Acta Pmdiatr 81:407-12.1992
Intrauterine growth in children with cerebral palsy P Uvebrant and G Hagberg Deportment of Puediurrics II, University of Goteborg. Sweden
Uvebrant P, Hagberg G. Intrauterine growth in children with cerebral palsy. Acta Pzediatr 1992;81: 407-12. Stockholm. ISSN 0803-5253 The risk of cerebral palsy in connection with intrauterine growth retardation has been analysed in a case-control study. The case series comprised 519 children with cerebral palsy born in 1967-1982 in the west health-care region of Sweden and the control series 445 children born during the same years in the same region. The risk of cerebral palsy in small-for-gestational-age infants was significantly increased in term and moderately preterm infants. The highest proportion among infants with cerebral palsy born at term was found in tetraplegia, followed by diplegia and dyskinetic cerebral palsy. It was concluded that small for gestational age on the one hand reflects early prenatal brain damage, and on the other mediates prenatal risk factors compatible with foetal deprivation of supply and also potentiates adverse effects of birth asphyxia and neonatal hypoxia. 0 Cerebral palsy, intrauterine growth retardation, small for gestational age P Uvebrant, Department of Paediatrics II, East Hospital, University of Gothenberg, S-41685 Gothenburg, Sweden
The role of multifactorial pre- and perinatal compromising events in cerebral palsy (CP) has attracted increased attention during recent years (1-4). The debate is that intrauterine growth retardation (IUGR) and antenatal events are more important than perinatal events such as birth asphyxia and neonatal hypoxia or that IUGR and antenatal events make an infant more vulnerable to perinatal risk factors. This study estimates, retrospectively, the gestational age-specific risk of CP in children with low birth weight (BW) for gestational age (GA). It also analyses the impact of prenatal events known to be associated with IUGR, as well as perinatal events suspected of increasing the risk in the presence of IUGR.
Definitions Birth at term was defined as that occurring after completion of the 37th week of gestation ( 2259 days) and preterm birth that occurring before the 37th week ( < 259 days). Small for gestational age (SGA) was defined as BW < -2 SD, and near-SGA as < - 1 SD from the mean BW for GA on a Swedish growth chart (5). Severe wasting was defined as BW < -2 SD and moderate wasting as < - 1 SD from the mean BW for birth length (BL). Preeclamptic signs were defined as blood pressure 2 140 or 2 90 mmHg and proteinuria, or 2 160 or 2 110 mmHg, both on two occasions. Foetal depriuation of supply (FDS) was considered in the case of preeclamptic signs, uterine bleeding, placental infarction or multiple pregnancy. Birth asphyxia was considered in the case of placental ablation or an Apgar
score < 6 at 1 min or later, and neonatal hypoxia in the case of respiratory disturbances during the first week of life, ranging from cyanotic spells to respiratory treatment. Cerebral palsy was defined as a non-progressive disorder of movement and posture due to a defect or lesion of the immature brain ( 6 ) .
Material and methods Computation of the risk of CP in connection with intrauterine growth was based on an unselected, population-based series of 519 children with CP (1-3) and a series of445 controls (7,8) born between 1967 and 1982 in the west health-care region of Sweden. From the total of 584 children in the CP series, 14 with a GA of less than 28 weeks, five with missing pre- or perinatal information and 46 with a postnatal (after first week of life) cause of CP were excluded. The risk of CP by wasting was restricted to the birth year period 1975-82 as we had not recorded BLs earlier. The control series had been selected earlier to match the subjects in a study of hemiplegic C P (7) and another of infantile hydrocephalus (8). The method of paired sampling was used, so that one control was sampled for each term and two controls for each preterm hemiplegic and hydrocephalic case. Term controls were chosen by selecting the term, liveborn infant born next after a term case in the same delivery unit and of the same sex. Preterm controls were selected from the birth registers of three neonatal wards, chosen as representative of the quality of the neonatal care in the study area, and matched for GA, sex, birth date and type of delivery
408
P Uuehrani und G Hugberg
ACTA PEDIATR 8 I ( I 992)
unit. In order to overcome difficulties in matching for all four variables in infants born before 32 weeks of gestation, GA was given first priority followed by birth date, sex and level of neonatal care. The final number of controls comprised 269 term and 176 preterm infants. The GA and the sex distribution of controls matched well that of CP cases in this study. Thus, the mean GA of term CP cases was 39.9 (median 40) weeks and that of term controls 39.9 (median 40) weeks, and that of preterm CP cases 32.4 weeks (week 28-33, mean 30.9, median 3 I ; week 34-36, mean 35.2, median 35) and preterm controls 33.2 weeks (28-33, mean 31.2, median 3 1; week 34-36, mean 35.0, median 35). The male excess ratio was 1.3: 1 in term C P cases and in term controls 1.4: 1, and in preterm CP cases and preterm controls 1.5 : 1 and 1.3: i, respectively. Controlsborn moderately preterm were not considered a representative group. Normally in Sweden, newborns with a BW of approximately 2500 g are cared for on the obstetric ward together with their mothers, and not in neonatal units, whereas our sampling procedure had selected preterm control infants from neonatal units, resulting in an overrepresentation of low BWs. Information on growth parameters and pre- and perinatal compromising events were in all cases and controls taken retrospectively from obstetric and paediatric records. Observer bias was not considered as only objective information that had been given by persons unaware of this study was noted in a structured protocol with the same items checked in both cases and controls. In the statistical analysis, conventional chi-square tests with Yate’s correction were used for comparison between proportions, the Wald test W2 was used for odds ratio (9) and Fisher’s exact test for four-fold tables whenever appropriate. The odds ratio for disease was used as a measure of the association between compromising factors and risk of CP, and calculated as the ratio between the factor not present and present in controls divided by the same ratio in cases (9). An odds ratio of 1 expresses an equal risk, above 1 an increased risk and less than 1 a decreased risk of CP. Adjusted odds ratio were calculated with the Mantel-Haenszel’s
method (9). The crude and adjusted risks of C P by FDS, SGA and birth asphyxia/neonatal hypoxia (A/H) was restricted to GAS of 2 34 weeks in order to avoid confounding by very preterm birth.
Results Of the 519 children with CP, 344 were born at term, 59 were moderately preterm (GA 34-36 weeks), and 116 were very preterm (GA 28-33 weeks). The overall risk of CPin infants SGA, near-SGA and large for GA (BW for GA > 2 SD) was significantly increased (Table I). The gestational age-specific risk was significantly increased in term SGA and near-SGA infants, and in moderately preterm SGA infants. The risk was. increased, but not significantly for SGA infants born very preterm. Thirty-eight (7.3%) of the 519 infants were twins, compared to 18 of 445 controls (4.0%). The difference (3.3%) was significant (t =2.17, ~ 6 0 . 0 5 ,95% confidence interval (CI) 6.2-0.4). Relatively more twins were found among cases than among controls at all GAS (2.9%vsl.I%, 11.9%vs4.4%and 1 8 . 1 % 12.8%for ~~ children born at term, moderately and very preterm, respectively). CP twins were more often SGA than twin controls (12/38 (32%) and 2/18 (1 1 X),respectively; the difference was not significant, t = 1.65, 0.10 < p < 0.05). Exclusion of twin births from the analysis did not change the risk of C P in SGA infants more than marginally, the odds ratio being 7.4, 4.3 and 2.5 in the term, moderately preterm and very preterm groups, respectively. The overall risk of CP by severe and moderate wasting was significantly increased (Table 2). The gestational age-specific risk of C P in infants with severe wasting and moderate wasting was significantly increased in term infants. The distribution of SGA and near-SGA in the different CP syndromes compared with controls is given in Table 3. In term CP children, the proportion of SGA and near-SGA was significantly increased in the majority’bf syndromes. Exceptions were SGA in term ataxia and near-SGA in term tetraplegia. The highest propor-
Table 1. Gestational age-specific risk of CP by birth weight for gestational age (BW/GA) in 519 children with CP and 445 controls (Co). Computation of odds ratio (OR) contrasts cases versus controls relative to infants with normal BW/GA.
Gestational age (weeks) 28-33
34-36
co CP OR Co CP (n=86) (n=116) (n =90) (n= 59)
OR
Co
> - 2 S D t o -ISDto +ISDto, +2SD
3 I1 61 8 3
I1 I1 74 13 7
*P-ISDto + ISD to < +2SD > +2SD
34-36
co (n=46)
CP (n=61)
OR
2 11 28 4
6 10 38 6 I
2.2 0.7 I 1.1 0.7
1
Co (n =40)
37+
CP (n= 34)
OR
2
3.8 2.3 I
I 9 30 0 0
I1 16 4 I
Co CP (n= 126) (n=152) 0 5 92 24 5
8 36 90 16 2
Total OR
co CP (n=212) (n=247)
*
-
7.4*** I 0.7 0.4
16 57
3 25 150 28 6
144
26 4
OR 5.6** 2.4** 1 1.O 0.7
*PG0.05, **p
Intrauterine growth in children with cerebral palsy P Uvebrant and G Hagberg Deportment of Puediurrics II, University of Goteborg. Sweden
Uvebrant P, Hagberg G. Intrauterine growth in children with cerebral palsy. Acta Pzediatr 1992;81: 407-12. Stockholm. ISSN 0803-5253 The risk of cerebral palsy in connection with intrauterine growth retardation has been analysed in a case-control study. The case series comprised 519 children with cerebral palsy born in 1967-1982 in the west health-care region of Sweden and the control series 445 children born during the same years in the same region. The risk of cerebral palsy in small-for-gestational-age infants was significantly increased in term and moderately preterm infants. The highest proportion among infants with cerebral palsy born at term was found in tetraplegia, followed by diplegia and dyskinetic cerebral palsy. It was concluded that small for gestational age on the one hand reflects early prenatal brain damage, and on the other mediates prenatal risk factors compatible with foetal deprivation of supply and also potentiates adverse effects of birth asphyxia and neonatal hypoxia. 0 Cerebral palsy, intrauterine growth retardation, small for gestational age P Uvebrant, Department of Paediatrics II, East Hospital, University of Gothenberg, S-41685 Gothenburg, Sweden
The role of multifactorial pre- and perinatal compromising events in cerebral palsy (CP) has attracted increased attention during recent years (1-4). The debate is that intrauterine growth retardation (IUGR) and antenatal events are more important than perinatal events such as birth asphyxia and neonatal hypoxia or that IUGR and antenatal events make an infant more vulnerable to perinatal risk factors. This study estimates, retrospectively, the gestational age-specific risk of CP in children with low birth weight (BW) for gestational age (GA). It also analyses the impact of prenatal events known to be associated with IUGR, as well as perinatal events suspected of increasing the risk in the presence of IUGR.
Definitions Birth at term was defined as that occurring after completion of the 37th week of gestation ( 2259 days) and preterm birth that occurring before the 37th week ( < 259 days). Small for gestational age (SGA) was defined as BW < -2 SD, and near-SGA as < - 1 SD from the mean BW for GA on a Swedish growth chart (5). Severe wasting was defined as BW < -2 SD and moderate wasting as < - 1 SD from the mean BW for birth length (BL). Preeclamptic signs were defined as blood pressure 2 140 or 2 90 mmHg and proteinuria, or 2 160 or 2 110 mmHg, both on two occasions. Foetal depriuation of supply (FDS) was considered in the case of preeclamptic signs, uterine bleeding, placental infarction or multiple pregnancy. Birth asphyxia was considered in the case of placental ablation or an Apgar
score < 6 at 1 min or later, and neonatal hypoxia in the case of respiratory disturbances during the first week of life, ranging from cyanotic spells to respiratory treatment. Cerebral palsy was defined as a non-progressive disorder of movement and posture due to a defect or lesion of the immature brain ( 6 ) .
Material and methods Computation of the risk of CP in connection with intrauterine growth was based on an unselected, population-based series of 519 children with CP (1-3) and a series of445 controls (7,8) born between 1967 and 1982 in the west health-care region of Sweden. From the total of 584 children in the CP series, 14 with a GA of less than 28 weeks, five with missing pre- or perinatal information and 46 with a postnatal (after first week of life) cause of CP were excluded. The risk of CP by wasting was restricted to the birth year period 1975-82 as we had not recorded BLs earlier. The control series had been selected earlier to match the subjects in a study of hemiplegic C P (7) and another of infantile hydrocephalus (8). The method of paired sampling was used, so that one control was sampled for each term and two controls for each preterm hemiplegic and hydrocephalic case. Term controls were chosen by selecting the term, liveborn infant born next after a term case in the same delivery unit and of the same sex. Preterm controls were selected from the birth registers of three neonatal wards, chosen as representative of the quality of the neonatal care in the study area, and matched for GA, sex, birth date and type of delivery
408
P Uuehrani und G Hugberg
ACTA PEDIATR 8 I ( I 992)
unit. In order to overcome difficulties in matching for all four variables in infants born before 32 weeks of gestation, GA was given first priority followed by birth date, sex and level of neonatal care. The final number of controls comprised 269 term and 176 preterm infants. The GA and the sex distribution of controls matched well that of CP cases in this study. Thus, the mean GA of term CP cases was 39.9 (median 40) weeks and that of term controls 39.9 (median 40) weeks, and that of preterm CP cases 32.4 weeks (week 28-33, mean 30.9, median 3 I ; week 34-36, mean 35.2, median 35) and preterm controls 33.2 weeks (28-33, mean 31.2, median 3 1; week 34-36, mean 35.0, median 35). The male excess ratio was 1.3: 1 in term C P cases and in term controls 1.4: 1, and in preterm CP cases and preterm controls 1.5 : 1 and 1.3: i, respectively. Controlsborn moderately preterm were not considered a representative group. Normally in Sweden, newborns with a BW of approximately 2500 g are cared for on the obstetric ward together with their mothers, and not in neonatal units, whereas our sampling procedure had selected preterm control infants from neonatal units, resulting in an overrepresentation of low BWs. Information on growth parameters and pre- and perinatal compromising events were in all cases and controls taken retrospectively from obstetric and paediatric records. Observer bias was not considered as only objective information that had been given by persons unaware of this study was noted in a structured protocol with the same items checked in both cases and controls. In the statistical analysis, conventional chi-square tests with Yate’s correction were used for comparison between proportions, the Wald test W2 was used for odds ratio (9) and Fisher’s exact test for four-fold tables whenever appropriate. The odds ratio for disease was used as a measure of the association between compromising factors and risk of CP, and calculated as the ratio between the factor not present and present in controls divided by the same ratio in cases (9). An odds ratio of 1 expresses an equal risk, above 1 an increased risk and less than 1 a decreased risk of CP. Adjusted odds ratio were calculated with the Mantel-Haenszel’s
method (9). The crude and adjusted risks of C P by FDS, SGA and birth asphyxia/neonatal hypoxia (A/H) was restricted to GAS of 2 34 weeks in order to avoid confounding by very preterm birth.
Results Of the 519 children with CP, 344 were born at term, 59 were moderately preterm (GA 34-36 weeks), and 116 were very preterm (GA 28-33 weeks). The overall risk of CPin infants SGA, near-SGA and large for GA (BW for GA > 2 SD) was significantly increased (Table I). The gestational age-specific risk was significantly increased in term SGA and near-SGA infants, and in moderately preterm SGA infants. The risk was. increased, but not significantly for SGA infants born very preterm. Thirty-eight (7.3%) of the 519 infants were twins, compared to 18 of 445 controls (4.0%). The difference (3.3%) was significant (t =2.17, ~ 6 0 . 0 5 ,95% confidence interval (CI) 6.2-0.4). Relatively more twins were found among cases than among controls at all GAS (2.9%vsl.I%, 11.9%vs4.4%and 1 8 . 1 % 12.8%for ~~ children born at term, moderately and very preterm, respectively). CP twins were more often SGA than twin controls (12/38 (32%) and 2/18 (1 1 X),respectively; the difference was not significant, t = 1.65, 0.10 < p < 0.05). Exclusion of twin births from the analysis did not change the risk of C P in SGA infants more than marginally, the odds ratio being 7.4, 4.3 and 2.5 in the term, moderately preterm and very preterm groups, respectively. The overall risk of CP by severe and moderate wasting was significantly increased (Table 2). The gestational age-specific risk of C P in infants with severe wasting and moderate wasting was significantly increased in term infants. The distribution of SGA and near-SGA in the different CP syndromes compared with controls is given in Table 3. In term CP children, the proportion of SGA and near-SGA was significantly increased in the majority’bf syndromes. Exceptions were SGA in term ataxia and near-SGA in term tetraplegia. The highest propor-
Table 1. Gestational age-specific risk of CP by birth weight for gestational age (BW/GA) in 519 children with CP and 445 controls (Co). Computation of odds ratio (OR) contrasts cases versus controls relative to infants with normal BW/GA.
Gestational age (weeks) 28-33
34-36
co CP OR Co CP (n=86) (n=116) (n =90) (n= 59)
OR
Co
> - 2 S D t o -ISDto +ISDto, +2SD
3 I1 61 8 3
I1 I1 74 13 7
*P-ISDto + ISD to < +2SD > +2SD
34-36
co (n=46)
CP (n=61)
OR
2 11 28 4
6 10 38 6 I
2.2 0.7 I 1.1 0.7
1
Co (n =40)
37+
CP (n= 34)
OR
2
3.8 2.3 I
I 9 30 0 0
I1 16 4 I
Co CP (n= 126) (n=152) 0 5 92 24 5
8 36 90 16 2
Total OR
co CP (n=212) (n=247)
*
-
7.4*** I 0.7 0.4
16 57
3 25 150 28 6
144
26 4
OR 5.6** 2.4** 1 1.O 0.7
*PG0.05, **p
