### Results

#### Characteristics of the Study Population

The study population consisted of 51 patients, 3 males and 48 females, with a mean age at the time of the first radiograph of 37 years (range 17-60) and at the time of the last radiograph of 64 years (range 44-80).

The mean duration of the natural history was 27 years (range 9-53). All but 8 of the patients reported low back pain, 22 reported nerve root pain, and 4 were pain-free. The various topographies were 30 single major curves (19 lumbar and 11 thoracolumbar), 18 double curves (double thoracic and lumbar), and 3 triple curves (in accordance with the Scoliosis Research Society classification).

The mean height of the thoracic rib hump was 11 mm (range 4-30), and the mean height of the lumbar or thoracolumbar hump was 16 mm (range 4-36). All the patients had at least 1 rotatory subluxation, except 1. The sites of rotatory subluxation were as follows: 2 at T10-T11, 9 at T11-T12, 14 at T12-L1, 5 at L1-L2, 5 at L2-L3, 34 at L3-L4, and 11 at L4-L5. Overall, there were 51 thoracolumbar and lumbar curves, including 30 single major curves. The values of the studied parameters are presented in Table 1 .

#### Results of Linear Regression Analysis Applied to Each Curve

Linear regression analysis was very significant in all but 5 of the patients ( Table 2 , Table 3 and Table 4 ). The slope of the regression line represents the rate of progression of each curve. The mean slope for lumbar and thoracolumbar curves (single curves and lumbar component of double curves) was 1.23°/y (range 0.34° to 3.82°). The mean slope of the thoracic curve in patients with double curves was 0.68°/y (range 0.3° to 0.97°) ( Table 4 ). The mean slope of the lumbar component in patients with double curves was 0.96°/y (range 0.34° to 1.67°) ( Table 2 ). The mean slope of the lumbar or thoracolumbar single curve was 1.45°/y (range 0.46° to 3.82°) ( Table 3 ). The double curves should be treated as a unit. We can only note that there were 6 cases where both curves progressed in the same way, only 1 where thoracic curve progressed more quickly than the lumbar curve. In 3 cases, lumbar curve progressed faster than thoracic curve. In 10 cases, it is difficult to conclude because the lower rate of progression of thoracic curves was not statistically significant, possibly because of a limited number of radiographs, indicating the need for longer follow-up or a larger number of radiographs. But in these 10 cases, the lumbar curve progressed faster.

#### Diagrams Visualizing Various Types of Progression of Scoliosis

Two main types of progression of adult scoliosis can be distinguished. Type A corresponds to adolescent scoliosis, which continues to progress after skeletal maturity at a rate specific to each curve. Thirteen patients presented this type of scoliosis (Figures 1-4, 9). Type B corresponds to a degenerative scoliosis, which progresses late in adulthood: either a preexisting stable adult scoliosis that progresses late or a *de novo* late-onset scoliosis. Twenty patients presented this type of scoliosis (Figures 5-9). Eleven of these 20 patients progressed certainly at the time of menopause (type BM, which is scoliosis that progresses at menopause) (Figures 6-8). BM is a subtype of B.

Figure 1.

Type A: Female, double major, men: 0, age rotatory subluxation 32 years (n = 7).

Figure 2.

Type A: Female, double major, men: 46 years, age rotatory subluxation: 32 years (n = 10) (insufficient number of thoracic points). L indicates lumbar single curve; TL, thoracolumbar single curve.

Figure 3.

Type A: Female, double major, men: 50 years, age rotatory subluxation: 47 years (n = 7).

Figure 4.

Type A: Male, lumbar single curve, age rotatory subluxation: 45 years (n = 10).

Figure 5.

Type B: Female, lumbar single curve, age men: unknown, age rotatory subluxation: 65 years (n = 8).

Figure 6.

Type BM *de novo*: Female, thoracolumbar single curve (TL), men: 50 years, age rotatory subluxation: 52 years (n = 7). L indicates lumbar single curve.

Figure 7.

Type BM *de novo*: Female, thoracolumbar single curve, men: 51 years, age rotatory subluxation: 52 years (n = 6).

Figure 8.

Type BM: Female, lumbar single curve, stable adolescent scoliosis then progressing at menopause, men: 50 years, age rotatory subluxation: 53 years (n = 13).

Figure 9.

Graph of all the curves type A and B.

The remaining 13 patients belonged to 1 of the aforementioned types, but the distinction between types A and B was impossible due to the absence of radiographs during early adulthood.

Other types of scoliosis also probably exist. For example, Figure 10 shows a very rapid progression from a small angle compared to type B but early in adulthood. Figure 10 shows a progression of 2.09°/y, with a Cobb 1 of 24° on the first radiograph at the age of 35 years. The x-intercept of these lines is also observed after skeletal maturity.

Figure 10.

Other type: Female, thoracolumbar single curve, men: 0, age rotatory subluxation: 46 years (n = 4).

No change of slope was observed on any of the 46 progression diagrams. In particular, in 8 women with type A scoliosis with a long progression comprising menopause, no change of slope was observed at menopause (Figures 2, 3), and no change of slope was observed after the development of rotatory subluxation in type A (Figures 1-3).

We did not find any correlation between the initial Cobb angle and slope of progression in the overall population, as a marked scatter of the slope was observed for each angle. We also did not observe any significant correlation in patients with type A scoliosis.

#### Comparison of the Parameters of Types A and B Scoliosis

Patients with type B scoliosis were all women and exclusively presented a lumbar or thoracolumbar single curve. Type A patients predominantly presented double curves, in 9 cases, with 1 triple curve, and 3 single curves (3 lumbar). The delay between the first radiograph and last radiograph was not significantly different between type A (mean 30 years; range 18-53) and type B (mean 24 years; range 9-41). The number of radiographs per patient was also not significantly different between type A (mean 7; range 3-10) and type B (mean 7; range 4-12). However, the following parameters were found to be significantly different between types A and B:

Loss of height: 5 cm for type A (0.18 cm/y) and 10 cm for type B (0.46 cm/y) (

*P*= 0.001).

Lumbar or thoracolumbar slope (or rate of progression): 0.82°/y (0.34-1.65) for type A and 1.64°/y (0.77-3.82) for type B (

*P*= 0.004).

Mean age at the time of the first radiograph (age 1): 24 years for type A and 46 years for type B.

Mean Lumbar and thoracolumbar Cobb angle of first radiograph (lumbar single and thoracolumbar single Cobb 1): 37° for type A (range 22° to 52°) and 20° for type B (range 3° to 35°) (

*P*= 0.000).

Mean age of first rotatory subluxation (age rotatory subluxation): 42 years for type A and 56 years for type B (

*P*= 0.000).

Mean Cobb angle of first rotatory subluxation (Cobb rotatory subluxation): 52° for type A and 29 for type B (

*P*= 0.000).

The lumbar single and thoracolumbar single Cobb angle of the last radiograph and thoracic kyphosis were at the limit of significance for type A *versus* type B. No significant difference was observed for the other parameters: coronal imbalance, sacral inclination, lumbar lordosis, number of vertebrae involved in the lordosis, and pelvic incidence.

The first rotatory subluxation occurred after an average of 18 years of progression in type A but was present at the time of onset of type BM. The only patient without rotatory subluxation presented type A scoliosis.

Spine. 2007;32(11):1227-1234. © 2007 Lippincott Williams & Wilkins

## Comments