MICROSURGICAL TREATMENT OF LUMBAR ADHESIVE ARACHNOIDITIS

(A Patient Information Service)

We have developed considerable experience as a result of evaluating many patients with symptomatic lumbar adhesive arachnoiditis who have been referred to us following a previous surgical procedure for disc disease and/or decompression for lumbar spinal stenosis. Over fifty (50) of these patients have undergone microsurgical intradural neurolysis and insertion of a dural decompressive patch graft since 1973. The indications for investigation and surgical treatment have been significant, persistent pain and neurological deficit following "disc surgery" or "decompression" particularly if there was a history or evidence of a dural laceration. On occasion, postoperative patients have been referred with severe incapacitating pain as the primary problem. Progressive loss of neurological function, particularly bladder control, has also been a cause for referral and among the indications for surgery.

The operation consists of a wide laminectomy extending to just above the level of arachnoidal involvement. Microsurgical neurolysis of the cauda equina then follows. An attempt is made to free each nerve root to its exit zone. However, in some cases with extensive matting this may be incompletely accomplished. With rare exception, these patients undergo some form of dural patch grafting in an effort to prevent further scarring of the roots and adhesions to the dura. This has been an important feature of the treatment since it allows for the re-establishment of cerebrospinal fluid circulation around the nerve roots and thus reduces the risk for recurrence of scarring.

Females and males have been affected equally. All patients, except one, had undergone previous lumbar operative procedures. Sixty (60%) percent of patients had multiple lumbar spinal operations prior to this operation. Although all of the patients had previous myelograms, 75% of them had two or more myelograms prior to referral. Two (2) patients had a previous chemonucleolysis and six (6) had intrathecal steroids injected. One patient underwent myeloscopy in addition to having suffered a dural laceration and having undergone seven intrathecal steroid injections and myelographic procedures before being referred to us from Canada.

The single patient who did not undergo a previous lumbar spine operation seems to have developed arachnoiditis consequent to four (4) Pantopaque myelograms which were initially done for cervical disc disease prior to referral. Twenty-five percent (25%) of patients were known to have suffered dural laceration at the time of their previous procedures. Four (4) patients (including 2 with dural lacerations) suffered disastrous neurological sequelae prior to referral. All of the patients had severe pain as a major preoperative complaint. At least 25% of patients had bladder and rectal sphincter dysfunctions. Two (2) of these were incontinent. Patients were divided into three categories, according to the extent and severity of arachnoidal involvement and degree of disability. Involvement of more than three vertebral levels constituted the "severely" involved group and included 50% of patients. Involvement of 2 to 3 vertebral levels constituted the "moderately" involved group and included 30% patients. Involvement at one vertebral level constituted the "mild" group and included 20% of patients.

In assessing our results, we have noted that 40% of patients in the "severe" category, 65% of the patients in the "moderate" category and 85% in the "mild" category have had a "good" result. (They have minimal or no pain which is easily controlled with non-narcotic analgesics.) They are either capable of or have returned to work or full regular activities. A "fair" result was characterized as persistent pain which was substantially less than the preoperative condition and controlled by medication. This "fair" result was achieved in 33% of patients in the "severe" group, 35% in the "moderate" group and 15% in the "mild" group.

All patients who suffered preoperative bladder disabilities enjoyed improvement of function except for those who were incontinent preoperatively. The patients in the "severe" group had the most extensive cauda equina involvement and consequently demonstrated the least improvement. Eighty-five percent (85%) of patients in the "severe" group had significant neurological deficits preoperatively. All patients in the "severe" category still had some pain postoperatively requiring medication. One patient's neurological deficit was worse postoperatively. While all the other patients are improved after surgery, the improvement in 27% of the "severe" group is limited and their overall result is considered to be poor. Thirty-three percent (33%) of patients with severe preoperative neurological deficit are improved postoperatively. Thirty percent (30%) of patients still use crutches or a cane to assist ambulation.

The 30% of all patients who were assigned to the "moderate" group were primarily concerned with pain and had mild neurological deficit preoperatively. All are improved postoperatively but with persistent mild or intermittent pain. The neurological deficits improved postoperatively to a nearly normal status. The 20% of all patients placed in the "mild" category all had severe pain preoperatively but only two had neurological deficit. Sixty-five percent (65%) are pain-free postoperatively, while 35% have mild pain.

A "poor" result was noted in 27% of patients in the "severe" group. Although there may have been some improvement in the pain relief and preoperative neurological status, it was minimal. Overall an 88% improvement in the patients' pain and neurological deficits were noted. Of those patients, 58% needed to return to the work force, 42% were able to do so. Another 36% of the patients were either involved in housework or are retired. Maximum improvement was generally noted to have been achieved by one year postoperatively. There have been three cases of worsening status and documented progression of the arachnoiditis on long-term follow-up. One case involved extension of the arachnoiditis above the level of our initial neurolysis procedure while the other was a similar extension of the arachnoiditis below the operative site. One other case experienced recurrence of pain and scarring (at the operative site) within two months of an initial neurolysis. All three patients underwent a repeat operation and have improved.

The major complication was the occurrence of cerebrospinal fluid leakage postoperatively in relation to the dural graft. This occurred in 17% of patients. Repeat operations for closure of the fistula were carried out in each instance. One of these patients developed a postoperative wound infection which responded to appropriate antibiotics. Another patient received a one unit blood transfusion and two months later presented with serum hepatitis.

Inspection of the previous operative area beneath the silastic dural reconstruction graft in two of the patients demonstrated a satisfactory situation characterized by a lack of scarring. A "pseudo-dura" forms beneath the silastic but is separated from the cauda equina by cerebrospinal fluid. Five (5) patients have undergone postoperative myelography. We have been reluctant to carry out this procedure unless there has been a strong neurosurgical indication (such as suspected cerebrospinal fluid fistula) since we would prefer to limit the introduction of any instrumentation or agents which could be related to the etiology of arachnoiditis.

There are several aspects of this difficult clinical problem that have impressed us. Causative factors seem to have been surgical misadventures during lumbar disc operations with either over-zealous retraction of nerve roots or the laceration of dura along with injury to the arachnoid. Additionally, there seemed to be some relationship to the use of oil-based myelographic dyes which are no longer available. We have been very impressed by the extent of scarring associated with the use of intrathecal steroids. The "carrier" substance has been found to be imbedded in very dense scar around matted nerve roots and in several instances was directly involved in an intradural ossification process which incorporated the nerve roots as well. This has led us to strongly caution against the use of intrathecal steroids for any reason.

There have been several cases referred to us where the lumbar nerve roots were found in dense scar outside the dura. Most of these patients had undergone a surgical procedure for a herniated lumbar disc up to five or ten years previously.

Postoperative cerebrospinal fluid fistula indicates an injury to the arachnoid membrane. There is a small risk that this could lead to arachnoiditis at a later date. Injuries to the pia mater (usually by surgical trauma) can lead to the development of arachnoiditis as can rare cases of intradural bacterial infection.

Our present policy when considering potential surgical candidates requires recent myelography with post-myelographic, high-resolution CT scanning as well as urodynamic studies. Most candidates for surgery must have either well documented advancing neurological deficit resulting from the arachnoiditis and/or incapacitating, agonizing pain that has not responded to aggressive reasonable methods of management. We are aware that early intervention may limit the spread of the arachnoiditis. It has been clear to us that when three or more lumbar levels are involved with this process, that the results are not nearly so satisfactory. We now have up to 30 years follow up on patients with the results being reliable and successful in cases where the original result was positive for the relief of symptoms.

1. Lumbar Adhesive Arachnoiditis. Letter to the editor, Neurosurgery 5:771-772, 1979 Lazar, M.L. and Bland, J.E.

2. Treatment of Pain Syndromes. Current Treatment of Neurological Diseases edited by Rosenberg, R., Spectrum Publications, Jamaica, NY, 1979, pp. 526-604. Bland, J.E., Lazar, M.L. and Naarden, A.L.

This page last edited on 2/20