Sixty patients met the criteria for enrollment in the study, as established by physical examination and confirmed with imaging studies and by the symptoms. These criteria consisted of a single intracanalicular disc herniation at the level between the second and third, the third and fourth, or the fourth and fifth lumbar vertebrae or between the fifth lumbar vertebra and the first sacral segment, with associated radiculopathy; a herniation not exceeding one-half of the anteroposterior diameter of the spinal canal; an absence of central or lateral osseous or ligamentous stenosis; accessibility of the disc for both arthroscopic microdiscectomy and laminotomy; failure to respond to nonoperative measures; more pain in the lower extremities than in the back; the presence of positive tension signs with or without an accompanying neurological deficit; a dermatomal distribution of pain in the lower extremities matching that seen on imaging studies and specific nerve-root involvement; no previous operation on the low back; and the absence of any litigation or Workers' Compensation claim involving the disc herniation.
The criteria for exclusion from the study included central or lateral stenosis of the spinal canal; severe degenerative narrowing of the intervertebral disc space at the index level; evidence, on imaging, of global bulging of the intervertebral disc10 associated with central or lateral stenosis; a sequestered herniation that had migrated; a large central or extraligamentous herniation between the fifth lumbar and first sacral vertebrae; drug dependency; and known psychological disorders.
The study comprised sixty patients: thirty who had an open laminotomy and discectomy (Group 1) and thirty who had a video-assisted arthroscopic microdiscectomy (Group 2). The ages of the seventeen male and thirteen female patients in Group 1 ranged from eighteen to sixty-seven years (mean, forty years), whereas those of the twenty-two male and eight female patients in Group 2 ranged from fifteen to sixty-six years (mean, thirty-nine years). Forty-four (73 percent) of the sixty patients were between the ages of twenty-five and fifty years.
The open laminotomy and discectomy (Group 1) was performed between the third and fourth lumbar vertebrae in one patient, between the fourth and fifth lumbar vertebrae in twenty-three patients, and between the fifth lumbar and first sacral vertebrae in six patients (Table I). The arthroscopic microdiscectomy (Group 2) was performed between the second and third lumbar vertebrae in one patient, between the third and fourth lumbar vertebrae in six patients, between the fourth and fifth lumbar vertebrae in nineteen patients, and between the fifth lumbar and first sacral vertebrae in four patients (Table I).
Although spinal stenosis is the most common cause of lumbar radiculopathy in elderly patients, the patients in this study did not have clinical or imaging characteristics of spinal stenosis, and they met the criteria for inclusion mentioned earlier.
All sixty patients had been managed nonoperatively elsewhere and were referred to our institution for consideration of operative management. However, when the nonoperative measures appeared to have been suboptimum, additional nonoperative treatment was employed. Because of the methods that were used, no data are available on the percentage of patients who responded to nonoperative therapy. The minimum duration of nonoperative treatment in both groups was fourteen weeks. None of these patients had worsening of the neurological deficit, which would have necessitated earlier operative intervention. In addition to the nonoperative treatment, nine patients in Group 1 (open laminotomy and discectomy) and twelve patients in Group 2 (arthroscopic microdiscectomy) had received steroids orally, and seven patients in each group had received epidural injections of steroids before being referred to our clinic.
The herniation had penetrated the boundary of the posterior longitudinal ligament in four patients in Group 1 and in five patients in Group 2. The herniated disc fragments were retrieved successfully from all patients. None of these herniations were displaced cephalad or caudad.
The nonoperative measures included short-term partial rest with avoidance of lifting, bending, climbing, or long periods of standing or sitting. All patients received nonsteroidal anti-inflammatory medication for approximately six weeks and participated in a standard physical therapy and exercise program before the operative procedure. Nonoperative management was considered to have failed in patients who had persistent or recurrent radicular symptoms in a specific dermatomal distribution, with pain in the lower extremities that was more disabling than pain in the low back; those who had positive tension signs (a positive straight-leg-raising test with the patient supine, a positive Lasègue sign, and a positive straight-leg-raising test with the patient seated); and those who had unremitting pain and a neurological deficit.
Neurological findings included reflex abnormalities in nine patients in Group 1 (open laminotomy and discectomy) and in twelve in Group 2 (arthroscopic microdiscectomy). These abnormalities included a unilateral absence of the tibialis posterior reflex, observed in three patients who had a disc herniation between the fourth and fifth lumbar vertebrae. In addition, twenty-eight patients in Group 1 and twenty-six in Group 2 had sensory deficits, and twenty-six in Group 1 and twenty-four in Group 2 had motor weakness (Table I).
The imaging studies were reviewed carefully by radiologists and the operating surgeon. An intervertebral disc was considered to be herniated when there was evidence of localized expansion of the external contour of the disc that was effacing the nerve root or the dural sac1,17.
Sixty sealed envelopes containing the words arthroscopic microdiscectomy or laminotomy were prepared and randomly placed in the files of patients who had lumbar radiculopathy that potentially necessitated operative intervention. This was done by the clinic's office personnel at the time of the first interview. If the examining physician thought that the patient was not a candidate for operative treatment, the envelope was removed and was returned to the secretarial area for future use. Before the operation, the envelope was obtained from the patient's chart for the selection of either arthroscopic microdiscectomy or laminotomy and discectomy.
The study was not blinded because of legal restrictions imposed by the patient's right to know and the fact that, postoperatively, the size and site of the skin incision alerted the patient to the nature of the operative procedure. The advantages and disadvantages of arthroscopic microdiscectomy compared with laminotomy and discectomy and with nuclear debulking procedures were carefully reviewed with the patients and their families. Appropriate consent, the wording of which was approved by the institutional review board, was obtained, and the procedure was performed in the operating suite.
Patients who had been referred for a specific operative procedure and those who insisted on having either a laminotomy or an arthroscopic microdiscectomy were excluded from the study. The envelope was then returned to the secretarial area and was placed randomly in a new patient file.
Operative Technique
The open laminotomy and discectomy was performed in a standard fashion. A four-centimeter posterior midline incision was made, and a small laminotomy and discectomy was performed at the specified level. The arthroscopic microdiscectomy was performed with use of an oval five by eight-millimeter (internal diameter) cannula (Figs. 1-A and 1-B), introduced into the triangular working zone as described previously by one of us (P. K.) and colleagues15-17. Under anteroposterior and lateral fluoroscopic control, an 18-gauge needle was introduced in an anteromedial direction from a distance of about eleven centimeters lateral to the midline. The tip of the needle was positioned in the triangular working zone that is bordered anterolaterally by the exiting nerve root, medially by the traversing nerve root and the dura, and caudally by the vertebral plate of the caudad lumbar segment. This was followed by the introduction of a soft-tissue dilator and the positioning of a universal five-millimeter (internal diameter) cannula. The annular surface in the triangular working zone and the contents of the spinal canal were inspected through a 0-degree working channel arthroscope to ensure that the annulotomy was performed adjacent to the spinal canal, just under the posterior longitudinal ligament (Figs. 2 [top left and right] and 3). (The working channel arthroscope is a specially designed endoscope that has a separate channel for introduction of instruments such as forceps, a knife, or a curet. It also provides channels for introduction and retrieval of saline solution during the operation. The whole unit fits inside the cannula.) The conversion of the universal cannula to an oval cannula was then accomplished by the introduction of a half-moon-shaped soft-tissue dilator next to the previously positioned universal soft-tissue dilator. This permitted the smooth passage of the oval cannula into the triangular working zone and the subligamentous region of the intervertebral disc at the index level.
The oval cannula is shaped so that it fits within the boundaries of the triangular working zone between the traversing and exiting nerve roots. This permits the simultaneous insertion of a 0-degree or 30-degree discoscope and an upbiting forceps for visualization and removal of the compressive elements. In contrast to the laminotomy and discectomy, with arthroscopic microdiscectomy the herniated disc fragments are pulled back into the intervertebral disc space and then are withdrawn. The anterior aspect of the dura or the fibers of the posterior longitudinal ligament may be inspected with a 30 or 70-degree arthroscope at the end of the procedure (Fig. 2 [bottom]).
Two patients who had evidence of a large central extraligamentous herniation that had not migrated needed a biportal access for the retrieval of herniated disc fragments.
The arthroscopic microdiscectomy was performed on an outpatient basis, whereas the laminotomy and discectomy necessitated one night of hospitalization. Hospitalized patients received no physical therapy, and they were discharged when they were able to walk without assistance.
Evaluation of the Patients
Follow-up visits were scheduled for both groups at two weeks, three months, six months, one year, and two years postoperatively.
The outcome analysis was based on the patient's self-evaluation before and after the procedure, the preoperative and postoperative findings on physical examination, and the patient's ability to return to a functional status6,31. Preoperatively, the patients completed a two-page questionnaire (a modification of the Rush-Presbyterian-St. Luke's lumbar spine analysis form31) that included information regarding symptoms (the presence or absence of numbness or weakness, the status and duration of disability, the types and dosages of medications and their frequency of use, the patient's ability to work, and diagrams showing the dermatomal distribution of the pain). The patients rated pain with use of the method of Houde6. The postoperative evaluation consisted of physical examination by an independent physician as well as questionnaires that focused on patient satisfaction, pain and use of pain medications, weakness and numbness, disability, the duration until the patient was able to return to work or to normal activity, and the patient's ability to work.
The mean duration of follow-up was thirty-one months (range, nineteen to forty-two months) for the patients in Group 1 (open laminotomy and discectomy) and thirty-two months (range, twenty-one to forty-two months) for the patients in Group 2 (arthroscopic microdiscectomy).
The postoperative management in Group 1 included intravenous patient-controlled administration of morphine sulfate for twenty-four hours. This was followed by oral administration of Percocet (oxycodone hydrochloride and acetaminophen). The mean duration of use of narcotics in this group was twenty-five days (range, seven to fifty-six days). In addition, at the time of the latest follow-up evaluation, six patients (20 percent) reported occasional use of codeine derivatives for control of the low-back pain or the pain in the buttocks, or both.
The patients in Group 2 did not need injectable medications postoperatively, and the mean duration of use of orally administered narcotics was seven days (range, three to fourteen days).
The mean duration of postoperative disability—that is, the time lost from work or until the patient was able to resume normal activity—was forty-nine days in Group 1 (Table II); this group included two retired individuals, two students, and two homemakers who resumed their normal activities. In Group 2, a mean of twenty-seven days was lost; this group included one student and one homemaker who returned to their preoperative daily activities. It also included one patient who had been receiving disability compensation preoperatively because of a reason unrelated to the disc herniation, and this patient remained disabled after the operative procedure. A patient in Group 1 had leakage of spinal fluid, which necessitated exploration and repair of the dural sac two weeks postoperatively. This patient had a long period of postoperative disability with satisfactory closure of the dural leak, but there was residual disability associated with pain in the low back and buttocks. Another Group-1 patient continued to have radicular symptoms and was considered to have had failure of the operative procedure. One patient in Group 2 later needed a two-level laminotomy and partial facetectomy for the treatment of mild lateral stenosis that had not been recognized at the time of the index operation. Postoperative complications such as infection or neurovascular injuries were not observed in either group of patients.
The postoperative findings on physical examination included reflex abnormalities in six patients in Group 1 and in seven patients in Group 2. Sensory deficits were found in eighteen patients in Group 1 and in sixteen patients in Group 2. Motor weakness was detected in ten patients in Group 1 and in five patients in Group 2. All patients who had a satisfactory outcome had negative postoperative tension signs.
The outcome was considered to be excellent if the radicular symptoms had ceased, the tension signs had become negative, the patient had returned to his or her previous occupation or to normal activity, and the patient expressed satisfaction with the result of the operative procedure. The outcome was considered to be good if the criteria just mentioned were met but the patient had residual back pain and had had to modify his or her occupation. The operation was considered to have failed if the patient had persistent radicular symptoms or needed an additional operative procedure. An excellent or good result was considered a satisfactory outcome. On the basis of the patient's preoperative and postoperative self-evaluation, the findings on physical examination, and the patient's ability to return to work or to normal activity, twenty-eight patients (93 percent) in Group 1 (open laminotomy and discectomy) and twenty-nine patients (97 percent) in Group 2 (arthroscopic microdiscectomy) were considered to have a satisfactory outcome.
The postoperative questionnaire included a 4-point satisfaction scale, with 1 point indicating that the patient was very satisfied with the outcome of the operation; 2 points, that the patient was satisfied; 3 points, that the patient was dissatisfied; and 4 points, that the patient was very dissatisfied. Twenty patients (67 percent) in Group 1 and twenty-two patients (73 percent) in Group 2 reported that they were very satisfied with the outcome. The patients also rated pain, with use of a 10-point scale, with 0 points indicating that they were pain-free and 10 points indicating that they had severe and incapacitating pain. The mean postoperative pain score was 1.9 points for Group 1 and 1.2 points for Group II (Table II).
The patient populations in the two groups were similar. In most of the patients, the herniated disc was between the third and fourth or the fourth and fifth lumbar segments. In Group 1 (open laminotomy and discectomy), twenty-four patients had extraction of a disc between the third and fourth or the fourth and fifth lumbar levels, whereas in Group 2 (arthroscopic microdiscectomy) twenty-five patients had the operation at one of these sites.
Although the numbers of patients who had a satisfactory outcome were similar in the two groups, the rate of postoperative morbidity was lower in the patients who had the minimally invasive, video-assisted arthroscopic microdiscectomy than in those who had the open laminotomy and discectomy. All of the arthroscopic procedures were performed on an outpatient basis, whereas the patients who had the laminotomy and discectomy needed at least one day of hospitalization. The postoperative use of narcotics was less and the overall satisfaction score was higher after the arthroscopic microdiscectomies than after the laminotomies and discectomies.
Although both macrodiscectomy and microdiscectomy are simple and acceptable methods for the treatment of a symptom-producing disc herniation, the current randomized study suggests that, for selected patients who meet specific criteria, a video-assisted arthroscopic microdiscectomy may be useful for the operative treatment of a lumbar disc herniation.
The shorter period of postoperative disability for the patients who had a video-assisted arthroscopic microdiscectomy may be attributed to the absence of the epidural fibrosis and tethering of nerve roots that are commonly observed after laminotomy and intracanalicular approaches3,4,23,26. In addition, the epidural venous systems7,21 are not disturbed during video-assisted microdiscectomy. This helps to prevent the postoperative development of venous stasis and chronic nerve-root edema. The minimum operative trauma inflicted on myoligamentous structures24,30 also may play an important role in the rapid recovery of patients after video-assisted microdiscectomy.
Arthroscopic visualization of the contents of the spinal canal in the patients in Group 2 made it possible to gain access to the herniated lumbar disc with use of a posterior subligamentous approach without entering the spinal canal15-17. The midpedicular placement of instruments adjacent to the spinal canal under discoscopic magnification and illumination made it possible to sweep the inserted instrument anteriorly near the traversing nerve root and the lateral aspect of the dura and to evacuate the herniated disc fragments.
Although the details of the operative technique are not within the scope of this report12,15-17, it should be noted that most orthopaedic surgeons have some familiarity with the posterolateral approach and have used this access for bone biopsies2 and nuclear debulking procedures5,20,29. In contrast, mastery of arthroscopic access to the disc fragments requires commitment and patience and cannot be learned in a short time-period. Particular attention to detail, including selection of patients with use of the criteria described earlier, preoperative planning, and operative technique, is essential to the success of this operative procedure.
During preoperative planning, a decision should be made about whether a single portal, inserted from the symptomatic side of the patient, is sufficient, or whether introduction of a second portal from the opposite side will be necessary for access and retrieval of herniated disc fragments.
Although use of an oval cannula (Figs. 1-A and 1-B) has greatly limited the need for bilateral, biportal access, the removal of a disc causing a large central herniation or one that has penetrated the boundary of the posterior longitudinal ligament requires biportal access for intradiscal triangulation and clear visualization of herniated fragments and the anterior surface of the dural sac.
Positioning of the instruments adjacent to the spinal canal must be emphasized. Although it is advisable, at the onset of the procedure, to make a quick arthroscopic inspection of the contents of the spinal canal to ensure posterior subligamentous positioning of the annular fenestration, epidural bleeding and adipose tissue invariably obstruct proper visualization of neurovascular structures and interfere with intracanalicular procedures performed through a foraminal approach. Inspection of the contents of the spinal canal is accomplished by tilting the external end of the inserted cannula anteriorly (with the patient in a prone position) (Figs. 2 [top left] and 3). The external end of the cannula then is tilted posteriorly (Fig. 2 [top right]) and is held against the annulus in preparation for annular fenestration and retrieval of herniated disc fragments. Final inspection of the intradiscal area through a 30 or 70-degree arthroscope (Fig. 2 [bottom]) will demonstrate the exposure of the anterior aspect of the dura and the adequate retrieval of the herniated disc fragments.
The data from this randomized, prospective study suggest that, in properly selected patients who meet the criteria for inclusion described earlier, a video-assisted arthroscopic microdiscectomy may be useful for the operative treatment of a lumbar disc herniation. However, arthroscopic microdiscectomy is a demanding technique and should not be attempted without specific instruction and training. Similar studies of larger groups of patients are needed to confirm these results statistically.