Endoscopic release of the carpal tunnel: A randomized prospective multicenter study

ORIGINAL COMMUNICATIONS Endoscopic release of the carpal tunnel: A randomized prospective multicenter study A 10-center randomized prospective multicenter study of endoscopic release of the carpal tunnel was carried out. Surgery was performed with a new device for transecting the transverse carpal ligament while control hands were treated with conventional open surgery. There were 122 patients in the study; 25 had carpal tunnel surgery on both hands and 97 had surgery on one hand. Of the surgical procedures, 65 were in the control group and 82 were in the device group. The endoscopic device was coupled to a fiberoptic light and a video camera. A trigger-activated blade was used to incise the transverse carpal ligament. After surgery, the best predictors of return to work and to activities of daily living were strength and tenderness variables. For patients in the device group with one affected hand, the median time for return to work was 21111 days less than that for the control group . Two patients treated with the endoscopicdevice required reoperation by open surgical decompression; only one of these had incomplete release with the device. Two patients in the device group experienced transient ulnar neurapraxia. (J HAND SURG 1992;17A:987-95.)

John M . Agee, MD, Sacramento, Calif , H. Relton McCarroll , Jr. , MD, San Francisco, Calif., Richard D. Tortosa , MD, Yuba City, Calif., Donald A. Berry, PhD, Durham , N.C ., Robert M. Szabo, MD, Sacramento, Calif., and Clayton A . Peimer, MD, Buffalo , N .Y.

Carpal tunnel syndrome is the most commonly diagnosed peripheral entrapment neuropathy. I The technical aspects of its surgical care are controversial? and varied. I , 3 Although conventional open surgical release of the carpal tunnel predictably relieves the symptoms of median nerve compression , it is frequently associated with postoperative pain , tenderness , and weakness that delay the patient's return to activities of daily living (ADL) and emplo yment. In contrast to the open surgical technique , som e surgeons have advocated a transverse skin incision at the

Supported in part by the 3M Orth opedic Products Division, 51. Paul , Minn. Received for publicat ion OCI. 24, 1991; accepted in revised form May 20 , 1992. One or more of the authors have received or will receive benefits for personal or professional use from a commercial party related directly or indirectl y to the subject of this article. Reprint requests : John M. Agee , MD, Director, Hand Biomechanics Laboratory, Inc" 77 Scripps Dr. , Suite 101, Sacramento, CA 95825 .

3/1/38959

wrisr' combined with a blind surgical release of a transverse carpal ligament by means of knives, scissors, and special instruments. It is our hypothesis that a surgical approach that combines endoscopic visualization of the transverse carpal ligament with a limited surgical incision will offer the benefits of decreased postoperative morbidity without unacceptable risk to the patient.

Materials and methods Endoscopic device. The endoscopic device (3M Orthopedic Products Division , St. Paul, Minn.) and the surgical technique were developed at the Hand Biomechanic s Laboratory in Sacramento, Calif. The design of the device was developed in concert with the surgical technique," with each evolving through the use of a series of prototypes on fresh cadaver spe cimens. The device has a pistol grip handp iece with an integral trigger mechanism from which extends a blade assembly desig ned in size and shape to be inserted down the length of the carpal tunnel. A rectangular window on the flat upper surface near the tip of the blade assembly permits viewing of the transverse carpal ligament through an endoscope positioned just proximal to the

THE JOURNAL OF HAND SURGERY

987

988

Agee et al,

Fig. 1. The probe is inserted through an incision in forearm fascia. All instruments and the device follow the ring finger ray.

window. A trigger-actuated retractable blade immediately distal to the window elevates 3.5 mm above the surface of the blade assembly. The flat upper surface of the blade assembly is designed to avoid injury to the median nerve and digital flexor tendons when the window is snugly applied to the deep side of the ligament. A video camera and a fiberoptic light source coupled to the endoscope provide direct viewing and illumination similar to those provided by other endoscopic/ athroscopic equipment. Surgical technique. With the use of tourniquet control and regional block or general anesthesia, a 2 em skin incision is placed in the wrist flexion crease between the flexor carpi radialis and flexi carpi ulnaris tendons. A spreading subcutaneous dissection protects the underlying palmar cutaneous nerve. A probe inserted through a distally based V-shaped incision in the forearm fascia helps define a proximal-to-distal path down the palmar-ulnar aspect of the carpal tunnel. With the device aimed at the ring finger and with the wrist in extension, the blade assembly of the device is inserted along the same path (Fig. 1). Through the window near the tip of the device's blade assembly, a strip of transverse carpal ligament that is

The Journal of HAND SURGERY

free of all other structures is visualized. Once the distal edge of the ligament is clearly identified and the ulnar position and ring finger aim are checked, the trigger is pulled to elevate the blade in a triangular area defined by the ulnar half of the distal edge of the transverse carpal ligament, the ulnar border of the median nerve and/ or its digital branches, and the proximal edge of the superficial palmar arch (Fig. 2). Application of upward pressure to hold the window against the deep side of the ligament as the device is gently withdrawn excludes flexor tendons and the median nerve during incision of the transverse carpal ligament (Fig. 3). Additional passes may be necessary for complete release of the transverse carpal ligament, with endoscopic control permitting visualization of its cut edges. Study design. A ten-center study compared conventional open primary carpal tunnel release with closed endoscope-assisted release. The study design was reviewed and approved by each center's institutional review board, and informed consent was obtained from each patient. Patients were required to meet entry criteria designed to identify idiopathic carpal tunnel syndrome supported by abnormal nerve studies but normal electromyograms. Patient inclusion/ exclusion criteria included the absence of generalized peripheral neuropathy as well as diabetes mellitus or thyroid disease, anatomic abnormalities of the wrist and hand (including previous displaced fractures), inflammatory joint diseases, a history of median nerve injury from trauma (including contusion, fracture of the distal radius or wrist joint, etc.), previous surgery on the wrist, vasospastic disorders such as Raynaud's disease or sympathetic dystrophies and their equivalents, psychiatric disorders, chronic renal disease requiring dialysis, previous carpal tunnel release, thenar weakness sufficient to require tendon transfer to support thumb opposition at the time of carpal tunnel release (Camitz or other), and any other surgery required at the time of carpal tunnel release. There were 122 patients in the study; 25 had carpal tunnel surgery on both hands and 97 had surgery on one hand. Thus a total of 147 carpal tunnel procedures were performed. Of these, 82 were in the device group and 65 were in the conventional open surgical control group. Of the patients with unilateral procedures, 79 were workers; of these, 49 were in the device group and 30 were in the control group. The only significant difference between the two treatment groups was the use of diuretics during preoperative treatment of the carpal tunnel syndrome in 2 of 82 patients in the device group and 7 of 65 patients in the control group. Some differences were expected on the basis of chance alone; however, these differences should not have had a material effect on the results.

Vol. 17A, No.6 November 1992

Endoscopic relase of carpal funnel

989

Blade Assembly

Fig. 2. Safe zone of blade elevation. The safe zone is within a triangle defined by the ulnar half of the distal edge of the transverse carpal ligament (a), the ulnar border of the median nerve, i.e., its common digital branch (b), and the ulnar proximal margin of the superficial palmar arch (c) .

....

Fig. 3. Transverse carpal ligament incision on withdrawal of device.

In patients who had bilateral carpal tunnel syndrome, each hand was operated on in a separate procedure. In the early part of the study, the first hand to be treated surgically was randomized into the control or device groups. Frequently, those patients who had the first carpal tunnel release performed endoscopically refused to have the other hand treated with the conventional open surgical approach. Later in the study, patients with bilateral carpal tunnel syndrome had the control surgery first, followed by surgery with the endoscopic device on the opposite hand. Clinical examination. Before surgery, the following tests were performed on the affected and unaffected hands: grip strength with Jamar dynamometer readings at five settings; pinch strength, both lateral/key and

pulp; Semmes-Weinstein monofilament sensory mapping; Phalen's wrist flexion test; Tinel's test at the carpal tunnel; and manual motor testing, with grades 0 through 5, for thumb abduction. In addition, all patients had preoperative wrist x-ray films, including a carpal tunnel view (normal films were required for patient participation). All patients were questioned about any drug therapy they had received in the 4 weeks preceding entry into the study. At postoperative weeks 1, 2, 3, 6, 9, 13, and 26, the following data were collected: employment status, return of hand use for ADL, postoperative symptoms (same as preoperative), and clinical examination data for comparison with the preoperative data (grip and pinch measurements, monofilament sensory examina-

990

Agee et al.

tion, motor testing of thenar muscles, etc.) Grip strength was never tested at week 1 because of wound healing. In addition, scar tenderness and radial and ulnar pillar tenderness were graded from 0 to 4 (none to very tender), and a record of medication use was completed. Data analysis. For evaluation of the effect of surgery within each treatment group at each visit, the hypothesis of no change from preoperative evaluation was tested. For quantitative variables, such as percentage change from preoperative values, paired t tests" were used. For categorical variables, such as normal! abnormal, McNemar's test" was used. Retired persons, housewives, and persons who had bilateral surgical procedures were excluded from analysis of times for return to work. In the comparison of times of return to ADL, patients who had bilateral surgery were included with those who had unilateral surgery. The data for return to ADL by the two treatment groups of bilateral surgery patients also were compared separately by the Wilcoxon two-sample rank test. 6 The Kaplan-Meier method? was used to estimate median times of return to work and to ADL. Treatment groups were compared on the basis of times of return to work and to ADL by means of a survival analysis version of the Wilcoxon test. 8 Results Strength, sensory, and motor testing. Hands in both treatment groups were weaker in the early postoperative period than they were before surgery. Hands in both' groups became stronger with time. Patients in the device group returned to preoperative or greater strength more quickly than those in the control group (p < 0.05, Table I). There was no difference in sensory and motor (thenar) test results between the device and control groups (Table II). Within both treatment groups, sensory test results improved (p < 0.05) at week 2 as compared with preoperative measurements, except those for the small finger, which showed little if any subsequent improvement. Motor tests showed continual improvement in both groups, with both groups eventually exceeding preoperative levels. Phalen's and Tinel's test scores improved dramatically and similarly in the two groups (p < 0.05, Table III). Patient assessments. As shown in Table IV, the percentage of patients with pain, tingling, weakness, and numbness decreased over the course of the study within both treatment groups (p < 0.05). However, one of four patients in each group was still reporting some pain 6 months after surgery. Similarly, 6 months after surgery, about one patient in eight reported tingling,

The Journal of HAND SURGERY

one in three was reporting weakness, and one in six was reporting numbness. Table V shows that patients in the device group experienced less scar tenderness throughout the study than did controls, with the difference being statistically significant at weeks 1,2,3, and 9. The device group also experienced less radial pillar tenderness 3 and 9 weeks after surgery, and less ulnar pillar tenderness 1 week after surgery, while tenderness reported during the remaining visits was comparable in the two groups. Median times of return to work and to ADL. The 50% line in Fig. 4 shows that the median time for patients with unilateral surgery to return to work was 25 days for the device group and 46.5 days for the control group (p < 0.01). Patients with workers' compensation took much longer to return to work than did patients without workers' compensation. The median time to return to work for patients on workers' compensation was 71 days in the device group and 78 days in the control group (Table VI). Among patients not receiving workers' compensation, the device group had a greater advantage in time for return to work over the control group (p < 0.01) than was observed when all patients in the two groups were compared. Median time to return to work for unilateral surgery patients without workers' compensation in the device group was 16.5 days as compared to 45.5 days in the control group (Fig. 5 and Table VI). Also, 75% of the patients in the device group had returned to work by 42 days whereas it took 52 days for 75% of the control patients to return to work. (Fig. 5). The median times for return to ADL for all patients, with and without workers' compensation, are shown in Table VI. For the patients without workers' compensation, these times were 5 and 13 days, respectively, for the device and control groups (p < 0.01). Of the 25 patients with bilateral operations, 16 returned to ADL sooner with the device-treated hand than they did with the control hand. The difference was not statistically significant. Center effect. In a multicenter study, there is a possibility that a few centers could influence the results in a given direction whereas other centers could have the opposite effect. When data from each center were examined separately, there was reasonable uniformity in the results across centers, with the direction of the effect being the same for virtually all centers. Complications. TWo patients in the device group had persistent signs and symptoms of carpal tunnel syndrome that required reoperation by conventional open surgery. One patient, with an incomplete release of the proximal extent of the transverse carpal ligament, had complete resolution of symptoms after surgery. The

Vol. 17A, No.6 November 1992

Endoscopic relase of carpal IUIl/Iel

991

Table I. Strength by time since surgery Mean percent changefrom preoperative value Postoperative week Strength measurement" group Jamar grip Device Control Lateral I key pinch Device Control Pulp pinch Device Control Sample sizet Device Control

I NA NA

2

I

6

I

3

I

I

9

/3

I

26 14~

-55

-2U -38

-12 -22

-16§ -32

-211 -19

6

m

14~

21~

-52

-I

3

10

14~

-26 -52

-28

-II

I

10 7t

9 17

1411 17

74 51

74 60

79 57

73 57

72

59

64 48

-33~

-27~

-4

-II§

-4 -14

6t

72

-2

9

55

'Strength was measured in pounds and expressed as percent change from preoperative value. tSample sizes are maxima; for some variables, a small number of observations was missing. *Significant improvement compared with control, p -c 0.001. §Significant improvement compared with control, p <: 0.05. [Significant improvement compared with control, p < 0.01. ~Signifieanl improvement compared with preoperative. p <: 0.05.

Table II. Sensory and motor testing as a function of time Postoperative week Test. group Sensory, thumb Device'

Control"

I

Preoperative

2

Device»

Control" Sensory. little finger Device Control Sample size, sensory testing Device Control Motor Device Control Sample size, motor testing Device Control

3

,

6

I

9

I

I3

I

26

Percent of patients testing normal'[

30 30

NA NA

47 44

41 41

NA NA

61 52

43 36

NA NA

60 56

77

0 0

76 71

67

0 0

74 59

60 48

30 20

43 33

82 65

71 54

74

Sensory, index finger Control" Sensory, long finger Device'"

I

67 77

60

50 44 Percent of patients testing normait 67 57 61 64 65 66 Percent of pal/ems testing normali 68 75 68 53 61 62 Percent of patients testing normati 82 86 86 79 79 95 65 52

79 56

43 55

44 42

44 34 Percent of patients testing normaii 57 71 74 60 44 73

79 57

72

58

74 56

43 48

72

62 52

74 81

65 67

90 76

76 85

90 90

37 27

57 42

81 74

81 83

74 55

63 47

64

'Significantly different (p <: 0.05) from preoperative values at all or essentially all postoperative time periods. t A normal value was 2.83 by Semmes-Weinstein monofilament sensory mapping. *On a scale of I 10 5, a manual motor lest score of 5 was normal.

second patient had persistent signs and symptoms of carpal tunnel syndrome that were unrelieved by open surgical decompression, which demonstrated that there had been prior complete release of the transverse carpal ligament. Two patients in the endoscopic device group

had transient ulnar neurapraxia, which was presumed to be secondary to retraction at the wrist. In the control group, one patient suffered injury to the deep motor branch of the ulnar nerve but refused surgical reexploration. Another control patient had

The Journal of 992

HAND SURGERY

Agee et al.

Table III. Phalen's and Tinel's testing as a function of time Postoperative week Test group

I

Preoperative

Phalen's Device" Control"

11 6

NA NA

NA NA

30 42

NA NA

NA NA

Sample size Device Control

82 65

0 0

0 0

6

I

13

I

9

Percent of patients testing negative NA NA 80 NA 93 NA Percent of patients testing negative NA NA 81 NA NA 80

Tinel's Device" Control"

I

3

I

2

0 0

0 0

74 56

i

26

NA NA

94

NA NA

88 83

0 0

46

92

64

·Significantly different (p < 0.05) from preoperative values at 9 and 26 weeks aftersurgery.

Table IV. Patient assessments of symptoms as a function of time Percent of patients with symptom present Postoperative week Symptom group

I

2

I

I

6

I

I

26

Preoperative

1

85 89

43 59

51 58

67 58

63 68

50 54

42 49

25 27

95 89

34 21

31 32

27 25

29 19

27 18

20 9

14 15

82 86

64 71

69 76

63 82

54 63

49 54

32 44

20 35

98 97

32 39

34 29

32 21

20 19

26 14

22 13

12 19

83 89

13 21

11

18 18

8 8

8 8

11

11

11

8 8

68 72

15 12

22 19

16 16

16 20

22 18

65 62

20 20

23 29

18 25

15 17

20 14

13

12 12

82 65

77

56

74 62

79 57

73 59

74 57

74 55

6S 48

3

9

13

I

Pain Devlce"

ControlTingling Devices Control"

Weakness Device" Control"

Numbness Device" Controls

Night symptoms Device· Control" Dropping items Device· Control"

Fine dexterity loss Device's Control" Number of patients]' Device Control

18 11

15

14 10

·Signiflcant improvement (p < 0.05) compared with preoperative values at all or essentially all postoperative time periods. tNumbers of patients are maxima; forsome variables, a small number of observations was missing.

bowstringing of the digital flexor tendons, and two control patients had wound dehiscence. Discussion

The hypothesis tested in this study was that postoperative morbidity after carpal tunnel release can be decreased by a surgical approach that divides only the

transverse carpal ligament, leaving the overlying skin, subcutaneous fat, palmar fascia, and palmaris brevis muscle intact. A surgical approach that views and divides the transverse carpal ligament without separately decompressing the median nerve by means of neurolysis seems justified. Gelberman et al. 9 and Mackinnon et al. 2, 10 dem-

Vol. 17A, No.6

Endoscopic relase of carpal tunnel

November 1992

993

Table V. Patient assessments of tenderness by time since surgery Average tenderness score" Postoperative week Tenderness measurement group

2

I

Scar Device Control Radial pillar Device Control Ulnar pillar Device Control Number of patients'[ Device Control

I

l.4t 2.0

1.1§ 1.8

0.8 1.0

0.8 0.9

0.9*

0.5 0.7

1.3 77 55

3

I

6

Lot

9

I

I

13

26

0.8 1.2

0.4* 0.8

0.3 0.5

0.1 0.3

1.1

0.6 0.9

0.4* 0.8

0.3 0.5

0.2 0.3

0.7 0.8

0.6 0.6

0.4 0.5

0.3 0.3

0.1 0.2

o.n

79 57

73 60

I

73 59

74 55

72

65 47

55

'Scored from 0 (no tenderness) to 4 (very tender). tSignificance of difference between device and control group, p < 0.05. :iSignificance of difference between device and control group, p < 0.01. §Significance of difference between device and control group. p < 0.001. llThe numbers of patients are maxima; for some variables, a small number of observations was missing.

Table VI. Median time (in days) for unilaterally treated patients to return to work and to activities of daily living Return 10 work (unilaterally treated patients)

Group Device Control

All 25' (49)t 46.5 (30)

I

Workers' compensation

I

71 (11) 78 (10)

'Significance of difference between device and control, p tNurnber in parentheses is number of patients.

Return to ADL (all patients)

No workers' compensation

All

16.5* (38) 45.5 (20)

9 (81) 13 (63)

I

Workers' compensation 16 (23) 12 (22)

I

No workers'

compensation 5* (58) 13 (41)

< 0.01.

onstrated that neurolysis was not a necessary or beneficial addition to surgical treatment of carpal tunnel syndrome. In the last two decades, multiple areas of surgery have benefitted from approaches that combine arthroscopic and/ or endoscopic technology with new instrumentation to decrease morbidity by minimizing surgical trauma. This technique differs from arthroscopic surgery in that the operative site is a synovium-lined tunnel and the endoscope views the transverse carpal ligament through air rather than through a liquid medium. Okutsu et al." and Chow!" 13 also have reported on endoscopic carpal tunnel release. Okutsu et al. 11 inserted a clear plastic tube into the carpal tunnel and then introduced a standard arthroscope down its length. A knife passed along the ulnar side of the tube divided the transverse carpal ligament, with the adjacent endoscope viewing the ligament incision through the clear

wall of the tube. Chow" released the transverse carpal ligament by inserting a metal tube, open on its palmar surface, through the carpal tunnel to exit through a second incision in the middle of the palm. With the transverse carpal ligament applied to the slit in the palmar surface of the tube, an endoscope introduced distally viewed the ligament during division by proximally introduced knives. In a second study, Okutsu et al." used catheter measurements of carpal tunnel pressures to demonstrate that they can be lowered effectively by endoscopic surgery. Conventional carpal tunnel surgery is one of the surgeon's most successful operations, with virtually all patients recovering. IS However, a significant concern is the length of recovery time. Return to work is one measure of recovery from carpal tunnel surgery. Recovery as assessed by return to work has multiple factors and is especially dependent on the decision maker,

The Journal of 994

Agee et al.

HAND SURGERY

100,-------------------, ~

a: a ~ 75

a

I-

z rr.

::) 50 IW

a: I-

z

w a:

o

25

w

a..

O+--,.-L--r..L.--,---,-----,---,.---,---,-------,r------j

o

20

40

60

80 100 120 140 160 180 200

DAYS POSTOPERATIVE Fig. 4. Time for all patients with unilateral surgery to return to work. The device group had 49 patients, and the median time for return to work was 25 days. The control group had 30 patients, and the median time for return to work was 46V2 days.

:::.:::

a:

a o I-

s:

75

z

a:

:::J I-

50

W

a: IZ W

o

25

a:

w

o,

O+-----'-,r--,..--.----,r----r----.--r----r--.-~

o

20

40

60

80 100 120 140 160 180 200

DAYS POSTOPERATIVE Fig', 5. Time for patients with unilateral surgery and no workers' compensation to return to work. The device group had 38 patients, and the median time for return to work was 16V2 days; the control group had 20 patients, and the median time for return to work was 45'/2 days.

which is variable (i.e., patient, surgeon, employer, insurance company, etc.). Improvement in most of the variables measured translated into earlier return to work and to ADL. Strength and tenderness variables were the best predictors of return to both work and ADL. These are the variables for which the device group most clearly benefitted over the controls, and they explain the earlier return to work and ADL by the device-treated patients than by the control patients.

One contributing factor to the variability in time for return to work is that workers' compensation benefits were associated with delayed return to work. Patients who are not receiving workers' compensation are motivated to return to work and are therefore more likely to do so when they are physically able. The median time for the device group of patients on workers' compensation to return to work was 71 days, and for noncompensation patients the median time was 16.5 days. In the control group, the median time for workers'

Vol. 17A, No.6 November 1992

compensation patients to return to work was 78 days, while in the noncompensation group, the median time was 45.5 days. For these reasons, the effectiveness of any approach may be assessed more clearly in patients who are not receiving workers' compensation. The improvement in median time of return to work for these patients in the device group was 29 days (45.5 minus 16.5 days), which is about 1 month of work gained for each carpal tunnel patient who required surgery. An effective endoscopic procedure could allow third-party payers to modify their recommended time off work for carpal tunnel surgery. This would greatly benefit workers, employers, compensation insurance companies, and industry. Strength and tenderness variables are the best predictors of return to work. Gellman et al." found similar trends in recovery of postoperative strength. These are the variables for which the device group showed the clearest benefit over the controls, and they serve to explain why patients treated with the endoscopic device returned to work sooner than did the controls. Strength and tenderness variables were also the best predictors of return to ADL, where again the device group showed the clearest benefit over the controls. Significant potential risks attend carpal tunnel decompression by all methods (Nancollas MP, Peimer CA, Wheeler DR, Sherwin FS. Long-term results of carpal tunnel release. American Association of Orthopaedic Surgeons meeting. March 7-12,1991, Anaheim, Calif.)." The 82 patients treated with the device suffered no injury to the median or ulnar nerves or to the superficial palmar arterial arch. However, median nerve injuries have occurred with the device in patients outside this study. The endoscopic device used for this study was manufactured and marketed by the Orthopedic Products Division, 3M Health Care Group, St. Paul, MN. Subsequent to the study, the product was withdrawn from the market. The blade assembly has been redesigned to permit accurate endoscopic visualization of the point of blade entry into the transverse carpal ligament. This is a preliminary study with limited follow-up. Whether release of the transverse carpal ligament alone will result in an increased incidence of recurrent and/ or persistent carpal tunnel syndrome is not known. Follow-up studies on these patients will be done and the results will be reported. We acknowledge the study participation of David P. Green, MD, Jack A. McCarthy, MD, William H. Neuman MD, Edward R. North, MD, and Gary G. Poehling, MD.

Endoscopic relase of carpal tunnel

995

REFERENCES 1. Pfeffer GB, Gelberman RH, Boyes JR, Rydevik B. The

2.

3. 4.

5.

6.

7.

8. 9.

10.

11.

12.

13.

14.

15. 16.

17.

history of carpal tunnel syndrome. J HAND SURG 1988;13B:28-34. Mackinnon SE, Dellon AL. Evaluation of microsurgical internal neurolysis in a primate median nerve model of chronic nerve compression. J HAND SURG 1988; 13A:357-63. Louis OS, Greene TL, Noellert RC. Complications of carpal tunnel surgery. J Neurosurg 1985;62:352-6. LearmonthJ. The principleof decompression in the treatment of certain diseases of peripheral nerves. Surg Clin North Am 1933;13:905-13. North ER. Endoscopic carpal tunnel release. In: Gelberman RH, ed. Operative nerve repair and reconstruction. Philadelphia: JB Lippincott, 1991:913-20. Berry DA, LindgrenBW. Statistics: theory and methods. Pacific Grove, California: Brooks/Cole, 1990:541-3, 582-4, 536-9. Kaplan EL, Meier P. Nonparametric estimationfrom incomplete observation. J Am Statist Assoc 1958;53:45781. Prentice RL. Linear rank tests with right censored data. Biometrika 1978;65:167. Gelberman RH, Pheffer GB, Galbraith RT, Szabo RM, Rydervick B, Dimick M. Results of treatment of severe carpal tunnel syndrome withoutinternal neurolysis of the median nerve. J Bone Joint Surg 1987;69A:896-903. Mackinnon SE, McCabe S, Murray JF, et al. Internal neurolysis fails to improve the results of primary carpal tunnel decompression. J HAND SURG 1991;16A:372-4. Okutsu I, NinomiyaS, TakatoriY, OgawaY. Endoscopic management of carpal tunnel syndrome. Arthroscopy 1989;5:11-8. Chow JC. Endoscopic release of the carpal ligament: a new technique for carpal tunnel syndrome. Arthroscopy 1989;5:19-24. Chow JC. Endoscopic release of the carpal ligament for carpal tunnel syndrome: 22-month clinical result. Arthroscopy 1990;6:266-96. Okutsu 1, Ninomiya S, Hamanaka I, Kuroshima N, Inanami H. Measurement of pressure in the carpal canal before and after endoscopicmanagementof carpal tunnel syndrome. J Bone Joint Surg 1989;71A:679-83. Phalen GS. Reflections on 21 years' experience with carpal tunnel syndrome. JAMA 1970;212:1365-7. Gellman H, Kan 0, Gee V, Kuschner SH, Botte MJ. Analysis of pinch and grip strength after carpal tunnel release. J HAND SURG 1989;14A:863-4. Kuschner SH, Brien WW, Johnson 0, GellmanH. Complications associated with carpal tunnel release. Orthop Rev 1991;20:346-52.