrasound have been proposed as part of the diagnostic workup after failed CTR.^2,13-17 In the setting of recurrent or persistent median nerve symptoms, MRI has the ability to evaluate both median nerve and carpal tunnel morphology.

The purpose of this investigation was to describe structural and morphological features of the median nerve and carpal tunnel on MRI studies obtained before, immediately after, 6 weeks after, and 6 years following successful ECTR. We aimed to compare preoperative MRI findings with MRI results 6 years after ECTR.

Materials and Methods

Institutional review board approval was obtained for this prospective cohort investigation. Patients over the age of 18 years with a diagnosis of CTS were eligible to participate in this investigation. Patients were excluded if they had any prior hand surgery or any comorbid conditions associated with CTS including diabetes mellitus, gout, or inflammatory arthritis. All patients were evaluated by the senior author (J.C.K.) in the outpatient clinic, part of a rural, academic, level I trauma center. Our sample size was determined by departmental funding received for this investigation, which allowed for 9 patients to participate. A total of 9 patients agreed to participate during their initial outpatient visit in 2012.

Carpal tunnel syndrome was diagnosed by history and physical examination, with a CTS-6 score of greater than 12 considered diagnostic. Upon agreeing to participate, patients were scheduled for an MRI before planned ECTR. Patients then underwent ECTR with a single, fellowship trained hand surgeon in the manner described by Ruch and Poehling. ¹⁸ Immediately after surgery, all patients underwent a second wrist MRI. Follow-up MRI studies were then obtained 6 weeks and 6 years postoperatively.

All MRIs were performed on a 1.5-T GE magnet (Fairfield, Connecticut) with a dedicated 8-channel wrist coil in the same manner as described previously by Beck et al. ¹⁵ Magnetic resonance images were analyzed by a musculoskeletal trained radiologist. We assessed for a discrete gap or division in the transverse carpal ligament (TCL) on the axial images from the level of the pisiform to the hook of hamate. Similar to Campagna et al, we defined retinacular regrowth as continuous, linear areas of low signal intensity superficial to the median nerve in the area of release. ¹³ This retinacular regrowth could include areas of incomplete TCL release, scar tissue, and retinacular regrowth. ¹³ At both the level of the pisiform and the hook of the hamate, we quantified the median nerve height, width cross-sectional area (CSA), and width-to-height ratio. We also measured the TCL width and the anteroposterior dimension of the carpal tunnel, which was determined by measuring the distance between the floor and the roof in the center of the carpal tunnel. These were determined both at the level of the pisiform and hook of hamate.

Statistics

Descriptive statistics were generated for MRI measurements of each variable of interest at each time point. The mean and standard deviation and minimum-maximum range were reported for continuous variables. In addition, pairwise differences in MRI measurements were calculated between each of the time points for the variables of interest. A 1-sample t test, Wilcoxon signed-rank/sign test were conducted to test the significance of the pairwise differences, where appropriate. Differences of P < .05 were considered statistically significant. A post hoc power analysis was conducted to determine what minimum detectable pairwise mean difference in median nerve CSA at the level of pisiform would have had 80% statistical power to be detected with the sample size (N = 9) when comparing between any 2 time points.

Results

A total of 9 patients participated in this investigation, and each patient completed MRI studies at all time points. Mean patient age was 58 years (range: 36-72), and 6 (66%) were women. Five patients underwent ECTR on the right hand and the dominant hand was involved in 5 patients. There were no postoperative complications. Carpal tunnel syndrome-6 score was greater than or equal to 12.5 for all patients preoperatively. All patients reported complete symptom resolution at the time of the 6-week postoperative MRI with continued relief at final follow-up (CTS-6 scores of 0 in all cases).

On the preoperative T2 weighted MRI images, the intact TCL was a continuous curvilinear dark band of signal. In all 9 patients, the immediate postoperative T2-weighted MRI images demonstrated TCL discontinuity with an intervening fluid bright gap between the opposed ends of the TCL. On all the 6-week postoperative T2-weighted MRI images, the gap in the TCL could be identified, but had begun to fill in with a curvilinear band of signal that morphologically resembled the native TCL (retinacular regrowth), but was brighter, or more intermediate in signal intensity, rather than the dark black in signal seen in the native TCR. In all cases, the appearance of the TCL on T2-weighted MR images at the 6-year postoperative period resembled and could not otherwise be distinguished from the native, intact, dark TCL seen on the preoperative studies (Figure 1). There was no evidence of perineural fibrosis on any of the postoperative MRIs.

Figure 1.

Axial T2 (transverse relaxation time)-weighted fat-saturated magnetic resonance imaging from the same patient before endoscopic carpal tunnel release (ECTR) (a), immediately after ECTR (b), 6 weeks (c), and 6 years (d) after ECTR.

Note. Arrows indicate the region of the transverse carpal ligament. Immediately after ECTR, there is a distinct gap in the flexor retinaculum. By 6 weeks, there is evidence of retinacular regrowth. At 6 years after ECTR, the flexor retinaculum appears similar to the preoperative image. HH = hook of hamate.

Table 1 includes median nerve and carpal tunnel measurements for the preoperative, immediate postoperative, 6-week postoperative, and 6-year postoperative MRIs. In comparing the preoperative and immediate postoperative MRIs, the median nerve CSA at the level of the pisiform increased from 10.9 to 15.0 mm² (P = .02). A-P dimension of the carpal tunnel at the level of the hook of the hamate also increased from 12.3 to 14.5 mm, but appeared to return to the preoperative baseline at the time of the final follow-up (12.0 mm). A post hoc power analysis demonstrated that with the sample size (N = 9), there would have been 80% power to detect a minimum pairwise mean difference of 4.3 mm² in median nerve CSA at the level of pisiform when comparing between any 2 time points.

Table 1.

Preoperative and Subsequent Postoperative Measurements for Patients With CTS That Underwent ECTR.

Variable	Preoperative	Immediate postoperative	6-wk postoperative	6-y postoperative
Median nerve cross-sectional area (in mm2) at the level of
Pisiform
Mean (SD)	10.9 (3.9)	15.0 (5.2)	14.0 (3.6)	14.2 (3.7)
Min, max	6.5, 16.6	10.4, 27.6	7.6, 20.7	8.2, 19.8
Change from preoperative, %	–	38%	28%	30%
Hook of hamate
Mean (SD)	10.0 (2.5)	11.7 (3.6)	12.3 (4.0)	14.3 (3.2)
Min, max	6.0, 13.2	8.5, 20.3	4.5, 15.8	11.1, 18.8
Change from preoperative, %	–	17%	23%	43%
Median nerve width (in mm) at the level of
Pisiform
Mean (SD)	6.4 (1.3)	6.9 (1.4)	6.9 (1.3)	6.4 (1.5)
Min, max	5.0, 8.6	5.4, 10.0	4.4, 9.2	4.7, 9.5
Change from preoperative, %	–	8%	8%	0%
Hook of hamate
Mean (SD)	5.8 (1.2)	6.1 (1.1)	6.3 (1.4)	6.4 (0.9)
Min, max	4.1, 7.5	4.8, 8.0	3.3, 7.6	4.9, 8.2
Change from preoperative, %	–	5%	9%	10%
Median nerve height (in mm) at the level of
Pisiform
Mean (SD)	2.5 (0.4)	2.8 (0.6)	2.7 (0.2)	2.9 (0.7)
Min, max	1.8, 2.9	2.0, 3.8	2.5, 3.1	2.2, 4.5
Change from preoperative, %	–	12%	8%	16%
Hook of hamate
Mean (SD)	2.3 (0.4)	2.5 (0.5)	2.4 (0.3)	3.2 (0.7)
Min, max	1.7, 2.9	2.0, 3.4	2.0, 2.8	2.6, 4.8
Change from preoperative, %	–	9%	4%	39%
Median nerve ratio (width/height) at the level of
Pisiform
Mean (SD)	2.6 (0.4)	2.5 (0.7)	2.5 (0.3)	2.3 (0.7)
Min, max	2.0, 3.2	1.8, 4.2	1.8, 3.0	1.1, 3.8
Change from preoperative, %	–	4%	4%	12%
Hook of hamate
Mean (SD)	2.5 (0.5)	2.5 (0.5)	2.5 (0.4)	2.1 (0.5)
Min, max	1.7, 3.1	1.9, 3.3	1.5, 2.8	1.3, 2.8
Change from preoperative, %	–	0%	0%	16%
TCL width (in mm)
Pisiform
Mean (SD)	1.3 (0.2)	1.2 (0.4)	1.4 (0.5)	1.6 (0.5)
Min, max	1.0, 1.8	0.6, 1.6	0.7, 2.1	0.6, 2.2
Change from preoperative, %	–	8%	8%	23%
Hook of hamate
Mean (SD)	1.4 (0.3)	1.6 (0.4)	1.6 (0.4)	1.8 (0.5)
Min, max	0.9, 2.0	1.2, 2.2	1.2, 2.5	1.0, 2.5
Change from preoperative, %	Ref	14%	14%	29%
Anteroposterior dimension of the carpal tunnel (in mm)
Pisiform
Mean (SD)	11.1 (1.3)	12.6 (1.6)	12.2 (2.3)	11.3 (2.3)
Min, max	8.8, 12.7	10.5, 14.6	8.8, 14.7	8.2, 15.6
Change from preoperative, %	Ref	14%	10%	2%
Hook of hamate
Mean (SD)	12.3 (2.5)	14.5 (2.3)	14.0 (2.3)	12.0 (2.1)
Min, max	9.0, 15.7	11.3, 18.7	11.1, 18.0	9.4, 15.4
Change from preoperative, %	Ref	18%	14%	2%

Note. CTS = carpal tunnel syndrome; ECTR = endoscopic carpal tunnel release; TCL = transverse carpal ligament.

Discussion

In our series, median nerve CSA at the level of the pisiform increased immediately after ECTR. We believe that this represents removal of compression at the inlet to the carpal tunnel, allowing for expansion of the nerve after decompression. There has been conflicting evidence regarding changes in median nerve CSA at the carpal tunnel inlet (level of the pisiform) after CTR. Inui et al found that median nerve CSA decreased 1 year after open CTR as measured on ultrasound (US). ¹⁹ Twelve weeks after ECTR, Abicalaf et al noted significantly decreased median nerve CSA at the level of the pisiform as measured on US. ²⁰ In contrast, Ng et al performed MRIs on 35 wrists preoperatively as well as at 3 and 12 months after ECTR. ²¹ These authors noted postoperative increases in median nerve CSA of greater than 10% at the carpal tunnel inlet and outlet. ²¹ In a prior study, our group obtained MRIs 3 months after ECTR and compared these with a control group without CTS, noting that the median nerve CSA at the level of pisiform and hamate were larger for patients who had undergone ECTR. ¹⁵ In our previous investigation, the median nerve CSA at the level of the pisiform was 14.1 mm² 3 months after ECTR compared with 14.0 mm² 6 weeks after ECTR in our current series. ¹⁵ Also, in our current investigation, median nerve CSA increased after ECTR and these changes were maintained during the 6-year study period. While none of the patients in our series had evidence of persistent or recurrent CTS, it is our opinion that these data suggest that increased median nerve CSA above diagnostic thresholds may not be associated with symptomatic CTS (persistence) after surgery. Previous studies have defined a range of 9 to 13 mm² as a diagnostic cutoff for CTS.^20-25 Established diagnostic thresholds defined for primary CTS on MRI and US may not be applicable to postoperative patients.

Numerous investigations analyzing revision CTR have identified “incomplete release” as a main cause of symptom persistence after CTR.^6,26-28 Both Momose et al and Beck et al noted the absence of a distinct gap in flexor retinaculum for all patients on MRIs 3 months after successful ECTR.^15,29 All patients in this investigation had evidence of complete TCL release on the immediate postoperative MRI, but by 6 weeks, there was early evidence of retinacular regrowth. These findings, specific to ECTR, highlight the difficulty in assessing the adequacy of release for patients with persistent carpal tunnel symptoms after surgery. While incomplete release of the TCL can occur, our results highlight the potential difficulty in using advanced imaging to determine an incomplete release after ECTR, particularly when performed at any time point greater than 6 weeks after the index surgery. The immediate postoperative MRI can confirm complete release, but retinacular regrowth and unreleased TCL can appear similar on MRI. Magnetic resonance imaging may offer little additional information for symptom persistence beyond 6 weeks.

Postoperative changes to median nerve and carpal tunnel morphology after ECTR appear to be maintained at 6-year follow-up. On the immediate postoperative MRIs, the A-P dimension of the carpal tunnel at the level of the hook of the hamate increased and remained above the preoperative level throughout the study period. While the method by which the carpal tunnel volume increases after decompression is controversial, prior authors have noted increased A-P dimension on advanced imaging after ECTR.^30,31 Previous investigations employing advanced imaging modalities after CTR have typically been limited to a follow-up period of 12 months or less. In our series, we note that changes to the A-P dimension of the carpal tunnel are maintained 6 years after endoscopic decompression in patients without evidence of symptom recurrence.

Limitations of this pilot investigation include a small sample size with only 9 patients, despite a follow-up period of 6 years. Our post hoc power analysis for median nerve CSA indicated there would have been 80% power to detect a minimum pairwise mean difference of 4.3 mm². Our observed mean difference was 4.1 mm² and while this 0.2-mm² difference may not be clinically meaningful, it suggests our design was underpowered. In addition, given the small sample size, our design was likely underpowered for additional comparisons of statistical significance, and thus, we reported descriptive statistics and the percentage of change from preoperative values. In addition, it is uncertain if these findings are generalizable to open or mini-open CTRs. Furthermore, our investigation was limited to patients who had complete symptom resolution after ECTR and it remains uncertain if these measurements would differ in patients with persistent or recurrent CTS symptoms. Future investigations using advanced imaging for patients with recurrent or persistent CTS symptoms would more accurately assess structural and morphological changes and allow for comparisons to normative values in asymptomatic postoperative patients. It is clear that MRI is not indicated as part of the routine diagnostic workup of primary CTS. As a diagnostic adjuvant, US has continued to emerge as a more appealing option due to the ability to perform the procedure in the office. Before the initiation of this investigation in 2012, some authors had suggested that MRI may have a potential role in the evaluation of recurrent or persistent CTS.^16,17 We agree with more recent authors who have suggested the role for MRI may be limited to instances where masses, tumors, or space occupying lesions are part of the differential for recurrent or persistent CTS. ¹¹ However, MRI has the ability to define structural and morphological changes in both the median nerve and carpal tunnel, and understanding these changes 6 years after operative decompression may have clinical implications.

For patients with symptom resolution after ECTR, median nerve CSA as measured on MRI appears to increase immediately after surgery, and these changes persist at 6 years postoperatively. Changes in median nerve and carpal tunnel gross morphology that occur immediately after surgery appear to be maintained at mid-term follow-up in asymptomatic patients. Given the persistent increases in median CSA and absence of a distinct retinacular gap in patients with symptom resolution 6 years after ECTR, MRI appears to be of limited clinical utility in the workup of persistent or recurrent CTS.