DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY REVIEW

Contribution of socio-economic status on the prevalence of

cerebral palsy: a systematic search and review
MYRILL SOLASKI 1 | ANNETTE MAJNEMER 2 | MARYAM OSKOUI 3
1 Department of Family Medicine, McGill University, Montreal, QC; 2 Department of Physical and Occupational Therapy, McGill University, Montreal, QC;
3 Departments of Pediatrics and Neurology/Neurosurgery, McGill University, Montreal, QC, Canada.
Correspondence to Maryam Oskoui, Montreal Children’s Hospital, 2300 Tupper Street, Room A-512, Montreal, QC H3H 1P3, Canada. E-mail: maryam.oskoui@mcgill.ca

PUBLICATION DATA AIM The association between socio-economic status (SES) and cerebral palsy (CP) remains
Accepted for publication 9th February 2014. controversial. Preterm birth, low birthweight, and postnatal injuries are accepted mediating risk
Published online 19th April 2014. factors for CP, but the question remains whether SES confers additional risk. The aim of this
study was to analyse existing knowledge on the relationship between SES and the risk of CP.
ABBREVIATIONS METHOD We conducted a systematic search and review of potentially relevant research
SES Socio-economic status relating to SES and CP published from 1980 to 2012. Heterogeneity between studies did not
RGSC Registrar General’s Social Class allow for data aggregation or meta-analysis; therefore, a narrative review was used to
summarize the findings.
RESULTS Twelve studies were included in the systematic review. Of these, eight found low
SES to be a risk factor for increased CP prevalence. Three studies detected statistically
significant associations even after controlling for birthweight and gestational age as
variables. Two of these studies also accounted for additional confounding variables (multiple
births and timing of CP acquisition) and continued to detect contributory effects of SES.
Linear negative correlations between CP prevalence and SES were shown by three studies.
INTERPRETATION Evidence suggests that the effect of SES on CP prevalence goes beyond
that of the mediating factors preterm birth, low birthweight, and postnatal trauma. These
associations were seen in area-based and, to a lesser extent, individual measures of SES. A
better understanding of mediating factors is imperative in developing targeted public health
intervention programmes to reduce the prevalence of CP.

A socio-economic gradient in child health has been shown
in preterm birth, low birthweight, and traumas acquired
postnatally.1–3 Each of these factors is a part of the causal
pathway to cerebral palsy (CP);4,5 however, a socio-
economic gradient in CP has not been well established. As
the most common cause of childhood physical disability,
with an aetiological profile reflecting maternal and infant
medical care, an association between socio-economic dis-
parities and risk of CP would inform targeted public health
prevention programmes. It is likely that socio-economic
status (SES) acts through these known causal pathways
of CP but it is not known whether or not it confers
additional risk. A better understanding of this is important
in establishing preventative strategies for subpopulations at
risk. Our goal was to analyse existing knowledge on the
association between SES and the risk of CP, with particular
attention to preterm birth, low birthweight, and postnatal
trauma as mediating factors.
METHOD
Systematic search strategy
Two bibliographic databases (MEDLINE and Embase)
were searched to identify all potential citations on the
association between SES and CP prevalence published
between 1980 and 2012. Only studies presenting data col-
lected after 1980 were included because of the considerable
changes in perinatal care, more uniform definition of CP,
and routine use of imaging after this date.6 The search
strategy was developed with the help of a health sciences
librarian. Additional searches were conducted using multi-
ple proxy terms for SES (education, social class, social sta-
tus, income, poverty, deprivation, occupation) and cross-
referenced with the initial search results to ensure a com-
prehensive search strategy.
References from MEDLINE and Embase were com-
bined and downloaded into a reference manager (EndNote,
Thomson Reuters, Philadelphia, PA, USA). Abstracts of all
references were screened independently by two reviewers
to select relevant population-based studies on the associa-
tion between SES and CP prevalence. Studies published in
French or English were included. Conference proceedings,
interventional studies, and review articles were excluded.
When multiple articles reported data from the same study
population, the most comprehensive study was selected.
The references of the selected studies were manually
searched to identify further potential references.

© 2014 Mac Keith Press DOI: 10.1111/dmcn.12456 1043

Disagreement pertaining to inclusion of articles was
resolved by consensus between the two reviewers.
Data analysis
The heterogeneity of measures used to assess and define
SES did not allow for data aggregation and meta-analysis.
A narrative review was used to summarize findings. Ele-
ments for quality appraisal are provided for each study;
however, quality assessment and grading of the level of evi-
dence was not performed.
RESULTS
Selected studies
References retrieved from MEDLINE and Embase were
combined to give a total of 541 articles. Of these, 129 were
duplicated and therefore were removed, which left 412 arti-
cles that met the search criteria. The references of the arti-
cles that were selected were searched and a further six
articles were added, yielding a total of 418 abstracts to be
screened. Twenty-one abstracts were selected for full-text
review, of which nine were excluded: four were conference
proceedings, three were descriptive studies with no compari-
son groups, one explored only racial disparities without SES
indicators, and one was excluded because it reported data
from the same population as another, more comprehensive
study. Thus, total of 12 studies were included in the system-
atic review. The selection process is outlined in Figure 1.
Identification
What this paper adds
• There is evidence that socio-economic status (SES) contributes to CP preva-
lence beyond the effects of known mediating risk factors.
• Area-based indicators of SES appear to have a greater effect than individual
measures.
• The heterogeneity of measures of SES prevents more effective aggregation
and interpretation of data.
Measures of socio-economic status
The measures of SES that were used in the selected studies
were heterogeneous. There is no consensus on a definition
for SES in the research community,7 making clear and dis-
cerning findings and comparative data difficult to come by.
Nine studies in this review used individual measures of
SES through questionnaires, interviews, and demographic
data obtained from census and hospital files.9–11,15–20 The
individual measures often combined factors such as paren-
tal education, occupation, marital status, and prenatal care,
making it difficult to decipher which of these was the main
cause of the association. Two studies used area-based mea-
sures: the median income from census data13 and the Car-
stairs Index,12 which was applied to electoral wards divided
into quintiles from the most to the least deprived. One
study used a combination of both individual and area-based
measures: the Registrar General’s Social Class (RGSC)
assignment at birth based on father’s occupation and a
postal code-based enumeration district classification at the
time of study.8

Records identified through Additional records identified

database searching through selected
(n=541) bibliographies (n=6)

Records after duplicates removed

Screening

(n=418)

Records screened Records excluded

(n=418) (n=397)
Eligibility

Full-text articles Full-text articles

assessed for eligibility excluded, see text
(n=21) ('Results')
(n=9)

Studies included in
qualitative synthesis
Included

(n=12)

Figure 1: Article selection flow diagram.

1044 Developmental Medicine & Child Neurology 2014, 56: 1043–1051

Association between socio-economic status and cerebral
palsy
Twelve studies explored the association between SES and
CP (Table I). Of these, eight found low SES to be a risk
factor for CP8–13,15,16 and four found no association
between SES and prevalence of CP.17–20
Three studies found statistically significant negative
associations between SES and overall CP prevalence even
after controlling for birthweight and gestational age as
variables.8–10 Two of these studies also accounted for addi-
tional confounding effects (multiple births and timing of
CP acquisition).8,9 A linear correlation between CP preva-
lence and SES was shown in three studies.8,9,11 This dose–
response type curve is what would be expected of a true
SES effect.
In a retrospective cohort study in the UK using an area-
based measure of SES,12 an increased risk of acquired CP
(postnatal) was seen in lower SES groups. The difference
in relative risk between the most and least deprived regions
was 1.86 (95% confidence interval [CI] 1.19–2.88). Non-
acquired CP was found to follow an SES gradient in only
two of the five regions and, therefore, the overall associa-
tion was not statistically significant (relative risk difference
of 1.16 [95% CI 1.00–1.35]). When infants with non-
acquired CP and of normal birthweight were analysed sep-
arately, a lower SES gradient was seen. Although it was
more likely that children from lower SES families would
be of very low or moderately low birthweight, among all
infants with low birthweight low SES did not increase the
risk of CP. The authors suggest maternal clinical and sub-
clinical infection and its role in preterm birth as a plausible
mediating factor. Interestingly, among the group with non-
acquired CP, the SES gradient was greatest in patients
with spastic bilateral CP or with severe intellectual impair-
ment.12
In a retrospective cohort study in Ireland,11 which also
used an individual measure of SES, there was a clear SES
gradient in risk of CP among infants with birthweight
greater than 2500g (prevalence of 0.89, 0.91, 1.37, 1.48,
and 1.56 per 1000 births in decreasing social class groups;
p<0.01). The gradient was not seen in low-birthweight
infants beyond that which was expected from the propor-
tion of low-birthweight births in each social class group.
There was also a clear SES gradient in children with CP
who were non-ambulant. Severity was measured as ‘severe
enough to prevent walking by age 4 years’ and was more
prevalent with increasing disadvantage. Patients with post-
natal (acquired) CP were excluded. A statistically signifi-
cant upwards trend for prevalence of hemiplegia and
diplegia with decreasing social class was among the associ-
ations detected by this study.11
In a retrospective cohort study in Sweden using individ-
ual indicators of SES,9 a linear association between SES
and CP (excluding injuries and malformations) was
observed. Children from low-SES households had a 50%
greater chance of being discharged from a hospital with a
CP diagnosis than children in the highest SES category.
Adjustments made for gestational age, perinatal asphyxia,
and small-for-gestational age status reduced this associa-
tion but did not eliminate it completely. Preterm birth was
the most important mediating factor. In the initial multi-
variate analysis, the odds ratio for risk of CP in low- versus
high-SES groups was 1.49 (95% CI 1.16–1.91). After con-
trolling for gestational age, perinatal asphyxia, and small-
for-gestational age status, the odds ratio decreased only to
1.36 (95% CI 1.05–1.75), remaining significant. Gesta-
tional age accounted for 65% of this change. Therefore,
the authors concluded that an association between SES
before birth and the development of CP exists indepen-
dently of perinatal risk factors. Analysis of the patients
excluded because their CP was caused by injuries or mal-
formations did not show an association between SES and
injuries in this study.9
The third study in which a true gradient in CP preva-
lence was seen across SES levels was a retrospective cohort
study in the UK.8 The association remained significant for
only the area-based measure of SES (based on enumeration
district, with a quintile derived from postal codes) and not
for the individual measure of SES RGSC, after adjustment
for gestational age and birthweight using a multivariate
logistic regression. The risk of CP in the lowest social class
(5) compared with the highest social class (1) was 2.57
(95% CI 1.27–5.21) for the individual-level RGSC measure
of SES and 1.65 (95% CI 1.14–2.39) for the area-based
(enumeration district) measure of SES when considering
all singleton births. After adjustment for known risk fac-
tors, the association for the individual level measure of
SES RGSC was no longer significant (odds ratio [OR]
2.11, 95% CI 1.0–4.30 points; p=0.160) and the association
for the area-based measure of SES (enumeration district)
remained significant (OR 1.55 [95% CI 1.06–2.25];
p=0.046).8
In a cross-sectional, multi-site study from the USA,13
racial differences in prevalence and variations by neigh-
bourhood SES were explored. Children were identified
using the Autism and Developmental Disabilities Monitor-
ing Network, and included postnatal causes of CP. The
average CP prevalence was 3.6 per 1000 8-year-old chil-
dren, which is higher than previously reported, suggesting
perhaps a lack of specificity in the definition used.14 Preva-
lence rates were highest among black non-Hispanic chil-
dren (4.2 per 1000, 95% CI 3.6–4.9 per 1000). Low- and
middle-income neighbourhoods had a combined preva-
lence of CP (4.1 per 1000) that was 70% higher than that
of high-income neighbourhoods (2.4 per 1000). There was
no adjustment for birthweight or gestational age. The
authors postulate possible genetic predisposition as a con-
tributing factor to ethnic disparities in prevalence rates.13
In a retrospective cohort study from California, USA,10
ethnic disparities in CP prevalence were observed. Females
who had received no prenatal care and females of low
insurance status or low educational attainment had an
Review 1045
1046
Table I: Selected studies on association between socio-economic status and cerebral palsy prevalence

Number
of
Total patients Number of Measure
number with comparison of Acquired Adjusted
of patient cerebral participants social-economic Data collection cerebral mediating Birth
Study Design (n) palsy (n) (n) status method palsy factors Country cohorts Results

Beqaj- Cross-sectional 81 150 Individual Parental Included Sex, rural/ Kosovo N/A Increased risk in
Zhjeqi15 case–control questionnaire urban lower maternal
education: no
maternal
education in 44%
of patients with
cerebral palsy vs
6% of comparison
participants
Boyle et al.17 Cross-sectional 305 Individual Parental Included None USA 1997–2008 No association
interview
Dolk et al.12 Retrospective 1 657 569 3758 Area-based Carstairs Index Included Birthweight UK 1984–1997 For acquired CP,
cohort (total by and electoral risk ratio of low
census) ward quintiles area-based socio-

Developmental Medicine & Child Neurology 2014, 56: 1043–1051

economic status
1.86 (1.19–2.88)
points
Dowding Retrospective 150 189 258 Individual Registrar Excluded Birthweight, Ireland 1976–1981 Clear socio-
and Barry11 cohort (total by General’s non-ambulation economic status
census) Social Class gradient in typical
birthweight (0.89,
0.91, 1.37, 1.48,
1.56 prevalence
per 1000 by
decreasing socio-
economic status,
p<0.01)
Emond Nested 40 16 751 Individual Registrar Excluded Gestational UK 1958, No association
et al.18 case–control 41 16 136 General’s age, 1970
Social Class birthweight
Hjern and Retrospective 805 543 1437 Individual Statistics Excluded Gestational Sweden 1987–1993 Low
Thorngren- cohort Sweden age, socio-economic
Jerneck9 classification asphyxia, status odds ratio
small for 1.49 points
gestational (1.16–1.91),
age adjusted odds
ratio 1.36 points
(1.05–1.75)
Nelson and Prospective 45 559 189 Individual Score derived Excluded Birthweight USA 1959–1966 No association
Ellenberg19 cohort from NCPP
(antecedents) database
Petridou Case–control 103 254 Individual Parental Included None Greece 1984–1988 No association
et al.20 interview
 Table I: Continued

Number
of
Total patients Number of Measure
number with comparison of Acquired Adjusted
of patient cerebral participants social-economic Data collection cerebral mediating Birth
Study Design (n) palsy (n) (n) status method palsy factors Country cohorts Results

Sciberras Case–control 55 200 134 134 Individual Parental Included None Malta 1981–1990 Increased risk in
and interview low
Spencer16 socio-economic
status (unskilled
manual workers
p<0.001)
Sundrum Cross-sectional 295 760 293 Individual Registrar Included Gestational UK 1982–1997 Adjust odds ratio
et al.8 and General’s age, for area-based
area-based Social Class birthweight socio-economic
and postcode status 1.55
classification (1.06–2.25) points
Wu et al.10 Retrospective 6.2 8397 Individual Hospital charts Excluded Ethnic USA 1991–2001 Risk ratio 1.33
cohort million group, (1.20–1.49) points
birthweight,
gestational
age
Yeargin- Cross-sectional 114 897 416 Area-based Median income Included Ethnic group USA 2002 Increased risk in
Allsopp on census files low
et al.13 socio-economic
status: low and
middle income
neighbourhoods
4.1 vs high
income
neighbourhoods
2.4 prevalence
per 1000

N/A, not available; NCPP, Collaborative Perinatal Project of the National Institute of Neurological and Communicative Disorders and Stroke.

Review
1047
increased risk of having a child with CP. Maternal educa-
tional attainment had an inverse dose–response relationship
with risk of CP, but only in white and Hispanic females.
Mothers who had not attended school or had attended
only primary school had the greatest increased risk (risk
ratio [RR] 1.33, 95% CI 1.20–1.49), while mothers who
graduated from high school or college had a more modest
increase in risk (RR 1.21 and 1.14 respectively). An
increased prevalence of CP among black children has been
found previously,13 but the aetiology of this difference
remains unclear despite speculation on a socio-economic
mediator. This 2011 study found that black infants were
more likely than white infants to have been of very or
moderately low birthweight (two- to three-fold higher) or
to be born preterm (<37wks’ gestation). After controlling
for these differences through a logistic regression model,
ethnic group was no longer a risk factor for CP. In fact,
black preterm or low-birthweight infants were less likely
than white infants born preterm or with low birthweight to
have CP. The authors concluded that the increased risk of
CP among black children is the result of increased preva-
lence of low birthweight.10
A retrospective cohort study of 81 patients with CP from
the National Institute of Public Health in Kosovo15 detected
a statistically significant increased risk of CP in offspring
whose parents had lower levels of education and whose
mothers attended fewer prenatal appointments. Mothers
without any education constituted 44.4% of the group with
CP, compared with only 6.0% of a group of comparison par-
ticipants, whereas mothers with a medium or high levels of
education accounted for only 23.5% of the group with CP
compared with 73.3% of the comparison participants
(p<0.01). The mean number of prenatal visits to consultation
institutions was 1.65 in the CP group and 4.61 in the com-
parison group (p<0.01). The groups in this study were
matched for sex and rural or urban place of residence only
and the study included patients with acquired CP.15
A case–control study in Malta of children born from
1981 to 199016 found an increased risk of CP in offspring
of unskilled manual workers on univariate analysis
(v2 goodness of fit test p<0.0001). This included partici-
pants with postnatal CP. There was no adjustment for
confounders and there was an increased risk of CP with
preterm birth and low birthweight.16
In a cross-sectional survey in the USA using self-
reported diagnosis on a survey and an individual measure
of SES,17 there was a non-significant increased prevalence
of CP in the lower SES group. Low SES was defined as
having public health insurance (p<0.05), income lower than
200% of the poverty level (p<0.05), and maternal educa-
tional achievement of high school level (p<0.05), and each
was significantly associated with an increased prevalence of
‘any disability’. There was no adjustment for gestational
age or birthweight and patients with postnatal CP were
included.17
An older nested case–control study from the UK,18
based on births from two separate 1-week periods in 1958
1048 Developmental Medicine & Child Neurology 2014, 56: 1043–1051
and 1970, found no differences in CP prevalence based on
father’s occupation, marital status, ethnic group, or educa-
tional background. Two groups of comparison participants
were used. The first group was matched on mother’s age,
parity, social class, and marital status and no differences in
SES were observed. The second group was matched on
birthweight, gestational age, certainty of dates, sex, and
plurality and showed no significant differences from the
index group on other important variables such as region,
household overcrowding, ethnic group, or educational
background.18
A prospective cohort study in the USA on 189 partici-
pants with CP among births from 1959 to 196619 found
no association between individual measures of SES and risk
of CP on univariate analysis. The individual measures
included maternal education, marital status, and maternal
working status. Among children with low birthweight, the
risk of CP was higher in the offspring of mothers who
attended fewer than six prenatal appointments than in
those whose mothers attended seven or more. Of the chil-
dren with CP and low birthweight, 75% were born to
mothers who had fewer than six prenatal appointments.19
A case–control study from Greece20 looking at risk fac-
tors for CP found no association between risk of CP and
maternal or paternal occupation on univariate analysis.
However, the group of 254 comparison participants was
selected from siblings or neighbours of the 103 participants
studied and could be expected to share area-based SES
measures.20
DISCUSSION
The relationship between SES and preterm birth is well
established.1 Two studies in this systematic review adjusted
for gestational age, which reduced the effect size without
eliminating it.8,9 The increased prevalence of CP among
preterm infants is apparent in developing and developed
countries alike.16 Many mediators have been suggested to
explain the pre- and perinatal effects of SES on CP preva-
lence, including maternal clinical and subclinical infection,
nutrition, and prenatal health care. Risk factors for preterm
and small-for-gestational age births can be included in this
category. Mechanisms that may cause hypoxic–ischaemic
injury to the fetus due to placental damage have been
hypothesized as mediators.9 Increased vulnerability to
stress may also manifest via, as yet unquantifiable, biopsy-
chosocial pathways of adverse pregnancy outcomes.21
Stressful intrauterine conditions and their effect on the
central nervous system have recently been increasingly
studied and documented. High levels of stress result in
delayed fetal maturation, impaired cognitive performance
during infancy, and decreased brain volume in areas associ-
ated with learning and memory.22 Measures of early gesta-
tional maternal anxiety have been correlated with
reductions in grey matter volume in several areas of the
cerebral cortex and the cerebellum.23 Unfortunately, the
preponderance of pre- and perinatal CP, as well as the
difficulty in determining time of onset for so many
participants, makes analysis among categories difficult and
confounds the search for causal associations.
Two studies explored the risk of postnatal CP alone.
One found an increased risk of CP among the most
deprived group12 and the other found no association
between the factors.9 Although postnatal CP constitutes
only a small percentage of the total number of patients
with CP,24 approximately 5% to 17% in developed coun-
tries,4 childhood injuries, including accidents, have a strong
association with SES,3 indicating a causal pathway that is
likely and providing strong evidence on the effect of SES
on CP acquired postnatally. It was surprising that, given
this well-associated link, the Swedish study9 found no asso-
ciation. Of note is that in this study the group consisted of
participants with CP caused by severe head injury only and
as a whole the group was not included in the main study
but was instead analysed separately to determine effect.
Some studies, such as the Swedish study, excluded CP
acquired postnatally from analysis and found correlations
between SES and CP.9 It is likely that this heterogeneity
between studies led to a decreased effect size for the stud-
ies that excluded it and an increased observed effect of
SES in the studies that included it. The majority of cases
of CP acquired postnatally were caused by infection, sur-
gery-related vascular incidents, or either accidental or non-
accidental head injury, all of which are worthy targets of
public health intervention strategies.
Four studies using individual measures of SES did not
find an association between SES and risk of CP,17–20 while
five studies using individual measures showed an increased
risk in the most deprived groups.9–11,15,16 Individual mea-
sures of SES include family income and parent education
level, both of which contribute to possible causal pathways
such as the ability to afford care and knowledge that it is
important to seek out high-quality prenatal care. For
example, individually measured proxies for SES such as
periodontal health have been linked to preterm birth and
small-for-gestational age infants. This was the case in both
lower- and middle-class females, indicating a physiological
mechanism of action such as subclinical infection.25 Both
studies using an area-based measure found an increased
risk in more deprived groups,12,13 whereas the study using
both an individual and area-based measure found an associ-
ation only in the area-based measure, when gestational age
and birthweight were adjusted for.8 This would suggest
that contextual environmental factors imply additional risk
beyond differences in individual-level SES characteristics.
There is already strong evidence linking neighbourhood or
‘area-based’ characteristics to childhood health outcomes.
Access to well-funded libraries, affordable produce, clean
green spaces, and well-educated social contacts may miti-
gate the effects of SES stemming from low income, poor
parental educational attainment, or lack of personal sav-
ings. A recent study looking at the role of built, socio-eco-
nomic, and social environments on health outcomes found
that access to parks, social ties, and neighbourhood afflu-
ence were correlated with healthy behaviours.26 In the case
of CP, these healthy behaviours could manifest in
decreased levels of CP through lower maternal stress as a
result of exposure to less violence, access to more green
spaces, improved nutrition from access to affordable pro-
duce, and improved prenatal care stemming from social
and hard copy (e.g. library/Internet) knowledge acquisition.
In essence, despite relatively low personal income or
employment, living in a neighbourhood designated as hav-
ing higher SES could be expected to reap significant social
capital and components of human capital when conceptual-
izing SES in terms of the more modern definition encom-
passing material, human, and social capital.7 Conversely,
research on individual versus area-based measures of SES
has shown individual measures of SES to have a stronger
association with childhood health outcomes than area-
based measures.27,28
The relationship between SES and low birthweight is
well established. Mediating factors are likely to include
nutrition, smoking, and maternal illness. Two studies strat-
ified analysis based on birthweight and found the increased
risk attributed to low SES dissipated when infants with low
birthweight were compared, but the gradient persisted
among normal-birthweight infants.11,12 Two studies also
adjusted their analysis by birthweight and reduced the
effect of SES on increased prevalence of CP but did not
eliminate it.8,9 Dolk et al.12 proposes three plausible expla-
nations why children with normal birthweight with CP
may show an SES gradient but low-birthweight children
with CP do not. First, low birthweight and CP may be
precipitated by the same insult, which occurs more fre-
quently in low-SES populations, and therefore controlling
for low birthweight would eliminate the measurable effect.
Second, CP in infants of low birthweight may be predomi-
nantly an effect of being small, while CP in infants of
normal birthweight may be mediated by a different SES-
related factor. Third, the causes of CP in children of nor-
mal and low birthweight may be unrelated either to each
other or to birthweight and while typical birthweight
mediators may be linked to SES, those of low birthweight
CP may not. The finding that very low-birthweight and
moderately low-birthweight infants are not more likely to
have CP if they are born in deprived areas12 is consistent
with the assertions above.
The low-birthweight paradox is the phenomenon
whereby low-birthweight infants have a higher chance of
perinatal survival if they are born into low-SES families
than if they are born into high-SES families. Typically the
low-birthweight paradox is applied to the finding that the
mortality rate in infants of low birthweight born to moth-
ers who smoke is lower than that of low-birthweight
infants born to mothers who do not, the explanation being
that the cause of the low birthweight in mothers who do
not smoke confers more serious effects than smoking.29
Infants of low birthweight born into low-SES families were
found to have a lower prevalence of CP than infants with
normal birthweight born into low-SES families.11,12 The
cause of CP in infants of normal birthweight, therefore,
Review 1049
can be hypothesized to be the result of an SES-linked vari-
able other than those causing low birthweight. Despite the
fact that the effect of SES was restricted to normal-birth-
weight infants,11 it is important to remember that the
higher number of low-birthweight births in low-income
areas accounts for a large percentage of the prevalence of
CP.30 Therefore, interventions targeted at mediating the
disparity in low-birthweight infants between SES catego-
ries will also decrease CP prevalence overall.
Strengths and limitations
This review is based on rigorous search strategy and data
abstraction of the selected studies by two reviewers (M.S.
and M.O). Efforts were made to report this systematic
search and review based as much as possible on the PRIS-
MA statement.31 Owing to the narrative nature of this
research, special precautions were taken in the analysis of
the selected studies. There was high heterogeneity of vari-
ables, methodologies, and baseline characteristics of popu-
lations being studied, specifically the measures of SES and
variations in controlling for time of acquisition, birth-
weight, and gestational age, which prevented aggregation
of data and meta-analysis. We sought to assess the rela-
tionship between study results and key aspects and com-
pared these across studies. The strength of evidence and
effect size, as well as robustness and methodological quality
of the primary studies were carefully considered.
CONCLUSION
Existing studies show mixed results, with either no associ-
ation or a protective effect of higher SES. Differences in
study design and target population may contribute in part
to these differences. The direction of the association,
however, suggests an additional risk of CP in children of
lower SES, seen with both individual and area-based mea-
sures, beyond the risk associated with preterm birth, low
birthweight, or postnatal trauma. Future studies are
needed to explore the affect of SES on the natural history
of CP such as ambulation, contractures, scoliosis, and
other associated comorbidities. Females of lower SES may
benefit from clinical vigilance during pregnancy and close
developmental surveillance of their infants. A better
understanding of mediating factors at both an individual
and a community level is imperative in developing tar-
geted public health intervention programmes to prevent
the risk of CP.

REFERENCES
1. Raisanen S, Gissler M, Saari J, Kramer M, Heinonen S.
Contribution of risk factors to extremely, very and mod-
erately preterm births – register-based analysis of
1,390,742 singleton births. PLoS One 2013; 8: e60660.
2. Spencer N, Bambang S, Logan S, Gill L. Socioeco-
nomic status and birth weight: comparison of an area-
based measure with the Registrar General’s Social Class.
J Epidemiol Community Health 1999; 53: 495–98.
3. Petrou S, Kupek E, Hockley C, Goldacre M. Social
class inequalities in childhood mortality and morbidity
in an English population. Paediatr Perinat Epidemiol
2006; 20: 14–23.
4. Stanley F, Blair E, Alberman E. Cerebral Palsies: Epide-
miology and Causal Pathways. London: Mac Keith,
Cambridge University Press, 2000.
5. Pharoah PO, Cooke T, Rosenbloom L. Acquired cere-
bral palsy. Arch Dis Child 1989; 64: 1013–16.
6. Stoelhorst GM, Rijken M, Martens SE, et al. Changes
in neonatology: comparison of two cohorts of very pre-
term infants (gestational age <32 weeks): the project on
preterm and small for gestational age infants 1983 and
the leiden follow-up project on prematurity 1996–1997.
Pediatrics 2005; 115: 396–405.
7. Oakes JM, Rossi PH. The measurement of SES in
health research: current practice and steps toward a new
approach. Soc Sci Med 2003; 56: 769–84.
8. Sundrum R, Logan S, Wallace A, Spencer N. Cerebral
palsy and socioeconomic status: a retrospective cohort
study. Arch Dis Child 2005; 90: 15–8.
9. Hjern A, Thorngren-Jerneck K. Perinatal complications
and socio-economic differences in cerebral palsy in
Sweden – a national cohort study. BMC Pediatr 2008;
8: 49.
10. Wu YW, Xing G, Fuentes-Afflick E, Danielson B,
Smith LH, Gilbert WM. Racial, ethnic, and socioeco-
nomic disparities in the prevalence of cerebral palsy.
Pediatrics 2011; 127: e674–e681.
11. Dowding VM, Barry C. Cerebral palsy: social class dif-
ferences in prevalence in relation to birthweight and
severity of disability. J Epidemiol Community Health 1990;
44: 191–95.
12. Dolk H, Pattenden S, Bonellie S, et al. Socio-economic
inequalities in cerebral palsy prevalence in the United
Kingdom: a register-based study. Paediatr Perinat Epi-
demiol 2010; 24: 149–55.
13. Yeargin-Allsopp M, Van Naarden Braun K, Doernberg
NS, Benedict RE, Kirby RS, Durkin MS. Prevalence of
cerebral palsy in 8-year-old children in three areas of
the United States in a multisite collaboration. Pediatrics
2002; 2008: 547–54.
14. Oskoui M, Coutinho F, Dykeman J, Jette N, Pringsheim
T. An update on the prevalence of cerebral palsy: a sys-
tematic review and meta-analysis. Dev Med Child Neurol
2013; 55: 509–19.
15. Beqaj-Zhjeqi V. Etiological factors in cerebral palsy in
Prishtina region. Med Arh 2008; 62: 20–24.
16. Sciberras C, Spencer N. Cerebral palsy in Malta 1981 to
1990. Dev Med Child Neurol 1999; 41: 508–11.
17. Boyle CA, Boulet S, Schieve LA, et al. Trends in the
prevalence of developmental disabilities in US children,
1997–2008. Pediatrics 2011; 127: 1034–42.
18. Emond A, Golding J, Peckham C. Cerebral palsy in two
national cohort studies. Arch Dis Child 1989; 64: 848–52.
19. Nelson KB, Ellenberg JH. Antecedents of cerebral palsy:
I. Univariate analysis of risks. Am J Dis Child 1985; 139:
1031–38.
20. Petridou E, Koussouri M, Toupadaki N, et al. Risk fac-
tors for cerebral palsy: a case–control study in Greece.
Scand J Soc Med 1996; 24: 14–26.
21. Sheehan TJ. Stress and low birth weight: a structural
modeling approach using real life stressors. Soc Sci Med
1998; 47: 1503–12.
22. Sandman CA, Davis EP, Buss C, Glynn LM. Prenatal
programming of human neurological function. Int J Pept
2011; 2011: e837596.
23. Buss C, Davis EP, Muftuler LT, Head K, Sandman
CA. High pregnancy anxiety during mid-gestation is
associated with decreased gray matter density in 6–9-
year-old children. Psychoneuroendocrinology 2010; 35:
141–53.
24. Cans C, McManus V, Crowley M, et al. Cerebral palsy
of post-neonatal origin: characteristics and risk factors.
Paediatr Perinat Epidemiol 2004; 18: 214–20.
25. Pitiphat W, Joshipura KJ, Gillman MW, Williams PL,
Douglass CW, Rich-Edwards JW. Maternal periodonti-
tis and adverse pregnancy outcomes. Community Dent
Oral Epidemiol 2008; 36: 3–11.
26. Carroll-Scott A, Gilstad-Hayden K, Rosenthal L, et al.
Disentangling neighborhood contextual associations with
child body mass index, diet, and physical activity: the
role of built, socioeconomic, and social environments.
Soc Sci Med 2013; 95: 106–14.
27. Pardo-Crespo MR, Narla NP, Williams AR, et al. Com-
parison of individual-level versus area-level socioeco-
nomic measures in assessing health outcomes of children
in Olmsted County, Minnesota. J Epidemiol Community
Health 2013; 67: 305–10.
28. Singh GK, Ghandour RM. Impact of neighborhood
social conditions and household socioeconomic status on

1050 Developmental Medicine & Child Neurology 2014, 56: 1043–1051

 behavioral problems among US children. Matern Child
Health J 2012; 16: S158–69.
29. Gisselmann MD. Education, infant mortality, and low birth
weight in Sweden 1973–1990: emergence of the low birth
weight paradox. Scand J Public Health 2005; 33: 65–71.
30. Dolk H, Pattenden S, Johnson A. Cerebral palsy, low
birthweight and socio-economic deprivation: inequalities
in a major cause of childhood disability. Paediatr Perinat
Epidemiol 2001; 15: 359–63.
31. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred
reporting items for systematic reviews and meta-analy-
ses: the PRISMA statement. BMJ 2009; 339: b2535.
Review 1051