Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=itxc20 Critical Reviews in Toxicology ISSN: 1040-8444 (Print) 1547-6898 (Online) Journal homepage: www.tandfonline.com/journals/itxc20 Maternal blood biomarkers and adverse pregnancy outcomes: a systematic review and meta-analysis J. Gomes, F. Au, A. Basak, S. Cakmak, R. Vincent & P. Kumarathasan To cite this article: J. Gomes, F. Au, A. Basak, S. Cakmak, R. Vincent & P. Kumarathasan (2019) Maternal blood biomarkers and adverse pregnancy outcomes: a systematic review and meta- analysis, Critical Reviews in Toxicology, 49:6, 461-478, DOI: 10.1080/10408444.2019.1629873 To link to this article: https://doi.org/10.1080/10408444.2019.1629873 Published online: 11 Sep 2019. Submit your article to this journal Article views: 697 View related articles View Crossmark data Citing articles: 15 View citing articles https://www.tandfonline.com/action/journalInformation?journalCode=itxc20 https://www.tandfonline.com/journals/itxc20?src=pdf https://www.tandfonline.com/action/showCitFormats?doi=10.1080/10408444.2019.1629873 https://doi.org/10.1080/10408444.2019.1629873 https://www.tandfonline.com/action/authorSubmission?journalCode=itxc20&show=instructions&src=pdf https://www.tandfonline.com/action/authorSubmission?journalCode=itxc20&show=instructions&src=pdf https://www.tandfonline.com/doi/mlt/10.1080/10408444.2019.1629873?src=pdf https://www.tandfonline.com/doi/mlt/10.1080/10408444.2019.1629873?src=pdf http://crossmark.crossref.org/dialog/?doi=10.1080/10408444.2019.1629873&domain=pdf&date_stamp=11 Sep 2019 http://crossmark.crossref.org/dialog/?doi=10.1080/10408444.2019.1629873&domain=pdf&date_stamp=11 Sep 2019 https://www.tandfonline.com/doi/citedby/10.1080/10408444.2019.1629873?src=pdf https://www.tandfonline.com/doi/citedby/10.1080/10408444.2019.1629873?src=pdf REVIEW ARTICLE Maternal blood biomarkers and adverse pregnancy outcomes: a systematic review and meta-analysis J. Gomesa , F. Aua,b, A. Basaka, S. Cakmakb , R. Vincentb,c and P. Kumarathasana,b,d aFaculty of Health Science, Interdisciplinary School of Health Sciences, Ottawa, Canada; bEnvironmental Health Science and Research Bureau, Healthy Environments and Consumer Safety Branch, Health Canada, Ottawa, Canada; cDepartment of Biochemistry, Microbiology and Immunology, Faculty of Medicine, University of Ottawa, Ottawa, Canada; dAnalytical Biochemistry and Proteomics Laboratory, Mechanistic Studies Division, Environmental Health Science and Research Bureau, Healthy Environments and Consumer Safety Branch, Ottawa, Canada ABSTRACT Background: Pregnancy is a vulnerable period for the mother and the infant and exposures to envir- onmental chemicals in utero can influence neonatal morbidity and mortality. There is a momentum toward understanding and exploring the current maternal biological mechanisms specific to in utero effects, to improve birth outcomes. This study aims to examine the current understanding of the role of biomarkers that may be associated with term of pregnancy, infant birth weights and infant develop- ment in utero. Methods: Electronic searches were conducted in PubMed, Embase, OvidMD, and Scopus databases; and all relevant research articles in English were retrieved. Studies were selected if they evaluated maternal blood plasma/serum biomarkers proposed to influence adverse birth outcomes in the neo- nate. Data were extracted on characteristics, quality, and odds ratios from each study and meta-analysis was conducted. Results: A total of 54 studies (35 for meta-analysis), including 43,702 women, 50 plasma markers and six descriptors of birth outcomes were included in the present study. The random effect point esti- mates for risk of adverse birth outcomes were 1.61(95%CI: 1.39–1.85, p< 0.0001) for inflammation- related biomarkers and 1.65(95%CI: 1.22–2.25, p¼ 0.0013) for growth factor/hormone-related bio- markers. All subgroups of plasma markers showed significant associations with adverse birth outcomes with no apparent study bias. Conclusions: The two subsets of plasma markers identified in this study (inflammation-related and growth factor/hormone-related) may serve as potentially valuable tools in the investigation of maternal molecular mechanisms, especially select pathways underlying inflammatory and immunological medi- ation in terms of modulating adverse infant outcomes. Future large, prospective cohort studies are needed to validate the promising plasma biomarkers, and to examine other maternal biological matri- ces such as cervicovaginal fluid and urine. ARTICLE HISTORY Received 30 July 2018 Revised 3 June 2019 Accepted 6 June 2019 KEYWORDS Maternal blood biomarkers; preterm birth; infant birth weight; small-for-gestational age pregnancy Table of contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462 Literature search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462 Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463 Study quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . 463 Data extraction and data cleaning . . . . . . . . . . . . . . . . . 465 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465 Inflammation-related biomarkers . . . . . . . . . . . . . . . . . . . 466 Growth factor/hormone-related biomarkers . . . . . . . . . . 468 Inflammation-related biomarkers and SGA or PTB outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469 Growth factor/hormone-related biomarkers and SGA or PTB outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 470 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 Declaration of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 ORCID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474 Introduction Pregnancy is a very sensitive period both for the mother and the infant; numerous studies have examined various factors that may have negative influence on pregnancy outcomes. Perinatal health outcomes are considered important markers of both child and adult health (Perera et al. 2007; Stillerman et al. 2008; Woodruff et al. 2009). Lifestyle factors like smoking and alcohol-consumption during pregnancy, are just CONTACT James Gomes jgomes@uottawa.ca Interdisciplinary School of Health Sciences, 25 University Pvt, Ottawa, ON K1N 6N5, Canada � 2019 Informa UK Limited, trading as Taylor & Francis Group CRITICAL REVIEWS IN TOXICOLOGY 2019, VOL. 49, NO. 6, 461–478 https://doi.org/10.1080/10408444.2019.1629873 http://crossmark.crossref.org/dialog/?doi=10.1080/10408444.2019.1629873&domain=pdf&date_stamp=2019-11-07 http://orcid.org/0000-0002-9818-5984 http://orcid.org/0000-0001-9921-2107 http://www.tandfonline.com a few examples of the many risk factors that have demon- strated strong correlations to several birth outcomes such as preterm birth (PTB), low birth weight (LBW), small-for-gesta- tional age (SGA), intrauterine growth restriction (IUGR), and preeclampsia (Godfrey et al. 1996; Mathews et al. 1999; Jolly et al. 2000; King 2003; Wigle et al. 2008a, 2008b; Abu-Saad and Fraser 2010; Gelson et al. 2011; Hackshaw et al. 2011; Howe et al. 2012). Biomarkers are defined as parameters that can be meas- ured in a biological sample and provide information on the level of exposure, or the effects of exposure on an individual (Benford et al. 2000). The status of biomarkers of effect might reflect an earlier stage of a disease, which may be predictive of an eventual adverse health effect, though such biomarkers are dependent upon the understanding of the etiology of the disease in question (Benford et al. 2000). There are reports on maternal markers of oxidative stress and inflam- mation that lead to an increased risk for preterm birth and low infant birth weight outcomes (Buhimschi et al. 2010; Conde-Agudelo et al. 2011). However, there is paucity of information and lack of concordance between maternal blood biomarkers and subsequent pregnancy outcomes (Dagouassat et al. 2012; Conde-Agudelo et al. 2013). A number of suggestions have been proposed to explain the possible pathways through which adverse birth outcomes manifest and how biomarkers help to predict birth outcomes. Even though each purported pathway for adverse birth out- come has unique genetic and environmental predispositions, many have been identified to be immunologically mediated or influenced by inflammatory mechanisms (Buhimschi et al. 2010). The prediction of potentially negative birth outcomes is important because it could allow for the identification of high risk health issues in both the mother and the infant; and, possibly to determine risk-specific interventions that could be used to have better gestation and better health outcomes (Kumarathasan et al. 2014). The ability to predict cases may also help acquire important insights into the maternal mechanistic drivers or biochemical pathways that are involved in adversely affecting birth outcomes that are critical for child health during early development. Consequently, the study of maternal biomarkers has gained growing interest in improving the diagnosis and employing preventive interventions during pregnancy. Although emerg- ing evidence suggests that multiple underlying pathogenic pathways and factors affect pregnancy outcomes, the eti- ology of preterm birth, along with other adverse birth out- comes, still remains largely elusive. The objective of this work was to better understand the nature of relationships between maternal biomarker levels in blood and maternal biochemical changes that could affect the development of the fetus in utero. Two measures that are typically considered in the assess- ment of the quality of an infant’s development are birth weight and gestational age at delivery. Low birth weight (<2500 g) has been associated with several chronic health consequences in adulthood including hypertension, diabetes mellitus and obesity (Goldenberg et al. 2005). Normal term pregnancy is expected to last between 37–41 completed weeks, while preterm birth (PTB) is defined as live birth before 37weeks (Bhat et al. 2014; Ferguson et al. 2014). Low infant birth weight can be the result in either preterm birth or intra uterine growth restriction (IUGR), or a combination of the two. The most common surrogate measurement for IUGR is small-for-gestational age (SGA), which is generally defined as birth weight below the 10th percentile for gestational age in comparison with a reference population (Brou et al. 2012; Coussons-Read et al. 2012). Elevated circulating levels of endothelin-1, a vasoactive peptide, and high blood pressure (BP) in pregnant women are reported to be related to IUGR resulting in low infant birth weights (Curry et al. 2007). Similarly, oxidative stress has also been reported to cause maternal and fetal morbidity and is implicated in preeclamp- sia leading to PTB (Goldenberg et al. 2005; Ernst et al. 2011). Previous work on environmental pollutants have suggested that air contaminants can trigger increases in markers of oxi- dative stress and, subsequently, have been linked to adverse pregnancy outcomes (Fleischer et al. 2014). Observational studies investigating the relationship between maternal biomarkers found in maternal blood sam- ples and adverse infant health outcomes have produced inconsistent results. To our knowledge until now there has been no comprehensive synthesis of research on maternal blood biomarkers and adverse infant birth outcomes. This systematic review and meta-analysis was undertaken to iden- tify and confirm the associations between maternal blood biomarkers and adverse infant birth outcomes such as low birth weight, preterm birth and small-for-gestational age infants through comprehensive synthesis of literature and subsequent meta-analyses of the relevant data. Procedures developed in our laboratory were followed to collect relevant existing data and systematically review it and conduct meta- analyses of relevant data. Methodology This systematic review was conducted following a prospect- ively prepared protocol, and is reported using widely recom- mended guidelines for systematic reviews and meta-analyses (Moher et al. 2009a, 2009b, 2009c; Moher et al. 2015; Shamseer et al. 2015; Stewart et al. 2015). The initial search was conducted until December 2017 (Table 1). Literature search The search strategy was designed to identify observational and experimental studies that described the association between maternal blood biomarkers and the risk of adverse birth outcomes. An initial search was performed in PubMed using a combination of keywords and text words related to biomarkers (“cytokine,” “chemokine,” “inflammatory markers,” “interferon-gama (IFN-c),” “vascular endothelial growth factor (VEGF),” “vascular cell adhesion molecule (VCAM),” “intercellular adhesion molecule (ICAM),” “tissue necrosis fac- tor-alpha (TNF-a),” “granulocyte macrophage colony stimulat- ing factor (GM CSF),” “C-reactive protein (CRP),” “matrix metalloproteinase (MMP),” “monocyte chemoattractant pro- tein (MCP-1),” “macrophage inhibitory protein (MIP-1),” 462 J. GOMES ET AL. “endothelin(ET),” “interleukin (IL)”) and hormone-related bio- markers including angiopoietin (ANGPT2), (BDNF), Cortisol, (CRH), (FGF), (IGFBP), (b-hCG), (PAPP-A), (NT), and (VEGF) were assessed and evaluated (Table 2(a)). The reported relationship between growth factor and/or hormone related factors (BDNF, CRH and b-hCG) and birth outcomes (“preterm,” “birth weight,” “preterm birth,” “small-for-gestational age,” “SGA,” and “gestational age”) was assessed and evaluated (Table 2(a)). The aforementioned biomarkers were chosen because previous research has linked PTB, LBW, and SGA with inflam- matory and growth factor and/or hormone related pathways (Conde-Agudelo et al. 2011, 2013). Search terms, as well as inclusion and exclusion criteria, were used to filter for more relevant studies. As per the inclu- sion criteria, studies were included if they identified bio- markers in maternal blood (whole blood, plasma or serum), as marker of exposure or adverse birth outcomes. The inclusion criteria were selected to explore a relatively less-studied bio- logical matrix (whole blood, plasma, or serum) as previous reports on maternal predictors of adverse birth outcome have mainly focused on amniotic fluid and cervicovaginal fluid. Studies were excluded if they assessed biomarkers in other biological samples such as urine, cervicovaginal fluid, saliva, or if the study was focused on a therapy or therapeutics. In the initial search, we chose those biomarkers identified in maternal blood (whole blood, plasma or serum) for the pre- diction of adverse birth outcomes such as preterm birth, low, or/and small for all the years until 2017 (Table 2(a)). Language restrictions were applied to include only English full text. Limitations were applied in the search strategy to include only human data, although animal data was drawn upon to formu- late conceptual pathways linking biomarkers and outcome. The initial search criterion was developed using PubMED and when the criterion was operational, it was applied to other databases. A comprehensive search was then conducted using SCOPUS, ToxLine, and OvidMD databases using keywords and text words for each of the biomarkers identified in the initial search and keywords and text words for adverse infant birth outcomes described previously. Selected articles were collected in RefWorks and catego- rized according to the type of study. The categories include observational studies (case-control, cohort studies and cross- sectional studies) and reviews. Furthermore, studies that were referenced in the articles being reviewed were hand-searched for potentially relevant studies that could be included to yield six additional studies. References for excluded studies can be obtained from the authors. Study selection One reviewer (F. A.) screened titles and abstracts of all identi- fied citations and selected potentially eligible studies. Following review of the abstract all full-text articles that were relevant were assessed by the same reviewer for inclusion and data extraction and 10% sample of the papers was examined by a second independent reviewer (JG). We resolved any disagreements by discussion and consensus. For multiple or duplicate publication of the same data set, we included only the most recent or complete study. Studies were included in the review if (i) they were cohort, cross-sectional, case–control, or review studies that evaluated the biomarkers in relation to intrauterine growth restriction (IUGR), preterm birth, low birth weight, or small-for-gesta- tional age (SGA) infants in women at any level of risk (ii) the biological samples were collected before the clinical onset of outcome and; for the meta-analysis: (iii) they allowed for the assessment of cases with higher biomarker levels and adverse birth outcomes. Studies were excluded if (i) they were case series or reports, editorials, or comments; and for the meta- analyses, if (ii) biomarker data were reported only as mean or median values or (iii) they reported insufficient data on cases exposed or control groups or (iv) if information was available on odds ratio or confidence intervals. The biomarkers that were included in this study were selected based on previous evidence suggesting that inflammation-related biomarkers and immunological biomarkers are key regulators of labor as well as infant growth and development (Goldenberg et al. 2000; Ekelund et al. 2008; Curry et al. 2009; Georgiou et al. 2011; Coussons-Read et al. 2012; Ferguson et al. 2014). Study quality assessment The methodological quality of the studies in the systematic review was assessed by at least one reviewer using a 15- point checklist (Appendix – Table A1) adapted from Downs and Black; because we have found it be the most compre- hensive and a thorough method to evaluate study quality (Downs and Black 1998; Wigle et al. 2008a, 2008b). The qual- ity of each study was evaluated based on external and internal validity, methodology, confounders, and population demographics. Two authors (F. A. and J. G.) reviewed the articles and assigned a numerical value representative of its quality and reported in (Appendix – Table A2). The scores were categorized into four groups. Studies with scores between 0 and 3 were classified as “poor quality” studies, 4–7 were deemed “low quality,” 8–11 were “medium quality,” and 12–15 were “good quality” studies. Among the studies that were reviewed 43 studies were deemed medium quality and the remaining 11 were categorized as good quality stud- ies (there were no studies in the poor quality category). Table 1. Search terms used for initial search for systematic literature review. Biomarkers Outcome Inclusion criteria Exclusion criteria Cytokine Chemokine Inflammatory markers IFN-c VEGF VCAM ICAM TNF-a GM CSF CRP MMP MCP-1 MIP-1 Endothelin Interleukin Preterm Birth weight Preterm birth Small-for-gestational age SGA Gestational age Maternal blood Maternal serum Maternal plasma In utero Therapy Treatment Therapeutic Infection Pulmonary Bronchopulmonary Cervicovaginal Urinary Urine Placenta CRITICAL REVIEWS IN TOXICOLOGY 463 Table 2. (a) Maternal biomarkers measured at different time points during pregnancy. Biomarkers 1st Trimester 2nd Trimester 3rd Trimester Delivery 1. Inflammatory-related biomarkers C-reactive protein (CRP) � � Eotaxin � � Ferritin � Granulocyte macrophage colony-stimulating factor (GM-CSF) � Granulocyte colony-stimulating factor (G-CSF) � Intercellular cell adhesion molecule (ICAM) � � Interferon-c (INFc) � Interleukin-1b, (IL-1b) � Interleukin-1R (IL-1R) � � IL-2, � � IL-4, � IL-5, � IL-6, � IL-6R � IL-7, � � IL-8, � IL-9, � � IL-10 � IL-12, � IL-13, � IL-17 � IL-18 � Matrix metalloproteinase(MMP-2) � MMP-3 � MMP-7 � MMP-8 � MMP-9 � Macrophage inflammatory protein -1 (MIP) � Macrophage migration inhibitory factor (MIF) � Monocyte chemotactic protein-1 and -3 (MCP) � Tissue inhibitor metalloproteinase-1, -2, and -4 (TIMP) � Transforming growth factor-a and –b (TGFa / b) � Tumour necrosis factor-a (TNFa) � Tumour necrosis factor receptor-1 (TNF R1) � Vascular cell adhesion molecule (VCAM) 2. Growth factor/hormone-related biomarkers Angiopoietin2 (ANG-PT2) � b-human chorionic gonadotropin (b-hCG) � � Brain derived neurotrophic factor (BDNF) � Corticotropin releasing hormone (CRH) � Cortisol � Fetal growth factor (FGF) � Insulin-growth-factor-binding protein-1 (IGF-BP � Neurotrophin-3 and -4 (NT) � Pregnancy-associated Plasma Protein A (PAPP) � � Vascular endothelial growth factor (VEGF) � Table 2(b). Relationship between biomarkers and adverse birth outcomes identified in the literature search (N¼ 54). Biomarker Number of studies� PTB LBW SGA IUGR Inflammatory-related biomarkers IL-1b 620,22,34,51,61,63 þþþþþ þ IL-2 522,24,32,51,64 þþþþþ IL-6 1920,23,24,26,27,32,45,47,49,50,51,58,61,63,64,66,67,70,74 þþþþþþþþþþþþþþþþþ þ þ IL-8 521,22,26,51,65 þþþ þ þ IL-12 434,49,61,65 þþ þ þ IL-17 236,49 þþ IL-18 333,49,61 þþþ IL-4, -5, -7, -9, -13 634,49,51,56,64,65 þþþ þ þþ TGF-a/b 249,61 þþ IFN-c 824,26,32,34,49,56,61,65 þþþþþ þ þþ TNF-a/-b 1320,22,23,24,32,37,45,49,51,61,63,66,67 þþþþþþþþþþþ þ þ ICAM/VCAM 227,39 þþ CRP 1820,25, 27,31,40,42,45,46,50,53,58,65,67,68,69,70,71,72 þþþþþþþþþþ þþþ þþþ þ GM-CSF 624,32,42,48,52,73 þþþþþ þ þ G-CSF 427,34,52,61 þþþ þ Ferritin 327,43,44 þþ þ MMP-2 360,70,77 þþ þ MMP-3/7 222,48 þþ MMP-8 270,77 þþ MMP-9 938,48,49,58,60,62,70,76,77 þþþþþþ þ þþ (continued) 464 J. GOMES ET AL. Study quality was conducted in order to exclude poor quality studies, however, there were no studies that were of poor quality; and no reporting discrepancies between the good and the medium study qualities were observed. Data extraction and data cleaning Potentially relevant articles were acquired, and data were extracted from all reports, and recorded on a form designed independently by the reviewers. There was no blinding of authorship. The following information was extracted from each article: study characteristics (design and prospective or retrospective data collection); participants (sample size, coun- try where research was conducted and date of publication); description of the biomarkers (gestational age at sampling, and biological sample); and reference standard used (preterm definition, SGA definition, and LBW definition). Both data extraction and quality assessment were conducted individu- ally and any differences of opinions or disagreements were resolved by discussion among authors. This study entails the analysis of data collected from multiple databases. All data extracted were imported from Microsoft Excel v.14.5.7 (Microsoft Office, Redmond, WA) into Comprehensive Meta-Analysis Software (CMA) v(0).2 (Biostat, Eaglewood, NJ) (Biostat 2011; Microsoft Office 2011). The biomarkers identified in this study as associated with birth outcomes are described in Table 2(a). The relevance of the biomarkers in the different tri- mesters and its association with adverse birth outcome was also explored (Tables 2(a,b)). The data extracted from the literature was categorized as inflammation-related or growth factor/hor- mone-related (Figure A1 and Tables A3 and A4). Statistical analysis Data extracted from each study were arranged in CMA and odds ratios were analyzed using Comprehensive Meta- Analysis Software v3 (Biostat, Inc., Englewood, NJ). Meta- analyses were performed using subgroups of studies with similar biomarker families to minimize clinical heterogeneity. Random effects model meta-analyses was conducted using CMA and the overall mean of all the studies was estimated as the weighted mean where the weight assigned to each study is the inverse of that study’s variance. Results The searches produced 6159 citations, of which 391 were considered relevant. We identified a total of 267 empirical studies of all study types (both observational and review studies) after filtering out 124 studies due to lack of human data or because English full text was unavailable. After reviewing the abstracts of the 267 selected articles, we excluded 213 studies either because there were no compari- sons of biomarker levels and outcomes, or because the bio- marker analysis was not done in maternal blood (whole blood, plasma, or serum) which resulted in 53 observational studies and 1 review article for final analyses (Figure 1). A total of 54 studies met the inclusion criteria (30 case–- control, 22 cohort, 1 systematic review, 1 cross-sectional) and were included for the systematic review. Of the 54 studies, 35 studies (18 case–control and 17 cohort), evaluating a total of 50 biomarkers were identified for the meta-analysis after exclusion of 19 studies due to insufficient information per- taining to elevated biomarker levels and corresponding out- comes, or lack of information on odds ratios or confidence intervals; as per the selection criteria for meta-analyses. The biomarkers identified in this study and select adverse birth outcomes were assessed in this review (Table 2(a)). Publication bias was conducted for all the studies included in these meta-analyses and no specific trends were observed. The quality scores for the studies included in the systematic review are shown in Appendix Table A2. The quality scores for all the studies ranged between 9(60%) and 13 (86.7%). The median score for all the studies was 11(73.7%). Further Table 2b. Continued. Biomarker Number of studies� PTB LBW SGA IUGR TIMP-1, -2, -4 322,48,77 þþþ Eotaxin 234,39 þ þ MIP-1 421,22,39,49 þþþþ MCP-1/3 321,34,49 þþ þ MIF 345,49,57 þþþ IL-1R 322,34,47 þþ þ TNF-R1 421,22,49,62 þþþþ þ IL-6R 321,22,49 þþþ IL-10 1020,27,45,47,49,56,61,63,66,70 þþþþþþþ þ þþ Growth factor/hormone-related biomarkers CRH 227,45 þþ BDNF 239,49 þþ NT-3/4 149 þ b-hCG 541,53,54,55,75 þþþ þþ PAPP-A 641,53,54,55,62,75 þþþþ þþþ Cortisol 227,45 þþ FGF/VEGF 122 þ IGFBP-1 139 þ ANGPT2 221,22 þþ PTB: pre-term birth; LBW: low birth weight; SGA: small-for-gestational age; IUGR: inter-uterine growth restriction. þ indicates a single study which addresses the specific biomarker and outcome of interest. Underlined study: excluded from the meta-analysis due to insufficient data.�numbers show the number of studies for that biomarker and the superscript provides the reference. CRITICAL REVIEWS IN TOXICOLOGY 465 analyses of the studies based on their qualitative scores were not conducted. Our findings from this review identified the relationships between the different birth outcomes (preterm birth (PTB), low birth weight (LBW), small of gestational age (SGA) and intra- uterine growth restriction (IUGR) and maternal inflammatory biomarkers and hormonal biomarkers measured at different time points of pregnancy (Tables 2a, 2b, A3, and A4). The dif- ferent biomarkers were associated with different birth out- comes. The number of studies addressing each of the biomarker and the different birth outcomes PTB (133 citations), LBW (48 citations), SGA (10 citations), and IUGR (2 citations) are shown in Table 2(b). The studies identified in this review were mostly conducted in Europe (18 studies) and in North America (13 studies) and the rest of the studies were conducted in other countries. The sample size in cohort studies ranged from 62 to 9450. The number of case participants enrolled in case–- control studies ranged from 29 to 3539 and the corresponding number of controls ranged from 21 to 46,262. The biological samples that were analyzed for biomarkers in this study were obtained from maternal serum (24 studies) and maternal plasma (11 studies). Nineteen studies provided data on preterm birth before 37weeks, four studies provided data on preterm birth before 34weeks, two studies provided data on preterm birth before 33weeks of gestation, and four studies provided data on preterm birth prior to 32weeks. Only five studies provided data on small-for-gestational age infants, one study reported findings on low birth weight and also one study reported findings on IUGR. The findings on IUGR were included for the systematic review but the study did not contain sufficient information to be included in the meta-analysis. The biomarkers were categorized as inflammation-related or growth factor/hormone-related depending on their mechanism of action. Inflammation-related biomarkers The different inflammation-related biomarkers of interest in this study were CRP (Goldenberg et al. 2001; Sattar et al. 2004; Pitiphat et al. 2005; Lohsoonthorn et al. 2007; Ernst et al. 2011; Bakalis et al. 2012; Moghaddam et al. 2012; Pearce et al. 2010; Ferguson et al. 2014), Eotaxin (Georgiou et al. 2011; Laudanski et al. 2012), Ferritin (Goldenberg et al. 2001; Paternoster et al. 2002; Ozgu-Erdinc et al. 2014), G-CSF (Goldenberg et al. 2000; Keith et al. 2000), GM-CSF (Curry et al. 2007, 2009; Tsiartas et al. 2012), ICAM (Goldenberg et al. 2001), IFN-c (Curry et al. 2007, 2009; Georgiou et al. 2011; Tsiartas et al. 2012; Pearce et al. 2016), IL-2 (Camgil Arikan et al. 2012; Dibble et al. 2014; von Minckwitz et al. 2000; Murtha et al. 1998; Curry et al. 2007, 2009; Brou et al. 2012; Tsiartas et al. 2012), IL-4 (von Minckwitz et al. 2000; Tsiartas et al. 2012), IL-6 (Turhan et al. 2000; von Minckwitz et al. 2000; Goldenberg et al. 2001; Curry et al. 2007, 2009; Pearce et al. 2010; Ruiz et al. 2012; Ferguson et al. 2014), IL-8 391 Potentially relevant studies identified in initial searches by applying the search criteria to different databases and by applying the inclusion and exclusion criteria 267 Studies retrieved for title and abstract screening for relevance 54 Articles included in this systematic review 35 Articles included in Meta-Analysis 18 Case-control studies 17 Cohort studies 213 Studies excluded because they did not contain infant outcome of interest or biomarker data 124 Studies excluded after applying filters from the search engines (full text, English, human data) 16 Did not have full text available 21 Were not English articles 87 Did not contain human data Figure 1. Study selection process. 466 J. GOMES ET AL. (von Minckwitz et al. 2000; Brou et al. 2012; Bhat et al. 2014), IL-10 (Goldenberg et al. 2001; Pearce et al. 2010; Ruiz et al. 2012; Tsiartas et al. 2012; Ferguson et al. 2014; Pearce et al. 2016), IL-12 (Georgiou et al. 2011; Tsiartas et al. 2012), IL-13 (Pearce et al. 2016), IL-17 (Hee et al. 2011; Tsiartas et al. 2012), IL-18 (Ekelund et al. 2008; Tsiartas et al. 2012), IL-1b (von Minckwitz et al. 2000; Brou et al. 2012; Ferguson et al. 2014), IL-1Ra (Georgiou et al. 2011; Brou et al. 2012; Ruiz et al. 2012), IL-6Ra (Turhan et al. 2000; Brou et al. 2012; Bhat et al. 2014), MCP (Brou et al. 2012; Tsiartas et al. 2012), MIF (Pearce et al. 2008; Tency et al. 2012; Tsiartas et al. 2012), MIP (Brou et al. 2012; Laudanski et al. 2012; Tsiartas et al. 2012; Bhat et al. 2014), MMP (Botsis et al. 2006; Brou et al. 2012; Tency et al. 2012; Tsiartas et al. 2012), TNF-a (von Minckwitz et al. 2000; Curry et al. 2007, 2009a, 2009b; Tency et al. 2012; Tsiartas et al. 2012; Ferguson et al. 2014; Jelliffe- Pawlowski et al. 2014), TGF-b (Tsiartas et al. 2012), TNF-R1 (Brou et al. 2012; Tsiartas et al. 2012; Bhat et al. 2014) and TNF-b (Tsiartas et al. 2012). The pooled risk estimates and the 95% confidence interval for all the inflammatory biomarkers was 0.845 (95%CI: 0.83–0.86) (Table 3). A total of eight studies had reported on the biomarker CRP and the significant pooled estimate for CRP was OR¼ 1.199; 95%CI:1.04–1.35 (Table 3). The pooled estimates for Eotaxin from two studies and Ferritin from six studies were also significant OR¼ 2.56; 95%CI:1.53–3.60 and OR¼ 1.05; 95%CI: 1.02–18, respectively (Table 3). Pooled risk estimates for ICAM and IFN were also significant at OR¼ 2.09; 95%CI:1.21–2.98 and OR¼ 0.27; 95%CI: 0.23–0.32, respectively. The pooled risk estimates for IL-10 was significant at OR¼ 0.60; 95%CI: 0.49–0.71. The pooled estimates for IL-17 and IL-18 were significant at OR¼ 0.47; 95%CI: 0.23–0.71 and OR¼ 0.55; 95%CI: 0.28–0.82, respectively. The pooled risk esti- mate for IL-1Ra was significant at OR¼ 3.02; 95%CI: 2.09–3.95. Pooled risk estimates for IL-4, IL-6, and IL-6R were all signifi- cant OR¼ 0.57; 95%CI: 0.33–0.80; OR¼ 1.20; 95%CI: 1.13–1.28, Table 4. Weighted mean estimates of odds ratio for growth factor/hormone-related biomarkers associated with adverse birth outcomes. Effect size and 95% confidence interval Test of null (2 tail) Heterogeneity Tau squared Biomarker # stds Point estimate SE Var Lower limit Upper limit Z value p-value Q value p value I2 Tau2 SE Var Tau2 ANGPT 2 0.335 0.092 0.009 0.145 0.534 3.277 0.001 7.515 0.057 60.08 0.208 0.367 0.134 0.456 BDNF 2 3.431 0.688 0.473 2.082 4.880 4.986 0.000 0.157 0.984 0.00 0.000 1.859 3.456 0.000 Cortisol 3 1.367 0.229 0.052 0.916 1.89116 5.945 0.000 3.686 0.595 0.00 0.000 0.276 0.076 0.000 CRH 3 1.867 0.309 0.095 1.262 1.262 6.039 0.000 5.989 0.307 76.525 0.161 0.625 0.391 0.401 B-hCG 5 1.624 0.142 0.020 1.345 1.345 11.413 0.000 2.775 0.972 0.00 0.000 0.102 0.011 0.000 PAPP 6 0.546 0.122 0.001 0.522 0.522 94.741 0.000 70.491 0.000 84.395 0.019 0.027 0.007 0.138 NT 2 0.453 0.019 0.008 0.273 0.273 4.933 0.000 11.698 0.000 74.356 0.233 0.336 0.113 0.482 All Fix 26 0.554 0.011 0.000 0.551 0.577 46.576 0.000 228.72 0.000 77.702 0.048 0.059 0.003 0.220 Ran 1.070 0.071 0.005 0.931 1.210 15.050 0.000 The bold represent statistically significant values (p< 0.05). Table 3. Weighted mean estimates of odds ratio for inflammation-related biomarkers and the overall mean of these estimates. Biomarker # of studies Effect size and 95% confidence interval Test of null (2 tail) Heterogeneity Tau squared Point estimate Std error Variance Lower limit Upper limit Z-value p-value Q-value p-value I2 Tau2 Std error Var Tau CRP 8 1.1999 0.0789 0.0062 1.0452 1.3546 15.2036 0.0000 26.7681 0.0307 43.9631 0.1027 0.1088 0.0118 0.3204 Eotaxin 2 2.5666 0.5285 0.2793 1.5308 3.6025 4.8565 0.0000 0.5240 0.9136 0.0000 0.0000 1.2170 1.4810 0.0000 Ferritin 6 1.0518 0.0159 0.0003 1.0206 1.0830 66.1022 0.0000 12.9059 0.2995 14.7674 0.0010 0.0032 0.0000 0.0317 G-CSF 4 1.2346 0.1251 0.0157 0.9894 1.4799 9.8680 0.0000 4.4081 0.7318 0.0000 0.0000 0.0755 0.0057 0.0000 GMCSF 114 0.9517 0.0463 0.0021 0.8610 1.0424 20.5666 0.0000 54.0852 0.0000 75.9638 0.0983 0.0565 0.0032 0.3135 CRP 3 1.3112 0.1672 0.0280 0.9834 1.6390 7.8406 0.0000 2.8808 0.7184 0.0000 0.0000 0.1167 0.0136 0.0000 ICAM 2 2.0931 0.4512 0.2035 1.2089 2.9774 4.6395 0.0000 2.6631 0.4465 0.0000 0.0000 1.3112 1.7192 0.0000 IFN 9 0.2723 0.0223 0.0005 0.2287 0.3159 12.2283 0.0000 333.621 0.0000 94.9044 0.3800 0.2830 0.0801 0.6164 IL-10 7 0.6019 0.0571 0.0033 0.4901 0.7138 10.5475 0.0000 93.8387 0.0000 86.1464 0.3564 0.2730 0.0745 0.5970 IL-12 2 0.7400 0.1862 0.0347 0.3750 1.1051 3.9736 0.0001 11.6274 0.0088 74.1989 1.0305 1.3962 1.9495 1.0151 IL-17 2 0.4733 0.1226 0.0150 0.2329 0.7137 3.8589 0.0001 9.1100 0.0279 67.0691 0.3529 0.4841 0.2343 0.5941 IL-18 2 0.5501 0.1395 0.0195 0.2767 0.8234 3.9440 0.0001 7.3348 0.0620 59.0992 0.2936 0.5017 0.2517 0.5418 IL-1B 3 1.0142 0.0652 0.0043 0.8864 1.1421 15.5507 0.0000 5.7688 0.3294 13.3276 0.0064 0.0311 0.0010 0.0799 IL-1Ra 3 3.0238 0.4714 0.2222 2.0999 3.9477 6.4146 0.0000 3.4848 0.6257 0.0000 0.0000 1.0433 1.0884 0.0000 IL-2 8 1.0857 0.0512 0.0026 0.9854 1.1859 21.2210 0.0000 11.5171 0.7151 0.0000 0.0000 0.0160 0.0003 0.0000 IL-4 2 0.5687 0.1195 0.0143 0.3344 0.8029 4.7578 0.0000 5.7673 0.1235 47.9830 0.0720 0.1359 0.0185 0.2684 IL-6 13 1.2030 0.0397 0.0016 1.1252 1.2808 30.2988 0.0000 13.2717 0.9730 0.0000 0.0000 0.0120 0.0001 0.0000 IL-6R 3 2.4402 0.3935 0.1548 1.6690 3.2113 6.2017 0.0000 0.7308 0.9812 0.0000 0.0000 0.6577 0.4326 0.0000 IL-8 3 1.8897 0.4041 0.1633 1.0977 2.6817 4.6767 0.0000 5.7791 0.3283 13.4818 0.2977 1.4124 1.9948 0.5456 MCP 2 0.5409 0.1400 0.0196 0.2665 0.8152 3.8644 0.0001 8.1324 0.0434 63.1106 0.3625 0.6064 0.3677 0.6021 MIF 3 3.9432 0.5501 0.3026 2.8650 5.0214 7.1680 0.0000 0.8078 0.9765 0.0000 0.0000 1.2909 1.6665 0.0000 MIP 5 1.8947 0.2392 0.0572 1.4260 2.3635 7.9222 0.0000 5.2408 0.8128 0.0000 0.0000 0.3009 0.0905 0.0000 MMP 3 2.3375 0.2526 0.0638 1.8425 2.8325 9.2555 0.0000 11.7690 0.0381 57.5157 0.8934 1.4020 1.9656 0.9452 TIMP 4 0.8622 0.0131 0.0002 0.8365 0.8878 65.9118 0.0000 54.2969 0.0000 87.1079 0.0103 0.0079 0.0001 0.1013 TNF 16 0.2732 0.0187 0.0003 0.2367 0.3098 14.6403 0.0000 446.594 0.0000 93.0586 0.4852 0.3593 0.1291 0.6966 ALL Fix 126 0.8450 0.0080 0.0001 0.8292 0.8607 105.184 0.0000 2223.84 0.0000 88.7582 0.1378 0.0539 0.0029 0.3712 Random 1.16 0.036 0.013 1.090 1.232 32.799 0.0000 77.07 The bold represent statistically significant values (p< 0.05). CRITICAL REVIEWS IN TOXICOLOGY 467 and OR¼ 2.44; 95%CI: 1.67–3.21, respectively. Pooled risk esti- mate for IL-8 was statistically significant at OR¼ 1.89; 95%CI: 1.09–2.68. For MCP the pooled risk estimate was significant at OR¼ 0.54; 95%CI: 0.27–0.82. Both MIF and MIP were estimated to have significant pooled risk estimates at OR¼ 3.94; 95%CI: 2.87–5.02 and OR¼ 1.89; 95%CI: 1.43–2.36, respectively. MMP was estimated to have a pooled risk estimate of OR¼ 2.34; 95%CI: 1.84–2.83. Both TMP and TNFa were observed to have a significant pooled risk estimate at OR¼ 0.82; 95%CI: 0.84–0.89 and OR¼ 0.27; 95%CI: 0.24–0.31, respectively (Table 3). Growth factor/hormone-related biomarkers The second group of biomarkers (growth factors/hormone- related) is a combination of growth factors and hormones pertinent to pregnancy and parturition (Table 4). Hormone- related biomarkers including ANGPT2 (Brou et al. 2012; Bhat et al. 2014), BDNF (Spencer et al. 2008; Kim et al. 2011), Cortisol (Goldenberg et al. 2001; Pearce et al. 2010), CRH (Goldenberg et al. 2001; Pearce et al. 2010), FGF (Brou et al. 2012), IGFBP (Laudanski et al. 2012), b-hCG (Goetzinger et al. 2010; Kirkegaard et al. 2011; van Ravenswanij et al. 2011; Dane et al. 2013), PAPP-A (Spencer et al. 2008; Goetzinger et al. 2010; Kirkegaard et al. 2011; Dane et al. 2013), NT Table 6. Weighted mean estimates of odds ratio for SGA or PTB based on growth factors/hormone related biomarkers. Effect size and 95% confidence interval Test of null (2-Tail) Heterogeneity Tau squared # of stds Point estimate Std error Var Lower limit Upper limit Z-value p-value Q-value p-value I2 Tau2 Std error Var Tau SGA� 4 0.5536 0.0121 0.0001 0.5298 0.5774 45.5699 0.0000 83.7814 0.0000 91.6449 0.0237 0.0324 0.0010 0.1540 PTB <37w† 13 0.6472 0.0551 0.0030 0.5392 0.7552 11.7471 0.0000 128.090 0.0000 80.4825 0.3884 0.2576 0.0664 0.6232 PTB <35w‡ 4 1.5405 0.1770 0.0313 1.1936 1.8874 8.7042 0.0000 1.6203 0.9779 0.0000 0.0000 0.1464 0.0214 0.0000 PTB <34w¶ 2 2.5302 0.8380 0.7022 0.8879 4.1726 3.0196 0.0025 3.3754 0.3373 11.1205 1.2311 9.1108 83.0068 1.1095 PTB <32w§ 4 2.3101 0.3600 0.1296 1.6046 3.0157 6.4172 0.0000 2.4891 0.9279 0.0000 0.0000 0.7046 0.4964 0.0000 ALL| Fix 26 0.5546 0.0119 0.0001 0.5313 0.5779 46.5768 0.0000 228.726 0.0000 77.7026 0.0487 0.0590 0.0035 0.2207 Ran 1.0707 0.0711 0.0051 0.9313 1.2102 15.0502 0.0000 �b-hCG (OR þ 1.91; 95%CI:1.20–3.02), PAPP-A (OR ¼ 2.35:95%CI:1.05–5.63); OR ¼ 0.54; 95%CI:0.50–0.57). †CRH, BDNF (OR ¼ 3.16; 95%CI:1.13–8.84); (OR ¼ 0.70; 95%CI:1.60–8.20), NT-3 (OR ¼ 2.30; 95%CI:1.10–4.70), NT-4 (OR ¼ 0.40; 95%CI:0.20–0.85), b-hCG, PAPP-A (OR ¼ 1.95; 95%CI:1.39–2.73), Cortisol, FGF, VEGF (OR ¼ 2.83; 95%CI:1.46–10.09), IGFBP-1, ANGPT2 (OR ¼ 5.19; 95%CI:1.50–17.28) (OR ¼ 0.38; 95%CI:0.12–0.89). ‡CRH, b-hCG, PAPP-A, Cortisol. ¶b-hCG, PAPP-A (OR ¼ 7.0; 95%CI: 1.80–27.70). §CRH, b-hCG, PAPP-A, Cortisol. |CRH, BDNF, NT-3, NT-4, b-hCG, PAPP-A, Cortisol, FGF, VEGF, IGFBP-1. The bold represent statistically significant values (p< 0.05). Table 5. Weighted mean estimates of odds ratio for SGA or PTB based on inflammation related biomarkers. Effect size and 95% confidence interval Test of null (2- Tail) Heterogeneity Tau squared # Stu Point estimate Std error Var Lower limit Upper limit Z-value p-value Q-value p-value I2 Tau2 Std error Var Tau SGA� 8 0.0931 0.0214 0.0005 0.0511 0.1350 4.3439 0.0000 83.5867 0.0000 82.0546 0.0697 0.0797 0.0064 0.2640 PTB <37w† 70 1.0562 0.0133 0.0002 1.0301 1.0823 79.3373 0.0000 385.2806 0.0000 63.9224 0.0588 0.0414 0.0017 0.2426 PTB <35w‡ 7 0.7836 0.1005 0.0101 0.5867 0.9806 7.7991 0.0000 26.7907 0.0133 51.4757 0.1937 0.1797 0.0323 0.4401 PTB <34w¶ 17 0.8916 0.0122 0.0001 0.8678 0.9155 73.2654 0.0000 117.5125 0.0000 71.9179 0.0149 0.0094 0.0001 0.1220 PTB <33w§ 2 0.7479 0.1019 0.0104 0.5482 0.9476 7.3411 0.0000 24.4715 0.0000 87.7409 0.3853 0.4128 0.1704 0.6208 PTB <32w| 10 0.9770 0.1168 0.0136 0.7481 1.2059 8.3646 0.0000 36.7001 0.0086 48.2290 0.3002 0.2578 0.0665 0.5479 PTB <30w# 11 0.8230 0.0501 0.0025 0.7248 0.9212 16.4266 0.0000 34.7732 0.0299 39.6086 0.0377 0.0311 0.0010 0.1942 ALL††Fix 125 0.8450 0.0080 0.0001 0.8292 0.8607 105.184 0.0000 2223.848 0.0000 88.7582 0.1378 0.0539 0.0029 0.3712 Ran 1.1607 0.0360 0.0013 1.0902 1.2312 23.2791 0.0000 �IL-12 (3.62; 95%CI: 1.74–7.54), IL-13 (0.11; 95%CI: 0.02–0.66), IL-1Ra (5.32; 95%CI: 2.40–11.79), IFN-c (3.72; 95%CI: 1.54–9.03). †IL-1b (OR ¼ 3.74; 95%CI: 1.31-10.70), IL-1Ra (OR ¼ 3.08; 95%CI: 1.16–8.20), (OR ¼ 2.55; 95%CI: 1.24–5.24), IL-2 (OR ¼ 4.02; 95%CI: 1.53–10.52), IL-4, IL-6 (OR ¼ 1.29; 95%CI: 1.07–1.56), (OR ¼ 1.27; 95%CI: 1.02–1.59), (OR ¼ 30.8; 95%CI: 3.93–241.20), IL-6R (OR ¼ 2.72; 95%CI: 1.15–6.48), (OR ¼ 2.70; 95%CI: 1.20–5.80), IL-8 (OR ¼ 3.75; 95%CI: 1.22–11.50), (OR ¼ 5.25; 95%CI: 1.14–24.27), IL-10 (OR ¼ 1.28; 95%CI: 1.01–1.62), (OR ¼ 4.60; 95%CI: 2.20–9.70), IL-12, IL- 17, IL-18, TGF-b, IFN-c, TNF-a (OR ¼ 6.86; 95%CI: 2.42–19.42), TNF-b (OR ¼ 2.90; 95%CI: 1.40–6.0), TNF-R1 (OR ¼ 6.42; 95%CI: 2.24–18.41), CRP (OR ¼ 8.96; 95%CI: 4.60–17.43), GM-CSF (OR ¼ 0.4; 95%CI: 0.2–0.8), G-CSF, MMP-7, MMP-9 (OR ¼ 33.75; 95%CI: 5.84–194.8), (OR ¼ 6.0; 95%CI: 2.90–12.70), TIMP-1, Eotaxin, MIP-1a (OR ¼ 2.98; 95%CI: 1.13–7.86), MIP-1b (OR ¼ 2.50; 95%CI: 1.20–5.10), MCP-3 (OR ¼ 4.44; 95%CI: 1.58–12.46), MIF-1 (OR ¼ 4.16; 95%CI: 1.83–9.48), (OR ¼ 4.70; 95%CI: 2.20–9.90). ‡IL-6, IL-10, ICAM-1, CRP, G-CSF, Ferritin (OR ¼ 4.85; 95%CI: 1.60–14.70). ¶IL-2, IL-6, IL-18 (OR ¼ 3.40; 95%CI: 1.20–9.80), IFN-c, TNF-a, CRP, GM-CSF, MMP-3, MMP-9 (OR ¼ 2.21; 95%CI: 1.59–3.07), TIMP-1(OR ¼ 0.89; 95%CI: 0.84–0.95), TIMP-2 (OR ¼ 0.78; 95%CI{0.77-0.84), TIMP-4. §IL-17, G-CSF (OR ¼ 1.42; 95%CI: 1.01–1.99). |IL-6, IL-10, ICAM-1 (OR ¼ 4.90; 95%CI: 1.30–18.81), CRP, G-CSF, Ferritin, MMP-9. #IL-2, IL-6, IFN-c, TNF-a, GM-CSF. ††IL-1b, IL-1Ra, IL-2, IL-4, IL-6, IL-6R, IL-8, IL-10, IL-12, IL-13, IL-17, IL-18, TGF-b, IFN-c, TNF-a, TNF-b, TNF-R1, CRP, GM-CSF, G-CSF, Ferritin, MMP-3, MMP-7, MMP-9, TIMP-1, TIMP-2, TIMP-4, Eotaxin, MIP-1a, MCP-3, and MIF-1. The bold represent statistically significant values (p< 0.05). 468 J. GOMES ET AL. (Tsiartas et al. 2012) and VEGF (Brou et al. 2012) were eval- uated by seven studies (Table 4). The overall pooled esti- mated risk for all hormonal biomarkers was OR¼ 0.55; 95%CI: 0.55–0.57 using a random model (Table 4). The estimated pooled risk for ANGPT and BDNF were sig- nificant at OR¼ 0.34: 95%CI: 0.15–0.53 and OR¼ 3.43; 95%CI: 2.08–4.88, respectively. The pooled risk estimates were also significant for CRH and b-hCG at OR¼ 1.87; 95%CI: 1.26–2.47 and OR¼ 1.62; 95%CI: 1.34–1.90, respectively. For PAPP-A and NT, the odds ratios were OR¼ 0.54; 95%CI: 0.52–0.57 and OR¼ 0.45; 95%CI: 0.27–0.63, respectively. Inflammation-related biomarkers and SGA or PTB outcomes The inflammation-related biomarkers of interest for SGA were IL-10, IL-12, IL-13 and IL-1Ra, IFN-c, CRP, and Eotaxin (Table 5). The contribution from each of these biomarkers to SGA is shown in Table 5. The overall pooled risk estimate for SGA was OR¼ 0.09; 95%CI: 0.05–0.14 (Table 5). The length of gestation analyses was conducted for preterm births at <37weeks, <35weeks, <34weeks, <33weeks, <32weeks, and <30weeks inflammation-related biomarkers (Table 5). Significant contributors for PTB <37weeks were IL-1b, IL-1Ra, IL-2, IL-6, IL-6Ra, IL-8, IL-10, TNF-b, TNF-R1, CRP, GM-CSF, MMP-9, MIP-1a, MCP-3, and MIF-1 (Table 5). The overall pooled risk estimate for PTB <37weeks was OR¼ 1.05; 95%CI: 1.03–1.08 (Table 5). Significant contributors to PTB <35weeks was Ferritin only (Table 5) and the overall pooled risk estimate for PTB <35weeks was OR¼ 0.78; 95%CI: 0.58–0.98. For PTB <34weeks, significant contributors were IL-18, MMP2, TMP1, and TMP2 (Table 5). The overall risk esti- mate was OR¼ 0.89; 95%CI: 0.87–0.92. The only significant contributor for PTB <33weeks is G-CSF and the overall risk estimate for PTB <33weeks was OR¼ 0.75; 95%CI: 0.55–0.95. For PTB <30weeks, the overall risk estimate was OR¼ 0.82; 95%CI: 0.72–0.92 (Table 5). Growth factor/hormone-related biomarkers and SGA or PTB outcomes The relationships between outcomes and growth factors/hor- mone-related biomarkers were also assessed (Table 6). For SGA statistically significant biomarkers were low b-hCG and low PAPP-A and the overall pooled risk estimate was (OR¼ 1.07; 95%CI: 0.93–1.21) and the I squared value of 77%. The length of gestation analyses was conducted for preterm births at <37weeks, <35weeks, <34weeks, <33weeks, <32weeks, and <30weeks growth factor/hormone-related biomarkers (Table 6). For PTB < 37weeks, significant contrib- utors were ANGPT2, BDNF, PAPP-A, NT-3, NT-4, and VEGF (Table 6). The overall risk estimate for PTB <37weeks was (OR¼ 0.63; 95%CI: 0.54–0.76). For PBT <35weeks, the overall risk estimate was (OR¼ 1.54; 95%CI: 1.19–1.89). The only sig- nificant contributor for PTB <32weeks was PAPP-A and the overall risk estimate was (OR¼ 2.31; 95%CI: 1.60–3.02) (Table 6). Discussion Preterm birth is a growing concern in developed countries and research has reported that billions of dollars are spent annually to treat the 540,000 premature infants that are delivered each year (Buhimschi et al. 2010). Since a shortened gestational age and decreased birth weight may lead to increased disease burden and growth restriction, preterm birth has become one of the leading causes of perinatal mor- tality and morbidity (Conde-Agudelo et al. 2011). Despite sig- nificant research being conducted and compelling evidence to suggest several underlying pathogenic pathways and fac- tors, the etiologies of PTB and SGA remain largely elusive. Over the past decade, it has become increasingly clear that infection/inflammation has a strong association with undesir- able pregnancy outcomes (Goldenberg et al. 2005). Inflammation at the maternal-fetal interface is one of the most well established causes of preterm birth (Ferguson et al. 2014). Pro-inflammatory cytokines are the most well studied pathway; however, reduced levels of anti-inflamma- tory cytokines may also be involved in mechanisms as markers of premature birth (Ferguson et al. 2014). Increased concentrations of IL-6 have been associated with increased odds of preterm birth (significantly different from control group), while lower MMP-2 levels have been associated with PTB (von Minckwitz et al. 2000; Goldenberg et al. 2001; Kouck�y et al. 2010; Kuc et al. 2010). Significantly higher levels of IL-2 were significantly associated with PTB outcome (Brou et al. 2012), while others also observed non-statistically sig- nificant increase in expression of IL-2 (von Minckwitz et al. 2000; Makhseed et al. 2003; Curry et al. 2007, 2009). While some researchers found statistically significant association (Goldenberg et al. 2001; Curry et al. 2009; Pearce et al. 2010; Brou et al. 2012; Tsiartas et al. 2012), others found non-sig- nificant associations between PTB and other inflammatory biomarkers (von Minckwitz et al. 2000; Curry et al. 2007; Ferguson et al. 2014). Three studies all found IL-1Ra to be significantly increased in PTB cases (Georgiou et al. 2011; Brou et al. 2012; Ruiz et al. 2012). IL-18, a cytokine capable of inducing Th1 and Th2 immune responses, was proposed to interact synergistically with IL-12 to induce an immunological response leading to the premature rupture of membranes and ultimately, PTB (Ekelund et al. 2008). An increase in IL-12 has been suggested to down regulate IFN-c (Kumarathasan et al. 2014), while elevated levels of CRP are inversely related to birth weight and high levels in the mother increase the odds of SGA infants, PTB as well as LBW (Ernst et al. 2011; Georgiou et al. 2011; Bhat et al. 2014). Matrix metalloproteinases, a family of zinc-dependent pro- teinases that play a role in the degradation of extracellular matrix during tissue remodeling, have been identified as important players during embryo implantation, placentation and cervical dilation during pregnancy (Mathews et al. 1999; Curry et al. 2007; Wigle et al. 2008a, 2008b; Tency et al. 2012; Dane et al. 2013). Imbalance in MMPs have been correlated with PTB and lower birth weights (Tu et al. 1998; Sesso and Franco 2010; Sorokin et al. 2010). MMP-9 at higher levels was determined to be significantly associated with higher risk of preterm births (Botsis et al. 2006; Poon et al. 2009; Kouck�y CRITICAL REVIEWS IN TOXICOLOGY 469 et al. 2010; Kuc et al. 2010; Sesso and Franco 2010; Tency et al. 2012; Tsiartas et al. 2012), but not with SGA or LBW (Sorokin et al. 2010; Karampas et al. 2014). According to Lyon et al., IL-1b increases the production of MMPs and TNF-a induces MMP activity and IL-8 aids in cer- vical remodeling membrane lysis (Lyon et al. 2010). Poon et al. (2009), observed an association between MMP-9 and TNF-R1, suggesting the important role of inflammation on PTB (Poon et al. 2009). Conversely, low levels of IL-10 com- bined with high levels of IL-Ra1 greatly increased the prob- ability of PTB (Ruiz et al. 2012). MMP-8 and IL-10 were positively correlated with CRP levels and increase in these markers correlated negatively with gestational age, however, MMP-8 did not exhibit significant increases when comparing cases to controls (Kouck�y et al. 2010; Kuc et al. 2010). Consequently, inflammatory biomarkers appear to be key regulators during pregnancy for normal infant growth and development. Pro-inflammatory biomarkers are associated with being especially important during the onset of labor and the pathways leading to labor. Often linked with hyper- tension and preeclampsia, inflammation, oxidative stress and vascular dysfunction are crucial to the analysis of preterm birth, which may result in SGA infants along with other birth complications (Buhimschi et al. 2010). A better understanding of pro-inflammatory markers, and the anti-inflammatory markers that regulate them, may help us achieve a more comprehensive biochemical pathways leading to the onset of adverse health outcomes at birth and in adulthood. Numerous hormones are responsible for the normal growth and development of a fetus, and any imbalance in in utero maternal hormones, may negatively affect the infant birth out- come. It is important to understand the hormonal biomarkers associated with pregnancy since the lack, or abundance, of any factor affecting fetal growth may potentially result in IUGR, preterm birth, and SGA infants Mattos et al. 2013. Elevated G-CSF levels were found to be associated with PTB and LBW, while high estriol levels predicted higher risk of PTB (Whitcomb et al. 2009; Ruiz et al. 2012). Conversely, several studies found low PAPP-A to be associated with SGA infants as well as PTB, and Poon et al. identified decreased levels of PAPP-A to be linked with increased MMP-9 and risk for PTB (Poon et al. 2009). Similarly, extensive research has been con- ducted on b-hCG however, only one study reported a signifi- cant association between low b-hCG levels and SGA (Kirkegaard et al. 2011). PAPP-A and b-hCG are both produced in the placenta and are present upon implantation, and its lev- els increase as the pregnancy progresses. As such, dramatic fluctuations in either biomarker may critically impact preg- nancy term and infant development. Since the maternal biochemical profile and the physiology change constantly during pregnancy, the evaluation of bio- markers must be conducted at several stages during the pregnancy. This review of literature strengthens preexisting hypotheses suggesting a link between common biomarkers, such as interleukins and various inflammatory biomarkers, and the pathways that lead to preterm birth or IUGR (de Steenwinkel et al. 2013; Haedersdal et al. 2013; Visentin et al. 2014). Although the results are still far from conclusive, fur- ther research on novel longitudinal high-content biomarkers and then the investigation of causal pathways may aid in identifying true risk factors and preventative screening meth- ods which are reliable. This systematic review demonstrates that molecular mech- anisms in the mother can play a role by affecting vascular remodeling and function as well as maternal inflammatory and immunological responses, which consequently, impact birth outcome. However, there is still no conclusive pathway linking maternal biomarkers and adverse infant outcomes. This study confirms that biological processes such as inflam- mation and hormonal changes have been strongly associated with influencing pregnancy and birth outcomes. In conclu- sion, this review may aid in identifying subtle, but clinically important effects which may not have been apparent in indi- vidual studies. Further research is required to assess the dif- ferent outcomes in relation to diverse maternal mechanistic pathways and maternal physiological parameters and bio- chemical changes should be investigated for relationships with adverse infant-related birth outcomes. Limitations First, the research question was focused on proteomic mater- nal blood biomarkers and did not include genetic biomarkers or biomarkers in other biological samples such as amniotic fluid, saliva, urine, cord blood or cervicovaginal fluid. Second, only neonatal outcomes were investigated and maternal out- comes such as preeclampsia and gestational diabetes were omitted. As a result, this systematic review does not cover information on other biomarkers found in maternal samples that may have potentially shown significant relationships with adverse birth outcomes. Also a number of studies, namely for IUGR, were excluded because they did not report sufficient information to calculate pooled odds ratios, which may have resulted in a potential loss of relevant data. Thirdly, although noted in the summary or articles, due to lack of information in selected studies, this study could not factor in the potential role of race and ethnicity, pregnancy stage, co-morbidities, and maternal age into the assessment of outcome association. Finally, several biomarkers, such as the matrix metallopro- teinases, are still very novel and the resulting number of studies was too small to draw a definitive conclusion. This review and meta-analysis attempts to broaden the under- standing of maternal blood biomarkers in hopes of aiding the discovery of mechanistic pathways underlying pregnancy and birth complications. This study reveals that biomarkers are complex in nature and all markers, are in some way net- worked, influencing one another, so further research should be conducted to investigate the interdependencies among markers. Predicting adverse birth outcomes is important because it could allow for the identification of women at highest risk, for whom appropriate, risk-specific interventions could be developed (Goldenberg et al. 2005). Future work should include multiple biological samples and should also investigate maternal outcomes such as preeclampsia or ges- tational diabetes. 470 J. GOMES ET AL. Conclusion Elevated levels of pro-inflammatory biomarkers in maternal plasma namely, CRP, Eotaxin, G-CSF, TNF-a, IFN-c, IL-1b, IL-6, IL-6R, IL-8, MIF, MIP, and MMP have all revealed positive asso- ciations with adverse infant outcomes such as PTB and LBW. Upregulation of hormonal biomarkers such as BDNF, CRH, cortisol, and b-hCG are also associated with PTB and LBW. The ability to predict adverse birth outcome cases based on biomarker composition may help acquire important insights into the mechanisms or pathways that lead to a preterm birth or small-for-gestational age infants. Moreover, the study of maternal biomarkers has become a growing interest to improve the diagnosis of PTB and IUGR cases. Acknowledgements The authors acknowledge the reviewers for their comments on the manuscript. Their suggestions were helpful and contributed to the clarity and accuracy of our publication. The authors thank Christine Absi for technical assistance in collecting the data for this review. Declaration of interest The authors declare to have no conflict of interest. No funding was received specifically for this project. This critical review was conducted during the normal course of the authors’ employment using institutional funding. No outside funds were used to prepare the review. This review is professional work of the authors and the views expressed are not necessarily the views of their employers. None of the authors have appeared during the last 5 years in any regulatory or legal proceedings related to the contents of this paper. Departmental funds were used to support the study. ORCID J. Gomes http://orcid.org/0000-0002-9818-5984 S. Cakmak http://orcid.org/0000-0001-9921-2107 References Abu-Saad K, Fraser D. 2010. Maternal nutrition and birth outcomes. Epidemiol Rev. 32:5–25. Bakalis SP, Poon LC, Vayna AM, Pafilis I, Nicolaides KH. 2012. C-reactive protein at 11–13 weeks’ gestation in spontaneous early preterm deliv- ery. J Matern Fetal Neonatal Med. 25:2475–2478. Benford DJ, Hanley AB, Bottrill K, Oehlschlager S, Balls M, Branca F, Castegnaro JJ, Descotes J, Hemminiki K, Lindsay D, et al. 2000. Biomarkers as predictive tools in toxicity testing: the report and rec- ommendations of ECVAM workshop. Altern Lab Anim. 28:119–131. Bhat G, Williams SM, Saade GR, Menon R. 2014. Biomarker interactions are better predictors of spontaneous preterm birth. Reprod Sci. 21: 340–350. Biostat 2011. Comprehensive Meta-Analysis version, 2. Botsis D, Makrakis R, Papagianni V, Kouskouni E, Grigoriou O, Dendrinos S, Creatsas G. 2006. The value of cervical length and plasma pro- MMP-9 levels for the prediction of preterm delivery in pregnancy women presenting with threatened preterm labor. Eur J Obstet Gynecol Reprod Biol. 128:108–112. Brou L, Almli LM, Pearce BD, Bhat G, Drobek CO, Fortunato S, Menon R. 2012. Dysregulated biomarkers induce distinct pathways in preterm birth. BJOG. 119:458–473. Buhimschi CS, Schatz F, Krikun G, Buhimschi IA, Lockwood CJ. 2010. Novel insights into molecular mechanisms of abruption-induced pre- term birth. Expert Rev Mol Med. 12:1–19. Cemgil Arikan D, Aral M, Coskun A, Ozer A. 2012. Plasma IL-4, IL-8, IL-12, interferon-gamma and CRP levels in pregnant women with pree- clampsia, and their relation with severity of disease and fetal birth weight. J Matern Fetal Neonatal Med. 25:1569–1573. Conde-Agudelo A, Papageorghiou AT, Kennedy SH, Villar J. 2011. Novel biomarkers for predicting spontaneous preterm birth phenotype: a systematic review and meta-analysis. BJOG. 118:1042–1054. Conde-Agudelo A, Papageorghiou AT, Kennedy SH, Villar J. 2013. Novel biomarkers for predicting intrauterine growth restriction: a systematic review and meta-analysis. Bjog. 120:681–694. Coussons-Read ME, Lobel M, Carey JC, Kreither MO, D’Anna K, Argys L, Ross RG, Brandt C, Cole S. 2012. The occurrence of preterm delivery is linked to pregnancy-specific distress and elevated inflammatory markers across gestation. Brain Behav Immun. 26:650–659. Curry AE, Thorsen P, Drews C, Schendel D, Skogstrand K, Flanders WD, Hougaard D, Olsen J, Vogel I. 2009. First-trimester maternal plasma cytokine levels, pre-pregnancy body mass index, and spontaneous preterm delivery. Acta Obstet Gynecol Scand. 88:332–342. Curry AE, Vogel I, Drews C, Schendel D, Skogstrand K, Dana Flanders W, Hougaard D, Olsen J, Thorsen P. 2007. Mid-pregnancy maternal plasma levels of interleukin 2, 6, and 12, tumor necrosis factor-alpha, interferon-gamma, and granulocyte-macrophage colony-stimulating factor and spontaneous preterm delivery. Acta Obstet Gynecol Scand. 86:1103–1110. Dagouassat M, Lanone S, Boczkowski J. 2012. Interaction of matrix metal- loproteinases with pulmonary pollutants. Eur Respir J. 39:1021–1032. Dane B, Dane C, Batmaz G, Ates S, Dansuk R. 2013. First trimester mater- nal serum pregnancy-associated plasma protein-A is a predictive fac- tor for early preterm delivery in normotensive pregnancies. Gynecol Endocrinol. 29:592–595. de Steenwinkel FDO, Hokken-Koelega ACS, de Man YA, de Rijke YB, de Ridder MAJ, Hazes JMW, Dolhain RJEM. 2013. Circulating maternal cytokines influence fetal growth in pregnant women with rheumatoid arthritis. Ann Rheum Dis. 72:1995–2001. Dibble S, Andersen A, Lassen MR, Cunanan J, Hoppensteadt D, Fareed J. 2014. Inflammatory and procoagulant cytokine levels during preg- nancy as predictors of adverse obstetrical complications. Clin Appl Thromb Hemost. 20:152–158. Downs SH, Black N. 1998. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions . J Epidemiol Commun Health. 52:377–384. Ekelund CK, Vogel I, Skogstrand K, Thorsen P, Hougaard DM, Langhoff- Roos J, Jacobsson B. 2008. Interleukin-18 and interleukin-12 in mater- nal serum and spontaneous preterm delivery. J Reprod Immunol. 77: 179–185. Ernst GDS, de Jonge LL, Hofman A, Lindemans J, Russcher H, Steegers EAP, Jaddoe VWV. 2011. C-reactive protein levels in early pregnancy, fetal growth patterns, and the risk for neonatal complications: the Generation R Study. Am J Obstet Gynecol. 205:132.e1–132.12. Ferguson KK, McElrath TF, Chen YH, Mukherjee B, Meeker JD. 2014. Longitudinal profiling of inflammatory cytokines and C-reactive pro- tein during uncomplicated and preterm pregnancy. Am J Reprod Immunol. 72:326–336. Fleischer NL, Merialdi M, van Donkelaar A, Vadillo-Ortega F, Martin RV, Betran AP, Souza JP. 2014. Outdoor air pollution, preterm birth, and low birth weight: analysis of the world health organization global sur- vey on maternal and perinatal health. Environ Health Perspect. 122: 425–430. Gelson E, Curry R, Gatzoulis MA, Swan L, Lupton M, Steer P, Johnson M. 2011. Effect of maternal heart disease on pregnancy outcomes. Expert Rev Obstet Gynecol. 117:886–891. Georgiou HM, Thio YS, Russell C, Permezel M, Heng YJ, Lee S, Tong S. 2011. Association between maternal serum cytokine profiles at 7–10 weeks’ gestation and birthweight in small-for-gestational age infants. Am J Obstet Gynecol. 204:415.e1–415.e12. Godfrey K, Robinson S, Barker DJ, Osmond C, Cox V. 1996. Maternal nutrition in early and late pregnancy in relation to placental and fetal growth. BMJ. 312:410–414. CRITICAL REVIEWS IN TOXICOLOGY 471 Goetzinger KR, Cahill AG, Macones GA, Odibo AO. 2010. Association of first-trimester low papp-a levels with preterm birth. Prenat Diagn. 30: 309–313. Goldenberg RL, Andrews WW, Mercer BM, Moawad AH, Meis PJ, Iams JD, Das A, Caritis SN, Roberts JM, Miodovnik M, et al. 2000. The preterm prediction study: granulocyte colony-stimulating factor and spontan- eous preterm birth. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol. 182:625–630. Goldenberg RL, Goepfert AR, Ramsey PS. 2005. Biochemical markers for the prediction of preterm birth. Am J Obstet Gynecol. 192:36–46. Goldenberg RL, Iams JD, Mercer BM, Meis PJ, Moawad A, Das A, Miodovnik M, VanDorsten PJ, Caritis SN, Thurnau G, et al. 2001. The preterm prediction study: toward a multiple-marker test for spontan- eous preterm birth. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol. 185:643–651. Hackshaw A, Rodeck C, Boniface S. 2011. Maternal smoking in pregnancy and birth defects: a systematic review based on 173 687 malformed cases and 11.7 million controls. Hum Reprod Update. 17:589–604. Haedersdal S, Salvig JD, Aabye M, Thorball CW, Ruhwald M, Ladelund S, Eugen-Olsen J, Secher NJ. 2013. Inflammatory markers in the second trimester prior to clinical onset of preeclampsia, intrauterine growth restriction, and spontaneous preterm birth. Inflammation. 36:907–913. Hee L, Kirkegaard IDA, Vogel IDA, Thorsen P, Skogstrand K, Hougaard DM, Uldbjerg N, Sandager PUK. 2011. Low serum interleukin-17 is associated with preterm delivery. Acta Obstet Gynecol Scand. 90: 92–96. Howe LD, Matijasevich A, Tilling K, Brion M-J, Leary SD, Smith GD, Lawlor DA. 2012. Maternal smoking during pregnancy and offspring trajecto- ries of height and adiposity: comparing maternal and paternal associa- tions. Int J Epidemiol. 41:722–732. Jelliffe-Pawlowski LL, Ryckman KK, Bedell B, O’Brodovich HM, Gould JB, Lyell DJ, Borowski KS, Shaw GM, Murray JC, Stevenson DK, et al. 2014. Combined elevated midpregnancy tumor necrosis factor alpha and hyperlipidemia in pregnancies resulting in early preterm birth. Am J Obstet Gynecol. 211:141.e1–141.e9. Jolly M, Sebire N, Harris J, Robinson S, Regan L. 2000. The risks associ- ated with pregnancy in women aged 35 years or older. Hum Reprod. 15:2433–2437. Karampas G, Eleftheriades M, Panoulis K, Rizou M, Haliassos A, Hassiakos D, Vitoratos N, Rizos D. 2014. Maternal serum levels of neutrophil gelatinase-associated lipocalin (NGAL), matrix metalloproteinase-9 (MMP-9) and their complex MMP-9/NGAL in pregnancies with pree- clampsia and those with a small for gestational age neonate: a longi- tudinal study. Prenat Diagn. 34:726–733. Keith JC Jr, Pijnenborg R, Luyten C, Spitz B, Schaub R, Van Assche FA. 2000. Maternal serum levels of macrophage colony-stimulating factor are associated with adverse pregnancy outcome. Eur J Obstet Gynecol Reprod Biol. 89:19–25. Kim MA, Lee BS, Park YW, Seo K. 2011. Serum markers for prediction of spontaneous preterm delivery in preterm labour. Eur J Clin Invest. 41: 773–780. King JC. 2003. The risk of maternal nutritional depletion and poor out- comes increases in early or closely spaced pregnancies. J Nutr. 133: 1732S–1736S. Kirkegaard I, Henriksen TB, Tørring N, Uldbjerg N. 2011. Papp-a and free b-hCG measured prior to 10 weeks is associated with preterm delivery and small-for-gestational-age infants. Prenat Diagn. 31:171–175. Kouck�y M, Germanov�a A, Kalousov�a M, Hill M, Cindrov�a-Davies T, Pa�r�ızek A, Svarcov�a J, Zima T, H�ajek Z. 2010. Low maternal serum matrix met- alloproteinase (MMP)-2 concentrations are associated with preterm labor and fetal inflammatory response. J Perinat Med. 38:589–596. Kuc P, Lemancewicz A, Laudanski P, Kretowska M, Laudanski T. 2010. Total matrix metalloproteinase-8 serum levels in patients labouring preterm and patients with threatened preterm delivery. Folia Histochem Cytobiol. 48:366–370. Kumarathasan P, Vincent R, Das D, Mohottalage S, Blais E, Blank K, Karthikeyan S, Vuong NQ, Arbuckle TE, Fraser WD, et al. 2014. Applicability of a high-throughput shotgun plasma protein screening approach in understanding maternal biological pathways relevant to infant birth weight outcome. J Proteomics. 100:136–146. Laudanski P, Raba G, Kuc P, Lemancewicz A, Kisielewski R, Laudanski T. 2012. Assessment of the selected biochemical markers in predicting preterm labour. J Matern Fetal Neonatal Med. 25:2696–2699. Lohsoonthorn V, Qiu C, Williams MA. 2007. Maternal serum C-reactive protein concentrations in early pregnancy and subsequent risk of pre- term delivery. Clin Biochem. 40:330–335. Lyon D, Cheng C-Y, Howland L, Rattican D, Jallo N, Pickler R, Brown L, McGrath J. 2010. Integrated review of cytokines in maternal, cord, and newborn blood: Part I – associations with preterm birth. Biol Res Nurs. 11:371–376. Makhseed M, Raghupathy R, El-Shazly S, Azizieh F, Al-Harmi JA, Al-Azemi M. 2003. Pro-inflammatory maternal cytokine profile in preterm deliv- ery. Am J Reprod Immunol. 49:308–318. Mathews F, Yudkin P, Neil A. 1999. Influence of maternal nutrition on outcome of pregnancy: prospective cohort study. Br Med J. 319: 339–343. Mattos SS, Chaves MEC, Costa SMR, Ishigami ACM, Rêgo SB, Souto Maior V, Severi R, de Lima Filho JL. 2013. Which growth criteria better pre- dict fetal programming? Arch Dis Child Fetal Neonatal Ed. 98:F81–F84. Moghaddam BL, Mohamadi B, Asghari Jaafarabadi M, Aliyan Moghadam N. 2012. Maternal serum C-reactive protein in early pregnancy and occurrence of preterm premature rupture of membranes and preterm birth. J Obstet Gynaecol Res. 38:780–786. Moher D, Liberati A, Tetzlaff J, Altman DG, and PRISMA Group. 2009a. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ (Clin Res Ed.). 339:b2535. Moher D, Liberati A, Tetzlaff J, Altman DG, and PRISMA Group. 2009b. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 6:e1000097. Moher D, Liberati A, Tetzlaff J, Altman DG, and PRISMA Group. 2009c. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol. 62:1006–1012. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA, and PRISMA-P Group. 2015. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 4:MR000031. Murtha AP, Greig PC, Jimmerson CE, Herbert WN. 1998. Maternal serum interleukin-6 concentration as a marker for impending preterm deliv- ery. Obstet Gynecol. 91:161–164. Ozgu-Erdinc AS, Cavkaytar S, Aktulay A, Buyukkagnici U, Erkaya S, Danisman N. 2014. Mid-trimester maternal serum and amniotic fluid biomarkers for the prediction of preterm delivery and intrauterine growth retardation. J Obstet Gynaecol Res. 40:1540–1546. Paternoster DM, Stella A, Gerace P, Manganelli F, Plebani M, Snijders D, Nicolini U. 2002. Biochemical markers for the prediction of spontan- eous pre-term birth. Int J Gynaecol Obstet. 79:123–129. Pearce BD, Garvin SE, Grove J, Bonney EA, Dudley DJ, Schendel DE, Thorsen P. 2008. Serum macrophage migration inhibitory factor in the prediction of preterm delivery. Am J Obstet Gynecol. 199:46e1–4636. Pearce BD, Grove J, Bonney EA, Bliwise N, Dudley DJ, Schendel DE, Thorsen P. 2010. Interrelationship of cytokines, hypothalamic-pituitary- adrenal axis hormones, and psychosocial variables in the prediction of preterm birth. Gynecol Obstet Invest. 70:40–46. Pearce BD, Nguyen PH, Gonzalez-Casanova I, Qian Y, Omer SB, Martorell R, Ramakrishnan U. 2016. Pre-pregnancy maternal plasma cytokine levels and risks of small-for-gestational-age at birth. J Matern Fetal Neonatal Med. 1:1–5. Perera FP, Tang D, Rauh V, Tu YH, Tsai WY, Becker M, Stein JL, King J, Priore GD, Lederman SA, et al. 2007. Relationship between polycyclic aromatic hydrocarbon-DNA adducts, environmental tobacco smoke, and child development in the World Trade Center cohort. Environ Health Perspect. 115:1497–1502. Pitiphat W, Gillman MW, Joshipura KJ, Williams PL, Douglass CW, Rich- Edwards JW. 2005. Plasma C-reactive protein in early pregnancy and preterm delivery. Am J Epidemiol. 162:1108–1113. Poon LC, Nekrasova E, Anastassopoulos P, Livanos P, Nicolaides KH. 2009. First-trimester maternal serum matrix metalloproteinase-9 (MMP-9) and adverse pregnancy outcome. Prenat Diagn. 29:553–559. 472 J. GOMES ET AL. Ruiz RJ, Jallo N, Murphey C, Marti CN, Godbold E, Pickler RH. 2012. Second trimester maternal plasma levels of cytokines IL-1Ra, Il-6 and IL-10 and preterm birth. J Perinatol. 32:483–490. Sattar N, McConnachie A, O’Reilly D, Upton MN, Greer IA, Smith GD, Watt G. 2004. Inverse association between birth weight and C-reactive pro- tein concentrations in the MIDSPAN Family Study. ATVB. 24:583–587. Sesso R, Franco M. 2010. Abnormalities in metalloproteinase pathways and IGF-I axis: a link between birthweight, hypertension, and vascular damage in childhood. Am J Hypertens. 23:6–11. Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Petticrew M, and PRISMA-P Group. 2015. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explan- ation. BMJ (Clin Res Ed.). 350:g7647. Sorokin Y, Romero R, Mele L, Wapner RJ, Iams JD, Dudley DJ, Spong CY, Peaceman AM, Leveno KJ, Harper M, et al. 2010. Maternal serum inter- leukin-6, c-reactive protein, and matrix metalloproteinase-9 concentra- tions as risk factors for preterm birth <32 weeks and adverse neonatal outcomes. Am J Perinatol. 27:631–640. Spencer K, Cowans NJ, Avgidou K, Molina F, Nicolaides KH. 2008. First-tri- mester biochemical markers of aneuploidy and the prediction of small-for-gestational age fetuses. Ultrasound Obstet Gynecol. 31: 15–19. Stewart LA, Clarke M, Rovers M, Riley RD, Simmonds M, Stewart G, Tierney JF, and PRISMA-IPD Development Group. 2015. Preferred reporting items for systematic review and meta-analyses of individual participant data: the PRISMA-IPD Statement. JAMA. 313:1657–1665. Stillerman KP, Mattison DR, Giudice LC, Woodruff TJ. 2008. Environmental exposures and adverse pregnancy outcomes: a review of the science. Reprod Sci. 15:631–650. Tency I, Verstraelen H, Kroes I, Holtappels G, Verhasselt B, Vaneechoutte M, Verhelst R, Temmerman M. 2012. Imbalances between matrix met- alloproteinases (MMPs) and tissue inhibitor of metalloproteinases (TIMPs) in maternal serum during preterm labor. PLoS One. 7:e49042. Tsiartas P, Holst RM, Wennerholm UB, Hagberg H, Hougaard DM, Skogstrand K, Pearce BD, Thorsen P, Kacerovsky M, Jacobsson B, et al. 2012. Prediction of spontaneous preterm delivery in women with threatened preterm labour: a prospective cohort study of multiple proteins in maternal serum. BJOG. 119:866–873. Tu FF, Goldenberg RL, Tamura T, Drews M, Zucker SJ, Voss HF. 1998. Prenatal plasma matrix metalloproteinase-9 levels to predict spontan- eous preterm birth. Obstet Gynecol. 92:446–449. Turhan NO, Karabulut A, Adam B. 2000. Maternal serum interleukin 6 lev- els in preterm labor: prediction of admission-to-delivery interval. J Perinat Med. 28:133–139. Van Ravenswanij R, Tesselaar-van der Goot M, de Wolf S, van Leeuwen- Spruijt M, Visser GHA, Schielen P. 2011. First-trimester serum papp-a and fB-hCG concentrations and other maternal characteristics to establish logistic regression-based predictive rules for adverse preg- nancy outcome. Prenat Diagn. 31:50–57. Visentin S, Lapolla A, Londero AP, Cosma C, Dalfr�a M, Camerin M, Faggian D, Plebani M, Cosmi E. 2014. Adiponectin levels are reduced while markers of systemic inflammation and aortic remodelling are increased in intrauterine growth restricted mother-child couple. Biomed Res Int. 2014:401595. von Minckwitz G, Grischke EM, Schwab S, Hettinger S, Loibl S, Aulmann M, Kaufmann M. 2000. Predictive value of serum interleukin-6 and -8 levels in preterm labor or rupture of the membranes. Acta Obstet Gynecol Scand. 79:667–672. Whitcomb BW, Schisterman EF, Luo X, Chegini N. 2009. Maternal serum granulocyte colony-stimulating factor levels and spontaneous preterm birth. J Womens Health (Larchmt). 18:73–78. Wigle DT, Arbuckle TE, Turner MC, Berube A, Yang Q, Liu S, Krewski D. 2008a. Epidemiologic evidence of relationships between reproductive and child health outcomes and environmental chemical contaminants. J Toxicol Environ Health B Crit Rev. 11:373–517. Wigle DT, Turner MC, Gomes J, Parent M-E. 2008b. Role of hormonal and other factors in human prostate cancer. J Toxicol Environ Health B Crit Rev. 11:242–259. Woodruff TJ, Parker JD, Darrow LA, Slama R, Bell ML, Choi H, Glinianaia S, Hoggatt KJ, Karr CJ, Lobdell DT, et al. 2009. Methodological issues in studies of air pollution and reproductive health. Environ Res. 109: 311–320. CRITICAL REVIEWS IN TOXICOLOGY 473 5.9 APPENDIX Table A1. Study Quality Checklist Factors for Evaluation of Study Quality Scores a) Reporting 1. Is the hypothesis/aim/objective of the study clearly described? /1 2. Are the main outcomes to be measured clearly described in the Introduction or Methods section? /1 3. Are the exposure measures clearly mentioned and distinguished? (case vs. control) /1 4. Have actual probability values been reported for the main outcomes except where the probability value is less than 0.001? /1 Subtotal for Reporting /4 b) External Validity 1. Were the study subjects recruited to participate in the study representative of the entire population of interest? /1 2. Were those subjects who participated in the study similar to those who did not and were they representative of the entire population? /1 Subtotal for External Validity /2 c) Internal Validity – Bias 1. Were the research subjects and the interviewers blinded when conducting the study? /1 2. Were the results of this study based on data dredging or re-analyzed data? /1 3. Were the time periods between exposure and outcome consistent and constant for the subjects? /1 4. Were the statistical methods appropriate for the study? /1 5. Were the exposures determined reliably and were clearly defined and based on pre-existing or documented information? /1 6. Were the outcome measures accurately defined and valid? /1 Subtotal for Internal Validity – Bias /6 d) Internal Validity - Confounding 1. Were the different exposure groups and unexposed groups recruited from the same population? /1 2. Were the different exposure groups recruited at the same time? /1 3. Were confounders accounted for and adequate adjustments made for confounding in the data analysis process? /1 Subtotal for Internal Validity – Confounding /3 Total Score /15 Table A2. Summative Quality Scores. # Author, Year Reporting External Validity Internal Validity TotalBias Confounding 1 Bakalis et al., 2012 4 1 4 2 11 2 Bhat et al., 2014 4 1 4 2 11 3 Botsis et al., 2006 3 1 4 1 9 4 Brou et al., 2012 4 1 4 2 11 5 Curry, Thorsen et al., 2009 4 1 4 2 11 6 Curry, Vogel et al., 2007 4 1 4 2 11 7 Ekelund et al., 2008 4 1 4 2 11 8 Ernst et al., 2011 4 1 5 2 12 9 Ferguson et al., 2014 4 1 4 2 11 10 Georgiou et al., 2011 4 1 4 1 10 11 Goldenberg, Andrews et al., 2000 4 1 5 2 12 12 Goldenberg, Iams et al., 2001 4 1 5 2 12 13 Hee et al., 2011 4 1 4 2 11 14 Jeliffe-Paulowsi et al., 2014 4 1 4 2 11 15 Laudanski et al., 2012 4 1 4 2 11 16 Lohsoonthorn et al., 2007 4 1 5 1 11 17 Moghaddam Banaem et al., 2012 4 1 4 2 11 18 Ozgu-Erdinc et al, 2014 4 1 4 2 11 19 Paternoster et al., 2002 4 1 5 1 11 20 Pearce, Garvin et al., 2008 4 1 5 1 11 21 Pearce, Grove et al., 2009 4 1 4 2 11 22 Pearce, Nguyen et al., 2016 4 1 5 1 11 23 Pitiphat et al., 2005 4 1 4 2 11 24 Ruiz et al., 2012 4 1 4 2 11 25 Tency et al., 2012 4 1 4 2 11 26 Tsiartas et al., 2000 4 1 5 2 12 27 Turhan et al., 2000 4 1 4 2 11 28 Von Minckwitz et al. 2000 4 1 5 1 11 29 Whitcomb et al., 2009 4 1 4 1 10 30 Dane et al., 2013 4 1 4 2 11 31 Kirkegaard et al., 2011 4 1 4 2 11 32 Goetzinger et al., 2010 4 1 4 2 11 33 Spencer et al, 2008 4 1 4 2 11 34 Sorokin et al, 2010 4 2 5 1 12 35 Van Ravenswaaij 4 2 5 2 13 36� Cemgil Arkan et al., 2012 4 1 4 1 10 37� Coussons-Read et al., 2012 4 1 5 2 12 38� De Steenwinkel et al., 2013 4 0 5 2 11 39� Dibble et al., 2014 4 1 5 1 11 40� Fransson et al., 2011 3 1 4 2 10 41� Haedersdal et al., 2013 4 2 5 1 12 (continued) 474 J. GOMES ET AL. Table A2. Continued. # Author, Year Reporting External Validity Internal Validity TotalBias Confounding 42� Karampas et al., 2014 4 2 4 1 11 43� Keith et al., 2000 4 1 4 1 10 44� Kim et al., 2011 4 1 4 2 11 45� Koucky et al., 2010 4 1 5 1 11 46� Ku�c et al., 2010 3 1 5 1 10 47� Lyon et al., 2010 3 1 5 2 11 48� Makhseed et al., 2003 3 1 4 1 9 49� Mattos et al., 2011 4 1 4 1 10 50� Murtha et al., 1998 3 1 5 3 12 51� Poon et al., 2009 4 2 5 1 12 52� Sattar et al., 2004 4 2 4 2 12 53� Sesso et al., 2010 4 1 4 2 11 54� Visentin et al., 2014 4 1 4 1 10 � Not included in meta-analysis due to insufficient data or odds ratios, confidence intervals and p-values were not reported�� Used modified Black & Downs checklist as see in supplementary material 1. Inflammatory-related biomarkers Interleukin-1β, -2, -4, -5, -6, -7, -8, -9, -12, -13, -17 and -18 (IL) Transforming growth factor-α and –β (TGFa / b) Interferon-γ (INFg) Tumour necrosis factor-α (TNFa) Intercellular cell adhesion molecule (ICAM) Vascular cell adhesion molecule (VCAM) C-reactive protein (CRP) Granulocyte macrophage colony-stimulating factor (GM-CSF) Granulocyte colony-stimulating factor (G-CSF) Ferritin Matrix metalloproteinase-2, -3, -7, -8, and -9 (MMP) Tissue inhibitor metalloproteinase-1, -2, and -4 (TIMP) Eotaxin Macrophage inflammatory protein -1 (MIP) Monocyte chemotactic protein-1 and -3 (MCP) Macrophage migration inhibitory factor (MIF) Interleukin-1R (IL-1R) Tumour necrosis factor receptor-1 (TNF R1) IL-6R and -10 (IL) 2. Growth factor/hormone-related biomarkers Corticotropin releasing hormone (CRH) Brain derived neurotrophic factor (BDNF) Neurotrophin-3 and -4 (NT) β-human chorionic gonadotropin (b-hCG) Pregnancy-associated Plasma Protein A (PAPP) Cortisol Fetal growth factor (FGF) Vascular endothelial growth factor (VEGF) Insulin-growth-factor-binding protein-1 (IGF-BP) Angiopoietin2 (ANG-PT2) Figure A1. Biomarkers Related to Adverse Birth Outcomes Identified in Literature. CRITICAL REVIEWS IN TOXICOLOGY 475 https://doi.org/ Table A3. Summary of Findings for Elevated Inflammation-Related Biomarkers Found in Maternal Blood. Outcome Biomarkera Odds Ratio 95%CI LL 95%CI UL Time of Sampling Study SGA IL-10 0.18 0.01 3.04 Unknown Pearce, Nguyen et al. IL-12 3.62� 1.74 7.54 Unknown Georgiou et al. IL-13 0.11� 0.02 0.66 Unknown Pearce, Nguyen et al. IL-1Ra 5.32� 2.40 11.79 Unknown Georgiou et al. IFN-c 3.72� 1.54 9.03 Unknown Georgiou et al. IFN-c 0.06� 0.01 0.75 Unknown Pearce, Nguyen et al. CRP 1.76 1.00 3.11 <18w Ernst et al. Eotaxin 2.32 0.98 5.46 Unknown Georgiou et al. LBW CRP 1.05 0.51 2.17 <18w Ernst et al. PTB <37w IL-1b 1.00 0.83 1.20 Unknown Ferguson et al. IL-1b 3.74� 1.31 10.70 Delivery Brou et al. IL-1b 1.25 0.37 4.23 Delivery Von Minckwitz et al. IL-1Ra 2.55� 1.24 5.24 22-24w Ruiz et al. IL-1Ra 3.08� 1.16 8.20 Delivery Brou et al. IL-2 1.20 0.93 1.55 <16w Curry, Thorsen et al. IL-2 0.99 0.73 1.33 <24w Curry, Vogel et al. IL-2 4.02� 1.53 10.52 Delivery Brou et al. IL-2 2.37 0.12 47.54 Delivery Von Minckwitz IL-4 0.90 0.50 1.70 Delivery Tsiartas et al. IL-4 0.32 0.01 16.79 Delivery Von Minckwitz et al. IL-6 1.29� 1.07 1.56 Unknown Ferguson et al. IL-6 1.15 0.89 1.48 <16w Curry, Thorsen IL-6 1.45 0.93 2.28 22-24w Ruiz et al. IL-6 1.31 0.99 1.75 <24w Curry, Vogel et al. IL-6 1.50 0.80 2.90 <24w Pearce, Grove et al. IL-6 1.27� 1.02 1.59 Delivery Turhan et al. IL-6 30.80� 3.93 241.20 Delivery Von Minckwitz et al. IL-6R 2.72� 1.15 6.48 Delivery Bhat et al. IL-6R 2.02 0.71 5.70 Delivery Brou et al. IL-6R 2.70� 1.20 5.80 Delivery Tsiartas et al. IL-8 3.75� 1.22 11.50 Delivery Bhat et al. IL-8 1.50 0.58 3.87 Delivery Brou et al. IL-8 5.25� 1.14 24.27 Delivery Von Minckwitz et al. IL-10 2.52 0.50 12.78 22-24w Ruiz et al. IL-10 2.30� 1.20 4.50 <24w Pearce, Grove et al. IL-10 4.60� 2.20 9.70 Delivery Tsiartas et al. IL-12 0.60 0.20 1.65 Delivery Tsiartas et al. IL-17 1.80 0.80 3.80 Delivery Tsiartas et al. IL-18 0.50 0.20 1.20 Delivery Tsiartas et al. TGF-b 1.70 0.80 3.40 Delivery Tsiartas et al. IFN-c 0.84 0.64 1.10 <16w Curry, Thorsen et al. IFN-c 1.36� 1.03 1.81 <24w Curry, Vogel et al. IFN-c 1.60 0.80 3.10 Delivery Tsiartas et al. TNF-a 1.16 0.81 1.66 Unknown Ferguson et al. TNF-a 1.32� 1.03 1.70 <16w Curry, Thorsen et al. TNF-a 1.10 0.83 1.47 <24w Curry, Vogel et al. TNF-a 2.40� 1.30 4.70 <24w Pearce, Grove et al. TNF-a 6.86� 2.42 19.42 Delivery Brou et al. TNF-a 1.60 0.80 3.20 Delivery Tsiartas et al. TNF-a 3.09 0.06 159.83 Delivery Von Minckwitz et al. TNF-b 2.90� 1.40 6.00 Delivery Tsiartas et al. TNF-R1 1.21 0.45 3.28 Delivery Bhat et al. TNF-R1 6.42� 2.24 18.41 Delivery Brou et al. TNF-R1 3.70� 1.60 8.50 Delivery Tsiartas et al. CRP 3.71 1.99 6.91 <37w Sorokin et al CRP 1.07 0.84 1.37 Unknown Ferguson et al. CRP 2.04� 1.13 3.69 Unknown Lohsoonthorn et al. CRP 8.96� 4.60 17.43 Unknown Moghaddam Banaem et al. CRP 1.34 0.67 2.66 <18w Ernst et al. CRP 1.56 0.89 2.75 <19w Pitiphat et al. CRP 1.90 1.00 3.70 <24w Pearce, Grove et al. GM-CSF 1.35� 1.05 1.75 <16w Curry, Thorsen et al. GM-CSF 1.11 0.83 1.48 <24w Curry, Vogel et al. GM-CSF 0.40� 0.20 0.80 Delivery Tsiartas et al. G-CSF 1.50 0.60 3.90 22-24w & 28-29w Goldenberg, Andrews et al. Ferritin 1.05� 1.00 1.09 16-22w Ozgu-Erdinc et al. Ferritin 2.12 0.60 7.50 24w Paternoster et al. MMP-7 1.62 0.63 4.20 Delivery Brou et al. MMP 096 0.46 2.0 <37w Sorokin et al MMP-9 33.75� 5.84 194.80 Unknown Botsis et al. MMP-9 6.00� 2.90 12.70 Delivery Tsiartas et al. TIMP-1 2.31 0.89 6.04 Delivery Brou et al. Eotaxin 3.16� 1.13 8.84 Unknown Laudanski et al. MIP-1a 2.62 0.95 7.21 Unknown Laudanski et al. (continued) 476 J. GOMES ET AL. Table A3. Continued. Outcome Biomarkera Odds Ratio 95%CI LL 95%CI UL Time of Sampling Study MIP-1a 1.36 0.49 3.79 Delivery Bhat et al. MIP-1a 2.98� 1.13 7.86 Delivery Brou et al. MIP-1a 1.70 0.80 3.40 Delivery Tsiartas et al. MIP-1b 2.50� 1.20 5.10 Delivery Tsiartas et al. MCP-3 4.44� 1.58 12.46 Delivery Brou et al. MCP-1 0.50 0.20 1.20 Delivery Tsiartas et al. MIF-1 4.16� 1.83 9.48 Unknown Pearce, Garvin et al. MIF-1 3.50� 1.80 6.70 <24w Pearce, Grove et al. MIF-1 4.70� 2.20 9.90 Delivery Tsiartas et al. PTB <35w IL-6 1.10 0.48 2.50 20w & 24w Goldenberg, Iams et al. IL-10 0.40 0.12 1.20 20w & 24w Goldenberg, Iams et al. ICAM-1 1.90 0.89 4.05 20w & 24w Goldenberg, Iams et al. CRP 1.20 0.50 3.11 20w & 24w Goldenberg, Iams et al. G-CSF 1.00 0.40 2.49 22-24w & 28-29w Goldenberg, Andrews et al. Ferritin 1.50 0.62 3.63 20w & 24w Goldenberg, Iams et al. Ferritin 4.85� 1.60 14.70 24w Paternoster et al. PTB <34w IL-2 1.13 0.85 1.49 <16w Curry, Thorsen et al. IL-2 1.02 0.74 1.41 <24w Curry, Vogel et al. IL-6 1.11 0.84 1.47 <16w Curry, Thorsen IL-6 1.07 0.78 1.48 <24w Curry, Vogel et al. IL-18 3.40� 1.20 9.80 Delivery Ekelund et al. IFN-c 1.00 0.76 1.33 <16w Curry, Thorsen et al. IFN-c 1.25 0.91 1.70 <24w Curry, Vogel et al. TNF-a 0.97 0.73 1.29 <16w Curry, Thorsen et al. TNF-a 1.13 0.82 1.54 <24w Curry, Vogel et al. CRP 1.06 0.41 2.71 11-13w Bakalis et al. GM-CSF 1.09 0.82 1.45 <16w Curry, Thorsen et al. GM-CSF 1.03 0.75 1.42 <24w Curry, Vogel et al. MMP-3 1.02 0.78 1.35 Delivery Tency et al. MMP-9 2.21� 1.59 3.07 Delivery Tency et al. TIMP-1 0.89� 0.84 0.95 Delivery Tency et al. TIMP-2 0.78� 0.73 0.84 Delivery Tency et al. TIMP-4 1.08 0.97 1.19 Delivery Tency et al. PTB <33w IL-17 0.37 0.11 1.26 19w Hee et al. G-CSF 1.42� 1.01 1.99 Unknown Whitcomb et al. PTB <32w IL-6 2.43� 1.29 4.59 Unknown Sorokin et al. IL-6 1.30 0.32 5.07 20w & 24w Goldenberg, Iams et al. IL-10 0.50 0.11 1.99 20w & 24w Goldenberg, Iams et al. ICAM-1 4.90� 1.30 18.81 20w & 24w Goldenberg, Iams et al. CRP 3.71� 1.99 6.91 Unknown Sorokin et al. CRP 1.70 0.39 7.71 20w & 24w Goldenberg, Iams et al. G-CSF 0.80 0.20 3.10 22-24w & 28-29w Goldenberg, Andrews et al. Ferritin 8.00 0.94 67.46 20w & 24w Goldenberg, Iams et al. Ferritin 1.20 0.20 7.10 24w Paternoster et al. MMP-9 0.96 0.46 2.00 Unknown Sorokin et al. PTB <30w IL-2 0.94 0.58 1.52 <16w Curry, Thorsen et al. IL-2 1.35 0.72 2.51 <24w Curry, Vogel et al. IL-6 0.98 0.61 1.56 <16w Curry, Thorsen IL-6 1.15 0.62 2.13 <24w Curry, Vogel et al. IFN-c 0.61 0.36 1.03 <16w Curry, Thorsen et al. IFN-c 1.08 0.57 2.04 <24w Curry, Vogel et al. TNF-a 0.37 0.11 1.26 Unknown Jelliffe-Pawlowski et al. TNF-a 0.73 0.44 1.21 <16w Curry, Thorsen et al. TNF-a 1.29 0.71 2.35 <24w Curry, Vogel et al. GM-CSF 1.09 0.69 1.74 <16w Curry, Thorsen et al. GM-CSF 1.37 0.76 2.46 <24w Curry, Vogel et al. �p< 0.05 a For increased levels of biomarkers CRITICAL REVIEWS IN TOXICOLOGY 477 Table A4. Summary of Findings for Growth Factor/Hormone-Related Biomarkers Found in Maternal Blood. Outcome Biomarkera Odds Ratio 95%CI LL 95%CI UL Time of Sampling Study SGA low b-hCG 1.31� 1.05 1.63 Unknown Van Ravenswaij et al. low b-hCG 1.91� 1.21 3.02 8-13w Kirkegaard et al. low PAPP-A 0.54� 0.50 0.57 Unknown Spencer et al. low PAPP-A 2.30� 1.91 2.77 Unknown Van Ravenswaij et al. PTB <37w CRH 3.20� 1.70 6.20 <24w Pearce, Grove et al. BDNF 3.16� 1.13 8.84 Unknown Laudanski et al. BDNF 3.70� 1.60 8.20 Delivery Tsiartas et al. NT-3 2.30� 1.10 4.70 Delivery Tsiartas et al. NT-4 0.40� 0.20 0.85 Delivery Tsiartas et al. low b-hCG 1.50 0.95 2.38 8-13w Kirkegaard et al. low PAPP-A 1.95� 1.39 2.73 8-13w Kirkegaard et al. Cortisol 1.20 0.60 2.30 <24w Pearce, Grove et al. FGF basic 0.49 0.18 1.29 Delivery Brou et al. VEGF 3.83� 1.46 10.09 Delivery Brou et al. IGFBP-1 2.41 0.88 6.61 Unknown Laudanski et al. ANGPT2 0.32� 0.12 0.89 Delivery Brou et al. PTB <35w CRH 1.40 0.62 3.21 20w & 24w Goldenberg, Iams et al. low b-hCG 1.40 0.80 2.40 Unknown Goetzinger et al. low PAPP-A 2.00� 1.00 3.80 Unknown Goetzinger et al. Cortisol 1.70 0.73 3.87 20w & 24w Goldenberg, Iams et al. PTB <34w low b-hCG 2.00 0.25 16.30 <12w Dane et al. low PAPP-A 7.00� 1.80 27.70 <12w Dane et al. PTB <32w CRH 2.70 0.49 14.45 20w & 24w Goldenberg, Iams et al. low b-hCG 2.00 0.99 3.80 Unknown Goetzinger et al. low PAPP-A 2.70� 1.10 6.40 Unknown Goetzinger et al. Cortisol 5.40 0.61 48.40 20w & 24w Goldenberg, Iams et al. �p< 0.05 a For increased levels of biomarkers unless otherwise stated 478 J. GOMES ET AL. Abstract Introduction Methodology Literature search Study selection Study quality assessment Data extraction and data cleaning Statistical analysis Results Inflammation-related biomarkers Growth factor/hormone-related biomarkers Inflammation-related biomarkers and SGA or PTB outcomes Growth factor/hormone-related biomarkers and SGA or PTB outcomes Discussion Limitations Conclusion Acknowledgements Declaration of interest References