Objective We assessed baseline prevalence, risk factors and outcomes of microcephaly in a large population of neonates.
Design Retrospective cohort study.
Setting All hospitals in the province of Quebec, Canada.
Participants 794 microcephalic and 1 944 010 non-microcephalic infants born between 1989 and 2012.
Main outcome measures Baseline prevalence of microcephaly and occurrence of other congenital anomalies. We estimated the association of (1) pregnancy risk factors including TORCH infections (toxoplasmosis, rubella, cytomegalovirus, herpes, other), exposure to teratogens, diabetes and maternal congenital anomalies with risk of microcephaly, and (2) microcephaly with risk of infant mortality and severe morbidity, adjusted for maternal characteristics.
Results The overall prevalence of microcephaly was 4.1 per 10 000, ranging between 3.0 and 5.3 per 10 000 over time. Only 37% of microcephalic infants presented with other congenital anomalies. Maternal infection during pregnancy was the strongest risk factor, with 32 times the risk of microcephaly (prevalence ratio 32.38; 95% CI 22.42 to 46.75) compared with no infection. Exposure to teratogens was the next most important risk factor, with three times greater risk (prevalence ratio 3.10; 95% CI 2.37 to 4.07). Microcephaly was associated with 20 times the risk of infant mortality compared with no microcephaly (prevalence ratio 20.52; 95% CI 15.57 to 27.04) and significantly greater infant morbidity.
Conclusions In Canada, infectious exposure during pregnancy is a strong risk factor for microcephaly, and affected infants are at higher risk of poor birth outcomes. Better monitoring of microcephaly is needed in the event that Zika or other novel viruses affect future risk.
- Congenital Abnormalities
- Risk Factors
- Zika Virus
Statistics from Altmetric.com
What is already known on this topic?
Prevalence, risk factors and outcomes of microcephaly are understudied despite the potential for Zika virus to spread in North America.
What this study adds?
Congenital microcephaly is rare in Canadian newborns, and TORCH agents (toxoplasmosis, rubella, cytomegalovirus, herpes, other) are a principal risk factor.
Microcephalic newborns have high mortality and significant morbidity after delivery.
Better data are needed for surveillance of microcephaly, including risk factors and associated morbidity.
Zika virus’ rapid spread through the Americas1 2 is concerning for its link with microcephaly in offspring of pregnant women.3 4 The potential for Zika to spread locally and increase the risk of microcephaly in western countries cannot be ignored. Primary mosquito vectors of Zika, including Aedes aegypti and A. albopictus, are already present in southern and eastern parts of the USA,5 with the possibility of establishing in northern areas.2 6 A. albopictus can adapt to colder temperatures,7 and climate change may facilitate movement of mosquitoes unpredictably.6 The potential for other vectors to spread Zika is unknown.7 More importantly, Zika could be introduced through returning travellers, sexual contact and possibly even blood transfusion.4 Pregnant women travelling or with partners returning from affected areas are at particular risk.
Zika virus has potential to cause microcephaly,4 but there is little documented information on baseline prevalence of microcephaly or associated risk factors and outcomes. Prevalence varies between countries.8 In the USA, anywhere from 2 to 12 infants per 10 000 are born microcephalic, depending on the state.4 9 In Europe, the overall prevalence is approximately 2.6 to 2.9 per 10 000,10 although rates vary greatly between countries.8 The Canadian Congenital Anomalies Surveillance Network has yet to report the prevalence of microcephaly.11 Globally, only Australia reports microcephaly by maternal age and ethnicity, beyond simply tracking national prevalence.12 More informative baseline trends are needed should Zika infection ever increase in the USA, Canada or other parts of the world through local transmission or returning travellers. Our aim was to document epidemiological trends in prevalence, risk factors and outcomes of microcephaly in a large population with no evidence of Zika virus transmission.
We carried out a retrospective cohort study using discharge summaries for all hospital-born infants in the province of Quebec, Canada, between 1 April 1989 and 31 March 2012. A quarter of Canadians live in Quebec. We extracted data from the Maintenance and Use of Data for the Study of Hospital Clientele database, a compilation of hospital discharge summaries. The majority of Quebec births occur in hospital (99%). We extracted all maternal and infant charts at delivery, and linked maternal charts with corresponding infant charts using unique identification numbers. In addition, we used scrambled health insurance numbers to identify prior hospitalisations in women during childhood or adolescence. There were 1 100 083 women in this study, including 1 944 804 infants born alive at >20 weeks of gestation between 1989 and 2012. The sample excludes 7873 stillbirths for which discharge summaries were not available. We did not have information on pregnancy terminations or losses before 20 weeks.
The main outcome measure was microcephaly documented before discharge from hospital. We identified infants with congenital microcephaly using the 9th and 10th revisions of the International Classification of Diseases (ICD; 742.1, Q02).13 Microcephaly is a clinical sign, not a diagnosis.13 14 There is no strict consensus, but microcephaly is usually defined among Canadian paediatricians as an occipitofrontal head circumference 2 or 3 SD below the mean on age-matched and sex-matched curves.14–16 Congenital or primary microcephaly is diagnosed shortly after delivery, whereas secondary microcephaly develops during infancy after normal head circumference at birth.14 Congenital microcephaly is usually identified during paediatric examinations before hospital discharge, but can also be detected in utero during ultrasound screening.4 14 In this study, we focused on congenital microcephaly because we did not have information on head growth after discharge from hospital.
Other congenital anomalies
We examined whether microcephaly presented with other congenital anomalies, including anomalies of the central nervous system (spina bifida, anencephaly/encephalocoele, congenital hydrocephalus, other), eye, ear, orofacial clefts, heart (critical, non-critical),17 respiratory, digestive, urinary (hypospadias, renal agenesis, other), skeletal (congenital hip dislocation, clubfoot, limbs or digits, skull or facial bones, other) and chromosomes (Down syndrome, trisomy 13 or 18, autosomal or sex-linked, other). In addition, we assessed rare but potentially lethal structural anomalies that as a group may be associated with significant morbidity or mortality, including anencephaly, encephalocoele, omphalocoele, diaphragmatic hernia, lung agenesis, critical heart defects, renal agenesis, osteogenesis imperfecta, osteochondroplasia, trisomy 13 and trisomy 18.
We used the ICD to identify defects (see online supplementary table 1). To increase ascertainment, we used the Canadian Classification of Diagnostic, Therapeutic, and Surgical Procedures and the Canadian Classification of Health Interventions to document infants with procedures for spina bifida, congenital hydrocephaly, orofacial clefts, heart defects, gastroschisis, clubfoot and polydactyly/syndactyly (see online supplementary table 2).
Maternal risk factors
We considered demographical and clinical risk factors for microcephaly. We included maternal age (<25, 25–34, ≥35 years), parity (0, 1, ≥2 previous pregnancies), multiple birth (yes, no), rural place of residence (yes, no, unknown), material deprivation (low, low-moderate, moderate, moderate-high, high, unknown) and period (1989–1996, 1997–2004, 2005–2012). We considered clinical risk factors for microcephaly, including maternal congenital anomalies or inborn errors of metabolism,13 14 pre-existing type 1 or type 2 diabetes, exposure to teratogens determined by the presence of epilepsy or alcohol and drug abuse,14 and TORCH agents during pregnancy (toxoplasmosis, rubella, cytomegalovirus, herpes, other).14 We used the ICD to document whether women had these disorders at time of delivery (see online supplementary table 1). To improve ascertainment, we determined whether women were ever hospitalised for anomalies, diabetes or epilepsy/drug/alcohol abuse during childhood, adolescence or young adulthood prior to the index pregnancy.
Infant morbidity and mortality
We assessed infant morbidity and mortality after delivery for microcephalic versus non-microcephalic infants. Using data from discharge summaries, we determined whether infants died before discharge or were admitted to neonatal intensive care units. We used gestational age at birth to document very preterm birth (<32 weeks of gestation) or preterm birth (<37 weeks). We assessed length of stay after delivery, including ≥7 or ≥14 days. Using ICD codes, we identified infants with necrotising enterocolitis, intracranial haemorrhage, bronchopulmonary dysplasia, intrauterine growth restriction and birth hypoxia (see online supplementary table 1). We determined whether procedures indicative of morbidity were present, including C-section, intubation, transfusion and phototherapy (see online supplementary table 2).
We computed the prevalence of microcephaly per 10 000 live births. We estimated the proportion of infants with microcephaly who had other congenital anomalies, and the proportion with isolated microcephaly. In addition, we estimated the prevalence of microcephaly according to maternal characteristics, and rates of infant mortality and morbidity before discharge for microcephalic versus non-microcephalic infants.
Using log-binomial models, we computed prevalence ratios (PRs) and 95% CIs for the association between maternal exposures and risk of microcephaly, adjusting for age, parity, multiple birth, rural residence, material deprivation and period. Literature has identified these covariates as potential confounders of microcephaly and poor birth outcomes.10 18 19 We tested interaction terms between maternal exposures, such as maternal age and parity. In addition, we computed adjusted prevalence differences for the association between each maternal risk factor and microcephaly. We used similar models to estimate the association between microcephaly and each infant outcome, adjusting for the same covariates. We used robust error estimators in generalised estimating equations to account for women with more than one delivery during the study, and a Poisson distribution to facilitate model convergence for rare outcomes.20 In sensitivity analyses, we verified that models run using women instead of infants as the unit of analysis yielded similar results. We used SAS V.9.3 for data analysis (SAS Institute).
The overall prevalence of microcephaly was 4.08 per 10 000 live births (95% CI 3.80 to 4.37) between 1989 and 2012. Prevalence was relatively stable, fluctuating between 3.0 and 5.3 per 10 000 annually (figure 1). In 1989, there were 3.84 (95% CI 2.53 to 5.15) and in 2012, 5.05 (95% CI 3.54 to 6.56) microcephalic infants per 10 000.
Microcephaly was isolated for almost two-thirds of affected infants (62.8%, table 1). In the remaining third, heart defects were the most common associated anomaly (11.2%), most of which were non-critical. Skeletal defects (9.2%) were the next most common feature, followed by central nervous system defects (8.2%). Potentially lethal defects were present in 6.3% of microcephalic infants. Chromosomal defects were present in 5.9%, of which one-third were Down syndrome.
The strongest clinical risk factor for microcephaly was TORCH infection (table 2). In adjusted models, infants of women with TORCH infection had 32 times the risk of microcephaly (PR 32.38; 95% CI 22.42 to 46.75) compared with no infection. TORCH was associated with 118 excess cases of microcephaly per 10 000 infants (PD 118.21; 95% CI 75.85 to 160.58) compared with absence of TORCH. Exposure to teratogens due to epilepsy or alcohol and drug abuse was the next most important risk factor, with more than three times the risk of microcephaly (PR 3.10; 95% CI 2.37 to 4.07). Maternal congenital anomalies and errors of metabolism were associated with two times the risk of microcephaly (PR 2.11; 95% CI 1.46 to 3.05). Pre-existing diabetes increased the risk of microcephaly by only 49% (PR 1.49; 95% CI 1.00 to 2.22). Low parity and high material deprivation were the strongest demographical factors associated with microcephaly, and women in rural areas had lower risk compared with urban areas. Interaction terms were not statistically significant between any risk factors.
Postnatally, microcephalic infants had significantly unfavourable outcomes compared with non-microcephalic infants (table 3). Risk of mortality before discharge was over 20 times greater for infants with microcephaly (PR 20.52; 95% CI 15.57 to 27.04). Microcephaly was associated with 3.2 times the risk of very preterm birth (PR 3.19; 95% CI 2.23 to 4.56) and 3.1 times the risk of preterm birth (PR 3.05; 95% CI 2.66 to 3.50). Related outcomes such as neonatal intensive care unit admission, necrotising enterocolitis and bronchopulmonary dysplasia were all elevated. Risk of procedures such as intubation and transfusion was higher with microcephaly. Finally, microcephalic infants had a higher risk of hospitalisation for more than 7 or 14 days after birth. The mean length of stay was 8.6 days for microcephaly, but only 3.6 days for other neonates. C-section was the only outcome that was not significantly elevated in microcephalic infants.
There are major gaps in what is known about congenital microcephaly. This study provides a preliminary portrait of microcephaly in Quebec, a large population with nearly complete coverage of all neonates and no documented evidence of perinatal Zika infection during the period under study. In the absence of Zika, there were four microcephalic infants reported for every 10 000 live births, a prevalence comparable with other Western countries.8 9 The underlying aetiology of microcephaly was heterogeneous and two-thirds of cases were isolated. TORCH infection was the strongest maternal risk factor, followed by exposure to teratogens. Microcephalic infants had poor outcomes overall, but risk of mortality before discharge was particularly high. These data, to our knowledge, provide the first picture of clinically reported microcephaly in North America using a large population-wide database, demonstrating that microcephaly is prevalent, TORCH is a major risk factor, and that morbidity and mortality in microcephalic infants is significant.
Our results highlight the challenge of tracking microcephaly and its underlying aetiology.21 There currently is no standardised definition for reporting microcephaly.4 14 Clinically, the generally accepted guideline is that microcephaly consists of an occipitofrontal head circumference 2 SD below average for age and sex, with severe microcephaly below 3 SD.15 In hospital data, there is no way of knowing how physicians measure or report microcephaly, including which reference chart or threshold (2 or 3 SD) is followed.14–16 With 3 SD below the mean, the expected prevalence of microcephaly is around 10 per 10 000,22 a number higher than the 4 per 10 000 seen in our population and other countries.8 10 12 This suggests that reporting of microcephaly in Canada is reserved for the most severe cases, or that prenatal mortality due to termination or stillbirth results in fewer cases captured. Hospital discharge data therefore underestimate prevalence, as many cases may not be reported despite meeting the clinical definition of microcephaly. Despite this limitation, microcephaly in our data correlated strongly with major risk factors and birth outcomes, suggesting that those captured were true cases.
The causes of microcephaly are diverse, ranging from genetic and metabolic to structural and acquired.22–26 In our study, chromosomal defects were present in only 6% of infants with microcephaly, but this is probably an underestimate as use of chromosomal microarray is only recent. TORCH infections increased the risk of microcephaly more than 30 times despite wide rubella vaccine coverage. Although the absolute number affected was low, this risk aligns with preliminary evidence that Zika is associated with an average 20-fold higher chance of microcephaly in Brazil.27 28 Because hospital discharge summaries are based on ICD codes rather than laboratory data, we could not determine which specific TORCH agents were present in our population. With the bulk of evidence supporting Zika as a novel TORCH agent,27 better documentation of TORCH infections and travel is needed on discharge summaries should Zika become more frequent in Canadian women in the future.
There are few epidemiological studies of microcephaly.13 18 29 The focus of existing research has been on identifying coexisting anomalies and sometimes risk factors,13 18 29 with a lack of data on infant outcomes at delivery. In our data, microcephalic infants had very high mortality, but only a small proportion had potentially lethal congenital anomalies. The majority of microcephalic infants in fact presented without defects, resembling findings of previous research.13 29 Most morbidity was related to preterm birth and associated conditions, all which may contribute to increased mortality.
This study had several limitations. We relied on ICD codes to identify microcephaly, without data on head measurements. Despite not being able to confirm the presence of microcephaly, we found strong associations with risk factors and outcomes. Even with head measurements, the lack of a standardised definition complicates ICD reporting and has been an important limitation during the current Zika outbreak.4 We did not have information on microcephaly in stillbirths or pregnancy terminations. Although microcephaly may be difficult to identify in the second trimester, associated anomalies may be detected and result in pregnancy termination or stillbirth, increasing the proportion of isolated microcephaly at delivery. Hospital data do not perfectly capture exposures such as alcohol or drug use during pregnancy, and coding errors are possible. In 2006, hospitalisation data in Quebec transitioned from the 9th to the 10th revision of the ICD, potentially affecting recording over time. We did not have complete data on smoking or ethnicity, and cannot exclude residual confounding. We followed women over time, but could not do so for infants once discharged from hospital. Finally, we investigated a population representative of a large Canadian province, but generalisability to other regions requires more study.
This study provides an overview of the epidemiology of microcephaly in a large North American region before the Zika virus era. Prevalence, risk factors and outcomes of microcephaly in other countries are understudied and require greater attention. Measurement of microcephaly is challenging and hampered by lack of consensus on standardised diagnostic criteria. The current analysis shows that microcephaly is an important contributor to infant mortality and morbidity, and that the hospital discharge summaries are useful to detect severe cases and associated risk factors. More importantly, this study is a call for action for improved surveillance of microcephaly. The need for better data is only expected to grow in response to climate change and potential spread of novel viruses through international travel.
Contributors NA, JHP and LA contributed to study conception and design. NA and JHP analysed the data, and CQ and AML helped interpret the results. NA and JHP drafted the work, and all authors revised it critically for important intellectual content. All authors approved the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. NA is the guarantor. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Funding This study was funded by the Canadian Institutes of Health Research (MOP-130452) and the Fonds de recherche du Québec-Santé (career grants awarded to Auger 25128 and Quach 26873).
Disclaimer The funders did not participate in the study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The researchers are independent of the funders.
Competing interests None declared.
Ethics approval We obtained an ethics waiver from the review board of the University of Montreal Hospital Centre. The data were de-identified and the study conformed to requirements for ethical conduct of research on humans in Canada.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement No additional data are available.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.