How to determine the optimal proximal fusion level for Scheuermann kyphosis.

OBJECTIVE: To determine optimal proximal fusion levels for instrumented spinal fusion for Scheuermann kyphosis. METHODS: We reviewed 86 patients (33 women) who underwent corrective instrumented spinal fusion for Scheuermann kyphosis. All patients had long-cassette upright lateral radiographs taken preoperatively, postoperatively, and at 2 years and the last follow-up. Demographic, radiographic, and surgical parameters were compared between patients with and without PJK. RESULTS: PJK occurred in 28 patients (32%). The mean maximum Cobb angle was 85.8° ± 11.7° preoperatively, 54.8° ± 14.2° postoperatively, and 59.7° ± 16.8° at the last follow-up. Age and sex did not differ between the PJK and non-PJK groups (P > 0.05). The preoperative curve characteristics, fusion levels, and corrective ratio were similar in both groups (P > 0.05). The maximal Cobb angle at 2 years and the last follow-up significantly differed between the 2 groups (P < 0.05). The proportion of patients with the uppermost instrumented vertebra (UIV) at or above the proximal end vertebra (PEV) was similar in both groups (P > 0.05). The proportion of patients with UIV at or above T2 was significantly greater in the non-PJK group (P < 0.05). PJK was significantly associated with a C7 plumb line (C7PL)-sacrum distance ≥ 50 mm (P < 0.05). CONCLUSION: PJK is the main cause of postoperative correction loss. Proper fusion-level selection can reduce PJK occurrence. We recommend having the UIV at T2 or above, especially when the C7PL-sacrum distance ≥ 50 mm.

Geometric analysis of pedicle subtraction osteotomy (PSO) for Kyphosis correction: anterior lengthening may occur at the osteotomized body as well as at the discs above and below.

OBJECTIVE: To validate the authors kyphosis correction formula for pedicle subtraction osteotomy (PSO) cases. Additionally, to use the formula to evaluate the safety of PSO by determining if there is anterior lengthening. METHODS: Twenty-two patients with primarily kyphosis corrected by PSO and with clear landmarks on preoperative and postoperative x-rays were selected. Several anatomical lines and angle measurements were utilized as depicted previously in the Vertebral Column Resection formula (see below). Two approximations were calculated: the geometric approximation (G) = (tanG°*2 + 1)*15° and the rough approximation (R) which is about the same amount of actual shortening (x), if parallel length (y) ≥ 40; twice of x, if y < 40. For each patient, the change of segmental kyphosis angle (K°) was measured and compared with G° and R°, and the correlation between each value was analyzed. RESULTS: The absolute Mean ± SE for K - G and K - R was 2.33° ± 0.34 and 6.09° ± 0.58, respectively. K - G is < 3° (p = 0.03). K - R is < 8° (p = 0.001). In other words, K was close to G and R and thus can be predicted by these approximations. Average posterior shortening, anterior shortening, and kyphosis correction at each level were 20.8 ± 2.0 mm, - 3.64 ± 1.5 mm (which equates to anterior lengthening), and 31.05° ± 2.0, respectively. Anterior lengthening occurred in 13 cases (in 4 cases, both at the body as well as at the disc above and below.) The correlation between posterior and anterior shortening was 0.03 (p = 0.88). There were 3 cage insertion cases: 1 had anterior lengthening, while 2 had anterior shortening even with the cage. CONCLUSION: This study validated the geometric and rough approximations originally used in PVCR patients, for PSO patients. Additionally, this study found that anterior lengthening may occur in PSOs usually at the discs, but occasionally at the osteotomized body.

Key Role of Preoperative Recumbent Films in the Treatment of Severe Sagittal Malalignment.

STUDY DESIGN: Retrospective cohort study. OBJECTIVE: To determine if severe sagittal malalignment (SM) patients without fixed deformities require a three-column osteotomy (3CO) to achieve favorable clinical and radiographic outcomes. SUMMARY OF BACKGROUND DATA: 3CO performed for severe SM has significantly increased in the last 15 years. Not all severe SM patients require a 3CO. METHODS: Severe SM patients (sagittal vertical axis [SVA] >10 cm) who underwent deformity correction between 2002 and 2011. Patients with <33% change in their lumbar lordosis (LL) on a preoperative supine radiograph were classified as stiff deformities, whereas those with ≥33% change were categorized as flexible deformities. The clinical/radiographic outcomes were assessed at minimum two years postoperatively. RESULTS: Seventy patients met the inclusion criteria, 35 patients with flexible and 35 with stiff deformities. Eighteen flexible-deformity patients underwent a 3CO versus 22 stiff-deformity patients. The remaining patients in each group underwent spinal realignment without a 3CO. The flexible-deformity patients not undergoing a 3CO had overall improvement in all sagittal radiographic parameters. Preoperative LL (22°), LL-pelvic incidence (PI) mismatch (43), SVA (17 cm), and pelvic tilt (PT, 34°) improved to 46°, 18, 6 cm, and 26°, respectively, p < .05. Flexible-deformity patients who underwent a 3CO also had overall improvement in all radiographic parameters. Preoperative LL (8.5°), LL-PI mismatch (47), SVA (19 cm), and PT (37°) improved to 39°, 15, 7 cm, and 24°, respectively (p < .05). Stiff-deformity patients who underwent a 3CO had statistically significant improvement in all radiographic parameters. However, stiff-deformity patients who did not undergo a 3CO had suboptimal improvement in all radiographic parameters, except for SVA (14 cm-9 cm, p < .05). Flexible patients who did not undergo a 3CO had statistical improvement in the SRS domains of function and self-mage as well as in their ODI scores (p < .05). CONCLUSION: Severe SM that is flexible can be corrected without a 3CO without compromising clinical and radiographic outcomes. LEVEL OF EVIDENCE: Level III.

Spinal fusion for pediatric neuromuscular scoliosis: national trends, complications, and in-hospital outcomes.

OBJECTIVE The objective of this study was to determine if the recent changes in technology, surgical techniques, and surgical literature have influenced practice trends in spinal fusion surgery for pediatric neuromuscular scoliosis (NMS). In this study the authors analyzed recent trends in the surgical management of NMS and investigated the effect of various patient and surgical factors on in-hospital complications, outcomes, and costs, using the Nationwide Inpatient Sample (NIS) database. METHODS The NIS was queried from 2002 to 2011 using International Classification of Diseases, Ninth Edition, Clinical Modification codes to identify pediatric cases (age < 18 years) of spinal fusion for NMS. Several patient, surgical, and short-term outcome factors were included in the analyses. Trend analyses of these factors were conducted. Both univariate and multivariable analyses were used to determine the effect of the various patient and surgical factors on short-term outcomes. RESULTS Between 2002 and 2011, a total of 2154 NMS fusion cases were identified, and the volume of spinal fusion procedures increased 93% from 148 in 2002 to 286 in 2011 (p < 0.0001). The mean patient age was 12.8 ± 3.10 years, and 45.6% of the study population was female. The overall complication rate was 40.1% and the respiratory complication rate was 28.2%. From 2002 to 2011, upward trends (p < 0.0001) were demonstrated in Medicaid insurance status (36.5% to 52.8%), presence of ≥ 1 comorbidity (40.2% to 52.1%), and blood transfusions (25.2% to 57.3%). Utilization of posterior-only fusions (PSFs) increased from 66.2% to 90.2% (p < 0.0001) while combined anterior release/fusions and PSF (AR/PSF) decreased from 33.8% to 9.8% (< 0.0001). Intraoperative neurophysiological monitoring (IONM) underwent increasing utilization from 2009 to 2011 (15.5% to 20.3%, p < 0.0001). The use/harvest of autograft underwent a significant upward trend between 2002 and 2011 (31.3% to 59.8%, p < 0.0001). In univariate analysis, IONM use was associated with decreased complications (40.7% to 33.1%, p = 0.049) and length of stay (LOS; 9.21 to 6.70 days, p <0.0001). Inflation-adjusted mean hospital costs increased nearly 75% from 2002 to 2011 ($36,805 to $65,244, p < 0.0001). In the multivariable analysis, nonwhite race, highest quartile of median household income, greater preexisting comorbidity, long-segment fusions, and use of blood transfusions were found to increase the likelihood of complication occurrence (all p < 0.05). In further multivariable analysis, independent predictors of prolonged LOS included older age, increased preexisting comorbidity, the AR/PSF approach, and long-segment fusions (all p < 0.05). Lastly, the likelihood of increased hospital costs (at or above the 90th percentile for LOS, 14 days) was increased by older age, female sex, Medicaid insurance status, highest quartile of median household income, AR/PSF approach, long-segment fusion, and blood transfusion (all p < 0.05). In multivariable analysis, the use of autograft was associated with a lower likelihood of complication occurrence and prolonged LOS (both p < 0.05). CONCLUSIONS Increasing use of IONM and posterior-only approaches may combat the high complication rates in NMS. The trends of increasing comorbidities, blood transfusions, and total costs in spinal fusion surgery for pediatric NMS may indicate an increasingly aggressive approach to these cases.

Results of Revision Surgery for Proximal Junctional Kyphosis Following Posterior Segmental Instrumentation: Minimum 2-Year Postrevision Follow-Up.

STUDY DESIGN: A retrospective cohort study. OBJECTIVES: The aim of this study was to evaluate radiographic and patient-reported outcomes at minimum 2 years after revision surgery for proximal junctional kyphosis (PJK), correlating these results with PJK etiology. SUMMARY OF BACKGROUND DATA: There are no studies detailing the results of revision surgery for PJK following posterior segmental instrumentation. METHODS: Thirty-two consecutive patients treated with revision surgery after PJK above posterior fusions (25 women/7 men, average age at surgery 60.6 yrs) were reviewed for radiographic and patient-reported outcomes (mean follow-up, 4.5 yrs; range, 2-10 yrs). Patients were subdivided into fracture (F) and nonfracture (NF) groups on the basis of PJK etiology. RESULTS: Radiographic severity of PJK improved significantly with revision surgery and was maintained at ultimate follow-up (P < 0.001). However, initial sagittal vertical axis (SVA) correction was not maintained through ultimate follow-up (P = 0.04). There were significant postrevision improvements in mean Oswestry scores (P < 0.001) and SRS total scores (P < 0.001) in all patients. In patients with pelvic incidence-lumbar lordosis (PI-LL) mismatch < 11°, final PJK measurement was smaller than in patients with mismatch ≥11° (9.4° vs. 19.8°, P = 0.009). Six patients (19%) developed new postrevision PJK, with two (6%) requiring additional surgery. Patients who sustained PJK through a fracture had greater improvements in Oswestry (P = 0.004), total SRS (P = 0.04), pain (P < 0.001), and satisfaction (P = 0.05) scores, although the fracture patients had less maintained SVA correction (P = 0.002). CONCLUSION: Revision surgery for PJK following posterior instrumentation achieved acceptable radiographic and clinical outcomes at minimum 2-year follow-up. Patients with PI-LL mismatch <11° experienced more ultimate PJK correction than patients with mismatch ≥11°. Although the NF group experienced more sustained correction of sagittal balance, the F group reported greater improvements in patient-reported outcomes. Ultimate clinical outcomes after revision surgery for PJK were similar between patients with and without compression fractures. LEVEL OF EVIDENCE: 3.

Deformity Angular Ratio Describes the Severity of Spinal Deformity and Predicts the Risk of Neurologic Deficit in Posterior Vertebral Column Resection Surgery.

STUDY DESIGN: Retrospective review of prospectively collected data. OBJECTIVE: To assess the value of the deformity angular ratio (DAR, maximum Cobb measurement divided by number of vertebrae involved) in evaluating the severity of spinal deformity, and predicting the risk of neurologic deficit in posterior vertebral column resection (PVCR). SUMMARY OF BACKGROUND DATA: Although the literature has demonstrated that PVCR in spinal deformity patients has achieved excellent outcomes, it is still high risk neurologically. This study, to our knowledge, is the largest series of PVCR patients from a single center, evaluating deformity severity, and potential neurologic deficit risk. METHODS: A total of 202 consecutive pediatric and adult patients undergoing PVCRs from November 2002 to September 2014 were reviewed. The DAR (coronal DAR, sagittal DAR, and total DAR) was used to evaluate the complexity of the deformity. RESULTS: The incidence of spinal cord monitoring (SCM) events was 20.5%. Eight patients (4.0%) had new neurologic deficits. Patients with a high total DAR (≥25) were significantly younger (20.3 vs. 29.0 yr, P = 0.001), had more severe coronal and sagittal deformities, were more myelopathic (33.3% vs. 11.7%, P = 0.000), needed larger vertebral resections (1.8 vs. 1.3, P = 0.000), and had a significantly higher rate of SCM events than seen in the low total DAR (<25) patients (41.1% vs. 10.8%; P = 0.000). Patients with a high sagittal DAR (≥15) also had a significantly higher rate of SCM events (34.0% vs. 15.1%, P = 0.005) and a greater chance of neurologic deficits postoperatively (12.5% vs. 0, P = 0.000). CONCLUSION: For patients undergoing a PVCR, the DAR can be used to quantify the angularity of the spinal deformity, which is strongly correlated to the risk of neurologic deficits. Patients with a total DAR greater than or equal to 25 or sagittal DAR greater than or equal to 15 are at much higher risk for intraoperative SCM events and new neurologic deficits. LEVEL OF EVIDENCE: 3.

Long-term radiographic outcomes of a central hook-rod construct for osteotomy closure: minimum 5-year follow-up.

STUDY DESIGN: Retrospective study. OBJECTIVE: To evaluate long-term effectiveness of central hook-rod constructs for posterior spinal osteotomy closure. SUMMARY OF BACKGROUND DATA: During osteotomy site closure various techniques are used, including patient positioning, rod cantilevering, extending fixation points, and compressing through pedicle fixation points. All add premature stress on fixation points and may lead to loosening/eventual fixation failure. To avoid this, we often use a central compression hook-rod construct for osteotomy closure. METHODS: Fifty-six consecutive patients with fixed sagittal imbalance were treated with multilevel posterior column osteotomies (N = 19), pedicle subtraction osteotomy (N = 31), or vertebral column resection (N = 6). All 56 patients had undergone osteotomy closure using central compression hook-rod constructs and were analyzed at a follow-up of 5 years or more. Compression hooks were inserted into the fusion mass or lamina above/below the osteotomy and centrally attached to a short rod connected to pedicle screw-based rods via a cross-link. Diagnoses included sagittal imbalance associated with scoliosis (N = 39), degenerative sagittal imbalance (N = 14), ankylosing spondylitis (N = 2), and Scheuermann's kyphosis (N = 1). There were 55 revision cases and 1 primary. Radiographic/clinical analysis was performed to evaluate the efficacy/complications of this technique. RESULTS: Overall lumbar lordosis increased an average of 31.7° and local lordosis through the osteotomy site increased an average of 29.3°. Sagittal balance improved by an average of 92 mm. In all cases, osteotomy closures were performed without screw loosening or loss of correction intraoperatively. At a follow-up of 5 years or more, no failures of the hook-rod construct were seen, but there were 3 patients with partial implant failure; however, no symptomatic pseudarthroses at the osteotomy sites occurred. Seven patients developed pseudarthrosis below the central hook-rod construct. CONCLUSION: A central hook-rod construct is safe, controlled, and effective for applying compressive forces to close various spinal osteotomies without fixation failure or pseudarthrosis at the osteotomy site noted at a follow-up of 5 or more years. It adds fixation strength to the overall construct avoiding undue stress on pedicle screws. LEVEL OF EVIDENCE: 3.

Comparison of Anteroposterior to Posterior-Only Correction of Scheuermann's Kyphosis: A Matched-Pair Radiographic Analysis of 92 Patients.

STUDY DESIGN: Retrospective radiographic benchmark study. OBJECTIVE: To evaluate the amount of instrumented correction obtained from a combined anterior/posterior (A/P) versus posterior-only (post-only) approach for Scheuermann's kyphosis. SUMMARY OF BACKGROUND DATA: An A/P approach was thought to optimize correction; however, instrumentation advances using pedicle screws allow treatment through an all-posterior approach. METHODS: A total of 166 Scheuermann's kyphosis patients were treated between 2 centers: 90 by combined A/P approach at 1 center and 76 by post-only at the second center. From the 166 patients, a matched cohort of 92 (46 from each) was established according to preoperative sagittal (±10°) and hyperextension (HE) Cobb (±10°) measurements and matched for age and gender. RESULTS: In the matched-pair group, average preoperative sagittal Cobb angles were 75.9° for the A/P group versus 78.8° for the post-only group (p = .2). The HE Cobb angles were similar (52.4° vs. 51.1°; p = .6). They showed similar corrections (33.7° vs. 30.6°; p = .3) and postoperative Cobb measurements (43.4° vs. 47.1°; p = .2) as well. The number of fusion levels was 9 in the A/P group and 12 in the post-only group; the difference yielded significance (p = .02). CONCLUSIONS: The A/P and post-only approaches averaged similar degrees of correction. The A/P patients were likely to correct more than their preoperative HE sagittal Cobb measurement, whereas the post-only group corrected close to their preoperative HE measurement. The number of fusion levels was larger with the post-only group.

Predicting Kyphosis Correction During Posterior-Only Vertebral Column Resection by the Amount of Spinal Column Shortening.

STUDY DESIGN: Retrospective. OBJECTIVE: To investigate the relationship between the amount of correction achieved (K°) and extent of vertebral column shortening (mm) with posterior vertebral column resection (PVCR). SUMMARY OF BACKGROUND DATA: There is no scientific reference to the correlation between K° and column shortening (mm) with PVCR. METHODS: Based on simple geometry, we tested the hypothesis that we could predict the amount of actual kyphosis correction (K°) by calculation on 26 kyphotic PVCR patients. Using multiple linear measurements (mm), two angular approximations (°) were calculated: the geometric approximation (G°) using the geometric calculation (G-cal), and the rough approximation (R°) by more simplistic calculation (R-cal). Both G° and R° were compared against K° as measured on the pre- and postoperative radiographs. If calculated G° and R° is close to measured K°, we can use the calculations (G-cal and R-cal) in the clinical situation. RESULTS: The mean correction of K° was 38°. K°-G° and K°-R° were not significantly greater than 3° and 6°, respectively. As K° was very close to G° and R°, K° can replace G° and R°. Therefore, we can use G-cal and R-cal in the clinical setting and we can determine how much posterior shortening and what cage size is required to obtain a certain amount of K°. CONCLUSIONS: With two calculations (G-cal & R-cal), we can determine how much vertebral column shortening (mm) we need during PVCR to obtain the amount of kyphosis correction desired (K°). In order to obtain K°, using the formula deduced from G-cal and R-cal, we can determine the shortening between the upper and lower pedicle screws and cage size.

Can Intraoperative Spinal Cord Monitoring Reliably Help Prevent Paraplegia During Posterior Vertebral Column Resection Surgery?

STUDY DESIGN: Retrospective case series. OBJECTIVE: Analyze patients who underwent posterior vertebral column resection (PVCR) above the conus medullaris with intraoperative spinal cord monitoring (SCM) data loss. SUMMARY OF BACKGROUND DATA: PVCR is a powerful technique for treating severe spinal deformity but carries a high risk for major spinal cord deficits. METHODS: We assessed clinical, radiographic, and electrophysiologic monitoring and operative records of 90 consecutive adult and pediatric patients (mean age, 24.8 years; range, 7.5-76.8) who underwent PVCR above the conus medullaris for severe spinal deformity performed from 2002 to 2010 by one surgeon at one institution. RESULTS: Fifteen of 90 patients (16.7%) (10 male/5 female; mean age, 15 years) lost SCM (n = 13) or had data degradation meeting warning criteria (n = 2). Diagnoses were kyphoscoliosis (n = 8), angular kyphosis (n = 3), global kyphosis (n = 2), and severe scoliosis (n = 2). Seven were revisions. The average pre-/postoperative scolioses were 99° (range, 32°-152°) and 43° (range, 6°-76°), respectively. The average pre-/postoperative kyphoses were +100° (range, 60°-170°) and +54° (range, 28°-100°), respectively. SCM fluctuated during osteotomy on nine occasions, stabilizing with elevation of blood pressure in addition to anterior spinal cord decompression in four, correction of subluxation in one, and traction reduction in one. Seven patients had SCM changes during rod compression. Three required partial release of correction, two larger cage insertion, one subluxation correction, and one pedicle screw removal. One experienced changes during rod placement/removal, and another, as a result of hypothermia. Data returned in all after prompt intervention (mean, 10.1 minutes; range, 1-60) and all awoke with intact lower extremity function. CONCLUSION: The prevalence of SCM changes during PVCR above the conus medullaris was 16.7%, mostly during osteotomy and rod/screw compression. Data returned with prompt intervention and all had intact lower extremity motor function postoperatively. These SCM "saves" strongly emphasize the importance of multimodality neurophysiologic monitoring during high-risk cases, minimizing postoperative complications.

Progressive Myelopathy Patients Who Lack Spinal Cord Monitoring Data Have the Highest Rate of Spinal Cord Deficits Following Posterior Vertebral Column Resection Surgery.

OBJECTIVES: The authors analyzed patients who underwent posterior vertebral column resection (PVCR). All patients had spinal cord monitoring (SCM) attempted but some did not have predictable and usable tracings. SUMMARY OF BACKGROUND DATA: Posterior vertebral column resection is a powerful technique to correct severe spinal deformities but it has the potential for major neurologic complications. Spinal cord monitoring is extremely helpful in managing these difficult patients. METHODS: Spinal cord monitoring data, operative reports, charts, and radiographs of 112 consecutive adult and pediatric patients (mean age, 23.5 years; range, 5.8-74.0 years) who underwent PVCR were reviewed. All surgical procedures were performed between 2002 and 2010 by 1 surgeon at a single institution. RESULTS: Twenty patients (11 male, 9 female; mean age, 15.9 years) of 112 (17.9%) did not have detectable SCM tracings during surgery. Average preoperative and postoperative scoliosis for these 20 patients was 79.2° and 41.3°, respectively. Average preoperative and postoperative kyphosis was 106.6° and 59.8°, respectively. Thirteen of the 20 were revisions. Preoperative neurologic status included acute progressive myelopathy (n = 9), no lower extremity function (n = 6), chronic weak lower extremities (n = 2), chronic quadriparesis (n = 1), and normal (n = 2). Four of 9 patients with acute progressive myelopathy developed transient paraplegia postoperatively. They had angular kyphosis (mean, 116.3°) and 3 were revisions. Compared with the 92 patients who had obtainable intraoperative SCM and no spinal cord deficits, the risk of developing postoperative paraplegia in patients who had no SCM tracings was statistically higher (p = .0008). All 4 with spinal cord deficits after surgery regained varying degrees of lower extremity function and resumed ambulatory status at most recent follow-up. CONCLUSIONS: The prevalence of unobtainable intraoperative SCM during PVCR was 17.9% (20 of 112). Postoperative transient paraplegia occurred exclusively in patients with no monitorable data as a result of angular kyphosis with acute progressive myelopathy. The rate of transient spinal cord deficits was significantly higher when there was no obtainable SCM (4 of 20 vs. 0 of 92 with SCM; p = .0008).

Comparison of standard 2-rod constructs to multiple-rod constructs for fixation across 3-column spinal osteotomies.

STUDY DESIGN: Retrospective matched-cohort comparative study. OBJECTIVE: Compare radiographical outcomes after the use of a standard 2-rod construct (2-RC) versus a multiple-rod construct (multi-RC) across 3-column osteotomy sites in a matched cohort with severe kyphosis and/or scoliosis with minimum 2-year follow-up. SUMMARY OF BACKGROUND DATA: Three-column osteotomies are used for treating severe spinal deformities, typically with a standard 2-RC across the highly unstable osteotomy site. METHODS: Between 1996 and 2010, patients undergoing a 3-column osteotomy by a single surgeon were matched for age/diagnosis/vertebra(e) resected/levels fused and curve magnitude. Sixty-six control patients with a 2-RC were identified and appropriately matched to 66 consecutive patients with a multi-RC across the 3-column osteotomy site. Each group included 50 patients with lumbar pedicle subtraction osteotomy and 16 patients with vertebral column resection. Radiographs were measured using standard adult deformity criteria. RESULTS: Averages were compared for 2-RC versus multi-RC demonstrating no statistical differences in mean age at surgery, vertebrae resected, levels fused, bone morphogenetic protein used (patients), or average preoperative Cobb magnitude. There were significant differences in the occurrence of rod breakage and revision surgery for pseudarthroses at the 3-column osteotomy site (rod breakage: 2-RC: 11 vs. multi-RC: 2, P=0.002; and revision: 2-RC: 6 vs. multi-RC: 0, P=0.011). There was no complete implant failure in the multi-RC group but 2 patients had partial implant failure without symptomatic pseudarthrosis. Eight patients in each group (12%) developed a pseudarthrosis above or below the osteotomy site. CONCLUSION: The use of a multi-RC is a safe, simple, and effective method to provide increased stability across 3-column osteotomy sites to significantly prevent implant failure and symptomatic pseudarthrosis versus a standard 2-RC. We strongly recommend using a multi-RC to stabilize 3-column osteotomies of the thoracic and lumbar spine. LEVEL OF EVIDENCE: 3.

Risk factors for and assessment of symptomatic pseudarthrosis after lumbar pedicle subtraction osteotomy in adult spinal deformity.

STUDY DESIGN: Retrospective review of prospectively collected data. OBJECTIVE: To assess the prevalence, risk factors, and clinical outcomes for pseudarthrosis after a lumbar pedicle subtraction osteotomy (PSO). SUMMARY OF BACKGROUND DATA: There exists no large series that examines pseudarthrosis rates of PSOs. METHODS: Data of 171 consecutive patients with adult deformity who underwent a lumbar PSO by 2 surgeons at a single institution with a minimum 2-year follow-up were analyzed. Pseudarthrosis diagnosed through sagittal malalignment and instrumentation failure noted on radiograph was confirmed intraoperatively. RESULTS: Eighteen (10.5%) of 171 patients developed pseudarthrosis after a PSO. Eleven of the 18 patients (6.4% of all patients, 61.1% of the 18 patients with pseudarthrosis) had pseudarthrosis at the PSO site, L3 being the most common; other locations included the lumbosacral junction (4/18), thoracolumbar junction (2/18), and upper thoracic spine (1/18). Preoperative pseudarthrosis level was a predictor of the postoperative level of pseudarthrosis (93%). Fifteen of the 18 patients (83%) had no interbody fusion directly above or below the PSO site, 16 (88%) had a history of pseudarthrosis at the time of PSO surgery and 2 of 3 patients who had prior radiation to the lumbar region developed pseudarthrosis. Most pseudarthroses occurred within the first 2 years (n = 13/18), between 2 and 5 years (n = 3/18), and more than 5 years (n = 2/18) postoperatively. Prior pseudarthrosis (P < 0.0001), pseudarthrosis at the PSO site (P < 0.0001), prior decompression in the lumbar region (P = 0.0037), prior radiation to the lumbar region (P < 0.0001), and presence of inflammatory/neurological disorders (P < 0.0036) were identified as risk factors. All 18 patients with pseudarthroses required revision surgery (posterior-only surgery, n = 12; anteroposterior surgery, n = 6) due to loss of sagittal alignment and pain. The mean pre-revision Scoliosis Research Society score was 85, post-revision score was 95 (P = 0.0166), and the mean pre-revision Oswestry Disability Index score was 42.5, post-revision score was 34.5 (P = 0.0203). CONCLUSION: The overall prevalence of pseudarthrosis was 10.5% of which 61% occurred at the actual PSO site and Scoliosis Research Society and Oswestry Disability Index scores improved significantly after pseudarthrosis repair. LEVEL OF EVIDENCE: 4.

Results of revision surgery after pedicle subtraction osteotomy for fixed sagittal imbalance with pseudarthrosis at the prior osteotomy site or elsewhere: minimum 5 years post-revision.

STUDY DESIGN: Retrospective review of pedicle subtraction osteotomy (PSO) cases. OBJECTIVE: To report our results, radiographic and clinical outcomes at a minimum 5 years following revision surgery for pseudarthrosis after a PSO. SUMMARY OF BACKGROUND DATA: To our knowledge, there is no report on the results of revision surgery for pseudarthrosis after a PSO. METHODS: Eighteen consecutive patients with pseudarthrosis after PSO (16 females/2 males; average age at surgery, 49.8 yr) treated with revision surgery at one institution were analyzed (average follow-up, 6.5 yr; range, 5-12 yr). Radiographic and clinical outcomes analysis was performed. RESULTS: Sagittal vertical axis (SVA) and lumbar lordosis (LL) improved significantly after revision surgery (SVA, P = 0.000; LL, P = 0.024) and were maintained until ultimate post-revision follow-up (SVA, P = 0.170; LL, P = 0.729). Proximal junctional angle (P = 0.828), thoracic kyphosis (P = 0.828), and PSO angle (P = 0.717) achieved by the primary surgery were also maintained until ultimate post-revision. We increased the number of rods and/or changed them to 6.35-mm diameter in all patients. There were significant improvements post-revision in Oswestry Disability Index (45 vs. 37.9, P = 0.041) and Scoliosis Research Society pain subscale (2.6 vs. 3.1, P = 0.047) but not in Scoliosis Research Society total score or other subscales. Pelvic incidence greater than 60° demonstrated a trend toward poorer Oswestry Disability Index and Scoliosis Research Society scores (P > 0.05), but there were no significant differences between SVA greater or less than 11 cm. CONCLUSION: Revision surgery for pseudarthrosis after PSO can provide acceptable radiographic and clinical outcomes at a minimum 5 years post-revision. Successful surgical outcomes may be achieved by using an increased number or size of implants and ample bone graft for complete fusion after revision surgery. LEVEL OF EVIDENCE: 4.

The prevalence of abnormal preoperative neurological examination in Scheuermann kyphosis: correlation with X-ray, magnetic resonance imaging, and surgical outcome.

STUDY DESIGN: Retrospective. OBJECTIVE: The purpose of this study was to report the prevalence of abnormal neurological findings detected by physical examination in Scheuermann kyphosis and to correlate it to radiographs, magnetic resonance imaging (MRI) findings, and results of operative treatment. SUMMARY OF BACKGROUND DATA: There have been sporadic reports about abnormal neurological findings in patients with Scheuermann kyphosis. METHODS: Among 82 patients with Scheuermann kyphosis who underwent corrective surgery, 69 primary cases were selected. Patients' charts were reviewed retrospectively in terms of pre and postoperative neurological examinations. Sensory or motor change was defined as an abnormal neurological examination. Their duration, associated problems, and various parameters on preoperative radiographs and MRI examinations were also measured to search for any atypical findings associated with an abnormal neurological examination. RESULTS: There were 6 cases (9%) (group AbN), with an abnormal neurological examination ranging from severe myelopathy to a subtle change (e.g., sensory paresthesias on trunk). Five patients recovered to a normal neurological examination after corrective surgery. The remaining 1 patient with severe myelopathy also showed marked improvement and was ambulatory unassisted by 2-year follow-up. In patients with a normal neurological examination (group N, n = 63), only 1 patient had neurological sequelae because of anterior spinal artery syndrome after combined anterior-posterior correction. No preoperative radiographical parameters were significantly different between groups. Average age was 21.3 (AbN) and 18.6 (N) years (P = 0.55). Average preoperative T5-12 kyphosis was 69.0° (AbN) and 72.5° (N) (P = 0.61). Forty-two magnetic resonance images were obtained and all showed typical findings of Scheuermann kyphosis. Five patients in the AbN group (1 patient underwent computed tomography/myelography) and 37 patients in the N group underwent an MRI. CONCLUSION: The prevalence of abnormal neurological findings in Scheuermann kyphosis was 9%, emphasizing the importance of performing a detailed preoperative neurological examination. If congenital stenosis or a herniated thoracic disc is present, myelopathy can occur. No radiographical findings correlated with the abnormal preoperative neurological examinations. A normal MRI can exist in the face of an abnormal neurological examination, and conversely, a normal neurological examination can be seen with an abnormal MRI. Surgery was successful in alleviating abnormal neurological issues. LEVEL OF EVIDENCE: 4.

Does preoperative narcotic use adversely affect outcomes and complications after spinal deformity surgery? A comparison of nonnarcotic- with narcotic-using groups.

BACKGROUND CONTEXT: The role of preoperative (preop) narcotic use and its influence on outcomes after spinal deformity surgery are unknown. It is important to determine which patient factors and comorbidities can affect the success of spinal deformity surgery, a challenging surgery with high rates of complications at baseline. PURPOSE: To evaluate if preop narcotic use persists after spinal deformity surgery and whether the outcomes are adversely affected by preop narcotic use. STUDY DESIGN/SETTING: Retrospective evaluation of prospectively collected data. PATIENT SAMPLE: Two hundred fifty-three adult patients (230 females/23 males) undergoing primary spinal deformity surgery were enrolled from 2000 to 2009. OUTCOME MEASURES: Preoperative and postoperative (postop) narcotic use and changes in Oswestry Disability Index (ODI), Scoliosis Research Society (SRS) pain, and SRS total scores. METHODS: Preoperative, 2-year postop, and latest follow-up pain medication use were collected along with ODI, SRS pain, and SRS scores. Preoperative insurance status, surgical and hospitalization demographics, and complications were collected. All patients had a minimum 2-year follow-up (average 47.4 months). RESULTS: One hundred sixty-eight nonnarcotic (NoNarc) patients were taking no pain meds or only nonsteroidal anti-inflammatories preoperatively. Eighty-five patients were taking mild/moderate/heavy narcotics before surgery. The average age was 48.2 years for the NoNarc group versus 53.6 years for the Narc group (p<.005). There were significantly more patients with degenerative than adult scoliosis in the Narc group (47 vs. 28, p<.001; mild 19 vs. 24, p<.02; moderate 6 vs. 14, p<.0003; heavy 3 vs. 10, p<.0002). Insurance status (private/Medicare/Medicaid) was similar between the groups (p=.39). At latest follow-up, 137/156 (88%) prior NoNarc patients were still not taking narcotics whereas 48/79 (61%) prior narcotic patients were now off narcotics (p<.001). Significant postop improvements were seen in Narc versus NoNarc groups with regard to ODI (26-15 vs. 44-30.3, p<.001), SRS pain (3.36-3.9 vs. 2.3-3.38, p<.001), and overall SRS outcome (3.36-4 vs. 2.78-3.68, p<.001) scores. A comparison of change in outcome scores between the two groups showed a higher improvement in SRS pain scores for the Narc versus NoNarc group (p<.001). CONCLUSIONS: In adults with degenerative scoliosis taking narcotics a significant decrease in pain medication use was noted after surgery. All outcome scores significantly improved postop in both groups. However, the Narc group had significantly greater improvements in SRS pain scores versus the NoNarc group.

Radiographical and Clinical Outcomes of Posterior Column Osteotomies in Spinal Deformity Correction.

STUDY DESIGN: Prospectively enrolled, retrospectively analyzed case series. OBJECTIVE: To evaluate a large series of pediatric patients/patients with adult spinal deformity undergoing surgery with posterior column osteotomies (PCOs). SUMMARY OF BACKGROUND DATA: Osteotomies of the posterior column (Smith-Petersen or Ponté) are used to reduce kyphosis, increase lordosis, or increase spinal flexibility. However, little focused evidence exists regarding the efficacy and safety of this technique. METHODS: A total of 128 consecutive patients underwent posterior spinal fusion with PCOs with minimum 2-year follow-up. Seventy-five were primary surgical procedures; 53 were revisions. Data were collected from hospital charts, clinic notes, radiographs, and standardized questionnaires (Scoliosis Research Society-30 and Oswestry Disability Index). RESULTS: A total of 128 patients aged 37.6 ± 21 years underwent 518 PCOs (mean, 4.0 ± 2.2 yr) with 14.4 ± 3 mean instrumentation levels, with 3-year (range, 2-6.8 yr) average follow-up. PCOs were used for kyphosis correction in 49%, scoliosis correction at the apex of a curve in 13%, and both in 38%. One hundred six patients had complete radiographical data available for evaluation. Mean kyphosis correction per PCO was 8.8° ± 7.2°, varying with patient age (10.2° for those younger than 21 yr vs. 7.7° for those 21 yr or older, P < 0.0001) and region of the spine: thoracolumbar (T10-L2) 11.6°, lumbar (L2-S1) 9.4°, midthoracic (T6-T10) 7.2° and proximal thoracic (T1-T6) 3.6°. With PCOs at the apex of a curve, the maximum coronal Cobb decreased from 66° ± 21° to 31° ± 14° (P < 0.0001). Average estimated blood loss was 1419 ± 887 mL, correlating with greater age (P < 0.0001) and more instrumented levels (P < 0.0001), but not with the number of PCOs (P = 0.32). Complications occurred in 31 (24.2%) patients, including 4 radiculopathies (none attributable to PCOs). Complications did not correlate with the number of PCOs (P = 0.5). Six (4.7%) patients had loss of spinal cord monitoring or a failed wake-up test that could be attributed to overcorrection with PCOs, but none had postoperative deficits. Oswestry Disability Index scores improved (34.4 ± 17 to 23.6 ± 18, P < 0.0001), as did normalized Scoliosis Research Society-30 scores (63.7 ± 13 to 76.4 ± 15, P < 0.0001). CONCLUSION: Patients in this series undergoing posterior spinal fusion with PCOs achieved overall favorable outcomes for spinal deformity correction. The number of PCOs did not correlate with increased estimated blood loss or complications. The main technical concern was overcorrection, but neurological consequences associated with overcorrection were identified by intraoperative spinal cord monitoring and wake-up tests, and no patients experienced permanent neurological deficits related to PCOs. LEVEL OF EVIDENCE: 4.

Clinical and radiographic outcomes after 3-column osteotomies with 5-year follow-up.

STUDY DESIGN: Retrospective review of prospective database. OBJECTIVE: To investigate long-term results after 3-column osteotomies (3-CO). SUMMARY OF BACKGROUND DATA: Short-term studies have noted improved outcomes and alignment after 3-CO, but there is a paucity of long-term studies with a large group of patients. METHODS: An analysis of 126 patients who underwent a 3-CO (pedicle subtraction osteotomy [N = 101]/vertebral column resection [N = 25]) with minimum 5-year follow-up was performed at a single institution. The mean age was 48 years (range, 8-79 yr) and average follow-up was for 7 years (range, 5-14 yr). Oswestry Disability Index and Scoliosis Research Society (SRS) scores and radiographical parameters were assessed at baseline, 6 weeks, and 1, 2, 3, and/or 5 years postoperatively and complications were recorded. RESULTS: Sagittal alignment improved at all postoperative time points from baseline (mean, 117 mm), but diminished from 6 weeks (mean, 24 mm) to 5 years (mean, 41 mm; P = 0.03). Average coronal alignment was improved from baseline (27 mm) at 6 weeks (18 mm; P = 0.003) and 5 years postoperatively (19 mm; P = 0.007), with no deterioration between 6 weeks and 5 years postoperatively (P = 0.9). Major surgical complications occurred in 36% (n = 45) and major repeat surgery was performed in 28% (n = 35). Significant improvements (P < 0.05) in Oswestry Disability Index and all SRS domain scores were found at each time point. All mean outcome scores at 5 years postoperatively exceeded minimal clinically important difference thresholds except the SRS function domain. Improvement in outcomes at 5 years postoperatively was similar in groups with major surgical complications versus those without and in those with reoperation versus those without. CONCLUSION: This study of 126 patients undergoing 3-CO found significant and sustained improvements in Oswestry Disability Index and SRS scores and sagittal alignment at a minimum 5 years postoperatively. This demonstrates the durability of these complex spinal reconstructions, even surprisingly in those patients having a major complication and/or revision surgery. LEVEL OF EVIDENCE: 4.

Revision Versus Primary Patients Undergoing Vertebral Column Resection for Severe Spinal Deformities.

STUDY DESIGN: Retrospective comparative study. OBJECTIVES: To compare correction rates and complications of revision versus primary patients undergoing vertebral column resection (VCR). SUMMARY OF BACKGROUND DATA: Although an all-posterior VCR has obviated the need for a circumferential approach, it is technically demanding, especially in a revision setting. METHODS: Between 2002 and 2009, 55 revision patients underwent a posterior-only VCR. Diagnoses included severe scoliosis (n = 3), kyphoscoliosis (KS) (n = 29), global kyphosis (GK) (n = 13), and angular kyphosis (AK) (n = 10). Radiographic findings and complications were compared with 38 primary patients who underwent a VCR during the same period. All patients had a minimum 2-year follow-up (range, 2-6 years). RESULTS: The mean number of VCR levels were 1.6 in revision versus 1.2 in primary cases (p = .005). In the severe scoliosis and KS groups, major coronal curve correction was 48% in revision versus 63% in primary cases (p = .001). In the KS, GK, and AK groups, the major sagittal curve correction was 52% in revision versus 57% in primary cases (p = .27). Preoperative (p = .015) and postoperative (p = .002) sagittal imbalance was significantly greater in the revision group. There were no spinal cord-related complications, but 7 revision (13%) and 3 primary (8%) patients temporarily lost neuromonitoring data or failed wakeup tests; however, none had a permanent neurological deficit. Six revision patients (11%) required further revision surgery due to implant failure (3), progressive sagittal or coronal imbalance (2), and delayed deep wound infection (1) versus only 1 primary patient (3%) due to increased coronal imbalance. Preoperative and postoperative Scoliosis Research Society scores were not significantly different between groups. CONCLUSIONS: Vertebral column resections in revision patients may be more technically demanding than in primary patients but can be performed safely in conjunction with intraoperative spinal cord monitoring. Revision and primary patients undergoing a VCR showed improved clinical outcomes.

Comparison of TLIF with rhBMP-2 versus no TLIF and higher posterolateral rhBMP-2 dose at L5-S1 for long fusions to the sacrum with sacropelvic fixation in patients with primary adult deformity.

STUDY DESIGN: Comparvative case series. Data was prospectively entered and retrospectively analyzed. OBJECTIVE: To evaluate the need for distal lumbar interbody fusion when sufficient recombinant human bone morphogenetic protein-2 (rhBMP-2) is used posterolaterally at L5-S1 in long spinal constructs for adult deformity via costs and radiographical and patient-reported outcome comparisons. SUMMARY OF BACKGROUND DATA: Many authors and investigators have suggested that an interbody fusion is mandatory at L5-S1 with long fusion to the sacrum with sacropelvic fixation. Past studies have shown competitive fusion rates using rhBMP-2 versus iliac crest bone graft for long fusions. There are various advocates for anterior lumbar interbody fusion versus posterior lumbar interbody fusion versus transforaminal lumbar interbody fusion (TLIF). The optimal strategy remains elusive. METHODS: Fifty-seven patients were studied at one institution. Thirty-one patients had no interbody fusion (NI group) with 20 mg of rhBMP-2 posterolaterally on 10 mL of carrier and 26 patients had TLIF at L5-S1 (TLIF group) with 6 mg of rhBMP-2 in the interbody space along with local bone graft and 6 mg of rhBMP-2 on carrier posterolaterally at L5-S1. Patients were followed for 24 to 87 months (mean follow-up, 3.92 yr). Demographics of the 2 groups were similar. RESULTS: There were no detected nonunions at L5-S1 in either group. By our limited cost analysis, the expense of performing a TLIF at L5-S1 is higher than that of using extra rhBMP-2 posterolaterally at that segment. Improvement in outcomes scores, namely Scoliosis Research Society-22 and Oswestry Disability Index, were the same statistically in both groups. Blood loss was greater in the TLIF group than the NI group. There were no identified rhBMP-2 adverse events in either group. CONCLUSION: Utilization of 20 mg of rhBMP-2 at L5-S1 has the potential to be less expensive than an interbody fusion in most patients having a primary long fusion for adult spinal deformity. The apparent fusion rates at L5-S1 were identical in both groups. Both strategies were successful in regard to improving patient outcomes and achieving apparent solid arthrodesis at the lumbosacral junction, which was the focus of this study. LEVEL OF EVIDENCE: 2.