Abstract
Introduction: The predominant classification systems for scoliosis have been based either on etiology or curve pattern. Traditionally triple curves have not been included in either classification system.
Methods: We reviewed the records of all children with an idiopathic triple scoliotic curve seen between 1988 and 2001 at the Children’s Hospital of Eastern Ontario, Ottawa, a major pediatric referral centre. The triple curve had 3 structural curved segments in accordance with the classification of Lenke and associates. Scoliosis was defined by a curve greater than 10°, and the apical vertebra was determined according to Scoliosis Research Society guidelines. Laterality of the curve was determined from the curve with the greatest magnitude. Curve progression was an increase in magnitude of at least 5°. Management included observation, with bracing if any curve progressed to greater than 20°. Surgery was reserved for curves that were unstable, rapidly progressive and greater than 45°–50°.
Results: Thirteen children (9 girls, 4 boys) were treated for a triple scoliosis curve. The average age at the time the triple curve developed was 13 years and 9 months (range from 9 yr 1 mo–17 yr 5 mo). The average follow-up was 3 years and 6 months (range from 1–11 yr). One child required spinal fusion, and 6 years postoperatively the curve was stable with a good fusion mass. Of the remaining children, 2 were treated with an orthosis and 10 with observation.
Conclusion: The triple major curve is a neglected curve pattern that should be included in classifications of scoliosis.
The predominant curve patterns in idiopathic scoliosis are single and double curves; triple curves are rare. The triple curve involves 3 distinct scoliotic curves, occurring in the cervical, thoracic and lumbar vertebrae or in the thoracic, thoracolumbar and lumbar regions. In this type of scoliosis, the most significant curve is usually the middle one. This pattern frequently presents with a small upper thoracic or cervicothoracic curve with larger curves in either the lower thoracic, thoracolumbar or lumbar regions and a high incidence of lordosis throughout the thoracic and lumbar spine.1
Triple curves are rarely mentioned in the literature, reflecting the infrequency with which this pattern occurs. In a series of 141 children, Stokes and colleagues2 reported a frequency of idiopathic triple curves of 1.4%. The mean Cobb angle for all the curves was 21.7° (range from 5°–61°). In contrast, Cruickshank and associates1 reported a rate of 28%. Cases of triple curves in children with congenital scoliosis secondary to hemivertebrae have been reported,3 and a rate of 10.7% was reported in children with Friedrich’s ataxia.4 It has been suggested that idiopathic triple curves have a benign natural history and rarely progress, consequently treatment is often not required.5
In 1998, Coonrad and associates6 described a classification for scoliosis based on the coronal pattern of the curve on standard anteroposterior radiographs of the spine. Triple curves were classified as type 8 curves, and laterality was determined by the curve with the greatest magnitude. In their study, there were 19 left triple curves and 26 right triple curves among 2000 consecutive patients with idiopathic scoliosis. More recently, Lenke and associates7 proposed a new classification based on curve type, lumbar spine deformity and thoracic spine alignment. Type 4 curves were triple major ones in which each curve was structural and either of the 2 lower curves were the major curves. In a review of 315 surgically treated cases, they reported 8 (2.5%) triple curves and recommended posterior spinal fusion for surgical stabilization.
To investigate the natural history and treatment of triple scoliotic curves, we reviewed our experience with this curve pattern at a major pediatric referral centre, with emphasis on the clinical course and management.
Methods
We reviewed the clinical, radiologic and operative records of all children who presented for evaluation of scoliosis between 1988 and 2001, with either a triple curve or development of a triple curve during follow-up of other curve patterns. Our definition of the triple curve was one with 3 structural curved segments, in accordance with the classification of Lenke and associates.7 For inclusion in the study, the curve etiology was required to be idiopathic. All curves were measured according to the Cobb method, the presence of scoliosis was defined by a structural curve greater that 10°, and the apical vertebra was determined in accordance with Scoliosis Research Society guidelines. Laterality of the curve was determined on the basis of the curve with the greatest magnitude. Curve progression was defined as an increase in magnitude of at least 5°. The management strategy utilized was similar to that in the setting of any case of idiopathic scoliosis. Regular observation was employed initially and bracing instituted if any curve progressed to greater than 20°. Surgical intervention was reserved for curves that were unstable, rapidly progressive and greater than 45°–50° in magnitude.
The English medical literature was thoroughly reviewed for reports of triple curves, and the details were compared to those presented in this study.
Results
There were 13 children (9 boys, 4 girls), ranging in age from 1 to 11 years and having an idiopathic triple curve. The average length of follow-up was 3 years and 6 months.
Each curve was analyzed according to several parameters (Table 1). At the time of initial presentation, 5 children had a triple curve, 5 had a double curve, and 3 had a single curve. Of the 8 children who did not initially present with a triple curve, the last curve to appear was the upper curve in 4 and the lower curve in 3. Both the middle and lower curves appeared at the same time in 1 child. The average age at the time the triple curve developed was 13 years 9 months (range from 9 yr 1 mo–17 yr 5 mo). One boy (case 2) initially presented at 9 years 5 months of age with a single thoracic curve, convex left, with subsequent development of a triple curve (Fig. 1). Laterality, determined from the largest curve, was the upper curve in 6 children, the middle curve in 6, and the lower curve in 1 child. The curve pattern was left in 7 (54%) and right in 6 (46%).
The average magnitude of the upper curve was 25° (range from 10°–49°). The middle curve averaged 27° (range from 15°–62°), and the lower curve averaged 19° (range from 10°–38°) (Table 2). The most progressive curve was the upper curve in 3 cases, the middle curve in 3, the lower curve in 1 case, and there was no difference between the curves in 6 (46%). The most common level of the apical vertebra of the upper curve was at the T3–4 level (7 cases), of the middle curve at the T8–9 level (4 cases) and of the lower curve at the L2-3 level (4 cases).
Treatment consisted of periodic observation in 10 (77%) children. Bracing was used in 3 children (cases 3, 6 and 7) and was effective in preventing significant progression in 2. The triple curve developed subsequent to bracing in 2 children (cases 6 and 7). The third curve was structural in both children and no reason for its development after bracing could be determined. The remaining child required surgical intervention to stabilize the curve due to rapid progression, with the upper curve measuring 49°, the middle curve 62°, and the lower curve 36°. Posterior spinal fusion was performed at 11 years 10 months of age (Fig. 2). The fusion extended from the second thoracic vertebra to the first lumbar vertebra, excluding the lower curve, which was the smallest curve and corrected on bending films (Fig. 2). Good correction was obtained and the curve was stable 6 years postoperatively.
Of the 8 children who did not initially present with a triple curve, only 1 child had development of the triple curve before the onset of maximum adolescent growth. In this girl (case 3), the triple curve developed at the age of 9 years 1 month, and the curves progressed rapidly, necessitating surgical stabilization. In the remaining children the triple curve developed after maximum adolescent growth, at an average age of 13 years (range from 12 yr 7 mo–17 yr 5 mo). Each child was managed with either observation or bracing, since the curve was stable and did not progress rapidly.
Discussion
The predominant curve patterns in scoliosis are single and double curves, usually in the thoracic and lumbar regions. Triple curves are rare and most commonly involve a small upper thoracic curve with 2 larger curves in either the lower thoracic, thoracolumbar or lumbar region.1 In this study, the largest curve was the middle curve in 6, the upper in 6 and the lower curve in 1. Laterality of the curve was determined by the largest curve with 7 left and 6 right curves. There has been no substantial reported predilection for right or left curves — Coonrad and associates6 documented 26 right curves in a series of 45 children having triple curves.
Idiopathic scoliosis, the most common form of scoliosis in children, predominantly affects girls and has been divided into 3 age groups: infantile, juvenile and adolescent. In this study, there was a female:male ratio of 2:1, and there were 12 adolescent idiopathic triple curves and 1 juvenile idiopathic curve. Cruickshank and associates1 reported 25 cases in a series of 90 children (28%). Of all curve patterns in their series, the triple curve was the most prevalent. Stokes and colleagues2 reported 2 cases in a series of 141 children (1.4%) in whom the average Cobb angle of the curves was 21.7° (range from 5°–61°). The average magnitude of the curves was similar in this series: the upper curve averaged 25°, the middle curve 27° and the lower curve 19°. Lenke and associates7 reported 8 triple major curves among 315 (2.5%) surgically treated cases of scoliosis.
Triple curves have also been reported in association with neuromuscular disorders. Labelle and colleagues4 reported 6 cases in a series of 56 patients (10.7%) with Friedrich’s ataxia, but no specific details were provided. Review of their cases led to the conclusion that the course of the scoliosis was more consistent with idiopathic scoliosis than neuromuscular scoliosis because of similar curve patterns, a lack of significant progression, no relationship to muscle weakness and a poor prognosis in cases of prepubescent development of scoliosis. 4 These findings were corroborated by another study of triple curves in children with Friedrich’s ataxia.8
The natural history of triple curves has not been thoroughly investigated. Travaglini5 reported that this curve pattern displayed a benign natural history and rarely progressed, consequently treatment was not often required. Similarly, in this series 10 (77%) of the 13 children did not demonstrate significant curve progression, and only observation was required until skeletal maturity. Eight children did not initially present with a triple curve. Although single and double curves may progress to triple curves, an etiologic explanation for progression to a triple curve pattern could not be elucidated and warrants further study. Triple curves differ from the more common double major thoracic curve with a lumbar compensatory curve because the third curve is structural and therefore will not respond to correction of the other 2 curves. The child who had development of a triple curve before maximum adolescent growth was the only one who required surgical intervention. So it appears that, like most idiopathic curves, triple curves that develop before or during peak adolescent growth will more likely progress and require further intervention. Of the 3 children who required bracing, the triple curve developed subsequently in 2 (cases 6 and 7), at the ages of 13 years 5 months and 15 years 6 months. These children were past the maximum adolescent growth rate, so there was no significant curve progression and further treatment was not required. The possibility that the triple curve developed in response to the brace cannot be excluded; however, development of additional structural curves after bracing is not common among children with idiopathic scoliosis. The exact mechanism of development of the additional curves is unclear and requires further investigation. On the basis of the experience in this study, it would seem reasonable to apply similar management strategies for triple curves as those recommended for all idiopathic curves. Initially, routine observation should be employed. Bracing should be started if any curve progresses to a magnitude greater than 20°, with surgical stabilization being indicated for rapidly progressive curves, particularly before peak adolescent growth. The decision on whether to fuse all 3 curves or just the largest 2 should be based on the degree of correction of the smallest curve on bending films.
Mention of triple curves in the literature is rare, and in studies with reported cases the rate has been 11% or less.3,4,6,7 The exception to this is the series reported by Cruickshank and associates1 (28%). These authors rationalized their high rate of triple curves by reporting that inadequate examination of the upper thoracic spine resulted in a lack of appreciation of a small, structural curve. In their experience, examination of the upper thoracic spine allowed for recognition of this small curve, increasing the incidence of triple curves. There have been no subsequent studies in the literature that have corroborated this finding. In our series, 8 children did not initially present with a triple curve, but a triple curve developed during the follow-up. We recommend observation of all children with scoliosis until they reach skeletal maturity. It is uncertain if radiographic examination at the time of skeletal maturity would identify additional triple curves among children being followed for single or double curves and if additional curves were discovered, the significance of this is unclear and is an area for future study.
In their study, Cruickshank and associates1 reported lordosis throughout the affected vertebral segments in all cases of triple curves with a non-rotated vertebral body between adjacent curves. They postulated that an area of lordosis balanced by adjacent areas of kyphosis would produce a single curve, while an extended lordotic area throughout the spine resulted in a triple curve since there was no area of kyphosis to compensate for the forces applied to the spine.1 There have been no subsequent reports confirming these findings. In this study, there were no instances of excessive lordosis or spinal deformity.
Treatment of triple curves most commonly consists of monitoring the curve for progression. Although the numbers in this series are small, 10 of the 13 cases were managed with observation. In curves with a more significant magnitude, bracing should be employed, with surgical stabilization being reserved for unstable, rapidly progressive curves. In each of the 8 cases of triple curves reported by Lenke and associates7 posterior spinal fusion was performed. To date, the surgical treatment of choice for triple major curves, based on the cases reported by Lenke and associates7 and in our cases, is posterior fusion. Further investigation of a larger number of triple curves is required, however, to further define the optimum surgical management.
Conclusions
Although our series comprised a relatively small number of patients, several trends can be reported. A 2–1 female preponderance, as seen in idiopathic scoliosis, is associated with these triple curves. The largest and most progressive curve is most commonly the middle curve. The lower curve seems to be the least progressive. A small number of single and double curve patterns may eventually progress to triple curves. Triple curves that develop before skeletal maturity will more likely progress and require fusion. In our experience, spinal fusion should be directed at fusing the 2 largest and most progressive curves. Further investigations are required to study the true incidence, clinical course, and management of triple curves.
Footnotes
Competing interests: None declared.
- Accepted January 7, 2003.