NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
National Guideline Alliance (UK). Cerebral palsy in under 25s: assessment and management. London: National Institute for Health and Care Excellence (NICE); 2017 Jan. (NICE Guideline, No. 62.)
Review question: What are the most important risk factors for developing cerebral palsy with a view to informing more frequent assessment and early recognition?
4.1. Introduction
Risk factors are events or circumstances that increase the risk of brain injury or malformation that then results in cerebral palsy. A risk factor does not always mean that the child will develop cerebral palsy. It means that the chances are higher than if the risk factor was not present. The absence of risk factors does not ensure that the child will not develop cerebral palsy. While many features can have associated risk of developing cerebral palsy it is important not to create unnecessary anxiety or increased surveillance for children who may develop typically.
Knowing the risk factors may help in preventing or effectively treating and managing risks.
The early identification and diagnosis of cerebral palsy is important for many reasons, not only to guide intervention but also to advise on prognosis and family planning.
The guideline has investigated the most important risk factors to target surveillance for those at risk of developing cerebral palsy. It focused on 3 specific timings of when the injury or dysfunction can occur in the developing brain: during the antenatal period (before birth); around the time of birth (perinatal factors); and after birth, most commonly within the first year of life (postnatal factors). The list of potential risk factors can be very large, including prematurity, infection and trauma, so it was important to identify the most significant ones.
The aim of this evidence review was to identify the most important risk factors for developing cerebral palsy with the view to providing information for parents and/or carers and to inform the need for more frequent assessment and early intervention.
The Committee prioritised the risk factors that were most commonly seen in clinical practice as the view was that it was neither practical nor useful to assess all possible risk factors. Only papers published after the year 2000 were included in the review to account for the changes in clinical practice and interventions available after this time.
Those prioritised were:
- Antenatal factors
- infections (for example, rubella, toxoplasmosis, cytomegalovirus [CMV], herpes simplex)
- multiple pregnancy
- intrauterine growth restriction
- haemorrhagic events
- Perinatal
- hypoxic-ischaemic events at term/post-term
- neonatal encephalopathy
- Apgar score at 10 min (low/very low below 4/3)
- neonatal sepsis
- Postnatal
- extremely preterm – 24 to 27 +6 weeks gestational age
- preterm – 28 to 31 +6 weeks gestational age
- late preterm – (32 to 37 weeks gestational age)
- infections: meningitis and encephalitis
- clotting disorders/hypercoagulation in mother
- trauma/non-accidental injury.
Individual systematic reviews were undertaken for each of these and the results are reported below, grouped by antenatal, perinatal and postnatal factors.
4.2. Description of clinical evidence: antenatal risk factors
Nine observational studies have been identified for this review (Bear & Wu 2016, Beaino 2010, Dammann 2001, Himpens 2010, Laptook 2005, Livinec 2005, Miller 2013, Streja 2013, Wu 2013). Four were retrospective cohorts using national registries as data sources (Bear & Wu 2016, Miller 2013, Streja 2013, Wu 2013). Five studies were prospective cohorts, of which two were based on the EPIPAGE cohort (Beaino 2010, Livinec 2005), and included babies born between 22 and 32 weeks of gestational age; 1 study (Himpens 2010) included children assessed at 1 centre for developmental disorders and referred from NICU; 1 study (Laptook 2005) was multicentre, including 14 different centres participating in the same network and it looked at very low birthweight babies; and 1 study included long-term survivors of a regional cohort of very low birthweight newborns (Dammann 2001).
Sample sizes ranged from 407 to 6,018,504 children.
Four studies reported on maternal infections as a risk factor for cerebral palsy: 1 study (Streja 2013) reported adjusted odds ratios for all infections, vaginal infections and urinary infections; for vaginal infections, it also presented the data separately for at-term and preterm babies. One study (Wu 2013) reported adjusted odds ratios for infections of the genitourinary system and for any other infections. One study (Miller 2013) reported adjusted estimates for any hospital reported maternal infection separately for preterm and at-term babies; and 1 study (Bear & Wu 2016) presented adjusted odds ratios for genitourinary infections other than chorioamnionitis, and respiratory infections.
Three studies reported on multiple pregnancies as a risk factor for cerebral palsy (Beaino 2010, Himpens 2010, Laptook 2005).
One study reported results on haemorrhagic events as an antenatal risk factor for developing cerebral palsy (Livinec 2005).
One study reported on fetal growth retardation as a risk factor for developing cerebral palsy (Dammann 2001).
Outcomes are reported as described in the original papers, so reflect the variation in reporting. Only studies presenting adjusted analyses have been considered for this review.
Studies were heterogeneous with regards to population and subgroups considered, risk factors studied and covariates included in the multivariate models. For these reasons, it was decided not to pool the data together. Therefore, forest plots presented in Appendix I do not report meta-analysed data but they have been produced to help the readers to visualise the direction of the effect sizes.
For this review, quality appraisal of the evidence has been conducted using the NICE manual methodology checklists. Quality appraisal has been conducted by study, and not by outcome. For full details see section 4.9.4 on quality of evidence.
The quality of each study was assessed using the NICE manual methodology checklists. Please see section 4.9.4 on quality of the evidence for more details.
For full details see the review protocol in Appendix D. See also the study selection flow chart in Appendix F, forest plots in Appendix I, study evidence tables in Appendix J and the exclusion list in Appendix K.
4.2.1. Summary of included studies and results
A summary of the studies included in this review and their results for antenatal factors are presented in Table 11.
4.3. Evidence statements
4.3.1. Maternal infections
High-quality evidence from 1 study with 6,018,504 participants (mother-infant dyads) reported an increased risk of cerebral palsy in children whose mothers had a hospital discharge diagnosis of genitourinary infection other than chorioamnionitis (OR=1.4) and whose mothers had a hospital discharge diagnosis of respiratory infection (OR=1.9).
Moderate-quality evidence from 1 study with 588,936 singletons showed an increased risk for cerebral palsy in children whose mothers had infections of the genitourinary system during pregnancy, but not for those whose mothers had ‘any other infections’ during pregnancy.
Low-quality evidence from 1 study with 81,066 singletons showed an increased risk for cerebral palsy in children whose mothers had vaginal infections during pregnancy; when looking at the risk of developing spastic cerebral palsy, an association was found for babies born at term but not for preterm babies. The same study reported no association between ‘all infections’, urinary infections and cerebral palsy.
Very low-quality evidence from 1 study with 440,564 singletons showed no association between maternal infections and cerebral palsy in both preterm and at-term babies.
4.3.2. Multiple pregnancy
High-quality evidence from 1 study with 1,473 very low birthweight babies showed an increased risk of cerebral palsy in babies born from multiple pregnancy. However, high-quality evidence from another study with 984 high-risk babies showed no association between multiple pregnancy and cerebral palsy.
Moderate-quality evidence from 1 study with 2,357 preterm babies showed no association between multiple pregnancy and development of cerebral palsy.
4.3.3. Haemorrhagic events
High-quality evidence from 1 study with 2,382 preterm babies showed no association between the occurrence of maternal haemorrhagic events and the development of cerebral palsy.
4.3.4. Intrauterine growth retardation
Moderate-quality evidence from 1 study with 324 very low birthweight babies showed that being small for gestational age was associated with a reduced risk of developing bilateral spastic cerebral palsy. However, the same study did not find the same association when looking at subsamples of babies born at 24 to 31 weeks, 28 to 31 weeks, and when using an age-matched sample.
4.4. Description of clinical evidence: perinatal risk factors
Fifteen studies have been identified for this review (Ahlin 2013, Alshaikh 2013, 2014, Bear & Wu 2016, Han 2012, Himpens 2010, Laptook 2005, Mitha 2013, Nasef 2013, Natarajan 2013, Pappas 2014, Shatrov 2010, Soraisham 2013, Sukhov 2012, Wang 2014). One study was a meta-analysis of 17 observational studies including very low birthweight infants (Alshaikh 2013). One study was a meta-analysis of 15 observational studies (Shatrov 2010). Six studies were prospective cohorts including children referred from the neonatal intensive care unit (NICU) from 1 centre for developmental disorders in Belgium (Himpens 2010), preterm survivors from 1 centre in Korea (Han 2012), very low birthweight babies from 14 centres (Laptook 2005), very low birthweight and preterm babies from 18 tertiary referral centres in Taiwan (Wang 2014), children of 22 to 32 weeks of gestational age from the EPIPAGE study (Mitha 2013), and 1 study (Pappas 2014) included preterm babies from 16 centres. Five studies were retrospective cohorts that used 3 different state databases (Sukhov 2012): a neonatal database of a single centre (Alshaikh 2014), hospital charts (Bear & Wu 2016, Nasef 2013), and children from 1 regional NICU (Soraisham 2013). One study was a secondary analysis of RCT data (Natarajan 2013) including children who had hypoxicischaemic events. One study (Ahlin 2013) used a case-control design using data from a national registry in Sweden.
Sample sizes ranged from n=174 to 6.1 million children.
Three studies reported on hypoxic-ischaemic events or birth asphyxia as a risk factor for developing cerebral palsy (Han 2012, Himpens 2010, Sukhov 2012).
One study reported on neonatal encephalopathy as a risk factor indicating cerebral palsy (Ahlin 2013).
One study reported on Apgar score at 10 minutes as a risk factor for cerebral palsy (Natarajan 2013).
Six studies reported on neonatal sepsis as a risk factor for developing cerebral palsy (Alshaikh 2013, 2014, Han 2012, Laptook 2005, Mitha 2013, Wang 2014).
Five studies reported specifically on chorioamnionitis as a risk factor for developing cerebral palsy (Bear & Wu 2016, Nasef 2013, Pappas 2014, Shatrov 2010, Soraisham 2013). This risk factor has not been specified in the protocol, but it has been recognised as in important perinatal feature to be reviewed.
The quality of each study was assessed using the NICE manual methodology checklists. Please see section 4.9.4 on quality of the evidence for more details.
For full details see review protocol in Appendix D. See also the study selection flow chart in Appendix F, forest plots in Appendix I, study evidence tables in Appendix J and exclusion list in Appendix K.
4.4.1. Summary of included studies
A summary of the studies that were included in this review and their results for perinatal factors are presented in Table 12.
4.5. Evidence statements
4.5.1. Hypoxic-ischaemic events or birth asphyxia
High-quality evidence from 1 study with 984 high-risk babies showed an increased risk of cerebral palsy in children who experienced birth asphyxia (collected and defined using medical records); the same study showed an increased risk of developing non-spastic cerebral palsy compared to spastic cerebral palsy in these children.
High-quality evidence from 1 study with 437 preterm babies showed no association between hypoxic-ischaemic event (defined as a 10-min Apgar score <6 and combined hypoxia identified by means of a blood test) and development of cerebral palsy.
Low-quality evidence from 1 study with 6.1 million children showed an increased risk of developing cerebral palsy in children who experienced mild to severe birth asphyxia (collected and defined using ICD classification).
4.5.2. Neonatal encephalopathy
Low-quality evidence from 1 study with 927 children showed an increased risk of both spastic and dyskinetic cerebral palsy in children with neonatal encephalopathy.
4.5.3. Apgar score at 10 min
Moderate-quality evidence from 1 study with 174 children who had hypoxic-ischaemic event showed a decreased risk of developing cerebral palsy for each point increase in Apgar score at 10 min.
4.5.4. Neonatal sepsis
High-quality evidence from 1 meta-analysis of 11 studies showed an increased risk of developing cerebral palsy in children with neonatal sepsis.
High- to moderate-quality evidence from 5 studies with 10,704 high-risk children showed no association between neonatal sepsis and cerebral palsy.
4.5.5. Chorioamnionitis
High-quality evidence from 1 meta-analysis with 15 observational studies showed an increased risk of cerebral palsy in children born after pregnancy with clinical evidence of chorioamnionitis; the same study also found increased risk of cerebral palsy in children who showed histological chorioamnionitis.
High-quality evidence from 1 study with 6,018,504 participants (mother-infant dyads) reported an increased risk of cerebral palsy in children whose mothers had a hospital discharge diagnosis of chorioamnionitis (OR=3.1). Moderate-quality evidence from 1 study with 384 preterm babies showed an increased risk of cerebral palsy in children born after pregnancy with histological evidence of chorioamnionitis; however, moderate-quality evidence from another study with 2,390 preterm babies showed no association between histological chorioamnionitis and cerebral palsy.
Low-quality evidence from 1 study with 2,390 preterm babies showed no association for both histological and clinical chorioamnionitis and development of cerebral palsy.
4.6. Description of clinical evidence: postnatal risk factors
Six studies have been identified for this review (Beaino 2010, Bonellie 2005, Himpens 2010, Petrini 2009, Stoll 2004, Sukhov 2012). Two studies were prospective cohorts, of which 1 was based on the EPIPAGE cohort (Beaino 2010) and included babies born between 22 and 32 weeks of gestational age; 1 study (Himpens 2010) included children assessed at 1 centre for developmental disorders and referred from NICU. Four studies used a retrospective design: 1 used 3 different state databases (Sukhov 2012), 1 used hospitalisation and outpatient databases from the Northern California Kaiser Permanente Medical Care Program (Petrini 2009), 1 used a registry of very low birthweight infants maintained by the National Institute of Child health and Human Development Neonatal Research Network (Stoll 2004), and 1 study used a national database (Bonellie 2005).
Sample sizes ranged from n=646 to 6.1 million children.
Five studies reported on gestational age as a risk factor for cerebral palsy (Beaino 2010, Bonellie 2005, Himpens 2010, Petrini 2009, Sukhov 2012).
One study reported on neonatal infections as a possible risk factor for cerebral palsy (Stoll 2004).
No evidence was retrieved for trauma or non-accidental injuries, or clotting disorders.
The quality of each study was assessed using the NICE manual methodology checklists. Please section 4.9.4 on quality of the evidence for more details.
For full details see the review protocol in Appendix D. See also the study selection flow chart in Appendix F, forest plots in Appendix I, study evidence tables in Appendix J and the exclusion list in Appendix K.
4.6.1. Summary of included studies
A summary of the studies that were included in this review and their results for postnatal factors are presented in Table 13.
4.7. Evidence statements
4.7.1. Gestational age
High-quality evidence from 1 study with 984 high-risk babies showed an association between longer gestational age and type of cerebral palsy; children with higher gestational age were at increased risk of developing non-spastic cerebral palsy versus spastic cerebral palsy, as well as of developing unilateral cerebral palsy versus bilateral.
Moderate-quality evidence from 1 study with 2,357 preterm babies showed no association between gestational age and cerebral palsy.
Low-quality evidence from 1 study with 646 children showed an increased risk of cerebral palsy for extreme preterm (24 to 27 weeks), preterm (28 to 31 weeks), and late preterm babies (32 to 36 weeks) compared to babies born 37+ weeks in both singletons and twins. Another study with low-quality evidence with 6.1 million children showed an increased risk of cerebral palsy for extreme preterm (<28 weeks), and preterm (28 to 31 weeks) babies compared to babies born 37+ weeks; however, no association was found for late preterm babies (32 to 36 weeks) compared to babies born 37+ weeks.
Low-quality evidence from 1 study with 141,321 children showed an increased risk of developing cerebral palsy in children with gestational age at birth of 30 to 33 weeks and 34 to 36 weeks compared to children born 37 to 41 weeks; no association was found between children born at 42+ weeks and those born at 37 to 41 weeks.
4.7.2. Neonatal infections (meningitis and encephalitis)
Moderate-quality evidence from 1 study with 7,892 babies found no association between meningitis with or without sepsis and development of cerebral palsy.
4.7.3. Trauma/non-accidental injuries
No evidence was retrieved for this risk factor.
4.7.4. Clotting disorders
No evidence was retrieved for this risk factor.
4.8. Economic evidence
This review question is not relevant for economic analysis because it does not involve a decision between alternative courses of action.
No economic evaluations on identifying the most important risk factors for cerebral palsy were identified in the literature search conducted for this guideline. Full details of the search and economic article selection flow chart can be found in Appendix E and Appendix F, respectively.
4.9. Evidence to recommendations
4.9.1. Relative value placed on the outcomes considered
The aim of this review was to identify the most important risk factors for developing cerebral palsy with the view to providing information for parents and/or carers and to inform the need for more frequent assessment and early intervention. The Committee prioritised the following risk factors, based on those commonly perceived to be implicated and expert opinion:
- Antenatal factors
- infections (for example, rubella, toxoplasmosis, CMV, herpes simplex)
- multiple pregnancy
- intrauterine growth retardation
- haemorrhagic events.
- Perinatal
- hypoxic-ischaemic events at term/post-term
- neonatal encephalopathy
- Apgar score at 10 min (low/very low below 4/3)
- neonatal sepsis.
- Postnatal
- extreme prematurity 24 to 27 weeks (+6 days) weeks gestational age)
- premature babies 28 to 31 weeks (+6 days) weeks gestational age
- late premature babies (32 to 37 weeks gestational age)
- infections: meningitis and encephalitis
- clotting disorders/hypercoagulation in mother
- trauma/non-accidental injury.
4.9.2. Consideration of clinical benefits and harms
The Committee recognised that cerebral palsy is aetiologically a multifactorial condition and in any affected person, a number of other clinical and socioeconomic risk factors may have contributed to the outcome. Thus, children born preterm may be at risk because of prematurity but may also have a risk arising from infection. Most of the studies analysed the magnitude of independent risk factors by using adjusted analyses. As part of the protocol for this evidence review, the Committee agreed that they wanted to understand the evidence for independent risk factors for cerebral palsy.
The Committee considered that studies in this area published before 2000 should not be included in the review because of changes in antenatal and neonatal clinical practice since then may have had a significant impact on relative risk factors for cerebral palsy.
The Committee agreed that low birthweight was frequently a proxy for preterm birth in the literature. The Committee noted that very low birthweight infants were the population considered for some of the risk factors such as multiple pregnancies and neonatal sepsis. Therefore, the Committee decided to add low birthweight to the recommendation as a risk factor itself given that it was frequently reported in the populations included in the studies.
4.9.2.1. Antenatal risk factors
Maternal infections
The evidence showed that effect sizes reached significance when vaginal or genitourinary infections were analysed separately from all the other infections during pregnancy. Evidence was also provided of increased risk for developing cerebral palsy associated with genitourinary and respiratory tract infections in the mother that was recognised in a hospital setting. Specific evidence was retrieved in other studies, indicating a direct association with chorioamnionitis, and the Committee agreed that it should be listed as an independent risk factor for cerebral palsy.
Multiple pregnancy
The evidence showed conflicting results with regards to multiple pregnancies acting as a risk factor for cerebral palsy, with 1 study showing an increased risk, a second study showing reduced risk, and a third showing no significant risk. The Committee noted that 1 study looked at a population of low birthweight infants and another at population of preterm infants and these found different results in relation to multiple pregnancies. This would not have been expected as low birthweight is a proxy for prematurity.
The Committee agreed that infants born in multiple pregnancies are more likely to be preterm and have low birthweight. The studies included in the review adjusted for that.
The Committee agreed that the evidence did not support including multiple pregnancies as an independent risk factor for the development of cerebral palsy.
Intrauterine growth restriction (IUGR)
Only 1 study met the inclusion criteria for this review. This study suggested that being small for gestational age was associated with a reduced risk of developing bilateral spastic cerebral palsy. The value of this study was limited by the fact that it was carried out in a population of very low birthweight babies. The Committee discussed how not all small for gestational age infants will be growth restricted. The category of very low birthweight babies includes infants of varying gestational ages, some of whom will be appropriate weight for gestation, some of whom will be more developed but small for gestation. The Committee agreed that this made the study findings more difficult to interpret and therefore agreed not to develop a specific recommendation that IUGR be considered an independent risk factor in cerebral palsy.
The Committee were aware of 3 other studies on intrauterine growth restriction that were not included in the evidence review as they did not meet the inclusion criteria specified in the protocol. Main reasons for exclusion included date of publication (before 2000) and lack of comparative data (all children had cerebral palsy). A study by Jarvis 2003 concluded that preterm babies either below the 10th percentile or above the 97th percentile were more likely to have cerebral palsy than those in a reference band between the 25th and 75th percentile; however, they did not adjust for IUGR as an independent risk factor. Two other studies (Uvebrant 1988 and Blair & Stanley 1990) showed that the risk of cerebral palsy was associated with poor intrauterine growth and dependent on gestation at delivery; however, these studies were done prior the 2000 cut off specified in the review protocol, and the Committee considered that changes in neonatal care made the findings less appropriate in modern practice.
Maternal haemorrhagic events
One study showed increased and reduced risk of cerebral palsy in twins and singletons, respectively; however, both estimates were not statistically significant. Based on the reviewed evidence, the Committee agreed that haemorrhagic events should not be considered as an independent risk factor for the development of cerebral palsy.
4.9.2.2. Perinatal risk factors
Hypoxic-ischaemic events (HIE)
Two of the 3 included studies included for HIE showed an increased risk of developing cerebral palsy for babies who had a hypoxic-ischaemic event. However, it was not clear from the studies what was to be considered as an HIE, making the evidence difficult to interpret. In addition, the studies differed in how they measured and reported HIE; for example, Han 2002 measured HIE based on a low Apgar score at 10 minutes and combining this with means of blood tests, while Himpens 2010 used medical records to collect data on this risk factor.
Neonatal encephalopathy
Only 1 study was included that assessed neonatal encephalopathy as a risk factor. Although this evidence was of low quality, the Committee was persuaded of its importance, as it showed a very large effect. They therefore recommended that neonatal encephalopathy be recognised as an independent risk factor for developing cerebral palsy.
Apgar score
One study showed that increasing Apgar score at 10 min was associated with a reduced risk of developing cerebral palsy. However, it did not identify the risks associated with particular Apgar scores at 10 min (as it was indicated in the review protocol), and so the Committee was not able to recommend a specific Apgar score as a risk factor.
Neonatal sepsis
Five studies showed an association (albeit non-significant) in terms of an increased risk between history of neonatal sepsis and cerebral palsy, all carried out in populations of preterm infants. In addition, a meta-analysis did show a significant association between a history of neonatal sepsis and an increased risk of cerebral palsy, again in populations of preterm infants.
The Committee noted that neonatal sepsis occurred more frequently in preterm infants as reflected in these studies, and there was a lack of evidence in relation to term infants. Despite this lack of published evidence, the Committee believed that neonatal sepsis was an independent risk factor for cerebral palsy in neonates generally, and so they recommended that it be recognised as such.
Chorioamnionitis
One high-quality meta-analysis and one high-quality cohort with large sample size showed an increased risk of cerebral palsy in babies born with a history of chorioamnionitis. The Committee were in agreement that chorioamnionitis should be recognised as an independent risk factor for the development for cerebral palsy.
4.9.2.3. Postnatal risk factors
Gestational age
Five studies were presented that examined the association between gestational age and risk of developing cerebral palsy. The Committee agreed that the evidence suggested an increased risk of cerebral palsy with reducing length of gestation. This was particularly high when considering a gestational age at birth of less than 28 weeks and was also increased in those born between 28 and 32 weeks gestation
The Committee pointed out that, although not shown in the retrieved evidence, it was their view that preterm delivery increased the risk of different forms of cerebral palsy differently. The Committee agreed that in high-risk infants delivered closer to full term the resultant motor subtype of cerebral palsy was more likely to be dystonic rather than spastic in nature, and unilateral rather than bilateral in distribution. Conversely, in early preterm cohorts the motor pattern was more likely to be spastic and bilateral.
The Committee noted the guidance provided in the NICE guideline on preterm labour and birth. Management to prevent preterm birth in at-risk labour (section 1.8), administration of maternal corticosteroids to mature fetal lung (section 1.9) and the use of magnesium sulfate as a neuroprotective mechanism (section 1.10) were all discussed.
Neonatal infection
One study showed a small increased risk for the development of cerebral palsy in very low birthweight babies who had suffered from meningitis. The Committee recognised the lack of evidence in relation to higher birthweight infants, but believed that clinical experience showed meningitis to be a serious risk factor. Again, the lack of evidence for the latter group reflected the fact that infection is more common in very preterm infants. Given the lack of evidence, the Committee decided not to make a specific recommendation for neonatal infection as a risk factor.
Traumanon-accidental injuries
The Committee was made aware that a few papers evaluated the association between neonatal seizures and adverse neurological outcomes, including the development of cerebral palsy. However, the Committee were in agreement that this information was more relevant as part of the ‘causes of cerebral palsy’ review. Given the lack of evidence on other trauma or non-accidental injuries, the Committee decided not to make a specific recommendation for these as risk factors.
Clotting disorders
No evidence was found for this as a risk factor.
The Committee discussed how limited the evidence base was when looking at whether choriamnionitis, other genito-urinary infections and respiratory tract infections requiring admission to hospital were significant risk factors for the child of a pregnancy being given a diagnosis of cerebral palsy. They agreed that high-priority research to look at the effects of different antibiotic regimens for treating genito-urinary infections in pregnant women on subsequent rates of cerebral palsy was needed.
4.9.3. Consideration of economic benefits and harms
Knowing the most important risk factors for developing cerebral palsy may lead to better prediction and identification (and thus more timely management) and has, therefore, indirectly, potentially important resource implications. However, this was an epidemiological review question and economic analysis was not applicable as it does not involve a comparison of competing alternatives.
4.9.4. Quality of evidence
The quality of each study was assessed using the NICE methodology checklist (2012) for prognostic studies, the NICE methodology checklist (2012) for systematic reviews and the NICE methodology checklist (2012) for cohort studies. Meta-analyses of observational studies and cohort studies were the most appropriate study designs for addressing this question, so were initially assigned high quality and downgraded based on potential sources of bias. Prospective and retrospective cohorts were both initially assigned high quality, as most of the retrospective studies used very large national databases. Only studies presenting adjusted analyses were included in the review, and the following covariates were indicated as the most relevant: gestational age, multiple birth, socioeconomic status, hypoxic events and neonatal sepsis. Studies were downgraded when their multivariate analysis included less than 3 of these covariates.
4.9.4.1.1. Quality of studies on antenatal risk factors
- fetal growth retardation: 1 study, moderate quality
- haemorrhagic events: 1 study, high quality
- maternal infections: 3 studies, moderate to very low quality
- multiple pregnancies: 3 studies, high to moderate quality.
4.9.4.1.2. Quality of studies on perinatal risk factors
- hypoxic-ischaemic events: 3 studies, high to low quality
- neonatal encephalopathy: 1 study, low quality
- neonatal sepsis: 6 studies, high to moderate quality
- choriomanionitis: 4 studies, high to low quality
- Apgar score at 10 min: 1 study, moderate quality.
4.9.4.1.3. Quality of studies on postnatal risk factors
- gestational age: 5 studies, high to low quality
- neonatal infections: 1 study, moderate quality.
4.9.5. Other considerations
The recommendations related to this evidence review were based on the evidence and the Committee’s clinical experience.
4.9.6. Key conclusions
The Committee concluded that multiple factors play a key role in the aetiology of cerebral palsy, but that most of the studies analysed the magnitude of independent risk factors by using adjusted analyses. Clear evidence was shown for the following factors that have an independent role in contributing to the aetiology of cerebral palsy: gestational age, birthweight, serious maternal infections, neonatal encephalopathy and neonatal sepsis.
4.10. Recommendations
- Recognise the following as independent risk factors for cerebral palsy:
- perinatal factors:
- low birth weight
- chorioamnionitis
- neonatal encephalopathy
- neonatal sepsis (particularly with a birth weight below 1.5 kg)
- maternal respiratory tract or genito-urinary infection treated in hospital
- postnatal factors:
- meningitis.
- Provide an enhanced clinical and developmental follow-up programme (see recommendations 12 to 19) for children who have any of the risk factors listed in recommendation 1.
4.11. Research recommendations
- What is the association between different antibiotic regimes to treat genito-urinary and respiratory tract infections in pregnant women and subsequent rates of cerebral palsy in children?
Footnotes
- i
The NICE guideline on developmental follow-up of preterm babies (publication expected August 2017) will contain more information about risk factors specific to preterm birth.
- j
The NICE guideline on preterm labour and birth covers preventing or delaying preterm birth, steroid treatment for maturation of fetal lungs and neuroprotection for the baby.
Tables
Table 11Summary of included studies
| Study | Data source | Sample and population studied | Risk factor(s) studied | Adjustment for: | Results | Quality of the study |
|---|---|---|---|---|---|---|
| Beaino 2010 | EPIPAGE cohort | N=2357 born 22 to 32 weeks of gestational age | Multiple pregnancy | GA, sex, small for GA, multiple pregnancy and PROM, neonatal factors (RDS). | aOR=0.67 (0.43-1.03)
aOR=1.6 (0.7-3.8) | Moderate |
| Himpens 2010 | Children assessed at the Centre for Developmental Disorders, Ghent | N=984 high-risk children (referred from NICU) | Multiple pregnancy | GA, gender, MG, BA, MV, WM disease and DGM lesion. | n=48/278, aOR=1.3 (0.8-2.1) | High |
| Laptook 2005 | 14 centres of the National Institute of Child Health and Human Development Neonatal Research Network | N=1473VLBW babies | Multiple pregnancy | Prenatal variables, birthweight, gender, multiple births, pneumothorax, late-onset sepsis, ventilation. | aOR=1.6 (1.1-2.5) | High |
| Livinec 2005 | EPIPAGE cohort | N=2382 born 22 to 32 weeks of gestational age | Maternal haemorrhagic events | For singletons = pregnancy complications, sex, GA, prenatal steroids; for twins = pregnancy complications, type of placentation, in utero vital status of co-twin, sex, GA, prenatal steroids. |
| High |
| Miller 2013 | Danish National Birth Register (National Registry) | N=440,564 singletons born 1997-2003 and resided in Denmark up to Dec 2008 | Maternal infections | Maternal age, smoking, parental income, calendar year. |
| Very low |
| Streja 2013 | Danish National Birth Cohort | N=81,066 singletons | Maternal infections | Maternal age, alcohol consumption, binge drinking, combined SES, season of birth, year of birth, number per household, smoking. |
Stratified analysis by GA
| Low |
| Wu 2013 | Danish Medical Birth Register (National Registry) | N=588,936 first-born singletons born 1982-2004 | Maternal infections | Maternal age, sex, maternal education, maternal marital status, birth year, family income, maternal infection before birth. |
| Moderate |
| Bear & Wu 2016 | California Office of State-wide Health Planning and Development | N=6,018,504 Californian births over an 11-year period | Maternal infections | Maternal age, family origin, education and SES; maternal hospital diagnosis of obesity, and infant sex. |
| High |
| Dammann 2001 | Regional cohort of VLBW babies | N=324 followed up until age 6 years | Fetal growth retardation (measured as SGA) | GA, foreign background, caesarean section, sepsis and PROM. |
| Moderate |
CP cerebral palsy, sCP spastic cerebral palsy, aOR adjusted odds ratio, aHR adjusted hazard ratio, GA gestational age, BW birthweight, VLBW very low birthweight, SES socioeconomic status, RDS respiratory distress syndrome, BPD bronchopulmonary dysplasia, PDA patent ductus arteriosus, IVH intra-ventricular haemorrhage, PVL peri-ventricular haemorrhage, PROM premature rupture of membranes, NICU neonatal intensive care unit, HIE hypoxic-ischaemic event, SGA small for gestational age, AGA appropriate birthweight for gestational age, EOS early onset sepsis, LOS late onset sepsis, MG multiple gestation, BA birth asphyxia, MV mechanical ventilation, WM white matter, DGM deep grey matter, RCT randomised controlled trial
Table 12summary of included studies
| Study | Data source | Sample and population studied | Risk factor(s) studied | Adjustment for: | Results | Quality of the study |
|---|---|---|---|---|---|---|
| Ahlin 2013 | Swedish Medical Birth Registry (national registry) | N=309 cases and 618 controls | Neonatal encephalopathy | All risk factors from univariate analyses attaining p<0.1 forCP were included in a stepwise multiple logistic regression analysis | Neonatal encephalopathy
| Low |
| Alshaikh 2013 | Meta-analysis | 17 studies involving N=15,331 VLBW infants | Neonatal sepsis | n/a | Pooled OR for CP from 11 studies=2.09 (1.78-2.45) I-squared=36.9%, p=0.064 | High |
| Alshaikh 2014 | Neonatal database of single centre | N=332 preterm babies | Neonatal sepsis | GA, severe IVH, chorioamnionitis and postnatal steroids | CoNS sepsis: aOR=0.63 (0.24-1.64) | Moderate |
| Laptook 2005 | 14 centres of the National Institute of Child Health and Human Development Neonatal Research Network | N=1473VLBW babies | Neonatal sepsis | Prenatal variables, BW, gender, multiple births, pneumothorax, LOS and ventilation | LOS: aOR=1.2 (0.8-1.7) | High |
| Mitha 2013 | EPIPAGE | N=2665 born 22 to 32 weeks of GA | Neonatal sepsis | For EOS:
For LOS:
|
| Moderate |
| Wang 2014 | Children admitted to NICUof 18 tertiary referral centres in Taiwan | N=5807 VLBW and preterm | Neonatal sepsis | GA, birthweight, sex, and retinopathy of prematurity >stage III | Neonatal sepsis aOR=1.22 (0.59-2.62) p=0.71 | Moderate |
| Han 2002 | Children born in 1 centre in Korea | N=437 preterm survivors |
| GA, BW, PROM or preterm labour, frequent miscarriage, birth asphyxia, neonatal sepsis, respiratory distress syndrome, neonatal seizures, ventriculomegaly, brain atrophy, periventricular echodensity, IVH, grade3 IVH, PVL |
| High |
| Nasef 2013 | Hospital charts | N=274 preterm babies <30 weeks admitted to NICU | Chorioamnionitis | Mode of delivery and presence of PROM |
| Low |
| Pappas 2014 | 16 centres | N=2390 preterm babies <27 weeks | Chorioamnionitis | Maternal age, multiple birth, parity, antenatal steroids, maternal hypertension, antepartum haemorrhage, sex, GA, small for GA, insurance, race and centre. |
| Moderate |
| Shatrov 2010 | Meta-analysis | 15 studies included | Chorioamnionitis | n/a |
| High |
| Soraisham AS 2013 | 1 regional NICU | N=384 preterm <29 weeks | Chorioamnionitis | Gestational age, maternal hypertension, PROM >24 hours, multiple pregnancy | Histological chorioamnionitis vs no HCA
| Moderate |
| Bear & Wu 2016 | California Office of State-wide Health Planning and Development | N=6,018,504 Californian births over an 11-year period | Chorioamnionitis | Maternal age, family origin, education, and socioeconomic status; maternal hospital diagnosis of obesity and infant sex. | OR=3.1 (2.9-3.4) | High |
| Himpens 2010 | Children assessed at the Centre for Developmental Disorders, Ghent | N=984 high-risk children (referred from NICU) | Hypoxic ischaemic events or birth asphyxia | GA, gender, MG, BA, MV, WM disease and DGM lesion | Birth asphyxia n=32/113:
| High |
| Suchov 2012 | 3 databases (state databases) | N=6.1 million (all children born in California 1991-2001) | Hypoxic ischaemic events or birth asphyxia | Maternal age, parity, maternal education, payer-source, family origin/ethnicity, timing of initiation of prenatal care, number of prenatal visits, GA, BW, and obstetric and neonatal comorbidities | Mild to severe birth asphyxia aOR=5.98 (5.28-6.58) | Low |
| Natarajan 2013 | Secondary analysis of RCT data | N=174 children with HIE | Apgar score at 10 min | BW, GA, gender, outborn status, hypothermia treatment and centre | Association between each point increase in Apgar at 10 min and CP aOR=0.69 (0.63-0.89) p<0.001 | Moderate |
CP cerebral palsy, sCP spastic cerebral palsy, aOR adjusted odds ratio, aHR adjusted hazard ratio, GA gestational age, BW birthweight, VLBW very low birthweight, SES socioeconomic status, CoNS coagulase-negative staphylococcus, RDS respiratory distress syndrome, BPD bronchopulmonary dysplasia, PDA patent ductus arteriosus, IVH intra-ventricular haemorrhage, PVL peri-ventricular haemorrhage, PROM premature rupture of membranes, NICU neonatal intensive care unit, HIE hypoxic-ischaemic event, SGA small for gestational age, EOS early onset sepsis, LOS late onset sepsis, MG multiple gestation, BA birth asphyxia, MV mechanical ventilation, WM white matter, DGM deep grey matter, RCTrandomised controlled trial.
Table 13summary of included studies
| Study | Data source | Sample and population studied | Risk factor(s) studied | Adjustment for: | Results | Quality of the study |
|---|---|---|---|---|---|---|
| Beaino 2010 | EPIPAGE | N=2357 born 22-32 weeks of gestational age | GA | GA, sex, small for GA, multiple pregnancy and PROM, neonatal factors (RDS) | Gestational age aO =1.00 (0.89-1.12)
| Moderate |
| Bonellie 2005 | Scottish registry | N=646 | GA | Not specified | Singletons (reference = 37+ wk.):
| Low |
| Himpens 2010 | Children assessed at the Centre for Developmental Disorders, Ghent | N=984 high-risk children (referred from NICU) | GA | GA, gender, MG, BA, MV, WM disease and DGM lesion | GA, n=25/165; aOR=1.1 (0.9-1.1) p=0.05
| High |
| Petrini 2009 | Hospitalisation and outpatient databases from the Northern California Kaiser Permanente Medical Care Program | N=141,321 children ≥30 weeks born 2000-2004 with follow-up June 2005 | GA | Maternal family origin and/or ethnicity, sex, plurality and size for gestational age status |
| Low |
| Suchov 2012 | 3 databases (state databases) | N=6.1 million (all children born in California 1991-2001) | GA | Maternal age, parity, maternal education, payer-source, family origin and/or ethnicity, timing of initiation of prenatal care, number of prenatal visits, GA, BW, and obstetric and neonatal comorbidities |
| Low |
| Stoll B. 2004 | Registry of VLBW infants maintained by the National Institute of Child Health and Human Development Neonatal Research Network | N=7892 eligible, 6314 available at follow-up | Neonatal infections | Infection group, study centre, GA, BW, sex, family origin and/or ethnicity, PROM more than 24 hours before delivery, mode of delivery, MB, antenatal antibiotic and steroids use, postnatal surfactant and steroids use, RDS, BPD, PDA, IVH, PVL and maternal age at time of delivery | Meningitis with or without sepsis n=184/5740; aOR=1.6 (1.0-2.5) | Moderate |
CP cerebral palsy, sCP spastic cerebral palsy, aOR adjusted odds ratio, aHR adjusted hazard ratio, GA gestational age, BW birthweight, VLBW very low birthweight, SES socioeconomic status, RDS respiratory distress syndrome, BPD bronchopulmonary dysplasia, PDA patent ductus arteriosus, IVH intra-ventricular haemorrhage, PVL peri-ventricular haemorrhage, PROM premature rupture of membranes, NICU neonatal intensive care unit, HIE hypoxic-ischaemic event, SGA small for gestational age, EOS early onset sepsis, LOS late onset sepsis, MG multiple gestation, BA birth asphyxia, MV mechanical ventilation, WM white matter, DGM deep grey matter, RCT randomised controlled trial.
Table 14Research recommendation rationale
| Research question | What is the association between different antibiotic regimes to treat genito-urinary and respiratory tract infections in pregnant women and subsequent rates of cerebral palsy in children? |
|---|---|
| Why this is needed | |
| Importance to ‘patients’ or the population | Treatment of infection in pregnancy is of prime importance for the health of the mother. There is potential for beneficial and adverse effects on the feus. In large population studies of pregnant women, choriamnionitis, other genitourinary infections and respiratory tract infections requiring admission to hospital are significant risk factors for the child of that pregnancy being given a diagnosis of cerebral palsy. The mechanisms are uncertain but include cytokine-induced damage to developing white matter leading to periventricular leukomalacia and sensitisation of the fetal brain to damage from hypoxia. Chorioamnionitis may precipitate preterm labour. Other infections are a risk to the mother’s general health. Pyrexia during labour is a risk factor for neonatal encephalopathy and cerebral palsy. |
| Relevance to NICE guidance | High priority: Minimising known risk factors for development of cerebral palsy |
| Relevance to the NHS | Very large, if cases of cerebral palsy were reduced this would reduce the requirement in health, social and educational settings |
| National priorities | |
| Current evidence base | Conflicting |
| Equality | Risks of maternal infections are recognised at different prevalence in diff social-economic groups |
Table 15Research recommendation statements
| Criterion | Explanation |
|---|---|
| Population | Large multi-centre cohort of children and their mothers delivered in a number of the regions of the UK |
| Intervention | Data collection: Maternal infection and specific anti-biotic use from: Primary care and hospital data Neonatal and maternal discharge information Looking at outcomes: Developmental outcome via national screening programme at age 2 and 5 |
| Comparator | No cerebral palsy |
| Outcome | Rates/risk of cerebral palsy |
| Study design | A prospective multi-centre study collecting prospective primary care and hospital data then linked to neonatal discharge diagnosis and outcome |
| Timeframe | Within 5 years |