Debate: To Fuse or Not to Fuse to the Sacrum, the Fate of the L5–S1 Disc

Abstract

In adult scoliosis deciding the caudal extent of the fusion can be problematic. A case is presented that represents the dilemma. The opposing viewpoints are presented representing the dilemma of the current lack of compelling evidence to support or refute either position. The two debaters argue for their viewpoints based on literature review and clinical experience. Case for Discussion: David W. Polly, Jr., MD A 51-year-old woman was diagnosed with scoliosis as a child but did not wear a brace. She has lost 2 inches of height since high school. Over the last 2 years, she has been progressively limited by her back. She is comfortable when sitting but has limited standing ability. She initially had good relief with epidural steroid injections. Physical examination shows her adult scoliosis with acceptable global sagittal balance. The neurologic examination is normal except for a slight decreased sensation in the right L5 dermatome. Radiographic examination shows a right lumbar curve with marked degenerative changes in the upper lumbar spine (Figure 1). The global sagittal balance is normal, but there is local kyphosis in the lumbar spine. The L5–S1 disc height is reasonably well preserved and reasonably parallel (Figure 2). On MRI, the disc height is good and the morphology normal for age with good hydration on T2-weighted images (Figure 3).Figure 1: Coronal and sagittal radiographs of 51-year-old woman diagnosed with scoliosis as a child but not treated with a brace. She has had progressive height loss and worsening pain. (Case courtesy of Dr. Joe Perra.)Figure 2: The spot lateral radiograph demonstrates good L5–S1 disc height and no evidence of listhesis.Figure 3: The MRI shows reasonable disc health with good alignment. There is minimal to no endplate signal change and there is no neural compression at L5–S1.The patient feels that nonoperative management is no longer acceptable and desires to proceed with surgical treatment. Stop at L5: Christopher L. Hamill, MD Our goals in spinal deformity surgery are the prevention of progression, maintenance of lumbar lordosis, restoring global balance, and achieving pain relief. The advantages of stopping long fusions at L5 have been well established. These include the retention of lumbosacral motion, absence of SI joint stress, decreased operative time, and lower pseudarthrosis rate. The fact of not having to cross the lumbosacral junction can also lead to decreased instances of instrumentation complication. The ability to fuse to L5 does require a relatively normal L5–S1 disc. This means a well-hydrated L5–S1 disc and a disc that has not lost its height. The relatively normal disc is usually found in younger patients who also have good bone stock. Overall lumbar lordosis in these patients needs to be well maintained. If the patient meets these criteria of a well-hydrated disc with good lumbar lordosis, the fusion need not be extended to the sacrum. Without having to fuse to the sacrum, operative times will be decreased. This may be due to less levels to be exposed. The fusion, if it does not cross the lumbosacral junction, will obviate the need for circumferential surgery as well as pelvic fixation. Edwards et al,1 in Spine in 2004, looked at a matched cohort analysis. They had 27 patients fused to L5 and 12 patients fused to S1. They reviewed radiographs, complications, and outcomes forms in the SRS-24 (Table 1). Their matching criteria were for age, smoking status, status of the L5–S1 disc, plumb line, and number of levels fused. They found their complication rate to be 22 major complications in the L5–S1 fusion versus 75 complications in the S1 fusion. The total procedures required were at 1.7 in the L5 patients versus 2.8 in the patients fused to S1. The SRS-24 outcome forms showed no significant difference between the two patient populations. Edwards et al,2 in Spine in 2003, looked at the fate of the L5–S1 disc in patients undergoing long fusions. They had 34 patients with an average followup of 5.6 years. They found that 19 (61%) developed advanced disc degeneration and 4 required revision to the sacrum. However, their functional outcomes that included the SRS-24 showed no significant difference between the two groups, with 88% showing good or excellent results. They did find that patients with preexisting degenerative disc disease correlated with a lower self-image, function satisfaction, and total SRS score. Their conclusion was that fusion to the L5 did have a good outcome but was dependent on L5–S1 disc competency. Those patients that had undergone advanced L5–S1 disc degeneration did have a loss of sagittal balance. The clinical relevance of this was mixed. They also found that patients who had a loss of fixation were associated with a deep-seated L5 vertebral body.Table 1: Complication RatesThere were other studies looking at fusion to the sacrum, including the Emami et al3 in Spine in 2002 where they looked at 54 patients fused to the sacrum; 11 of them fused to the sacrum using the Luque Galveston technique, 36 with Isola and iliac wing screws, and 12 patients bicortical S1 screws. They found a significant pseudarthrosis rate in Group 1 of 36%, Group 2 of 14%, and Group 3 of 8.5%. Grubb et al,4 in Spine in 1994, looked at 28 patients requiring fusion to the sacrum and found a 17.5% pseudarthrosis rate with their type of technique. In summary, the ability to stop at L5–S1 does depend on the L5–S1 disc. This requires a good disc height and hydration. This is most commonly found in a younger patient who also has good bone stock. The support of the L5 screws can be enhanced by the use of anterior column support in the form of an anterior structural graft. This can be placed as an anterior lumbar interbody fusion (ALIF) or a transforaminal lumbar interbody fusion (TLIF). There is some thought of bicortical L5 screws; however, because of vascular and neurologic structures anteriorly, this can be difficult, if not dangerous, to perform. Those patients in which the fusion stops at L5 should have good overall lumbar lordosis and a well-maintained sagittal balance. Extend to the Pelvis: Keith H. Bridwell, MD Whether to Stop at L5 or the Sacrum? To Advocate Stopping at the Sacrum. The goals of surgical treatment of adult lumbar scoliosis are to improve back and leg pain and accomplish a solid fusion with good coronal and sagittal balance. The most common presentation is one of a rotatory subluxation at L3–L4 and fixed tilt at L4–L5. Therein, it is usually not possible to stop at L3 or L4 in a patient in his or her 50s. The question comes whether to stop at L5 or the sacrum. I think that most surgeons would advocate stopping at the sacrum if there is severe disc degeneration at L5–S1 or if spondylolisthesis or spinal stenosis exists at that segment. The other absolute indication for extending to the sacrum is if there is an oblique take-off at L5–S1. However, there is considerable controversy whether to stop at L5 or the sacrum if the disc appears relatively well hydrated and has good height at L5–S1. In almost all circumstances, there will be some disc degeneration at this segment. The perceived disadvantages of extending to the sacrum rather than L5 are as follows: A bigger operation is needed when fusing to the sacrum. More exposure is required, and usually anterior column support is advocated at the distal lumbar spine through either a posterior or anterior approach. If motion is lost at L5–S1, this might alter the patient's gait. If fusion is extended to the sacrum, many advocate the use of iliac screws. Many feel that iliac screws are imperfect and lead to a high rate of “painful instrumentation” that requires ultimate removal. Fusing to the sacrum runs the risk of pullout of the sacral screws and a higher pseudarthrosis rate at L5–S1. The concern about stopping at L5 is that subsequently the L5–S1 disc will degenerate in a high percentage of cases1,2 and that ultimate extension to the sacrum will be required in many patients. Further, as the L5–S1 disc degenerates below a long fusion, this often leads to a substantial forward shifting of the C7 plumb relative to the lumbosacral disc. There is no perfect study analyzing gait differences between fusions stopping at L5 or the sacrum. However, a pilot study by Engsberg et al5 showed no difference in the gait of subjects whether their fusion was stopped at L5 or the sacrum. On the other hand, there was a dramatic difference in gait whether the patient was in neutral or positive sagittal balance.5,6 Certainly, the data from our institution by Kim et al7–9 verify that the pseudarthrosis rate is higher when the fusion is extended to the sacrum versus stopping at L5. However, if pseudarthrosis occurs at L5–S1, this can usually be salvaged by a relatively small operation revising the implants at the lumbosacral junction and achieving fusion on revision using bone morphogenetic protein (BMP). If the nonunion is identified early, than the patient does not lose sagittal balance. Many surgeons think that iliac screws are prominent and need to be removed. We have found there is a subset of adult patients (approximately 15%) who complain about their iliac screws and do benefit from their removal. Removing the iliac screws is a relatively small operation. One potential consequence of not having iliac protection of sacral screws is pullout of the sacral screws and stress fracture through the sacrum. I have anecdotally been sent many cases of this complication throughout the country, so I tend to think it is a bigger problem that what is currently reported in the literature. Our analysis of sacropelvic fixation with >5-year follow-up found the complication of iliac screws to be relatively small. We saw no cases of disabling pain or sacroiliac joint degeneration.10 We have found that the lowest pseudarthrosis rate at L5–S1 is associated with complete sacropelvic fixation and surgery in patients under the age of 55 years.9 By complete sacropelvic fixation, we mean bilateral bicortical S1 screws out through the promontory anteriorly, bilateral iliac screws and anterior column support at L5–S1. There are three potential ways of accomplishing anterior column support: TLIF, posterior lumbar interbody fusion (PLIF), or ALIF. There is no substantial report in the literature about this option just yet, but I have anecdotally seen several cases in which the TLIF cage has not been placed ideally and failure has occurred through the sacral screws. A thoracoabdominal approach is often used to achieve exposure of all segments of the lumbar spine. However, a disadvantage of this approach is that one does not access the L5–S1 disc as directly anteriorly as through a paramedian approach. A paramedian approach carries less morbidity than a thoracoabdominal approach, although there is still some potential for abdominal muscle denervation with the approach. Horton et al11 describe the morbidity of thoracoabdominal versus lumbar oblique versus paramedian-type approach. There was some morbidity to all three approaches but far less through the paramedian approach. Although we do not have substantial numbers to document this point, to date my observation has been the problems with a somewhat bulgy abdomen are more likely to occur if segments above L4–L5 and L5–S1 are exposed, and it is relatively uncommon if only L4–L5 and L5–S1 are exposed below. Of the three options, my personal preference is ALIF by the paramedian approach, although the TLIF approach is seemingly the most “trendy” at the present time. In the work by Edwards et al1,2 and a subsequent work that is being presented at the Scoliosis Research Society meeting in September 2006 by Kuhns et al, we have found a substantial amount of late disc degeneration at L5–S1, even if the disc is relatively radiographically normal before surgery by the Weiner classification.1,2 Therefore, we feel that strong consideration should be given toward extending the fusion to the sacrum in this circumstance. We do not think that the morbidity is unacceptably high. Any sagittal plane anomaly will seemingly increase the likelihood of the L5–S1 disc subsequently degenerating, as reported in our papers on the results of pedicle subtraction osteotomy.12 So, we advocate extending to the sacrum with bilateral sacral screws out through the promontory anteriorly and protected with bilateral iliac screws. We would try to do our best at burying the iliac screws. If they are prominent in a thin patient, we would advise that potentially at 2 years postoperation the iliac screws can be removed. We would also protect the sacral screws with cages performed anteriorly through a paramedian approach. Our preference would be the large trapezoidal titanium mesh cages. And we would perform the “fusion” with BMP rather than harvesting additional autogenous bone. Case Follow-up The patient underwent an L1–L5 posterior fusion with decompression and transforaminal lumbar interbody fusion at L4–L5 and a TLIF at L3–L4 (Figures 4, 5). Segmental pedicle screw instrumentation was used. Bone graft was local bone, bone marrow aspirate, and allograft. Her postoperative course was benign. At 1 year postoperatively, her visual analog scale pain ratings went from a 7–8 to a 1–2. She was working full time, able to touch her toes, and able to go on long motorcycle rides, something that was very important to her. At 2 years after the surgery, she was still quite functional but had significant pain when she overdoes it. At 2 years 9 months after surgery, she had significantly limiting pain. Sacroiliac joint injection gave her no relief. Imaging showed progressive degeneration of the L5–S1 disc. She underwent a second surgery with extension of the fusion to the pelvis (Figure 6). She had a bilateral TLIF at L5–S1 using a polyetheretherketone (PEEK) cage, bone morphogenetic protein (rh-BMP-2), and local bone. She had bilateral S1 pedicle screws and sub-S1 screws through a sacral plate.Figure 4: The detail lumbar films after fusion to L5 show the segmental instrumentation and transforaminal lumbar interbody fusions at L3–L4 and L4–L5.Figure 5: The long films show good global coronal and sagittal balance after the surgery.Figure 6: The patient subsequently had more than 2 years of excellent pain relief and function. She then had further deterioration requiring extension of the fusion to the sacrum. This was accomplished with S1 and alar fixation. She has again had improvement in pain and some restoration of function but is more limited than after the initial surgery.The patient states she was very happy with the pain relief and functional improvement she had for the 2 years after surgery and would do the same thing all over again. It is too early to tell about the results of the extension of the fusion to the sacrum. Acknowledgment The authors thank Dr. Joe Perra for providing the case history and radiographs for this debate.