Intellectual and Perceptual-Motor Characteristics of Treated Myelomeningocele Children Pegeen \s=b\ We
L.
Soare, PhD, Anthony J. Raimondi, MD
studied 173 children with my-
elomeningocele, 133 of whom developed hydrocephalus (and had shunt procedures) and 40 of whom did not. Eighty siblings were tested as a control group. Sixty-three percent of children with hydrocephalus had IQs above 80; 87% of those without hydrocephalus had IQs above 80. Children who had associated hydrocephalus were significantly less intelligent than their siblings, whereas those without hydrocephalus were not. When patients and siblings were matched by age and IQ, the former scored significantly lower on a perceptual-motor functioning test. When patients with and without hydrocephalus were similarly matched, those with hydrocephalus scored lower. Inverse relationships between sac location and IQ, and sensory level and IQ, were found to be dependent on the association of higher sac levels and of sensory loss with hydrocephalus. Patient IQ was related to family income and education. (Am J Dis Child 131:199-204, 1977)
few
exceptions,'-7 published reports With intelligence myelomeningocele of the
of children with (MM/H) and without (MM) hydro¬ cephalus have been limited to analyz¬ ing the natural history of nonoperatively treated hydrocephalus. By and large, these studies have been retro¬ spective, based on an accumulation of varying numbers of survivors. Simi¬ larly, the appraisals of survival, quali¬ ty of survival, and impact of the affected child on his home and enviFrom the Division of Neurological Surgery, Northwestern University Medical School, and Children's Memorial Hospital, Chicago. Reprint requests to Division of Neurological Surgery, Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614 (Dr Raimon-
di).
ronment have been made from anal¬ yses of groups of children whose management has not been adequately
described.
METHOD As a first step in studying these children established the criteria for diagnosing hydrocephalus in the child with myelomeningocele, determining precisely when a we
shunt was no longer functioning, compar¬ ing intellectual quotients with the presence or absence of hydrocephalus (treated, in the former event), using the siblings as the norm, and identifying (neuroradiologically) the nature and degree of brain dyspla¬ sia."'3 The two groups studied were myelomeningocele (MM) and shunted hy¬
drocéphalie myelomeningocele (MM/HS)
children. The abbreviation MM/H refers to children who did not undergo a shunt
procedure. There is,
at the present time, consid¬ erable interest in the question of whether to treat such children at all, or whether treatment should be offered only to those children who may be expected to "live a meaningful life." Criteria on which such a decision may be based are lacking, though some arbitrary criteria (such as level of the
myelomeningocele sac at birth, paraplegia,
and hydrocephalus) have been suggested. In fact, one author has carried prediction to an extreme,' reporting such a poor prog¬ nosis for myelomeningocele children with a thoracolumbar lesion, for example, irre¬ spective of the presence of hydrocephalus, that he recommends withholding all forms of treatment. This article presents the results of our study of the intellectual development of 173 children with myelomeningocele, 133 who had a shunt for hydrocephalus (MM/ HS) and 40 in whom hydrocephalus never developed (MM). These data are compared with those for untreated children (MM/H) and to siblings, because intellectual and social potentials are two important factors -
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
by
which the quality of survival may be evaluated. It is desirable that this informa¬ tion be presented now that we are attempt¬ ing to establish criteria on which to base a decision to treat or not to treat. We identi¬ fied three goals: (1) to explore the interre¬ lationships between the site of the sac, presence or absence of hydrocephalus, and IQ; (2) to establish norms for our sample of patients, giving due consideration to what might be expected of each child according to his socioeconomic background; (3) to determine whether perceptual-motor defi¬ cits are present and, if so, whether they are present in MM, MM/HS, or both groups.
SUBJECTS This series consists of 173 children with myelomeningocele, 133 of whom were shunted for hydrocephalus (77%) and 40 without hydrocephalus. The cases were selected only on the basis of survival to the time of testing. Since patients were not routinely tested until 16 to 18 months of age, children who were too young (5), or who were scheduled to be seen for the first time in the next six months (18), and/or who died before they were old enough to be tested (68) are not included. We have also excluded 73 children who were lost to follow-up before we could test them the first time (almost all had moved). There¬ fore, these 173 children represent about half (51%) of the total number managed by us.
Because Children's Memorial Hospital receives many referrals from outside the immediate vicinity, our sample is quite heterogeneous. Seventy-eight percent of the children were white; 11%, SpanishAmerican; and 11%, black. Thus far we have tested 61 siblings (two half-siblings) of MM/HS and 19 siblings of MM chil¬ dren. Seven of the MM/HS patients are currently without shunts, two following a posterior fossa decompression and the others following ventriculitis. Five MM/ HS children died since they were initially
tested. The mean IQ of those five was 60. The average age at the time of death was 3 years, 3 months. If the myelomeningocele sac was covered with skin, no surgery was recommended at birth. If it was not covered with skin, repair was routinely performed within 24 hours of birth. If, for reasons such as delayed referral, the child could not be operated on during the first day of life, the sac was treated with ten-minute surgical scrubs every six hours (daily cultures were taken) until the area gave negative bacte¬ rial growth results for five consecutive days, at which time the sac was closed surgically. There was no correlation be¬ tween surgical closure of the sac and sur¬ vival, the onset of hydrocephalus, or neuro¬
logic deficit. The diagnosis of hydrocephalus, and the distinction between aqueductal occlusion and constrictive hydrocephalus, was made neuroradiologically in all children, as was
the identification of the Arnold-Chiari II anomaly, associated tentorial dysplasia,
mesencephalic hyperplasia, microgyria,
and buckling of the medulla oblongata on the medulla spinalis.s All children who had a diagnosis of hydrocephalus received shunts immediately, followed closely there¬ after by revisions performed on an emer¬ gency basis when indicated. We consider "arrested" hydrocephalus a pathologic con¬ dition. ''
MATERIALS The instruments used to evaluate these children were (1) the Stanford-Binet Intel¬ ligence Scale, Form L-M; (2) the Cattell
Infant Intelligence Scale, a downward extension of the Stanford-Binet; (3) the Vineland Social Maturity Scale, which measures self-help skills; (4) the Goodenough-Harris Draw-a-Man test; and (5) the Beery-Buktenica Developmental Test of Visual-Motor Integration, or VMI," which measures perceptual-motor skills. The latter was chosen over other percep¬ tual-motor tests because it can be applied to a wide age range. The average chronologic age at the time of the first psychologic test for the MM/HS children was 3 years 11 months, with a range from 11 months to 12 years and 4 months. For the 40 MM patients the average chronologic age was 7 years 4 months, with a range from 10 months to 18 years and 11 months. We obtained information on (1) size of family; (2) occupation of the breadwinner; (3) income; (4) age of parents at birth of patient; and (5) education of the parents. The parents were requested to bring in the next older sibling of the patient for testing. It is generally accepted that there
correlation of about .50 between the It has been shown that the correlation of IQ with socioeconomic status is in the range of .20 to .40.'" Thus, the rationale for getting the sibling's IQ and the socioeconomic data was to determine as accurately as possible the patient's potential for development apart from the disease entity. In order to assess the effects of socioeco¬ nomic variables, an index or composite score, based on the mother's education and the family income, was developed. The US Bureau of the Census classification system was used so that our sample could be compared to national averages. Data on education and family income were based on figures from 1969. These composite scores were then split at the sample median, and the IQs of those above the median were compared with those below it. The VMI'4 was used as an indicator of perceptual-motor functioning. This test yields an age equivalency score, which for normal children should be equivalent to the chronologic age. Fifty-nine of the 133 MM/ HS children and 38 siblings had VMI scores. Twenty-nine of the 40 MM children and 12 of their siblings had VMI scores. The children who did not have VMI scores were either too young or too retarded. To assess the effect of length of hospital¬ ization on perceptual-motor functioning, the ten children with the greatest percep¬ tual-motor disabilities were compared with the ten with the least perceptual-motor disabilities for days spent in the hospital during the first 3 years of life a length of time chosen to represent the early, most crucial years, when lack of stimulation could be presumed to have the most effect. Only MM/HS children were included as the larger of the two MM groups. Each child was placed in one of the following groups, based on the site of the is
a
IQs of full siblings in a family.'
myelomeningocele
saç
'
(1) occipito-cervical
and cervical; (2) thoracolumbar and thoracolumbosacral; (3) lumbar; and, (4) lumbosacral and sacral. A few children were listed as having thoracic lesions, but the numbers were too small to be considered separately. All patients with sensory levels
(recorded
at most recent
examination)
at
T12 and above were combined into one group, and those with no sensory loss into another. Based on the IQ scores, an a posteriori decision was made to combine levels L1-L4 into one group and levels LO¬ SS into another.
RESULTS
Children whose myelomeningocele repaired on the day of birth, or the following day, were classified as
was
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
within the first 24 hours. As may be seen from Table 1, 43% of the children were operated on within the first 24 hours. The mean age at the time of initial shunt placement for MM/HS children was 2.6 months. In a previous study we showed that IQ is not related to the number of shunt revisions." For the sample of 133 MM/HS patients described here, the mean number of shunt revisions per child was 3.6. Table 2 shows the distribution of shunt revisions over time after the initial shunt placement. All patients who were at least 2 years of age at the time the analysis was done were included in the "2-5 year" groups. Only 75 patients were over 5 years of age and included in that group. Because this is a longitudinal study, each child is reexamined at approxi¬ mately yearly intervals. New patients are added continuously. Thus far, 2 patients have been seen five times, 9 have been seen four times, 23 have been seen three times, 56 have been seen twice, and 82 have been seen only once. Between the first and most recent tests, 41 patients increased IQ, 45 decreased, and 4 showed no change. The average change for all patients was only +.33 IQ points. Of those tested more than once, 629c changed ± 10 points, 26% changed from 11 to 20 points, and 12% changed by more than 20 points. The mean IQs for the MM and MM/HS patients and their siblings, as recorded at their most recent test, are presented in Table 3. The MM group was significantly brighter than the MM/HS (P < .001). Of the latter group, 63% had IQs above 80, while 87% of the former had IQs above 80. Sixty-one siblings of MM/HS pa¬ tients were tested. The IQs of the 61
being repaired
patient-sibling pairs were examined. The means were significantly differ¬
5.50, < .01, and the correla¬ tion between the two groups was only .21 (not significant). Nineteen siblings of MM children were tested; they did not differ significantly from the patients, and the correlation between the two groups was .52 (P < .01), which is what would be expected in the normal population. A socioeconomic index, based on ent, t
=
Table With
1.—Age
at
The ten children with the greatest
Repair of Myelomeningocele
Hydrocephalus
Without
Hydrocephalus
_No. (%)_No. (%)_Total 24 hr 15
(45) 23(17) 29(23) 20 (15)" 60
1
day-1 wk 8 days-1 mo 1
mo
'One child with hydrocephalus
was never
Time After Initial Shunt < 6 mo 6 mo-1 yr 1-2 yr
0.59 0.61 1.04 0.47
2-5 yr > 5 yr
Mean Cumula¬ tive No. of Revisions 1.4
1.73 2.34 3.38 3.63
Of the total number of patients with shunts, 24% had no revision, 38% had one to three revisions, 20% had four to six revisions, and 18% had more than six revisions.
mother's education and family in¬ was devised for our total sample, and these index scores were split at the median. The mean IQ of the 85 patients above the median was 100.6 (SD 20.0), and that of those below the median 79.8 (SD 24.8). The difference between the two means come,
significant (t 6.0, < .001), indicating that the group above the median had higher IQs. Approximate¬ ly equal numbers of MM and MM/HS was
=
above and below the median. We were interested in determining how the perceptual-motor age scores compared with the chronologic age (Table 4). A t test for related means showed a significant difference be¬ tween VMI ages and chronologic ages for the MM/HS (t 9.21, < .001) and MM children (f 4.41, < .001), the VMI ages being lower. However, a t test between chronologic ages and VMI ages for the siblings was also significant (t 3.08, < .01). Be¬ cause the correlation coefficient of VMT age with chronologic age was .70 and that with IQ .51, a different statistical procedure had to be used to compare patients with siblings on the VMI. This was necessary because the differences in IQ and chronologic age were
=
=
=
life. The difference between the two groups was not significant. When the two groups
repaired.
Table 2.—Mean Number of Shunt Revisions Over Time for 133 Patients*
Mean No. of Revisions 1.4
75 (43) 32(18) 36(21) 29 (17)
(38) 9(22) 7(18) 9 (22)
perceptual-motor difficulties were compared with the ten with the least difficulties as to days spent in the hospital during the first 3 years of were
compared
as
to
days
spent in the hospital during the first 5
Table 3.-IQ Scores of Myelomeningocele Children With and Without Hydrocephalus and Their Siblings MM/ HS Mean 87.7 SD 24.8 133
Siblings 109.5 17.4 61
MM 102.3 19.9
Siblings
40
19
119.9 13.6
between the patients and siblings would bias the results. A selection of patients and siblings was made on the basis of those who could be matched both according to chronologic age (within +2 years of each other) and IQ (within ± 15 points of each other). This procedure pro¬ vided 14 pairs. The difference scores (chronologic age minus VMI age) of
both groups were then compared. Table 5 shows that the matching procedure was successful in that the mean chronologic age and IQ of the patients and siblings were very close. A t test for related means showed that there was a significant differ¬ ence between patients and siblings (Table 5), indicating that the patients did more poorly on the perceptualmotor tests than did their siblings. In order to compare the perceptualmotor skills of MM and MM/HS chil¬ dren, the same type of matching procedure was used. Patients were selected from the two groups and matched according to the criteria of age and IQ (the means are presented in Table 6). This resulted in 26 pairs. The difference scores (chronologic age minus VMI) were then compared. A t test for related means showed that there was a significant difference between MM and MM/HS children (Table 6), indicating that the MM/HS children had more perceptual-motor difficulties than did the MM chil¬ dren.
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
years of
life, the groups were signifi¬ cantly different (P < .05), the group with least difficulties having spent fewer days in the hospital. The relationship between percep¬
tual-motor deficit and level of the myelomeningocele sac was then stud¬ ied. Only one of the ten children with the most deficit had a lesion in the lumbosacral area. All others were in the lumbar or thoracic area. Of the children with the least deficit, six had lesions in the lumbosacral area and four within the lumbar and thoracic areas. The Fisher Exact Probability test'7 showed that there was a signifi¬ cant association between higher le¬ sions and greater perceptual-motor difficulties (P < .05, one-tailed). Another area of interest was to determine whether there was a rela¬ tionship between the level of the myelomeningocele sac, sensory level, and IQ, independent of the effects of hydrocephalus. In order to test for this, we analyzed the MM and MM/HS groups separately. The mean IQ scores for the MM and MM/HS chil¬ dren, listed according to the site of the
myelomeningocele
sac,
are
presented
in Table 7. Excluded from Table 7 were three MM/HS children with a thoracic level sac (IQs: 94, 106, 84) and three MM children with a thoracic level (IQs: 64, 114, 77); because the numbers were too small for analysis. The group with lesions in the occipitocervical areas was excluded from the analysis of variance because the requirement of homogeneity of variance was not met (the standard deviation was too different from the others). The dele¬ tion of this group does not affect the relationship. An analysis of variance was done on the IQ scores of the MM/ HS group of children and was not
significant. For the MM group of children a t was done between the lumbar
test
groups and the lumbosacral and sacral groups and also was not significant. When an analysis of variance was done on the means for both groups together, it was significant at the .01 level, indicating that the IQs of chil¬ dren with myelomeningocele differ significantly according to the level of
the
Table 4—Mean IQ, Chronologie Age, and VMI Scores for Myelomeningocele Children With and Without Hydrocephalus and Their Siblings
method of comparison, groups were formed on the basis of the sensory level. These data are pre¬ sented in Table 8. Analyses of variance were done on the mean IQ scores of the MM/HS and MM groups separately and were not significant. When the means of both groups were analyzed together, a significant difference was found (P < .05), indicating that the IQs of the groups based on sensory level were significantly different from each other.
Mean
SD
Our results, showing that MM/HS children performed at a significantly lower intellectual level than MM chil¬ dren, are consistent with results reported by others16-1"·20 for untreated children with myelomeningocele and hydrocephalus, although our differ¬ ences are of a lesser magnitude. In fact, our data show that 63% of the MM/HS children have IQs above 80, while Merrill et al" reported a 10% frequency and Stephen1* one of 40% for untreated hydrocéphalie MM chil¬ dren. Although it is difficult to compare the intellectual outcome of treated MM/HS and untreated MM/ HS hydrocéphalie children because of the different survival rates, it is true that the presence of hydrocephalus in a child with myelomeningocele de¬ creases his chances for normal intelli¬ gence. Though surgical treatment of the hydrocephalus increases the chances for normal intelligence, my¬ elomeningocele children with hydro¬
cephalus, as a group, are still signifi¬ cantly less intelligent than their normal siblings and also less intelli¬ gent than myelomeningocele children without hydrocephalus.
The percentage of MM/HS children with IQs above 80 (63%) is higher than that reported by Lorber' (52% and 57% for two series), Ames and Schuf
(n
All
29)
=
Siblings (n
=
50)
109.0 15.4
95.0 20.1
101.4
8.0 2.4
9.6 3.5
10.0
5.7 1.8
7.7 2.5
9.2 2.2
20.0
Chronologic age, yr Mean SD
a
COMMENT
MM
59)
=
IQ
sac.
As
MM/HS (n
Scores
VMI, yr Mean SD
Table 6.—Mean Chronologie Age (CA), IQ, and Difference Scores for MM and MM/HS Children*
Table 5.—Mean Chronologie Age (CA), IQ, and Difference Scores for Patients (MM/HS and MM) and Siblings* Patients 9.3 105.5
CA, yr IQ Difference score, yr (CA-VMI)t *N
tf
=
=
14
pairs.
3.73, P
L.
Soare, PhD, Anthony J. Raimondi, MD
studied 173 children with my-
elomeningocele, 133 of whom developed hydrocephalus (and had shunt procedures) and 40 of whom did not. Eighty siblings were tested as a control group. Sixty-three percent of children with hydrocephalus had IQs above 80; 87% of those without hydrocephalus had IQs above 80. Children who had associated hydrocephalus were significantly less intelligent than their siblings, whereas those without hydrocephalus were not. When patients and siblings were matched by age and IQ, the former scored significantly lower on a perceptual-motor functioning test. When patients with and without hydrocephalus were similarly matched, those with hydrocephalus scored lower. Inverse relationships between sac location and IQ, and sensory level and IQ, were found to be dependent on the association of higher sac levels and of sensory loss with hydrocephalus. Patient IQ was related to family income and education. (Am J Dis Child 131:199-204, 1977)
few
exceptions,'-7 published reports With intelligence myelomeningocele of the
of children with (MM/H) and without (MM) hydro¬ cephalus have been limited to analyz¬ ing the natural history of nonoperatively treated hydrocephalus. By and large, these studies have been retro¬ spective, based on an accumulation of varying numbers of survivors. Simi¬ larly, the appraisals of survival, quali¬ ty of survival, and impact of the affected child on his home and enviFrom the Division of Neurological Surgery, Northwestern University Medical School, and Children's Memorial Hospital, Chicago. Reprint requests to Division of Neurological Surgery, Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614 (Dr Raimon-
di).
ronment have been made from anal¬ yses of groups of children whose management has not been adequately
described.
METHOD As a first step in studying these children established the criteria for diagnosing hydrocephalus in the child with myelomeningocele, determining precisely when a we
shunt was no longer functioning, compar¬ ing intellectual quotients with the presence or absence of hydrocephalus (treated, in the former event), using the siblings as the norm, and identifying (neuroradiologically) the nature and degree of brain dyspla¬ sia."'3 The two groups studied were myelomeningocele (MM) and shunted hy¬
drocéphalie myelomeningocele (MM/HS)
children. The abbreviation MM/H refers to children who did not undergo a shunt
procedure. There is,
at the present time, consid¬ erable interest in the question of whether to treat such children at all, or whether treatment should be offered only to those children who may be expected to "live a meaningful life." Criteria on which such a decision may be based are lacking, though some arbitrary criteria (such as level of the
myelomeningocele sac at birth, paraplegia,
and hydrocephalus) have been suggested. In fact, one author has carried prediction to an extreme,' reporting such a poor prog¬ nosis for myelomeningocele children with a thoracolumbar lesion, for example, irre¬ spective of the presence of hydrocephalus, that he recommends withholding all forms of treatment. This article presents the results of our study of the intellectual development of 173 children with myelomeningocele, 133 who had a shunt for hydrocephalus (MM/ HS) and 40 in whom hydrocephalus never developed (MM). These data are compared with those for untreated children (MM/H) and to siblings, because intellectual and social potentials are two important factors -
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
by
which the quality of survival may be evaluated. It is desirable that this informa¬ tion be presented now that we are attempt¬ ing to establish criteria on which to base a decision to treat or not to treat. We identi¬ fied three goals: (1) to explore the interre¬ lationships between the site of the sac, presence or absence of hydrocephalus, and IQ; (2) to establish norms for our sample of patients, giving due consideration to what might be expected of each child according to his socioeconomic background; (3) to determine whether perceptual-motor defi¬ cits are present and, if so, whether they are present in MM, MM/HS, or both groups.
SUBJECTS This series consists of 173 children with myelomeningocele, 133 of whom were shunted for hydrocephalus (77%) and 40 without hydrocephalus. The cases were selected only on the basis of survival to the time of testing. Since patients were not routinely tested until 16 to 18 months of age, children who were too young (5), or who were scheduled to be seen for the first time in the next six months (18), and/or who died before they were old enough to be tested (68) are not included. We have also excluded 73 children who were lost to follow-up before we could test them the first time (almost all had moved). There¬ fore, these 173 children represent about half (51%) of the total number managed by us.
Because Children's Memorial Hospital receives many referrals from outside the immediate vicinity, our sample is quite heterogeneous. Seventy-eight percent of the children were white; 11%, SpanishAmerican; and 11%, black. Thus far we have tested 61 siblings (two half-siblings) of MM/HS and 19 siblings of MM chil¬ dren. Seven of the MM/HS patients are currently without shunts, two following a posterior fossa decompression and the others following ventriculitis. Five MM/ HS children died since they were initially
tested. The mean IQ of those five was 60. The average age at the time of death was 3 years, 3 months. If the myelomeningocele sac was covered with skin, no surgery was recommended at birth. If it was not covered with skin, repair was routinely performed within 24 hours of birth. If, for reasons such as delayed referral, the child could not be operated on during the first day of life, the sac was treated with ten-minute surgical scrubs every six hours (daily cultures were taken) until the area gave negative bacte¬ rial growth results for five consecutive days, at which time the sac was closed surgically. There was no correlation be¬ tween surgical closure of the sac and sur¬ vival, the onset of hydrocephalus, or neuro¬
logic deficit. The diagnosis of hydrocephalus, and the distinction between aqueductal occlusion and constrictive hydrocephalus, was made neuroradiologically in all children, as was
the identification of the Arnold-Chiari II anomaly, associated tentorial dysplasia,
mesencephalic hyperplasia, microgyria,
and buckling of the medulla oblongata on the medulla spinalis.s All children who had a diagnosis of hydrocephalus received shunts immediately, followed closely there¬ after by revisions performed on an emer¬ gency basis when indicated. We consider "arrested" hydrocephalus a pathologic con¬ dition. ''
MATERIALS The instruments used to evaluate these children were (1) the Stanford-Binet Intel¬ ligence Scale, Form L-M; (2) the Cattell
Infant Intelligence Scale, a downward extension of the Stanford-Binet; (3) the Vineland Social Maturity Scale, which measures self-help skills; (4) the Goodenough-Harris Draw-a-Man test; and (5) the Beery-Buktenica Developmental Test of Visual-Motor Integration, or VMI," which measures perceptual-motor skills. The latter was chosen over other percep¬ tual-motor tests because it can be applied to a wide age range. The average chronologic age at the time of the first psychologic test for the MM/HS children was 3 years 11 months, with a range from 11 months to 12 years and 4 months. For the 40 MM patients the average chronologic age was 7 years 4 months, with a range from 10 months to 18 years and 11 months. We obtained information on (1) size of family; (2) occupation of the breadwinner; (3) income; (4) age of parents at birth of patient; and (5) education of the parents. The parents were requested to bring in the next older sibling of the patient for testing. It is generally accepted that there
correlation of about .50 between the It has been shown that the correlation of IQ with socioeconomic status is in the range of .20 to .40.'" Thus, the rationale for getting the sibling's IQ and the socioeconomic data was to determine as accurately as possible the patient's potential for development apart from the disease entity. In order to assess the effects of socioeco¬ nomic variables, an index or composite score, based on the mother's education and the family income, was developed. The US Bureau of the Census classification system was used so that our sample could be compared to national averages. Data on education and family income were based on figures from 1969. These composite scores were then split at the sample median, and the IQs of those above the median were compared with those below it. The VMI'4 was used as an indicator of perceptual-motor functioning. This test yields an age equivalency score, which for normal children should be equivalent to the chronologic age. Fifty-nine of the 133 MM/ HS children and 38 siblings had VMI scores. Twenty-nine of the 40 MM children and 12 of their siblings had VMI scores. The children who did not have VMI scores were either too young or too retarded. To assess the effect of length of hospital¬ ization on perceptual-motor functioning, the ten children with the greatest percep¬ tual-motor disabilities were compared with the ten with the least perceptual-motor disabilities for days spent in the hospital during the first 3 years of life a length of time chosen to represent the early, most crucial years, when lack of stimulation could be presumed to have the most effect. Only MM/HS children were included as the larger of the two MM groups. Each child was placed in one of the following groups, based on the site of the is
a
IQs of full siblings in a family.'
myelomeningocele
saç
'
(1) occipito-cervical
and cervical; (2) thoracolumbar and thoracolumbosacral; (3) lumbar; and, (4) lumbosacral and sacral. A few children were listed as having thoracic lesions, but the numbers were too small to be considered separately. All patients with sensory levels
(recorded
at most recent
examination)
at
T12 and above were combined into one group, and those with no sensory loss into another. Based on the IQ scores, an a posteriori decision was made to combine levels L1-L4 into one group and levels LO¬ SS into another.
RESULTS
Children whose myelomeningocele repaired on the day of birth, or the following day, were classified as
was
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
within the first 24 hours. As may be seen from Table 1, 43% of the children were operated on within the first 24 hours. The mean age at the time of initial shunt placement for MM/HS children was 2.6 months. In a previous study we showed that IQ is not related to the number of shunt revisions." For the sample of 133 MM/HS patients described here, the mean number of shunt revisions per child was 3.6. Table 2 shows the distribution of shunt revisions over time after the initial shunt placement. All patients who were at least 2 years of age at the time the analysis was done were included in the "2-5 year" groups. Only 75 patients were over 5 years of age and included in that group. Because this is a longitudinal study, each child is reexamined at approxi¬ mately yearly intervals. New patients are added continuously. Thus far, 2 patients have been seen five times, 9 have been seen four times, 23 have been seen three times, 56 have been seen twice, and 82 have been seen only once. Between the first and most recent tests, 41 patients increased IQ, 45 decreased, and 4 showed no change. The average change for all patients was only +.33 IQ points. Of those tested more than once, 629c changed ± 10 points, 26% changed from 11 to 20 points, and 12% changed by more than 20 points. The mean IQs for the MM and MM/HS patients and their siblings, as recorded at their most recent test, are presented in Table 3. The MM group was significantly brighter than the MM/HS (P < .001). Of the latter group, 63% had IQs above 80, while 87% of the former had IQs above 80. Sixty-one siblings of MM/HS pa¬ tients were tested. The IQs of the 61
being repaired
patient-sibling pairs were examined. The means were significantly differ¬
5.50, < .01, and the correla¬ tion between the two groups was only .21 (not significant). Nineteen siblings of MM children were tested; they did not differ significantly from the patients, and the correlation between the two groups was .52 (P < .01), which is what would be expected in the normal population. A socioeconomic index, based on ent, t
=
Table With
1.—Age
at
The ten children with the greatest
Repair of Myelomeningocele
Hydrocephalus
Without
Hydrocephalus
_No. (%)_No. (%)_Total 24 hr 15
(45) 23(17) 29(23) 20 (15)" 60
1
day-1 wk 8 days-1 mo 1
mo
'One child with hydrocephalus
was never
Time After Initial Shunt < 6 mo 6 mo-1 yr 1-2 yr
0.59 0.61 1.04 0.47
2-5 yr > 5 yr
Mean Cumula¬ tive No. of Revisions 1.4
1.73 2.34 3.38 3.63
Of the total number of patients with shunts, 24% had no revision, 38% had one to three revisions, 20% had four to six revisions, and 18% had more than six revisions.
mother's education and family in¬ was devised for our total sample, and these index scores were split at the median. The mean IQ of the 85 patients above the median was 100.6 (SD 20.0), and that of those below the median 79.8 (SD 24.8). The difference between the two means come,
significant (t 6.0, < .001), indicating that the group above the median had higher IQs. Approximate¬ ly equal numbers of MM and MM/HS was
=
above and below the median. We were interested in determining how the perceptual-motor age scores compared with the chronologic age (Table 4). A t test for related means showed a significant difference be¬ tween VMI ages and chronologic ages for the MM/HS (t 9.21, < .001) and MM children (f 4.41, < .001), the VMI ages being lower. However, a t test between chronologic ages and VMI ages for the siblings was also significant (t 3.08, < .01). Be¬ cause the correlation coefficient of VMT age with chronologic age was .70 and that with IQ .51, a different statistical procedure had to be used to compare patients with siblings on the VMI. This was necessary because the differences in IQ and chronologic age were
=
=
=
life. The difference between the two groups was not significant. When the two groups
repaired.
Table 2.—Mean Number of Shunt Revisions Over Time for 133 Patients*
Mean No. of Revisions 1.4
75 (43) 32(18) 36(21) 29 (17)
(38) 9(22) 7(18) 9 (22)
perceptual-motor difficulties were compared with the ten with the least difficulties as to days spent in the hospital during the first 3 years of were
compared
as
to
days
spent in the hospital during the first 5
Table 3.-IQ Scores of Myelomeningocele Children With and Without Hydrocephalus and Their Siblings MM/ HS Mean 87.7 SD 24.8 133
Siblings 109.5 17.4 61
MM 102.3 19.9
Siblings
40
19
119.9 13.6
between the patients and siblings would bias the results. A selection of patients and siblings was made on the basis of those who could be matched both according to chronologic age (within +2 years of each other) and IQ (within ± 15 points of each other). This procedure pro¬ vided 14 pairs. The difference scores (chronologic age minus VMI age) of
both groups were then compared. Table 5 shows that the matching procedure was successful in that the mean chronologic age and IQ of the patients and siblings were very close. A t test for related means showed that there was a significant differ¬ ence between patients and siblings (Table 5), indicating that the patients did more poorly on the perceptualmotor tests than did their siblings. In order to compare the perceptualmotor skills of MM and MM/HS chil¬ dren, the same type of matching procedure was used. Patients were selected from the two groups and matched according to the criteria of age and IQ (the means are presented in Table 6). This resulted in 26 pairs. The difference scores (chronologic age minus VMI) were then compared. A t test for related means showed that there was a significant difference between MM and MM/HS children (Table 6), indicating that the MM/HS children had more perceptual-motor difficulties than did the MM chil¬ dren.
Downloaded From: http://archpedi.jamanetwork.com/ by a University of Michigan User on 06/15/2015
years of
life, the groups were signifi¬ cantly different (P < .05), the group with least difficulties having spent fewer days in the hospital. The relationship between percep¬
tual-motor deficit and level of the myelomeningocele sac was then stud¬ ied. Only one of the ten children with the most deficit had a lesion in the lumbosacral area. All others were in the lumbar or thoracic area. Of the children with the least deficit, six had lesions in the lumbosacral area and four within the lumbar and thoracic areas. The Fisher Exact Probability test'7 showed that there was a signifi¬ cant association between higher le¬ sions and greater perceptual-motor difficulties (P < .05, one-tailed). Another area of interest was to determine whether there was a rela¬ tionship between the level of the myelomeningocele sac, sensory level, and IQ, independent of the effects of hydrocephalus. In order to test for this, we analyzed the MM and MM/HS groups separately. The mean IQ scores for the MM and MM/HS chil¬ dren, listed according to the site of the
myelomeningocele
sac,
are
presented
in Table 7. Excluded from Table 7 were three MM/HS children with a thoracic level sac (IQs: 94, 106, 84) and three MM children with a thoracic level (IQs: 64, 114, 77); because the numbers were too small for analysis. The group with lesions in the occipitocervical areas was excluded from the analysis of variance because the requirement of homogeneity of variance was not met (the standard deviation was too different from the others). The dele¬ tion of this group does not affect the relationship. An analysis of variance was done on the IQ scores of the MM/ HS group of children and was not
significant. For the MM group of children a t was done between the lumbar
test
groups and the lumbosacral and sacral groups and also was not significant. When an analysis of variance was done on the means for both groups together, it was significant at the .01 level, indicating that the IQs of chil¬ dren with myelomeningocele differ significantly according to the level of
the
Table 4—Mean IQ, Chronologie Age, and VMI Scores for Myelomeningocele Children With and Without Hydrocephalus and Their Siblings
method of comparison, groups were formed on the basis of the sensory level. These data are pre¬ sented in Table 8. Analyses of variance were done on the mean IQ scores of the MM/HS and MM groups separately and were not significant. When the means of both groups were analyzed together, a significant difference was found (P < .05), indicating that the IQs of the groups based on sensory level were significantly different from each other.
Mean
SD
Our results, showing that MM/HS children performed at a significantly lower intellectual level than MM chil¬ dren, are consistent with results reported by others16-1"·20 for untreated children with myelomeningocele and hydrocephalus, although our differ¬ ences are of a lesser magnitude. In fact, our data show that 63% of the MM/HS children have IQs above 80, while Merrill et al" reported a 10% frequency and Stephen1* one of 40% for untreated hydrocéphalie MM chil¬ dren. Although it is difficult to compare the intellectual outcome of treated MM/HS and untreated MM/ HS hydrocéphalie children because of the different survival rates, it is true that the presence of hydrocephalus in a child with myelomeningocele de¬ creases his chances for normal intelli¬ gence. Though surgical treatment of the hydrocephalus increases the chances for normal intelligence, my¬ elomeningocele children with hydro¬
cephalus, as a group, are still signifi¬ cantly less intelligent than their normal siblings and also less intelli¬ gent than myelomeningocele children without hydrocephalus.
The percentage of MM/HS children with IQs above 80 (63%) is higher than that reported by Lorber' (52% and 57% for two series), Ames and Schuf
(n
All
29)
=
Siblings (n
=
50)
109.0 15.4
95.0 20.1
101.4
8.0 2.4
9.6 3.5
10.0
5.7 1.8
7.7 2.5
9.2 2.2
20.0
Chronologic age, yr Mean SD
a
COMMENT
MM
59)
=
IQ
sac.
As
MM/HS (n
Scores
VMI, yr Mean SD
Table 6.—Mean Chronologie Age (CA), IQ, and Difference Scores for MM and MM/HS Children*
Table 5.—Mean Chronologie Age (CA), IQ, and Difference Scores for Patients (MM/HS and MM) and Siblings* Patients 9.3 105.5
CA, yr IQ Difference score, yr (CA-VMI)t *N
tf
=
=
14
pairs.
3.73, P
