- Email: [email protected]
Low-intensity pulsed ultrasound improves spinal fusion
The Spine Journal 1 (2001) 246–254
Low-intensity pulsed ultrasound improves spinal fusion Stephen D. Cook, PhD*, Samantha L. Salkeld, MSE, Laura Popich Patron, BSE, John P. Ryaby, BSE, Thomas S. Whitecloud, III, MD Department of Orthopaedic Surgery. SL32, Tulane University Medical School, 1430 Tulane Avenue, New Orleans, LA 70112, USA Received 22 September 2000; revised 30 October 2000; second revision 22 February 2001; accepted 12 May 2001
Abstract
Background context: Increasing the incidence of solid bony fusion is a primary goal in spine surgery. Daily low-intensity pulsed ultrasound therapy has been shown to improve and accelerate the bone healing process. Purpose: The purpose of this study was to evaluate the efficacy of daily low-intensity pulsed ultrasound therapy to improve the rate and quality of spinal fusion. Study design: Canine fusion model prospective study. Patient sample: Fourteen adult male dogs were used. Outcome measures: Radiographic grading of plain films, computed tomography (CT) and magnetic resonance imaging (MRI), gross palpation, torsional stiffness and histologic grading were used to determine the presence or absence of fusion. Methods: Posterior noninstrumented bilateral fusions were evaluated at the L2–L3 and L5–L6 levels. Treatment with low-intensity pulsed ultrasound for 20 minutes per day over the fusion site (stimulated) was compared with fusion sites that received no stimulation (nontreated controls) at 6 and 12 weeks after surgery. Plain film radiographs, CT and MRI, mechanical torsion testing and histologic examination were performed. Results: At 6 weeks, ultrasound treated sites were more frequently fused compared with nontreated controls, although the difference in fusion rate was not statistically significant. At 12 weeks after surgery complete radiographic and histologic fusion occurred in 100% of ultrasound-treated sites. In the nontreated control sites 78% had achieved complete radiographic fusion and 44% had complete histologic fusion. Compared with control sites, the histological and mechanical fusion rate was significantly greater in ultrasound-treated sites (P.05) at 12 weeks. A statistically significant increase in mechanical stiffness in ultrasound-treated sites was also found at 12 weeks after surgery. Conclusions: Low-intensity pulsed ultrasound therapy may be a useful means to ensure successful spine fusion. © 2001 Elsevier Science Inc. All rights reserved.
Keywords:
Ultrasound; Spinal fusion; Bone formation; Histology; Mechanical testing
Introduction Posterior spinal fusion is the most common procedure for spinal stabilization [1]. Improved mechanical stability can help relieve symptoms in patients with degenerative arthritis of the spine, spinal deformity or instability. Posterior spinal fusion generally includes onlay grafting of autogenous corticocancellous iliac crest bone after decortication of the bony surfaces of the posterior vertebral elements [2]. Failure of spinal fusion can result in significant morbidity. The incidence of pseudoarthrosis can range from 5% to This paper was supported by a research support grant from Exogen, Inc. * Corresponding author. Tel.: (504) 588-2273; fax: (504) 584-2722. E-mail address: [email protected] (S.D. Cook)
40% in a large series, depending on many factors, including whether instrumentation is used, the number of levels fused, surgical technique, smoking history, age and method of evaluation [1]. The clinical consequences of pseudoarthrosis can be significant and include pain, progression of curvature in scoliosis surgery and unpredictable results of operations for segmental instability. Increasing the incidence and rapidity of solid bony fusion is a primary goal in spine surgery. Ultrasound is a form of mechanical energy that is transmitted into the body as high-frequency acoustical pressure waves [3]. Low-intensity ultrasound can be used to produce micromechanical stresses and strains in tissues. The micromechanical force applied to bone tissue is less than 10 mg,
1529-9430/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S1529-9430(01)00 0 8 6 - 9
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
which can penetrate to depths of greater than 10 cm despite the attenuation effects of overlying soft tissue. The intensity of the ultrasound (less than 30 milliwatts per square centimeter) is similar to that used in diagnostic imaging. Ultrasound promotes bone formation in a manner comparable to that first postulated and described by Wolff’s law [4]. The physiological response of bone to mechanical force, including effects on bone healing of early weight bearing, dynamic compression techniques and impact loading, have been described in the literature [5–7]. In placebo-controlled animal studies, Duarte [3] reported that daily low-intensity pulsed ultrasound therapy accelerated bone healing in both fibular osteotomy and femoral defect models. Pilla et al. [8] reported a statistically significant increase in the mechanical strength of low-intensity ultrasound-treated osteotomy sites versus control sites. Wang et al. [9] and Yang et al. [10] reported that use of low-intensity ultrasound produced a statistically significant increase in fracture callus amount, stiffness and strength in a bilateral, placebo-controlled femoral fracture model that used an intramedullary metal rod for fixation. Molecular biology experiments indicate that specific cartilage genes, including aggrecan mRNA and DNA binding proteins, are expressed when exposed to ultrasound stimulation [10]. Recent research with real-time studies, using fluorescent markers, have demonstrated increases in cellular calcium flux [11]. In addition, low-intensity pulsed ultrasound produces dramatic and immediate increases in the microvasculature and blood flow at the fracture site. Ryaby et al. [12–15] have demonstrated that ultrasound has multifunctional cellular effects, such as increasing 45Ca uptake, modulating transforming growth factor–beta (TGF-) synthesis and parathyroid hormone (PTH) response in mesenchymal or osteoblast cells. These reports suggest a direct effect on bone formation and resorption, noncollagenous protein synthesis and expression of collagen phenotypes and are consistent with the observation of increased cartilage and advanced endochondral ossification. A pulsed, low-intensity ultrasound device has been evaluated in Food and Drug Administration (FDA)–approved multicenter, prospective, randomized, double-blind and placebo-controlled clinical trials (Sonic Accelerated Fracture Healing System; Exogen, Inc., Piscataway, NJ). The ultrasound therapy was administered daily at home for 20 minutes per day. A statistically significant decrease in clinical and radiographic healing time was reported for diaphysis bone (tibial shaft) fractures (active device, 96 days; placebo device, 154 days [P.0001) [16]. The healing time in tibial shaft fractures was found to be reduced 41% in smoking patients and 26% in nonsmokers [17]. In a second clinical trial of the ultrasound device for metaphyseal (posterior displaced distal radius) fractures, the healing time was also reduced 38% (active device, 61 days; placebo device, 98 days [P.0001) [18]. The healing times in distal radius fractures were reduced 51% in smokers and 34% in nonsmokers [17]. These studies confirmed the results obtained in laboratory
247
and animal studies of the stimulatory effect on fracture healing of low-intensity pulsed ultrasound therapy. The purpose of this study was to determine the efficacy of daily pulsed, low-intensity ultrasound therapy in improving the rate of spinal fusion using a canine noninstrumented posterior autogenous bone graft onlay model. Materials and methods Experimental design Posterior spinal fusions were evaluated in 14 adult male dogs. Each animal was fused bilaterally at the L2–L3 and L5–L6 levels with autogenous bone graft harvested from the iliac crests. In 11 animals, one site received active ultrasound stimulation for 20 minutes per day, whereas the other site received no stimulation (Table 1). To evaluate the effect of fusion level, two animals received no stimulation at either site, whereas one animal received ultrasound stimulation at both sites. Six animals were sacrificed at 6 weeks and eight animals at 12 weeks after surgery. The quality of fusion, new bone formation and bone graft incorporation were evaluated radiographically using plain films, computed tomography (CT) and magnetic resonance imaging (MRI). The postretrieval fusion masses were further evaluated by manual manipulation and mechanical testing in torsion and analyzed histologically for evidence of bony fusion using nondecalcified techniques. Animal model and surgical procedure Skeletally mature adult mongrel dogs weighing approximately 24 to 30 kg were randomly divided into two groups corresponding to implantation periods of 6 weeks (six animals) and 12 weeks (eight animals; Table 1). Before surgery, the animals were allowed to adapt to vivarial conditions for 2 weeks and were treated with antibiotics and antiparasitics to ensure uniform good health. The animals were radiographically screened to ensure the absence of skeletal abnormalities. All surgical procedures and animal
Table 1 Experimental design: distribution of group sample sizes Evaluation period 6 weeks
Treatment group
Levels fused
Sample size
Autogenous bone graft treated with daily ultrasound therapy
L2–L3 L5–L6 Total L2–L3 L5–L6 Total
3 3 6 3 3 6
L2–L3 L5–L6 Total L2–L3 L5–L6 Total
3 4 7 5 4 9
Autogenous bone graft (control)
12 weeks
Autogenous bone graft treated with daily ultrasound therapy Autogenous bone graft (control)
248
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
care were performed according to National Institutes of Health guidelines for the care and use of laboratory animals (publication number 85-23, revision 1985) with the supervision of a licensed veterinarian. Using standard aseptic techniques, a midline skin incision was made from levels L1 to L7. The dissection was then carried down to the spinous processes of L2–L3 and L5–L6 using electrocautery. The paraspinous muscles were dissected free of the posterior bony elements and displaced laterally. The fusion bed was then cleaned of soft tissue using a Cobb elevator. Decortication of the spinous process, lamina, articular surfaces and processes, and the facet joints was performed using a high-speed pneumatic bur until punctuate bleeding was observed. The procedure was repeated on the contralateral side. Each fusion spanned one motion segment, with two intervening vertebral levels left undisturbed, using the vertebrae L2–L6. A muscle tissue barrier was maintained between the fusion sites. No internal fixation was used. Autograft bone consisted of corticocancellous iliac crest bone harvested at the time of surgery. The iliac crest was cleaned, trimmed, cut into matchstick pieces and placed across the fusion bed. Although all animals were of uniform size and body weight, care was taken to place the same volume of bone graft in each fusion site. The bone graft volume was 4 cm3 per fusion site and is consistent with our previous use of this model [19,20]. In order to reproducibly locate the fusion sites for daily ultrasound treatment, a surgical staple was placed at the top of the spinous processes of L2 and L5 before closure. After closing the wound, each treatment location was marked directly above each fusion site on the skin by permanent tattoo. With a locating device in place over the sites, postoperative radiographs were then taken to verify proper placement of the skin marks directly over the fusion bed. Animals were administered antibiotics (0.08 ml/kg body weight; Flo-cillin; Fort Dodge Laboratories, Fort Dodge, IA) for 7 days after surgery. Butorphanol tartrate (0.01 mg/ kg body weight; Torbutrol; Fort Dodge Laboratories, Fort Dodge, IA) was given as necessary for pain control after surgery. After surgery, the animals were transferred to standard cages to restrict motion until full weight bearing was demonstrated. The animals were then transferred to 8 feet by 3 feet runs and allowed unrestricted motion. Ultrasound therapy and test device description The ultrasound therapy used (Exogen, Inc., Piscataway, NJ) is a noninvasive, FDA-approved external device indicated for accelerating healing of fresh fractures. The unit delivers a low-level acoustic pressure wave signal with an intensity of 30 milliwatts per square centimeter to the skin at the site. Because ultrasound waves cannot propagate through air, the ultrasound is coupled to the skin at the site by means of a coupling gel. The signal is composed of a burst width of 200 microseconds containing 1.5 megahertz
sine waves, with a repetition rate of one kilohertz. The round ultrasound transducer head has a diameter of 2.22 cm and emits a field with an effective geometric radiating area of 3.88 cm2. The field of acoustic energy retains a beam pattern through soft tissue without spread of the beam pattern for the first 12 cm of penetration. The depth from the site of application at the skin to the fusion bed was in the range of 5 to 7 cm. Because two levels intervened between the two experimental fusion sites, the distance between them was in the rage of 6 to 8 cm and was adequate to isolate the ultrasound treatment over the desired fusion site. Treatment consisted of one 20-minute application per day, with a transducer head centered over each unilateral fusion mass. Placebo control units were identical to the active units, except no ultrasound signal was emitted. Each unit was individually numbered and periodically tested throughout the study to ensure proper functioning. The active units provided a 20-minute ultrasound treatment identical to that used in the clinical trials of tibia shaft and distal radius fractures and that was proven to significantly accelerate fracture and bone healing [16–18]. Beginning on the third postoperative day, one site received ultrasound stimulation for 20 minutes per day, whereas the other site received no stimulation from a placebo unit. Ultrasound treatment was randomly assigned to the fusion site in order to avoid possible positional bias. Two animals (one at each postoperative time period) received no stimulation to assess the potential effect on the placebo control fusion site from the ultrasound used on the other fusion site. These animals were compared with control sites from stimulated animals and previous results obtained using this spine fusion model [19,20]. Radiographic evaluations Routine anteroposterior and lateral radiographs were taken immediately after surgery to ensure proper surgical placement. Additional radiographs were obtained of all surviving animals at 6 and 12 weeks after surgery. At the end of the study period, animals were sacrificed using an intravenous injection of concentrated sodium pentobarbital (1 ml/4.5 kg body weight; Beuthanasia-D Veterinary Euthanasia Solution; Schering-Plough Animal Health, Kenisworth, NJ). The operative section of the spine was excised en bloc, sprayed with physiologic saline, and wrapped in saline-soaked towels. After retrieval, CT and MRI studies were done using previously described protocols [19,20]. Briefly, CT images of 3-mm-thick sections were obtained in 3-mm increments, allowing for continuous image acquisition. Axial images of the posterior elements were obtained using both standard and detail processing algorithms. Sagittal T1-weighted and T2 variable echo multiplantar (VEMP) MRI studies were obtained of 4-mm-thick sections in 4-mm increments using a 1.5 Tesla scanner. Axial T1-weighted MRI studies of the posterior elements were obtained of 3-mm-thick sections and in 3-mm increments.
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
249
A previously developed grading scale was used to evaluate incorporation of graft material and development of bony fusion [19,20]. CT and MRI were graded using a 0 to 5.5 scale where 0 no new bone formation, 1 minimal new bone formation around the spinous processes only, 2 continuous new bone formation adjacent to the spinous process but not extending to the facet joint, 3 moderate new bone formation throughout the entire fusion site but not continuous with the facets or within the graft material, 4 moderate new bone formation continuous within the graft materials and near the facet joint and spinous processes but with the facets not fused, and 5 extensive new bone formation and continuity throughout the fusion site including facet joint (5a complete fusion, weighted as 5 points; 5b complete fusion with advanced remodeling, weighted as 5.5 points). Each unilateral mass of each fusion segment was graded separately for each animal. Two observers, blinded to treatment group and time in situ, evaluated each fusion site. The grades from each observer were combined, and the mean and standard deviation between the two grades was obtained. Interobserver error in radiographic grading was within 0.5 grade.
were dehydrated in graduated ethyl alcohol solutions from 70% to 100%, each for 24 hours. The specimens were then placed in methyl methacrylate monomer, allowed to polymerize and cut on a high-speed, water-cooled diamond saw (CS600-A Mark V Laboratories, Inc., East Grandby, CT) into axial sections approximately 750 to 1,000 m thick. A total of 15 to 20 axial sections were obtained from each fusion mass. The final number of sections available was dependent on the size and length of the vertebral segments involved. These sections were mounted on acrylic slides and ground to a 50-m thickness using a metallurgical grinding wheel. All sections were stained with basic fuchsin and used to determine static parameters of fusion. Each section was evaluated for the presence of new bone or fibrous tissue, bone graft incorporation, remodeling of the fusion mass and continuity of the fusion mass. The sections were evaluated by observers blinded to the time in situ and the treatment group. Histologic fusion was defined as the presence of new bone formation in continuity with both of the facets joints and across the fusion bed with evidence of graft incorporation and remodeling.
Mechanical testing
All mechanical testing data were initially screened for statistical outliers and then pooled. Data screening was performed by identifying the maximum and minimum value for each variable and calculating the z-score for both values. The z-score for a data point is defined as the difference between the data point value and the variable mean divided by the variable standard deviation. A statistical outlier was defined as a data point having a z-score equal to or greater than 3.0 [21]. After outliers were eliminated, analysis of variance (ANOVA) was used to determine the effect of animal-to-animal variation or location of the fusion mass. Thereafter, all data were pooled and ANOVA was used to examine the effects of ultrasound treatment and postoperative time period. The mean and standard deviation of the radiographic grades and torsional stiffness were calculated. When possible, paired t tests were performed. Radiographic grading was evaluated by the KruskalWallis one-way ANOVA and multiple comparisons to examine the effects of treatment group and time in situ. All statistical analysis was performed with statistical analysis software for a personal computer (1990 BMDP; BMDP Statistical Software, Inc., Los Angeles, CA). The incidence of a radiographic, mechanical (by means of manual palpation) or histologic fusion was evaluated at each evaluation period by a two-by-two chi-squared test to determine if the ultrasound treatment significantly affected the fusion rate at each evaluation period. A P value .05 or less indicated a significant effect.
After radiographic analysis, the spine was stripped of soft tissues. Each spinal fusion segment was isolated and prepared for mechanical testing by sectioning the segments through the vertebral foramen using a watercooled diamond saw producing two segments per site. The stability of each unilateral fusion mass was subjectively evaluated by manual examination of the involved segments. A zero- to two-point scale was used to characterize each mass as to whether the site was solidly fused without any motion between the segments (a grade of two), partially fused with slight but significantly limited motion between the segments (grade of one) or not fused (grade of zero) having motion similar to the adjacent nonoperative segments of the spine. Only a fusion mass with a grade of two was considered completely fused. The stiffness of the fusion mass was then determined by nondestructive mechanical testing. Each unilateral fusion site was secured in a specially designed test fixture with polymethymethacrylate (Fastray; Harry J. Bosworth Company, Skokie, IL). The specimens were nondestructively tested in torsion on a closed-loop hydraulic test machine (Model 810; MTS Systems Corporation, Minneapolis, MN) operated in stroke control, at a constant displacement rate of 47.5 degrees per minute. Torsional stiffness was calculated by determining the linear portion of the torque-angular rotation curve between 5 and 10 degrees of rotation.
Statistical analysis
Histology Each fusion segment was prepared for histologic analysis using undecalcified techniques. After mechanical testing, the specimens were fixed by immersion in 10% formalin solution for at least 24 hours. After fixation, the specimens
Results There were no postoperative complications. All animals were ambulatory within 48 hours of surgery. Ultrasound
250
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
Fig. 1. Axial computed tomography (CT) sections from (A) the placebo control fusion site and (B) the fusion site treated with daily low-intensity pulsed ultrasound at 6 weeks after surgery from the same animal. Axial CT sections from (C) the placebo control fusion site and (D) the fusion site treated with daily lowintensity pulsed ultrasound at 12 weeks after surgery from the same animal. In nontreated control animals new bone formation was observed primarily at the 12-week time period. Early new bone bridging and graft consolidation was observed as early as 6 weeks after surgery in the ultrasound-treated sites.
therapy was successfully applied at least six times within a 7-day period throughout the study. Gross pathologic examinations at sacrifice revealed no adverse effects related to the ultrasound treatment. Overall, ANOVA of all radiographic mechanical and histologic data demonstrated no significant difference that could be attributed to the fusion level or interanimal bias. Therefore, all data were pooled by treatment and time in situ for further analysis. Radiographic evaluation Radiographic differences were subtle within evaluation periods for the two treatment groups and were difficult to appreciate from plain radiographs and MRI scans. CT scans were helpful in determining the presence of new bone development and differences between ultrasound stimulated and nonstimulated sites. CT scans demonstrated more advanced, dense new bone formation and graft continuity with the host bone at the facets in ultrasound-treated sites compared with controls at both time periods. At 12 weeks after surgery, the degree of graft incorporation and amount of new bone formation and remodeling was advanced in ultrasound-treated sites compared with control sites. Fig. 1 demonstrates the typical appearance of a CT slice through the fusion mass in control and ultrasound-treated sites at 6 and 12 weeks after surgery. Statistical analysis demonstrated that the right and left radiographic grades could be pooled and a single mean grade for each fusion mass was obtained. At 6 weeks after sur-
gery, both the ultrasound-treated sites and control sites demonstrated evidence of new bone formation throughout the fusion mass. However, complete fusion (radiographic grade greater than four) was observed in only one of six control sites (16%; Table 2). Three of six ultrasound-treated sites (50%) had a radiographic grade greater than four at 6 weeks after surgery. By 12 weeks, the new bone mass and density increased in most sites and was often continuous with the facets and spinous process. Seven of nine control sites were considered fused (78%). In comparison, all seven sites that were treated with ultrasound were fused by 12 weeks. However, differences in fusion rate resulting from time or experimental treatment were not statistically significant. The mean radiographic grades for both ultrasound-treated and control sites increased from 6 to 12 weeks after surgery although not statistically significant. The mean radiographic grades for the ultrasound-treated sites at both 6 and 12 weeks (4.10.9 and 5.10.2, respectively) were greater than corresponding control sites (3.90.7 and 4.80.5). By 12 weeks, a complete fusion was observed in all ultrasound-treated sites. Within-animal comparisons of ultrasound-treated and control sites often demonstrated marked differences at both time periods. However, the differences in the mean radiographic grades for each treatment group were not statistically significant. Mechanical testing At 6 weeks, manual palpation of the fusion sites indicated that three of six ultrasound-treated sites (50%) were solidly fused (grade of two), whereas two of the six control
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
251
Table 2 Presence of radiographic fusion by treatment group Rate of radiographic fusion1/number of sites (%)
Treatment group 6 weeks 12 weeks
Ultrasound Control Ultrasound Control
3/6 (50) 1/6 (16) 7/7 (100) 7/9 (78)
P value (chi squared)
Mean radiographic grade ( standard deviation, N·m/degree) 4.1 0.9 3.9 0.7 5.1 0.2 4.8 0.5
NS NS
P value (t test) NS NS
1
Radiographic fusion was defined as a radiographic grade greater than four per fusion mass. NS not statistically significant.
sites (33%) were solidly fused (Table 3). At 12 weeks, six of seven ultrasound treated sites (86%) were solidly fused compared with three of nine control sites (33%). Five of the remaining six control sites were partially fused at 12 weeks. Chi-squared analysis demonstrated that ultrasound significantly improved the fusion rate at 12 weeks (P.05) compared with control sites, although not at 6 weeks (Table 3). Statistical analysis demonstrated that the right and left stiffness values could be pooled, and a single mean for each fusion mass was obtained. A stiffness measurement from one animal was found to be an outlier (z3.0) and was eliminated from the analysis. The stiffness of ultrasoundtreated sites increased from 6 to 12 weeks. Change in stiffness with time was not statistically significant. Overall, the mechanical stiffness at 6 and 12 weeks after surgery of fusion sites treated with ultrasound (53.122.3 and 59.918.9 Nm/degrees, respectively) was greater when compared with the control sites (51.820.4 and 38.416.1 Nm/degrees; Table 3). The 12-week ultrasound-treated sites had the greatest mean torsional stiffness, which was approximately one third greater than the mean stiffness of the 12-week control sites (Table 3). The difference in mean stiffness was statistically significant (P.0282) at 12 weeks, although not at 6 weeks. The decrease in torsional stiffness from 6 weeks to 12 weeks in control sites was not statistically significant, and the difference was primarily the result of one control site that exhibited little bony healing and mechanical stiffness.
rity in animal-to-animal and treatment group comparisons. Qualitative observations included greater maturity, new bone formation and advanced graft incorporation in sites treated with ultrasound when compared with control sites at both time periods. From 6 to 12 weeks, the frequency of new bone formation, bone graft incorporation and remodeling increased in both groups (Fig. 2). At 6 weeks postoperative, evidence of histologic fusion, defined as new bone formation, graft remodeling and continuity of bone with the facet joints and across the fusion bed, was observed in two of six ultrasound-treated sites (Table 4). No control site demonstrated a complete histologic fusion at 6 weeks after surgery. At 12 weeks after surgery, a complete histologic fusion demonstrating new bone and bone graft remodeling was observed in seven of seven (100%) ultrasound-treated sites. Although histologic fusion was present in four of nine control sites, remodeling was not observed or was not as advanced in control sites compared with ultrasound-treated sites at 12 weeks. Comparisons of the histologic fusion rates demonstrated that the incidence of bony fusion in ultrasound-treated sites was significantly greater than control sites at 12 weeks (P.005, Table 4), although the histologic fusion rate was not significantly different at 6 weeks. Within-animal comparisons demonstrated that graft incorporation with the host bone and new bone maturity was more advanced in ultrasound-treated sites compared with control sites.
Histologic evaluation
Discussion
Histologic evaluation revealed different stages of bone formation, bone graft incorporation and fusion mass matu-
Failure or delay in spinal fusion can result in significant morbidity. The cause of failure is usually multifactorial but
Table 3 Mechanical testing results by treatment group Rate of mechanical fusion1/number of sites (%)
Treatment group 6 weeks 12 weeks 1
Ultrasound Control Ultrasound Control
3/6 (50) 2/6 (33) 6/7 (86) 3/9 (33)
Manual palpation score of two of two possible points. NS not statistically significant.
P value (chi squared) NS .05
Mean stiffness ( standard deviation N·m/degree) 53.1 22.3 51.8 20.4 59.9 18.9 38.4 16.1
P value (t test) NS .0282
252
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
Fig. 2. Histologic sections from within the same animal at 6 weeks after surgery from (A) the placebo control fusion site. The residual anterior portion of the caudal articular process (CAP) and the cranial articular process (cap) and their articulating surfaces (*). Bone graft (G) is present posterior to host bone bed, but there is little new bone formation. A fibrous tissue layer (f) lies between the host bone, the bone graft pieces. (B) The fusion site within the same animal treated with daily low-intensity pulsed ultrasound. The anterior host bone bed is located just to the left of the image (arrows) and is continuous with new bone (nb) and bone graft (G). The fusion mass is continuous with the spinous process (borders the top of the image). Twelve-week histologic sections from within the same animal from (C) the placebo control fusion site. Viable bone graft (G) is present with evidence of early new bone formation and mineralizing cartilage seams (c) that have begun to bridge the bone graft and the anterior host bone bed. (D) A fusion site from within the same animal treated with daily lowintensity pulsed ultrasound showing nearly complete fusion with new bone (nb). New bone bridges from the graft to the anterior portion of the cranial articular process (cap). A cartilage seam (c) joins bone graft (G) to the caudal articular process.
often includes problems in new bone formation and healing. The radiographic, mechanical and histologic results of this study suggest that the use of a low-intensity, pulsed ultrasound device in this model may have an accelerative effect on spinal fusion. At 6 and 12 weeks after surgery, a higher incidence of fusion was observed in ultrasound-treated sites compared with control sites, although only significant at 12 weeks. The qualitative radiographic and histologic observations of ultrasound-treated sites at 6 weeks was similar to those observed in control sites at 12 weeks. However, there were no statistically significant differences in radiographic grades or radiographic fusion rate at either time period attributed to the use of ultrasound stimulation. Comparison of the radiographic, mechanical and histologic results from control sites in ultrasound-stimulated animals and the nonstimulated animals indicates the stimulatory effect is localized to the area directly below the ultrasound transducer and may necessitate the use of multiple transducers for maximum effect in a multilevel spinal fusion. The lack of significant improvement of the ultrasoundtreated fusion sites compared with control sites at 6 weeks after surgery may be indicative of the early healing stage of the fusion mass and, perhaps, the effect of ultrasound ther-
apy was below the detectable limits of the analysis methods used in this study. A large variance within the treatment groups was observed in the mean mechanical stiffness results at 6 weeks (standard deviation was approximately 50% of the mean stiffness). Interestingly, statistically significant improvement in fusion was not detected by radiographic means (plain X-ray, CT and MRI) at either the 6- or 12week evaluations. Only the manual palpation, stiffness measurements and histologic examination found evidence of a significant increase in fusion rate at 12 weeks (P.05 in all comparisons). This is in agreement with the clinical difficulty in assessing fusion success by radiographic means Table 4 Histologic analysis of fusion by treatment group Rate of histologic fusions/number of sites (%)
Treatment group 6 weeks 12 weeks
Ultrasound Control Ultrasound Control
2/6 (33) 0/6 (0) 7/7 (100) 4/9 (44)
NS not statistically significant.
P value (chi squared) NS .005
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
alone, particularly during early postoperative assessments. Despite increased new bone formation compared with the control sites, remodeling and incorporation of the bone graft with the host bone did not yet appear to be complete in radiographic and histologic evaluations at 6 weeks in the ultrasound-treated sites. However, at 12 weeks, incorporation of bone graft with the new bone and the host bone was observed more frequently in the ultrasound-treated sites resulting in statistically significant improvement in fusion rate compared with nontreated control sites (P.005). Incorporation and remodeling of the fusion mass with the host bone bed is the essential occurrence in increasing the rigidity of the fused construct using the posterior noninstrumented surgical model of spine fusion. The results obtained for autogenous graft control sites in this study were in agreement with those reported in our previous investigations [19,20]. A 33% radiographic and mechanical fusion rate at 6 weeks and 33% radiographic fusion rate and 66% mechanical fusion rate at 12 weeks for autograft controls has been reported previously [20]. Although some new bone formation and graft incorporation occurs at 6 and 12 weeks after surgery in sites treated with bone graft alone, a complete fusion with extensive remodeling of the bone graft is generally not observed until 26 weeks after surgery [19,20]. In the present study a 100% histologic fusion rate was observed in all ultrasound-stimulated sites at 12 weeks after surgery and an 86% fusion rate was observed upon manual palpation, suggesting an acceleration of fusion over nontreated autograft control sites. The exact mechanism of action of the ultrasound device is unknown. It has been theorized that ultrasound pressure waves may mediate biologic activity by mechanical deformation of the cell membrane or extracellular matrix, indirectly by an electrical effect caused by cell deformation or by modifying the streaming potential in the extracellular fluid [16]. Studies have also shown effects on blood flow and vascularization as well as significant effects on cellular function [10–15]. However, studies in animals and humans have shown repeatedly its effectiveness in accelerating the bone repair process [3,8–11,16–18]. In a study by Glazer et al. [22], the rabbit posterolateral intratransverse process model was used to examine the effect of low-intensity pulsed ultrasound on spine fusion. Using radiographic and mechanical methods, they found a statistically significant increase in fusion rate from 65% to 93% with the use of ultrasound over nontreated controls at 6 weeks after surgery. A statistically significant increase in fusion segment stiffness was also observed with the use of ultrasound therapy. Similar to the present study, Glazer et al. found a qualitative increase in the new bone formation in histologic sections from ultrasound-treated sites compared with nontreated controls. Future clinical application of ultrasound therapy could be aimed at improving fusion rates by means of a noninvasive treatment where bone healing may be compromised, for example, in patients with a history of tobacco use. Pseudoar-
253
throsis after spinal fusion occurs at least four times more often in patients who smoke compared with patients who do not smoke [23–25]. Ultrasound stimulation was shown to substantially reduce the healing time of patients with tibia and distal radius fracture and significantly reduce the incidence of delayed union and nonunion [17]. The healing time of smoking patients in these studies with ultrasound stimulation was found to be equivalent to the healing time of placebo control nonsmokers [17]. In summary, the results of this study have demonstrated that the use of a low-intensity, pulsed ultrasound device may result in an earlier, more consistent spinal fusion. Although not statistically significant at 6 weeks, the histologic findings and mechanical analysis at 12 weeks advance the concept that low-intensity pulsed ultrasound can modulate bone healing, bone graft incorporation and, potentially, the stimulation of fusion. The limitations of the present study, specifically the small sample size of each experimental group and the lack of statistically significant radiographic results at either 6 or 12 weeks, prevent drawing strong conclusions regarding the efficacy of ultrasound in accelerating canine noninstrumented posterior spine fusions. Future studies to characterize and optimize the effects of the ultrasound signal may further demonstrate that ultrasound therapy may be a useful adjunct in spinal fusion surgery.
References [1] Cotler JM, Star AM. Complications of spinal fusions. In: Cotler JM, Cotler HB, eds. Spinal fusion: science and technique. New York: Springer-Verlag, 1990;361–87. [2] Muschler GF, Lane JM, Dawson EG. The biology of spinal fusion. In: Cotler JM, Cotler HB, eds. Spinal fusion: science and technique. New York: Springer-Verlag, 1990;9–21. [3] Duarte LR. The stimulation of bone growth by ultrasound. Arch Orthop Trauma Surg 1983;101:153–9. [4] Wolff J. The law of bone remodeling [Das Gesetz der Transformation der Knochen]. Originally published by Verlag von August Hirshwald, Berlin, 1869. English translation by Maquet P and Fulong R, published by Springer Verlag, Berlin, 1982. [5] Buckley MJ, Banes AJ, Levin LG, et al. Osteoblasts increase their rate of division and align in response to cyclic, mechanical tension in vitro. Bone Min 1988;4:225–36. [6] Goodship AE, Kenwright J. The influence of induced micromovement upon the healing of experimental tibial fractures. J Bone Joint Surg Br 1985; 67:650–5. [7] Rubin CT, Langon LE. Regulation of bone formation by applied dynamic loads. J Bone Joint Surg Am 1984;66:397–402. [8] Pilla AA, Mont MA, Nasser PR, et al. Non-invasive low-intensity pulsed ultrasound accelerates bone healing in the rabbit. J Orthop Trauma 1990;4:246–53. [9] Wang SJ, Lewallen DG, Bolander ME, et al. Low intensity ultrasound treatment increases strength in a rat femoral fracture model. J Orthop Res 1994;12:40–7. [10] Yang KH, Parvizi J, Wang SJ, et al. Exposure to low-intensity ultrasound increases aggrecan gene expression in a rat femur fracture model. J Orthop Res 1996;14:802–9. [11] Parvizi J, Wu CC, Lewallen DG, Greenleaf JF, Bolander ME. Lowintensity ultrasound stimulates proteoglycan synthesis in rat chondrocytes by increasing aggrecan gene expression. J Orthop Res 1999;17: 488–94.
254
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
[12] Ryaby JT, Bachner EJ, Bendo JA, Dalton PF, Tannenbaum S, Pilla AA. Low intensity pulsed ultrasound increases calcium incorporation in both differentiating cartilage and bone cell cultures. Trans Orthop Res Soc 1989;14:15. [13] Ryaby JT, Duarte-Alves P, Pilla AA. Ultrasound and static mechanical force affect second messenger activity in bone cell cultures. Trans Orthop Res Soc 1990;15:131. [14] Ryaby JT, Matthew J, Duarte-Alves P. Low intensity pulsed ultrasound affects adenylate cyclase and transforming growth factor-beta synthesis in osteoblastic cells. Trans Orthop Res Soc 1992;17:590. [15] Ryaby JT, Matthew J, Pilla AA. Low intensity pulsed ultrasound modulates adenylate cyclase activity and transforming growth factor beta synthesis. In: Brighton CT, Pollock SR, eds. Electromagnetics in biology and medicine. San Francisco: San Francisco Press, 1991:95–100. [16] Heckman JD, Ryaby JP, McCabe J, Frey JJ, Kilcoyne RF. Acceleration of tibial fracture-healing by non-invasive, low-intensity pulsed ultrasound. J Bone Joint Surg Am 1994;76:26–34. [17] Cook SD, Ryaby JP, McCabe J, Frey JJ, Heckman JD, Kristiansen TK. Acceleration of tibia and distal radius fracture healing in patients who smoke. Clin Orthop 1997;337:198–207. [18] Kristiansen TK, Ryaby JP, McCabe J, Frey JJ, Roe LR. Accelerated
[19]
[20]
[21] [22] [23] [24]
[25]
healing of distal radial fractures with the use of specific, low-intensity ultrasound. A multicenter, prospective, randomized, double-blind, placebo-controlled study. J Bone Joint Surg Am 1997;79:961–73. Cook SD, Dalton JE, Prewett AB, Whitecloud TS. In vivo evaluation of demineralized bone matrix as a bone graft substitute for posterior spinal fusion. Spine 1995;20:877–86. Cook SD, Dalton JE, Tan EH, Whitecloud TS, Rueger DC. In vivo evaluation of recombinant human osteogenic protein (rhOP-1) implants as a bone graft substitute for spinal fusions. Spine 1994;19:1655–63. Rosner B. Fundamentals of biostatistics, 3rd ed. Boston: PWS-Kent Publishing Company, 1990:200–6. Glazer PA, Heilman MR, Lotz JC, Bradford DS. Use of ultrasound in spinal arthrodesis. A rabbit model. Spine 1998;23:1142–8. Blumenthal SL, Baker J, Dossett A, Selby DK. The role of anterior lumbar fusion for internal disc disruption. Spine 1988;13:566–9. Brown CW, Orme TJ, Richardson HD. The rate of pseudoarthrosis (surgical nonunion) in patients who are smokers and patients who are nonsmokers: a comparison study. Spine 1986;11:942–3. Hanley EN, Levy JA. Surgical treatment of isthmic lumbosacral spondylolisthesis. Analysis of variables influencing results. Spine 1989;14:48–50.
Low-intensity pulsed ultrasound improves spinal fusion Stephen D. Cook, PhD*, Samantha L. Salkeld, MSE, Laura Popich Patron, BSE, John P. Ryaby, BSE, Thomas S. Whitecloud, III, MD Department of Orthopaedic Surgery. SL32, Tulane University Medical School, 1430 Tulane Avenue, New Orleans, LA 70112, USA Received 22 September 2000; revised 30 October 2000; second revision 22 February 2001; accepted 12 May 2001
Abstract
Background context: Increasing the incidence of solid bony fusion is a primary goal in spine surgery. Daily low-intensity pulsed ultrasound therapy has been shown to improve and accelerate the bone healing process. Purpose: The purpose of this study was to evaluate the efficacy of daily low-intensity pulsed ultrasound therapy to improve the rate and quality of spinal fusion. Study design: Canine fusion model prospective study. Patient sample: Fourteen adult male dogs were used. Outcome measures: Radiographic grading of plain films, computed tomography (CT) and magnetic resonance imaging (MRI), gross palpation, torsional stiffness and histologic grading were used to determine the presence or absence of fusion. Methods: Posterior noninstrumented bilateral fusions were evaluated at the L2–L3 and L5–L6 levels. Treatment with low-intensity pulsed ultrasound for 20 minutes per day over the fusion site (stimulated) was compared with fusion sites that received no stimulation (nontreated controls) at 6 and 12 weeks after surgery. Plain film radiographs, CT and MRI, mechanical torsion testing and histologic examination were performed. Results: At 6 weeks, ultrasound treated sites were more frequently fused compared with nontreated controls, although the difference in fusion rate was not statistically significant. At 12 weeks after surgery complete radiographic and histologic fusion occurred in 100% of ultrasound-treated sites. In the nontreated control sites 78% had achieved complete radiographic fusion and 44% had complete histologic fusion. Compared with control sites, the histological and mechanical fusion rate was significantly greater in ultrasound-treated sites (P.05) at 12 weeks. A statistically significant increase in mechanical stiffness in ultrasound-treated sites was also found at 12 weeks after surgery. Conclusions: Low-intensity pulsed ultrasound therapy may be a useful means to ensure successful spine fusion. © 2001 Elsevier Science Inc. All rights reserved.
Keywords:
Ultrasound; Spinal fusion; Bone formation; Histology; Mechanical testing
Introduction Posterior spinal fusion is the most common procedure for spinal stabilization [1]. Improved mechanical stability can help relieve symptoms in patients with degenerative arthritis of the spine, spinal deformity or instability. Posterior spinal fusion generally includes onlay grafting of autogenous corticocancellous iliac crest bone after decortication of the bony surfaces of the posterior vertebral elements [2]. Failure of spinal fusion can result in significant morbidity. The incidence of pseudoarthrosis can range from 5% to This paper was supported by a research support grant from Exogen, Inc. * Corresponding author. Tel.: (504) 588-2273; fax: (504) 584-2722. E-mail address: [email protected] (S.D. Cook)
40% in a large series, depending on many factors, including whether instrumentation is used, the number of levels fused, surgical technique, smoking history, age and method of evaluation [1]. The clinical consequences of pseudoarthrosis can be significant and include pain, progression of curvature in scoliosis surgery and unpredictable results of operations for segmental instability. Increasing the incidence and rapidity of solid bony fusion is a primary goal in spine surgery. Ultrasound is a form of mechanical energy that is transmitted into the body as high-frequency acoustical pressure waves [3]. Low-intensity ultrasound can be used to produce micromechanical stresses and strains in tissues. The micromechanical force applied to bone tissue is less than 10 mg,
1529-9430/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S1529-9430(01)00 0 8 6 - 9
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
which can penetrate to depths of greater than 10 cm despite the attenuation effects of overlying soft tissue. The intensity of the ultrasound (less than 30 milliwatts per square centimeter) is similar to that used in diagnostic imaging. Ultrasound promotes bone formation in a manner comparable to that first postulated and described by Wolff’s law [4]. The physiological response of bone to mechanical force, including effects on bone healing of early weight bearing, dynamic compression techniques and impact loading, have been described in the literature [5–7]. In placebo-controlled animal studies, Duarte [3] reported that daily low-intensity pulsed ultrasound therapy accelerated bone healing in both fibular osteotomy and femoral defect models. Pilla et al. [8] reported a statistically significant increase in the mechanical strength of low-intensity ultrasound-treated osteotomy sites versus control sites. Wang et al. [9] and Yang et al. [10] reported that use of low-intensity ultrasound produced a statistically significant increase in fracture callus amount, stiffness and strength in a bilateral, placebo-controlled femoral fracture model that used an intramedullary metal rod for fixation. Molecular biology experiments indicate that specific cartilage genes, including aggrecan mRNA and DNA binding proteins, are expressed when exposed to ultrasound stimulation [10]. Recent research with real-time studies, using fluorescent markers, have demonstrated increases in cellular calcium flux [11]. In addition, low-intensity pulsed ultrasound produces dramatic and immediate increases in the microvasculature and blood flow at the fracture site. Ryaby et al. [12–15] have demonstrated that ultrasound has multifunctional cellular effects, such as increasing 45Ca uptake, modulating transforming growth factor–beta (TGF-) synthesis and parathyroid hormone (PTH) response in mesenchymal or osteoblast cells. These reports suggest a direct effect on bone formation and resorption, noncollagenous protein synthesis and expression of collagen phenotypes and are consistent with the observation of increased cartilage and advanced endochondral ossification. A pulsed, low-intensity ultrasound device has been evaluated in Food and Drug Administration (FDA)–approved multicenter, prospective, randomized, double-blind and placebo-controlled clinical trials (Sonic Accelerated Fracture Healing System; Exogen, Inc., Piscataway, NJ). The ultrasound therapy was administered daily at home for 20 minutes per day. A statistically significant decrease in clinical and radiographic healing time was reported for diaphysis bone (tibial shaft) fractures (active device, 96 days; placebo device, 154 days [P.0001) [16]. The healing time in tibial shaft fractures was found to be reduced 41% in smoking patients and 26% in nonsmokers [17]. In a second clinical trial of the ultrasound device for metaphyseal (posterior displaced distal radius) fractures, the healing time was also reduced 38% (active device, 61 days; placebo device, 98 days [P.0001) [18]. The healing times in distal radius fractures were reduced 51% in smokers and 34% in nonsmokers [17]. These studies confirmed the results obtained in laboratory
247
and animal studies of the stimulatory effect on fracture healing of low-intensity pulsed ultrasound therapy. The purpose of this study was to determine the efficacy of daily pulsed, low-intensity ultrasound therapy in improving the rate of spinal fusion using a canine noninstrumented posterior autogenous bone graft onlay model. Materials and methods Experimental design Posterior spinal fusions were evaluated in 14 adult male dogs. Each animal was fused bilaterally at the L2–L3 and L5–L6 levels with autogenous bone graft harvested from the iliac crests. In 11 animals, one site received active ultrasound stimulation for 20 minutes per day, whereas the other site received no stimulation (Table 1). To evaluate the effect of fusion level, two animals received no stimulation at either site, whereas one animal received ultrasound stimulation at both sites. Six animals were sacrificed at 6 weeks and eight animals at 12 weeks after surgery. The quality of fusion, new bone formation and bone graft incorporation were evaluated radiographically using plain films, computed tomography (CT) and magnetic resonance imaging (MRI). The postretrieval fusion masses were further evaluated by manual manipulation and mechanical testing in torsion and analyzed histologically for evidence of bony fusion using nondecalcified techniques. Animal model and surgical procedure Skeletally mature adult mongrel dogs weighing approximately 24 to 30 kg were randomly divided into two groups corresponding to implantation periods of 6 weeks (six animals) and 12 weeks (eight animals; Table 1). Before surgery, the animals were allowed to adapt to vivarial conditions for 2 weeks and were treated with antibiotics and antiparasitics to ensure uniform good health. The animals were radiographically screened to ensure the absence of skeletal abnormalities. All surgical procedures and animal
Table 1 Experimental design: distribution of group sample sizes Evaluation period 6 weeks
Treatment group
Levels fused
Sample size
Autogenous bone graft treated with daily ultrasound therapy
L2–L3 L5–L6 Total L2–L3 L5–L6 Total
3 3 6 3 3 6
L2–L3 L5–L6 Total L2–L3 L5–L6 Total
3 4 7 5 4 9
Autogenous bone graft (control)
12 weeks
Autogenous bone graft treated with daily ultrasound therapy Autogenous bone graft (control)
248
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
care were performed according to National Institutes of Health guidelines for the care and use of laboratory animals (publication number 85-23, revision 1985) with the supervision of a licensed veterinarian. Using standard aseptic techniques, a midline skin incision was made from levels L1 to L7. The dissection was then carried down to the spinous processes of L2–L3 and L5–L6 using electrocautery. The paraspinous muscles were dissected free of the posterior bony elements and displaced laterally. The fusion bed was then cleaned of soft tissue using a Cobb elevator. Decortication of the spinous process, lamina, articular surfaces and processes, and the facet joints was performed using a high-speed pneumatic bur until punctuate bleeding was observed. The procedure was repeated on the contralateral side. Each fusion spanned one motion segment, with two intervening vertebral levels left undisturbed, using the vertebrae L2–L6. A muscle tissue barrier was maintained between the fusion sites. No internal fixation was used. Autograft bone consisted of corticocancellous iliac crest bone harvested at the time of surgery. The iliac crest was cleaned, trimmed, cut into matchstick pieces and placed across the fusion bed. Although all animals were of uniform size and body weight, care was taken to place the same volume of bone graft in each fusion site. The bone graft volume was 4 cm3 per fusion site and is consistent with our previous use of this model [19,20]. In order to reproducibly locate the fusion sites for daily ultrasound treatment, a surgical staple was placed at the top of the spinous processes of L2 and L5 before closure. After closing the wound, each treatment location was marked directly above each fusion site on the skin by permanent tattoo. With a locating device in place over the sites, postoperative radiographs were then taken to verify proper placement of the skin marks directly over the fusion bed. Animals were administered antibiotics (0.08 ml/kg body weight; Flo-cillin; Fort Dodge Laboratories, Fort Dodge, IA) for 7 days after surgery. Butorphanol tartrate (0.01 mg/ kg body weight; Torbutrol; Fort Dodge Laboratories, Fort Dodge, IA) was given as necessary for pain control after surgery. After surgery, the animals were transferred to standard cages to restrict motion until full weight bearing was demonstrated. The animals were then transferred to 8 feet by 3 feet runs and allowed unrestricted motion. Ultrasound therapy and test device description The ultrasound therapy used (Exogen, Inc., Piscataway, NJ) is a noninvasive, FDA-approved external device indicated for accelerating healing of fresh fractures. The unit delivers a low-level acoustic pressure wave signal with an intensity of 30 milliwatts per square centimeter to the skin at the site. Because ultrasound waves cannot propagate through air, the ultrasound is coupled to the skin at the site by means of a coupling gel. The signal is composed of a burst width of 200 microseconds containing 1.5 megahertz
sine waves, with a repetition rate of one kilohertz. The round ultrasound transducer head has a diameter of 2.22 cm and emits a field with an effective geometric radiating area of 3.88 cm2. The field of acoustic energy retains a beam pattern through soft tissue without spread of the beam pattern for the first 12 cm of penetration. The depth from the site of application at the skin to the fusion bed was in the range of 5 to 7 cm. Because two levels intervened between the two experimental fusion sites, the distance between them was in the rage of 6 to 8 cm and was adequate to isolate the ultrasound treatment over the desired fusion site. Treatment consisted of one 20-minute application per day, with a transducer head centered over each unilateral fusion mass. Placebo control units were identical to the active units, except no ultrasound signal was emitted. Each unit was individually numbered and periodically tested throughout the study to ensure proper functioning. The active units provided a 20-minute ultrasound treatment identical to that used in the clinical trials of tibia shaft and distal radius fractures and that was proven to significantly accelerate fracture and bone healing [16–18]. Beginning on the third postoperative day, one site received ultrasound stimulation for 20 minutes per day, whereas the other site received no stimulation from a placebo unit. Ultrasound treatment was randomly assigned to the fusion site in order to avoid possible positional bias. Two animals (one at each postoperative time period) received no stimulation to assess the potential effect on the placebo control fusion site from the ultrasound used on the other fusion site. These animals were compared with control sites from stimulated animals and previous results obtained using this spine fusion model [19,20]. Radiographic evaluations Routine anteroposterior and lateral radiographs were taken immediately after surgery to ensure proper surgical placement. Additional radiographs were obtained of all surviving animals at 6 and 12 weeks after surgery. At the end of the study period, animals were sacrificed using an intravenous injection of concentrated sodium pentobarbital (1 ml/4.5 kg body weight; Beuthanasia-D Veterinary Euthanasia Solution; Schering-Plough Animal Health, Kenisworth, NJ). The operative section of the spine was excised en bloc, sprayed with physiologic saline, and wrapped in saline-soaked towels. After retrieval, CT and MRI studies were done using previously described protocols [19,20]. Briefly, CT images of 3-mm-thick sections were obtained in 3-mm increments, allowing for continuous image acquisition. Axial images of the posterior elements were obtained using both standard and detail processing algorithms. Sagittal T1-weighted and T2 variable echo multiplantar (VEMP) MRI studies were obtained of 4-mm-thick sections in 4-mm increments using a 1.5 Tesla scanner. Axial T1-weighted MRI studies of the posterior elements were obtained of 3-mm-thick sections and in 3-mm increments.
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
249
A previously developed grading scale was used to evaluate incorporation of graft material and development of bony fusion [19,20]. CT and MRI were graded using a 0 to 5.5 scale where 0 no new bone formation, 1 minimal new bone formation around the spinous processes only, 2 continuous new bone formation adjacent to the spinous process but not extending to the facet joint, 3 moderate new bone formation throughout the entire fusion site but not continuous with the facets or within the graft material, 4 moderate new bone formation continuous within the graft materials and near the facet joint and spinous processes but with the facets not fused, and 5 extensive new bone formation and continuity throughout the fusion site including facet joint (5a complete fusion, weighted as 5 points; 5b complete fusion with advanced remodeling, weighted as 5.5 points). Each unilateral mass of each fusion segment was graded separately for each animal. Two observers, blinded to treatment group and time in situ, evaluated each fusion site. The grades from each observer were combined, and the mean and standard deviation between the two grades was obtained. Interobserver error in radiographic grading was within 0.5 grade.
were dehydrated in graduated ethyl alcohol solutions from 70% to 100%, each for 24 hours. The specimens were then placed in methyl methacrylate monomer, allowed to polymerize and cut on a high-speed, water-cooled diamond saw (CS600-A Mark V Laboratories, Inc., East Grandby, CT) into axial sections approximately 750 to 1,000 m thick. A total of 15 to 20 axial sections were obtained from each fusion mass. The final number of sections available was dependent on the size and length of the vertebral segments involved. These sections were mounted on acrylic slides and ground to a 50-m thickness using a metallurgical grinding wheel. All sections were stained with basic fuchsin and used to determine static parameters of fusion. Each section was evaluated for the presence of new bone or fibrous tissue, bone graft incorporation, remodeling of the fusion mass and continuity of the fusion mass. The sections were evaluated by observers blinded to the time in situ and the treatment group. Histologic fusion was defined as the presence of new bone formation in continuity with both of the facets joints and across the fusion bed with evidence of graft incorporation and remodeling.
Mechanical testing
All mechanical testing data were initially screened for statistical outliers and then pooled. Data screening was performed by identifying the maximum and minimum value for each variable and calculating the z-score for both values. The z-score for a data point is defined as the difference between the data point value and the variable mean divided by the variable standard deviation. A statistical outlier was defined as a data point having a z-score equal to or greater than 3.0 [21]. After outliers were eliminated, analysis of variance (ANOVA) was used to determine the effect of animal-to-animal variation or location of the fusion mass. Thereafter, all data were pooled and ANOVA was used to examine the effects of ultrasound treatment and postoperative time period. The mean and standard deviation of the radiographic grades and torsional stiffness were calculated. When possible, paired t tests were performed. Radiographic grading was evaluated by the KruskalWallis one-way ANOVA and multiple comparisons to examine the effects of treatment group and time in situ. All statistical analysis was performed with statistical analysis software for a personal computer (1990 BMDP; BMDP Statistical Software, Inc., Los Angeles, CA). The incidence of a radiographic, mechanical (by means of manual palpation) or histologic fusion was evaluated at each evaluation period by a two-by-two chi-squared test to determine if the ultrasound treatment significantly affected the fusion rate at each evaluation period. A P value .05 or less indicated a significant effect.
After radiographic analysis, the spine was stripped of soft tissues. Each spinal fusion segment was isolated and prepared for mechanical testing by sectioning the segments through the vertebral foramen using a watercooled diamond saw producing two segments per site. The stability of each unilateral fusion mass was subjectively evaluated by manual examination of the involved segments. A zero- to two-point scale was used to characterize each mass as to whether the site was solidly fused without any motion between the segments (a grade of two), partially fused with slight but significantly limited motion between the segments (grade of one) or not fused (grade of zero) having motion similar to the adjacent nonoperative segments of the spine. Only a fusion mass with a grade of two was considered completely fused. The stiffness of the fusion mass was then determined by nondestructive mechanical testing. Each unilateral fusion site was secured in a specially designed test fixture with polymethymethacrylate (Fastray; Harry J. Bosworth Company, Skokie, IL). The specimens were nondestructively tested in torsion on a closed-loop hydraulic test machine (Model 810; MTS Systems Corporation, Minneapolis, MN) operated in stroke control, at a constant displacement rate of 47.5 degrees per minute. Torsional stiffness was calculated by determining the linear portion of the torque-angular rotation curve between 5 and 10 degrees of rotation.
Statistical analysis
Histology Each fusion segment was prepared for histologic analysis using undecalcified techniques. After mechanical testing, the specimens were fixed by immersion in 10% formalin solution for at least 24 hours. After fixation, the specimens
Results There were no postoperative complications. All animals were ambulatory within 48 hours of surgery. Ultrasound
250
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
Fig. 1. Axial computed tomography (CT) sections from (A) the placebo control fusion site and (B) the fusion site treated with daily low-intensity pulsed ultrasound at 6 weeks after surgery from the same animal. Axial CT sections from (C) the placebo control fusion site and (D) the fusion site treated with daily lowintensity pulsed ultrasound at 12 weeks after surgery from the same animal. In nontreated control animals new bone formation was observed primarily at the 12-week time period. Early new bone bridging and graft consolidation was observed as early as 6 weeks after surgery in the ultrasound-treated sites.
therapy was successfully applied at least six times within a 7-day period throughout the study. Gross pathologic examinations at sacrifice revealed no adverse effects related to the ultrasound treatment. Overall, ANOVA of all radiographic mechanical and histologic data demonstrated no significant difference that could be attributed to the fusion level or interanimal bias. Therefore, all data were pooled by treatment and time in situ for further analysis. Radiographic evaluation Radiographic differences were subtle within evaluation periods for the two treatment groups and were difficult to appreciate from plain radiographs and MRI scans. CT scans were helpful in determining the presence of new bone development and differences between ultrasound stimulated and nonstimulated sites. CT scans demonstrated more advanced, dense new bone formation and graft continuity with the host bone at the facets in ultrasound-treated sites compared with controls at both time periods. At 12 weeks after surgery, the degree of graft incorporation and amount of new bone formation and remodeling was advanced in ultrasound-treated sites compared with control sites. Fig. 1 demonstrates the typical appearance of a CT slice through the fusion mass in control and ultrasound-treated sites at 6 and 12 weeks after surgery. Statistical analysis demonstrated that the right and left radiographic grades could be pooled and a single mean grade for each fusion mass was obtained. At 6 weeks after sur-
gery, both the ultrasound-treated sites and control sites demonstrated evidence of new bone formation throughout the fusion mass. However, complete fusion (radiographic grade greater than four) was observed in only one of six control sites (16%; Table 2). Three of six ultrasound-treated sites (50%) had a radiographic grade greater than four at 6 weeks after surgery. By 12 weeks, the new bone mass and density increased in most sites and was often continuous with the facets and spinous process. Seven of nine control sites were considered fused (78%). In comparison, all seven sites that were treated with ultrasound were fused by 12 weeks. However, differences in fusion rate resulting from time or experimental treatment were not statistically significant. The mean radiographic grades for both ultrasound-treated and control sites increased from 6 to 12 weeks after surgery although not statistically significant. The mean radiographic grades for the ultrasound-treated sites at both 6 and 12 weeks (4.10.9 and 5.10.2, respectively) were greater than corresponding control sites (3.90.7 and 4.80.5). By 12 weeks, a complete fusion was observed in all ultrasound-treated sites. Within-animal comparisons of ultrasound-treated and control sites often demonstrated marked differences at both time periods. However, the differences in the mean radiographic grades for each treatment group were not statistically significant. Mechanical testing At 6 weeks, manual palpation of the fusion sites indicated that three of six ultrasound-treated sites (50%) were solidly fused (grade of two), whereas two of the six control
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
251
Table 2 Presence of radiographic fusion by treatment group Rate of radiographic fusion1/number of sites (%)
Treatment group 6 weeks 12 weeks
Ultrasound Control Ultrasound Control
3/6 (50) 1/6 (16) 7/7 (100) 7/9 (78)
P value (chi squared)
Mean radiographic grade ( standard deviation, N·m/degree) 4.1 0.9 3.9 0.7 5.1 0.2 4.8 0.5
NS NS
P value (t test) NS NS
1
Radiographic fusion was defined as a radiographic grade greater than four per fusion mass. NS not statistically significant.
sites (33%) were solidly fused (Table 3). At 12 weeks, six of seven ultrasound treated sites (86%) were solidly fused compared with three of nine control sites (33%). Five of the remaining six control sites were partially fused at 12 weeks. Chi-squared analysis demonstrated that ultrasound significantly improved the fusion rate at 12 weeks (P.05) compared with control sites, although not at 6 weeks (Table 3). Statistical analysis demonstrated that the right and left stiffness values could be pooled, and a single mean for each fusion mass was obtained. A stiffness measurement from one animal was found to be an outlier (z3.0) and was eliminated from the analysis. The stiffness of ultrasoundtreated sites increased from 6 to 12 weeks. Change in stiffness with time was not statistically significant. Overall, the mechanical stiffness at 6 and 12 weeks after surgery of fusion sites treated with ultrasound (53.122.3 and 59.918.9 Nm/degrees, respectively) was greater when compared with the control sites (51.820.4 and 38.416.1 Nm/degrees; Table 3). The 12-week ultrasound-treated sites had the greatest mean torsional stiffness, which was approximately one third greater than the mean stiffness of the 12-week control sites (Table 3). The difference in mean stiffness was statistically significant (P.0282) at 12 weeks, although not at 6 weeks. The decrease in torsional stiffness from 6 weeks to 12 weeks in control sites was not statistically significant, and the difference was primarily the result of one control site that exhibited little bony healing and mechanical stiffness.
rity in animal-to-animal and treatment group comparisons. Qualitative observations included greater maturity, new bone formation and advanced graft incorporation in sites treated with ultrasound when compared with control sites at both time periods. From 6 to 12 weeks, the frequency of new bone formation, bone graft incorporation and remodeling increased in both groups (Fig. 2). At 6 weeks postoperative, evidence of histologic fusion, defined as new bone formation, graft remodeling and continuity of bone with the facet joints and across the fusion bed, was observed in two of six ultrasound-treated sites (Table 4). No control site demonstrated a complete histologic fusion at 6 weeks after surgery. At 12 weeks after surgery, a complete histologic fusion demonstrating new bone and bone graft remodeling was observed in seven of seven (100%) ultrasound-treated sites. Although histologic fusion was present in four of nine control sites, remodeling was not observed or was not as advanced in control sites compared with ultrasound-treated sites at 12 weeks. Comparisons of the histologic fusion rates demonstrated that the incidence of bony fusion in ultrasound-treated sites was significantly greater than control sites at 12 weeks (P.005, Table 4), although the histologic fusion rate was not significantly different at 6 weeks. Within-animal comparisons demonstrated that graft incorporation with the host bone and new bone maturity was more advanced in ultrasound-treated sites compared with control sites.
Histologic evaluation
Discussion
Histologic evaluation revealed different stages of bone formation, bone graft incorporation and fusion mass matu-
Failure or delay in spinal fusion can result in significant morbidity. The cause of failure is usually multifactorial but
Table 3 Mechanical testing results by treatment group Rate of mechanical fusion1/number of sites (%)
Treatment group 6 weeks 12 weeks 1
Ultrasound Control Ultrasound Control
3/6 (50) 2/6 (33) 6/7 (86) 3/9 (33)
Manual palpation score of two of two possible points. NS not statistically significant.
P value (chi squared) NS .05
Mean stiffness ( standard deviation N·m/degree) 53.1 22.3 51.8 20.4 59.9 18.9 38.4 16.1
P value (t test) NS .0282
252
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
Fig. 2. Histologic sections from within the same animal at 6 weeks after surgery from (A) the placebo control fusion site. The residual anterior portion of the caudal articular process (CAP) and the cranial articular process (cap) and their articulating surfaces (*). Bone graft (G) is present posterior to host bone bed, but there is little new bone formation. A fibrous tissue layer (f) lies between the host bone, the bone graft pieces. (B) The fusion site within the same animal treated with daily low-intensity pulsed ultrasound. The anterior host bone bed is located just to the left of the image (arrows) and is continuous with new bone (nb) and bone graft (G). The fusion mass is continuous with the spinous process (borders the top of the image). Twelve-week histologic sections from within the same animal from (C) the placebo control fusion site. Viable bone graft (G) is present with evidence of early new bone formation and mineralizing cartilage seams (c) that have begun to bridge the bone graft and the anterior host bone bed. (D) A fusion site from within the same animal treated with daily lowintensity pulsed ultrasound showing nearly complete fusion with new bone (nb). New bone bridges from the graft to the anterior portion of the cranial articular process (cap). A cartilage seam (c) joins bone graft (G) to the caudal articular process.
often includes problems in new bone formation and healing. The radiographic, mechanical and histologic results of this study suggest that the use of a low-intensity, pulsed ultrasound device in this model may have an accelerative effect on spinal fusion. At 6 and 12 weeks after surgery, a higher incidence of fusion was observed in ultrasound-treated sites compared with control sites, although only significant at 12 weeks. The qualitative radiographic and histologic observations of ultrasound-treated sites at 6 weeks was similar to those observed in control sites at 12 weeks. However, there were no statistically significant differences in radiographic grades or radiographic fusion rate at either time period attributed to the use of ultrasound stimulation. Comparison of the radiographic, mechanical and histologic results from control sites in ultrasound-stimulated animals and the nonstimulated animals indicates the stimulatory effect is localized to the area directly below the ultrasound transducer and may necessitate the use of multiple transducers for maximum effect in a multilevel spinal fusion. The lack of significant improvement of the ultrasoundtreated fusion sites compared with control sites at 6 weeks after surgery may be indicative of the early healing stage of the fusion mass and, perhaps, the effect of ultrasound ther-
apy was below the detectable limits of the analysis methods used in this study. A large variance within the treatment groups was observed in the mean mechanical stiffness results at 6 weeks (standard deviation was approximately 50% of the mean stiffness). Interestingly, statistically significant improvement in fusion was not detected by radiographic means (plain X-ray, CT and MRI) at either the 6- or 12week evaluations. Only the manual palpation, stiffness measurements and histologic examination found evidence of a significant increase in fusion rate at 12 weeks (P.05 in all comparisons). This is in agreement with the clinical difficulty in assessing fusion success by radiographic means Table 4 Histologic analysis of fusion by treatment group Rate of histologic fusions/number of sites (%)
Treatment group 6 weeks 12 weeks
Ultrasound Control Ultrasound Control
2/6 (33) 0/6 (0) 7/7 (100) 4/9 (44)
NS not statistically significant.
P value (chi squared) NS .005
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
alone, particularly during early postoperative assessments. Despite increased new bone formation compared with the control sites, remodeling and incorporation of the bone graft with the host bone did not yet appear to be complete in radiographic and histologic evaluations at 6 weeks in the ultrasound-treated sites. However, at 12 weeks, incorporation of bone graft with the new bone and the host bone was observed more frequently in the ultrasound-treated sites resulting in statistically significant improvement in fusion rate compared with nontreated control sites (P.005). Incorporation and remodeling of the fusion mass with the host bone bed is the essential occurrence in increasing the rigidity of the fused construct using the posterior noninstrumented surgical model of spine fusion. The results obtained for autogenous graft control sites in this study were in agreement with those reported in our previous investigations [19,20]. A 33% radiographic and mechanical fusion rate at 6 weeks and 33% radiographic fusion rate and 66% mechanical fusion rate at 12 weeks for autograft controls has been reported previously [20]. Although some new bone formation and graft incorporation occurs at 6 and 12 weeks after surgery in sites treated with bone graft alone, a complete fusion with extensive remodeling of the bone graft is generally not observed until 26 weeks after surgery [19,20]. In the present study a 100% histologic fusion rate was observed in all ultrasound-stimulated sites at 12 weeks after surgery and an 86% fusion rate was observed upon manual palpation, suggesting an acceleration of fusion over nontreated autograft control sites. The exact mechanism of action of the ultrasound device is unknown. It has been theorized that ultrasound pressure waves may mediate biologic activity by mechanical deformation of the cell membrane or extracellular matrix, indirectly by an electrical effect caused by cell deformation or by modifying the streaming potential in the extracellular fluid [16]. Studies have also shown effects on blood flow and vascularization as well as significant effects on cellular function [10–15]. However, studies in animals and humans have shown repeatedly its effectiveness in accelerating the bone repair process [3,8–11,16–18]. In a study by Glazer et al. [22], the rabbit posterolateral intratransverse process model was used to examine the effect of low-intensity pulsed ultrasound on spine fusion. Using radiographic and mechanical methods, they found a statistically significant increase in fusion rate from 65% to 93% with the use of ultrasound over nontreated controls at 6 weeks after surgery. A statistically significant increase in fusion segment stiffness was also observed with the use of ultrasound therapy. Similar to the present study, Glazer et al. found a qualitative increase in the new bone formation in histologic sections from ultrasound-treated sites compared with nontreated controls. Future clinical application of ultrasound therapy could be aimed at improving fusion rates by means of a noninvasive treatment where bone healing may be compromised, for example, in patients with a history of tobacco use. Pseudoar-
253
throsis after spinal fusion occurs at least four times more often in patients who smoke compared with patients who do not smoke [23–25]. Ultrasound stimulation was shown to substantially reduce the healing time of patients with tibia and distal radius fracture and significantly reduce the incidence of delayed union and nonunion [17]. The healing time of smoking patients in these studies with ultrasound stimulation was found to be equivalent to the healing time of placebo control nonsmokers [17]. In summary, the results of this study have demonstrated that the use of a low-intensity, pulsed ultrasound device may result in an earlier, more consistent spinal fusion. Although not statistically significant at 6 weeks, the histologic findings and mechanical analysis at 12 weeks advance the concept that low-intensity pulsed ultrasound can modulate bone healing, bone graft incorporation and, potentially, the stimulation of fusion. The limitations of the present study, specifically the small sample size of each experimental group and the lack of statistically significant radiographic results at either 6 or 12 weeks, prevent drawing strong conclusions regarding the efficacy of ultrasound in accelerating canine noninstrumented posterior spine fusions. Future studies to characterize and optimize the effects of the ultrasound signal may further demonstrate that ultrasound therapy may be a useful adjunct in spinal fusion surgery.
References [1] Cotler JM, Star AM. Complications of spinal fusions. In: Cotler JM, Cotler HB, eds. Spinal fusion: science and technique. New York: Springer-Verlag, 1990;361–87. [2] Muschler GF, Lane JM, Dawson EG. The biology of spinal fusion. In: Cotler JM, Cotler HB, eds. Spinal fusion: science and technique. New York: Springer-Verlag, 1990;9–21. [3] Duarte LR. The stimulation of bone growth by ultrasound. Arch Orthop Trauma Surg 1983;101:153–9. [4] Wolff J. The law of bone remodeling [Das Gesetz der Transformation der Knochen]. Originally published by Verlag von August Hirshwald, Berlin, 1869. English translation by Maquet P and Fulong R, published by Springer Verlag, Berlin, 1982. [5] Buckley MJ, Banes AJ, Levin LG, et al. Osteoblasts increase their rate of division and align in response to cyclic, mechanical tension in vitro. Bone Min 1988;4:225–36. [6] Goodship AE, Kenwright J. The influence of induced micromovement upon the healing of experimental tibial fractures. J Bone Joint Surg Br 1985; 67:650–5. [7] Rubin CT, Langon LE. Regulation of bone formation by applied dynamic loads. J Bone Joint Surg Am 1984;66:397–402. [8] Pilla AA, Mont MA, Nasser PR, et al. Non-invasive low-intensity pulsed ultrasound accelerates bone healing in the rabbit. J Orthop Trauma 1990;4:246–53. [9] Wang SJ, Lewallen DG, Bolander ME, et al. Low intensity ultrasound treatment increases strength in a rat femoral fracture model. J Orthop Res 1994;12:40–7. [10] Yang KH, Parvizi J, Wang SJ, et al. Exposure to low-intensity ultrasound increases aggrecan gene expression in a rat femur fracture model. J Orthop Res 1996;14:802–9. [11] Parvizi J, Wu CC, Lewallen DG, Greenleaf JF, Bolander ME. Lowintensity ultrasound stimulates proteoglycan synthesis in rat chondrocytes by increasing aggrecan gene expression. J Orthop Res 1999;17: 488–94.
254
S.D. Cook et al. / The Spine Journal 1 (2001) 246–254
[12] Ryaby JT, Bachner EJ, Bendo JA, Dalton PF, Tannenbaum S, Pilla AA. Low intensity pulsed ultrasound increases calcium incorporation in both differentiating cartilage and bone cell cultures. Trans Orthop Res Soc 1989;14:15. [13] Ryaby JT, Duarte-Alves P, Pilla AA. Ultrasound and static mechanical force affect second messenger activity in bone cell cultures. Trans Orthop Res Soc 1990;15:131. [14] Ryaby JT, Matthew J, Duarte-Alves P. Low intensity pulsed ultrasound affects adenylate cyclase and transforming growth factor-beta synthesis in osteoblastic cells. Trans Orthop Res Soc 1992;17:590. [15] Ryaby JT, Matthew J, Pilla AA. Low intensity pulsed ultrasound modulates adenylate cyclase activity and transforming growth factor beta synthesis. In: Brighton CT, Pollock SR, eds. Electromagnetics in biology and medicine. San Francisco: San Francisco Press, 1991:95–100. [16] Heckman JD, Ryaby JP, McCabe J, Frey JJ, Kilcoyne RF. Acceleration of tibial fracture-healing by non-invasive, low-intensity pulsed ultrasound. J Bone Joint Surg Am 1994;76:26–34. [17] Cook SD, Ryaby JP, McCabe J, Frey JJ, Heckman JD, Kristiansen TK. Acceleration of tibia and distal radius fracture healing in patients who smoke. Clin Orthop 1997;337:198–207. [18] Kristiansen TK, Ryaby JP, McCabe J, Frey JJ, Roe LR. Accelerated
[19]
[20]
[21] [22] [23] [24]
[25]
healing of distal radial fractures with the use of specific, low-intensity ultrasound. A multicenter, prospective, randomized, double-blind, placebo-controlled study. J Bone Joint Surg Am 1997;79:961–73. Cook SD, Dalton JE, Prewett AB, Whitecloud TS. In vivo evaluation of demineralized bone matrix as a bone graft substitute for posterior spinal fusion. Spine 1995;20:877–86. Cook SD, Dalton JE, Tan EH, Whitecloud TS, Rueger DC. In vivo evaluation of recombinant human osteogenic protein (rhOP-1) implants as a bone graft substitute for spinal fusions. Spine 1994;19:1655–63. Rosner B. Fundamentals of biostatistics, 3rd ed. Boston: PWS-Kent Publishing Company, 1990:200–6. Glazer PA, Heilman MR, Lotz JC, Bradford DS. Use of ultrasound in spinal arthrodesis. A rabbit model. Spine 1998;23:1142–8. Blumenthal SL, Baker J, Dossett A, Selby DK. The role of anterior lumbar fusion for internal disc disruption. Spine 1988;13:566–9. Brown CW, Orme TJ, Richardson HD. The rate of pseudoarthrosis (surgical nonunion) in patients who are smokers and patients who are nonsmokers: a comparison study. Spine 1986;11:942–3. Hanley EN, Levy JA. Surgical treatment of isthmic lumbosacral spondylolisthesis. Analysis of variables influencing results. Spine 1989;14:48–50.