A systematic review of the effects of maternal obesity on neonatal outcomes in women with gestational diabetes

Maternal obesity and gestational diabetes mellitus (GDM) prevalence are increasing, with both conditions associated with adverse neonatal outcomes. This review aimed to determine the risk of adverse outcomes in women with obesity and GDM, compared with women with obesity alone. A systematic search identified 28 eligible articles. Meta‐analysis was conducted using a random effects model, to generate pooled estimates (odds ratios, OR, or mean difference, MD). Compared with normal‐weight controls, women with obesity had increased risks of large for gestational age (LGA, OR 1.98, 95% CI: 1.56, 2.52) and macrosomia (OR 2.93, 95% CI: 1.71, 5.03); the latter's risk almost double in women with obesity than GDM. Birth weight (MD 113 g, 95% CI: 69, 156) and shoulder dystocia (OR 1.23, 95% CI: 0.85, 1.78) risk was also higher. GDM significantly amplified neonatal risk in women with obesity, with a three‐ to four‐fold risk of LGA (OR 3.22, 95% CI: 2.17, 4.79) and macrosomia (OR 3.71, 95% CI: 2.76, 4.98), as well as higher birth weights (MD 176 g, 95% CI: 89, 263), preterm delivery (OR 1.49, 95% CI: 1.25, 1.77), and shoulder dystocia (OR 1.99, 95% CI: 1.31, 3.03), when compared with normal‐weight controls. Our findings demonstrate that maternal obesity increases serious neonatal adverse risk, magnified by the presence of GDM. Effective strategies are needed to safeguard against neonatal complications associated with maternal obesity, regardless of GDM status.

associated with both GDM and maternal obesity.These include increased birth weight, jaundice, shoulder dystocia, neonatal intensive care unit (NICU) stays, and fetal death. 3,4[7][8] Despite the abundance of published data on how both conditions can contribute to adverse neonatal outcomes, it is not clear whether maternal obesity augments the adverse neonatal risks associated with GDM, or if obesity and GDM are independent risk factors for poor pregnancy outcomes. 9,10At the time of this review, there were no published systematic reviews on this subject.Therefore, the aim of this study was to systematically review the literature, to establish the difference in neonatal outcomes of women with obesity, with or without GDM, and where appropriate pool results to determine risk for specific outcomes.

| METHOD
This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 11The protocol for this systematic review is registered with PROSPERO (CRD42021260217).

| Search strategies
A systematic search of the databases MEDLINE and EMBASE via OVID, Scopus, CINAHL, and the Web of Science Core Collection was conducted.Search strategies were not restricted by language or year, and they were developed with the assistance of a medical librarian.A combination of subject-indexed terms and free text search terms in the title and abstract fields were used, covering GDM, maternal obesity, and neonatal outcomes.An example of search terms for MED-LINE is provided in the Supporting Information.

| Study selection
Two independent authors (T.L.W. and M.M.) reviewed the titles and abstracts to identify relevant studies and exported them to EndNote X7.All duplicated records were removed.Full-text articles were then exported to Covidence systematic review software for screening, against prespecified eligibility criteria.The references for the included studies were also reviewed, to identify any articles that were not identified in the preliminary search.Additional information was sought from study authors when full text was not available, or when information was required to resolve questions about eligibility.Any disagreement between study authors was resolved by discussion or by consultation with the study adjudicator (S.J.G.).
To be eligible for inclusion, studies needed to include women with maternal obesity, defined as a pre-pregnancy body mass index (BMI) above or equal to 30 kg/m 2 , as per WHO.Studies also needed to provide the criteria used to diagnose GDM.Women with GDM and maternal obesity formed the exposure group, and women with maternal obesity alone formed the control group.
Neonatal outcomes included large for gestational age (LGA; defined as birth weight above the 90th percentile for gestational age) and macrosomia (defined as birth weight > 4000 g), perinatal mortality, serious morbidity (e.g., birth trauma and shoulder dystocia), congenital anomalies including neural tube defects, small for gestational age (SGA; defined as birth weight below the 10th percentile), preterm delivery or premature birth (defined as birth prior to 37 weeks), neonatal hypoglycemia (defined as a blood glucose level < 2.6 mmol/L), and respiratory distress syndrome.Studies needed to include at least one outcome measure for inclusion.All studies needed to provide sufficient data for quantitative measurement of exposure and comparator group numbers and outcome measures.
The following exclusion criteria were applied: editorials, reviews, or comments; animal studies; studies involving participants < 18 years old; or those that exclusively studied in vitro fertilization (IVF) pregnancies, women who received bariatric surgery, or those with a diagnosis of polycystic ovary syndrome (PCOS).The decision to exclude studies exclusively dedicated to PCOS aimed to maintain a broader representation of the general population in our systematic review.By incorporating studies that encompassed both women with and without PCOS, our methodology sought to offer a comprehensive understanding of how obesity influences pregnancy outcomes across diverse populations.

| Study quality
Quality and risk of bias for each included study were independently assessed by the initial reviewers (T.L.W. and M.M.), using the quality appraisal tool Newcastle-Ottawa Scale (NOS), 12 as recommended in the Cochrane Handbook of Systematic Reviews.The NOS uses a star system, assessing three categories: selection of the study groups, comparability of the groups, and ascertainment of either the exposure or outcome.Studies score between 0 and 9.We defined scores between 0 and 3 as poor quality, scores of 4 and 6 as moderate quality, and scores between 7 and 9 as high quality.

| Data extraction
All data were extracted independently by both investigators in a predesigned form on Covidence; conflicts were resolved by consensus.
Data extracted from each study included first author, contact details, journal citation, year of publication, study design, sample size, study location and duration, participant characteristics, definition of GDM, cut-off values for BMI categories, statistical analysis used, and measures of association, including any reported odds ratios, relative risks, and confidence intervals between maternal obesity, GDM, and neonatal outcomes.

| Statistical analysis
Pooled statistics were reported as either odds ratio (OR) for categorical variables, or mean difference (MD) for continuous variables, with 95% confidence intervals (CI), using a random effects model.This model was chosen based on the heterogeneity between individual studies.Heterogeneity was quantified using Tau 2 , I 2 , and Q statistics.
We regarded heterogeneity as substantial if I 2 ≥ 50% and either Tau 2 was greater than zero, or if the Q statistics was significant (P < 0.05).
Random effects models weigh studies more equally and generally produce wider confidence intervals than fixed effect models, resulting in a more conservative estimate of obesity and GDM effect on outcomes.When 10 or more studies were available, publication bias was assessed using funnel plots and Egger's regression test.All analysis was performed in R Statistical Software, using the metafor package (Version 3.0-2). 13| RESULTS

| Data search results
The electronic database search identified 6,809 potential citations, and of these, 113 were selected for full-text review; 85 were excluded as they did not meet eligibility criteria, leaving 28 studies for final inclusion in the systematic review and meta-analysis (Figure 1).

| Study characteristics
The main characteristics of the studies are summarized in Table S1.
Out of the included studies, 20 were cohort studies (14 retrospective and 6 prospective), and 8 were case-controlled studies.Eleven were from Europe (United Kingdom, France, Finland, Sweden, Slovenia, Spain, Germany), nine were from the United States, five were from the Asia Pacific region (China, Australia), two were from Saudi Arabia, and one was an international study. 14

| Quality assessment
The rating of the included studies according to the NOS assessed quality based on selection, comparability, and outcome, shown in Tables S3 and S4.No studies were found to be of poor quality, three were found to be of moderate quality, and the remainder were of high quality.

| Large for gestational age
Seventeen studies reported on the risk of LGA.Two studies were excluded from meta-analysis, as outcomes were reported by BMI subclass or ethnicity. 29,37][18]20,22,23 The associated risk of LGA increased to over three-fold when both obesity and GDM were present (OR 3.22, 95% CI 2.17-4.79,P < 0.0001; Figure 4).Furthermore, meta-analysis of 15 studies demonstrated that women with both obesity and GDM had a higher risk of LGA offspring compared to women with obesity alone (OR 1.47, 95% CI 1.16-1.87,P = 0.0014; Figure 6).In this meta-analysis, women with either obesity or GDM demonstrated a slight, but non-significant, increased risk of LGA in women with obesity compared to those with GDM alone (OR 1.10; 95% CI 0.91-1.33; Figure 5).

| Macrosomia
17]28 However, unlike LGA, the risk of macrosomia appeared to be higher in women with maternal obesity, with an almost three-fold increased risk in women with obesity versus normal-weight women (OR 2.93, 95% CI 1.71-5.03,P < 0.0001; Figure 7), compared with a 1.4 increased risk in women with GDM F I G U R E 2 Forest plot of pooled effect estimates for large for gestational age in normal-weight, euglycemic women versus women with obesity.OR, odds ratio; CI, confidence interval.
F I G U R E 3 Forest plot of pooled effect estimates for large for gestational age in normal-weight, euglycemic women versus women with GDM.OR, odds ratio; CI, confidence interval.
F I G U R E 4 Forest plot of pooled effect estimates for large for gestational age in normal-weight, euglycemic women versus women with obesity and GDM.OR, odds ratio; CI, confidence interval.
F I G U R E 6 Forest plot of pooled effect estimates for large for gestational age in women with obesity versus women with obesity and GDM.OR, odds ratio; CI, confidence interval.
F I G U R E 5 Forest plot of pooled effect estimates for large for gestational age in women with GDM versus women with obesity.OR, odds ratio; CI, confidence interval.
F I G U R E 7 Forest plot of pooled effect estimates for macrosomia in normal-weight, euglycemic women versus women with obesity.OR, odds ratio; CI, confidence interval.
versus normal-weight women (OR 1.40, 95% CI 1.21-1.62,P < 0.0001; Figure 8).This increased risk was also seen when directly comparing macrosomia outcomes in women with obesity alone versus women with GDM alone (OR 1.81, 95% CI 1.67-1.97,P < 0.0001; Figure 10).When combined, maternal obesity was found to augment macrosomia risk in women with GDM (OR 3.71, 95% CI 2.76-4.98; Figure 9).Pooled estimates from 10 studies demonstrated a non-significant difference in macrosomia between women with obesity and GDM, and women with obesity alone (OR 1.11, 95% CI 0.75-1.63; Figure 11).Consistent with the trends observed in macrosomia, an analysis of 11 studies exploring the influence of obesity and GDM on birth weight revealed that neonates born to women with obesity alone were 113 g heavier than normal-weight women (95% CI 69-156, P < 0.0001; Table 1).When combined, women with both obesity and GDM had neonates 176 g heavier than normal-weight women (95% CI 89-263, P < 0.0001), mainly attributed to obesity.Women with obesity alone gave birth to neonates 94 g heavier than women with GDM alone (95% CI 41-148, P = 0.0006).This aligns with the observed patterns of macrosomia, reinforcing the significant impact of obesity, whether alone or in conjunction with GDM, on neonatal birth weight.

| Small for gestational age
Seven studies reported on SGA.After meta-analysis no results were found to be statistically significant (Table 1).However, there was a trend towards a positive association between SGA risk in women with GDM compared with normal-weight women (OR 1.07, 95% CI 0.38-3.01).This association was increased when both obesity and GDM were present (OR 1.25, 95% CI 0.41-3.79).

| Preterm delivery
7][28][29] Two were excluded from meta-analysis as one presented the results as obesity subclasses, and the other study did not provide absolute/ unadjusted numbers to allow for pooling of estimates. 10,29like fetal overgrowth conditions, when compared with normalweight women, women with obesity were not at an increased risk of preterm delivery (OR 1.06, 95% CI 0.91-1.24).In contrast, GDM was associated with a higher risk of preterm delivery compared with normal-weight women (OR 1.36, 95% CI 1.11-1.67,P = 0.0028; Table 1).Similar to other outcomes, the combination of obesity and GDM increased the risk to the neonate of preterm delivery (OR 1.49, 95% CI 1.25-1.77,P < 0.0001).The combination of obesity and GDM was also associated with a higher preterm delivery risk when compared with obesity alone (OR 1.20, 95% CI 1.06-1.35,P = 0.0032).
When compared against each other, women with obesity alone had F I G U R E 1 0 Forest plot of pooled effect estimates for macrosomia in women with GDM versus women with obesity.OR, odds ratio; CI, confidence interval.
F I G U R E 9 Forest plot of pooled effect estimates for macrosomia in normal-weight, euglycemic women versus women with obesity and GDM.OR, odds ratio; CI, confidence interval.
F I G U R E 1 1 Forest plot of pooled effect estimates for macrosomia in women with obesity versus women with obesity and GDM.OR, odds ratio; CI, confidence interval.

| Gestational age
Estimates of gestational age were pooled from seven out of nine studies.The remaining two papers reported measures as median and IQR, 25,26 and hence, they were excluded as the distribution of the original data was unclear.Pooled estimates from seven studies found that there was no significant difference in gestational age between the neonates of women with obesity and normal-weight women (MD 0.27, 95% CI À0.90 to 1.44).As with preterm delivery, women with GDM were more likely to give birth earlier than normalweight women (2.54 days, 95% CI 1.77-3.30,P < 0.0001).Women with GDM also delivered earlier than women with obesity alone (2.38 days, 95% CI 1.07-3.70,P = 0.0004).
Similar to other outcomes, when both obesity and GDM were present, there was an increased risk of earlier gestational age, with neonates of mothers with obesity and GDM born 3.40 days earlier than offspring of women with normal-weight women (95% CI 1.60-5.20,P = 0.0002).When compared with obesity alone, neonates of mothers with obesity and GDM were born 4.46 days earlier (95% CI 2.38-6.54,P < 0.0001).

| Shoulder dystocia/birth injury
There was an associated greater risk of shoulder dystocia in the neonates of women with both obesity and GDM, compared with the neonates of normal-weight women (OR 1.99, 95% CI 1.31-3.03,P = 0.001), and those of women with obesity alone (OR 1.57, 95% CI 1.04-2.35,P = 0.030) (Table 1).There was also a positive, but non-significant, association of shoulder dystocia in the neonates of women with obesity alone, compared with normal-weight women (97/3639 [2.7%] vs. 72/4025 [1.8%]; Tables 1 and S2).

| Congenital malformations
Of the four studies examining congenital malformations, only two provided sufficient data to be included for meta-analysis, with a nonsignificant increase in malformations seen in babies born to mothers with obesity and GDM, compared with obesity alone (OR 1.59, 95% CI 0.54-1.69,P = 0.3987).Further meta-analysis was unable to be performed due to insufficient study numbers and differences in reported measures.

| Respiratory distress syndrome
Though the data suggested that maternal obesity with or without GDM was associated with increased risk of RDS, after meta-analysis, none of these results were statistically significant (Tables 1 and S2).

| Neonatal hypoglycemia
There was a strong association between obesity and GDM and neonatal hypoglycemia risk, when compared with obesity alone (OR 5.13, 95% CI 2.60-10.12,P < 0.0001; Table 1).Only two studies reported on normal-weight women with GDM.Pooled estimates showed that women with obesity alone had a lower risk of neonatal hypoglycemia compared with women with GDM alone (265/6850 [3.9%]   vs. 233/2308 [10.1%],OR 0.25, 95% CI 0.07-0.91,P = 0.0348) (Tables 1 and S2).Notably, women with obesity and GDM had the highest rate, with 436 out of 2967 (14.7%) neonates experiencing hypoglycemia.Only one study reported on normal-weight women, so no meta-analysis was able to be performed.This study did report much lower rates of neonatal hypoglycemia 125/8573 (1.5%).

| Perinatal mortality
Meta-analysis was conducted on three studies reporting on perinatal mortality, with one study excluded, 10 as it did not provide unadjusted values to allow for comparison.Pooled results revealed an almost three-fold increased risk of perinatal mortality in women with obesity compared with normal-weight women with GDM (OR 2.85, 95% CI 1.04-7.82,P = 0.0417; Table 1).There was no significant difference in perinatal mortality in women with obesity and GDM versus obesity alone (OR 0.94, 95% CI 0.46-1.89).As with neonatal hypoglycemia, only one study reported perinatal mortality in normal-weight women, and thus, meta-analysis was unable to be performed.

| DISCUSSION
This systematic review and meta-analysis found that maternal obesity and GDM both independently increase the risk of LGA and macrosomia.Together, maternal obesity and GDM increase the risk of LGA and macrosomia by an alarming three-to four-fold, compared with normal-weight women.Compared with obesity alone, concurrent obesity and GDM are associated with an increased risk of LGA, but not macrosomia.Interestingly, the likelihood of macrosomia is significantly greater for obesity than for GDM (OR 2.93 vs. 1.40).This could be accounted for by earlier timing of delivery, given the higher rates of preterm delivery in women with GDM, regardless of concurrent obesity.As expected, given the known effects of maternal hyperglycemia on fetal hyperinsulinemia, GDM increased the risk of neonatal hypoglycemia.Importantly, perinatal mortality was higher in women with obesity compared to women with GDM, identifying women with obesity as an important group to target to prevent this catastrophic outcome.
As outlined above, one of the most interesting findings of this study is the differences seen in LGA and macrosomia by group comparison.Compared with obesity alone, LGA, but not macrosomia, was 1.5 times more common in women with obesity and GDM.Increased obstetric intervention in the form of earlier planned births in women with a diagnosis of GDM is known to reduce macrosomia and is the most likely reason why fewer cases of macrosomia were observed with obesity and GDM versus obesity alone. 38Preterm delivery was in fact 1.2 times more likely in women with obesity, if concurrent GDM was present.This phenomenon is likely driven by the increased frequency of comorbidities among women with obesity, such as hypertensive disorders, which drive an increase in both spontaneous and planned deliveries. 27though unnecessary obstetric intervention should be avoided, with planned births under 39 weeks associated with an elevated risk of poor child development, 39 our review suggests that this practice does reduce rates of macrosomia, a known risk factor for adverse delivery outcomes, and poor metabolic health outcomes in later life, including childhood obesity. 40Nonetheless, LGA was still more common in women with both conditions present, suggesting that intensive glycemic and metabolic management during pregnancy is required for this high-risk group.
There was a two-fold increase rate of shoulder dystocia in neonates of women with obesity and GDM, compared with normalweight women.In addition, despite the lower rate of macrosomia, shoulder dystocia was over 1.5 times more likely in women with obesity and GDM versus obesity alone.Previous studies confirm that the risk of birth injury increases with higher BMI categories. 10Unfortunately, few studies in this systematic review reported overall birth injuries, and so we were unable to make a distinction between risk of birth injury between groups.The rate of respiratory distress syndrome was similar between all group comparisons.We found that neonatal hypoglycemia was magnified over five times in women with obesity and GDM versus women with obesity alone.Conversely, women with obesity had a four-fold lower risk of neonatal hypoglycemia compared with those with GDM alone.Ascertainment bias may account for this difference to some degree, as most guidelines for the management of the neonate stipulate the need to check blood glucose levels in the neonate born to women with GDM. 41Nonetheless, our data support the known physiological impact of maternal hyperglycemia on neonatal hyperinsulinemia, resulting in transient neonatal hypoglycemia after birth. 42Moreover, our data support existing evidence indicating that obesity increases the likelihood of hypoglycemia in the neonate, with or without the presence of GDM. 42,43Possible mechanisms may relate to higher levels of free fatty acids and amino acids seen in women with obesity, which can influence fetal β-cell secretagogues and insulin secretion. 43Therefore, women with obesity and GDM should be a target group to prevent and monitor for this important neonatal complication.
Though the specifics of GDM management are hotly debated, for example, glucose levels for diagnosis, glucose targets for management, use of oral glucose-lowering agents, and timing of delivery, the diagnosis and management of GDM are known to protect the neonate from adverse outcomes, including perinatal death. 44Our study found that neonates of mothers with obesity were over 2.5 times more likely to die in the perinatal period compared with neonates of mothers with GDM.This in part may be related to the more intensive management women with GDM receive compared with their obese counterparts.
Early indicators for concern may also be more difficult, with surveillance of fetal growth and well-being less reliable in women with obesity.Other postulated explanations relate to the increased presence of comorbidities in the obese population, leading to higher rates of fetal hypoxia and placental atherosclerosis as a result of maternal metabolic derangement. 45ere was a non-significant trend for increased risk of SGA in women with both obesity and GDM, compared with obesity alone.
Increased study numbers, as well as adjustment for confounders including smoking and gestational weight gain, may provide further evidence in relation to this association into the future.Limited study numbers also prohibited a sub-analysis to explore the effect of different BMI subclasses, with prior studies demonstrating a U-shaped risk profile, with a higher risk of SGA in women at the extremes of BMI. 46,47

| Strengths and limitations
To our knowledge, this is the first systematic review to examine the impact of GDM on adverse neonatal outcomes, in women with obesity.We have summarized the results from both prospective and retrospective cohort and case-control studies, many with large sample sizes, with diverse racial and ethnic populations and countries, increasing the validity of our results.However, there are several limitations to our study.First, moderate to high heterogeneity was observed between studies.To mitigate this, we employed a random effects model rather than a fixed effects model, adjusting for between-study variance, weighing studies based on sample size, resulting in more conservative estimates, but producing more generalizable results.While addressing the observed heterogeneity, we conducted a sensitivity analysis by systematically removing two studies 14,19 that focused on pregnancy outcomes in untreated GDM populations.This strategic exclusion, however, did not yield a noteworthy alteration in the outcome measure, and the I 2 and Q statistics exhibited minimal change.Recognizing the inherent limitations of our meta-analysis, including variations in diagnostic criteria for GDM across studies, we attempted to assess the impact of changes in criteria over time.
Despite reordering studies based on the year of publication as a surrogate marker, this did not reveal discernible temporal trends in the reported outcome measures.
There were low study numbers for many pre-specified outcomes, impacting the robustness of results for SGA, birth injury, and congenital malformations.Interestingly, three studies on congenital heart disease risk reported a higher incidence of cardiac defects in neonates born to mothers with both obesity and GDM, compared with obesity alone. 29,34,35Likewise, Anderson et al. 36 found that women with both maternal risk factors also had higher rates of CNS defects.More robust data are required to pool and synthesize data in this area of great interest.
Moreover, limited data hindered our ability to conduct a thorough analysis of the effects of insulin treatment in women with GDM and its potential interaction with maternal obesity in influencing maternal outcomes.The lack of granularity in the available data also restrained our capacity to explore the nuanced interplay between treatment choices and outcomes.Specifically, information on insulin treatment and its potential impact on maternal outcomes in the context of maternal obesity was rarely reported across studies, highlighting a critical gap in the available literature.
Publication bias was assessed when there were at least 10 studies and funnel plots and Egger's regression test did not indicate bias.However, it is important to note that these methods do not entirely exclude the potential for publication bias.Finally, we were unable to conduct subgroup analyses for outcomes owing to the small number of available studies and incomplete data, despite attempts at contacting authors, prevented analysis on the impact of confounding variables, such as ethnicity, age, and socioeconomic background.In particular, the effect of ethnicity requires further analysis, with a disproportionate number of women from certain ethnic backgrounds, such as those of Asian descent, at a higher risk of GDM and adverse outcomes.Bowers et al. 37 demonstrated that although GDM and maternal obesity were jointly associated with an increased risk of LGA, the degree of risk varied across races.In addition, there exists valid concern regarding the universal application of the WHO definition of obesity as a BMI ≥ 30 kg/m 2 across all ethnic groups.
Although the use of ethnic-specific cut-offs may assist in determining an individual's level of risk, our diverse, multicultural world, can make categorizing BMI by race and ethnicity difficult.

| CONCLUSION
This systematic review highlights the significant impact maternal obesity places on the unborn child.Generally, current high-risk antenatal care focuses attention on women with obesity only when a diagnosis of GDM is also present, leaving women with obesity and a normal glucose tolerance test without crucial lifestyle advice and closer obstetric and medical monitoring that women with GDM often receive.With the growing burden of obesity in our society, it is time that we shift our focus to this group of at-risk individuals.In women with GDM, our review highlights the important comorbid influence of obesity on perinatal outcomes.Maternal obesity has a significant impact on perinatal outcomes, starkly transparent when perinatal mortality is compared in women with obesity versus GDM.Attention to interventional strategies that can improve the perinatal outcomes for pregnancies affected by maternal obesity is urgently needed to mitigate both short-and longer-term risks to the unborn child.
strated that both obesity and GDM are independently and similarly associated with a two-fold increased risk of LGA (OR 1.98, 95% CI F I G U R E 1 PRISMA flow diagram of the study selection process.T A B L E 1 Neonatal outcomes stratified by obesity and GDM status.Neonatal outcome Normal vs. obesity Normal vs. GDM Normal vs. obesity and GDM GDM vs. obesity Obesity vs. obesity and GDM AUTHOR CONTRIBUTIONS Tessa L. Weir and Sarah J. Glastras contributed to the study concept and design.Tessa L. Weir performed literature search.Tessa L. Weir and Monica Majumder selected studies, with Sarah J. Glastras settling any conflicts.Tessa L. Weir and Monica Majumder extracted data from relevant studies.Tessa L. Weir prepared initial figures and data analysis and wrote the first draft of the manuscript.All authors contributed to critically revised successive drafts of the manuscript and approved the final version.ACKNOWLEDGMENTS T.L.W. acknowledges stipend support from the Douglas and Lola Douglas Scholarship, and S.J.G. acknowledges fellowship support from the National Health and Medical Research Council.No funding sources were used in the design or conduct of this review.Open access publishing facilitated by The University of Sydney, as part of the Wiley -The University of Sydney agreement via the Council of Australian University Librarians.