Risk factors for preterm labor: An Umbrella Review of meta-analyses of observational studies

Preterm birth defined as delivery before 37 gestational weeks, is a leading cause of neonatal and infant morbidity and mortality. Understanding its multifactorial nature may improve prediction, prevention and the clinical management. We performed an umbrella review to summarize the evidence from meta-analyses of observational studies on risks factors associated with PTB, evaluate whether there are indications of biases in this literature and identify which of the previously reported associations are supported by robust evidence. We included 1511 primary studies providing data on 170 associations, covering a wide range of comorbid diseases, obstetric and medical history, drugs, exposure to environmental agents, infections and vaccines. Only seven risk factors provided robust evidence. The results from synthesis of observational studies suggests that sleep quality and mental health, risk factors with robust evidence should be routinely screened in clinical practice, should be tested in large randomized trial. Identification of risk factors with robust evidence will promote the development and training of prediction models that could improve public health, in a way that offers new perspectives in health professionals.


Introduction
Preterm Birth (PTB) is de ned as delivery before 37 gestational weeks and is a leading cause of infant morbidity and mortality [1][2][3][4]. 15 million babies are estimated to be born preterm every year and the PTB rate ranges between 5-18% worldwide [3] (PTB rates in USA:12-13% [1,2]; in Europe is 5-9% [2]). Advances in neonatology and the administration of corticosteroids before birth have improved signi cantly the prognosis of babies born preterm. In contrast, although vigorous research, costing millions of dollars, was carried out during the last 40 years, focusing in the prediction and prevention of preterm birth its incidence remains relatively unchanged. The most probable explanation is that preterm birth is a syndrome, rather than a single disease and many different causes may be responsible [161].
Numerous systematic reviews and meta-analyses have assessed various, non-genetic risk factors of preterm labor. Several environmental and clinical parameters such as present pregnancy characteristics, previous pregnancy history [4], infections [7,8], environmental exposures, pharmaceutical factors [9,10] and surgical interventions have been proposed as plausible factors related to PTB. Identifying robust risk factors for PTB should either help us de ne a study population for speci c interventions, allocate available resources effectively and allow risk-speci c treatment and understanding the mechanism leading to PTB [1]. However, the exact causes of this syndrome are still mostly unknown and the contribution of every risk factor in terms of prevention is still questionable.
To our knowledge there is no previous effort to summarize existing evidence of meta-analyses of non-genetic risk factors for PTB. We conducted an umbrella review across published meta-analyses of observational studies with the goal to map the existing evidence and critically evaluate the reported associations applying stringent criteria that assess potential systematic biases and we highlight previously studied associations that provide robust evidence of association.

Description of Eligible meta-analyses
The search identi ed 2769 items, of which 2239 were excluded after review of the title and abstract (Fig. 1, PRISMA Flowchart). Of the remaining 530 articles that were reviewed in full text, eight articles did not report the appropriate information for the calculation of excess of statistical signi cance (either because the total sample size was missing or the study-speci c relative risk estimates were missing), and 98 articles were excluded because a larger systematic review or meta-analysis investigating the same risk factor was available. From the 223 comparisons, we further excluded the ones that included one or two studies (53 comparisons). Therefore, 219 articles were analyzed, of which 133 were systematic reviews without any quantitative component and 86 were meta-analyses. The 86 eligible meta-analyses [116][117][118] included data on 170 comparisons and 1511 primary studies.

Summary Effect-sizes And Signi cant Findings
Three to 152 studies, with a median of nine studies, were included per meta-analysis. The median number of case and control subjects in each study was 88 and 529, respectively. The median number of case and control subjects in each meta-analysis was 98 and 807, respectively. The number of cases was greater than 1000 in 98 comparisons. Overall, 578 (45%) individual studies observed nominally statistically signi cant results. 40 meta-analyses used the Newcastle-Ottawa Scale to assess qualitatively the included primary studies. One meta-analysis used assessment criteria for non-randomized observational studies adapted from Duckitt and Harrington, 3 meta-analyses used the Methodological Index for Non-Randomized Studies (MINORS) and 38 meta-analyses used other assessment tools. Four meta-analyses did not perform any quality assessment. Details of the 170 comparisons that included 1511 individual study estimates are summarized in Supplemental Table 1. encourages women who experienced a previous miscarriage to wait for a minimum of 6 months before the next conception to achieve optimal outcome and reduce obstetric complications such as preterm birth [119]. Contrary to the ndings of the research on which WHO based its recommendations, some studies reported that the risk of adverse obstetric outcomes including preterm birth is lower in women who conceived less than 6 months after a pregnancy loss [120,125,126], while synthesizing all available data provided the same conclusion [105]. This metaanalysis included eight studies, performed two analyses: one including the study of Conde Agudelo 2004 [121] and one excluding it, and robust results were obtained after excluding the study. While this was a large retrospective study on which the WHO guidelines for delaying pregnancy for at least 6 months [119] are based, it did not differentiate between induced and spontaneous abortions and used data from many countries where induced abortion is illegal [121], therefore should be interpreted with caution. After a miscarriage, there is a very small burden on the folate reserve and thus miscarriage is not very likely to lead to folate de ciency in the postpartum period, so miscarriage and delivery later in pregnancy can have differential effects on subsequent pregnancy. This could explain the reduced risk of adverse outcomes in a short IPI after a miscarriage [122] but not after delivery. In support of this hypothesis, there is evidence to suggest that late miscarriages (after 12 weeks of gestation) are associated with worse outcomes in the subsequent pregnancy [123]. In addition, most women who attempt another pregnancy soon after a miscarriage are likely to be motivated to take better care of their health and consequently result in better pregnancy outcomes [124]. Another plausible reason may be that those who conceive soon after a miscarriage are naturally more fertile and younger and consequently have better pregnancy outcomes.
Another association with robust evidence was pregnant women with sleep breathing disorders. This meta-analysis clearly demonstrated the increased risk pro le of women who experience SBD not only for preterm birth but for other pregnancy outcomes. Regarding plausible mechanisms, the association between SDB and intermittent maternal hypoxia as well as the link with conditions synonymous with impaired placental function such as pre-eclampsia suggest a multifactorial cause, with both physiologic changes associated with pregnancy and placental dysfunction involved. This robust association has clear implications for obstetric practice. First, given the rapidly increasing worldwide obesity rates, SDB is likely to become more prevalent in the pregnant population and is worthy of being screened for. Second, the increased risk for both adverse intrapartum and perinatal outcomes demonstrated in this review strongly support the need for increased surveillance of this cohort. Third, public health education programs must take into account the speci c maternal and perinatal risks and promote education about the signi cance of obstructive sleep apnea symptoms and the need for women to discuss this with their obstetric caregivers. In alignment with this suggestion, women with personality disorders could be identi ed early through mental health screening, where targeted health interventions and multidisciplinary management can be implemented in order to reduce poor outcomes for the baby/child and woman. This early identi cation and support also have the potential to enable the prevention of maladaptive development trajectories within the mother infant relationship [128,129]. Regarding induced termination of pregnancy with vacuum aspiration, our results should be interpreted with caution because it is unclear whether two of the ve included studies come from the same population, therefore the variance of the pooled estimate may be arti cially narrower.
Furthermore, it is important that clinical examination and medical history includes risk factors which are not well known, identi ed in metaanalysis with highly suggestive evidence. To be more speci c regarding highly suggestive evidence, there were a few that are well known and used to classify pregnancies as high risk for PTB such as therapies for cervical intraepithelial neoplasia, advanced maternal age, placental pathology, race, rst trimester bleeding and maternal comorbidities. There were also included factors that are not routinely screened in the obstetric population such as intimate partner violence, cancer survivors and being unmarried. When it comes to intimate partner violence exposure during pregnancy, this meta-analysis included 30 studies examining the risk of PTB [28]. Two possible pathways have been described which could lead to adverse perinatal outcomes [140,141]. One is the direct exposure to violence consisted of either physical assault directly to the abdomen or sexual abuse. Direct exposure has been associated with pregnancy complications such as premature rupture of membranes, uterine contractions and placental damage, too [140,142]. On the other hand, indirect exposure to violence trigger biological mechanisms, such as smoking, alcohol or drug use, inadequate prenatal care and weight gain that contribute to adverse birth outcomes [140,[145][146][147][148][149][150][151][152][153][154][155][156]. Women with history of abuse by their partner is believed to have less support, lower levels of self-esteem and higher levels of stress, too [140,142,145,153,160]. All these factors contribute to the indirect mechanism "theory" associated to preterm birth. As a result, healthcare professionals/institutes follow screening protocols in some nations or clinical guidelines, in order to detect and take care of these cases [157][158][159]. Another association that demonstrates highly suggestive evidence is pregnant women, whom survived cancer. This meta-analysis included fourteen studies which described the incidence of PTB [30]. Regarding plausible mechanisms, it is believed that radiotherapy treatment protocols for cancer, especially irradiation of the abdomen is harmful both for the uterine vasculature and the uterus muscular development. This leads to a reduction in uterine elasticity and uterine volume [143,144]. Uterine volume can also be smaller due to hormonal de ciency, caused by ovarian failure [144]. This could lead to preterm delivery. However, there is a possible association between the dosage of radiotherapy and risk of PTB, something which still has not been examined due to the obscuring of pooling dosages in previous studies. Higher radiations doses may re ect to higher risk of PTB. In addition, we should highlight the fact that the population of cancer survivors following advancing treatment grows and the prevalence of PTB cases in these groups is going to rise, in regards. Maternal marital status plays also a role in PTB, but healthcare professionals rarely consider it as a risk factor. This meta-analysis consists of 21 studies comparing unmarried women to married ones, identifying an increased risk of PTB [43]. Regarding the ways in which unmarried women are associated to PTB, it is suggested that the quality of relationship between biological maternal and paternal gures is more important than their legal status [136,137]. Moreover, a biological father might be more caring or supportive of the birth compared to another family member or partner. Mother psychosocial stress level depends on the support that she receives from her familiar environment [138, 139], but a variety of other factors should be taken into consideration before interpreting these results. With regards to health practitioner's point of view, the importance of obtaining social history information during clinical exam, lies in identifying pregnancies at risk for PTB and offering new perspectives. This information should be focused on rarely screened factors in every-day routine, which support highly suggestive evidence.
Regarding environmental risk factors, increased residential greenness was associated with a protective effect on the risk of PTB. Although this nding was categorized as having suggestive evidence, the p-value of the random effect estimate was very close to the stringent threshold of < 10 − 6. Acknowledging the detrimental projected effect of climate change in greenness and given that it is one of the few protective risk factors for PTB, serious efforts should be made to maintain and grow residential greenness. Possible mechanisms include among others amelioration of the effects of air pollutants, reduction of stress and increase in physical activity [116]. There were also suggestive evidence for early pregnancy exposure to PM 2.5 and the risk of PTB. This association has been debated in the literature with con icting results about the timing and magnitude of effect and is less robust than other associations that have been shown to have strong evidence for associations [130] such as birthweight.
In the current umbrella review, we applied a transparent and replicable set of criteria and statistical tests to evaluate and categorize the level of existing observational evidence. Although, 58,8% of associations in the included meta-analyses report a nominally (P < 0.05) statistically signi cant random-effects summary estimate, when stringent P value was considered (P < 10 − 6 ), the proportion of signi cant associations decreased to 24,1%. 94 (55,3%) associations had large or very large heterogeneity, while when we calculated the 95% prediction intervals, which further account for heterogeneity, we found that the null value was excluded in less than half of the associations. Only seven (4.1%) of the assessed risk factors found to provide robust evidence, indicating that several published meta-analyses of observational studies in the eld could be susceptible to biases and the reported associations in the existing studies are often exaggerated.
The ability to modify those factors, mainly those related to mental health and sleep quality screening, through screening and clinical interventions or public health policy measures remains to be established. Furthermore, there is no guarantee that even a convincing observational association for a modi able risk factor would necessarily translate into large preventive bene ts for preterm birth if these risk factors were to be modi ed [93]. With obesity becoming a global epidemic, the assessment of the strength of the evidence supporting the impact of overweight and obesity in sleep breathing disorders could allow the identi cation of women at high risk for adverse outcomes and allow better prevention. Obesity is generating an unfavorable metabolic environment from early gestation; therefore, initiation of interventions for weight loss during pregnancy might be belated to prevent or reverse adverse effects, which highlights the need of weight management strategies before conception [68, 103,104,131]. PTB does not only increase the risk for maternal and infant complications, but also signi cantly increases a woman's risk of cardiovascular disease (CVD) after pregnancy, therefore primary prevention [12,[132][133][134] is extremely important.
Our assessment has certain limitations. Umbrella reviews focus on existing systematic reviews and meta-analyses and therefore some studies may have not been included either because the original systematic reviews did not identify them, or they were too recent to be included. In the current assessment we used all available data from observational studies, therefore the meta-analysis estimates may partly re ect the biases from which the original studies suffer from. Statistical tests of bias in the body of evidence (small study effect and excess signi cance tests) offer hints of bias, not de nitive proof thereof, while the Egger test is di cult to interpret when the between-study heterogeneity is large. These tests have low power if the meta-analyses include less than 10 studies and they may not identify the exact source of bias [23,25,135]. More speci cally, in our study, all robust evidence applied to meta-analyses with less than 10 studies, therefore the results of publication bias should be interpreted with caution. Furthermore, we did not appraise the quality of the individual studies on our own, since this should be included in the original meta-analysis and it was beyond the scope of the current umbrella review. However, we recorded whether and how they performed a quality assessment of the synthesized studies. Lastly, we cannot exclude the possibility of selective reporting for some associations in several studies. For example, perhaps some risk factors were more likely to be reported, if they had statistically signi cant results.

Conclusion
The present umbrella review of meta-analyses identi ed 170 unique risk factors for preterm birth. Our analysis identi ed seven risk factors with robust evidence and strong epidemiological credibility pertaining to isolated single umbilical artery, amphetamine exposure, maternal personality disorder, sleep breathing disorders, induced termination of pregnancy with vacuum aspiration, low gestational weight gain and interpregnancy interval following miscarriage of less than 6 months. As previously suggested, the use of standardized de nitions and protocols for exposures, outcomes, and statistical analyses may diminish the threat of biases, allow for the computation of more precise estimates and will promote the development and training of prediction models that could promote public health.

Methods
We conducted an umbrella review which is a comprehensive and systematic approach that collects and critically evaluates all systematic reviews and meta-analyses performed on a speci c research topic [11]. We used previously described, standardized methods that have been already used in previously published umbrella reviews referring to risk factors related to various outcomes [13][14][15][16] and have been elaborated below.
A protocol for this umbrella review was registered in the International prospective register of systematic review (PROSPERO 2021 CRD42021227296)

Search Strategy
Two researchers (A.E., I.M.) independently searched PubMed database from inception to December 2020, in order to identify systematic reviews and meta-analyses of studies that examine the association between risk factors and preterm birth. The search strategy included combinations of the Medical Subject Headings (MESH) terms, key words and word variants for terms "preterm birth" AND ("systematic review" OR "metaanalysis"). Titles and abstracts were screened and potentially eligible articles were retrieved for full text evaluation. A detailed description of our search strategy is provided in the supplement (Supplemental Table 3).

Eligibility Criteria And Data Extraction
We included systematic reviews with meta-analyses investigating the association between various types of exposures and PTB. Speci cally, we included studies with singleton pregnancies and studies where PTB was evaluated as primary outcome.
Case report or series and individual participant data meta-analyses were excluded. We also excluded studies that set time limits on time span or were performed on a restricted setting (i.e. conducted for one speci c country). Furthermore, we excluded studies that assessed PTB as a secondary outcome, studies including multiple pregnancies, and studies that assessed genetic or over -omics features as risk factor for PTB. All studies were compared to avoid the possibility of duplicate or overlapping samples. If more than one meta-analysis referring to the same research question were eligible, the one with the largest amount of component studies with data on individual studies' effect sizes retained for the main analysis Publications whom the estimates of the studied associations, such as relative risks (RR) and 95% con dence intervals (CIs), were not reported or could not be retrieved/calculated were excluded from the analysis. For the non-environmental risk factors, we also excluded meta-analyses that did not provide the number of cases in the exposed and non-exposed groups, which is used for the calculation of the excess signi cance tests.
For the environmental risk factors, since most commonly they report the results as per unit(s) increase in exposure and everyone is exposed, we included them even if they did not report the number of cases and total sample size.
Eligible articles were screened by four independent reviewers (AE/IM and EB/TK). Any disagreement between reviewers was resolved by consensus or after evaluation of a third author (SP or EE). The data of eligible studies were extracted in a prede ned data extraction form recording for each study the rst author, journal, year of publication, the examined risk factors and the number of reviewed studies. Either the study speci c relative risk estimates (risk ratio, odds ratio, hazard ratio, incidence rate ratio) and the con dence intervals were extracted or the mean and the standard deviation for continuous outcomes were also noted in this form. We also extracted exposed and control group used; outcome assessed; study population; exposure characteristics; number of studies in the meta-analysis; meta-analysis metric and method; effect estimate with the corresponding 95% con dence interval; number of cases and total sample size; I2 metric and the corresponding χ2 p-value for the Q test; and Egger's regression P-value.

Assessment Of Summary Effect And Heterogeneity
We re-calculated summary effects and 95% Con dence Intervals (CIs) for each meta-analysis via xed and random effects model [17,18]. 95% prediction intervals (PI) were also computed for the summary random-effects estimates, which further account for between-study heterogeneity indicating the uncertainty for the effect that would be expected in a new study examining the same correlation [19,20]. A PI describes the variability of the individual study estimates around the summary effect size and represents the range in which the effect estimate of a new study is expected to lie.
The largest study considered as the most precise with a difference between the point estimate and the upper or lower 95% con dence interval less than 0.20. If the largest study presented a statistically signi cant effect, then we recorded this as a part of the grading criteria.
Between study heterogeneity was assessed and P-value of the χ 2 -based Cochran Q test and the I 2 metric for inconsistency (re ecting either diversity or bias) was reported, too. I 2 metric were used to indicate the ratio of between study-variance over the sum of within and between-study variances, ranging from 0-100% [21]. Values exceeding 50% or 75% are usually considered to represent large or very large heterogeneity, respectively. 95% Con dence intervals were calculated as per Ioannidis et al. [22].

Assessment Of Small-study Effect
Small studies tend to give substantially larger estimates of effect size when compared to larger studies. We evaluated the evidence of the presence of the small study effect, in order to identify publication and other selective reporting biases. They can also re ect genuine heterogeneity, chance, or other reasons for differences between small and large studies [23]. We evaluated whether smaller (less precise) studies lead to in ated effect estimates comparted to than larger studies. We used the regression asymmetry test proposed by Egger, that examines the potential existence of small study effects via funnel plot asymmetry [24]. Egger's test ts a linear regression of the study estimates on their standard errors weighted by their inverse variance. Indication of small study effects based on the Egger's asymmetry test was claimed when P-value ≤ 0.10. This is considered as an indication of publication bias, Indication of small study effects based on the Egger's asymmetry test was claimed when Pvalue ≤ 0.10 and the random effects summary estimate was larger compared to the point estimate of the largest (most precise) study in the meta-analysis.

Excess Statistical Signi cance Evaluation
The excess signi cant test was applied to evaluate the existence of relative excess of signi cant ndings in the published literature for any reason (e.g. publication bias, selective reporting of outcomes or analyses). The number of expected positive studies is estimated by a chi-squared-based test and being compared to the observed number of studies with statistically signi cant results (P < 0.05) [25]. A binomial test evaluated whether the number of positive studies in a meta-analysis was too large according to the power that these studies have to detect plausible effects at α = 0.05. In brief, observed versus expected studies for each meta-analysis were compared separately and this comparison also extended to groups of many meta-analysis after summing the observed and expected studies from each meta-analysis. The power of each component study was calculated using the xed-effects summary, the random effects summary, or the effect size of the largest study (smallest SE) as the plausible effect size [15]. An algorithm using non-central t distribution was used to calculate the power of each study [26]. Excess statistical signi cance for single meta-analyses was claimed at P < 0.10 (one-sided P < 0.05, with observed > expected as previously proposed), given the power to detect a speci c excess will be low, especially with few positive studies. [25] Grading Of Evidence We followed a 4-level grading (robust, highly suggestive, suggestive and weak) to evaluate the strength of the evidence based on the following criteria: number of cases, summary random-effects P-value, between-studies heterogeneity, 95% PI, small study effects bias and excess statistical signi cance [110]. This grading approach based on these parameters was used because it allows for an objective, standardized classi cation of the level of evidence and has been previously shown that provides consistent results with other more subjective grading schemes [111,112]. As most of the environmental risk factors included meta-analyses did not report the number of cases or the sample size of the studies included, we were unable to estimate the power of each meta-analysis and the excess signi cance test for these factors so we did not include excess statistical signi cance in the grading of these evidence.
Brie y, meta-analyses were considered to be supported by robust evidence if: the association was supported by more than 1000 cases, a highly signi cant association (the random effects model had a P-value ≤ 10 − 6 , a threshold that is considered to substantially reduce false positive ndings) [113,114,115], there was absence of high heterogeneity based on I2 < 50%, the 95% PI excluded the null value, and there was no evidence of small study effects or excess statistical signi cance. Highly suggestive evidence required more than 1000 cases, a highly signi cant association (a random-effects P-value ≤ 10 − 6 ), and the largest study in the meta-analysis was nominally signi cant. Associations based on metaanalyses a random-effects P-value ≤ 10 − 3 and included more than 1000 cases [113,114,115] were graded as suggestive evidence. The remaining nominally signi cant associations were graded as weak evidence (P < 0.05). We need to highlight that this speci c grading scheme focuses on the reduction of false positive ndings and the evaluation of potential biases in the studied associations. Therefore, the set of criteria used here is not ideal for a detailed evaluation of non-signi cant associations and to distinguish insu cient evidence from robust evidence of no association. That would require a different approach and another set of criteria altogether that would focus on the power of the meta-analyses to observe a signi cant effect, which was beyond the scope of our review.
Statistical analyses were performed using STATA version 14 (StataCorp, Texas, USA)

Declarations
Reporting summary: Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Data availability: Relevant data to our study are mainly included in the article, tables and supplemental material. However, we will share the original dataset after reasonable requests.
Code availability: The statistical code supporting the ndings of our study will be available upon reasonable requests SupplementalTables.docx