Article Text
PDF
Abstract
Objective To explore associations between maternal body mass index (BMI) in early pregnancy and childhood infections.
Design Birth cohort study linked to primary care records.
Setting Bradford, UK.
Participants Live singleton births within the Born in Bradford cohort study between 2007 and 2011.
Exposures Maternal BMI in early pregnancy.
Main outcome measures The total number of infections between birth and ~14 years of age with subgroup analysis by infection type and age.
Results A total of 9037 mothers and 9540 children were included in the main analysis. 45% of women were of Pakistani ethnicity and 6417 women (56%) were overweight or obese. There was an overall trend for an increasing infection rate with increasing maternal BMI. In adjusted models, only those with obesity grade 2–3 had offspring with significantly higher rates of infection during the first year of life (RR 1.12 (95% CI 1.05 to 1.20)) compared with women of healthy weight. However, by age 5 to <15 years, children born to overweight women (RR 1.09 (95% CI 1.02 to 1.16)), obese grade 1 women (RR 1.18 (95% CI 1.09 to 1.28)) or obese grade 2 women (RR 1.31 (95% CI 1.16 to 1.48)) all had significantly higher rates of infection compared with those born to healthy weight mothers. Respiratory tract and skin/soft tissue infections made up the majority of excess infections.
Conclusions Maternal BMI was positively associated with rates of offspring infection in this study cohort, and suggests that we should be supporting women to achieve a healthy weight for pregnancy. Future research should investigate whether this is replicated in other populations, whether there is a causal association and the potential mechanisms and areas for intervention.
- Child Health
- Epidemiology
- Infectious Disease Medicine
- Obesity
Data availability statement
Data may be obtained from a third party and are not publicly available. Born in Bradford allows researchers to apply to access the study data through the Born in Bradford Executive Group. Researchers need to submit an EOI form to borninbradford@bthft.nhs.uk and the EOI will be reviewed at the monthly Born in Bradford Executive. More information about how to access Born in Bradford data can be found on the study website: https://borninbradford.nhs.uk/research/how-to-access-data/.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Increased inflammation in utero is associated with changes to the developing fetal immune system which may leave children at increased risk of infection.
Recent studies suggest that high body mass index (BMI) in pregnancy is associated with increased infection-related hospital admissions in early childhood.
WHAT THIS STUDY ADDS
Maternal obesity was associated with increased rates of less severe infections treated in primary care settings.
Respiratory and skin/soft tissue infections accounted for the majority of excess infections associated with maternal BMI.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Future research could focus on understanding whether this is a causal association.
This study adds to a growing body of literature to suggest that policies helping clinicians to provide adequate information and support on weight management from pre-conception to the postnatal period may be important for the longterm health of offspring.
Introduction
The proportion of obese women giving birth has more than doubled in recent decades and, in 2017, 27.4% and 21.6% of women giving birth in England were overweight or obese, respectively.1 There is growing evidence to support a link between maternal overweight and obesity in pregnancy and an increased risk of adverse long-term outcomes for offspring, such as obesity,2 metabolic and cardiovascular disease,3 4 poor developmental outcomes5 and asthma.6
There is also emerging evidence to suggest that maternal overweight and obesity during pregnancy is associated with an increased risk of offspring infections. Studies in Sweden7 and Australia8 reported an increased risk of infection-related hospital admissions during early childhood in those born to overweight or obese mothers compared with healthy weight peers. Multiple studies have looked at respiratory tract infections in children born to obese women compared with those born to healthy weight women and reported inconsistent findings.9–11
To date, studies within this area have either analysed hospital admission data or used parental self-report outcome measures and only followed children up to the age of 5 years, with few conducted in the UK. With rates of obesity continuing to rise, it is important to understand the long-term risk of infection associated with maternal overweight and obesity in pregnancy and to identify which infections are most strongly associated. The aims of this study were to describe the most common types of childhood infection associated with maternal body mass index (BMI) in pregnancy and to explore the associations between maternal BMI and offspring infections using data from the Born in Bradford cohort study linked to primary care data.
Methods
Population and data source
Born in Bradford is a multi-ethnic longitudinal UK birth cohort study that has followed up 12 500 families across Bradford since 2007.12 13 Bradford is a socially deprived and ethnically diverse city in the north of England. Women planning to give birth at the Bradford Royal Infirmary between 2007 and 2011 were eligible to participate. Women were recruited during their oral glucose tolerance test at ~26–28 weeks gestation, which is universally offered to all women in Bradford.
The Born in Bradford cohort data have been linked to maternity health records, hospital admission data (Bradford Royal Infirmary only) and primary care data for both mother and child. Women were included if they had a live singleton birth and available maternal height and weight data. Children were followed up from birth to 3 October 2022 or withdrawal. Women were excluded if they moved outside the Bradford Local Authority Area shortly after birth; had missing or implausible height, weight, or gestation at booking; or if their child had a condition unrelated to maternal BMI that left them at high risk of infection (see online supplemental table S1).
Supplemental material
Exposures
Maternal height was recorded at recruitment (~26 weeks gestation) by a research nurse. Maternal weight had been recorded by a midwife during the woman’s booking appointment (~10 weeks gestation). Weight recorded during the first trimester is considered representative of a woman’s pre-pregnancy weight.14 If weight was missing or recorded beyond the first trimester, BMI from 1 year prior to pregnancy or during the first trimester from primary care records was used. BMI was calculated by dividing weight by height squared (kg/m2) and categorised as underweight (<18.5), healthy weight (18.5–<25), overweight (25–<30), obese grade 1 (30–<35) and obese grade 2–3 (≥35). Because women of Asian descent experience metabolic disease at lower BMIs than white women,15 different definitions were used for women of Pakistani heritage (see online supplemental table S2).
Outcomes
An infection was defined as a primary care record with either (1) an infection-related diagnosis code (Read code (Clinical Term Version 3 (CTV3)) or (2) a prescription for an antibiotic.
Infections were grouped by diagnosis code and, in cases with no diagnosis code, the antibiotic prescription into the following categories based on existing literature16 17 (see online supplemental information B and C): upper respiratory tract infections (URTIs), lower respiratory tract infections (LRTIs), skin and soft tissue (SST) infections, genitourinary, gastrointestinal and multisystem viral infections. Multiple infections recorded in the same GP visit or day were only counted once within the total infections, but both were counted in their respective infection categories (Sections S2–S3). Cases with no diagnosis code and a broad spectrum antibiotic prescription were only counted within the total infection category and no specific infection type was identified.
Supplemental material
Supplemental material
Statistical analysis
Maternal and child characteristics were compared across maternal BMI categories using χ2 tests. Person-years-at-risk (PYR) for each child were calculated as time from birth to study end or withdrawal. Crude infection rates per 100 PYR were calculated for each category of maternal BMI and repeated for each age band (<1, 1–<2, 2–<5, 5–15 years).
Negative binomial regression models were used to estimate rate ratios (RR) and 95% CIs for infections within each category of maternal BMI (healthy weight as reference category). Models were clustered by mother to account for correlation between siblings. The models were adjusted for confounders identified within the Directed Acyclic Graph including maternal age, parity, socioeconomic position, index of multiple deprivation, maternal education, ethnicity and smoking during pregnancy (see online supplemental figure S1). Confounders were grouped and models were adjusted sequentially. To explore if the rate of infection changed over time, an interaction term between BMI and child age was included and the Wald test was performed to assess significance. The analysis was then repeated for each age group and infection type. Due to low counts in some infection categories, BMI was also explored with only four categories (underweight, healthy weight, overweight and obese).
Multiple sensitivity analyses were performed: (1) using ‘western’ BMI cut-offs for Asian women; (2) including maternal weights measured within the first 16 weeks of pregnancy; (3) including all women with available weight measures, adjusted for expected weight gain by gestational age at booking; (4) excluding high-risk pregnancies; (5) adjusting for maternal health-seeking behaviour; (6) restricting analysis to fully vaccinated children (see online supplemental table S3); (7) adjusting for potential mediators (mode of birth, gestational age at birth, child BMI (4–5 years) and breastfeeding); (8) restricting infections to those with a diagnosis code only (see online supplemental table S4); and (9) exploring coding differences with commonly prescribed multi-use antibiotics (Section S5).
Data were evaluated for patterns of missing data and deemed missing at random. Missing data were imputed using chained equations with 20 imputations and results were combined using Rubin’s rules.18 All variables in the main analysis model were included in the imputation model, with the addition of the following auxiliary variables: breastfeeding, gestational age, mode of birth and child sex. Distributions of imputed and observed data were similar (see online supplemental tables S5 and S6). A complete case analysis was conducted as a sensitivity analysis. All analyses were conducted in Stata15.
Results
A total of 12 453 women were recruited which, after exclusions, resulted in 9037 women and 9540 children within the main analysis cohort (figure 1).
Flow chart of sample population. aMaternal weight adjusted based on gestation at booking (see Appendix X for more detail). bMaternal weight measured within the first 16 weeks of pregnancy. cMaternal weight measured within the first trimester and this will be the sample used for the main analysis. BMI, body mass index; LAA, Local Authority Area.
Table 1 shows the characteristics of children in the main analysis cohort (n=9540). The highest proportion were born to mothers of Pakistani ethnicity (45%), followed by white British children (40%). 5% of children had underweight mothers, 30% had overweight mothers and 26% had obese mothers. Underweight mothers were more likely to be <30 years old, of Pakistani ethnic origin, living in socially deprived areas, having their first baby and to have given birth vaginally. Mothers who were obese were more likely to be older, higher parity, of Pakistani ethnic origin, of lower socioeconomic position and to have given birth via caesarean section.
Descriptive statistics by maternal BMI for the main analysis cohort (N=9540)
There were 135 582 infection events recorded in primary care during 119 468 PYR of follow-up (mean follow-up 12.6 years and mean infection rate of 1.14 /year). The number of infections per child ranged from 0 to 221 (median 11, IQR 5–18) between birth and 15 years. Almost all (94%) children had evidence of at least one infection in their primary care records. The highest rates of infections were seen at <1 year and decreased with age (see table 2 and online supplemental table S7).
Crude rates of infections stratified by age group and maternal body mass index
There was evidence of an interaction (p<0.0001) between maternal BMI and infections over time. RRs and 95% CIs for childhood infections by maternal BMI and child age are shown in figure 2. There was an overall trend for increasing infection rates with increasing maternal BMI; this relationship appeared to strengthen with increasing age of the child. In the fully adjusted models, only women with obesity grade 2–3 had offspring with significantly higher infection rates at <1 year (RR 1.12 (95% CI 1.05 to 1.20)) compared with women of healthy weight. However, by age 5–<15 years, children born to overweight women (RR 1.09 (95% CI 1.02 to 1.16)), obese grade 1 women (RR 1.18 (95% CI 1.09 to 1.28)) or obese grade 2 women (RR 1.31 (95% CI 1.16 to 1.48)) all had significantly higher rates of overall infection. These trends were consistent across all sensitivity analyses (see online supplemental tables S8–S15).
Unadjusted and adjusted* IRRs and 95% CI for maternal body mass index (BMI), total infection, stratified by age. Healthy BMI is the reference category. *Adjusted for maternal age, parity, ethnicity, smoking in pregnancy, socioeconomic position, maternal education and Index of Multiple Deprivation score. SES, socioeconomic position + maternal education + Index of Multiple Deprivation score.
URTIs were the most common infection, with the highest rates at <1 year (105/100 PYR), followed by SST infections (88.9/100 PYR). However, diagnosis codes for SST infections were less commonly recorded, with most identified through prescription data only (see online supplemental tables S16 and S17). Similarly, when amoxicillin prescriptions were considered to indicate either URTIs or LRTIs, the rates of these two infections increased significantly (see online supplemental tables S18 and S19).
Adjusted RRs and 95% CIs for specific infection types are shown in figure 3. A positive association between maternal BMI and subgroups of infection was seen in URTIs, LRTIs and SST infections only, with the highest rates seen between the ages of 5 and <15 years. LRTIs were most strongly associated with maternal obesity, with RRs of 1.19 (95% CI 1.03 to 1.37) and 1.39 (95% CI 1.16 to 1.67) at <1 year and by 5–<15 years, respectively. Interestingly, compared with healthy weight women, being born to an underweight mother was associated with lower rates of genitourinary infections in children <5 years old. The results remained similar in the complete case analysis (see online supplemental tables S20 and S21).
Adjusted* IRRs and 95% CIs for subgroups of infection stratified by age. Healthy body mass index (BMI) is the reference category. *Adjusted for maternal age, parity, ethnicity, smoking in pregnancy, socioeconomic position, maternal education and Index of Multiple Deprivation score.
Discussion
This study explored the association between maternal BMI in early pregnancy and offspring infection from birth to approximately 14 years of age. We observed a small but consistent positive association between maternal BMI and childhood infection. While the crude rates of infection decreased with age, the association with BMI was strongest after 5 years of age. The majority of excess infections in children born to women with obesity were URTIs, LRTIs and SST infections.
Our findings are in keeping with existing literature. A study conducted in Sweden found that children of overweight or obese mothers had increased infection-related hospital admissions up to the age of 5 years.7 Another study conducted in Israel found that maternal obesity was significantly associated with offspring infection-related hospital admissions from birth to 18 years.19 The increased inflammatory state that obesity creates, which then modulates the development of the fetal immune system,20 21 may be responsible for this relationship. In addition, maternal obesity is associated with alterations to the microbiome of the infant, which can impact immune function.22
We found respiratory infections were associated with maternal BMI and accounted for the majority of excess infections. This was also reported by authors of a UK-based prospective cohort study9 and the Swedish study based on hospital admissions.7 In contrast, a study conducted in Norway reported no relationship; however, exposures and outcomes were self-reported.11 SST infections were positively associated with maternal BMI in our study, in contrast to the Swedish hospital-based study.7 Given that most SST infections will be minor, this may account for the variation in the findings. In a sensitivity analysis we found most SST infections had a topical antibiotic prescription rather than a diagnosis code, suggesting they were minor in nature.
Maternal underweight during early pregnancy was associated with lower rates of genitourinary infections at <5 years of age, in contrast to existing literature.7 This result might be due to an under-representation of infections within this group23 based on their characteristics.
The strength of the relationship between BMI and infection appeared to increase with childhood age and remained a consistent relationship in all sensitivity analyses. Maternal BMI is strongly associated with socioeconomic position in women,24 so it is possible that this relationship is due to either residual confounding or the cumulative impact of lower socioeconomic position. The cumulative inequality theory suggests that the impact of socioeconomic position may increase as children age.25 Similar trends were found in relation to maternal BMI and childhood wheeze,6 and in a study exploring breastfeeding and educational outcomes.26
Strengths and limitations
This cohort is from a UK city with high social deprivation and a high proportion of patients of Pakistani ethnicity, providing a unique insight into associations between different social and health-related characteristics. The longitudinal cohort data with high coverage and reliability recorded through well-documented routine healthcare in pregnancy, including rich data on ethnicity and socioeconomic position, is a key strength. The long-term follow-up of children meant we were able to look at rates of infection into mid-adolescence and stratify by age. Extensive linkage to external data sources across primary and secondary care also means we had access to objective measures of infection, particularly the addition of prescription data which enabled more inclusive estimates of infection due to the documented variation with coding infections in primary care records.27
With regard to limitations, there was variation in when women attended their first antenatal appointment and weight during the first trimester is only a proxy for pre-pregnancy weight. Women gain approximately 0.5–2 kg during the first trimester,28 so BMIs around boundaries may be misclassified. While antibiotic prescription coding is automatic when a prescription is issued, disease coding within consultations can be variable.29 Therefore, mild self-limiting infections not requiring a prescription may have been missed. Antibiotic prescribing behaviour for childhood infections is complex and influenced by multiple factors, including socioeconomic position, language barriers and level of social support,30 and prescribing patterns in and within Bradford may not reflect UK-wide patterns. As UK primary care is free at the point of use, there is potential for parental health-seeking behaviour causing informed presence bias (if someone is more likely to visit their GP, they are more likely to have a recorded diagnosis code).31 We included the rate of maternal GP visits per year as a proxy for this in the sensitivity analysis, but it is a relatively crude marker. We were unable to account for some potential confounders and mediators (eg, maternal and child nutrition). While we adjusted for socioeconomic position, it is also possible that residual confounding may be at play.
Implications
While the findings suggest that there is a modest effect of BMI, if causally related it could potentially impact large numbers of women and children. Therefore, the findings highlight the need for women of reproductive age to ensure they have a healthy body weight before conception. This could include education and support by healthcare professionals and public health campaigns to help them achieve this.32 It is also important to recognise that many healthcare professionals may need training in how to do this, due to the sensitive nature of bodyweight.33 Further research is needed to fully explore the potential mechanisms and identify areas for intervention if women with obesity become pregnant.
Conclusions
Obesity during early pregnancy was consistently associated with a modest increased risk of infection in offspring. Respiratory and SST infections made up the majority of excess infections. Women should be given advice and support to help them achieve a healthy weight for pregnancy. Future research should investigate potential mechanisms and areas for intervention from pre-conception to the postnatal period.
Data availability statement
Data may be obtained from a third party and are not publicly available. Born in Bradford allows researchers to apply to access the study data through the Born in Bradford Executive Group. Researchers need to submit an EOI form to borninbradford@bthft.nhs.uk and the EOI will be reviewed at the monthly Born in Bradford Executive. More information about how to access Born in Bradford data can be found on the study website: https://borninbradford.nhs.uk/research/how-to-access-data/.
Ethics statements
Patient consent for publication
Ethics approval
Bradford Research Ethics Committee approved the following studies that form Born in Bradford: 07/H1302/112, 15/YH/0455, 17/YH/0202. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We are grateful to the following organisations and individuals who were involved in this study: The Nuffield Department of Population Health for funding this project; Born in Bradford for collecting, linking and providing access to the data; and the Patient and Public Involvement (PPI) group who provided feedback on the initial concept and design of the project. Finally, Born in Bradford is only possible because of the enthusiasm and commitment of the children and parents in BiB. We are grateful to all the participants, health professionals, schools and researchers who have made Born in Bradford happen. In addition to this, we gratefully acknowledge the contribution of TPP and the TPP ResearchOne team in completing study participant matching to GP primary care records and in providing ongoing informatics support.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors VC designed the study with input from HFA, CC, MAQ and GS. VC and GS were responsible for the acquisition of the data. VC was involved in data cleaning with input from CC, GS and MAQ. Statistical analysis was performed by VC. VC, CC, MAQ, HFA and GS were all involved in the interpretation of the data. VC was responsible for the initial draft of the manuscript. MAQ, HFA, GS and CC reviewed and contributed to drafts of the manuscript, and all authors have reviewed the final version. VC is the guarantor of this study.
Funding This research was funded by an Intermediate Research Fellowship from the Nuffield Department of Population Health at the University of Oxford, UK (H6D00010). The fellowship was awarded to the lead author, VC. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.