Vol 8, No 6, November/December 2000 339 Adcon-L (Gliatech, Inc, Cleveland, Ohio) is a sterile absorbable gel matrix composed of a gelatin in a carbohydrate polymer. The material is biodegradable, does not incite any inflammatory reaction, and is reported to disappear completely within 21 days after implantation. Animal studies have shown that Adcon-L (or its prototypes) can reduce the development of fibrotic epidural scars in rat, rabbit, and dog laminectomy models. 1,2 On the basis of subsequent studies in dog laminectomy and diskectomy models, as well as a large prospec- tive multicenter randomized trial performed in Europe, 3 the US Food and Drug Administration (FDA) recently approved Adcon-L as a device (rather than as a drug) for marketing in the United States. Since receiving approval in May 1998, Gliatech has been marketing Adcon-L with the following FDA- approved wording on its label: “(1) Adcon-L significantly reduces postsurgical scarring and adhe- sions. (2) Adcon-L improves pa- tient outcome and reduces activity- related pain as compared to the control group. (3) Adcon-L has no significant adverse events attrib- uted to its use.” Epidural Fibrosis With proper patient selection, more than 90% of patients who undergo surgical treatment for lumbar disk herniations can be expected to have an excellent or good result. Lack of pain relief initially after surgery may be attributed to improper diagnostic interpretation, surgical error, or improper patient selec- tion. If there is initial improvement and later recurrence of symptoms, the reason may be postoperative infection, recurrent disk herniation, arachnoiditis, or epidural fibrosis (Fig. 1). The exact cause of epidural fi- brosis following lumbar laminecto- my remains unknown. Key and Ford 4 first studied this issue in 1948 when they reported epidural scar formation after laminectomy and diskectomy. They theorized that the postoperative adhesions were derived from the surgically dam- aged anulus fibrosus. Nachemson 5 also supported this early theory and felt that leakage from the dis- rupted intervertebral disk was a causative factor in epidural fibrosis. However, LaRocca and Macnab 6 felt that this “laminectomy membrane” was actually a layer of fibrotic tis- sue that covered the laminectomy site as a result of migration of fibroblasts from the raw surface of the erector spinae musculature. Postoperative epidural fibrosis can surround nerve fibers, leading to increasing neural tension and adhesion to the nerve root. It has also been theorized that epidural scar increases the vulnerability of nerve roots to compression or ten- sion, which can occur with recurrent disk herniation or stenosis. Most clinical studies assess post- operative scar formation with quali- tative methods, such as graded his- Dr. Fischgrund is Attending Spine Surgeon, William Beaumont Hospital, Royal Oak, Mich. Reprint requests: Dr. Fischgrund, Suite 100, 16800 West Twelve Mile Road, Southfield, MI 48076-2176. Copyright 2000 by the American Academy of Orthopaedic Surgeons. Abstract Adcon-L is a biodegradable gel matrix that was recently approved by the Food and Drug Administration for use during single-level posterior lumbar laminec- tomy or laminotomy procedures. Experimental and clinical studies have sug- gested that the use of this product will decrease postoperative epidural scarring. However, the relationship between epidural fibrosis and patient outcome remains unclear. If the treating surgeon is of the opinion that there is a need to reduce epidural scar, the use of this product appears worthwhile. However, fur- ther studies are needed to evaluate clinical outcome and justify the use of this product on a routine basis. J Am Acad Orthop Surg 2000;8:339-343 Use of Adcon-L for Epidural Scar Prevention Jeffrey S. Fischgrund, MD Perspectives on Modern Orthopaedics Use of Adcon-L for Epidural Scar Prevention Journal of the American Academy of Orthopaedic Surgeons 340 tology and dissection. Abitbol et al 7 described a new experimental model in dogs that allowed objective biome- chanical quantification of the effect of postoperative scar. The ultimate load of postoperative adhesions on nerve roots and the dura was mea- sured. Their findings suggested a beneficial effect of hyaluronic acid in decreasing the biomechanical strength of extradural adhesions after laminotomy for nerve root exploration or disk excision com- pared with the use of fat graft or no interpositional membrane. One of the major questions re- garding the formation of epidural fibrosis is its causal relationship to clinical symptoms. The literature contains numerous anecdotal reports of patients with postoperative epi- dural fibrosis and radicular pain 8,9 ; however, the cause-and-effect rela- tionship remains to be delineated. Conversely, recent reports concern- ing computed tomography (CT) and magnetic resonance (MR) imaging studies on patients who have under- gone laminectomy or diskectomy have shown a high prevalence of epi- dural fibrosis even in patients with clinically excellent outcomes. 10,11 Both high-resolution CT and MR imaging have been used for postop- erative evaluation. Cervellini et al 12 examined 20 asymptomatic postop- erative patients and 20 patients with recurrent sciatic pain after disk sur- gery. No patients had bone steno- sis, recurrent stenosis, or any other identifiable cause of failed symptom relief. Their observations revealed no difference in the amount or loca- tion of fibrosis demonstrated by CT between symptomatic and asymp- tomatic patients. It was concluded that the degree and type of fibrosis were not related to recurrence of symptoms. Tullberg et al 13 reported a pro- spective study of 36 patients with radicular leg pain and lumbar her- niation who underwent single-level diskectomy. They found that 64% of the patients had evidence of postoperative epidural fibrosis (documented by gadolinium– diethylenetriamine-penta-acetic acid MR imaging). There was no correlation between the postopera- tive MR findings and the clinical outcome. Annertz et al 14 reported on 16 patients evaluated with MR imag- ing after lumbar diskectomy. Two matched groups of 8 patients each— one group being symptomatic and the other asymptomatic—were compared. No relationship was found between epidural fibrosis and sciatica in the two groups. Addi- tionally, the 8 symptomatic patients underwent reexploration and neu- rolysis of the adhesion; 6 (75%) had no statistically significant benefit from the second surgical procedure. The authors of this study concluded that epidural fibrosis alone was not the cause of the symptoms, as surgi- cal release of these adhesions did not result in patient relief. The prevention of epidural fi- brosis has been a field of intensive research by clinical investigators. Animal studies have included the placement of a gelatin foam sponge, Silastic sheeting, fat grafts, polylac- tic acid membrane, and hyaluronic acid in the laminectomy defect. 15-17 Most of these studies have shown a remarkable decrease in the amount of postoperative scarring; however, the clinical efficacy of these agents remains unproved. In contrast to the numerous ani- mal studies performed, there are few prospective human trials that demonstrate the greater efficacy of one interposition membrane over another. Jacobs et al 18 performed a clinical trial that compared gelatin foam sponge and free fat graft and found that those patients with a fat graft had a better clinical outcome. Their patients represented a mixed sample; 28% had previous opera- Figure 1 T1-weighted magnetic resonance images obtained before (A) and after (B) administration of gadolinium demonstrate scar formation around the left S1 nerve after L5-S1 lumbar diskectomy. A B Jeffrey S. Fischgrund, MD Vol 8, No 6, November/December 2000 341 tions, 34% underwent procedures involving multiple levels, and 58% underwent simple disk-herniation surgery. A prospective randomized study by MacKay et al 10 included 156 pa- tients who were treated surgically for lumbar disk herniations and randomly assigned to one of three groups. All patients underwent an uncomplicated one-level lumbar laminectomy for disk herniation. In two groups, either fat or gelatin foam sponge was placed in the laminectomy defect; in the third group, no interposition membrane was used. The patients were as- sessed at 1 year. Overall, the extent of epidural scar formation was rated excellent or good in 83% of the patients. The authors concluded that the clinical outcome after disk surgery did not correlate with the use or type of interposition mem- brane placed to prevent epidural fibrosis. Adcon-L Adcon-L is supplied in gel form in 5-g tubes. The package insert lists the indications for its use as single- level posterior lumbar laminectomy or laminotomy procedures in which nerve roots are exposed. The use of Adcon-L was evaluated in a recent publication by Ross et al, 19 who described the association be- tween epidural scar and recurrent radicular pain after lumbar diskec- tomy. This study (which was sup- ported by Gliatech) reported on 197 patients who underwent first-time single-level unilateral diskectomy for lumbar disk herniation. At 6 months after surgery, all patients recorded their pain on a visual ana- log pain scale. Additionally, all pa- tients underwent MR imaging 6 months postoperatively. The studies were evaluated by a neuroradiolo- gist who was blinded to the clinical findings. The amount of scar for- mation was graded from 0 to 4, with a score of 0 indicating no scar for- mation or only traces thereof and a score of 4 indicating that more than 75% of one quadrant was filled with scar. Overall, the authors concluded that the amount of scarring was a significant (P = 0.009) predictor of recurrent radicular pain. Those patients with extensive scar forma- tion (scar score of 4) were 3.2 times more likely to experience recurrent radicular pain than patients with a scar score of less than 4. Several components of pain were measured in that study, including pain when most severe, pain on average, and pain at the end of the day. Only the “pain on average” and the “pain when most severe” measures were analyzed. Only the latter measurement correlated sig- nificantly with scar formation. This study still leaves unanswered several questions about the role of epidural fibrosis in postoperative radicular pain. Only 14 (17%) of the 84 patients who had extensive scar (grade 4) had recurrent radicular pain (most severe). The other 83% of the patients with extensive scar reported no radicular pain. There- fore, there must be factors other than scar formation that also influence patient outcome. As part of the FDA approval for marketing of Adcon-L in the United States, the manufacturer, Gliatech, was required to complete the US clinical study and submit the results for inclusion in product labeling. In June 2000, the FDA reviewed these clinical results and issued a report that raised many concerns regard- ing the data from the US study. In this ongoing study, the researchers attempted to validate the postopera- tive MR imaging studies by reading and scoring 115 of the 324 MR stud- ies a second time. According to a re- cent story in The Wall Street Journal (August 31, 2000; Marketplace sec- tion:B11), scrutiny by the FDA investigators reportedly revealed that 32 of the 115 readings were allegedly different from the original readings, with 29 second scorings resulting in a more favorable out- come for Adcon-L. In addition, it appeared that 6 of the second scor- ings had erasures or write-overs, which resulted in a more favorable outcome for Adcon-L. The issues raised by this FDA audit call into question the accuracy of the entire study. In an animal study, Einhaus et al 1 examined seven adult mongrel dogs on which multiple-level unilateral hemilaminotomy and annular fen- estration were performed. In each animal, Adcon-L was applied to two randomly assigned sites around the diskectomy, the nerve roots, and the hemilaminotomy. Identical surgery was performed on a third site in each animal, but no interposi- tion membrane was placed. The rel- ative amount of anterior scarring around the diskectomy site and pos- terior scarring at the hemilaminot- omy site was assessed histologically, as was the extent of healing of the annular defect. The authors found that Adcon-L significantly (P<0.05) reduced epidural fibrosis in this model and did not inhibit healing of the anulus fibrosus. The large European clinical study by de Tribolet et al 3 deserves care- ful scrutiny. In that prospective, multicenter, randomized, double- blind trial at nine neurosurgical centers in three countries, the data on 269 patients who underwent primary lumbar surgery for re- moval of a unilateral single-level herniated disk were evaluated. Patients were randomly assigned to receive either Adcon-L or no interposition membrane (control group). Both preoperative and 6- month postoperative MR studies of all patients were obtained. In the treatment group, 3 to 5 g of Adcon-L gel was applied to all accessible surfaces of the nerve root between the thecal sac and the posterior lon- Use of Adcon-L for Epidural Scar Prevention Journal of the American Academy of Orthopaedic Surgeons 342 gitudinal ligament, between the dorsal surface of the thecal sac and the overlying muscle and fascia up to the lower surface of the lamina, and around the nerve root as it en- tered the foramina. The 6-month postoperative MR imaging study was used to quanti- tate the extent of epidural scar. The scar formation was given a score from 0 to 4 based on the percentage of fibrosis within a quadrant. Post- operative outcome was assessed with an activity-related pain ques- tionnaire, the straight-leg-raising examination, and a clinical assess- ment of pain. A single radicular fac- tor score for each patient was then determined on the basis of the scores for 5 of 12 specific daily activi- ties evaluated on the activity-related pain questionnaire. Analysis of the postoperative MR results demonstrated a significant (P = 0.002) reduction in epidural scar formation in patients treated with Adcon-L (Table 1), similar to the results in previous animal stud- ies. 1 These results were supported by the clinical findings in a small group of patients who underwent reoperation at some point between the first postoperative week and as long as more than 1 year later. Nine of the 11 patients had no nerve root adhesion or only minimal adhesion around the nerve root, and 8 had no anterior adhesions or only minimal adhesions. Residual Adcon-L gel was not observed in any patient who underwent reoperation more than 22 days after the initial surgery. Considerably more important than the MR findings is the postop- erative patient outcome, which unfortunately is difficult to assess from this single trial. Significantly (P = 0.013) fewer Adcon-L–treated patients reported increased pain when performing activities of daily living compared with the control group. However, careful review of the data indicates that there was no significant difference in the radicu- lar pain response from the baseline. That is, the amount of leg pain the patient was having postoperatively was clinically improved in both patient groups, and no intergroup difference was noted at 6 months postoperatively. Interestingly, Adcon-L–treated patients experienced less back pain. It is hard to explain this fact physiologically, because the major advantage of Adcon-L is reduction of epidural fibrosis around the nerve root, which has not been proved to affect low back pain. Furthermore, Adcon-L–treated patients also had improved scores on straight-leg-raising examination on the operative as well as the non- operative side. The authors found this improvement particularly important, because they believe this clinical test is the standard used by many surgeons for assess- ment of postoperative outcome. The European clinical study did not report any complications directly attributable to the use of Adcon-L. There was no effect on wound healing, and the incidence of postoperative infection was not increased compared with the con- trol group. However, there are the- oretical concerns regarding its use after incidental durotomy. Because this product decreases the inci- dence of postoperative epidural scar by decreasing the inflammatory response, there may also be persis- tence of dural leaks. Small dural lacerations can occur during rou- tine surgery and may go unnoticed by the operating physician. During the course of normal postoperative healing, these leaks rarely become clinically significant, as the inflam- matory reaction that occurs after a surgical procedure is often suffi- cient to seal the leak. Therefore, the surgeon should use this prod- uct very cautiously if an incidental dural laceration is suspected. Summary On the basis of the available infor- mation, it is difficult to advocate the routine use of Adcon-L after lum- bar laminectomy. Unfortunately, the manufacturer of Adcon-L has sponsored all experimental and clinical studies presented and pub- lished thus far, and recent FDA concerns have called into question the reliability of the data. These studies indicate that the use of this product will decrease the amount of postoperative fibrosis in the lam- inectomy defect. Because of this decrease in fibrosis, there is a theo- retical advantage to using this product in patients who require revision surgery (patients in whom Adcon-L was placed at the initial surgery). However, it has not yet been conclusively shown that this reduction in epidural scarring translates into improved patient outcome. If the treating surgeon is of the opinion that there is a need for a product to reduce epidural scar, use of this product may be worthwhile. Table 1 Distribution of Scar Based on MR Imaging in Study by de Tribolet et al 3 No. of Evaluable Patients (%) Extent of Scar Adcon-L–Treated (n = 127) Control (n = 139) Minimal or no scar 17 (13) 5 (4) Moderate 62 (49) 65 (47) Extensive 48 (38) 69 (50) Vol 8, No 6, November/December 2000 343 However, it should be noted that the current charge to hospitals for one tube of Adcon-L is $495. If it is estimated that 200,000 laminec- tomies are performed in the United States every year, routine use of Adcon-L would add a rather large expenditure to the national health- care bill. Further studies are needed to evaluate clinical outcomes and establish the cost-effectiveness of this procedure. Jeffrey S. Fischgrund, MD References 1. Einhaus SL, Robertson JT, Dohan FC Jr, Wujek JR, Ahmad S: Reduction of peridural fibrosis after lumbar lami- notomy and discectomy in dogs by a resorbable gel (ADCON-L). Spine 1997;22:1440-1447. 2. Robertson JT, Meric AL, Dohan FC Jr, Schweitzer JB, Wujek JR, Ahmad S: The reduction of postlaminectomy peridural fibrosis in rabbits by a carbo- hydrate polymer. J Neurosurg 1993;79: 89-95. 3. de Tribolet N, Porchet F, Lutz TW, et al: Clinical assessment of a novel anti- adhesion barrier gel: Prospective, ran- domized, multicenter, clinical trial of ADCON-L to inhibit postoperative peridural fibrosis and related symp- toms after lumbar discectomy. Am J Orthop 1998;27:111-120. 4. Key JA, Ford LT: Experimental inter- vertebral-disc lesions. J Bone Joint Surg Am 1948;30:621-630. 5. Nachemson A: Intradiscal measure- ments of pH in patients with lumbar rhizopathies. Acta Orthop Scand 1969; 40:23-42. 6. LaRocca H, Macnab I: The laminecto- my membrane: Studies in its evolu- tion, characteristics, effects and pro- phylaxis in dogs. J Bone Joint Surg Br 1974;56:545-550. 7. Abitbol JJ, Lincoln TL, Lind BI, Amiel D, Akeson WH, Garfin SR: Preventing postlaminectomy adhesion: A new experimental model. Spine 1994;19: 1809-1814. 8. Martin-Ferrer S: Failure of autologous fat grafts to prevent postoperative epidural fibrosis in surgery of the lumbar spine. Neurosurgery 1989;24: 718-721. 9. Burton CV, Kirkaldy-Willis WH, Yong-Hing K, Heithoff KB: Causes of failure of surgery on the lumbar spine. Clin Orthop 1981;157:191-199. 10. MacKay MA, Fischgrund JS, Herko- witz HN, Kurz LT, Hecht B, Schwartz M: The effect of interposition mem- brane on the outcome of lumbar lami- nectomy and discectomy. Spine 1995; 20:1793-1796. 11. Jensen TT, Asmussen K, Berg-Hansen EM, et al: First-time operation for lum- bar disc herniation with or without free fat transplantation: Prospective triple-blind randomized study with reference to clinical factors and enhanced computed tomographic scan 1 year after operation. Spine 1996;21: 1072-1076. 12. Cervellini P, Curri D, Volpin L, Ber- nardi L, Pinna V, Benedetti A: Com- puted tomography of epidural fibrosis after discectomy: A comparison be- tween symptomatic and asymptomatic patients. Neurosurgery 1988;23:710-713. 13. Tullberg T, Grane P, Isacson J: Gado- linium-enhanced magnetic resonance imaging of 36 patients one year after lumbar disc resection. Spine 1994;19: 176-182. 14. Annertz M, Jönsson B, Strömqvist B, Holtås S: No relationship between epidural fibrosis and sciatica in the lumbar postdiscectomy syndrome: A study with contrast-enhanced magnet- ic resonance imaging in symptomatic and asymptomatic patients. Spine 1995;20:449-453. 15. Songer MN, Rauschning W, Carson EW, Pandit SM: Analysis of peridural scar formation and its prevention after lumbar laminotomy and discectomy in dogs. Spine 1995;20:571-580. 16. DiFazio FA, Nichols JB, Pope MH, Frymoyer JW: The use of expanded polytetrafluoroethylene as an interpo- sitional membrane after lumbar laminectomy. Spine 1995;20:986-991. 17. Cook SD, Prewett AB, Dalton JE, Whitecloud TS III: Reduction in peri- neural scar formation after laminecto- my with polyactive membrane sheets. Spine 1994;19:1815-1825. 18. Jacobs RR, McClain O, Neff J: Control of postlaminectomy scar formation: An experimental and clinical study. Spine 1980;5:223-229. 19. Ross JS, Robertson JT, Frederickson RCA, et al: Association between peri- dural scar and recurrent radicular pain after lumbar discectomy: Magnetic resonance evaluation. Neurosurgery 1996;38:855-863. . single-level diskectomy. They found that 64% of the patients had evidence of postoperative epidural fibrosis (documented by gadolinium– diethylenetriamine-penta-acetic acid MR imaging). There was no correlation