Cerebral palsy (CP) is an umbrella term that describes impaired muscle control of movement and posture caused by early insults to the developing brain. The incidence is approximately 2.1 per 1000 live births. Normal intrauterine growth is essential for later normal growth and health,[3, 4] and intrauterine growth restriction has repeatedly been identified as an independent risk factor for CP.[5, 6] However, at the other extreme, excess growth resulting in high birthweight has also been shown to be a risk factor for CP.[6, 7] Children with above or below average birthweight who later develop CP will often exhibit the most severe motor and cognitive impairments.[8, 9]
Deviation in fetal growth is difficult to diagnose during pregnancy; thus, weight at birth is used as a proxy for fetal growth, and statistical cut-offs based upon birthweight charts are used to diagnose deviations. Moreover, children with deviating birthweight may be further divided into those with symmetrical and those with asymmetrical body proportions. In suspected fetal growth restriction, different body proportions are considered to reflect the cause: smoking in pregnancy is associated with symmetrical growth restriction and asymmetrical growth restriction could result from placental conditions, defined in this study by biparietal diameter and mean abdominal diameter. Children with low birthweight and asymmetrical body proportions are at increased risk of intrapartum complications and neonatal morbidity, including intraventricular haemorrhages of grades III to IV, compared with children whose birthweight is appropriate for gestational age. A study by Blair and Stanley concluded that, overall, spastic CP in 44% of all sufferers could be attributed to growth deviation at birth.
Despite extensive research, the causes of CP are known in only a minority of cases. In children born very preterm, the typical brain lesion is focal periventricular leukomalacia in deep white matter, leading to the characteristic spastic bilateral CP subtype with lower limb involvement (diplegia). In term-born infants, the most common CP subtype is spastic unilateral CP, considered to be mainly the sequela of antenatal or perinatal stroke. In contrast, global hypoxic–ischaemic injury in children born at term, for instance due to fetopelvic disproportion, is considered to be the main cause of dyskinetic and spastic bilateral subtypes with four-limb involvement (quadriplegia). However, in the majority of cases, CP probably has an antenatal cause, while only about 10% of cases of CP have intrapartum causes.
Previous studies on the relationship between growth deviations and the risk of CP have primarily studied decreased or increased birthweight.[6, 8, 9, 16] We wanted to expand previous knowledge by exploring the association between deviations of other measures of fetal growth and the risk of CP in singletons born at term, and, therefore, we assessed not only birthweight, but also birth length, head circumference, and body proportions. Since in most cases, CP in children born at term is thought to be of antenatal origin, we hypothesized that children with CP would be shorter, would have lower birthweight, and would have a smaller head circumference at birth than children without CP, and that these deviations would be the typical findings in children with unilateral and diplegic CP subtypes. In contrast, we hypothesized that children with dyskinetic and spastic quadriplegic CP subtypes would be larger and more often have asymmetrical body proportions, manifesting as low weight for length.
- Top of page
- Supporting Information
Figures 1, 2a,b and S3–S6 (online supporting information) show the excess risk of CP associated with growth deviation at birth, both for the total CP population and for the different CP subtypes. Table 1 presents the ORs for the different z-score bands underlying the figures. Among all children, those with the highest and lowest birth length and head circumference z-scores were at increased risk of CP, whereas only those with the lowest birthweight had an increased risk of CP compared with children in the reference groups. Increasing ponderal index gradually decreased the risk of CP (Table 2 and Figs 1, 2a,b).
Table 1. Odds ratios of cerebral palsy (CP) and subtypes of CP according to centiles of anthropometric measurements of children at birth in Norway between 1999 and 2006
|Measurement|| n ||Z-score bands (equivalent centiles)|
|<−1.28 (<10%)||−1.28 to <−0.67 (10 to <25%)||−0.67 to <0.67 (25–<75%)||0.67 to <1.28 (75 to <90%)||≥1.28 (≥90%)|
|OR (CI)|| p ||OR (CI)|| p ||OR||OR (CI)|| p ||OR (CI)|| p |
|CP total||398||2.1 (1.6–2.8)||<0.001||0.9 (0.7–1.3)||0.742||Reference||1.0 (0.7–1.3)||0.949||1.1 (0.7–1.5)||0.742|
|Unilateral||184||1.9 (1.3–2.8)||0.002||1.0 (0.6–1.5)||0.847||Reference||0.9 (0.6–1.5)||0.726||0.9 (0.5–1.5)||0.593|
|Bilateral||113||3.4 (2.1–5.5)||<0.001||1.1 (0.6–2.0)||0.735||Reference||1.2 (0.7–2.2)||0.555||1.5 (0.8–2.9)||0.170|
|Diplegic||74||3.1 (1.8–5.5)||<0.001||0.9 (0.4–2.0)||0.937||Reference||0.5 (0.2–1.3)||0.177||1.2 (0.6–2.7)||0.574|
|Quadriplegic||39||4.6 (1.8–11)||0.001||1.7 (0.6–4.9)||0.352||Reference||3.3 (1.4–8.0)||0.007||2.5 (0.9–7.3)||0.095|
|Dyskinetic||28||1.8 (0.6–5.0)||0.266||0.7 (0.2–2.5)||0.596||Reference||1.0 (0.3–2.9)||0.936||0.7 (0.2–3.1)||0.655|
|Ataxic||22||1.2 (0.4–4.4)||0.713||1.1 (0.4–3.4)||0.856||Reference||0.8 (0.2–3.0)||0.781||0.0||0.950|
|CP total||398||1.9 (1.4–2.5)||<0.001||0.8 (0.6–1.1)||0.224||Reference||1.1 (0.8–1.5)||0.612||1.7 (1.3–2.3)||<0.001|
|Unilateral||184||1.4 (0.9–2.2)||0.099||0.8 (0.5–1.2)||0.249||Reference||1.0 (0.6–1.6)||0.967||1.2 (0.8–2.0)||0.374|
|Bilateral||113||3.5 (2.1–5.8)||<0.001||1.3 (0.7–2.4)||0.422||Reference||1.5 (0.8–2.7)||0.240||2.9 (1.7–4.9)||<0.001|
|Diplegic||74||4.3 (2.4–7.6)||<0.001||0.8 (0.3–2.0)||0.826||Reference||1.7 (0.8–3.5)||0.162||2.3 (1.1–4.6)||0.024|
|Quadriplegic||39||2.0 (0.7–5.7)||0.180||2.2 (0.9–5.3)||0.079||Reference||1.0 (0.3–3.5)||0.989||4.1 (1.8–9.4)||0.001|
|Dyskinetic||28||2.2 (0.8–6.2)||0.140||0.6 (1.3–2.7)||0.498||Reference||1.5 (0.5–4.5)||0.514||2.2 (0.8–6.3)||0.134|
|Ataxic||22||1.1 (0.3–3.9)||0.850||0.0||0.940||Reference||0.3 (0.0–2.4)||0.261||1.5 (0.5–4.6)||0.458|
|CP total||398||2.5 (1.9–3.3)||<0.001||1.1 (0.8–1.5)||0.492||Reference||1.2 (0.9–1.7)||0.168||1.6 (1.2–2.2)||0.003|
|Unilateral||184||2.4 (1.6–3.5)||<0.001||0.8 (0.5–1.3)||0.374||Reference||1.1 (0.7–1.7)||0.719||1.3 (0.8–2.1)||0.271|
|Bilateral||113||3.7 (2.3–6.1)||<0.001||1.7 (1.0–3.0)||0.043||Reference||1.3 (0.7–2.4)||0.383||1.9 (1.0–3.5)||0.043|
|Diplegic||74||4.4 (2.4–8.1)||<0.001||1.7 (0.9–3.4)||0.114||Reference||1.8 (0.9–3.6)||0.095||1.3 (0.5–3.2)||0.565|
|Quadriplegic||39||2.5 (1.0–6.3)||0.043||1.8 (0.7–4.2)||0.206||Reference||0.5 (0.1–2.2)||0.354||2.9 (1.2–6.8)||0.018|
|Dyskinetic||28||2.0 (0.6–6.5)||0.229||1.2 (0.4–3.9)||0.731||Reference||1.0 (0.3–3.8)||0.950||3.5 (1.3–9.2)||0.011|
|Ataxic||22||1.5 (0.4–5.6)||0.519||0.9 (0.3–3.3)||0.919||Reference||0.7 (0.2–3.2)||0.639||2.0 (0.6–6.4)||0.243|
Table 2. Odds ratios of cerebral palsy (CP) and subtypes of CP according to centiles of anthropometric measurements of children at birth in Norway between 1999 and 2006
|Measurement|| n ||Z-score bands (equivalent centiles)|
|<−1.28 (<10%)||−1.28 to <−0.67 (10% to <25%)||−0.67 to <0.67 (25–<75%)||0.67 to <1.28 (75 to <90%)||≥1.28 (≥90%)|
|OR (CI)|| p ||OR (CI)|| p ||OR||OR (CI)|| p ||OR (CI)|| p |
|CP total||398||2.1 (1.6–2.7)||<0.001||1.4 (1.0–1.8)||0.026||Reference||0.9 (0.6–1.2)||0.365||0.6 (0.3–0.9)||0.011|
|Unilateral||184||2.0 (1.3–2.9)||0.001||0.9 (0.6–1.4)||0.704||Reference||1.0 (0.6–1.5)||0.966||0.3 (0.1–0.7)||0.005|
|Bilateral||113||2.9 (1.8–4.8)||<0.001||2.3 (1.4–3.7)||0.001||Reference||1.1 (0.6–2.1)||0.672||0.5 (0.3–1.7)||0.475|
|Diplegic||74||3.1 (1.7–5.7)||<0.001||2.2 (1.2–4.0)||0.012||Reference||1.4 (0.7–2.9)||0.337||0.8 (0.3–2.2)||0.626|
|Quadriplegic||39||2.7 (1.1–6.3)||0.025||2.5 (1.1–5.3)||0.024||Reference||0.7 (0.2–2.3)||0.523||0.7 (0.2–2.9)||0.590|
|Dyskinetic||28||2.1 (0.7–5.9)||0.167||2.5 (1.1–5.9)||0.037||Reference||0.0||0.939||0.8 (0.2–3.7)||0.812|
|Ataxic||22||1.8 (0.6–5.7)||0.305||1.2 (0.4–3.8)||0.739||Reference||0.3 (0.0–2.4)||0.254||0.9 (0.2–4.1)||0.902|
Figure 1. Odds ratios (y-axis) of cerebral palsy according to centile bands (x-axis) of anthropometric measurements in 398 children born as singletons at term in Norway between 1999 and 2006.
Download figure to PowerPoint
Figure 2. (a) Odds ratios (y-axis) of spastic unilateral CP according to centile bands (x-axis) of anthropometric measurements in 184 children born as singletons at term in Norway between 1999 and 2006. (b) Odds ratios (y-axis) of spastic bilateral CP according to centile bands (x-axis) of anthropometric measurements in 113 children born as singletons at term in Norway between 1999 and 2006.
Download figure to PowerPoint
Regarding CP subtypes, excess risk of spastic unilateral CP was found among children with the lowest weight (p=0.002), length (p=0.099), and head circumference (p<0.001) z-scores at birth, but not among those with the highest z-scores (Table 1 and Fig. 2a). In contrast, the risk of spastic bilateral CP was also increased for children with the highest z-scores for length and head circumference, as well as for children with low birthweight (Table 1 and Fig. 2b). Within the group of children with spastic bilateral CP, the excess risk of spastic diplegic CP was mainly associated with low z-scores for weight, length, and head circumference, while the risk of quadriplegic CP was also found to be significantly increased among children with the highest z-scores for all three measurements at birth (Table 1 and Fig. S6).
The associations between size at birth and the dyskinetic and the ataxic CP subtypes were not statistically significant (Table 1), but in these groups the total number of children with CP was low. However, when the dyskinetic and the spastic quadriplegic CP subtypes were combined, both low and high z-scores for all three growth parameters, as well as low ponderal index, were associated with an excess risk of CP (data not shown).
- Top of page
- Supporting Information
In this study, we have expanded previous knowledge on the association between low birthweight and excess risk of CP.[6, 8, 16, 20] We have demonstrated that the risk of CP is also increased in children who are short or long, as well as in children with small or large head circumferences at birth. The latter finding was, however, confined to children with the spastic bilateral subtype, while the excess risk of spastic unilateral CP was associated only with being small at birth. Within the group of children with the spastic bilateral subtype, the excess risk of diplegic CP was mainly associated with being small at birth. On the other hand the quadriplegic and dyskinetic subtypes were associated with excessive growth in length and head circumference at birth, but with low weight and ponderal index at birth.
The strengths of this study are the large study population and the prospective recording of data in the MBRN. The limited number of children with some of the CP subtypes (i.e. spastic quadriplegic, dyskinetic, and ataxic CP) resulted in low statistical power in some of the subgroup analyses. Thus, the results of the latter analyses should be interpreted with caution. In all children with CP, the diagnoses were confirmed after the age of 4 years, in accordance with European guidelines, and therefore misclassification is unlikely. We cannot exclude the possibility that a small number of children may have died before they were diagnosed with CP. However, these few children are unlikely to bias the results significantly. Confounding by preterm and/or multiple births was minimized by restricting the analyses to singletons born at term. However, the low ascertainment of cases in the CPRN and missing anthropometric data for gestational age at birth are limitations. During the observation period, 75% of all children with CP in Norway were included in the CPRN. We have previously reported that low ascertainment in the CPRN has mainly been the result of work overload in some centres, and that we found no difference in CP characteristics between counties with low and high ascertainment. Thus, cases not recorded in the CPRN are, in terms of severity and CP subtypes, most likely missing at random, and the study population is likely to be representative of the total Norwegian CP population. The anthropometric measurements were performed by midwives as part of the routine examination shortly after birth and were not further standardized. This may affect the measurements of head circumference and length at birth.
We have been able to identify only two studies with an approach similar to the present study: one study in Europe by Jarvis et al. and one in Australia by Blair and Stanley. The Australian study included children with spastic CP only, and the number of participants was lower than in our study. However, Blair and Stanley addressed the effects of various combinations of small head circumference, length at birth, and birthweight on the overall risk of CP, but they did not study CP subtypes. The excess risk of CP in the lowest birthweight group in our study is consistent with the results for 105 children born at term in the Australian study. Jarvis et al. included a total of 4503 children with CP, and reported an inverse J-shaped association between birthweight and CP for children born at term when they used customized growth charts. However, when they used conventional growth standards they found that excess risk of CP was associated only with low birthweight and not with high birthweight. Our results are consistent with this finding, although it should be emphasized that the increased risk of CP at high birthweight in that study was observed at birthweight z-scores above 1.88, while in our study children with z-scores above 1.28 constituted the highest birthweight group. In contrast to the study by Jarvis et al., we were able to adjust for parity in the calculation of birthweight z-scores, and we could use the same anthropometric data obtained from a control population born in the same period. Jarvis et al. also studied the association between birthweight and the main CP subtypes, and reported the same inverse J-shaped curve for all CP subtypes. In contrast, we found an inverse J-shaped curve for children with bilateral spastic CP, consistent with the findings of Jarvis et al., but we could not observe similar increased risk with high weight for the other CP subtypes, and in particular not for unilateral spastic CP.
Whereas Jarvis et al. did not report data on head circumference and length at birth in the European study, the association between deviations in these measurements and CP was a main objective of the Australian study. Blair and Stanley reported that the combination of being small for gestational age and being short and thin was associated with the highest OR for CP compared with children with typical growth. In children of a typical size for their gestational age, Blair and Stanley found, consistent with our study, that both small and large head circumference were associated with excess risk of CP. However, in contrast to our findings, they were not able to show an excess risk of CP associated with deviations in birth length in term-born children. Neither of the two aforementioned studies reported on the association between CP subtypes and deviations in head size and length at birth. The present study has therefore extended previous knowledge in the field by showing that spastic unilateral CP is associated mainly with reduced growth in all three measurements, and not with being large for gestational age at birth, while the spastic bilateral CP subtype is associated with both reduced and increased antenatal growth in length and head circumference as well as with reduced weight.
The finding that the excess risks for the spastic unilateral and spastic bilateral CP subtypes with lower limb involvement (diplegia) were mainly associated with growth restriction at birth is consistent with the hypothesis of an antenatal cause underlying these CP subtypes. Spastic diplegia is also the most common subtype seen in children with CP born at very low gestational age (i.e. before week 32 of pregnancy), and imaging studies have suggested that many children born at term with this CP subtype have white matter lesions similar to those seen in preterm children. Moreover, these findings are consistent with our previous study suggesting an antenatal cause in the majority of children with CP born small for gestational age at term.
In accordance with SCPE guidelines, CP in children is classified as either spastic or dyskinetic. However, some children exhibit mixed symptoms and differentiating between the two subtypes can be difficult. Some countries consider all children with mixed symptoms to have spastic CP, while in other countries the diagnosis depends on the dominant symptom, which is what the current SCPE guidelines recommend. Mixed symptoms may suggest a similar aetiology in many of the children with spastic and dyskinetic CP. Our study supports this as we found greater similarity of growth patterns between the spastic quadriplegic and dyskinetic subtypes than between the spastic quadriplegic and diplegic subtypes. Spastic quadriplegic and dyskinetic CP subtypes were associated with high length and large head circumference at birth, but with low birthweight. These two subtypes are considered to be the characteristic CP subtypes following adverse hypoxic–ischaemic intrapartum events and neonatal encephalopathy. Our results are consistent with this notion, since a complicated birth may be more likely in larger rather than in smaller infants. Interestingly, the combined risk of the dyskinetic and the spastic quadriplegic CP subtypes was significantly associated with large head circumference and length at birth, but with low weight and ponderal index. Since a low ponderal index may reflect reduced fat and glycogen storages, our results are also consistent with the suggestion that an intrapartum event may be more harmful to the brain in children with limited nutritional reserves at birth.
We found that children of low birthweight and those with low or high length or small or large head circumferences at birth were at increased risk of CP. Whereas reduced size measurements were mainly associated with the spastic unilateral and diplegic CP subtypes, large size at birth was particularly associated with the spastic quadriplegic and the dyskinetic CP subtypes. The findings are consistent with the notion that the most frequent causes of CP are adverse antenatal events, followed by restricted fetal growth, while large infants may be more vulnerable to adverse intrapartum events.