Sizing Up Scoliosis

Idiopathic scoliosis is a curious spinal disorder that is remarkable for its contrasts. In some patients, scoliosis can be physically evident and often striking on plain radiographs even to the untrained eye. Yet in other young patients it is not physically evident or symptomatic, so clinicians must try to predict the effects far into the future.

The landmark study by Weinstein et al¹ in this issue of THE JOURNAL can help with this prediction. Just as longitudinal studies of populations in Framingham, Mass,² and the Precursors Study in Baltimore, Md,³ have provided bases of medical understanding of cardiovascular disease, the Iowa studies of the natural history of musculoskeletal disorders form a cornerstone of orthopedic treatment.⁴ The senior authors have been carefully following these patients for 50 years and have published several earlier analyses.^{5 - 6} This will likely be the final major report, and it is unlikely that a natural history study of such length and breadth will be matched.

The authors point out that some earlier studies of scoliosis combined different forms of the condition, including patients with congenital, neuromuscular, and early onset.⁷ This has led to conclusions of morbidity, which were confounded by other factors. The authors of the current study focus solely on the most common form, late-onset idiopathic scoliosis (LIS), commonly called adolescent idiopathic scoliosis, which is diagnosed after age 10 years and is often heritable.⁸ The cohort in this study was diagnosed at the University of Iowa before or during 1948, and exhaustive attempts were made to review as many as possible at a mean age of 66 years. The cohort excludes 58 patients of the original 444 who underwent spinal fusion.

The authors have previously shown that scoliosis increases slowly (by a mean of 23°) during adulthood if the curve is greater than about 45° at maturity.⁶ Also, pulmonary function has been shown to be decreased by thoracic scoliosis. The clinical correlation is an increased risk of shortness of breath associated with thoracic scoliosis of greater than 80°. Using various assumptions about the patients who could not be located, the authors conclude no statistically significant difference in the probability of survival compared with the general population. However, when analyzing the causes of 36 deaths since their last follow-up in 1981, the authors found that scoliosis potentially contributed to 3 deaths, with each patient having had curves greater than 100°. One patient in the 1981 report also died at age 54 years of cor pulmonale attributed to scoliosis.⁶

Function in life did not appear to be severely impaired for patients with LIS. There was no difference compared with controls in the number of patients who married, or in the number of children they had. The vast majority were either homemakers or gainfully employed in middle age. The scores on a depression measure were equivalent. However, back pain was more prevalent in patients than controls, although rarely leading to use of narcotics. There was an increase in the number of patients who felt they had a disability and who reduced their work due to back pain. There also was a decrease in body satisfaction. There was no objective evidence of neurological impairment in this cohort, although it does not include 8 patients who underwent fusion in adulthood, some of whom may have had such problems.

Several features of this study merit comment. The majority of the analysis involves 117 patients of an initial group of 444. Fifty-eight of these had undergone fusion in prior years. Although the reasons for fusion are not given, the patients are likely to have been among the more severely affected, and might therefore represent a selection bias. The cause of death is unknown in 20 of the 69 patients who are known to have died. In addition, 127 currently eligible patients were not located despite remarkable effort. Since the original cohort yielded at least 4 patients whose death may be linked to LIS, important information is not available on the remainder, which might have provided a more precise estimation of disease-related mortality than simply comparing the survival probability with the general population.

Although the depression scores of the current population do not differ from scores of controls, the first 2 parts of the Iowa series include 3 suicides, all in patients having thoracic curves of 70° or greater, which are the most deforming.^{5 - 6} Importantly, the current series includes a number of patients who have minor curves. Individual curve data are not given, but graphs from the 1981 follow-up show that 28% of the curves were less than 45° at maturity. Some of the patients included in the present series still have curves of 15° to 20°. It is reasonable to wonder if the patients who have more severe curves might be considered separately from the patients who have minor curves, since the former are the focus of most treatment decisions.

Current practice of managing scoliosis varies widely but can be generally characterized as follows. School screening for primary and secondary prevention in early adolescence is widely practiced but is not mandatory and results in a significant number of false-positive referrals.^{9 - 11}

Brace treatment is considered in actively growing children with curves of more than 25° since these patients are more likely than not to have an increase in curve before maturity. The efficacy of bracing in preventing worsening of curves is supported by evidence from prospective center randomized controlled trials and multicenter nonrandomized studies,^{12 - 14} but there is also evidence to question this treatment.¹⁵ No other nonoperative therapy has been shown to change the natural history of scoliosis. Surgical correction of the curve with fusion is an option for adolescents with curves greater than 45° to prevent worsening in adulthood. Surgery is more strongly favored by many surgeons for thoracic curves, especially above 60°, because of the potential pulmonary effects of such curves. Some adults also seek surgery because of pain, neuropathy, or increasing deformity. Modern scoliosis surgery is effective in correcting curves and has a low complication rate, especially in adolescence.¹⁶ However, it is expensive and can be followed by pain at other levels of the spine. There are no randomized trials of bracing or surgery compared with nontreatment for patients with scoliosis.

The studies by Weinstein et al provide important information for advising patients as well as for policy. This series allows a framework for clinicians to predict future possibilities for patients. The patient who does not have a large thoracic curve may not have any increased risk of mortality due to the scoliosis. Some patients with large (Cobb angle >80°) thoracic curves may eventually develop shortness of breath or other pulmonary problems. The moderate-to-large (Cobb angle >45°) curves will likely increase slowly, along with any attendant deformity. Back pain is more prevalent than in the general population, but is not usually disabling. It varies greatly between patients for reasons that are still unknown. The risk of neuropathy is low. Since bracing and surgery are not problem-free, patients can be counseled to match treatment to their preferences.

Future goals for policy include refining screening programs to reduce the substantial over-referral by screening only the most critical population, setting appropriate criteria for positives, and managing them efficiently.⁹ Shared decision-making strategies with patients and parents can improve their ability to make well-informed choices.