Copyright 1982 by The Journal ofBone and Joint Surgery.

Incorporated

The Natural History of Congenital Scoliosis

A STUDY OF Two HUNDREDAND FIFTY-ONEPATIENTS
BY MICHAEL J. MCMASTER, M.D., F.R.C.S.*, AND KUNIYOSHI OHTSUKA, M.D.t, EDINBURGH, SCOTLAND

From the Edinburgh Scoliosis Unit, Edinburgh

ABSTRACT: The cases of 251 patients with untreated imbalance in the longitudinal growth of the spine. These
congenital scoliosis were studied, and 216 patients were vertebral anomalies develop during the first six weeks of
followed without treatment for an average of 5.1 years. intrauterine life, when the anatomical pattern of the spine
Most of the curves were first seen when the patient was is formed in mesenchyme . Once the mesenchymal mold is
either in the first two years of life or at puberty, when established, the cartilaginous and osseous stages follow
there was an increased rate of deterioration. An early that pattern. The vertebral abnormality is present at birth,
onset carried a bad prognosis. Of the 143 patients who but the clinical deformity may not become evident until
were last seen, without treatment, after the age of ten later in childhood when a scoliosis develops and the diag
years, fifty-one (36 per cent) had a curve of 40 to 60 de nosis can be made radiographically. Some anomalies
grees and forty (28 per cent) had a curve of more than cause so little deformity that they remain undetected, so
60 degrees. We concluded that the prognosis in such the true incidence of congenital scoliosis in the general
patients is even worse because sixty-seven patients in population remains unknown. Wynne-Davies found that
our series required treatment at or before the age of ten multiple vertebral anomalies were often hereditary in ori
years due to the severity of the curve: twenty of them gin, but the occurrence of an isolated single anomaly was
had a curve of 40 to 60 degrees and thirty-nine had a usually sporadic.
curve of more than 60 degrees. The radiographic appearance of the vertebral
Radiographically, 90 per cent of the curves could anomalies varies considerably and, as a result, congenital
be classified into five groups. The rate of deterioration scoliosis was for many years thought to be unpredictable in
and the ultimate severity of the curve were found to de its behavior. Some thought that it seldom required treat
pend on both the type of anomaly and the site at which ment―3.In 1952, Kuhns and Hormell reviewed the cases
it occurred. Increasingly severe and progressive of 165 children and concluded, as did many physicians at
scoliosis, regardless of the area of the spine affected, that time, that a congenital scoliosis usually is relatively
developed when there was a block vertebra, a wedge benign and progresses slowly, if at all. It was not until
vertebra, a single hemivertebra, two unilateral 1968 that Winter et al. , in a study involving 234 children,
hemivertebrae, a unilateral unsegmented bar, or, most firmly established the much more serious prognosis for
severe, a unilateral unsegmented bar with contralat certain types of congenital scoliosis . Apart from these two
eral hemivertebrae at the same level. For each of these papers , there have been very few reports of the natural his
types of anomaly, the rate of deterioration was usually tory of congenital scoliosis in large numbers of unselected
less severe if the abnormality was in the upper thoracic patients6.
region, more severe in the thoracic region, and most A congenital scoliosis often is rigid and correction
severe in the thoracolumbar region. The median yearly can be difficult. It therefore is important to be able to antic
rate of deterioration for each type and site of curve ipate when a congenital scoliosis is at risk for rapid de
without treatment before and after the patient was ten terioration and to initiate treatment when the curve is small
years old was evaluated. Secondary problems due to rather than to attempt the dangerous surgical salvage pro
tilting of the head, elevation of the shoulder line, de cedures that are necessary when the deformity is severe.
compensation of the trunk, pelvic obliquity, and the Planning such a prophylactic course of treatment requires a
formation of large secondary structural curves also oc more thorough knowledge of the natural history of all
curred, and contributed significantly to the over-all types of congenital scoliosis than is presently available.
disability and deformity. The purpose of this study was to investigate the
natural history of congenital scoliosis in a large number of
Congenital scoliosis is a lateral curve of the spine that patients . All came from one area and were seen at the
is due to the presence of vertebral anomalies that cause an Edinburgh Scoliosis Clinic, and most were followed for
long periods. This clinic is the only referral point for pa
* Princess Margaret Rose Orthopaedic Hospital, Fairmilehead,
tients with congenital scoliosis from a large population. It
Edinburgh EH1O 7ED, Scotland. Please address reprint requests to Dr. was hoped that this paper would provide indications as to
McMaster.
t Department of Orthopaedic Surgery, Shinshu University, Mat
the necessity for treatment of the various types of congeni
sumoto, Japan. tal scoliosis.

1128 THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1129
Classification Materials and Methods
We classified the types of vertebral anomaly causing A study was made of all patients seen at the Edin
congenital scoliosis in the patients in this study according burgh Scoliosis Clinic between 1958 and 198 1 who were
to the system of Winter et al. We added one subcategory, diagnosed as having congenital scoliosis . The majority of
as described by Nasca et al. , because it has a particularly these patients were seen after 1965 . Patients who had
severe prognosis. Under the category of simple anomalies, myelomeningocele, Klippel-Feil syndrome, or congenital
we included failure of segmentation, which consists of: (1) kyphosis or kyphoscoliosis in which the kyphosis was the
a unilateral unsegmented bar, (2) unilateral failure of seg predominant deformity were excluded because the natural
mentation with contralateral hemivertebrae at the same history of those deformities was very different from that of
level6 and (3) bilateral failure of segmentation (block ver the ones under study . There were 25 1 patients who had a
tebrae). true congenital scoliosis that was diagnosed radiographi
The other simple anomaly in our series was unilateral cally; that is, each had a lateral curve of the spine due to
failure of formation of a vertebral body, which was either: the presence of congenital vertebral anomalies . Radio
(1)complete
(hemivertebra)
or(2)partial
(wedgever graphs made elsewhere, prior to referral, were often avail
tebra). able, and sometimes provided evidence of progression of
Complex (unclassifiable) anomalies consisted of a the deformity.
jumble of vertebral anomalies that were too varied for There were 179 female and seventy-two male pa
separate classification. In some patients included in this tients, whose ages at diagnosis ranged from birth to
group it was impossible to identify the type of anomaly be twenty-two years (Fig . 1). They had a total of 269 curves
cause the severity of the curve was so great. (Table I).
Occasionally other congenital vertebral anomalies Two hundred and sixteen patients were followed for a
were present at levels in the spine other than those involv mean of 5. 1 years (range, six months to seventeen years).
ing the scoliotic curve, but if they did not contribute to the At their last visit to the clinic, 108 patients had not been
scoliosis they were ignored in the classification. Anoma treated, and forty-six of them had reached skeletal matur
lies of the ribs were very common, but they too were ig ity. The mean length of follow-up for these patients was
nored because they did not cause a scoliosis by them 7 .3 years . The sixty-two patients who were skeletally im
selves. mature were still attending our clinic, without treatment,
In addition to determining the type of vertebral after a mean of 5 . 1 years . The remaining 108 patients had
anomaly, the scoliosis was also classified according to the been followed without treatment for a mean of 4.2 years
site of the apex of the curve, which frequently coincided before eventually being treated either with a Milwaukee
with the site of the anomaly. A scoliosis whose apex lay brace or with spine fusion; the data on these patients relate
between the second and sixth thoracic vertebrae, inclu to the pretreatment interval.
sive, was termed an upper thoracic curve. A scoliosis No longitudinal data could be obtained on the natural
whose apex was located between the seventh and eleventh history of the remaining thirty-five patients, since twenty
thoracic vertebrae , inclusive , was a lower thoracic curve. three of them received immediate treatment when they
A thoracolumbar scoliosis was one with its apex at the were first seen because of the severity of the deformity and
twelfth thoracic or first lumbar vertebra. The apex of a the other twelve were skeletally mature when first seen.
lumbar scoliosis was between the second and fourth lum The findings in these patients are included in some of the
bar vertebrae, inclusive, and the apex of a lumbosacral analyses in this study because prior to attending our clinic
scoliosis was at the fifth lumbar vertebra8. they had been untreated, and therefore contributed data on
A patient with a single congenital scoliosis was con the natural history of congenital scoliosis.
sidered to have only one structural curve due to the pres The sites and types of congenital scoliosis were diag
ence of congenital vertebral anomalies . Several patients nosed by reviewing the radiographs of the spine , made
were classified as having a single congenital scoliosis even with the patient standing. Using the Cobb method, care
though a second structural scoliosis developed either was taken to remeasure from the same levels that were
above or below the primary congenital curve. This second used on the earliest radiograph. Pelvic obliquity was mea
curve was not classified as a congenital scoliosis because it sured on the radiographs of the spine by noting the angle
did not result from a congenital vertebral anomaly. subtended between a line drawn across the tops of the iliac
A patient was considered to have a double congenital crests and the bottom of the radiograph , which was hori
scoliosis if there were two opposing structural curves, zontal when the patient was standing . Decompensation or
each due to separate congenital vertebral anomalies that listing of the trunk and tilting of the head were assessed on
might or might not be similar. A few patients had more the basis of the spine radiographs and the clinical photo
than two structural curves because of separate congenital graphs made with the patient standing.
vertebral anomalies.
Skeletal maturity was determined radiographically Results
when there was complete ossification and fusion of the The incidences of the various sites and types of con
iliac @‘¿. genital scoliosis are shown in Table I.

VOL. 64-A, NO. 8, OCTOBER 1982

 1130 M. J. McMASTER
AND KUNIYOSHIOHTSUKA

TABLE I
SITES AND T@PES OF CONGENITAL
PATIENTS)Type
ScoLIosIS (Two HUNDRED AND FIFTY-ONE

Curves)UnilateralUnsegmentedUnilateralBarof Congenital Vertebral Anomaly (No. of

withSite
BlockWedgeCurveBarHemivertebrae
ofUnsegmentedContralateral
HemivertebraVertebraUnclassifiableUpperthoracic407 Vertebra
2315Lowerthoracic3911 7
21413Thoracolumbar1710 2
1338Lumbar60 3
1913Lumbosacral00 1
0 1200

A single congenital curve occurred in 238 patients: seen not only in older children but also at birth and in the
216 of the curves were due to simple anomalies and first year of life. Minor curves were seen in patients in all
twenty-two, to complex anomalies. age groups.
Two or more congenital curves occurred in thirteen
patients. In ten of the patients the curves were due to sim Severity of the Curve and Rate of Deterioration
plc anomalies and in three the anomalies were complex. Of the 25 1 patients in this study, 173 were seen, un
The ages of the patients and the sizes of the congeni treated, at or before the age of ten years. Thirty-eight pa
tal curves at the initial visit are shown in Figure 1. In pa tients (22 per cent) had a curve of 20 degrees or less,
tients with two or more congenital curves, only the largest forty-seven (27 per cent) had a curve of 2 1 to 39 degrees,
curve was used for tabulation. Most of our patients were forty-one (24 per cent) had a curve of 40 to 60 degrees,
diagnosed radiographically either in the first two years of and forty-seven (27 per cent) had a curve of more than 60
life (seventy-seven patients) or between the ages of nine degrees (range, 62 to 148 degrees).
and fourteen years (sixty-eight patients). The degree of Sixty-seven of these 173 patients were eventually
severity of the curve did not correlate with the age at pre treated at or before the age of ten years because of the
sentation. Severe curves of more than 80 degrees were severity of the deformity. Before treatment, eight had a

50.

40

size c@Curve

@ 800 +

30
@ 40°- 790

Number 0 o°
-39°
Patteits

16

11

1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Age at Dlagno.ia (year.)
FIG. 1
Ages of the patients and severity of the congenital scolioses at diagnosis.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1131

curve of 20 to 39 degrees, twenty had a curve of 40 to 60 Four patients had two hemivertebrae, but on opposite
degrees, and thirty-nine had a curve of more than 60 de sides and at different levels of the spine. In two of these
grees. patients (Cases 4 and 5) the hemivertebrae occurred in the
One hundred and forty-three patients were last seen, thoracic region, within one or two segments of each other
without treatment, after they were ten years old. Fifteen of and on opposite sides of the spine. These hemivertebrae
them (10 per cent) had a curve of 20 degrees or less, caused two small kinks that never became large and that
thirty-seven (26 per cent) had a curve of 21 to 39 degrees, balanced each other, producing a minimum deformity at
fifty-one (36 per cent) had a curve of4O to 60 degrees, and maturity. Two other patients (Cases 6 and 7) had opposing
forty (28 per cent) had a curve of more than 60 degrees hemivertebrae that were much more widely separated, on
(range, 64 to 156 degrees). opposite sides, and in different regions of the spine. These
Fifty-eight patients reached skeletal maturity without hemivertebrae produced bigger curves that were unbal
treatment. Of these patients, six (10 per cent) had a curve anced, causing a list of the trunk and requiring correction
of 20 degrees or less, twenty-two (38 per cent) had a curve and fusion at the ages of ten and fourteen years.
of 2 1 to 39 degrees , nineteen (33 per cent) had a curve of Two patients (Cases 8 and 10) had three hemiverte
40 to 60 degrees , and eleven ( 19 per cent) had a curve of brae and one patient (Case 9) had four hemivertebrae
more than 60 degrees (range, 66 to 156 degrees). which alternated on either side of the spine in the thoracic
All of the large curves that were followed without and thoracolumbar regions. The resulting curves tended to
treatment should have been treated at a much earlier stage, balance each other, but in one patient (Case 8) the upper
regardless of the patient's age. hemivertebra was at the first thoracic level and this caused
We found, as did Winter et al., that the rate of de an elevation of the shoulder line, requiring correction and
tenioration and the ultimate severity of a congenital fusion of the upper curve at the age of eight years.
scoliosis depended on both the type of vertebral anomaly Three patients had multiple small curves due to com
and the site of the curve. In addition, we found that the rate plex anomalies. These curves were closely associated with
of deterioration was not constant but tended to increase one another and tended to balance each other, causing little
after the age of ten years. deformity other than stunting of the spine. One patient,
In order to obtain a more complete understandingof however, had an elevated shoulder line that necessitated
the natural history of all types of congenital scoliosis, we correction and fusion of the upper thoracic curve at the age
studied the rate of deterioration in degrees per year and the of nine years.
severity of the curve , both before and after the age of ten
years , for each type of vertebral anomaly producing a Single Congenital Curves
scoliosis in each region of the spine (Tables II through For the purpose of analysis of the natural history of
VI). The degree of scoliosis was followed serially and re the curves, each of the six groups was divided into three
mained fairly constant until between the ages of nine and subgroups. Subgroup A comprised those patients for
eleven years, when it tended to increase, especially in pa whom serial observations had not been made because they
tients with unilateral failure of segmentation. The rate of were either treated immediately or were first seen, un
deterioration before and after the age of ten years was es treated, at skeletal maturity. Subgroup B contained those
. timated by subtracting the initial angle from patients who had been followed for variable periods before
the final angle

of the curve on the routine spine radiographs made with the age of ten years. Subgroup C contained those patients
the patient standing, and dividing the increment by the who had been followed at or after the age of ten years.
number of years and months. We calculated the median There was some overlap between Subgroups B and C.
rate of deterioration rather than the mean for each specific
Unilateral Failure of Segmentation
type of congenital scoliosis occurring at each specific site,
(Unilateral Unsegmented Bar) (Group 1)
and this value was found to be more representative of each
subgroup as a whole and less distorted by values that were A unilateral unsegmented bar was the cause of a con
greatly different from those of the majority (Fig. 8). genital scoliosis in ninety-nine patients and was always
present on the concavity of the curve. Ninety-six patients
Multiple Congenital Curves had a single congenital scoliosis and three (Table II, Cases
Before analyzing the single curves, we studied the 1, 2, and 3) had two separate and opposing unbalanced
cases of the thirteen patients with two or more congenital congenital curves, each due to a unilateral unsegmented
curves due to simple (Table II) and complex anomalies. bar but on opposite sides and at different levels of the
Three patients (Cases 1, 2, and 3) had double curves spine. The unsegmented bars in the ninety-nine patients
due to unilateral unsegmented bars on opposite sides and did not demonstrate a preference for side but were seen
in different regions of the spine. These double curves more frequently in the thoracic region (seventy-nine pa
showed a high rate of deterioration and became large and tients) than elsewhere. In three patients the unilateral un
unbalanced, causing a list of the trunk that necessitated segmented bar was not recognized in infancy and was only
correction and fusion soon after the patient was ten years diagnosed retrospectively, between the ages of three and
old. four years, when the bar began to ossify and appeared on

VOL. 64-A, NO. 8, OCTOBER 1982

1132 M . J . McMASTER AND KUNIYOSHI OHTSUKA

TABLE
NATURAL HISTORY OF TEN PATIENTS WITH Moan

Mos.)Length
AnomalyCurve of
CaseSexVertebral SeenTypeSite(Yrs. Side ExtentAge(Yrs.+ First SeenWhen
Follow-upWhen Last
Mos.)1FUnseg. +

62FUnseg. bar + 013 + 610 +

Unseg. barT4-T9 L2-L4L RT3-T12 L1-L53
13FUnseg. bar + 22 + 32 +
Unseg. barT2-T5 T11-L1L RT2-T5 Tll-L30
84FHemivert. bar + 211 + 01 +
Unseg. barT3-T5 Ll-L3R LTi-Ti T7.L38
95FHemivert. + 317 + 014 +
Hemivert.T3 T6L RT2-T4 T5-T72
36FHemivert. + 14 + 41 +
Hemivert.T2 T5R LT1-T3 T4-T63
07FHemivert. + 010 + 07 +
Hemivert.11 L2L RT2-T12 T12-L53
1None8MHemivert. 1-L4 + 1
Hemivert.Li L6L RTi LA-Si13

109FHemivert. Hemivert. TiO R T8-Tli + 38 + 11 +

Hemivert.Ti LiR LC7-T3 Ti1-LA6

0None10FHemivert. Hemivert. T9 R T8-Ti1 +

Hemivert. T12 L Tli-L1
Hemivert.T6 LAL RT4-T8 L3-L520

Hemivert. T8 R T7-T9 + 53 + 63 + 0
Hemivert.T5 LAL LT3-T6 Li-L50

radiographs . The rate of progression of these three curves years, ten of them had a curve of more than 30 degrees.
did not differ significantly from that of the other curves in Twelve patients were seen, untreated, at the age of ten
the same regions . The prognosis for congenital scoliosis years, when the mean curve was 33 degrees (range, 13 to
due to a unilateral unsegmented bar depended mainly on 76 degrees). Two patients had a more rapid rate of deterio
the site of the curve and partly on the extent of the bar. The ration before the age of ten years. One of them, who had a
extent ranged from two to eight vertebrae. However, there bar involving seven vertebrae, had a 5-degree curve at
was not always a direct relationship between the length of birth which deteriorated at a rate of 6 degrees per year until
the bar and the rate of progression of the curve. The longer it was 62 degrees at the age of ten years (Figs. 3-A and
unsegmented bars all tended to produce the larger curves 3-B). The second patient had a bar involving three verte
in a specific region, but occasionally a shorter bar pro brae which produced a 24-degree curve at the age of three
duced an equally large curve in the same region. years and deteriorated at a rate of 5 degrees per year until it
The seventy-two patients who were followed without was 53 degrees at the age of nine years . The case of this
treatment were considered to be in Groups lB and 1C patient demonstrated that even a relatively short unseg
(Table III). The twenty-four patients for whom serial ob mented bar can produce a large curve.
servations had not been made were in Group 1A and were After the patients were ten years old, the median rate
excluded from further analysis. of deterioration increased to 4 degrees per year, and six
teen patients required a spine fusion at a mean age of thir
Upper Thoracic Curves teen years (range, eleven to fifteen years), at which time
Of the thirty-eight patients who had an upper thoracic the mean curve was 59 degrees (range, 30 to 98 degrees).
curve , ten were in Group lA . Nine of them were treated The most rapid rates of deterioration occurred in four of
immediately when they were first seen, between the ages these patients, who were followed untreated prior to spine
of two and fifteen years , and had curves ranging from 26 to fusion for periods ranging from nine months to two years
85 degrees. One patient was seen at skeletal maturity, with and six months. A bar involving three to five vertebrae was
a curve of 68 degrees. present in three of these curves , which deteriorated at a
The remaining twenty-eight patients with an upper rate of 8 degrees per year. A bar involving seven vertebrae
thoracic curve were in Groups lB and 1C (Table Ill) and was present in one curve, which deteriorated at a rate of
were followed without treatment for a mean of 4.7 years more than 10 degrees per year and measured 90 degrees at
(range , nine months to fourteen years and two months). the age of twelve. Not all of the upper thoracic curves de
Before they reached the age of ten years , the median rate teriorated at the same rate, however, and five of the ten
of deterioration was 2 degrees per year. By the age of ten curves that were followed, untreated, to maturity then

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1133
II
THAN ONE CURVE DUE TO SIMPLE ANOMALIES

(Degrees)CommentsWhen Size of Curye

of Change
FirstYear22
SeenWhen Last SeenTotal ChangeRate per

old8 uppercurvecorrectedandfusedat 13yrs. and6 mos.

0122 75100 759.5 7Unbalanced;
old25 Milwaukee brace applied at 2 yrs. and 3 mos.
720 4012 336 16Unbalanced;
old2245231.5Balanced335219128 required correction and fusion at 11 yrs.
3240 5315 219 13Unbalanced;

2729
1Balanced39 281 11
old76 required correction and fusion at 10 yrs.
4454 6115 172 2Unbalanced;
old34 required correction and fusion at 13 yrs. and 11 mos.
42Unbalanced;
shoulder and upper curve required correction and fusion at 8 yrs.
35
old32 35 0 0 and 1 mo.
4134 450 40 2High

38
mature17161600Balanced151610181800 and skeletally
32Balanced

measured 30 to 40 degrees. The largest untreated curves at The remaining eight patients were followed without treat
maturity measured 68 and 88 degrees and were due to bars ment for a mean of 4.8 years (range, nine months to four
involving three and five vertebrae, respectively. teen years) . The median rate of deterioration for the five
Upper thoracic curves, especially those whose apex patients followed before the age of ten years was 6 degrees
lay at the second, third, or fourth thoracic vertebra, pro per year (range, 2 to 8 degrees) for the secondary structural
duced a significant cosmetic deformity due to elevation of curves and 3 degrees per year (range, 1 to 4 degrees) for
the shoulder line on the convex side of the curve. Tilting of the primary congenital curves. Five patients were followed
the head toward the concavity also occurred in those pa after the age of ten years and then the median rate of de
tients whose curves extended beyond the cervicothoracic terioration was 7 degrees per year (range, 3 to 10 degrees)
junction; the tilting resulted because a satisfactory com for the secondary structural curves and 4 degrees per year
pensatory curve failed to develop above the congenital (range, 2 to 8 degrees) for the primary congenital curves.
curve (Figs. 2-A and 2-B). The structural compensatory curve usually deteriorated at
In thirteen patients with an upper thoracic curve, with approximately twice the rate of the primary congenital
the apex at the fourth , fifth , or sixth thoracic vertebra, an curve. Six patients were treated between the ages of twelve
additional long structural curve developed in the lower and sixteen years , when the upper thoracic congenital
thoracic or thoracolumbar region (Figs. 3-A and 3-B). curves were a mean of 56 degrees (range, 30 to 85 de
Why such a curve should develop in some patients and not grees) and the lower secondary structural curves were a
in others is unknown, as it was not always associated with mean of 79 degrees (range , 53 to 104 degrees) . Three pa
the larger upper thoracic curves. This initially compensa tients reached skeletal maturity without treatment, at
tory curve, which involved no congenital anomalies, ap which time the upper thoracic congenital curves were 40,
peared secondarily; while it initially was compensatory 59, and 68 degrees and the lower secondary structural
and therefore could be corrected, later it became fixed and curves were 78, 75, and 82 degrees, respectively.
deteriorated even more rapidly, and it was soon more se
vere than the primary curve. The major deformity then was Lower Thoracic Curves
caused by the secondary structural curve, which was much Of the thirty-eight patients who had a lower thoracic
more rotated than the primary congenital curve and pro curve, nine were in Group 1A. Six of them were treated
duced a large rib hump. Five of the patients with this con immediately, between the ages of eleven and fourteen
dition were in Group lA; two were skeletally mature and years, with curves ranging from 41 to 96 degrees (mean,
three required immediate treatment of the secondary struc 62 degrees). Three were seen at skeletal maturity, with
tural curve between the ages of fourteen and sixteen years. curves measuring from 66 to 89 degrees.

VOL. 64-A, NO. 8. OCTOBER 1982

1134 M. J. MCMASTER AND KUNIYOSHI OHTSUKA

TABLE
UNTREATED SINGLE CONGENITAL Scouosls DUE

Years)No.Age (Patients Seen beforetheAge of Ten

of Curye (Degrees)*Rate
of When (Degrees)When Last Seen of Deterioration per Year
Site of CurveNo. CurvesGrouplB First Seen@Size First SeenWhenbefore 10 Yrs. Old123
910( 4 5 6 7 8
Yrs.)Upper

ILower thoracic28154 (0-9)29 (0.60)35 (13-76)634 1

4Thoracolumbar121
thoracic29212 (0-9)40 ( 12.60)61 (25-84)332 1 2 1 5
1Lumbar32(0- 13 (0-9)35 ( 15-75)58 (35-82)11 3 5
1)(8-62)(53-82)I1

* Median, with range in parentheses.

t The number of patients who were followed untreated until after the age of ten years but who were first seen before the age of ten years, and who are
also included in Group lB. is in parentheses.

The remaining twenty-nine patients with a lower per year) and nine curves deteriorated at a rate of 7 or 8
thoracic curve were in Groups lB and lC (Table III) and degrees per year. Ten patients received treatment before
were followed without treatment for a mean of 3.7 years the age of ten years. By the age of ten, sixteen of twenty
(range, six months to eleven years and seven months). Be one curves were more than 40 degrees , and twelve of these
fore the patients reached the age of ten years , many of the were greater than 60 degrees (Figs. 4-A and 4-B). Ten pa
curves deteriorated rapidly (at a median rate of 5 degrees tients were seen, untreated, at the age of ten years, at

FIG. 2-A FIG. 2-B

A girl, thirteen years old, who had a unilateral unsegmented bar on the right, extending from the second to the fifth thoracic vertebra and producing
a right upper thoracic scoliosis measuring 60 degrees. She had a major cosmetic deformity due to elevation of the left shoulder line and slight tilting of
the head to the right.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1135

ifi
TO A UNILATERAL UNSEGMENTED BAR

Years)Size GroupiC (PatientsSeenat or after the Ageof Ten

of Curve (Degrees)*Age
(Degrees)No.tWhenFirst Seen at Last Seen When of Deterioration per Year
891020(7)32(i7-81)47(20-98)15(11-19)525
or after 10 Yrs. OldWhen or at Maturity234 Last Seen* 5 6 7
(Yrs.)Rate

3114 2 2
166
(6)43 (25-73)65 (33-96)14 (11-18)141 1
23i(0)2037141
(5)58 (35-80)78 (45-93)12 (ii-17)1

which time the mean curve was 48 degrees (range, 25 to 68 nosed at birth which, without treatment, became 77 de
degrees). Three of these curves were fused when the pa grees at the age of four years. The other two patients both
tients were ten years old. were first seen at the age of fourteen years with curves of
After the patients were ten years old, the median rate 57 and 50 degrees, which deteriorated rapidly to 70 and 96
of deterioration increased to 6.5 degrees per year. Ten pa degrees in the patients' fifteenth and sixteenth years , re
tients received treatment at a mean age of thirteen years spectively.
(range, eleven to sixteen years), at which time the mean
curve was 65 degrees (range , 40 to 96 degrees) . Three pa Thoracolumbar Curves
tients reached skeletal maturity without treatment; at that Of the fifteen patients who had a thoracolumbar
time the curves were 47 , 60, and 66 degrees. curve, three were in Group lA. Two of them were treated
The most extensive unsegmented bars in this region immediately, at two and thirteen years old, with curves
occurred in three patients and involved five to six verte measuring 35 and 85 degrees, respectively. One was seen
brae. In one of these patients a 45-degree curve was diag at skeletal maturity, with a curve of 8 1 degrees.

FIG. 3-A FIG. 3-B

Fig. 3-A: A girl, one month old, who had a radiograph made because of asymmetry of the thoracic rib cage. The radiograph shows absent ribs and a
unilateral unsegmented bar extending from the third to the seventh thoracic vertebra on the left.
Fig. 3-B: No treatment was given, and by the age of ten years the unilateral unsegmented bar had produced a 62-degree right upper thoracic
scoliosis. The major deformity, however, was due to a very severe secondary left structural thoracolumbar scoliosis that involved no congenital
anomalies.

VOL. 64-A, NO. 8, OCTOBER 1982

1136 M . J . MCMASTER AND KUNIYOSHI OHTSUKA

The remaining twelve patients with a thoracolumbar vere disability was then due to an apparent shortening of
curve were in Groups lB and lC (Table III); they were all one lower limb.
seen in the first decade of life and were followed without
Unilateral Failure of Segmentation (Unilateral
treatment for a mean of five years (range, six months to
Unsegmented Bar with Contralateral Hemivertebrae)
sixteen years) . Before they reached the age of ten years,
(Group 2)
the median rate of deterioration of the curves was 6 de
grees per year. Six patients were treated between the ages Twenty-eight patients had a single congenital
of two and seven years, at which time the mean curve was scoliosis due to a unilateral unsegmented bar on the con
64 degrees (range, 40 to 82 degrees). The remaining six cavity of the curve with one or more hemivertebrae on the
curves had a mean measurement of 59 degrees (range, 35 convexity at the same level . Twelve curves were convex to
to 80 degrees) at the age of ten years. the right and sixteen, to the left. These curves were
After the patients were ten years old, the median rate classified separately from those that were due to a unilat
of deterioration increased to 9 degrees per year. Five of six eral unsegmented bar alone because they had an even more
patients received treatment at a mean age of 11.8 years severe prognosis. By the age of five years, all of these
(range, eleven to twelve years), at which time the mean curves became very severe. The rate of deterioration de
curve was 78 degrees (range, 45 to 90 degrees). pended mainly on the site of the anomalies and partly on
The most extensive bars in this region occurred in the extent of the unsegmented bar. The bar involved from
three patients and involved six to eight vertebrae. In one two to eight vertebrae. The number of hemivertebrae
patient a 15-degree curve was diagnosed in the first year of ranged from one to seven. This type of anomaly occurred
life, and at skeletal maturity the untreated curve measured most frequently in the thoracic region (eighteen patients).
93 degrees . The second patient was first seen and treated at The hemivertebrae were most often opposite the bar and
thirteen years old with an 85-degree curve, and the third not separated by a number of normal vertebrae, as was
was first seen at skeletal maturity with an 80-degree curve. usually the case when two unilateral hemivertebrae oc
Pelvic obliquity or listing of the trunk to the side of curred in the absence of a bar. The anomaly was most eas
the unsegmented bar, or both, occurred in twelve of the fif ily recognized radiographically in the first few years of life
teen patients with a thoracolumbar curve. The degree of (Fig. 5-A), but as the curve rapidly increased the hemiver
pelvic obliquity was severe (more than 25 degrees) in four tebrae became obscured (Fig. 5-B) and were difficult to
(range, 26 to 30 degrees), moderate (10 to 25 degrees) in distinguish from a unilateral unsegmented bar alone. The
four, mild (less than 10 degrees) in two, and absent in five. three patients for whom serial observations were not made
The listing was severe in one, moderate in two, mild in were in Group 2A and were not included in this part of the
five, and absent in seven. analysis. The twenty-five patients who were followed
without treatment were in Group 2B (Table IV).
Lumbar Curves
Of the five patients with a lumbar curve, two were Upper Thoracic Curves
treated immediately and were in Group lA. These two pa The seven patients with an upper thoracic curve were
tients had bars involving four and eight vertebrae, produc in Group 2B (Table IV) and were followed without treat
ing curves of 64 and 100 degrees at the ages of sixteen and ment for a mean of 4.6 years (range, six months to twelve
nine years, respectively. years). Before they reached the age of ten years, the me
The remaining three patients with a lumbar curve dian rate of deterioration of the curves was 5 degrees per
were in Groups lB and 1C (Table III) and were followed year (range, 3 to 8 degrees). Four patients were treated be
without treatment for a mean of four years (range, one to tween the ages of three and five years . One of these pa
eight years). One patient had a bar involving three verte tients, with an unsegmented bar involving eight vertebrae
brae which produced a 62-degree curve at the age of one with seven contralateral hemivertebrae, had an 82-degree
year; the curve had deteriorated to 82 degrees in the second curve at birth which deteriorated at a rate of 8 degrees per
year of life, when it was treated. The second patient also year to become 108 degrees at the age of four years . The
had a bar involving three vertebrae but this produced only other three patients had unsegmented bars whose extent
an 8-degree curve in the first year of life; the curve de ranged from three to five vertebrae with one or two con
teriorated to 53 degrees at the age of ten years . The third tralateral hemivertebrae, and the curves measured 35, 52,
patient had a bar that involved two vertebrae and the curve and 54 degrees prior to treatment. One patient with a bar
deteriorated slowly . The curve measured 20 degrees when involving three vertebrae and one contralateral hemiver
the patient was eleven years old and then deteriorated more tebra had a 27-degree curve at the age of three months, and
rapidly, to reach 37 degrees when the patient was fourteen was untreated at the age of two years when the curve had
years old. increased to 35 degrees.
Pelvic obliquity and listing of the trunk occurred in all Two patients reached the age of ten years without
five of the patients in this group and was directly propor treatment. One, with a bar involving six vertebrae with
tional to the severity of the curve. In the larger curves the two contralateral hemivertebrae, had a 45-degree curve at
pelvic obliquity was the major deforming factor. The se one year old, which deteriorated at a rate of 5 degrees per

ThE JOURNALOF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1137
TABLE IV
UNTREATED SINGLE CONGENITAL SCOLIOSISDUE TO A UNILATERAL
UNSEGMENTED BAR WITH CONTRALATERAL HEMIVERTEBRAE

Years)No.Age 2B (Patients Seenbefore theAge of Ten

of Curve (Degrees)*Rate
of when (Degrees)WhenWhen Last Seen of Deterioration per Year
10>10Upper
Site of CurveNo. CurvesGroup First Seen* First Seen before 10 Yrs. Old123 4 5 6 7 8 9
(Yrs.)Size

thoracic (0-5) (12-82) 54 (35-108) 1 3 2

Lower thoracic 10 10 1 (0-6) 58 (23-76) 84 (27-108) 1 1 5 1
Thoracolumbar7 87 81 1 (0-2)45 58 (41-86) 79 (54-148)1 1 1 12 5

* Median, with range in parentheses.

year to become 83 degrees by the age of ten years. The ated at a rate of more than 5 degrees per year and measured
second patient, with a bar involving three vertebrae with between 59 and 103 degrees in the three patients who were
one contralateral hemivertebra, had a 40-degree curve at treated between the ages of three and five years . In one pa
the age of five years old, which deteriorated at a rate of 4 tient, who was treated at the age of thirteen years, the
degrees per year to become 59 degrees at the age of ten curve measured 72 degrees.
years. After these patients reached ten years old, the rate
of deterioration increased to S and 7 degrees per year, and Lower Thoracic Curves
both patients were treated with spine fusion at the ages of Of the eleven patients who had a lower thoracic
eleven and thirteen years old when the curves were 88 and curve, one was in Group 2A. This patient, with a 60-
82 degrees, respectively. degree curve, was treated immediately at the age of ten
Elevation of the shoulder line or tilting of the head, or years.
both, occurred in all of these patients and, as with other The remaining ten patients with a lower thoracic
types of upper thoracic congenital scoliosis, caused a sig curve were in Group 2B (Table IV), and were followed
nificant cosmetic deformity. In addition, in four patients in without treatment for a mean of 5.5 years (range, one year
whom the apex of the curve was at the fourth, fifth, or to ten years and six months). Because of the severity of the
sixth thoracic vertebra a secondary structural lower deformity, in eight patients the scoliosis was diagnosed in
thoracic or thoracolumbar scoliosis developed, but did not the first two years of life when the mean curve was 50 de
result from congenital anomalies. These curves deterior grees (range, 23 to 76 degrees). After the initial diagnosis,

FIG. 4-A FIG. 4-B

Fig. 4-A: A girl, five years and one month old, with a unilateral unsegmented bar on the right, extending from the eighth to the tenth thoracic
vertebra and producing a left lower thoracic scoiiosis measuring 61 degrees.
Fig. 4-B: No treatment was given, and by the age of eleven years and two months the lower thoracic curve had deteriorated to 91 degrees.

VOL. 64-A, NO. 8, OCTOBER 1982

1138 M. J. MCMASTER AND KUNIYOSHI OHTSUKA

the median rate of deterioration, without treatment, was 6 segmented bar with contralateral hemivertebrae.
degrees per year (range , 4 to 11 degrees) and all but one of
the patients received treatment before the age of ten years. Bilateral Failure of Segmentation
Three patients were treated at the age of nine years , with (Block Vertebrae) (Group 3)
curves of 72, 99, and 106 degrees. Block vertebrae were present in thirteen patients.
The most extensive vertebral anomalies in this region Thus , they were not considered to be a common cause of
occurred in four patients who had unsegmented bars in congenital scoliosis . In our series this type of anomaly
volving from seven to nine vertebrae with three to six con most frequently occurred in the upper thoracic region (sev
tralateral hemivertebrae . When the patients were three en patients) and was least common in the lumbar region
years old, all of these curves measured more than 70 de (one patient). All of these patients had a single congenital
grees (range, 76 to 102 degrees). One patient was followed scoliosis (eight left and five right) due to a single block of
without treatment and at the age of ten years the curve bilaterally unsegmented vertebrae which ranged from two
measured 108 degrees . The rate of deterioration then in to five segments in length. Four patients had a block of two
creased, and when the patient was thirteen years old the vertebrae; six , of three vertebrae; one , of four vertebrae;
curve measured 130 degrees. and two had a block of five vertebrae. No patient required
treatment and the mean follow-up was 7 .8 years (range,
Thoracolumbar Curves two years to sixteen years and six months) , with three pa
Of the ten patients who had a thoracolumbar curve, tients having reached skeletal maturity. There were nine
two were in Group 2A . Both of these patients required upper thoracic or lower thoracic curves . All of the curves
treatment in the first year of life, with curves of 68 and 75 behaved in the same manner and remained 21 degrees or
degrees. less (range , 10 to 2 1 degrees) . The rate of deterioration
The remaining eight patients with a thoracolumbar was 1 degree or less per year. Of the three curves that were
curve were in Group 2B (Table IV) and were followed followed to skeletal maturity , none exceeded 2 1 degrees.
without treatment for a mean of 3 .8 years (range, six There were three thoracolumbar curves that behaved in a
months to ten years and eight months). These curves were slightly different manner. Two of these curves had
found to have the worst prognosis of any type of congeni virtually no deterioration and measured 10 and 14 degrees
tal scoliosis occurring at any site. All of the scolioses were when the patients were nine and fourteen years old, re
diagnosed in the first two years of life , when the median spectively. The remaining curve, which was first seen
curve was 64 degrees (range , 4 1 to 86 degrees). The when the patient was ten years old, deteriorated a little and
curves deteriorated at a median rate of more than 10 de reached 35 degrees at skeletal maturity. The one lumbar
grees per year. All but three of the patients received treat curve deteriorated very slowly and measured 24 degrees
ment before the age of five years , when the median curve when the patient was ten years old.
was 70 degrees (range, 64 to 125 degrees). Two of the
three patients who were followed, untreated, for five years Unilateral Complete Failure of Formation
received treatment at the age of seven years , with curves of (Hemivertebrae) (Group 4)
90 and 148 degrees. The third patient was untreated at ten Hemivertebrae were the cause of a congenital
years old, with a 98-degree curve, which then deteriorated scoliosis in seventy-seven patients . Two of these patients
at a rate of 10 degrees per year to become 124 degrees at were seen at skeletal maturity and were in Group 4A . The
the age of twelve years and eight months (Figs. 5-A, 5-B, remaining seventy-five patients were followed without
and 5-C). treatment and were in Groups 4B and 4C. The hemiverte
The most extensive verte&ral anomalies in the brae were always on the convexity of the curve. Seventy
thoracolumbar region occurred in three patients who had patients had a single congenital scoliosis (thirty-six left
unsegmented bars involving seven to nine vertebrae, with and thirty-four right), of which fifty-six were due to a
three or four contralateral hemivertebrae . At the age of two single hemivertebra (Table V) and fourteen were due to
years , all of these curves were more than 80 degrees two hemivertebrae on the same side, producing a single
(range, 82 to 103 degrees). One patient was not treated congenital scoliosis (Table VI). Seven patients had
until the age of seven years, when the curve measured 148 hemivertebrae on opposite sides but at different levels of
degrees. the spine, producing two or more opposing congenital
Pelvic obliquity or listing of the trunk toward the side curves (Table II, Cases 4 through 10).
of the unsegmented bar, or both, occurred in nine of the Two patients had a hemivertebra that was initially
ten patients with a thoracolumbar curve. The degree of separate but later became synostosed with one of its
pelvic obliquity was severe (greater than 25 degrees) in neighboring vertebrae. Both were in the lumbar region and
two (32 and 50 degrees), moderate (10 to 25 degrees) in the resulting curves progressed less rapidly (less than 1
four, mild (less than 10 degrees) in two, and absent in two. degree per year) than the majority of the curves in this re
The listing was severe in four, moderate in one, mild in gion. Eight patients had an incarcerated hemivertebra; that
two, and absent in three. is, a small, poorly formed segment of bone tucked into a
There were no lumbar curves due to a unilateral un niche between adjacent normal vertebrae . Four of these in

ThE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURALHISTORYOF CONGENITAL
SCOLIOSIS 1139

2@1I
@ .v

FIG. 5-A FIG. 5-B FIG. 5-C

Fig. 5-A: A girl, two years and ten months old, who had a 56-degree left thoracolumbar scoliosis due to a unilateral unsegmented bar on the right
extending from the eighth to the twelfth thoracic vertebra with four adjacent contraiateral hemivertebrae at the ninth to the twelfth thoracic levels.
Figs. 5-B and 5-C: No treatment was given, and by the age of twelve years and eight months the scoliosis had increased to i24 degrees and the
contralateral hemivertebrae were no longer visible. There was pelvic obliquity, and a severe list of the upper part of the body to the right produced a
severe cosmetic deformity.

carcerated hemivertebrae occurred in the upper thoracic a single hemivertebra was 1 degree per year (range , zero to
region; one, in the lower thoracic region; one, in the 2 degrees), compared with 2 degrees per year (range, 1 to
thoracolumbar region; and two, in the lumbar region. The 2 degrees) in the curves that were due to two unilateral
resulting curves were all less than 20 degrees , with mini hemivertebrae. By the age of ten years, no curve that was
mum or no progression. due to a single hemivertebra was greater than 40 degrees,
There was a total of 102 hemivertebrae, which were whereas all three patients with two unilateral hemiverte
evenly distributed on either side of the spine (fifty-three on brae had a curve of 40 degrees or more (range , 40 to 47
the left and forty-nine on the right) and occurred at any degrees).
level, although they were seen slightly more frequently in After the patients reached ten years old, the median
the middle thoracic and lower lumbar regions . Of the pa rate of deterioration doubled (2 degrees per year) in the
tients with two unilateral hemivertebrae producing a single curves that were due to a single hemivertebra and in
congenital curve, the hemivertebrae were separated by creased to 2.5 degrees per year (range, 2 to 3 degrees) in
four normal vertebrae in two patients , by three normal ver those due to two unilateral hemivertebrae. Of six patients
tebrae in three patients , by two normal vertebrae in five with a single hemivertebra, three were followed without
patients, and by one normal vertebra in three patients; in treatment to skeletal maturity, at which time the curves
only one patient were the hemivertebrae adjacent. measured 25, 39, and 43 degrees, and the remaining three
patients had a spine fusion when the curves measured 28,
Upper Thoracic Curves 3 1, and 44 degrees . Two patients with two unilateral
Sixteen patients had a single congenital upper hemivertebrae were followed after the age of ten years and
thoracic curve, of which thirteen were due to a single both required a spine fusion during the adolescent growth
hemivertebra (Table V) and three, to two unilateral spurt , when the curves were 43 and 47 degrees.
hemivertebrae (Table VI). Follow-up without treatment In general , upper thoracic curves due to one or two
was for a mean of four years (range , one year to nine years unilateral hemivertebrae progressed relatively slowly , and
and ten months). The median rate of deterioration, without only a few became moderately severe. Although none be
treatment, before the age of ten years for the curves due to came very severe, they did, like other congenital upper

VOL. 64-A, NO. 8, OCTOBER 1982

1140 M. J. McMASTER AND KUNIYOSHI OHTSUKA

TABLE
SCOLIOSISSite UNTREATED SINGLE CONGENITAL

Years)No.Age 4B (Patients Seenbefore theAge of Ten

of Curve (Degrees)*Rate
of When (Degrees)WhenWhen Last Seen of Deterioration per Year
910Upper
of CurveNo. CurvesGroup First Seen* First Seen before 10 Yrs. Old< 11 2 3 4 5 6 7 8
(Yrs.)Size

2Lower thoracic13102 (0-7)16 ( 11-28) 18 ( I 1-36)53

17-50)24ThoracolumbarS31
thoracic964 (0-9)34 ( 17-48) 44 (
2Lumbar15122(0-9)24(0-37) (0-3)26 (20-35) 40 (21-45)1
2Lumbosacral1284 30(13-41)73
(2-9)18 (10-32) 19 (10-53)61 1

* Median, with range in parentheses.

t The number of patients who were followed untreated until after the age of ten years but who were first seen before the age of ten years , and who are
also included in Group 4B, is in parentheses.

thoracic curves, produce a significant cosmetic deformity were followed without treatment (Figs. 6-A and 6-B), five
due to elevation of the shoulder line on the convex side of required treatment between the ages of twelve and fifteen
the curve, and occasionally they produced tilting of the years, at which time the median curve was 64 degrees
head. In only one patient, with a single hemivertebra at the (range , 30 to 70 degrees) , and one reached skeletal matur
fifth thoracic level, did a secondary structural lower ity with a 50-degree curve.
thoracolumbar scoliosis develop. At skeletal maturity, this
curve measured 52 degrees and the upper thoracic congeni Thoracolumbar Curves
tal curve measured 43 degrees. Of the ten patients with a thoracolumbar curve, two
were in Group 4A. Both of these patients had a single
Lower Thoracic Curves hemivertebra and were seen at skeletal maturity with
Seventeen patients had a single congenital lower curves of 52 and 55 degrees. The remaining eight patients
thoracic curve, nine of which were due to a single with a thoracolumbar curve were followed without treat
hemivertebra (Table V) and eight, to two unilateral ment for a mean of five years (range, one year and seven
hemivertebrae (Table VI). These patients were followed months to seven years and seven months).
without treatment for a mean of 3.6 years (range, one year Five patients had a curve that was due to a single
to nine years and four months), and four reached skeletal hemivertebra (Table V) , and the mean rate of deterioration
maturity. The median rate of deterioration before the pa before the age of ten years was 2 degrees per year (range, 1
tients reached the age of ten years was 2 degrees per year to 2 degrees) and 3 .5 degrees thereafter (range , 3 to 4 de
(range, zero to 2 degrees) for the curves that were due to a grees). Before the patients were ten years old, one of three
single hemivertebra and 2 degrees per year (range, 2 to 6 curves was greater than 40 degrees and the two curves that
degrees) for the curves that were due to two unilateral were untreated at skeletal maturity measured 33 and 43
hemivertebrae. By the age of ten years, the majority of the degrees.
patients (four of six) with a single hemivertebra had a Three patients had a single congenital thoracolumbar
curve of between 40 and 50 degrees , whereas all of the six curve due to two unilateral hemivertebrae (Table VI).
patients with two unilateral hemivertebrae had a curve of These curves were all diagnosed in the first two years of
50 degrees or more. life , when they exceeded 50 degrees . The median rate of
After the patients reached the age of ten years, the deterioration before the age of ten years was S degrees per
median rate of deterioration for the lower thoracic curves year (range, 4 to 6 degrees) and all three patients required
due to a single hemivertebra increased to 2.5 degrees per treatment in the third or fourth year of life, when the me
year (range, less than 1 degree to 5 degrees), and in the dian curve was 63 degrees.
curves due to two unilateral hemivertebrae the median rate
was 3 degrees per year (range, 2 to 4 degrees). Of the six Lumbar Curves
patients with a single hemivertebra who were followed Fifteen patients had a single congenital lumbar curve,
after the age of ten years, three reached skeletal maturity all of which were due to a single hemivertebra at the third
without treatment, at which time the curves measured 37, or fourth lumbar level (Table V). None of these patients
38 , and 43 degrees; two other curves progressed more required treatment, and they were followed for a mean of
rapidly and required spine fusion, measuring 52 degrees 6.6 years (range, one year to fourteen years and four
when one patient was twelve years old and 60 degrees months) . The median rate of deterioration was less than 1
when the other was fourteen years old; and one patient had degree per year (range, zero to 2 degrees) before the age of
a 47-degree untreated curve at the age of eleven years. Of ten years and 1 degree per year thereafter (range, zero to 3
the six patients with two unilateral hemivertebrae who degrees) . Before the age of ten years , only one curve was

ThE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1141
V
DUE TOA SINGLE HEMI VERTEBRAGroup

Years)Size 4C(Patients Seen ator afterthe Age of Ten

of Curve (Degrees)*Age
(Degrees)No.tWhenFirst Seen at When Last Seen When of Deterioration per Year
or after 10 Yrs. Old
89106(3)30(15-40) or at Maturity< Last Seen* 112 3 4 5 6 7
(Yrs.)Rate

26(3)44(25-50) 35(25-44)16(11-17)4
12 45(37-60)16(11-19)111 2
18(5)32(14-52)
(0)30 (21-40) 41 (33-43)Both matureI
112(8)20(15-53) 36(16-55)17(13-18)232
25(15-83)15(11-18)512 3 1

greater than 40 degrees and six patients had curves that had (range, one year to ten years and seven months). In nine
remained virtually unchanged, at less than 30 degrees, patients the hemivertebra was at the fifth lumbar level and
from an early age. Five patients were followed, without in three patients the hemivertebra lay between the fifth
treatment, to skeletal maturity, when the median curve lumbar level and the sacrum; in one of these patients there
was 42 degrees (range, 22 to 55 degrees). was a failure of segmentation between the hemivertebra
In all of the patients with a lumbar curve the trunk and the sacrum. The congenital lumbosacral curves were
remained balanced, but in two there was mild pelvic all very short and extended from the fourth or fifth lumbar
obliquity. level to the sacrum. In all of the patients the pelvis re
mained level, and as a result the hemivertebra caused the
Lumbosacral Curves
lumbar spine to take off obliquely from the sacrum. To
This type of congenital scoliosis occurred only in as overcome this imbalance, in all of the patients a long sec
sociation with a single hemivertebra at the lumbosacral ondary thoracolumbar curve developed, extending from
junction (Table V). There were twelve such patients, who the fifth lumbar vertebra to the ninth, tenth, or eleventh
were followed without treatment for a mean of 6.8 years thoracic vertebra and soon becoming fixed. In eleven pa

FIG. 6-A FIG. 6-B

Fig. 6-A: A boy, three years and six months old, with a 40-degree right lower thoracic scoliosis due to two unilateral hemivertebrae at the seventh
and tenth thoracic levels.
Fig. 6-B: No treatment was given and the curve deteriorated to 70 degrees at the age of fourteen years, just prior to spine fusion.

VOL. 64-A, NO. 8. OCTOBER 1982

1 142 M. J. McMASTER AND KUNIYOSHI OHTSUKA

TABLE
UNTREATED SINGLE CONGENITAL ScoLlosls

Years)No.Age 4B (Patients Seen before the Age of Ten

of Curve (Degrees)*Rate
of When (Degrees)WhenWhen Last Seen of Deterioration per Year
Site of CurveNo. CurvesGroup First Seen*Size First Seen
910( before 10 Yrs. Old123 4 5 6 7 8
Yrs.)Upper

(40-47)12Lower
thoracic336 (3-9)42 (32-44) 44
(50-54)31Thoracolumbar331
thoracic842 (0-4)40 (30-48) 53
(0-2)55 (50-60) 63 (55-72)1 1 1

* Median, with range in parentheses.

t The number of patients who were followed untreated until after the age of ten years but who were first seen before the age of ten years, and who are
also included in Group 4B, is in parentheses.

tients this curve did not compensate sufficiently and as a exceeded 25 degrees. After the patients were more than ten
result the upper part of the body listed to a varying degree years old, the rates of deterioration of both curves in
to the side opposite that of the hemivertebra. In the one creased slightly, as did the tendency of the upper part of
patient in whom the body remained balanced, it did so be the body to list to one side. Five of these patients were fol
cause of a limb-length discrepancy that resulted in a corn lowed to skeletal maturity, without treatment, at which
pensatory pelvic obliquity. With time, all of these sec time neither the congenital lumbosacral curve nor the sec
ondary curves became structural, and three deteriorated to ondary thoracolumbar curve exceeded 36 degrees and the
become the major deformity. degree of listing was not sufficient to require treatment.
Not all of the secondary curves deteriorated at the Three patients had a hemivertebra that did not extend
same rate, and this appeared to depend on the extent of the across the midline, and this produced a greater list and a
hernivertebra. Nine patients had a hemivertebra that ex much larger secondary structural thoracolumbar curve,
tended across the midline, and in these patients both the which was severely rotated and produced a large rib hump.
congenital lumbosacral curve and its secondary thoraco Before the patients were ten years old, both the lumbosa
lumbar curve remained virtually static until the patients cral curve and the secondary thoracolumbar curve de
reached the age of ten years, at which time neither curve teriorated relatively slowly (at a rate of 1 to 3 degrees per

?@‘@

131@yrs

3@2yrs

42°

1000

32°

83°

FIG. 7-A FIG. 7-B FIG. 7-C

Fig. 7-A: A girl, three years and five months old, with a 32-degree left lumbosacral scoliosis due to a single hemivertebra at the lumbosacral
junction. There was also a 42-degree right thoracolumbar scoliosis in which no congenital anomalies were involved. The pelvis was level but the
upper part of the body listed to the right.
Figs. 7-B and 7-C: No treatment was given, and by the age of thirteen years and five months the congenital lumbosacral curve had deteriorated to
83 degrees. The thoracolumbar curve measured 100 degrees and had become structural and severely rotated. The list had increased but the pelvis
remained level. The major deformity was due to the severe thoracolumbar scoliosis which, despite its size, failed to compensate for the list to the
right.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1143
VI
DUE TO Two UNILATERAL HEMIVERTEBRAE

Years)Size Group 4C (Patients Seen at orafter the Age of Ten

of Curve(Degrees)*Age
(Degrees)NotWhen First Seen at Last Seen When of Deterioration per Year
89102 orafterlOYrs.OldWhen oratMaturity234 LastSeen* 5 6 7
(Yrs.)Rate
1-16)116
(2)42 (40-44)45 (43-47)12 (1
(12-17)2220(0)———
(2)51 (28-54)62 (30-70)13

year), but after the age of ten years a significant cosmetic treatment at the age of five years , when the curve mea
deformity developed due to the rib hump and an increasing sured 32 degrees. The remaining two patients were un
tendency of the upper part of the body to list to one side. treated, and the curves deteriorated at a rate of 2 degrees
The congenital lumbosacral curves in two of these patients per year until they measured 17 degrees, at which time one
measured 25 and 42 degrees at the age of fourteen years, child was thirteen years old. The other child had a 40-
and the secondary structural thoracolumbar curves mea degree curve at skeletal maturity.
sured 40 and 48 degrees , respectively . The most severe de
formity occurred in the third patient who, at the age of thir Lumbar Curves
teen years, had an 83-degree lumbosacral curve with a Only one lumbar curve was due to wedging, and it in
100-degree secondary structural thoracolumbar curve and volved the fifth lumbar vertebra. This curve deteriorated
a severe list (Figs. 7-A, 7-B, and 7-C). slowly, at a rate of less than 1 degree per year over three
years , and measured 12 degrees when the patient was
Unilateral Partial Failure of Formation
seven years old.
( Wedge Vertebrae) (Group 5)
Wedge vertebrae occurred in only nine patients. They Complex (Unclassifiable) Anomalies (Group 6)
all had a single congenital scoliosis (five left and four There were twenty-five patients with complex (un
right) . The most commonly affected vertebrae were in the classifiable) congenital vertebral anomalies causing a con
thoracic region (six patients). Seven patients had a single genital scoliosis. Twenty-two of these patients had a single
wedge vertebra and two patients had a pair of adjacent congenital curve (twelve right and ten left) and three had
thoracic vertebrae that were affected on the same side. multiple congenital curves.
Nineteen patients had a single congenital scoliosis
Upper Thoracic Curves
due to a jumble of vertebral anomalies, and they were fol
There was only one upper thoracic congenital lowed without treatment for a mean of seven years (range,
scoliosis that was due to wedging, and it involved the one to seventeen years) . Two patients had an upper
fourth thoracic vertebra. This curve deteriorated at a rate thoracic curve; eight patients, a lower thoracic curve;
of slightly less than 2 degrees per year, from 23 degrees seven patients , a thoracolumbar curve; and two patients , a
when the patient was twelve years old to 3 1 degrees at lumbar curve. All of these curves were much more unpre
skeletal maturity. dictable in their behavior than the congenital curves that
were due to simple anomalies in the same regions , but in
Lower Thoracic Curves
general they tended to progress relatively slowly regard
There were four lower thoracic congenital curves , of less of their location. Fifteen patients were untreated at the
which two were due to a single wedge vertebra and two, to age of ten years, at which time eight curves were less than
a pair of adjacent wedge vertebrae. Three patients were 20 degrees; three curves, between 20 and 30 degrees; five
followed, without treatment, for a mean of 6.4 years curves, between 3 1 and 40 degrees; and only two curves
(range, one year and eight months to nine years and nine measured more than 40 degrees (5 1 and 73 degrees). After
months) and required treatment at nine, eleven, and twelve the age of ten years, the rate of deterioration usually in
years old, when the curves measured 33, 39, and 41 de creased. Seven patients reached skeletal maturity without
grees , respectively . One patient with two wedge vertebrae treatment, at which time the curves ranged from 8 to 56
was first seen at skeletal maturity, when the curve mea degrees (mean, 3 1 degrees). The largest untreated curve in
sured 49 degrees. the remaining five skeletally immature patients was in the
thoracolumbar region and measured 82 degrees at the age
Thoracolumbar Curves of twelve years.
There were three thoracolumbar congenital curves, Three patients who were seen at skeletal maturity had
which were followed for a mean of 3 .4 years (range , one a single congenital scoliosis , but the congenital anomaly
year to seven years and four months) . One patient required was partially obscured and could not be accurately

VOL. 64-A, NO. 8, OCTOBER 1982

1144 M . J . McMASTER AND KUNIYOSHI OHTSUKA

AnomalyBlock of Congenital

Unilateral Un
segmented Bar
Site of Unsegmented and Contralateral
CurvatureType VertebraWedged
VertebraHemivertebraUnilateral
HemivertebraeSingleDoubleUpper Bar.

—¿6°Lower
thoracic° —¿ 1 °*I °—¿
2@2° —¿
2.5°2° —¿
4°5°

thoracic@1 °—¿
—¿7°Thoraco 1°@2°
—¿
2°2°—¿
2.5°2°—¿
3°5° —¿
6.5°6°

lumbar<1 °—¿ 1°1 .5°

*Lumbarzl°—*<1@@*.ci°—i°I*>5°—* —¿
2°2°—¿
3.5°5° —¿
*6° —¿
9°> 10° —¿

,*Lumbo

sacral**@1° 1.5°***

required E:JMayrequire
t@JNotreatment spinalfusion 0 Require
spinalfusion
* Too few or no curves
FIG. 8
Median yearly rate of deterioration (in degrees) without treatment for each type of single congenital scoliosis in each region of the spine (i87
patients). The numbers on the left in each column refer to patients who were seen before the age of ten years; the numbers on the right refer to patients
who were seen at or after the age of ten years.

classified. All three patients had a thoracic or thoracolum mal growth of the spine with the pathological anatomy of
bar curve of more than 100 degrees. the various types of congenital vertebral anomalies . Nor
mally, longitudinal growth of the spine is the sum total of
Discussion the growth occurring at the end-plates on the upper and
The prognosis for a patient with congenital scoliosis lower surfaces of the vertebral bodies , which occurs
can vary considerably. Some patients are first seen with equally on either side of the spine so that the spine remains
small curves, many of which progress minimally, whereas straight and without a scoliosis2 . A congenital vertebral
others are first seen with larger curves that deteriorate anomaly can, however, cause a growth imbalance due to a
rapidly and cause extreme deformity . Of the 25 1 patients deficiency in either the number of end-plates or their rate
in this study, 143 were last seen, untreated, after the age of of growth on one side of the spine. The lateral curve that
ten years, at which time fifty-one (36 per cent) had a curve results is of a severity proportional to the degree of the
of 40 to 60 degrees and forty (28 per cent) had a curve of growth imbalance.
more than 60 degrees. We concluded that the prognosis for In classifying a congenital scoliosis with regard to its
untreated curves after the age of ten years generally be prognosis , the most important feature is , therefore, the
comes more unfavorable because in our series most of the growth imbalance caused by the vertebral anomalies that
other 108 patients (that is, an additional sixty-seven pa predominate on one side of the spine. The radiographic
tients) required treatment at or before the age of ten years classification that was used in this study was found to be
because of the severity of the curve. Twenty of these generally satisfactory in this respect, although not all of
curves measured 40 to 60 degrees and thirty-nine, more the curves progressed as expected and the growth potential
than 60 degrees . These findings are very different from could not be exactly predicted radiographically. The ver
those of Kuhns and Hormell, who found that only 38 per tebral anomalies in 90 per cent of the patients could be
cent of eighty-five children who were followed to skeletal classified into one of five specific groups , but some curves
maturity without treatment had curves of more than 30 could not be classified (Group 6) because they either in
degrees . Our findings are more in agreement with those of volved a complex jumble of anomalies or, on occasion,
Winter et al. , who found that 84 per cent of thirty-eight because the severity was so great as to obscure the radio
children who were followed without treatment beyond the graphic characteristics of the anomaly.
age of ten years had curves of more than 40 degrees. Congenital scoliosis occurred significantly more often
To understand the variable prognosis for congenital in girls than in boys in our series . The congenital curves
scoliosis , it is necessary to correlate the principles of nor often were present at birth and were most frequently diag

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1145

nosed in either the first few years of life or between the side of the spine. In our series, the longer unsegmented
ages of nine and fourteen years , probably because the bars tended to produce the larger curves in a specific re
periods of most rapid growth of the spine &e in utèrO, gion, but occasionally a shorter bar produced an equally
from birth to the age of three years, and at puberty7 . The large curve in the same region. In three children the un
scolioses that presented as a clinical deformity in the first segmented bar was not recognized radiographically until
year of life had the worst prognosis, as this indicated a they were three and four years old, when it became more
marked growth imbalance that would continue until ossified. As a result, these patients were initially misdiag
skeletal maturity, resulting in severe deformity. nosed as having infantile idiopathic scoliosis. Because of
A single congenital scoliotic curve occurred in 95 per the severe growth imbalance, all of the curves that were
cent of the 25 1 patients and was much more common than associated with an unsegméntedbar and that were present
the occurrence of two or more congenital curves (5 per during infancy deteriorated very rapidly, and all became
cent of the patients). The commonest type of vertebral very severe. The mean rate of deterioration in patients who
anomaly causing the scoliosis was a unilateral unseg were younger than ten years old ranged from 5 degrees per
mented bar (38 per cent of the 269 congenital curves; year for lower thoracic curves to 6 degrees per year for
Table I), followed by hemivertebrae (33 per cent), corn thoracolumbar curves (Table Ill). Of the thirty-six patients
plex anomalies ( 11 per cent) , a unilateral unsegmented bar with a lower thoracic, thoracolumbar, or lumbar curve
with contralateral hemivertebrae (10 per cent), and a block who were last seen without treatment at or before the age
vertebra (5 per cent); the least common was a wedge ver of ten years , ten had a curve that measured between 40 and
tebra (3 per cent) . The commonest site for the congenital 60 degrees and nineteen had a curve of more than 60 de
scoliosis was the lower thoracic region (33 per cent of the grees . This type of congenital scoliosis requires treatment
269 congenital curves; Table I), followed by the upper as soon as the anomaly is diagnosed.
thoracic (3 1 per cent), thoracolumbar (20 per cent), and It is important to recognize the anomaly of a unilat
lumbar regions ( 11 per cent); the least common was the eral unsegmented bar with contralateral hemivertebrae, as
lumbosacral region (5 per cent). described by Nasca et al. , because it has the worst prog
Like Winter et al . , we found that the rate of deterio nosis of any type of congenital vertebral anomaly. All of
ration and the ultimate severity of the congenital scoliosis these patients were seen before the age of two years in our
depended not only on the type of anomaly but also on the series. At that age the hemivertebrae could be seen radio
site at which it occurred. The site of curvature that had the graphically, but as the deformity progressed the hemiver
worst prognosis , for each type of vertebral anomaly , was tebrae tended to be obscured because of the severity of the
usually the thoracolumbar region, and the prognosis was curve, and the anomaly then became indistinguishable
only slightly less severe in the lower thoracic region. from an unsegmented bar alone. The median rate of de
These findings differ from those of Winter et al. , who terioration before the age of ten years ranged from 5 de
found that lower thoracic curves had a worse prognosis grees per year for upper thoracic curves to more than 10
than thoracolumbar curves . We agree with Winter et al. degrees per year for thoracolumbar curves . Sixteen of
that lower thoracic and thoracolumbar curves usually have eighteen lower thoracic and thoracolumbar curves ex
a worse prognosis than do lumbar curves, and that the ceeded 50 degrees once the patient was three years old
most benign curves occur in the upper thoracic region. (Table IV). This type of congenital scoliosis requires
The type of anomaly causing the most severe scoliosis treatment as soon as the anomaly is diagnosed.
in each region of the spine was a unilateral unsegmented A hemivertebra produces a scoliosis by acting as an
bar with contralateral hemivertebrae at the same level. enlarging wedge on the affected side of the spine , whereas
This was followed in severity by scoliosis caused by a uni in patients with a unilateral unsegmented bar there is re
lateral unsegmented bar alone, two unilateral hemiverte tarded growth on the affected side. The growth imbalance
brae, a single hemivertebra, and a wedge vertebra; the in patients with hemivertebrae is never as severe as in
least severe was scoliosis caused by a block vertebra (Ta those with a unilateral unsegmented bar. Not all hemiver
bles II through VI). tebrae, however, produce the same degree of growth
In addition, we found that the rate of deterioration of imbalance at the same site. Eight patients in our series had
the curves was not constant, but if the curve was present an incarcerated hemivertebra5 , which we defined as a
before the patient was ten years old it usually increased, in small, poorly formed, extra segment of bone tucked into
particular during the adolescent growth spurt. Although the spine between adjacent normal vertebrae. This type of
there was often a relatively wide range in the rate of de hemivertebra has no growth potential and the resulting
terioration, the majority of the curves that were due to curves were all less than 20 degrees , with minimum or no
radiographically similar anomalies and that occurred in the progression. Two patients had a hemivertebra that became
same region tended to deteriorate at approximately the synostosed with a neighboring vertebra (semisegmented),
same rate (Fig. 8). and this also produced a curve that progressed less rapidly
A unilateral unsegmented bar does not contain growth than the others in the same region. The majority of single
plates and therefore cannot grow longitudinally, whereas hemivertebrae, however, caused slowly progressive
normal or nearly normal growth may occur on the opposite curves . Before the patients were ten years old the median

VOL. 64-A, NO. 8, OCFOBER 1982

1146 M . J . MCMASTER AND KUNIYOSHI OHTSUKA

rate of deterioration for both upper thoracic and lumbar apex of the curve the more severe was the deformity. An
curves due to a single hemivertebra was 1 degree per year elevated shoulder line was most distressing to girls, and a
and no curve exceeded 40 degrees (Table V) . These curves 30-degree curve seemed to be the upper limit that the pa
usually do not require treatment. Single hemivertebrae in tients would tolerate.
the lower thoracic and thoracolumbar regions caused more An additional problem associated with thoracic
severe deformity. Before the patients were ten years old, curves , especially those with the apex at the fifth, sixth , or
the median rate of deterioration was 2 degrees per year, seventh thoracic vertebra, was the development of a long
and six of nine curves were between 40 and 50 degrees. secondary structural curve in the lower thoracic or
These patients often require treatment during the adoles thoracolumbar region. This curve, which did not involve
cent growth spurt. Two unilateral hemivertebrae produced any congenital anomalies , initially was compensatory and
a greater growth imbalance and caused a much more se was correctable , but later it tended to become fixed and to
vere deformity. The median rate of deterioration before the deteriorate even more rapidly than the primary (congeni
age of ten years ranged from 2 degrees per year for upper tal) curve (Figs. 3-A and 3-B). Moreover, the congenital
thoracic curves to 5 degrees per year for thoracolumbar thoracic curve usually had only a mild degree of rotation,
curves . All seven lower thoracic and thoracolumbar curves whereas the lower secondary curve often was severely ro
were greater than 50 degrees by the time the patient was tated. The large rib hump so produced usually constituted
ten years old. These curves usually require treatment be a major deformity. In nineteen of forty-five patients who
fore the patient is five years old. had a congenital curve due to a unilateral unsegmented
A wedge-shaped vertebra is due to a unilateral partial bar, with or without contralateral hemivertebrae in the
failure of formation of a vertebra, with retarded longitudi upper thoracic segments , a secondary structural lower
nal growth on the hypoplastic side. In two patients, two thoracic or thoracolumbar curve developed.
adjacent vertebrae were affected, but this did not unduly In patients with a lower thoracic, thoracolumbar, or
increase the severity of the scoliosis . The one upper lumbar curve, especially those due to a unilateral unseg
thoracic curve and one lumbar curve in our series de mented bar with or without contralateral hemivertebrae, a
teriorated slowly (less than 2 degrees per year) and did not compensatory curve that was adequate to balance the con
require treatment, whereas the seven lower thoracic and genital curve often failed to develop because there were
thoracolumbar curves deteriorated somewhat more rapidly too few normal mobile vertebrae between the anomaly and
(median rate of increase, 2 degrees per year) and three re either the upper end of the spine or the sacrum. As a result,
quired treatment during the adolescent growth spurt. in 80 per cent of the patients with a congenital
A block vertebra is due to bilateral failure of segmen thoracolumbar curve and in all with a lumbar curve in our
tation, and longitudinal growth is impaired on both sides series some degree of pelvic obliquity and an apparent
of the spine but not always symmetrically. The result was shortening of one lower limb developed (Figs. 5-B and
a mild degree of curvature that rarely exceeded 20 degrees, 5-C). Decompensationor listing of the upper part of the
and the thirteen patients in our series who were so affected body to one side was also a common finding associated
did not require treatment. with the more severe lower thoracic and thoracolumbar
The thirteen patients who had twoor more congenital curves and could result in a very severe deformity.
curves had a variable prognosis , depending on the site of The twelve lumbosacral curves in our series were all
the opposing anomalies . If the anomalies , either similar or due to a single hemivertebra at the lumbosacral junction
dissimilar, occurred within a few segments of each other but, unlike the situation with congenital lumbar or
and were in the same region, they tended to balance each thoracolumbar curves, the pelvis remained level. The
other and produced little deformity other than a kink in the hemivertebra, therefore, caused the lumbar spine to take
spine. If, however, they were widely separated and in dif off obliquely from the sacrum. In an attempt to keep the
ferent regions , the resulting curves tended to be unbal body balanced, a secondary lumbar or thoracolumbar
anced, producing decompensation or listing of the trunk curve developed, but unfortunately it usually was in
that required treatment. sufficient to prevent the trunk from listing to the side op
Apart from the rate of deterioration and ultimate se posite that of the hemivertebra. In nine patients, both the
verity of a congenital scoliosis , there were also a number list and the secondary curve remained small (less than 36
of important secondary features that should be emphasized degrees) and produced only a mild to moderate deformity.
because they contributed significantly to the over-all dis In three patients, however, the secondary curve became
ability and deformity of the patient. large, fixed, and rotated and caused a major cosmetic de
Upper thoracic curves, especially those that extended formity due to a large rib hump and decompensation of the
cranially beyond the cervicothoracic junction, commonly trunk (Figs. 7-B and 7-C).
produced a cosmetic deformity due to elevation of the In conclusion, we can state that congenital scoliosis is
shoulder or, less frequently, tilting of the head (Figs. 2-A a potentially serious condition, which can and often does
and 2-B) . Because congenital scoliosis occurs so fre result in severe curvature of the spine and malalignment of
quently in the upper thoracic region (3 1 per cent of our pa the body. We have shown that it is possible to anticipate
tients) the deformity is a common one, and the higher the the course of a congenital scoliosis if the type of vertebral

THE JOURNAL OF BONE AND JOINT SURGERY

 THE NATURAL HISTORY OF CONGENITAL SCOLIOSIS 1147

anomaly and its site are known. Ideally a congenital for progression can be recognized and an appropriate
scoliosis should be diagnosed while the patient is young prophylactic course of treatment can be planned so as to
and the curve is small. At that stage, a curve that is at risk prevent severe deformity.

References
1. ARICIN,A. M.: Conservative Management of Scoliosis. Clin. Orthop. , 1: 99-108, 1953.
2. BICK, E. M., and COPEL, J. W.: Longitudinal Growth of the Human Vertebra. A Contribution to Human Osteogeny. J. Bone and Joint Surg.,
32-A: 803-814,Oct. 1950.
3. BILLING, E. L.: Congenital Scoliosis: An Analytical Study of Its Natural History. In Proceedings of the Western Orthopedic Association. J.
Bone and Joint Surg. , 37-A: 404-405, April 1955.
4. KUHNS, J. G., and HORMELL, R. S.: Management of Congenital Scoliosis. Review of One Hundred Seventy Cases. Arch. Surg. , 65: 250-263,
1952.
5. MOE,J. H.; WINTER,R. B.; BRADFORD,
D. S.; and LONSTEIN,
J. E.: Scoliosisand OtherSpinal Deformities.Philadelphia,W. B. Saunders,
1978.
6. NASCA,R. J.; STELLING,F. H., III; and STEEL, H. H.: Progression of Congenital Scoliosis Due to Hemivertebrae and Hemivertebrae with Bars.
J. Bone and Joint Surg. , 57-A: 456-466, June 1975.
7. TANNER, J. M.; WHITEHOUSE,R. H.; and TAKAISHI,M.: Standards from Birth to Maturity for Height, Weight, Height Velocity, and Weight
Velocity: British Children, 1965. Arch. Dis. Child., 41: 454-471, 613-635, 1966.
8. TERMINOLOGY COMMITTEE OF THE SCOLI0SIS RESEARCH SOCIETY: A Glossary of Scoliosis Terms. Spine, 1: 57-58, 1976.
9. WINTER,R. B.; MOE, J. H.; and EILERS,V. E.: Congenital Scoliosis. A Study of 234 Patients Treated and Untreated. Part I: Natural History. J.
. Bone and Joint Surg. , 50-A: 1-15, Jan. 1968.

10. WYNNE-DAVIES,RUTH: Congenital Vertebral Anomalies: Aetiology and Relationship to Spina Bifida Cystica. 3. Med. Genet. , 12: 280-288,
1975.
11. ZAOUSSIS,A. L., and JAMES,J. I. P.: The Iliac Apophyses and the Evolution of Curves in Scoliosis. J. Bone and Joint Surg. , 40-B(3): 442-453,
1958.

VOL. 64-A, NO. 8, OCTOBER 1982