Materials and Methods
The population-based case-control study providing the data for this report was designed to investigate factors associated with risk of CP, defined as a disorder of movement and/or posture and motor function because of a nonprogressive interference/lesion or abnormality of the developing brain. Persons registered with the Western Australian Register of Developmental Anomalies–Cerebral Palsy were selected if born between Jan. 1, 1980, and Dec. 31, 1995, excluding those whose CP was acquired postneonatally or resulted in minimal motor impairment.
Controls and perinatal deaths were selected from the 380,918 births between 1980 and 1995 registered on the Maternal Child Health Research Database, which links statutory birth and death registries with statutory pregnancy and delivery information and includes more than 99.5% of registered births in Western Australia. Neonatally surviving controls not registered as CP were individually matched to CP cases for gestational age (within 1 week), date of birth (within 12 months), and plurality. Representative samples of intrapartum stillbirths and neonatal deaths (live births dying within 28 days) were selected by taking 7 and 4 year birth cohorts, respectively, distributed within the 1980-1995 period.
Information was sought concerning pregnancy, delivery, neonatal, family, and maternal obstetric history from medical records of the hospitals of delivery and neonatal care and private obstetricians or general practitioners who provided antenatal care. These data were recorded by trained midwives or neonatal nurses. Analyses of FGR included singleton births and focused on children with CP and controls born after at least 35 weeks’ gestation.
Appropriateness of fetal growth was assessed with the proportion of optimal birthweight (POBW). This method compares the index birthweight with the median estimated weight of fetuses of the same gestational age subsequently live born at term to women without recorded exposure to common factors known to affect fetal growth and of the same height and parity as the index mother.
Fetal growth was categorized as mild if POBW was less than 85% (about the 10th percentile of all Western Australian live births) and marked if POBW was 77.3% or less (2 SD below the median of the optimally grown population, approximately the fifth percentile of all Western Australian live births) or, to minimize false-negative results, had been diagnosed neonatally as growth restricted.
PIH was defined as blood pressure reaching or exceeding 140/90 mm Hg, a systolic rise of 20 mm Hg, or a diastolic rise of 15 mm Hg during pregnancy. As found with other adverse perinatal outcomes, the presence of proteinuria did not affect the risk of CP associated with PIH and was not considered further.
Women were classified as smokers in pregnancy if they were reported to smoke after 20 weeks’ gestation; this datum was missing for about one-third of study subjects. The number of cigarettes smoked daily before and after 20 weeks’ gestation was also recorded if available (83% of smokers). Alcohol and recreational drug use was noted if recorded, but gross underreporting is likely.
Data concerning defects present at birth and recognized at any time before 6 years of age were obtained by linking with the Western Australian Register of Developmental Anomalies–Birth Defects, which records up to 10 diagnoses of birth defects together with the age at which each diagnosis was made. For this analysis birth defects were categorized hierarchically as cerebral defects; cardiac defects; other major defects; and deformational defects and minor defects (which included cosmetic defects and those unlikely to affect physical quality of life). Major defects excluded exclusively minor defects. Because CP is defined by motor disorders not present at birth, it was not considered a birth defect. Recorded teratogenic exposures and chromosomal or genetic defects were classified separately.
Sentinel events were defined as intrapartum events likely to enhance the potential for fetal asphyxia and included significant intrapartum hemorrhage, cord prolapse, uterine rupture, a tight nuchal cord, shoulder dystocia, and maternal collapse. A clinical diagnosis of birth asphyxia or hypoxic ischemic encephalopathy in the medical record was noted.
The risks of CP were estimated for each growth, PIH, and maternal smoking stratum for singleton births after at least 35 weeks of gestation. Associations were sought between CP within high-risk strata for major birth defects, sentinel events, a clinical diagnosis of birth asphyxia, and recreational drug use.
Statistical analysis
SAS version 9.3 (SAS Institute, Cary, NC) was used to generate descriptive statistics and odds ratios estimated from a conditional logistic regression, which effectively adjusts for the matching variables of gestational duration (within 1 week) and year of birth (within 12 months).
This study was approved by the Princess Margaret Hospital/King Edward Memorial Hospital Ethics Committee, the Confidentiality of Health Information Committee of the Western Australia Department of Health, individual hospital and regional ethics committees, the University of Sydney Human Research and Ethics Committee, and the Confidentiality of Health Information Committee of the Western Australia Department of Health, whose documents can be provided on request. All waived the requirement for individual patient consent.
Results
Among the 386,159 infants born in Western Australia from 1980 through 1995, there were 376,541 singleton births (97.5%) of whom 363,747 (96.6%) delivered at or after 35 weeks’ gestation. There were 493 singleton children with CP, 508 matched controls, and 176 perinatal deaths born at 35 weeks’ gestational age or later and 167 singleton children with CP, 148 matched controls, and 319 perinatal deaths born before 35 weeks for whom appropriateness of growth could be assessed.
One quarter of controls (25.9%), 22.9% of CP, and 14.9% of perinatal deaths who were classified as markedly growth restricted were so classified on the basis of a neonatal diagnosis of growth restriction and did not have a birthweight more than 2 SD below their estimated optimal. Five of the 19 CP thus categorized infants were diagnosed as congenitally hydrocephalic. For the others, weight was either low relative to length, particularly in CP, or, more rarely, the length was low relative to the head circumference. With the exception of 2 CP infants with marked congenital hydrocephaly, all had birthweights below their estimated optimal.
Association of FGR with risk of CP
In term and late preterm infants, the odds of CP for those with marked FGR were increased almost 9-fold compared with infants of appropriate birthweight for gestational age (AGA) (odds ratio [OR], 4.16; 95% confidence interval [CI], 2.5–6.8). For mild FGR there was no significant increase in CP risk ( Table 1 ).
| Variable | Factor | Nco:Nca | OR (95% CI) |
|---|---|---|---|
| Total, n a | Fetal growth | 508:493 | |
| AGA | 399:325 | 1 (reference) | |
| Mild FGR | 82:85 | 1.21 (0.85–1.73) | |
| Marked FGR | 27:83 | 4.16 (2.5–6.8) b | |
| Total, n a | Maternal hypertension | 503:477 | |
| AGA | N | 348:256 | 1 (reference) |
| Mild FGR | N | 71:71 | 1.24 (0.84–1.8) |
| Marked FGR | N | 22:69 | 4.81 (2.7–8.5) b |
| AGA | Y | 47:61 | 1.56 (1.0–2.4) b |
| Mild FGR | Y | 10:11 | 1.61 (0.65–4.0) |
| Marked FGR | Y | 5:9 | 2.56 (0.78–8.4) |
| Total, n a | Maternal smoking | 332:331 | |
| AGA | N | 195:161 | 1 (reference) |
| Mild FGR | N | 31:25 | 1.35 (0.63–2.9) |
| Marked FGR | N | 13:33 | 7.82 (2.5–24) b |
| AGA | Y | 67:58 | 1.37 (0.81–2.34) |
| Mild FGR | Y | 22:26 | 1.11 (0.49–2.5) |
| Marked FGR | Y | 4:28 | 14.8 (3.3–66) b |
| Total, n a | Major birth defect | 508:493 | |
| AGA/mild FGR | N | 465:266 | 1 (reference) |
| Marked FGR | N | 25:39 | 2.77 (1.5–5.1) b |
| AGA/mild FGR | Y | 16:144 | 13.9 (7.5–25.7) b |
| Marked FGR | Y | 2:44 | 30.9 (7.0–136) b |
a Number of subjects with nonmissing data for all stratification variables
b OR differs statistically significantly from unity ( P < .05).
Normotensive and hypertensive FGR
Among markedly FGR singletons, about half of controls and children with CP born before 35 weeks, but a much larger proportion of controls, 81% (22 of 27), CP, 88% (69 of 78), and 71% (34 of 48) perinatal deaths born at or after 35 weeks, experienced normotensive pregnancies ( Figure 1 ). Marked FGR in normotensive gravidae was associated with statistically significantly increased risk of CP, whereas that in hypertensive gravidae was not ( Table 1 ). In AGA neonates born of hypertensive pregnancies, there was a relatively small elevation of CP risk of marginal statistical significance.
Thus, the large majority of term and late preterm infants with marked FGR who developed CP were products of normotensive pregnancies.
Maternal smoking
Maternal smoking data were missing for 34.6% of controls, 33.3% of CP, and 26.3% of perinatal deaths. In those with available data, there was no increase in CP risk associated with maternal smoking that was not accompanied by marked FGR ( Table 1 ).
To test the hypothesis that the marked growth restriction in CP births to smoking mothers resulted from greater exposure to cigarette smoke, we compared the quantities smoked between those with marked and those with mild or no growth restriction. The mothers of CP births and perinatal deaths with marked growth restriction did not smoke more than those with mild or no growth restriction: mean (SE) of the number of cigarettes smoked daily after 20 weeks’ gestation was 12.2 (1.2) in those with a marked restriction and 13.3 (1.0) in those without. This suggests that the excess of markedly FGR infants born to smoking mothers of CP births and perinatal deaths may be attributable to some characteristic of the woman or her environment that is associated with smoking in pregnancy rather than with smoking itself.
Substance abuse and social disadvantage are possible candidates for a factor associated with smoking in pregnancy that raises risk for adverse outcome. In those with nonmissing data, as anticipated, alcohol use was associated with smoking: in controls (OR, 3.5; 95% CI, 1.8–6.9), CP (OR, 2.0; 95% CI, 1.0–3.9), and perinatal deaths (OR, 3.4; 95% CI, 1.2–9.9). Fetal alcohol syndrome was identified in 4 CP cases (and no controls), 2 of whom were known to be born to smoking mothers but none of whom met criteria for either mild or marked FGR.
Recreational drug use (primarily heroin, cocaine, amphetamines, and marijuana) was prospectively reported for 8 of 293 controls (2.7%), 16 of 300 CP (5.3%), and 6 of 122 perinatal deaths (4.9%) and was also associated with smoking but more strongly in CP (OR, 17.9; 95% CI 4–80) than in controls (OR, 3; 95% CI, 0.7–12) or perinatal deaths (OR, 2.5; 95% CI, 0.5–13). Five of seven CP but only 1 of 6 perinatal deaths born to women who smoked and were reported to use recreational drugs were markedly FGR.
Of the total of 28 CP, 4 controls and 10 perinatal deaths who were both markedly FGR and born to mothers known to smoke, major birth defects were observed in 18 of the 28 CP but no control or perinatal death. Excluding 2 CP infants with congenital TORCH infection (toxoplasmosis, other [hepatitis B], rubella [German measles], cytomegalovirus, and herpes simplex virus) (and no reported recreational drug use), the distribution of the type of defects in the 4 for whom recreational drug use was reported was similar to the 12 for whom it was not, with half of each group sustaining a cerebral defect.
FGR and birth asphyxia
For infants born of normotensive pregnancies, the proportion of CP associated with a clinical diagnosis of birth asphyxia was not higher in markedly FGR infants (15.6%) than in AGA neonates (22.9%). Similarly, potentially asphyxiating birth events occurred less rather than more frequently during markedly FGR deliveries than AGA deliveries (3.0% in FGR, 10.2% in AGA; Table 2 ). Thus, asphyxial birth as identified by either of these criteria did not appear to be the factor linking growth restriction in normotensive pregnancy to CP.
| Variable | Diagnosis of birth asphyxia | Sentinel event | Major birth defects | ||||
|---|---|---|---|---|---|---|---|
| Fetal growth | PIH | Controls | CP | Controls | CP | Controls | CP |
| AGA | N | 0 (0/348) | 22.9 a (57/249) | 4.0 (14/348) | 10.2 (26/254) | 2.9 (10/348) | 34.8 a (89/256) |
| Mild FGR | N | 0 (0/71) | 20.6 a (14/68) | 5.6 (4/71) | 11.3 (8/71) | 5.6 (4/71) | 42.2 a (30/71) |
| Marked FGR | N | 0 (0/22) | 15.6 (10/64) | 9.1 (2/22) | 3.0 (2/67) | 9.1 (2/22) | 55.1 a (38/69) |
| AGA | Y | 2.1 (1/47) | 31.0 a (18/58) | 6.4 (3/47) | 8.2 (5/61) | 2.1 (1/47) | 31.2 a (19/61) |
| Mild FGR | Y | 0 (0/9) | 27 a (3/11) | 10 (1/10) | 9.1 (1/11) | 0 (0/10) | 36.4 a (4/11) |
| Marked FGR | Y | 20 (1/5) | 44.4 a (4/9) | 0 (0/5) | 0 (0/9) | 0 (0/5) | 55.6 a (5/9) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
