Recurrent thoracic outlet syndrome

Transcription

1 Recurrent thoracic outlet syndrome Richard J. Sanders, MD, Craig E. Haug, MD, and William H. Pearce, MD,* Denver, Colo., and Chicago, Ill. Recurrent symptoms develop in 15% to 20% of patients undergoing either first rib resection or scalenectomy for thoracic outlet syndrome. Over the past 22 years 134 operations for recurrence were performed in 97 patients. Four operations were used: transaxillary first rib resection (26); supradavicular first rib resection with neurolysis (15); scalenectomy with neurolysis (58); and brachial plexus neurolysis (35). Complications included temporary plexus injury (0.7%), temporary phrenic palsy (3.7%), and permanent phrenic palsy (1.4%). The combined primary success rate of all four operations for recurrence was 84% in the first 3 months. This fell to 59% at 1 to 2 years; 50% at 3 to 5 years; and 41% at 10 to 15 years. No significant difference was found in results between the four operations used for recurrence. When recurrence was caused by trauma the results of reoperatious were better than when recurrence was spontaneous. The primary success rates of three initial operations for thoracic outlet syndrome were compared to their secondary success rates (improved after reoperation). By use of life-table methods, reoperation improved the 5- to 10-year success rate of transaxiuary first rib resection from 69% to 86% and for scalenectomy from 69% to 84%. Reoperation is successful in most cases of recurrent thoracic outlet syndrome and better when recurrence is the result of a neck injury. (J VASC SURG 1990;12: ) First rib resection or scalenectomy improves symptoms in more than 70% of patients with thoracic outlet syndrome (TOS), whereas 15% to 20% experience recurrence. 1-4 The recurrent symptoms develop at varying time intervals after the primary operation and frequently mimic the initial presenting symptoms. These recurrent symptoms have been treated in a variety of ways. In patients who have undergone first rib resection, scalenectomy from a supraclavicular approach is advocated. Likewise, if the patient initially underwent a scalenectomy, first rib resection is performed for the recurring symptoms. The purpose of this article is to examine in detail the operative indications and outcomes of 134 operations performed for recurrent TOS over a 22- year period. The effect of reoperation on the success rates of primary operations for TOS is also examined. The statistical concept of primary and secondary success rates, similar to primary and secondary patency rates for vascular grafts, is introduced. From the Department of Surgery, Rose Medical Center and the University of Colorado Health Sciences Center, Denver, and Northwestern University Medical School, Chicago. ~ Supported in part by a grant from the Rose Foundation, Denver, Colo. Presented at the Fifth Annual Meeting of the Western Vascular Society, Coronado, Calif., Jan 25-28, 1990, Reprint requests: Richard J. Sanders, MD, 4545 E. 9th Ave., Denver, CO /6/ MATERIAL AND METHODS Between 1966 and 1987, 134 operations fgr recurrent TOS were performed in 97 patients; 18 patents had bilateral procedures. We had performed the initial operation in 76 cases, and in 58 instances someone else had performed the previous surgery. In 11 patients a total of 19 additional operations were carried out for repeat recurrence on the same side. Three operations were performed for persistent symptoms after transaxillary first rib resection. Ten operations were carried out for persistent hand and arm symptoms in patients who had improved head and neck symptoms after scalenectomy. The other 121 procedures were performed for recurrence after an interval of improvement. The patients' ages ranged from 18 years to 55 years with a mean of 32 years. There were 77 women (79%) and 20 men (21%). Fifty percent of recurrences developed within 6 months and 80% within 2 years of the primary operation. No difference was found in the time of recurrence after either scalenectomy or rib resection. The time between the initial operation and the reoperation ranged from 1 month to 20 years, with 50% occurring within 24 months of the first procedure. The duration of recurrent symptoms before reoperation was 1 to 45 months with a median of 11 months. Symptoms. The symptoms were often similar to those that preceded the initial operation, with slight

2 Volume 12 Number 4 October 1990 Recurrent thoracic outlet syndrome 391 Table I. Indications for different reoperations for recurrent TOS First or previous operation(s) Reoperation No. reoperations Scalenectorny TransaxiUary rib resection 26 Supraclavicular rib resection 15 Transaxillary first rib resection Scalenectomy and first rib resection Scalenectomy with neuolysis Brachial plexus neurolysis via supraclavicular route Neurolysis of C-8 and T via transaxillary route Total 134 variations. The most common symptoms were paresthesia in the hand (82%) and pain in the neck (73%), shoulder (71%), and arm (62%). The par-,'sthesia most often involved all five fingers, but fiequently the fom~ and fifth fingers were worse or involved alone. Occipital headaches, weakness of the arm, and aggravation of the symptoms when elevating the arms were other common complaints. Occasional patients had chest pain. Physical examination. Supraclavicular tenderness over the scalene muscles or above the scar where scalenectomy had been performed was present in 81% of the patients. Pressure in this area usually caused paresthesia or pain in the arm, elbow, or hand. Abducting the arms to 90 degrees in external rotation (90 degree AER position) reproduced the patient's symptoms in 84%. Loss of the radial pulse in the 90 degree AER position occurred in only 12%. Other common findings were reduced range of motion of the neck when rotating the neck from one side to the other and causing pain on the symptomatic side qf the neck by tilting the head to the opposite side. Other areas were evaluated for associated or differential diagnoses. These included the shoulder joint, the biceps/rotator cuff tendons, the cervical and dorsal spine, and the wrist area for carpal runnel compression. DIAGNOSTIC TESTS Scalene area block. In 86 patients a diagnostic test was carried out by the infiltration of 1% lidocaine hydrochloride (Xylocaine) into the anterior scalene muscle or the scar tissue in the anterior scalene area in patients ha whom scalenectomy had already been performed. A positive response was improvement in the range of motion of the neck and fewer symptoms with the arms in the 90 degree AER position. Radiography. Cervical spine films were obtained to detect cervical ribs and other bony abnormalities that might have been missed at the initial operation. Radiography also revealed the status of resected first ribs, the existence of a long posterior or anterior stump, signs of first rib regeneration, cervical ribs, cervical arthritis, or suggestions of a cervical disc. Angiography and plethysmography. Indications for arteriography and plethysmography were ischemic symptoms in the fingers, reduced arm pressure, or reduced pulses at rest. Indications for venography were swelling or distended arm veins. These findings were seldom present, so few patients received these studies. Neurophysiologic diagnostic studies. Indications for electromyography and nerve conduction velocities were clinical suspicion of ulnar nerve entrapment at the elbow, median nerve compression at the wrist, or primary nerve injury. SURGICAL PROCEDURES Four different reoperations were performed: scalenectomy; transaxillary first rib resection; supraclavicular first rib resection with brachial plexus neurolysis; and brachial plexus neurolysis alone. The choice of procedure depended on what had been performed at the previous operation as outlined in Table I. If the first operation was rib resection, reoperation was scalenectomy; if the first operation had been scalenectomy, reoperation was first rib resection. Early in the study rib resections were performed through the transaxillary approach. After 1980 most rib resections were performed by supraclavicular approach, so a brachial plexus neuroplasty could be achieved simultaneously. In patients who had already undergone both rib resection and scalenectomy, reoperation was initially a supradavicular brachial plexus neurolysis. If ulnar nerve compression symptoms persisted after supraclavicular neurolysis, transaxillary neurolysis of C-8 and T-1 was carried out at a later time. This was done only once in the course of this study. The techniques of scalenectomy and first rib re-

3 392 Sanders, Haug, and Pearce Journal of VASCULAR SURGERY section have been described previously and are similar to those used when that operation is performed for the first time. ss However, even though most operations for recurrence are performed through different incisions than the first operation, scar tissue is encountered around the brachial plexus and subclavian artery. This makes the reoperation more difficult than the primary operation and may require localized neurolysis by use of some of the techniques described below. Neurolysis was performed alone when both rib resection and scalenectomy had been previously accomplished. It was usually performed through the supraclavicular approach by use of a 7 to 10 cm incision 2 cm above the clavicle. Superior and inferior skin flaps were elevated, and the lateral edge of the sternocleidomastoid muscle was freed and retracted medially. It was not necessary to divide the lateral head of the sternocleidomastoid muscle. The omohyoid muscle was often encountered below the sternocleidomastoid muscle and divided, it was not reapproximated when closing. The scalene fat pad was bluntly separated to expose the lateral portion of the brachial plexus. The internal jugular vein was not sought. If this vein was seen, the dissection had proceeded too far medially. Usually the plexus was covered with scar tissue. Dissection began laterally, where C-5 and C-6 were often fused as the superior trunk of the plexus. A fine curved hemostat and pediatric right-angle hemostat were excellent tools. Scar tissue was gently freed and excised from each nerve. The anterior surfaces of C- 5, C-6, and C-7 were freed first. When previous supradavicular dissections had been done, the phrenic nerve was buried in scar tissue and difficult to identify. in these cases it was assumed that the phrenic nerve lay medially, and no attempt was made to dissect it from the scar tissue for fear of injuring it. As dissection proceeded from lateral to medial, vertical running fibers were tested with a nerve stimulator to see if they were the phrenic nerve. Once C-7 had been freed, the procedure usually did not proceed further medially in the plane anterior to the brachial plexus. After freeing C-5, C-6, and C-7 the dissection moved lateral and deep to C-5, looking for the neck of the first rib, if it was still evident by radiography. This was dissected free and excised. A Raney neurosurgical rongeur, a duck bill rongeur, and an Urschell first rib rongeur were used as needed to remove the posterior renmant of the rib back to its articulation with the transverse process. Safe exposure of the neck of the rib usually required mobilizing the lateral edge of C-5 and C-6 to their full extent so they could gently be pushed medially and not damaged. Next, scar tissue was removed from behind C-5, C-6, and C-7, the area of the middle scalene muscle. Any portion of the muscle that was remaining from previous surgery was also removed. The long thoracic nerve was in this area, deep to C-5, in the unresected portions of the middle scalene muscle. This nerve was usually identified and spared. However, if the long thoracic nerve was not easily identified, additional efforts to find it were not pursued. Scar tissue around C-8 was encountered medial and posterior to C-7. it was removed if it could be done with good vision and without stretching C-5, C-6, or C-7. If these nerves had to be stretched,'~ see C-8, the procedure was terminated without dissecting C-8. When a wide space existed between C- 6 and C-7, this plane was used to free C-8 instead of going lateral to C-5. The wound was closed by approximating the scalene fat pad over the nerves, leaving a suction drain deep to the plexus, and closing the skin with a running subcuticular suture. Statistical methods. Results were graded excellent, ifau symptoms were relieved; good, if the major symptoms were relieved; fair, if a few symtoms were relieved but some major symptoms persisted; and poor, if there was no significant improvement. Since the life-table method was used to express the staffstics, only success or failure is recognized. 9 Excellent and good results were considered successes; fair improvement was classified with poor results in the life table. The log rank test was used to evaluate differences in life-table results between each of the sever~" operations. 1 Chi-square analysis was used to test for statistical significance among other variables. Primary and secondary success. Success rates were defined as primary or secondary in a way similar to that used to define vascular graft patency rates. 1~ Primary success is improvement of symptoms with no further operations. Because this study dealt only with cases of recurrent or persistent TOS, the term primary success refers to the result of each of the 134 reoperations individually. Patients receiving more than one reoperation on a given side had each operation except the last one listed as a failure. If the last procedure was successful, it was the only one so classified. Secondary success rate includes each operative side only once. The results in the 11 patients who received 19 additional reoperations are graded as the

4 Vokune 12 Number 4 October 1990 Recurrent thoracic outlet syndrome 393 8O % Success 60 % ~... TRAUMA N=36 'Q ~-.,.,; 0 40 [] NOTRAUMA N=98... ~... I::] OPEN (C) ) =PRIMARY SUCCESS SOLID ( ) =SECONDARYSUCCESS 30 I I I 1-: Mos 1-2 Yrs 3-5 Yrs 5-10 Yrs Yrs Fig. 1. Success rate after initial reopcration (primary success, open symbols) and subsequent reoperations (secondary success, solid symbols). result of the last operation only: if the last was successful, that patient-side was counted as one success. The time at which the last procedure failed was recorded as the failure time for that patient-side. In Fig. 1, the definitions of primary and secondary success are slightly different because this figure describes the first operation in a group of 668 patients previously reported.~2 In these patients the definition of success using life-table method was good and excellent, as well as fair results. If the fair results had been classified as failures, success rates at each terval would be reduced about 10%. Seventy six patients (12%) in this group received reoperations. The term secondary success is applied to the improvement rate of the entire group as a result of reoperation. RESULTS Success rates. Initial improvement after all 134 reoperations was 84%. This diminished to 59% at 1 to 2 years and to 50% at 3 to 5 years. Even after 5 years recurrence was seen, further reducing the success rate to 41% at 10 to 15 years (Table II). No statistically significant difference was found between the results of the four operations (p = >0.05 for all comparisons). The life table for all 134 cases is presented in Table III. The three patients who had persistent rather than recurrent symptoms after their initial first rib resec- tions did poorly after reoperation by scalenectomy: two patients had inirnediate failure, and the third recurred 18 months after operation. Ten patients were reoperated on by transaxillary first rib resection because their primary anterior and middle scalenectomy had given them relief of their head and neck symptoms but did not help their hand and arm symptoms. Seven of these 10 patients enjoyed good longterm improvement after first rib resection. Trauma and primary and secondary success. A neck injury, invariably from a whiplash type automobile accident, was the cause of recurrence in 36 of the 134 cases (27%). In these patients the results of reoperation were better than when symptoms recurred spontaneously. This is not apparent from the primary success rate of trauma versus nontrauma cases in Fig. 1 until it is appreciated that haft of the failures in the traumatic group were due to another injury. They were successfiflly operated on again, and improvement is revealed in the secondary success rate of 71% at 1 to 2 years, which remained unaltered for the next 10 years. The difference between primary- and secondary success rates for nontrauma cases was not statistically significant (p --- >0.05), whereas the difference between the two for trauma cases was significant for the first year (p = <0.05). Statistical significance was not noted after the first year (p = >0.05), perhaps because the size of the samples became so small.

5 Journal of VASCULAR 394 Sanders, Haug, and Pearce SURGERY Table II. Success rate of each reoperative proccdure No. 10 to 15 Operation patients 1 to 3 mo I to 2 yr 3 to 5 yr 5 to 10 yr yr Transaxillary first rib re % (26)* 64% (13) 50% (11) 42%(8) 42% (4) section Supraclavicular rib resection 15 93% (15) 77% (10) 34% (3) 34% (3) & Neuolysis Scalenectomy and neuro % (58) 55% (33) 53% (25) 41% (22) 36% (14) lysis Brachial plexus neurolysis 35 80% (35) 52% (23) 48% (15) 43% (12) 43% (7) Total of all reoperations % (134) 59% (79) 50% (54) 42% (45) 41% (25) *Numbers in parentheses are cases at risk in that time interval. Table III. Results of 134 reoperations Not, No. followed to Good Time operations end and Percent failed in Cumulative (mo) at risk* period~ excellent Fair Poor period~: success (%) /134 (15.7%) 91% /108 (6.5%) 79% /94 (8.5%) 72% /85 (3.5%) 70% /79 (15%) 59% /64 (14%) 51% /50 (2%) 50% /40 (15%) 42% /18 (5.6%) 41% No. failed Life-table formula: Failure in an interval = No. at beginning - 1/2 withdrawals *Number of operations at risk = number of operations at beginning of time interval. tnot followed to end period = not followed to the end of the time interval; patient was either lost or operated on too recently for that time period. ~:Percent failed in period = combined fair and poor results are failures. Gttmulative success percent = cumulative success rate. The secondary success rate for trauma cases was significantly better than the primary success rate for nontrauma cases (p = <0.05). Effect of reoperation on the initial operation. The influence of reoperation on the results of three primary operations in 668 patients is shown in Fig. 2. Seventy-six patients underwent reoperation. With use of life-table methods, reoperation improved the 5- to 10-year success rate of transaxillary first rib resection from 69% to 86% and for scalenectomy (anterior and middle) from 69% to 84% (no statistically significant difference, p = >0.1). The difference between primary and secondary success rates for each of these two operations was statistically significant for all time intervals beyond 6 months (p = <0.05 to 2 years and <0.01 after 2 years). Also statistically significant was the difference between the secondary success rate of rib resection or scalenectomy compared to the primary success rate of the combined operation (p = <0.05 for all time inv_ vals after 9 months). Only 6% of the patients who initially received the combined operation were reoperated on (solid diamond in Fig. 1), so there is not much possibility for significant secondary success in this group (p = >0.05 at all time intervals). Supraclavicular first rib resection and scalenectomy had a 72% 5- to 10-year success rate that increased to 76% with reoperation. This increase is not comparable to that of the other two operations, because only 6% of the failures from the combined operation were reoperated on compared to 19% and 14%, respectively, for transaxillary rib resection and anterior and middle scalenectomy. Scalene block. A scalene area block was performed in 86 patients, 79 of whom had improved neck motion or fewer symptoms in the 90 degree AER position. Seventy-five of these 79 (95%) also had initial improvement after surgery.

6 Volume 12 Number 4 October 1990 Recurrent thoracic outlet syndrome 395 The indication for the block was clinical suspicion of recurrent TOS. The test was used to support the diagnosis of TOS in patients with borderline clinical pictures, because an earlier study had demonstrated good correlation between the response to the block and the results of surgery? Improvement after the block was used as one more factor favoring the diagnosis, whereas a poor response to the block made the diagnosis less likely. Patients with poor responses to the block were operated on only if their history and physical examination were typical of recurrent TOS. Seven patients in this category were operated on, and six had good responses to surgery, indicating that a negative response to the block did not eliminate the diagnosis of TOS when other clinical features were present. Complications. Nerve injuries to either the brachial plexus or the phrenic nerve were the most serious complications in this series. There were five (3.7%) temporary and two (1.4%) permanent phrenic nerve injuries. Temporary brachial plexus palsy occurred in one patient (0.7%). Permanent brachial plexus palsy did not occur. One patient awoke with a totally paralyzed arm. This followed what was thought to be a routine supraclavicular neuroplasty and first rib resection. This patient had complete recovery of both sensory and motor function over a 3- month period without residual effects. One of the first patients reoperated on suffered a recurrent laryngeal nerve palsy because the dissection entered a plane near the sternohyoid muscle, medial to the internal jugular vein, a plane now recognized as being too far medial. This patient had only partial recovery and has permanent hoarseness. Upper plexus versus lower plexus. On the basis of the location of hand paraesthesia, TOS patients can be divided into those with numbness in the fourth and fifth fingers, ulnar nerve distribution; numbness in the first to third fingers, median nerve distribution; numbness in all fingers, combined ulnar and median nerve involvement; and those with no complaints of paraesthesia. The highest success rate, 73%, was in patients with ulnar nerve symptoms, compared to success rates of 59%, 62%, and 58% for the other three groups. However, the differences between these groups were not statistically significant (p = >0.05). Results for transaxillary rib resection and for scalenectomy with neurolysis were almost identical in all four groups with no statistically signiiicant difference between them. Supraclavicular rib resection had a little higher success rate than the other procedures, but it also had the shortest follow-up period. DISCUSSION The results of surgery for recurrent TOS were quite similar to the results for primary TOS. Eightyfour percent showed initial improvement, which fell to 59% at 1 to 2 years. Beyond 2 years recurrences still occurred but less frequently. The results of reoperation were influenced by the cause of recurrence. Spontaneous recurrence is probably the result of progressive scarring in the area of previous surgery. In this situation, reoperation was less successful than when recurrence was elicited by an injury. Twentyseven percent of the reoperations in this study were in patients whose initial symptoms had been relieved surgically until they had a neck injury that precipitated recurrence. In these people the results of reoperations, sometimes more than one, were 71% successful (Fig. 2). Many surgeons have abandoned anterior scalenotomy because of high failure rates. 13'14 However, a previous study revealed that when anterior and middie scalenectomy replaced anterior scalen0tomy, the long-term success rates of first rib resection alone and scalenectomy alone were identical? z Recently another comparative study by Cikrit et al.15 noted that the results of scalenectomy were actually better and the complications fewer than those of transaxillary first rib resection. The present study indicates that when failure occurs the results of reoperation after either procedure are similar. Etiology. The precise cause of recurrence is unknown. It is probably due to the normal physiologic healing process which deposits scar tissue in any surgical area. Healing of the first TOS operation always leaves some scar tissue around the plexus. In 15 to 20% of patients, this scar tissue contracts in such a way as to reproduce the preoperative or similar symptoms. When the primary operation is transaxillary first rib resection, a posterior stump longer than 2 cm is frequently noticed. However, long posterior stumps are also seen in patients enjoying good results from the same operation, and very short stumps are seen in other patients with recurrence. The posterior stump may be one of several contributing factors to recurrence, but it is not necessarily the primary one. Its role, as well as that of a complete, unremoved rib, may be to act as a fixed, rigid anchoring point to which scar tissue can attach. Immediate failure versus recurrence. Immediate failure should be differentiated from recurrence. Patients who initially experience no improvement of any symptoms after combined scalenectomy and rib resection represent errors in diagnosis or another diagnosis dominating the picture. In these patients re-

7 396 Sanders, Haug, and Pearce Journal of VASCULAR SURGERY """"::: ,...,.... % of Success TRANSAXILLARY RIB RES. N=111 SCALENECTOMY N=279 OPEN (C) ) =PRIMARYSUCCESS ~1~ SUPRA CLAV. RIB RES. & SCAL N=278 SOLID ( O ) =SECONDARY SUCCESS 40 I -- I Mos 1-2 Yrs 3-5 Yrs 5-10 Yrs Yrs Fig. 2. Influence of reoperation on success rate of primary operation. operation is usually unsuccessful. On the other hand, patients who have symptomaric improvement initially and then develop recurrent symptoms can benefit from reoperation. Immediate failure after transaxillary rib resection can be due to a wrong diagnosis or scalene muscle anomalies, which cannot be released through the axilia. In this study, two of the three reoperated patients in this category were also immediate failures of reoperation. However, Qvarfordt et al. 13 noted good results from reoperation in all seven of their patients with persistent symptoms. Patients experiencing no immediate improvement after scalenectomy should be considered candidates for first rib resection. In addition, rib resection should also be offered to patients noting relief of head and neck symptoms but persistance of hand symptoms after scalenectomy. In these people, first rib resection at a later date has been quite successful, as noted in the present study. Diagnosis. The diagnosis of recurrent TOS is often easier to establish than is the initial diagnosis of TOS. Improvement of symptoms after the initial operation is one of the better criteria for confirming that the original diagnosis of TOS was correct. If the symptoms recur months later, the diagnosis of recun'ent TOS is usually correct. Differential versus associated diagnosis. Differential diagnoses include primary plexus injury, car- pal tunnel syndrome, cervical arthritis, cervical spine strain, and biceps/rotator cuff tendinitis. If any of these conditions are present, they should be treated first. The presence of another diagnosis does not exclude the diagnosis of recurrent TOS. Many patients have more than one diagnosis, which makes the diagnosis of TOS more difficult. Primary plexus injury is sometimes a diagnosis made by exclusion when there is no neurologic improvement after surgery for TOS, and all other areas of nerve compression (wrist, elbow, spine) have been ruled out, Treatment strategy. Initially, treatment of recurrent TOS, like primary TOS, should be conservative. This includes the various modalities of physical therapy and a home exercise program. Patients whose recurrence is due to right scar tissue around the plexus usually do not respond to conservative therapy. However, if some of the symptoms were due to associated diagnoses, improvement can occur. Patients who do not experience much improvement after a few months of conservative therapy, and in whom symptoms continue to be disabling, are considered candidates for reoperation. No single surgical approach for recurrence has proved to be superior to any other. The most important criterion in selecting an approach is whether the first rib was removed previously. If not, the operation for recurrence is to remove it through the supraclavicular or transaxillary routes. We have not

8 Volume 12 Number 4 October 1990 Recurrent thoracic outlet syndrome 397 used the posterior approach because the postopera- 'rive pain from dividing the parascapular muscles ~s greater than the pain from other approaches. Sessions ~6 used the posterior approach in four of his 60 cases of recurrent TOS, and all four failed; all improved after reoperation through other routes. After performing first rib resection as a reopera.tion through both transaxillary and supraclavicular approaches, we tend to select the transaxillary route if the symptoms are limited to the lower plexus. Also, if the indication for reoperation was improvement of head and neck symptoms but persistent hand and arm symptoms after scalenectomy, the transaxillary approach is appropriate. For upper plexus or diffuse plexus involvement we have used the supraclavicular a; oroach first, and in all but a few cases this has been Success~.ll. Upper versus lower plexus. Roos 17 has emphasized the importance of separating patients into those veith upper plexus, lower plexus, or entire plexus involvcment. Selecting the operative approach on the basis of the symptom complex is an appealing idea: transaxillary for ulnar nerve symptoms, supraclavicular for median nerve symptoms, and combining both approaches in one operation when there is diffuse involvement. However, the data in this study revealed that the success rate for scalenectomy for recurrence with ulnar nerve symptoms was equal to or better than the success rate for recurrence with mcdian nerve or combined symptoms. This supports the selection of scalenectomy with neurolysis, without transaxillary exploration, as the first operative choice when reoperation is to be performed for re- ~rrence after transaxillary rib resection regardless of the symptom complex. Middle scalenectomy. Middle scalenectomy should routinely be performed when doing supraclavicular exploration for recurrent TOS. Congenital ligaments or bands of scar tissue can lie hidden here within the muscle belly and are detectable only by removing the middle scalency Most patients with recurrent TOS have symptoms of neck pain and headache that are probably muscular symptoms from anterior or middle scalene muscle spasm and tightness rather than neurologic symptoms from plexus comprcssion. For this reason, whenever supraclavicular exploration is done, the middle scalene muscle is excised along with the anterior scalene. This decompresses C-7 and C-8 and may be the reason scalenectomy was as successful as rib resection in patients with lower plexus symptoms. This may also explain why reexploration of the lower plexus through the axilla was seldom needed in this study. Complications. The nature of first rib resection, by either transaxillary or supraclavicular routes, requires some retraction, pushing, or pulling on the nerves of the brachial plexus. There is no way to technically remove the rib without exemng some pressure on these nerves. Even in the hands of experienced surgeons, efforts to protect nerves can injure them. Although most injuries are temporary, a few (about 1%) arc permanent. Staging. Combining the data from this study on recurrent operations with data of earlier studies on primary operations suggests a better way to improve long-term results: 14 staging the operative treatment. The results of combining scalencctomy with first rib resection at the time of the primary operation have not been better than the results of each of these procedures alone. By performing these operations in two stages, the secondary success of either rib resection followed by scalenectomy or scalenectomy followed by rib resection yielded about a 15% higher success rate than the combined operation (statistically significant, p = <0.01). Therefore a reasonable plan of treatment when surgery is indicated for neurologic TOS (as differentiated from arterial or venous TOS) is to perform only one of these procedures at a time. The second procedure is reserved for recurrence. Our preference, which has been suggested by Moore, 19 is to perform anterior and middle scalenectomy as the primary procedure because its complications are fewer and the postoperative recovery is smoother and faster as compared to rib resection (through any route). If there is persistence or recurrence of symptoms, either txansaxillary or supraclavicular rib resection is offered. Prevention. Prevention of postoperative scarring after neurolysis is a problem that has yet to be solved. Covering the nerves with a layer of adipose tissue is a standard neurosurgical maneuver, but its value is unproved. In supraclavicular operations the scalenc fat pad is always replaced on top of the nerves at the conclusion of the procedure, but this has not prevented recurrence. At reoperation, the fat pad is usually found adherent to the nerves. The instillation of local steroids has been used in several patients, but this too has not prevented postoperative scarring. Recently, polytetrafluoroethylene has been used to cover the plexus to prevent adjacent tissues from attaching to it, although this does not prevent the nerves from sticking to each other. The long-term results are still tmknown) 6

9 398 Sanders, Haug, and Pearce journal of VASCULAR SURGERY The authors express their thanks to Darryl Jones, PhD, for assistance with statistical analyses. REFERENCES 1. Sanders RJ, Monsour JW, Gerber FG, Adams WRA, Thompson N. Scalenectomy versus first rib resection for treatment of the thoracic outlet syndrome. Surgery 1979;85: Wiley E. Discussion in Roos DB. The place for scalenectomy and first rib resection in thoracic outlet syndrome. Surgery 1982;92: Qvarfordt PG, Ehrenfeld WK, Stoney RJ. Supraclavicular radical scalenectomy and transaxillary first rib resection for the thoracic outlet syndrome. A combined approach. Am J Surg 1984;148: Martinez DB. Transaxillary first rib resection for thoracic outlet syndrome. (Submitted for publication.) 5. Sanders RJ, Monsour JW, Gerber WJ. Recurrent thoracic outlet syndrome following first rib resection. Vasc Surg 1979;13: Roos DB, Owens JC. Thoracic outlet syndrome. Arch Surg 1966;93: Sanders RJ, Monsour We, Baer SB. Transaxillary first rib resection for the thoracic outlet syndrome. Arch Surg 1968;97: Sanders RJ, Raymer S. The supraclavicular approach to scalenectomy and first rib resection: description of technique. J VASC SURG 1985;2: Colton T. Statistics in medicine. Boston: Little Brown & Co, 1975: Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized trials requiring prolonged observations of each patient. II. Analysis and examples. Br J Cancer 1977;35: Bandyk DF, Kaebnick HW, Stewart GW, Towne JB. Durability of the in situ saphenous vein arterial bypass: a comparison of primary and secondary patency. J VAsc SlntG 1987;5: Sanders RJ, Pearce WH. The treatment of thoracic oufle~ ~ syndrome: a comparison of different operations. J VASC SURG 1989;10: Raaf J. Surgery for cervical rib and scalenus anticus syn. JAMA 1955;157: Clagett OT. Presidential address: research and prosearch. J Thorac Cardiovasc Surg 1962;44: Cikrit DF, Haefner R, Nichols WK, Silver D. Transaxillary~ or supraclavicular decompression for the thoracic outlet syndrome: a comparison of the risks and benefits. Am Sur!~ -~ 1989;55: Sessions RT. Reoperation for thoracic outlet syndro~. J Cardiovasc Surg 1989;30: Roos DB. Recurrent thoracic outlet syndrome after first rib resection. Int Angiol 1984;3: Sadler TR Jr. Discussion in Urschel et al. Reoperation foi recurrent thoracic outlet syndrome. Ann Thorac Surg 1976~ 21: Moore WS. Discussion in Qvarfordt PG, Ehrenfeld WK, Stoney RJ. Supraclavicular radical scalenectomy and transaxillary first rib resection for the thoracic outlet syndrome. A combined approach. Am J Surg 1984;148:115. DISCUSSION Dr. David B. Roos (Denver, Colo.). First, it is difficult for me to evaluate this paper because the authors included patients with "fair" results with the "good to excellent" groups as successes. Since the patients would not consider an outcome that left them with their major symptoms unrelieved, as the authors state in their "fair" group, as being successful operations, most surgeons would include these "fair" results in the "poor" category. This shift would substantially change the successful and unsuccessful groups in the life-table analysis. As we are not given an adequate breakdown of these results in the authors' cases by which we can evaluate this variation for ourselves, I am unable to accept their conclusions that the results of all surgical procedures are statistically equal. Second, my own experience indicates substantial differences in the recurrence rate after scalenectomy versus first rib resection as the primary procedure. In the recent 5-year period, 19% of my original scalenectomy patients had severe recurrent symptoms compared to only 5% of my transaxillary first rib patients. These results are corroborated by Dr. Green from the University of Rochester and presented to the Society for Clinical Vascular Surgery last year. He found the recurrence rate after neck surgery for TOS was five times greater than recurrence from transaxillary rib operation. Although scalenectomy has its place when the syn~& ~ toms and physical findings are confined to the C-5, C-6, and C-7 nerves of the brachial plexus, the inordinate rate of recurrent scar tissue entrapment of these nerves after scalenectomy dictates that we should avoid the neck approach unless there are compelling reasons to explore these nerves. To use scalenectomy as a routine primary procedure for all cases of TOS seems to be ill advised. I think the surgical approach must be tailored to release the specific nerves involved: scalenectomy for the upper plexus, C-5, C-6, C-7 neuropathies and axillary first rib resection for lower plexus, C-8, T-1 ulnar nerve symptoms. Third, regarding lidocaine infiltration of the anterior scalene muscle as a diagnostic test for TOS, I think it is unnecessary and somewhat hazardous for three reasons. First, it gives no specific information that we cannot obtain by simpler clinical evaluation of the plexus, such as palpation, stretch test, and using the 3-minute elevated arm

10 Volume 12 Number 4 October 1990 Recurrent thoracic outlet syndrome 399 stress test combined with a careful history. Second, patients who have had previous neck surgery with scar formation qver the plexus may have poor absorption of the lidocaine because of a fibrotic barrier around the plexus, which could Cause false-negative results. Third, all patients who have distinct upper plexus symptoms, either primarily or after first rib resection, have significant congenital anomalies in that region often involving the brachial plexus itself. The C-5 nerve and sometimes the upper trunk of the plexus ~flay be congenitally displaced on the anterior surface of the anterior scalene muscle just lateral to the phrenic nerve in quite an abnormal location. I believe this test is not necessary, reliable, or specific, so I do not use it. Six of the seven patients in the authors' series who had a negative ~idocaine block test underwent operation anyway based on other more significant findings, and did well. This indicates the false-negative results of the test and the greater validity,of other clinical evaluations. in conclusion, I find it difficaflt to evaluate this paper on a statistical basis because of the questionable assignment of cases with persistent major symptoms into the successfial group. This makes the authors' main conclusion suspect, ~_~amely that all procedures to release thoracic outlet entrapment had equal results. Scalenectomy in most hands has a much higher rate of recurrence and therefore cannot be justified as a preferred primary operation for all cases of TOS. Also the authors' recommendation of routinely performing the opposite procedure for recurrence without regard to the specific nerves 'involved is too simplistic as well as anatomically and dinically unsound. It exposes the unaffected nerves of the plexus to the risk of later scar tissue entrapment in an area previously uninvolved. The use Of lidocaine block in the scalene muscle, first described in the i930s, has gained little acceptance over,,the past 50 years because it is so nonspecific. Dr. Richard Sanders. Dr. Roos raises some interest- 7-points. The results expressed here use the life-table method by which only success or failure can be recorded; there is no allowance for partial improvement. This method has the advantage of giving results at various time intervals, beyond 10 years. Dr. Roos's results do not provide this "time factor, which for TOS is quite important. The number of fair results was 4% to 8% for each operation. Were we to count only the good and excellent results as success, there still would be no significant difference between the results of rib resection and anterior and middle scalenectomy. When Dr. Roos talks about scalenectomy as a primary operation, he is referring to anterior scalenectomy only. It "must be stressed that the scalenectomies presented in this study were both anterior and middle. Our data for both primary and recurrent TOS operations indicate that the results for upper and lower plexus involvement are the same for anterior and middle scalenectomy and first rib resection. There is no evidence to sustain the view of Dr. Roos (unsupported by any data) that anterior and middle scalenectomy should be reserved for upper plexus symptoms only. The scalene block, like all diagnostic signs and tests for TOS, is not infallible. It is based on the hypothesis that the underlying abnormalities in most cases of TOS is scarring of the scalene muscles, a theory that is supported by the clinical finding of tenderness over this area in more than 90% of TOS patients. Having used the block in several hundred patients, we have found it to be one more piece of evidence that is helpfixl when positive. We have not seen any patient made worse by its use. Dr. Ronald P. Stoney (San Francisco, Calif.). I seem to have increasing experience with not only primary TOS operations but also referred patients with recurrences. I did not find a good description of the brachial plexus recurrent abnormality that induced the symptoms which led to the reoperation in the authors' manuscript. There may be some reasons for this. The route that was used may not have offered an adequate exposure to the plexus in each case, or there might have been findings identified that were thought to be the primary reason for the recurrence and other findings in the region of the plexus may not have been methodically sought out. I reviewed my own experience of 150 supradavicular operations, 25 were for recurrence. I noticed that there were two types of operations that led to reoperations. Transaxillary procedures had an incompletely removed first rib, or an inadequately removed first rib, or the second rib was taken out by mistake. The scar tissue appears between the first rib bed and the plexus and can cause a reattachment of the scalene muscles. After supraclavicular thoracic outlet operations persistent or inadequately removed rib and reattachment of the anterior scalene muscle were the main cause of brachial plexus irritation. As Dr. Sanders just said, the middle scalene is now an important feature of the TOS because it can be reattached with scar tissue and encroach or compress the plexus. I think the technique of the reoperation is very important, and I have slight differences with the author. I think mobilization of the scalene fat pad on a laterally based flap should be done so you can split this into two parts and use that to surround the brachial plexus after neurolysis. I do this in all primary operations, and I have been able to do it in secondary and even tertiary reoperations. I think it helps to minimize scarring around the nerves. Also, I do a neurolysis on all five roots of the plexus from the neural foramina to the clavicle each me including secondary and tertiary operations, without brachial plexus injury. The middle scalene muscle should be excised from a line parallel and inferior to the long thoracic nerve down to the first rib bed. Finally, at the suggestion of Dr. Roos I have put in a fine catheter into the wound and leave that in for the four to six days the patient remains in the hospital. Steroids are injected in the expectation that this may cut

11 400 Sanders, Haug, and Pearce Journal of VASCULAR SURGE]~,Y down on subsequent scar tissue formation. I have three questions for Dr. Sanders. Please explain why the reoperation rate for supraclavicular procedures was only 24%, whereas for transaxillary operations it was 66%, and for scalenotomy or scalenectomy patients it was 54%. Second, what is the disadvantage of using a supraclavicular approach for all recurrences? It is an effective approach for a primary operations, and I believe all mechanisms known to cause recurrence can always be reached from the neck. Finally, how long does a reopera~ion take, the axillary reoperation and the neck reoperation? Dr. Richard J. Sanders (closing). The abnormality in all cases of recurrent TOS was scar tissue around the brachial plexus. In those cases in which transaxillary rib resection was the first operation, the scalene muscles were uniformly found bound down to the top of the plexus and the subclavian artery. The figures you quoted for reoperation rate must have been misunderstood. Reoperation for combined scalenectomy and rib resection was only 6%, probably because there was not much to do at reoperation except remove scar tissue. The reoperation rate for the other two primary operations was 19% for transaxillary rib resection and 14% for anterior and middle scalenectomy. Regarding the exclusive use of the supraclavicular approach for all reoperations, I agree with Dr. Stoney; today, this is my standard practice. However, on reviewing thedata in this paper, it was noted that when anterior and middle scalenectomy was the first operation, there was no significant difference between the results of transaxillary and supradavicular rib resection as the reoperation. Finally, the time for transaxiuary rib resection is usually under 1 hour if the first operation was supraclavicular. It takes twice as long if the transaxillary route was used before. Supraclavicular reoperarions take 1 to 2 hours. LIEBIG FOUNDATION AWARD FOR VASCULAR SURGICAL RESEARCH, 1991 The Liebig Foundation announces the tenth annual competitive award of $5000 for the best essay on a problem in general vascular surgery. The investigative work shall be: 1. Clinical or experimental research 2. Original and unpublished (nor submitted elsewhere for publication) 3. Performed by a house officer in the United States, Canada, or Mexico with senior collaborators acting in a consulfive capacity 4. Submitted in English (6 copies of typed manuscript and 6 copies of glossy prints of illustrations) 5. Accompanied by a signed letter from the essayist's superior confirming the status of the essayist and complying with "Instructions to Authors" of the Jouv~a~ Previous winners were: 1990, Scott A. Berceli, MD, University of Pittsburgh School of Medicine, Pittsburgh, Pa.; 1989, Michael L. Marin, MD, College of Physicians and Surgeons, Columbia University, N.Y. The winning essay will be judged for presentation at the June 1991 meeting of the North American Chapter, International Society for Cardiovascular Surgery and will be considered for publication in the JOUI~AL. Further inquiries may be directed to the same address to which the essays must be sent, postmarked before Dec. 31, 1990: Dr. Richard Turner Award Committee Secretary 112 Bauer Drive Oakland, NJ USA (201)