How does back muscle strength change after posterior lumbar interbody fusion?

Sixty-five women with symptomatic lumbar degenerative diseases were recruited from the Department of Orthopedic Surgery at Samsung Medical Center between January 2010 and December 2010. The mean age was 58 years (range 43–68 years). The inclusion criteria stipulated that adequate nonsurgical treatments had failed to improve back pain and that patients had functional disability for at least 6 months, that they were candidates for 1- or 2-level PLIF between L-3 and S-1, and that they were available for complete follow-up visits that included assessments with quality of life (QOL) questionnaires. The exclusion criteria were the following: 1) a previous lumbar spinal operation, 2) a combined cervical or thoracic myelopathy, 3) an inability to respond to questions in the questionnaires and to properly follow the protocol of the strength test, and 4) a failure to perform activities of daily living because of comorbidities (e.g., cerebrovascular accident, Parkinson disease, psychotic disorders, or other orthopedic degenerative diseases). Men were not included in the present study because only a small number of potential participants were available, and there was the possibility of lower participation by men in clinic-based supervised postoperative exercises.

All diagnoses and procedures were performed by 1 senior surgeon at our institution (C.S.L.). In all cases, PLIF with a titanium cage and concomitant decompression were performed and augmented by posterior spinal instrumentation. When possible, an autogenous local bone graft obtained during surgery was used in the fusion, and an irradiated allograft bone was also used if insufficient local bone was available. We did not use bone morphogenetic protein, demineralized bone matrix, or any other kinds of bone substitutes in this study. For 3 months postoperatively, all patients wore a rigid lumbosacral orthosis.

At 3 months after PLIF, patients in the LSE group were instructed in an exercise protocol developed at our institution. The clinic-based LSE consisted of 36 sessions of 1 hour each and ran for 12 weeks. The LSE consists of a series of warm-up and main exercises of chiefly back stretching and lumbar stabilization^2,3,24,26 and is performed without a range-of-motion restriction in the lumbar spine. The warm-up exercises were performed for 15 minutes on a treadmill or bicycle, and the main exercise was closely supervised by 1 physical therapist (Y.G.S.) at our sports medicine center. The main exercise postures are shown in Fig. 1 and were divided into stretching positions (knee to chest, modified knee to chest, piriformis, lower body twist, cat and camel position, and prayer position) and strengthening positions (pelvic tilt, partial sit-up, bridge, prone arms and legs raise, and reverse crunch). The 33 patients in the control group had no exercise prescription and were asked to continue their usual exercise and physical activities without any restriction.

The isometric strength of back muscles was determined with a lumbar extension machine (MedX). Each test measured the maximal voluntary isometric strength of the lumbar extensor muscles at 7 angular positions (0°, 12°, 24°, 36°, 48°, 60°, and 72°) of lumbar flexion. All patients were thoroughly instructed in the method for accurate testing and performed warm-up exercises for 15 minutes before the test. The patients were positioned according to an established protocol and were then asked to increase the lumbar extension torque over a 2- to 3-second period. Once maximal tension had been achieved, the patients were instructed to maintain the contraction for an additional 1–2 seconds. After reaching maximum torque, they were asked to slowly decrease the torque. During the contractions, visual feedback was provided on a video display screen interfaced with the machine, and the patients were also verbally encouraged to give their maximum effort.^10,25 The patients were given a 10-second resting period between angle conditions, and the maximal isometric voluntary extension of the back muscles was measured (in foot-pounds). All patients were assessed by the same physical therapist at our sports medicine center who was blinded to other results. Repeated tests for measuring reliability were not conducted.

The body mass index and bone density were also measured. Interbody fusion was defined as an angulation of 5° or less and translation as 3 mm or less on full flexion-extension lateral radiographs and as continuous trabecular bone growth connecting the vertebral bodies assessed on reformatted CT scans at 12 months postoperatively.⁵

With a predefined statistical significance level of 5% and to achieve a statistical power of 80%, we calculated that the minimum sample size needed for this study was 60 evaluable patients.²¹ A medical statistical consultant performed the statistical analyses with SAS version 9.1 (SAS Institute Inc.). The data were expressed as mean ± SD and were analyzed with paired and unpaired Student t-tests, Mann-Whitney U-test, and Wilcoxon signed-rank test. A p < 0.05 was considered statistically significant.

No adverse effects of the LSE were observed. The LSE and control groups had similar extensor strengths both preoperatively (LSE group 90.5 ± 26.8 ft-lb vs control group 99.1 ± 27.7 ft-lb, p = 0.468) and 3 months postoperatively (LSE group 83.4 ± 25.6 ft-lb vs control group 91.5 ± 29.2 ft-lb, p = 0.488). However, after completion of the 12-week LSE program at 6 months postoperatively, the patients in the LSE group had significantly higher mean extensor strength (135.1 ± 24.9 ft-lb, corresponding to a 49.3% increase relative to the preoperative strength) than those in the control group (116.4 ± 38.6 ft-lb, corresponding to a 17.5% increase relative to the preoperative strength) (p = 0.012). At 12 months postoperatively, the mean strength was also significantly higher in the LSE group (148.6 ± 23.6 ft-lb, corresponding to a 64.2% increase relative to preoperative strength) than in the control group (120.6 ± 33.1 ft-lb, corresponding to a 21.7% increase relative to the preoperative strength) (p = 0.012).

In both groups, the mean preoperative VAS score for back pain (LSE group 58.6 ± 26.9 and control group 59.5 ± 26.7) had decreased significantly by the last follow-up at 12 months postoperatively (LSE group 24.5 ± 14.4 and control group 44.0 ± 21.9) (p = 0.004 and p = 0.031, respectively). However, the rate of the decrease in self-reported back pain was more significant in the LSE group than in the control group at 6 months postoperatively (decrease of 56.3% and 29.6%, respectively, p = 0.021) and at the last follow-up (decrease of 58.2% and 26.1% respectively, p = 0.013) (Fig. 3).

Preoperatively, scores on the 8 sections of the SF-36—that is, for physical functioning, physical role functioning, general health perceptions, vitality, social role functioning, emotional role functioning, and mental health—did not significantly differ between the LSE and control groups (p > 0.05), except for bodily pain (LSE group 36.5 ± 5.8 and control group 31.4 ± 6.5, p = 0.038). However, at the last follow-up, the scores for physical role functioning, bodily pain, general health perception, vitality, and social functioning were significantly higher in the exercise group than in the control group (p < 0.05) (Fig. 5).

None of the patients showed an angulation of 5° or greater or a translation of 3 mm or more on full flexion-extension lateral radiographs. On CT scans acquired 12 months postoperatively, 25 (96%) of the 26 patients in the exercise group and 31 (94%) of the 33 patients in the control group had complete bone bridge formation between the adjacent vertebral bodies. In each group, patients without a complete bone bridge did not have a continuous bone bridge on their CT scan but also had no radiographic evidence of instability or clinical correlation with nonunion.

This study had some limitations. First, we did not perform a proper randomization. However, we note that preoperative demographics were very similar between the 2 groups. Although most of the patients wanted to participate in the clinic-based postoperative LSE program, a true randomization was difficult because the program ran for 12 weeks and differences in work obligations and distances between the hospital and patients' residences would have influenced patient participation. Second, back strength was evaluated only as lumbar extensor strength and did not include flexion power of other trunk muscles. Although a balance between the flexion and extension strength is considered to be important in controlling back pain and functional disability,^30,32,33 we deemed evaluating back extensor strength important for patients with PLIF because this surgical approach mainly has an effect on the deterioration of lumbar extensor muscles, such as the longissimus or multifidus muscles. Last, the patients in the LSE group tended to be younger than the patients in the control group, and younger age might be an important factor influencing postoperative outcomes. Further studies with a longer followup and with a larger patient cohort are therefore needed. Nonetheless, the results of the present study were based on validated and quantitative measurement tools and provide new evidence that LSE is safe and of benefit to patients who underwent lumbar spinal fusion.