Agirl weighing 3420 grams was delivered after thirty-nine weeks of
gestation to a primiparous mother in an obstetrics clinic. The baby presented
in the double footling breech position. The right upper extremity was forcibly
manipulated because it was caught by the umbilical cord during the vaginal
delivery. On the day following birth, the infant did not move the right arm
and cried on passive movement of that extremity. On the third day after birth,
a radiograph made at the obstetrics clinic revealed a fracture of the right
clavicle (Fig. 1). Because the
mother later noticed that the child had a limitation of adduction and internal
rotation of the right shoulder, the infant was examined on the thirteenth day
after birth at an orthopaedic clinic and radiographs made there showed callus
around the proximal aspect of the humerus.
The patient was examined at our clinic on the following day. She maintained
the right shoulder in an abducted and externally rotated position, and a deep
cleft was visible in the skin of the lateral shoulder. A hard mass was
palpable in the anteromedial aspect of the shoulder, and the humeral head was
palpable under the acromion. An osseous mass, 2.5 cm in diameter, was noted at
the middle of the clavicle, and an elastic mass, 2 cm in diameter, was noted
in the mid-portion of the sternocleidomastoid muscle. No neurologic deficit
was present. Arthrography showed that the humeral head was in its normal
position in relation to the glenoid (Fig.
2). The infant was diagnosed as having epiphysiolysis of the
proximal aspect of the right humerus, a clavicular fracture, and myogenic
torticollis due to birth trauma.
The fractured bone had completely united without any limitation of
adduction seven weeks after the patient was born. The right shoulder, with the
arm at the side, had passive external rotation to 180° (the left had
100°) and 0° of internal rotation. When the child was one year and two
months old, marked rotational deformity was still seen on the radiographs
(Fig. 3). When the girl was one
year and nine months old, the right shoulder, with the arm at the side, had
passive external rotation to 160° (the left had 90°) and internal
rotation to the level of L1 (to the level of T3 on the left). With the arm in
90° of abduction, the right shoulder had 180° of passive external
rotation (the left had 120°) and 0° of internal rotation (90° on
the left). Thereafter, the range of external rotation with the arm at the side
gradually diminished and the range of internal rotation increased, but there
was no appreciable change in the range of external or internal rotation with
the arm in 90° of abduction. When the patient was twelve years old, the
right shoulder, with the arm at the side, had 120° of external rotation
(the left had 90°) and internal rotation to the level of T9 (to the level
of T3 on the left). With the arm in 90° of abduction, the right shoulder
had 140° of external rotation (the left had 105°) and 5° of
internal rotation (the left had 60°). Humeral head retroversion was
80° on the right side (37° on the left) as measured with use of
computed tomography scans according to the method of Symeonides et
al.8. When the
patient was sixteen years old, the right shoulder, with the arm at the side,
had 120° of external rotation (the left had 90°) and internal rotation
was to the level of T8 (to the level of T2 on the left). With the arm in
90° of abduction, the right shoulder had 140° of external rotation
(the left had 100°) and 10° of internal rotation (the left had
70°). When the patient was twenty years old, the range of motion did not
differ from that achieved when the patient was sixteen years of age; there was
no evidence of excess laxity in any direction, the load-and-shift test and the
apprehension test both yielded negative results, and the patient remained
entirely free from limitations of daily activities.
Humeral retroversion, measured on computed tomography scans, was 77° on
the right (30° on the left) at both the examination when she was sixteen
years old and when she was twenty years old. The difference in torsion between
the two humeral diaphyses was measured on the computed tomography scans that
had been made when the patient was sixteen years old (Figs.
4-A and 4-B). This measurement
demonstrated that the affected humerus was rotated uniformly throughout the
central 50% of the diaphysis. Throughout the follow-up period, the humerus was
consistently shorter on the right side, although the difference never exceeded
20 mm clinically or radiographically.
Birth trauma occurs frequently during breech extractions from primiparous
mothers9. Upper-limb
birth trauma occurs more often on the right side of the infant's body, with
the clavicle being the most frequently fractured
bone10. Proximal
humeral epiphysiolysis, which is the most frequent shoulder
injury11, is
usually a Salter-Harris type-I
injury12; the only
exception, to our knowledge, was a patient with a type-II pattern reported by
Jones et al.13.
An early, correct diagnosis of proximal humeral epiphysiolysis is often
difficult to make because ossification of the proximal humeral epiphysis is
visualized radiographically in only 50% of newborns at two weeks after
birth14.
Displacement of the proximal metaphysis in relation to the glenoid may be the
only radiographic
finding15,16.
Because callus becomes demonstrable on radiographs within five to seven days
after the injury, bone injury is often first diagnosed at that
point9,17.
Arthrography15,16,18,
ultrasonography16,19,
and magnetic resonance
imaging20 are
useful for making an early diagnosis. In our patient, the radiographic
evidence of callus and the arthrographic verification of the position of the
humeral head made it possible to establish the diagnosis.
Fractured long bones in children are widely recognized to have a high
potential to remodel an angular deformity. An angular deformity is corrected
independently in the physis according to the Hueter-Volkmann law and in the
fracture region according to Wolff's
law2. In a clinical
study on radial fractures, Gandhi et
al.1 noted that
correction of an angular deformity is most rapid and most complete when there
is separation at the distal epiphysis, where three-fourths of longitudinal
humeral growth takes place. The proximal physis of the humerus contributes 80%
of the longitudinal growth of that
bone21, so
fractures at that site exhibit considerable remodeling
potential5. However,
the ability of long bones to correct a rotational deformity remains undefined.
Although no conclusion has yet been drawn from extensive clinical studies on
fractures of the immature femoral
shaft4,6,7,
Murray et al.4, in a
report on an experimental investigation demonstrating a helical growth pattern
at the physis, stated that rotational deformities were able to correct as a
result of this helical pattern.
The fracture in our patient resulted in at least 80° of excessive
retroversion of the humerus at seven weeks, evidenced by the fact that
external rotation of the affected shoulder was 80° greater than that of
the unaffected shoulder. The retroversion of the humerus on the affected side,
which was measured on computed tomography scans performed when the patient was
sixteen and twenty years of age, was 47° greater than that on the
unaffected side. Consequently, the amount of correction of the rotational
deformity that was achieved with growth was 33°.
Krahl noted that retroversion of the humerus diminishes physiologically by
9° from the time of birth until
adulthood22, which
indicates that retroversion decreases normally with growth.
Edelson23 reported
that humeral retroversion in children reaches adult values by the age of
sixteen years. The osseous correction of rotational deformity that we observed
in our patient appears to represent, in part, the growth-associated
physiological decrease in the retroversion of the humerus. On the basis of the
uniform change in humeral shaft torsion that was observed on computed
tomography scans made when the patient was sixteen years of age, it appears
that the osseous remodeling proceeded almost uniformly during growth.
Siegel et al.24,
who followed patients after in situ fixation of a slipped capital femoral
epiphysis, suggested that soft-tissue stretching might account for the
recovery of motion despite only minimal change in the osseous structure. In
our patient, stretching of the adjoining soft tissues might have affected the
improvement in motion to some extent as the residual excessive retroversion
did not interfere with daily activities. Nevertheless, the soft-tissue
stretching did not generate any shoulder instability.
The case of our patient serves to emphasize that surgical intervention is
not required for a rotational deformity resulting from proximal humeral
epiphysiolysis due to birth trauma; rather, rotational remodeling of the
humerus and stretching of the adjoining soft tissues with growth can provide
adequate compensation. ?