Volume 162 Number I
21. Benesch R, Benesch RE. Intracellular organic phosphates as regulators of oxygen release by haemoglobin. Nature 1969;221:618-22. 22. Madsen H, Ditzel j. Correlation of serum unconjugated oestriol to red cell 2,3-diphosphoglycerate levels in diabetic pregnancy. Diabeto1ogia 1983;24: 152-4.
Fetal growth related to maternal oxygen transport
23. Daikoku NH, Johnson JWC, Graf C, Kearney K, Tyson jE. King TM. Patterns of intrauterine growth retardation. Obstet Gynecol 1979;54:211-9. 24. Moore LG. Rounds SS,jahnigen D, Grover R, ReevesjT. Infant birth weight is related to maternal arterial oxygenation at high altitude. J App1 Physiol 1982;52:695-9.
Intrauterine growth and spastic cerebral palsy I. Association with birth weight for gestational age Eve Blair, PhD, and Fiona Stanley, MD, MSc, MFCM
Nedlands, Western Australia Birth weight, gestational age at delivery, and other factors were collected for 171 white children with spastic cerebral palsy. Their birth weights were compared with the birth weight distribution expected for a population of the same race, gestation, sex, maternal height, and parity, born in the same geographic area, and during the same time period. Birth weights of children with spastic cerebral palsy tended to be significantly lower than the median birth weight of their comparison population. Analysis stratified by gestation at delivery suggested that if the reduced birth weight were causally associated with the spastic cerebral palsy, 22% of cases were attributable to being below the 10th percentile of the comparison population birth weight distribution. The risk of spastic cerebral palsy associated with poor intrauterine growth was dependent on gestation at delivery; poorly grown infants delivered between 34 and 37 weeks' gestation were at highest risk. Some probable pathways by which growth retardation could result in brain damage (intrapartum hypoxia, hypoglycemia, and hypothermia) were investigated. Only intrapartum hypoxia may have played a causal role but probably accounted for 37 34-37 < 34
< 10th
10th to 90th
,1 95 % confidencej PopulatuJl/ 1950/. Confidence 1 P opulatIOn Odds ratio lIIterval aU/lbutab/e mk ( % ) Odds mtlO mterval attn butable risk (% )
6.48 5.25 19.56 1.98
4.2-11 2.11-9.8 8.1-47 0 .3-9.8
14.1 (8.6-22 ) 11.3 (5.0-21) 35.8 (17.6-58) 2.86 (0-21)
and percentile birth weight. as is spastic cerebral palsy. Therefore both gestation and percentile birth weight must be controlled for in the estimation of risk of spastic cerebral palsy associated with these perinatal factors. This was achieved by analysis of all matched casecontrol sets by conditional logistic regression," with gestation and percentile birth weight as main effects in the regression equation as well as the factor of interest. The coefficient of the interaction term between percentile birth weight and the factor of interest was then estimated. This coefficient is the natural logarithm of the additional risk of cerebral palsy incurred by SGA infants over and above that of AGA infants on account of hypothermia (hypoglycemia or hypoxia) when these conditions are present. For example, if a part of the risk of spastic cerebral palsy associated with being SGA were caused by hypothermia, the interaction coefficient would be positive, indicating that the risk associated with being SGA would increase if the subject were also hypothermic.
3.8 1 3.27 10. 1 1.48
2.7-5.5 2.0-5.2 4.9-21 0.5-4.3
Odds ratio
2 1.9 (14 .5-31) 18.5 (9.4-30) 47.7 (30-68) 4.58 (0-25)
Because interaction terms seldom reach statistical significance in small data sets, particularly when controlling for multiple confounders, alternative analyses were also done. The gestational age distribution of cases of < 10th percentile birth weight was similar to that of all controls with < 10th percentile birth weight, as shown in the Appendix. Therefore the odds ratios and population-attributable risks associated with hypothermia and hypoglycemia could be estimated in this SGA group with little confounding by gestation. These estimates of odds ratio were compared with those obtained by matched analysis for the 10th to 90th percentile birth weight group. A different approach was possible for intrapartum hypoxia. Distribution of percentile birth weight in cases likely to have been caused by intrapartum hypoxia. The distribution of percentile birth weights of the 14 white cases whose disability we attributed to intrapartum hypoxia" was compared with that of other cases and with the population distribution by estimating odds
Intrauterine growth and spastic cerebral palsy 233
Volume 162 Number 1
Table III. Odds ratios for spastic cerebral palsy by percentile birth weight relative to 10th to 90th percentile birth weight statified by sex and plurality P ercentzle birth weight category 90th Odds ratIO
0.67 0.59 0.80 0.69
1.7-4.2 1.6-5.0 1.0-5.1 1.6-4.1 0.7-36
Confidence I 95%mteroal
0.3-1.4 0.2-1.5 0.2-2.3 0.3-1.4 No cases
Table IV. Numbers of cases and controls below 3rd percentile birth weight with odds ratios for spastic cerebral palsy associated with antecedents of poor intrauterine growth, by gestational age GestatIOnal age at delIVery 34-37 Completed wk
No.*
Total Antecedent Urinary tract infection Other infections Twinning Maternal age >34 years Antepartum hemorrhage Preeclampsia Miscarriage Oligohydramnios Placental! cord anomaly Poor placental condition:j: Placental separation Congenital anomaly TORCH infection
I
Odds ratIO
I
95% Confidence interoal
p
11146 1112 113 2/2 0/5 2110 5/22 3113 211 011 5/8 0/5 0/7 0/2
No.*
I
1
>37 Completed wk
Odds ratIO
95 % Confidence interoal
p
0.1-7 0.03-41 0-6 1.2-127 0.08-5 0.2-5 0.3-8 0.1-00 0-10 0.3-12.5 0-6 0-21 0.1-00
1.0 0.5 0.6 0.014t 1.0 1.0 0.5 0.2 1.0 0.4 0.6 1.0 0.2
13/50 0.29 1.4 4.7 0 0.80 0.91 0.95 9.4 0 3.8 0 0 0
0.01-2.4 0.02-20 0.3-73 0-4.7 0.07-4.9 0.2-4.2 0.1-4.8 0.5-602 0-163 0.7-20 0-4.7 0-2.9 0-23
0.4 1.0 0.2 0.6 1.0 1.0 1.0 0.09 1.0 0.1 0.6 0.3 1.0
217 112 0/4 4/2 2/9 4114 4110 110 0/3 3/6 0/4 0/2 110
I.I 2.0 0 10.1 0.8 I.I 1.8 00 0 2.2 0 0 00
TORCH, Toxoplasmosis. Rubella, Cytomegalovirus. Herpes. or other virus infection reported in the index pregnancy. *Number of cases/number of controls. tSignificant at p < 0.05 level. :j:Placenta described as macerated. necrotic. or infarcted.
ratios with maximum likelihood analysis 9 and calculating attributable risks. Results
Of the 171 cases 63.7% were male, and 4.7% were twins. The motor handicap was purely spastic in 88.3%, with hemiplegia in 37.4%, diplegia in 38.3%. and quadriplegia in 14.6%. The remaining cases were spastic in addition to being athetoid (5.3%). dystonic (4.7%), or primarily ataxic (1.8%). The severity was classified as mild in 48.0% of cases, moderate in 30.4%, severe in 17.5%. and unknown in 4.1 %. Sixty-eight cases (39.8%) were born before 37 completed weeks of gestation. including 7.0% born before 30 weeks; 39.8% had birth weights
21. Benesch R, Benesch RE. Intracellular organic phosphates as regulators of oxygen release by haemoglobin. Nature 1969;221:618-22. 22. Madsen H, Ditzel j. Correlation of serum unconjugated oestriol to red cell 2,3-diphosphoglycerate levels in diabetic pregnancy. Diabeto1ogia 1983;24: 152-4.
Fetal growth related to maternal oxygen transport
23. Daikoku NH, Johnson JWC, Graf C, Kearney K, Tyson jE. King TM. Patterns of intrauterine growth retardation. Obstet Gynecol 1979;54:211-9. 24. Moore LG. Rounds SS,jahnigen D, Grover R, ReevesjT. Infant birth weight is related to maternal arterial oxygenation at high altitude. J App1 Physiol 1982;52:695-9.
Intrauterine growth and spastic cerebral palsy I. Association with birth weight for gestational age Eve Blair, PhD, and Fiona Stanley, MD, MSc, MFCM
Nedlands, Western Australia Birth weight, gestational age at delivery, and other factors were collected for 171 white children with spastic cerebral palsy. Their birth weights were compared with the birth weight distribution expected for a population of the same race, gestation, sex, maternal height, and parity, born in the same geographic area, and during the same time period. Birth weights of children with spastic cerebral palsy tended to be significantly lower than the median birth weight of their comparison population. Analysis stratified by gestation at delivery suggested that if the reduced birth weight were causally associated with the spastic cerebral palsy, 22% of cases were attributable to being below the 10th percentile of the comparison population birth weight distribution. The risk of spastic cerebral palsy associated with poor intrauterine growth was dependent on gestation at delivery; poorly grown infants delivered between 34 and 37 weeks' gestation were at highest risk. Some probable pathways by which growth retardation could result in brain damage (intrapartum hypoxia, hypoglycemia, and hypothermia) were investigated. Only intrapartum hypoxia may have played a causal role but probably accounted for 37 34-37 < 34
< 10th
10th to 90th
,1 95 % confidencej PopulatuJl/ 1950/. Confidence 1 P opulatIOn Odds ratio lIIterval aU/lbutab/e mk ( % ) Odds mtlO mterval attn butable risk (% )
6.48 5.25 19.56 1.98
4.2-11 2.11-9.8 8.1-47 0 .3-9.8
14.1 (8.6-22 ) 11.3 (5.0-21) 35.8 (17.6-58) 2.86 (0-21)
and percentile birth weight. as is spastic cerebral palsy. Therefore both gestation and percentile birth weight must be controlled for in the estimation of risk of spastic cerebral palsy associated with these perinatal factors. This was achieved by analysis of all matched casecontrol sets by conditional logistic regression," with gestation and percentile birth weight as main effects in the regression equation as well as the factor of interest. The coefficient of the interaction term between percentile birth weight and the factor of interest was then estimated. This coefficient is the natural logarithm of the additional risk of cerebral palsy incurred by SGA infants over and above that of AGA infants on account of hypothermia (hypoglycemia or hypoxia) when these conditions are present. For example, if a part of the risk of spastic cerebral palsy associated with being SGA were caused by hypothermia, the interaction coefficient would be positive, indicating that the risk associated with being SGA would increase if the subject were also hypothermic.
3.8 1 3.27 10. 1 1.48
2.7-5.5 2.0-5.2 4.9-21 0.5-4.3
Odds ratio
2 1.9 (14 .5-31) 18.5 (9.4-30) 47.7 (30-68) 4.58 (0-25)
Because interaction terms seldom reach statistical significance in small data sets, particularly when controlling for multiple confounders, alternative analyses were also done. The gestational age distribution of cases of < 10th percentile birth weight was similar to that of all controls with < 10th percentile birth weight, as shown in the Appendix. Therefore the odds ratios and population-attributable risks associated with hypothermia and hypoglycemia could be estimated in this SGA group with little confounding by gestation. These estimates of odds ratio were compared with those obtained by matched analysis for the 10th to 90th percentile birth weight group. A different approach was possible for intrapartum hypoxia. Distribution of percentile birth weight in cases likely to have been caused by intrapartum hypoxia. The distribution of percentile birth weights of the 14 white cases whose disability we attributed to intrapartum hypoxia" was compared with that of other cases and with the population distribution by estimating odds
Intrauterine growth and spastic cerebral palsy 233
Volume 162 Number 1
Table III. Odds ratios for spastic cerebral palsy by percentile birth weight relative to 10th to 90th percentile birth weight statified by sex and plurality P ercentzle birth weight category 90th Odds ratIO
0.67 0.59 0.80 0.69
1.7-4.2 1.6-5.0 1.0-5.1 1.6-4.1 0.7-36
Confidence I 95%mteroal
0.3-1.4 0.2-1.5 0.2-2.3 0.3-1.4 No cases
Table IV. Numbers of cases and controls below 3rd percentile birth weight with odds ratios for spastic cerebral palsy associated with antecedents of poor intrauterine growth, by gestational age GestatIOnal age at delIVery 34-37 Completed wk
No.*
Total Antecedent Urinary tract infection Other infections Twinning Maternal age >34 years Antepartum hemorrhage Preeclampsia Miscarriage Oligohydramnios Placental! cord anomaly Poor placental condition:j: Placental separation Congenital anomaly TORCH infection
I
Odds ratIO
I
95% Confidence interoal
p
11146 1112 113 2/2 0/5 2110 5/22 3113 211 011 5/8 0/5 0/7 0/2
No.*
I
1
>37 Completed wk
Odds ratIO
95 % Confidence interoal
p
0.1-7 0.03-41 0-6 1.2-127 0.08-5 0.2-5 0.3-8 0.1-00 0-10 0.3-12.5 0-6 0-21 0.1-00
1.0 0.5 0.6 0.014t 1.0 1.0 0.5 0.2 1.0 0.4 0.6 1.0 0.2
13/50 0.29 1.4 4.7 0 0.80 0.91 0.95 9.4 0 3.8 0 0 0
0.01-2.4 0.02-20 0.3-73 0-4.7 0.07-4.9 0.2-4.2 0.1-4.8 0.5-602 0-163 0.7-20 0-4.7 0-2.9 0-23
0.4 1.0 0.2 0.6 1.0 1.0 1.0 0.09 1.0 0.1 0.6 0.3 1.0
217 112 0/4 4/2 2/9 4114 4110 110 0/3 3/6 0/4 0/2 110
I.I 2.0 0 10.1 0.8 I.I 1.8 00 0 2.2 0 0 00
TORCH, Toxoplasmosis. Rubella, Cytomegalovirus. Herpes. or other virus infection reported in the index pregnancy. *Number of cases/number of controls. tSignificant at p < 0.05 level. :j:Placenta described as macerated. necrotic. or infarcted.
ratios with maximum likelihood analysis 9 and calculating attributable risks. Results
Of the 171 cases 63.7% were male, and 4.7% were twins. The motor handicap was purely spastic in 88.3%, with hemiplegia in 37.4%, diplegia in 38.3%. and quadriplegia in 14.6%. The remaining cases were spastic in addition to being athetoid (5.3%). dystonic (4.7%), or primarily ataxic (1.8%). The severity was classified as mild in 48.0% of cases, moderate in 30.4%, severe in 17.5%. and unknown in 4.1 %. Sixty-eight cases (39.8%) were born before 37 completed weeks of gestation. including 7.0% born before 30 weeks; 39.8% had birth weights
