Secondary Logo

Journal Logo

Online Exclusive Spine Focus Section: Idiopathic Scoliosis

Factors in Surgical Decision Making for Thoracolumbar/Lumbar AIS: It’s About More Than Just the Curve Magnitude

Souder, Christopher MD*; Newton, Peter O. MD; Shah, Suken A. MD; Lonner, Baron S. MD§; Bastrom, Tracey P. MA; Yaszay, Burt MD

Author Information
Journal of Pediatric Orthopaedics: December 2017 - Volume 37 - Issue 8 - p e530-e535
doi: 10.1097/BPO.0000000000000746
  • Free


Adolescent idiopathic scoliosis (AIS) is a complex 3-dimensional distortion of the spine consisting of deviations in coronal and sagittal alignment as well as rotational abnormalities in the transverse plane.1 When evaluating AIS, surgeons typically base their decision on the severity of the deformity as determined by the major curve angle measurement of the coronal curvature on a PA radiograph. In skeletally mature patients, this leads to recommendations of observation versus surgical intervention based on the likelihood of curve progression in adulthood. There is concern for pulmonary compromise and osteoarthritis in curves that undergo significant progression throughout adulthood.2 The threshold surgeons commonly use for recommending surgical intervention is 50 degrees of coronal curvature.

Besides concerns for the future, the desire to lessen the clinical deformity is a major concern to most adolescents with scoliosis.3 Three main clinical abnormalities are seen in patients with scoliosis, including a rib hump or paraspinal prominence, shoulder asymmetry, and trunk shift.4 It is of no surprise that the typical patient’s satisfaction is based more on the improvement of the perceived deformity than possible future health benefits provided by surgical correction.2 Unfortunately, the clinical deformity in idiopathic scoliosis has not been shown to have a clear linear relationship with major curve angle measurements or vertebral rotation.4

In AIS, the major curve angle describes only the coronal plane deformity of a 3-dimensional disturbance of the spine. To fully evaluate a scoliotic deformity, the surgeon must assess the alterations of the spine in all 3 planes as seen on radiographs, as well as the clinical deformity as appreciated by the patient. Unfortunately, it is difficult to quantify clinical deformity as it is mostly a subjective measure determined by the patient and surgeon. With this known limitation, some patients undergo surgical intervention below the typical threshold of 50 degrees. This can especially be seen in thoracolumbar/lumbar (TL/L) curves. The purpose of this study was to compare operative and nonoperative patients with similar curve magnitudes to determine factors associated with surgical correction of “smaller” TL/L AIS curves.


A retrospective review of a prospectively collected AIS database was performed. Patients with major TL/L curves <50 degrees (Lenke 5 and 6) and a low risk of progression (Risser 3, 4, and 5) were selected. Patients were grouped based on their treatment (operative vs. nonoperative). Nonoperative patients belonged to one of 2 groups: (1) patient age 10 to 21 years, coronal curve ≥40 degrees, surgery had been offered and the patient elected to not proceed with surgery, or (2) patient age 10 to 21 years, curve magnitude ≥30 degrees, and skeletally mature.

Preoperative demographic, radiographic, Scoliosis Research Society (SRS)-22 outcomes questionnaire scores, and trunk shape data were obtained for the operative patients. Similar baseline values were obtained for the nonoperative cohort. These baseline values were obtained at the patient’s clinical visit during which the patient was identified by the surgeon as eligible and enrolled in the nonoperative arm of the study. Clinical measures of trunk shape obtained during the clinic visit were not available for the nonoperative cases. Therefore, only trunk shape measures able to be measured from radiographs were evaluated (shoulder height difference and trunk shift) and compared between groups. Shoulder height difference was determined by differences in the acromial height, whereas trunk shift was the difference between the center sacral vertical line and midpoint between the lateral margins of the ribs. These measures were evaluated as continuous variables and also grouped into those with shoulder height difference or trunk shift ≥2 cm versus those <2 cm.

All available variables were compared in a series of univariate analyses between the operative and nonoperative cohorts to identify individually significant variables eligible for entry in a multivariate analysis. The α was set as P<0.10 for entry into the multivariate binary logistic regression (predictors of surgical intervention). Two separate regression models were evaluated. The first consisted of only demographic and radiographic/trunk shape measures found to be eligible for entry. The second added the subjective SRS score to the model. This stepwise modeling was undertaken in an effort to isolate the predictive value of the patient and/or deformity characteristics on the decision to pursue surgical correction, followed up by the impact that the patients’ perceptions of their deformities had on predicting treatment selection by entering the outcome scores into the model. The α was set at P≤0.05 to declare significance within the regression models.


A total of 126 patients undergoing surgical intervention and 17 patients pursuing nonoperative treatment were analyzed. The operative group consisted of 126 patients with lumbar curves ranging from 35 to 49 degrees (average 43 degrees). The nonoperative group consisted of 17 patients with lumbar curves ranging from 26 to 49 degrees (average 39 degrees). At the time of enrollment, 6 of the 17 nonoperative subjects had been offered surgery and declined.

Sex and body mass index were similar between groups; the operative group was younger by 1 year (Table 1). The operative group had a slightly larger lumbar curve, trunk shift, and thoracolumbar apex translation as well as a smaller thoracic/lumbar curve ratio and larger thoracolumbar apex translation (Table 1). Other radiographic parameters were similar between the 2 groups. There were significantly more patients with a trunk shift ≥2 cm in the operative cohort (50%) compared with the nonoperative cohort (24%, P=0.04). SRS scores were significantly lower (more symptomatic) in the operative group for pain, self-image, function, mental health, and total (Table 2). An example of an operative and a nonoperative curve is shown in Figures 1 and 2, respectively.

Results of the Univariate Analysis
SRS Score Comparison Between Operative and Nonoperative Patients
An example of a patient in the operative group who had a 37-degree thoracolumbar/lumbar curve at the preoperative timepoint.
An example of a nonoperative patient who had a 32-degree thoracolumbar/lumbar curve at baseline.

Only lumbar curve magnitude and trunk shift remained significant predictors of surgery in the multivariate regression model of demographic, radiographic, and trunk shape parameters (Table 3). SRS total score was selected as the representative patient outcome score for entry into the second regression model. This 1 score was selected due to potential multicollinearity issues that arise when all domains are selected for entry. When SRS total score was entered into the regression, it was the only significant predictor of surgical intervention (Table 3). This inverse odds ratio is indicative of a decreased likelihood of surgical intervention with increasing SRS total score.

Variables Entered Into the Regression Models


Surgical intervention for AIS is most commonly initiated in patients with scoliotic curves that have a considerable risk of progression. Much of the information used to prognosticate the potential for progression of scoliotic curves is based on the information gathered in a landmark study by Weinstein et al5 evaluating long-term follow-up and prognosis for untreated scoliosis. It was found that most thoracic curves that reached 50 to 80 degrees at skeletal maturity continued to progress into adulthood. With this information, most surgeons recommend surgical intervention in patients with curves >50 degrees to prevent later progression. Gardner added additional factors to be taken into account when considering surgical intervention in young patients with scoliosis. These 5 factors are to be considered in decreasing order of importance: curve magnitude, age, flat thoracic spine, pain, and external appearance.6

Thoracolumbar and lumbar curves can behave differently, as seen in the classic article by Weinstein et al.5 In this population, thoracolumbar and lumbar curves were frequently found to progress once reaching the 35- to 40-degree mark. Thoracolumbar curves were also noted to be more commonly associated with marked translatory shifts between 2 vertebrae. Later, Weinstein7 noted that thoracolumbar curves >30 degrees had rapid progression initially followed by continual progression through the 40-year follow-up. Each curve was also associated with high degrees of apical vertebral rotation. Other authors have also noted that thoracic and thoracolumbar curves are more likely to progress and produce a more obvious clinical deformity.6,8

Factors influencing a patient’s desire for surgery below the typical thresholds of 50 degrees in skeletally mature patients have not been previously evaluated. Literature discussing decision making for surgery almost exclusively focuses on surgeon utilization of the major curve angle measurement of the coronal radiographic deformity for determining when surgery is indicated. There is a paucity of literature discussing the severity of clinical deformity in which surgery would be indicated, and this would most definitely vary based on individual patient perception. With this known limitation, most surgeons utilize the size of the curve as the driving force for recommending surgery. The clinical deformity and its effects as interpreted by the patient have an important role in the decision making for surgical intervention for a scoliotic deformity. As demonstrated by this study, worse self-perception as measured by SRS scores were seen in surgical patients with relatively small curve magnitudes.

It has been shown that the pattern and location of the curve are more closely associated with the clinical deformity than with the magnitude of curvature.9 Theologis et al3 found that major curve angle measurements of scoliotic deformities correlated poorly with clinical cosmetic scores. Although the rib hump was found to have the strongest correlation, it was not sufficient to fully quantify the cosmetic appearance of the back. In addition, Thulbourne and Gillespie4 found that the clinical deformity, identified as the rib hump, in idiopathic scoliosis has no clear linear relationship with major curve angle measurements or vertebral rotation. This is consistent with Gardner’s assessment that the success of surgical correction may be better measured by the shift of the apex of the curve and its attached ribs toward midline rather than by how much the major curve angle is decreased.10

The outward appearance of the patient is affected by more than simply curve magnitude as labeled by the major curve angle or even vertebral rotation. Subjective scores of cosmesis have been found to correlate with a combination of major curve angle, hump severity, asymmetry of waistline, circumference of chest, and obesity.11 In a long-term follow-up of patients with idiopathic scoliosis, Edgar and Mehta12 found that psychosocial disturbances were common, and that patients’ primary concerns arose from the rib hump and having an unbalanced appearance. This is supported by our study that found trunk shift to be a predictor for surgery.

The reason for the larger deformity produced by smaller curves has not been fully evaluated. Schwender and Denis13 inspected lumbosacral hemicurves (lumbosacral fractional curve/oblique takeoff of the lumbar spine) and their effect on coronal imbalance in AIS. They found that the fractional curve was associated with decompensation both preoperatively and postoperatively. Spinal imbalance will occur if a hemicurve does not have a compensatory hemicurve of equal magnitude. This has been found true in cases of coronal decompensation after selective thoracic fusion in which a fractional lumbar curve persisted.13,14 Both lateral trunk shift and coronal decompensation were observed resulting from a relatively small persistent hemicurve. Trigonometric calculation shows that a simple 3-degree obliquity located at the sacrum can produce an imbalance of 2.2 cm.13 In this same thought, a “hemicurve” can exist in isolation within the lumbar or thoracolumbar region. In this situation, if not paired with a compensatory curve above, an isolated “hemicurve” with a relatively small magnitude could easily produce a clinically significant trunk deformity. For many of our patients we believe the unbalanced hemicurve is responsible for the patients’ trunk shift and poor self-perception, and ultimately their desire for surgical intervention (Fig. 3). A better understanding of the hemicurve would be beneficial. In addition to limb discrepancy, other factors may influence its development. Although this was outside of the scope of the current study, future studies are needed to determine why patients develop hemicurves and are unbalanced.

Example of a patient with a “hemicurve” resulting in noticeable trunk shift both radiographically and clinically.

Adult literature evaluating idiopathic TL/L curves have shown a more ominous outcome compared with thoracic curves. When followed into adulthood, patients with major thoracolumbar or lumbar curves were found to have a higher incidence of back pain in multiple studies.15,16 In direct comparison with this, studies documenting outcomes in surgically managed adolescents with idiopathic scoliosis have revealed excellent subjective scores in those with TH/L curves. White et al17 evaluated patients’ perceived outcomes based on SRS scores after surgical intervention and found that patients with TL/L curves had the best results, with 85% of patients scoring the highest possible score of self-image testing. There are a few limitations to this study. The most significant limitation is that the exact motivation for surgery is not known. This is hindered by the retrospective nature of the study. Only a prospective study collecting specific questions regarding motivation for surgical intervention will be able to collect those exact data needed. Assessment of truncal deformities and SRS scores is only a surrogate measure of the motivation for surgery. As there was significant difference between the 2 groups in these measures we do believe that they may be responsible for some of the decision making for surgery. The other limitation is that the influence of the treating surgeon is not known. The multicenter and retrospective nature of the study makes it difficult to assess for the surgeon bias. However, the benefit of it being a multicenter study is that it prevents a single surgeon’s bias from influencing the results. Finally, the nonoperative cohort was not matched and not specifically collected to be a comparison group for the operative patients. However, these patients had a major thoracolumbar curve and were collected from the same group of surgeons that collected the operative group. We, therefore, believe the nonoperative cohort provided some insight into the motivation for nonoperative management.

In conclusion, this study suggests that poor self-perception and significant trunk shift drive some patients with smaller TL/L curves to seek surgery. The clinical deformity may likely result, in many patients, from an unbalanced hemicurve. However, further study is needed to support this hypothesis. Considering the available literature suggesting the poor natural history, the long-term follow-up may demonstrate a positive benefit in the surgical treatment of these select groups of small TL/L curves. Ultimately, the decision to perform surgery is a complex one, influenced by patient perceptions, surgeon perceptions, and parent perceptions. Indications for surgery still consists of curves with magnitudes >50 degrees and those with marked apical rotation or truncal shift.18,19 Although this study does suggest that poor self-perception is a factor for surgery in patients with smaller TL/L curves, we do not believe that this should be the main indicator for surgery.


1. Deacon P, Archer IA, Dickson RA. The anatomy of spinal deformity: a biomechanical analysis. Orthopedics. 1987;10:897–903.
2. Haher TR, Merola A, Zipnick RI, et al. Meta-analysis of surgical outcome in adolescent idiopathic scoliosis. A 35-year English literature review of 11,000 patients. Spine (Phila Pa 1976). 1995;20:1575–1584.
3. Theologis TN, Jefferson RJ, Simpson AH, et al. Quantifying the cosmetic defect of adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 1993;18:909–912.
4. Thulbourne T, Gillespie R. The rib hump in idiopathic scoliosis. Measurement, analysis and response to treatment. J Bone Joint Surg Br. 1976;58:64–71.
5. Weinstein SL, Zavala DC, Ponseti IV. Idiopathic scoliosis: long-term follow-up and prognosis in untreated patients. J Bone Joint Surg Am. 1981;63:702–712.
6. Gardner AFindlay G, Owen R. Surgical correction of scoliosis in the young. Surgery of the Spine. Oxford: Blackwell Scientific Publications; 1992:431–464.
7. Weinstein SL. Idiopathic scoliosis. Natural history. Spine (Phila Pa 1976). 1986;11:780–783.
8. Clarisse P. Prognostic evolutif des scolioses idiopathiques mineures de 10” to 29” en periode de croissance. Lyon: Université Claude Bernard Lyon 1; 1974.
9. Roach JW. Adolescent idiopathic scoliosis. Orthop Clin North Am. 1999;30:353–365. vii-viii.
10. Gardner A. Correction of adolescent idiopathic scoliosis by the heavy (6 mm) harstshill rectangle with particular reference to the behaviour of the ribs. 1991. In Proceedings of British Scoliosis Society Combined meeting with Scandinavian Scoliosis Societies. J Bone Joint Surg (Br). 1992;74-B:98.
11. Iwahara T, Imai M, Atsuta Y. Quantification of cosmesis for patients affected by adolescent idiopathic scoliosis. Eur Spine J. 1998;7:12–15.
12. Edgar MA, Mehta MH. Long-term follow-up of fused and unfused idiopathic scoliosis. J Bone Joint Surg Br. 1988;70:712–716.
13. Schwender JD, Denis F. Coronal plane imbalance in adolescent idiopathic scoliosis with left lumbar curves exceeding 40 degrees: the role of the lumbosacral hemicurve. Spine (Phila Pa 1976). 2000;25:2358–2363.
14. Richards BS. Lumbar curve response in type II idiopathic scoliosis after posterior instrumentation of the thoracic curve. Spine (Phila Pa 1976). 1992;17(suppl):S282–S286.
15. Jackson RP, Simmons EH, Stripinis D. Incidence and severity of back pain in adult idiopathic scoliosis. Spine (Phila Pa 1976). 1983;8:749–756.
16. Ascani E, Bartolozzi P, Logroscino CA, et al. Natural history of untreated idiopathic scoliosis after skeletal maturity. Spine (Phila Pa 1976). 1986;11:784–789.
17. White SF, Asher MA, Lai SM, et al. Patients’ perceptions of overall function, pain, and appearance after primary posterior instrumentation and fusion for idiopathic scoliosis. Spine (Phila Pa 1976). 1999;24:1693–1699. Discussion 1699-700.
18. Weinstein S, Flynn J. Lovell and Winter’s Pediatric Orthopaedics, 7th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2014.
19. Herring J. Tachdjian’s Pediatric Orthopaedics from the Texas Scottish Rite Hospital for Children, 5th ed. Philadelphia, PA: Elsevier Saunders; 2014.

adolescent idiopathic scoliosis; fusion; nonoperative; thoracolumbar scoliosis

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.