Association between maternal exposure to oil and gas extraction process with adverse birth outcomes: a systematic review and meta-analysis

Abstract

Background

Oil and gas extraction plays a critical role in global energy supply and economic development, but it is increasingly associated with adverse health outcomes, particularly during pregnancy. This systematic review and meta-analysis evaluates the relationship between maternal exposure to oil and gas pollutants and selected adverse birth outcomes.

Methods

A comprehensive literature search was conducted across MEDLINE/PubMed, Scopus, CINAHL, Web of Science, and the Cochrane Library up to June 16, 2024. We included studies assessing the effects of maternal exposure to oil and gas extraction processes on preterm birth (PTB), miscarriage, stillbirth, birth defects, small for gestational age (SGA), low birth weight (LBW), and birth weight (BW). Study quality was assessed using the Newcastle–Ottawa Scale and GRADE framework. From 4,235 screened articles, 24 studies met inclusion criteria.

Results

Our analyses revealed statistically significant associations between maternal exposure and increased odds of PTB (pooled OR: 1.07, 95% CI: 1.01–1.13), miscarriage (OR: 2.03, 95% CI: 1.60–2.58), SGA (OR: 1.23, 95% CI: 1.05–1.45), and reduced BW (mean difference: -30.36 g, 95% CI: -43.98 to -17.28). No significant associations were observed for stillbirth (OR: 0.98, 95% CI: 0.68–1.40), birth defects (OR: 1.14, 95% CI: 0.84–1.53), or LBW (OR: 1.07, 95% CI: 0.93–1.24). Substantial heterogeneity was present across most outcomes, and publication bias could not be ruled out in several analyses.

Conclusions

These findings suggest possible associations between maternal exposure to oil and gas extraction processes and several adverse birth outcomes, including PTB, miscarriage, SGA, and LBW. However, due to methodological variability, potential biases, and high heterogeneity among studies, these results should be interpreted with caution. Further research with standardized exposure assessments and larger, population-based cohorts is needed to confirm and refine these associations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-08022-z.

Introduction

The extraction and processing of oil and gas are essential components of the global energy supply, driving economic growth and development [1]. However, these activities are also major sources of environmental pollution, releasing a variety of toxic chemicals into the air and water through processes such as drilling, hydraulic fracturing (fracking), production, and transportation [2–4]. Oil and gas sites pose numerous environmental hazards that can have detrimental effects on human health [5, 6]. Living near active oil and gas wells or working at these sites may increase the risk of experiencing various adverse health effects, including reproductive and developmental disorders [5]. Pregnant women are particularly vulnerable to the effects of environmental exposures, as they can pass these toxins on to their developing fetus through the placenta [7].

Previous research has indicated that exposure to oil and gas extraction processes during pregnancy may lead to an increased risk of adverse birth outcomes [8, 9]. Even minimal exposure to these harmful chemicals can have negative impacts on the development and function of various systems in infants [6]. The developing fetus is especially susceptible to environmental toxins, and exposure to oil and gas-related air pollutants during pregnancy has been linked to various adverse birth outcomes, such as preterm birth (PTB), miscarriage, stillbirth, birth defects, small for gestational age (SGA), low birth weight (LBW), birth weight (BW), birth length (BL) and birth head circumference (BHC) [10, 11]. These pollutants can cross the placental barrier and build up in fetal tissues, contributing to adverse outcomes [7]. The mechanism behind this is that these pollutants can cause oxidative stress, DNA damage, and disrupt the endocrine system, affecting fetal growth [12]. The association between maternal exposure to oil and gas extraction processes and adverse birth outcomes is an important area of research that that has gained attention in recent years.

While epidemiological studies have investigated the link between maternal exposure to oil and gas activities and ABO, the evidence in this area reported inconsistent results [13–16]. Some studies have shown significant associations between exposure and adverse birth outcomes, while others have not found such relationships [17–21]. This inconsistency emphasizes the need for a comprehensive review of the available evidence to better understand the potential risks. It is important to comprehend the risks associated with maternal exposure to oil and gas extraction processes in order to develop effective strategies to protect pregnant women and their unborn children. By combining data from multiple studies, meta-analyses can provide more precise estimates of the association between maternal exposure to oil and gas activities and birth outcomes. This approach also allows for the examination of potential factors that may influence the relationship, enhancing our understanding of the underlying mechanisms. Therefore, this systematic review and meta-analysis, for the first time, was conducted to evaluate the association between maternal exposure to oil and gas process and birth outcomes.

Materials and methods

Design and registration

This systematic review and meta-analysis follows the Preferred Reporting Items for Systematic Review and Meta‐Analysis (PRISMA-P) 2020 guidelines [22] and is registered with the International Prospective Register of Systematic Reviews (CRD42024555447).

Search strategy

Search strategies included limiters, with some variation by database: English language, publication years up to 16 June 2024, global studies, and journal articles or review articles (the latter to facilitate reference hand searching). We performed systematic literature searches in five databases: MEDLINE/PubMed, Scopus (Elsevier), CINAHL (Cumulative Index to Nursing and Allied Health Literature), the Web of Science (WoS) (Clarivate Analytics), and the Cochrane Library (Wiley & Sons Inc.). The research librarian used EndNote reference management software for management and automated deduplication of records. Keywords and MeSH terms related to oil and gas extraction included terms such as"Oil Extraction,"OR"Gas Extraction,"OR"Natural Gas,"OR"Gas Extraction,"OR"Petroleum Industry,"OR"Fossil Fuels,"OR"Oil Drilling,"OR"Gas Drilling,"OR"Oil Refineries,"OR"Petroleum Production,"OR"Oil Spills,"OR"Extraction Techniques,"OR"Fracking,"OR"Offshore Drilling,"OR"Hydraulic Fracturing."Terms for birth outcomes, the terms included"Birth Outcomes,"OR"Preterm Birth,""PTB,"OR"Small for Gestational Age,"OR"SGA,"OR"Miscarriage,"OR"Birth Defects,""Stillbirth,"OR"Birth anthropometric measures,"OR"Low Birth Weight,"OR"Birth Weight,""Birth Length,"OR"Head Circumference,"OR"Chest Circumference."Exposure-related terms comprised"Residential Exposure,"OR"Occupational Exposure,"OR"Maternal Exposure,"OR"Prenatal Exposure,"OR"In Utero Exposure."The search strategy underwent a Peer Review of Electronic Search Strategies (PRESS) to ensure rigor and comprehensiveness.

In addition to the database searches, grey literature databases and manual searches of reference lists were conducted to minimize publication bias and include all relevant studies. In addition to published literature, we explored grey literature (i.e., unpublished studies) using the national grey literature collection (https://allcatsrgrey.org.uk). Preprint servers such as medRxiv (https://www.medrxiv.org) and PsyArXiv (https://psyarxiv.com) were also searched. Furthermore, we performed an incognito search in Google Scholar, reviewing the first 200 results for potential inclusion. The EThOS database (https://ethos.bl.uk) was consulted as well. Additionally, we manually reviewed the reference lists of potentially eligible papers to ensure that no relevant studies were overlooked. Studies published in English and involving human subjects were included to maintain relevance and applicability to the research question.

Eligible criteria

The eligibility criteria for studies included in the analysis were determined using the PECOS strategy, which stands for Participants, Exposure, Comparison, Outcome, and Study Design. In order to be included, studies had to meet the following criteria; Participants: Pregnant women at any stage of pregnancy. Exposure: Exposure to oil and gas extraction processes or associated environmental pollution, including residential or occupational exposure to petrochemical plants, oil refineries, unconventional gas development (UGD), hydraulic fracturing, or related emissions. Comparator: The comparator group included pregnant women with no exposure or minimal exposure to oil and gas extraction activities. This encompassed: (a) Populations residing in control areas without oil and gas operations, defined by distance thresholds (e.g., ≥ 10–30 km away), different wind patterns, or absence of industrial facilities. (b) Individuals not employed in oil and gas extraction, petrochemical industries, or associated occupational environments. Minimal exposure typically referred to baseline environmental background levels not attributable to industrial activity. Definitions of"non-exposed"and"low-exposed"were based on exposure categorizations used in the original studies and were documented during data extraction (Table 1). Outcomes: Adverse birth outcomes, including such as PTB, SGA, miscarriage, stillbirth, birth defects, LBW, and BW. Study Design: Observational studies (cohort, case–control, and cross-sectional designs). Both peer-reviewed studies and relevant grey literature meeting the inclusion criteria were considered.

Table 1.

Characteristics of included studies to evaluate impact of maternal exposure to oil and gas extraction processes on birth outcomes

First Authors, Year, (Ref)	Type of study	Country	Study population	Type of exposure	Outcomes	Results	Key findings
Axelsson et al., 1988, [23]	Retrospective cohort study	Sweden	For miscarriage: women who lived near petrochemical industries (n = 607) vs. women who lived in an area without petrochemical industries (n = 705) For birth defect: Women who lived near petrochemical industries (n = 1,255) vs. women who lived in an area without petrochemical industries (n = 1,527)	Residential exposure to petrochemical industries	Miscarriage, Birth defects and low birth weight	Miscarriage: A: OR: 1.15, 95%CI: 0.75–1.76 B: OR: 6.6, 95%CI: 2.3–19.2 Birth defect: observed/expected = 0.68 vs. congenital malformation; observed/expected = 0.79 Low birth weight: there were fewer infants with a birth weight lower than 2,500 g than anticipated	Miscarriage: (A) the miscarriage rate was slightly elevated in the exposed area (B) While a statistically significant increase in miscarriages was found for a small subset of women who worked for one of the petrochemical companies during pregnancy Birth defect: The numbers of malformed infants in the exposed area were lower than expected, but no statistical analyses were performed
Axelsson et al., 1989, [24]	Cross-sectional	Sweden	Women involved in laboratory work at a petrochemical plant during 1973 and 1987 (n = 55) vs. women who were not involved in laboratory work at the petrochemical plant (n = 47)	Occupational exposure to petrochemical plant	Miscarriage and birth weight	observed (n = 6) and expected number (n = 2) of miscarriages (p < 0.05)	A statistically significant difference between observed and expected number of miscarriages was seen during the period 1970–1974 They reported that the mean birth weight of the 40 live born infants whose mothers worked in laboratories at a petrochemical plant during pregnancy were slightly higher than the birth weights of infants born to mothers living in the same general area
Xu et al., 1998, [25]	Retrospective Cohort study	China	Women working in petrochemical plants and occupationally exposed to petrochemicals during the first trimester of pregnancy (n = 1620) vs. women who worked in petrochemical plants, but did not occupationally expose to petrochemicals during the first trimester of pregnancy (n = 1233)	Occupational exposure to petrochemical plants	Miscarriage	A: Adjusted for all confounders; OR: 2.7, 95%CI: 1.8–3.9 B: Self-reported exposure; OR: 2.9, 95%CI: 2–4 C: Excluding women with inconsistent exposure; OR: 2.9, 95%CI: 2–4 D: Exposure to benzene; OR: 2.5, 95%CI:1.7–3.7, E: Exposure to gasoline; OR:1.8, 95%CI:1.1–2.9 F: exposure to hydrogen sulphide; OR: 2.3, 95%CI: 1.2–4.4	An increased risk of miscarriage was found associated with the exposure to petrochemicals, including benzene, gasoline, and hydrogen sulphide
Bull et al., 1999, [26]	Cross-sectional	Norway	A: pregnancies-offshore operators (non-exposed, n = 86) vs. pregnancies-offshore mechanics (exposed, n = 59) B: pregnancies-offshore operators (non-exposed n = 86) vs. pregnancies-offshore drilling personnel (exposed, n = 172)	Occupational exposure to offshore oil processes	Miscarriage	A: OR: 1.1, 95% CI: 0.4–3.1 B: OR: 1.4, 95% CI: 0.6–3.2	No association was found; exposure to hydrocarbons in the occupations studied did not seem to have had a major influence on the incidence of miscarriage
Lin et al., 2001, [27]	Retrospective cohort study	Taiwan	49,673 unexposed women vs. 2,027 exposed women	Residential exposure to petrochemical municipalities	Preterm birth	OR: 1.41, 95%CI: 1.08–1.82	Significant association between exposure to petrochemical municipalities and increase the risk of preterm birth
Yang, et al., 2002, [28]	Retrospective cohort study	Taiwan	Women (n = 5,338) who had first-parity singleton deliveries and who lived in area with petrochemical complexes vs. A 10% random sample of all women (n = 51,789) in Taiwan who did not live in the study region	Residential exposure to petrochemical and petroleum industries	Preterm birth	OR: 1.18, 95%CI: 1.04–1.34	Significant association between exposure to petrochemical/petroleum industries and increase the risk of preterm birth
Yang et al., 2002, [29]	Retrospective cohort study	Taiwan	Women (n = 20,077) who had first-parity singleton deliveries living in petrochemical industrial municipality vs. matched reference (n = 19,673)	Residential exposure to petrochemical and petroleum industries	Preterm birth and Low birth weight (LBW)	Preterm birth: OR: 1.03, 95%CI: 0.94- 1.13 LBW: OR: 1.07, 95%CI: 0.95–1.22	No significant association was found between exposure to petrochemical/petroleum industries and preterm birth and LBW
Sebastián et al., 2002, [30]	Cross-sectional	Ecuador	Women living 9 communities within 5 km of an oil field, following a downstream direction (n = 791) vs. women living in 14 communities at least 30 km upstream from any oil field (n = 586)	Residential exposure to within 5 km of an oil field	Miscarriage and stillbirth	Miscarriage: OR: 2.47, 95%CI: 1.61–3.79 Stillbirth: OR: 0.85, 95%CI: 0.35–2.05	Significant association between exposure to oil filed and increase the risk of miscarriage. However, no association was found between stillbirth and exposure to oil filed
Oliveira et al., 2002, [31]	Case–control study	Brazil	Stillborn (> 500 g) selected from region close to petrochemical plant or wind condition (n = 230) vs. newborns weighing ≥ 2,500 g from the same hospital as case (n = 230) Newborns with birth defects selected region close to petrochemical plant or wind condition (n = 159) vs. newborns without birth defects as from the same hospital as control (n = 158) Region near plant or windy: 987 low birth weight newborns in Brazil vs. Reference region: 986 healthy newborns weighing ≥ 2,500 g	Residential exposure to petrochemical plant	Stillbirth Birth defect Low birth weight	Stillbirth: - Region close to petrochemical plant vs control: OR: 0.78, 95%CI:0.22–2.72 - Region with preferential wind direction vs control: OR:0.98: 95% CI:0.38–2.54 Birth defect: - Region close to petrochemical plant vs Reference region: OR = 0.30, 95%CI = 0.70–1.27 - Region with preferential wind direction vs Reference region: OR: 1.08, 95%CI: 0.30–3.88 Low birth weight: Region close to petrochemical plant vs Reference region: OR: 1.50, 95%CI:0.90–2.50 Region with preferential wind direction vs Reference region: OR: 1.42, 95%CI: 0.87–2.31	No significant association was found between exposure to petrochemical plant/wind condition and stillbirth, birth defect and low birth weight
Tsai, et al., 2003, [32]	Retrospective cohort study	Taiwan	Women (n = 14,545) who lived within a circle of 2 km radius around the industries vs. A 10% random sample of all women (n = 49,670) in Taiwan who did not live in the study region	Residential exposure to petrochemical petroleum, steel, and shipbuilding industries	Preterm birth	OR: 1.11, 95% CI: 1.02–1.21	Significant association between exposure to industries and increase the risk of preterm birth
Yang et al., 2004, [33]	Retrospective cohort study	Taiwan	Women (n = 7,095) who lived near three oil refinery plants vs. A 10% random sample of all women (n = 50,388) who did not live in the study region	Residential exposure to three oil refinery plants	Preterm birth	OR: 1.14, 95%CI: 1.01–1.28	Significant association between exposure to oil refinery plants and increase the risk of preterm birth
Chevrier et al., 2006, [34]	Cross-sectional	France	164 cleft lips with/without cleft palate (CL/P), 76 cleft palates (CP) vs. 236 controls includes Children hospitalized for treatment of some other disorders without any birth defects	A: Occupational exposure to petroleum B: Occupational exposure, to mixtures of organic solvents during pregnancy	Birth defects: the risk of non-syndromic oral clefts	A: OR: 3.6, 95%CI: 1.5–8.8 B: OR: 1.2, 95%CI: 0.3–4.9	The risk of oral clefts increased linearly with level of exposure within the three subgroups of oxygenated solvents
Desrosiers et al., 2012, [35]	Case–control study	USA	children with birth defects (n = 9998) who their parents worked in petroleum and gas vs children without birth defects who their parents did not work in petroleum and gas (n = 4066)	Occupational exposure to oil and gas operations	Birth defect	A: glaucoma/anterior chamber defects OR: 2, 95%CI: 0.8–5.1 B: Colonic atresia/stenosis OR:2.8, 95%CI: 0.9–9 C: intercalary limb deficiency OR: 2.6, 95%CI: 1.1–6.5 D: Atrial septal defect (ASD), secundum OR:1.6, 95%CI: 1.0–2.4	No significant association was found between exposure to oil/gas operations and glaucoma/anterior chamber defects or Colonic atresia/stenosis. However, significant association was observed between oil/gas operations and intercalary limb deficiency and ASD
McKenzie et al. 2014, [36]	Retrospective cohort study	USA	(A) Low exposure (1-tertile) (n = 18,884) vs. control group (n = 65,506) (B) Medium exposure (2-tertile) (n = 18,854) vs. control group (n = 65,506) (C) High exposure (3-tertile) (n = 19,384) vs. control group (n = 65,506)	Residential exposure in within 10-mile radius of natural gas development (NGD) wells	Preterm birth, low birth weight and birth defect	Preterm birth: (A) Low exposure vs. control group: OR:0.96, 95%CI: 0.89–1.0 (B) Medium exposure vs. control group: OR: 0.93, 95% CI: 0.87–1.0 (C) High exposure vs. control group: OR: 0.91, 95%CI: 0.85–0.98 Low birth weight: (A) Low exposure vs. control group: (OR: 1.0, 95% CI: 0.9–1.1) Mean difference: 5 (–2.2, 13) (B) Medium exposure vs. control group: (OR: 0.86, 95% CI: 0.77–0.95) Mean difference: 24 (17, 31) (C) High exposure vs. control group: (OR: 0.9, 95% CI: 0.8–1.0) Mean difference: 22 (15, 29) Birth defect; Congenital heart defects (CHDs) (A) Low exposure vs. control group: OR: 1.1, 95%CI: 0.93–1.3 (B) Medium exposure vs. control group: OR: 1.2, 95%CI: 1.0–1.3 (C) High exposure vs. control group: OR: 1.3, 955CI: 1.2–1.5 Birth defect; Neural tube defects (NTDs) (A) Low exposure vs. control group: OR: 0.65, 95%CI: 0.25–1.7 (B) Medium exposure vs. control group: OR: 0.80, 95%CI:0.34–1.9 (C) High exposure vs. control group: OR: 2.0, 95%CI: 1.0–3.9 Birth defect; Oral clefts (A) Low exposure vs. control group: OR: 0.65, 95%CI: 0.43–0.98 (B) Medium exposure vs. control group: OR: 0.89, 95%CI: 0.61–1.3 (C) High exposure vs. control group: OR: 0.82, 95%CI: 0.55–1.2	The results showed that the high exposure to the NGD was negatively associated with preterm birth. In addition, adjusted regression (OR) estimates for low birth weight suggest decreased risk of it with increasing exposure to NGD Prevalence of CHDs increased with exposure tertile, NTD prevalence was associated with the highest tertile of exposure compared with the absence of any gas wells within a 10-mile radius. No association was found between exposure and oral clefts
Stacy et al., 2015, [37]	Retrospective cohort study	USA	Women living in the area with an inverse distance weighted count of unconventional wells (A) second quartile: ≥ 0.87, but < 2.60 (n = 3,905), (B) third quartile: unconventional wells ≥ 2.60, but < 6.00 (n = 3,791), (C) fourth quartile: unconventional wells ≥ 6.00 (n = 4,151) Vs first quartile: Women living in the area with an inverse distance weighted count of unconventional wells > 0, but < 0.87 (n = 3,604)	Residential exposure unconventional gas drilling (UGD)	Preterm birth, small for gestational age (SGA) and birth weight	Preterm birth: (A) Second quartile vs. First quartile; OR:0.82, 95%CI: 0.68–0.98 (B) Third quartile vs. First quartile; OR≈1.1, 95%:CI crossing 1 (Data presented in figure, no actual number) (C) Fourth quartile vs. First quartile; OR≈1.0, 95%CI crossing 1) (Data presented in figure, no actual number) Small for gestational age A) Second quartile vs. First quartile; OR≈1.1, 95%CI crossing 1) (Data presented in figure, no actual number) (B) Third quartile vs. First quartile; OR≈1.2, 95% CI crossing 1 (Data presented in figure, no actual number) (C) Fourth quartile vs. First quartile; OR:1.54, 95%CI: 1.10–2.63) Birth weight: (A) Second quartile vs. First quartile; 3370.4 ± 540.5 vs 3343.9 ± 543.9 (Difference in mean: 26.5 g) p > 0.05 (B) Third quartile vs. First quartile; 3345.4 ± 553.5 vs 3343.9 ± 543.9 (Difference in mean: 1.5 g) p > 0.05 (C) Fourth quartile vs. First quartile; 3323.1 ± 558.2 vs 3343.9 ± 543.9 *(difference in mean: −20.8 g) p = 0.02	There was no significant association of proximity and density of UGD with preterm birth. However, revealed a higher incidence rate of SGA with higher exposure to UGD (4-quartile vs. 1-quartile)
Ha et al., 2015, [38]	Cross-sectional	USA	423,719 Singleton Births	Residential exposure to active power plants	Preterm birth and low birth weight	Preterm birth: OR: 1.018, 95%CI:1.013–1.023 Low birth weight: OR: 1.011, 95%CI:1.002–1.02	Results showed 1.8% increased odds for PTD and 1.1% for term LBW for each 5 km closer to any power plant
Casey et al., 2016, [39]	Retrospective cohort study	USA	(A) 2,648 Second quartile vs. 2,590 First quartile (B) 2,642 Third quartile vs. 2,590 First quartile (C) 2,616 Fourth quartile vs 2,590 First quartile	Residential exposure to the unconventional natural gas development (UNGD) activity	Preterm birth and birth weight (BW)	Preterm birth: A) Second quartile vs. First quartile; Adjusted for all covariate OR: 1.2, 95%CI: 0.9–1.6 Adjusted for year of birth OR: 1.3, 95%CI: 1–1.8 B) Third quartile vs. First quartile Adjusted for all covariate: OR: 1.3, 95%CI: 1–1.8 Adjusted for year of birth OR: 1.6, 95%CI:1.1–2.4 C) Fourth quartile vs. First quartile Adjusted for all covariate: OR: 1.4, 95%CI: 1–1.9 Adjusted for year of birth: OR: 1.9, 95%CI: 1.2–2.9 Birth weight: A) Mean difference: Second quartile vs. First quartile; Adjusted for all covariate: − 21 g (95% CI: − 46 to 5) Adjusted for year of birth: −16 g (95% CI: −44 to 11) B) Mean difference: Third quartile vs. First quartile Adjusted for all covariate: − 9 g (95% CI: − 35 to 16) Adjusted for year of birth: 1 g (95% CI: −34 to 36) C) Mean difference: Fourth quartile vs. First quartile Adjusted for all covariate: − 31 g (95% CI: − 57 to −5) Adjusted for year of birth: −20 g (95% CI: −56 to 16)	Prenatal residential exposure to unconventional natural gas development activity was associated with higher incidence of preterm birth. In addition, there was no associations of UNGD with birth weight (after adjustment for year)
Walker Whitworth et al., 2018, [40]	Case–control study	USA	Preterm cases (n = 13,549) and controls (n = 67,745)	Exposure to drilling and production activity related to unconventional gas Development (UGD)	Preterm birth	UGD Drilling activity 1-tertile vs. control OR: 1.03, 95%CI: 0.90–1.18 2-tertile vs. control OR: 1.03, 95%CI: 0.90–1.18 3-tertile vs. control OR: 1.20,95%CI: 1.06–1.37 UGD production activity 1-tertile vs. control OR: 1.07, 955CI: 0.97–1.17 2-tertile vs. control OR: 1.13, 95%CI: 1.02–1.24 3-tertile vs. control OR: 1.15, 95%CI: 1.05–1.26	Results showed that an increased odds of PTB in 3-tertile of the UGD drilling and 2-tertile and 3-tertile of the UGD production activity
Cushing et al., 2020, [41]	Retrospective Cohort study	USA	Women with singleton birth within rural areas of the 27 counties (n = 23,487)	Exposure to Flaring from Unconventional Oil and Gas Development (OGD) within 5 km	Preterm birth, SGA, birth weight	Preterm birth Low Flaring from OGD OR: 0.76, 95%CI: 0.6–0.97 High Flaring from OGD OR: 1.41, 95%CI: 1.11–1.69 SGA Low Flaring from OGD OR: 0.89, 95%CI: 0.71–1.14 High Flaring from OGD OR: 1.02, 95%CI: 0.82–1.29 Birth weight Low Flaring from OGD MD: −5.9, 95%CI: −36, 24.3 High Flaring from OGD MD: −9.2, 95%CI: −40.9, 22.5	Significant association was observed between exposure to high Flaring from Unconventional Oil and Gas Development (OGD) within 5 km and increased odds of preterm. While no significant association between low and high exposure to flaring OGD and SGA and birth weight
Gonzalez et al., 2020, [42]	Case–control	USA	Preterm birth cases (n = 27,913) vs. term birth controls (n = 197,461)	Exposure to oil and gas production well sites	Preterm birth	PTB in 20–27 weeks First trimester 1-quantile vs. unexposed OR: 1.01 (0.88, 1.15) 2-quantile vs. unexposed OR: 1.00 (0.88, 1.13) 3-quantile vs. unexposed OR: 1.09 (0.95, 1.24) Second trimester 1-quantile vs. unexposed OR: 0.98 (0.85, 1.12) 2-quantile vs. unexposed OR: 0.99 (0.87, 1.13) 3-quantile vs. unexposed OR: 1.08 (0.95, 1.23) PTB in 28–31 weeks First trimester 1-quantile vs. unexposed OR: 0.93 (0.83, 1.05) 2-quantile vs. unexposed OR: 0.95 (0.84, 1.07) 3-quantile vs. unexposed OR: 1.14 (1.01, 1.27) ** Second trimester 1-quantile vs. unexposed OR: 0.94 (0.83, 0.96) 2-quantile vs. unexposed OR: 0.96 (0.85, 1.07) 3-quantile vs. unexposed OR: 1.14 (1.01, 1.27) ** PTB in 32–36 weeks First trimester 1-quantile vs. unexposed OR: 0.93 (0.89, 0.97) 2-quantile vs. unexposed OR: 0.96 (0.92, 1.01) 3-quantile vs. unexposed OR: 0.99 (0.95, 1.04) Second trimester 1-quantile vs. unexposed OR: 0.93 (0.88, 0.97) 2-quantile vs. unexposed OR: 0.96 (0.92, 1.01) 3-quantile vs. unexposed OR: 0.99 (0.95, 1.04) Third trimester 1-quantile vs. unexposed OR: 0.93 (0.89, 0.97) 2-quantile vs. unexposed OR: 0.96 (0.92, 1.01) 3-quantile vs. unexposed OR: 1.00 (0.95, 1.04)	Exposure to oil and gas well sites was associated with increased risk of preterm birth in the first and second trimester in 28–31 weeks
Harville et al., 2021, [43]	Cohort study	USA	Women exposed n = 960 women/1255 pregnancies	Oil spill exposure	Preterm and low birth weight	Preterm birth OR: 2.27, 95%CI: 1.34–3.87 Low birth weight OR: 2.19, 95% CI, 1.29–3.71	Associations were seen with high levels of contact with oil for LBW and PTB
Oaka et al., 2021, [44]	Cohort study	Japan	Total number of pregnancies (n = 89,196) and total number of miscarriage cases (n = 356) Total number of pregnancies (n = 94,548) and total number of stillbirth cases (n = 161)	Occupational or daily maternal exposure to Kerosene, petroleum, benzene, or gasoline	Miscarriage and stillbirth	Miscarriage A: 1–3 times a month exposed vs. never exposed; OR: 1.07,95%CI: 0.72–1.61 B: Once a week and over exposed vs. never exposed; OR: 1.37,95%CI: 0.70–2.67 Stillbirth: A: 1–3 times a month exposed vs. never exposed; OR: 1.07, 95%CI: 0.64–1.78 B: Once a week and over exposed vs. never exposed; OR: 0.95, 95%CI: 0.35–2.60	Any type of exposure was not significantly associated with miscarriage and stillbirth
Tran et al., 2021, [45]	Retrospective cohort study	USA	Exposed (n = 1,192) vs. unexposed (n = 1,004,563)	Prenatal exposure to hydraulic fracturing (HF) in rural and urban mothers	Low birth weight, preterm birth, SGA and birth weight	Low birth weight Rural: OR: 1.74, 1.10–2.75 Urban: OR: 0.83, 0.63–1.07 Preterm birth Rural: OR: 1.17, 0.64–2.12 Urban: OR: 0.65, 0.48–0.87 SGA Rural: OR: 1.68, 1.42–2.27 Urban: OR: 1.23, 0.98–1.55 Birth weight Rural: Md: −73, −131, −15 Urban: Md: −2, −35, 31	Among rural mothers, HF exposure was associated with increased odds of LBW, SGA and lower BW and not associated with Preterm birth Among urban mothers, HF exposure was inversely associated with PTB and not associated with LBW, SGA, and BW
Cairncross et al., 2022, [46]	Retrospective cohort study	Canada	26 193 individuals with 34 873 unique pregnancies	Residential proximity to hydraulic fracturing sites (underground petroleum extraction process)	Preterm birth and small for gestational age (SGA)	Preterm birth: RR: 1.01, 95%CI: 0.88–1.16 SGA: RR: 1.12, 955CI: 1.03–1.23	SGA was significantly higher for individuals who lived within 10 km of at least 1 hydraulically fractured well. However, Risk of indicated preterm birth was not related to residential exposure to hydraulic fracturing sites

We excluded studies without full-text access, animal studies, clinical trials, letters-to-the-editor, conference papers, and posters if the main article was inaccessible or lacked methodological information. Additionally, studies published in languages other than English and those not involving human subjects were excluded. Studies not focusing on maternal exposure or the impact on birth outcomes were also excluded.

Study selection

To enhance the efficiency of managing search results and reviewing relevant studies, all gathered studies were transferred into EndNote software (version 20.2.1). This process involved eliminating duplicate entries and ensuring accurate citation details. Initially, two independent researchers (AVA & SA) assessed the titles and abstracts to identify studies relevant to the research questions. Subsequently, the same researchers thoroughly examined the full-text articles to determine their eligibility based on inclusion criteria. Studies meeting the criteria were included, while those not meeting them were excluded. To maintain consistency, the agreement between the researchers was evaluated, and any disagreements were resolved through discussion or consultation with a third researcher (FA). The evaluators demonstrated a strong level of agreement (Kendall's coefficient of agreement = 0.92 (P < 0.001)). The final search results were illustrated in a flow diagram following the PRISMA 2020 guidelines for transparency and clarity.

Risk of bias (quality) assessment

The methodological quality of the studies included in this study was assessed by two independent researchers (AVA & SA) using the Newcastle–Ottawa Scale (NOS), a tool designed for evaluating non-randomized studies like observational research [47]. The NOS evaluates studies based on three domains: selection of study groups, comparability of groups, and ascertainment of outcomes or exposure. For case–control and cohort studies, the NOS can award a maximum of nine points, while cross-sectional studies can receive up to eight points. This scoring system includes up to four points for selection, two points for comparability, and three points for outcomes [48]. The NOS results were then categorized as'good','fair', or'poor'quality according to the Agency for Healthcare Research and Quality (AHRQ) standards [49].

Additionally, the reliability of evidence from the included studies was assessed using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) framework [50]. This framework evaluates evidence quality based on factors such as risk of bias, imprecision, inconsistency, and indirectness. Initially, the studies were classified as high-quality evidence and then adjusted according to these criteria. The evidence was ultimately classified into four levels: high quality, moderate quality, low quality, and very low quality, indicating the level of certainty in the estimated impact.

Data collection

Data collection involved two review authors (AVA, SA) independently extracting study characteristics from included studies onto a data-collection template. An independent review author (FA) double-checked for accuracy. The extracted key information included study characteristics such as author(s) name, year of publication, country, study design, sample size, participant characteristics, and details of exposure to oil and gas extraction processes (type, time and duration of exposure).

In this study, we used the term “oil and gas extraction process” to encompass a broad range of activities related to the exploration, extraction, and processing of crude oil and natural gas. This includes conventional methods (e.g., drilling, offshore rigs, oil spills, oil refineries) and unconventional methods (e.g., hydraulic fracturing or “fracking”, unconventional gas development [UGD], and flaring operations). “Petrochemical industries” are defined as industrial plants involved in the downstream processing and refinement of petroleum and natural gas into chemicals such as benzene and kerosene. “Natural Gas Development (NGD)” refers to both conventional and unconventional extraction of natural gas, often associated with horizontal drilling and fracking. In addition, based on the data extracted from each study, exposures were classified into the following groups; (a) residential exposure: living within proximity (e.g., < 10 km) of oil and gas infrastructure such as wells, flaring sites, refineries, petrochemical complexes, or power plants, (b) occupational exposure: employment in or near oil and gas or petrochemical industries during pregnancy, (c) type of industrial activity that included; conventional oil and gas extraction (e.g., drilling, offshore rigs, oil refineries), unconventional methods such as hydraulic fracturing (fracking), UGD, and flaring, (d) petrochemical and petroleum industries (e.g., exposure to benzene, gasoline, solvents).

Moreover, all maternal and neonatal outcomes were defined using standardized criteria to ensure consistency across all included studies and were meticulously recorded. PTB was defined as birth occurring before 37 completed weeks of gestation, in accordance with WHO guidelines [51]. Miscarriage was defined as a spontaneous loss or abortion of a pregnancy before the 20th week of gestation [52]. Stillbirth was defined as fetal death occurring at or after 20 weeks of gestation [53]. Birth defects were defined as any physical, behavioral, or developmental abnormalities present at birth [54]. SGA was defined as a birth weight below the 10th percentile for a given gestational age, based on population- or ultrasound-derived growth standards [55]. LBW was defined as a birth weight of less than 2500 g, regardless of gestational age.

Results were documented in terms of effect estimates (e.g., odds ratios, β coefficients or mean difference) with corresponding confidence intervals (CIs), p-values, and adjustments for confounding factors. Inter-rater reliability was assessed throughout the data extraction process to maintain consistency between reviewers.

Synthesis of results and statistical analysis

We conducted a meta-analysis when at least two studies reported similar exposures and outcomes. For studies that did not meet this criterion, we systematically compiled, reviewed, and presented the data. Rigorous methods were used to estimate the overall effect size (odds ratios: ORs) for studies included in the meta-analysis. This involved aggregating individual study results using a random-effects model, calculating summary effect estimates, and determining confidence intervals. Heterogeneity among studies was assessed using Cochran’s Q test and quantified with the I-squared (I²) statistic. Leave-one-out sensitivity analyses were performed to ensure the robustness of the results. Publication bias was assessed using Begg's test, Egger's test, Funnel plots, and the trim-and-fill method. Meta-regression analysis was conducted to identify sources of heterogeneity, using the year of publication as the independent variable and the effect size (ES) as the dependent variable. The meta-analysis results were presented clearly and concisely, including forest plots, summary effect estimates, and confidence intervals. Significance was determined at a P-value < 0.05. All statistical analyses were conducted using STATA version 17 to ensure the reliability and robustness of the data interpretation.

Results

Search outcomes

Our initial search across Scopus, PubMed/MEDLINE, CINAHL, WoS, and the Cochrane Library databases identified a total of 5,220 articles. An additional 15 articles were located through bibliographies, bringing the total number of retrieved articles to 5,235. After eliminating duplicates and screening titles and abstracts, we assessed 41 full-text articles for eligibility. Among these, 17 articles were excluded as they did not meet the inclusion criteria, such as failing to report relevant birth outcomes (n = 6) or not considering the exposure of interest (n = 11). Ultimately, 24 papers were included in the current study (Fig. 1).

Fig. 1.

The literature searches results and the screening process based on PRISMA 2020 flowchart

Studies characteristics

The 24 studies included in this review varied in their geographic and methodological characteristics and exposure settings. Based on exposure classification, 18 studies (75%) focused on residential exposure [13–15, 17–21, 56–65] and 6 (25%) on occupational exposure [16, 66–70]. Regarding industrial activity types: petrochemical industries were the primary exposure source in 9 studies [19, 21, 56, 57, 60, 62, 64, 66–68], oil refineries in 1 study [59], conventional oil and gas extraction, including offshore drilling, in 2 studies [63, 70], unconventional operations, including fracking and UGD, in 5 studies [13–16, 61], and Flaring was explicitly examined in 1 study [20]. These categorizations are summarized in Table 1. Stratified synthesis was conducted wherever study characteristics and data allowed.

In addition, these studies were conducted in various geographic locations, including the USA (n = 10) [14, 15, 17–21, 61, 65, 70], Taiwan (n = 5) [56–60], Sweden (n = 2) [62, 66], China [67], Norway [69], Ecuador [63], Brazil [64], France [68], Japan [16], and Canada [13]. The study designs varied, with 15 studies being cohort studies [13–18, 20, 56–62, 67], five being case–control studies [19, 64, 65, 68, 70], and four being cross-sectional studies [21, 63, 66, 69]. The outcomes studied included preterm birth PTB (15 studies) [13–15, 17–21, 56–61, 65], miscarriage (6 studies) [16, 63, 66, 67, 69], stillbirth (3 studies) [16, 63, 64], birth defects (5 studies) [18, 62, 64, 68, 70], LBW (7 studies) [14, 15, 18, 21, 57, 62, 64], BT (5 studies) [15, 17, 20, 61, 66] and SGA (4 studies) [13, 15, 20, 61] (Supplementary File 1).

Risk of bias assessment

The risk of bias was assessed using the NOS, and based on the AHRQ standards, study quality was classified as good (≥ 6 points), fair (4–5 points), or poor (≤ 3 points). The risk of bias ratings for the included studies indicated that 16 (66.7%) were of good quality, 6 (25%) were of fair quality, and 2 (8.3%) were of poor quality (Supplementary File 2, Table S2-S4). Among the case–control studies (n = 5), three were rated as good quality [19, 65, 70] and two as fair quality [64, 68]. In the cohort studies (n = 15), eleven were rated as good quality [13–18, 20, 57, 59, 60, 67], three as fair quality [56, 58, 61], and one study by Axelsson et al. [62], was rated as poor quality. For the cross-sectional studies (n = 4), two were rated as good quality [21, 63], one as fair quality [69], and one as poor quality [66]. The GRADE assessment of the 32 included studies indicated moderate-certainty evidence of critical importance regarding the association between maternal exposure to oil and gas extraction processes and adverse birth outcomes (Supplementary File 2, Table S5). While the findings were largely consistent across diverse settings and study designs, concerns were identified in two key domains: risk of bias and indirectness. These stemmed from residual confounding, limitations in exposure assessment, and methodological variability, including self-reported data and incomplete adjustment for covariates. Although subgroup analyses and stratified evaluations were conducted to mitigate these concerns, the potential for uncontrolled confounding and exposure misclassification warranted a cautious interpretation of the strength of the evidence. Nevertheless, the presence of large and consistent effect sizes across studies, some evidence of dose–response relationships, and a lack of plausible alternative explanations suggest that any remaining biases would likely attenuate, rather than nullify, the observed associations. Taken together, these findings support a moderate level of certainty in the observed association, with a strong public health relevance.

Maternal exposure and adverse birth outcomes

Preterm birth (PTB)

The association between maternal exposure to the oil and gas extraction process and PTB was examined in 15 studies [13–15, 17–21, 56–61, 65]. With the exception of Gonzalez et al. [19], all studies reported PTB as any birth occurring before 37 weeks without considering of gestational age subcategories. However, this case–control study included data from 27,913 PTB cases and 197,461 term birth controls and investigated the relationship between different quantiles of exposure to oil and gas production well sites and PTB at 20–27, 28–31, and 32–36 weeks of gestational ages [19]. The results showed a significant increase in the odds of PTB associated with exposure to wells in the third quantile during the first and second trimesters for births at 28–31 weeks of gestation, with adjusted OR: 1.14, 95% CI: 1.01–1.27 and OR: 1.14, 95% CI: 1.01–1.27, respectively.

The meta-analysis included 14 studies that examined the relationship between maternal exposure to oil and gas extraction process and PTB [13–15, 17, 18, 20, 21, 56–61, 65]. The findings showed a significant association between exposure to these agents and an increased risk of PTB (OR: 1.07, 95% CI: 1.01–1.13, P = 0.03), with considerable heterogeneity among the studies (I²: 88.86%, P < 0.001) (Fig. 2).

Fig. 2.

Association between maternal exposure to oil and gas extraction process and preterm birth (PTB)

Analysis of publication bias through funnel plot inspection revealed some asymmetry, and statistical tests yielded mixed results. While Begg’s test did not indicate significant bias (P = 0.392), Egger’s test approached significance (P = 0.072), suggesting possible asymmetry. Application of the trim-and-fill method, which added six imputed studies, indicated that publication bias may influence the pooled estimate (adjusted OR: 1.05, 95% CI: 0.99–1.11) (Supplementary File 3, Figure S1). Although the adjusted estimate no longer reached statistical significance, the confidence interval remains compatible with a small increase in risk. These findings suggest that publication bias may have influenced the original estimate, and caution should be exercised in interpreting the strength of the association.

Sensitivity analyses supported the robustness of the findings, showing consistent results when individual studies were excluded. The Galbraith plot indicated moderate heterogeneity among most studies (Supplementary File 3, Figure S2). Subgroup analyses by publication year (before 2015 and after 2015) revealed no statistically significant association for studies published before 2015 (OR: 1.03, 95% CI: 0.96–1.10), although the upper bound of the confidence interval suggests a potential modest increase in risk. A significant association was found for studies published after 2015 (OR: 1.11, 95% CI: 1.00–1.22). Heterogeneity remained high in both groups (I²: 85.9% for studies before 2015, I²: 79.9% for studies after 2015) (Supplementary File 3, Figure S3). Univariate meta-regression indicated that publication year did not significantly contribute to the observed heterogeneity (β: −0.004, 95% CI: −0.013, 0.005), though the wide confidence interval suggests uncertainty in ruling out a minor role.

Miscarriage

The association between maternal exposure to oil and gas extraction processes and miscarriage was examined in six studies [16, 62, 63, 66–69]. Five of these studies [16, 62, 63, 67, 69] were included in a meta-analysis, which revealed a statistically significant association between exposure and increased risk of miscarriage (OR: 2.03, 95% CI: 1.60–2.58, P < 0.001), although with substantial heterogeneity (I² = 66.8%, P = 0.01) (Fig. 3A).

Fig. 3.

Association between maternal exposure to oil and gas extraction process and (A) miscarriage and (B) stillbirth

Funnel plot analysis suggested some asymmetry, potentially indicating small-study effects such as publication bias or methodological differences in smaller studies; however, formal statistical tests (Begg’s test P = 0.220; Egger’s test P = 0.845) did not support significant asymmetry. Application of the trim-and-fill method did not materially alter the pooled estimate, suggesting that any potential missing studies would not substantially affect the overall conclusion (Supplementary File 3, Figure S4). Sensitivity analyses further supported the robustness of the findings. The Galbraith plot confirmed moderate heterogeneity among the studies (Supplementary File 3, Figure S5). Subgroup analyses by publication year showed a stronger association in studies published before 2000 (OR: 2.22, 95% CI: 1.73–2.86) compared to those published after 2000 (OR: 1.55, 95% CI: 0.91–2.63), though neither subgroup showed homogeneity (I²: 59.4% and 71.9%, respectively) (Supplementary File 3, Figure S6). Meta-regression analysis indicated that publication year did not significantly explain the observed heterogeneity (β: −0.020, 95% CI: −0.044 to 0.004).

Taken together, while there is some visual suggestion of asymmetry in the funnel plots, the lack of statistical significance in Egger’s and Begg’s tests, along with the stability of the pooled effect in sensitivity and trim-and-fill analyses, suggests that the observed association is unlikely to be driven primarily by publication bias. Nevertheless, we acknowledge the limitations posed by potential small-study effects and have interpreted the findings cautiously in the context of possible biases.

Stillbirth

The association between maternal exposure to oil and gas extraction processes and stillbirth was evaluated in three studies [16, 63, 64]. All three reported null findings in terms of statistical significance; however, confidence intervals in each study were wide and compatible with both potential increases and decreases in risk. The pooled estimate for the effect of exposure to pollutants from oil and gas extraction on stillbirth yielded (OR: 0.98, 95% CI: 0.68–1.40, P = 0.92), suggesting that the data are consistent with little or no association, but do not rule out the possibility of a clinically meaningful effect due to the wide confidence interval. There was no heterogeneity among the studies (Fig. 3B).

Analysis of publication bias using funnel plot analysis revealed some bias, but both Begg's (P = 0.220) and Egger's (P = 0.611) tests did not indicate significant publication bias. Even after adjusting for three imputed studies using the trim-and-fill method, there were no significant changes in the results (Supplementary File 3, Figure S7). Sensitivity analysis showed that the findings remained consistent even when individual studies were excluded. The Galbraith plot demonstrated that all studies were within the expected range (Supplementary File 3, Figure S8).

Birth defects

The association between maternal exposure to oil and gas extraction activities and birth defects was examined in five studies [18, 62, 64, 68, 70]. Among them, only three provided disaggregated data by defect type (e.g., oral clefts, neural tube defects, congenital heart defects) [18, 68, 70], while the others reported aggregated “birth defect” outcomes without subtype specification [62, 64]. Due to the limited number of studies available per individual defect type, a stratified meta-analysis was not feasible. Therefore, a pooled analysis was conducted using overall birth defect outcomes, while individual associations by defect type are presented descriptively.

Axelsson et al. [62] analyzed birth registry records and found that the number of malformed infants in the exposed area was lower than expected (observed/expected = 0.68 based on the congenital malformation registry; observed/expected = 0.79 based on the medical birth registry), although no statistical analyses were conducted. Oliveira et al. [64] did not find any statistically significant differences when comparing regions near a petrochemical plant to a reference region for birth defects (OR: 0.30, 95% CI: 0.70–1.27), or when comparing regions with preferential wind direction to the reference region (OR: 1.08, 95% CI: 0.30–3.88).

Chevrier et al. [68] reported a significant association between occupational exposure to oil and gas pollutants and the risk of oral clefts, specifically cleft lip with (OR: 3.6, 95% CI: 1.5–8.8), but not cleft lip without cleft palate (OR: 1.2, 95% CI: 0.3–4.9). Desrosiers et al. [70] did not find statistically significant correlations between occupational exposure and birth defects such as glaucoma/anterior chamber defects (OR: 2, 95% CI: 0.8–5.1) and colonic atresia/stenosis (OR: 2.8, 95% CI: 0.9–9.00). However, they found a positive correlation between exposure and intercalary limb deficiency (OR: 2.6, 95% CI: 1.1–6.5) as well as ASD (OR: 1.6, 95% CI: 1.0–2.4). Mckenzie et al. [18], discovered that high (OR: 2.13, 95% CI: 1.2–1.5) and medium (OR: 1.2, 95% CI: 1.1–1.3) exposure to natural gas development (NGD) were linked to CHDs. They also found a significant increase in the odds of NTDs with high exposure to NGD (OR: 2, 95% CI: 1–3.9). However, they did not find a significant association between low, medium, and high exposure to NGD and oral cleft [18].

The pooled effect of exposure to pollutants from oil and gas extraction process on the occurrence of birth defects, when not specifying the type of birth defects, did not show a statistically significant association (OR: 1.14, 95% CI: 0.84–1.53, P = 0.41), with high levels of heterogeneity among the studies (I²: 93.5%, P < 0.001) (Fig. 4A).

Fig. 4.

Association between maternal exposure to oil and gas extraction process and (A) birth defects and (B) small for gestational age (SGA)

Analysis of publication bias using funnel plot analysis, Begg's test (P = 0.836), Egger's test (P = 0.152), and the trim-and-fill method did not suggest significant publication bias (Supplementary File 3, Figure S9). Sensitivity analysis indicated consistent findings even when individual studies were excluded. However, the Galbraith plot revealed that studies deviated from the expected range, showing high heterogeneity among the studies (Supplementary File 3, Figure S10). Subgroup analyses based on publication year (before 2013 and after 2013) revealed no significant association between maternal exposure to oil and gas extraction activities and the occurrence of birth defects in both groups of studies: before 2013 (OR: 1.47, 95% CI: 0.78–2.79) and after 2013 (OR: 1.14, 95% CI: 0.84–1.53). Significant heterogeneity was observed in both groups of studies (I²: 83.7% for studies before 2013, I²: 77.8% for studies after 2013) (Supplementary File 3, Figure S11). Univariate meta-regression analysis indicated that publication year did not significantly contribute to the observed heterogeneity (β: 0.034, 95% CI: −0.039 to 0.107).

Maternal exposure and birth anthropometric outcomes

Small for gestational age (SGA)

Four studies [13, 15, 20, 61] were examined to assess the association between maternal exposure to oil and gas extraction process and SGA infants. Stacy et al. [61] discovered a notable association between higher UGD exposure (in the fourth quartile) and an increased risk of SGA (OR: 1.54, 95% CI: 1.1–2.63) compared to lower exposure (first quartile). However, Cushing et al. [20], did not find a significant relationship between low (OR: 0.89, 95% CI: 0.71–1.14) and high (OR: 1.02, 95% CI: 0.82–1.29) levels of flaring from unconventional oil and gas development (OGD) within 5 km and SGA. Tran et al. [15] demonstrated an increased risk of SGA with prenatal exposure to hydraulic fracturing (HF) in rural mothers (OR: 1.68, 95% CI: 1.42–2.27) but not in urban mothers (OR: 1.23, 95% CI: 0.98–1.55). Additionally, Cairncross et al. [13], found a significantly higher relative risk of SGA associated with residential exposure to HF sites (RR: 1.12, 95% CI: 1.03–1.23).

A meta-analysis of four studies [13, 15, 20, 61] revealed a significant association between maternal exposure to oil and gas extraction process and SGA (OR: 1.23, 95% CI: 1.05–1.45, P = 0.01), with substantial heterogeneity among the studies (I²: 71.7%, P < 0.001) (Fig. 4B).

Evaluation of publication bias using funnel plot analysis and statistical tests indicated some bias in the meta-analysis. Nonetheless, the Begg's (P = 0.173) and Egger's (P = 0.184) tests did not show significant publication bias. Upon applying the trim-and-fill method and adding three imputed studies, the results changed significantly (Supplementary File 3, Figure S12). Sensitivity analysis supported the strength of the findings, with consistent results upon exclusion of individual studies. However, the Galbraith plot revealed that the studies exhibited substantial deviation from the expected range, suggesting high levels of heterogeneity (Supplementary File 3, Figure S13). Subgroup analyses based on publication year (before 2015 and after 2015) indicated a noteworthy association between maternal exposure to oil and gas extraction activities and SGA in studies published before 2015 (OR: 1.51, 95% CI: 1.16–1.97), with no heterogeneity noted among the studies. Conversely, no significant association was observed in studies published after 2015 (OR: 1.16, 95% CI: 0.95–1.41), with significant heterogeneity among the studies (I²: 81.4%, P < 0.001) (Supplementary File 3, Figure S14). Additionally, univariate meta-regression analysis indicated that publication year did not significantly contribute to the observed heterogeneity (β: −0.035, 95% CI: −0.099 to 0.029).

Low birth weight (LBW)

Seven studies examined the association between maternal exposure to oil and gas extraction process and LBW [14, 15, 18, 21, 57, 62, 64]. All studies, except for Axelsson et al. [62], reported the odds of LBW in relation to maternal exposure to oil and gas extraction activities. Axelsson et al. [62] only showed a lower than expected number of infants with LBW near a petrochemical plant. Oliveira et al. [64] did not find statistically significant differences when comparing regions near a petrochemical plant to a reference region for LBW (OR: 1.5, 95% CI: 0.9–2.5) or when comparing regions with preferential wind direction to the reference region (OR: 1.42, 95% CI: 0.87–2.31). Similarly, Yang et al. [57] found no significant differences among infants born to mothers near petrochemical industries compared to non-exposed mothers (OR: 1.07, 95% CI: 0.95–1.22). However, McKenzie et al. [18], revealed a negative association between mild exposure to NGD wells and LBW (OR: 0.86, 95% CI: 0.77–0.95). Ha et al. [21] a positive association between LBW and residential exposure to active oil and gas plants. Tran et al. [15] showed an increased risk of LBW with prenatal exposure to hydraulic fracturing (HF) in rural mothers (OR: 1.47, 95% CI: 1.1–2.75) but not in urban mothers (OR: 0.83, 95% CI: 0.63–1.07). Additionally, Harville et al. [14] identified a significantly higher risk of LBW associated with oil spill exposure (OR: 2.19, 95% CI: 1.29–3.71).

The meta-analysis included six studies [14, 15, 18, 21, 61, 62] on oil and gas extraction process and LBW. Overall, no significant association was found (OR: 1.07, 95% CI: 0.93–1.24, P = 0.34), with high heterogeneity among studies (I²: 92.2%, P < 0.001) (Fig. 5A). Publication bias analysis indicated some bias through funnel plot and Egger’s regression test (P = 0.002), but adjustments using the trim-and-fill method did not significantly change the results even after imputing three additional studies to adjust for missing data (Supplementary File 3, Figure S15). Sensitivity analysis and Galbraith plot confirmed the robustness of the findings and moderate heterogeneity, thereby enhancing the reliability of the results (Supplementary File 3, Figure S16). Subgroup analyses based on publication year (before 2014 and after 2014) revealed no significant association between maternal exposure to oil and gas extraction activities and the odds of LBW in both groups of studies: before 2014 (OR: 0.98, 95% CI: 0.89–1.09) and after 2014 (OR: 1.27, 95% CI: 0.83–1.95). Significant heterogeneity was observed in both groups of studies (I²: 64.5% for studies before 2014, I²: 90.1% for studies after 2014) (Supplementary File 3, Figure S17). Univariate meta-regression analysis indicated that publication year did not significantly contribute to the observed heterogeneity (β: −0.0023, 95% CI: −0.0267 to 0.0221).

Fig. 5.

Association between maternal exposure to oil and gas extraction process and (A) low birth weight (LBW) and (B) birth weight (BW)

Birth weight (BW)

Five studies [15, 17, 20, 61, 66] were reviewed to investigate the relationship between maternal exposure to air pollutants from oil and gas extraction activities and BW. Axelsson et al. [66], reported that infants born to mothers working in a laboratory at a petrochemical plant had slightly higher birth weights compared to those born to mothers in the general population, but this difference was not statistically significant. Stacy et al. [61], found that women living in areas with a higher density (fourth quartile) of UGD had significantly lower mean BW compared to those in areas with lower density (first quartile). Similarly, Casey et al. [17], observed significant associations between UNGD and BW, particularly in the fourth quartile compared to the first quartile. However, Cushing et al., [20] did not find a significant relationship between low (MD: −5.9 gr, 95% CI: −36 to 24.3) and high (MD: −9.2 gr, 95% CI: −40.9 to 22.5) levels of flaring from OGD within a 5 km and BW. Tran et al. [15] demonstrated a reduction in BW with prenatal exposure to HF among rural mothers (MD: −73 gr, 95% CI: −131 to −15) but not in urban mothers (MD: −2 gr, 95% CI: −35 to 31).

A meta-analysis of three studies [15, 17, 20] examining the impact of maternal exposure to oil and gas extraction process on BW revealed a significant inverse association (MD: −30.36 gr, 95% CI: −43.98 to −17.28), with no significant heterogeneity among the studies (I²: 23.6%, P = 0.25) (Fig. 5B). Analysis of publication bias using funnel plot analysis indicated some bias, however, both Begg's test (P = 0.23) and Egger's test (P = 0.32) did not suggest significant publication bias. Even after adjusting for two imputed studies using the trim-and-fill method, there were no significant changes in the results (Supplementary File 3, Figure S18). Sensitivity analysis showed consistent findings even when individual studies were excluded. The Galbraith plot revealed that most studies fell within the expected range (Supplementary File 3, Figure S19). Subgroup analyses based on publication year (before 2016 and after 2016) showed a significant association between maternal exposure to oil and gas extraction activities and BW MD in both groups: before 2016 (MD: −26.60 gr, 95% CI: −42.02 to −11.19) and after 2016 (MD: −42.71 gr, 95% CI: −69.41 to −16.02), without significant heterogeneity among the studies (Supplementary File 3, Figure S20).

Discussion

This systematic review and meta-analysis aimed to investigate the association between maternal exposure to oil and gas extraction activities and various adverse birth outcomes. By synthesizing findings from 24 eligible articles up to June 16, 2024, we provided a comprehensive assessment of the potential risks posed by these environmental exposures during pregnancy. Our meta-analyses revealed significant associations between exposure to oil and gas extraction process and PTB, miscarriage, SGA and BW. Our findings showed that the exposure was associated with a 7% increase in the likelihood of PTB, a twofold increase in the likelihood of miscarriage, a 23% increase in the likelihood of SGA, and a mean decrease in BW of 30.36 g. However, we did not find significant associations between exposure and stillbirth, birth defects, or LBW. Notably, our analysis highlighted substantial heterogeneity among studies, particularly for PTB, miscarriage, birth defects, SGA, and LBW.

In our review of the literature, we identified 15 studies examining the association between maternal exposure to oil and gas extraction activities and PTB [13–15, 17–21, 56–61, 65] with inconsistent results. Some studies, such as those by Lin et al. [60], Tsai et al. [56], Yang et al. [59], Ha et al. [21], and Harville et al. [14], suggested an increased risk of PTB among women exposed to petrochemical municipalities, oil refinery plants, active power plants, and oil spills. However, other studies by McKenzie et al. [18], Stacy et al. [69] and Cairncross et al. [13]) did not find a significant association between exposure to NGD, UGD or HF and PTB, respectively. Additionally, some studies found an association only with high exposure levels to oil and gas activity pollutants [15, 17, 19, 20, 65]. Our meta-analysis of 14 studies indicated a significant association between maternal exposure and a 7% increase in the odds of PTB [13–15, 17, 18, 20, 21, 56–61, 65]. However, the substantial heterogeneity among studies suggests considerable differences in study populations, methodologies, exposure assessments and location which could influence the overall effect estimate.

The association between maternal exposure to oil and gas extraction activities and miscarriage was examined in six studies [16, 62, 63, 66, 67, 69]. Studies by Axelsson et al. [62], Xu et al. [67] and Sebastián et al. [63] identified elevated risks of miscarriage, with a 2–sixfold increase in risk associated with residential and occupational exposures. However, studies by Axelsson et al. [66], Bull et al. [69], and Oaka et al. [16] did not find any association between exposure and miscarriage. Our meta-analysis further confirmed a significant association between exposure and a twofold increase in the risk of miscarriage, with moderate heterogeneity. In contrast, all three included studies that evaluated the association between stillbirth and maternal exposure, did not find any relationship [16, 63, 64].

The association between maternal exposure to oil and gas extraction activities and birth defects was examined in five studies [18, 62, 64, 68, 70]. Two studies by did not specify birth defect types, finding no significant association with exposure [62, 64]. Three studies examined specific birth defects, with conflicting results [18, 68, 70]. Chevrier et al. [68] found a significant association with oral clefts, specifically cleft lip with cleft palate. Desrosiers et al. [70] found significant associations with intercalary limb deficiency and atrial septal defects but not with glaucoma/anterior chamber defects or colonic atresia/stenosis. McKenzie et al. [18] reported significant associations with congenital heart defects and neural tube defects, but not with oral clefts. Overall, the pooled effect estimated from four studies [18, 64, 68, 70], did not indicate a significant correlation between exposure and birth defects, despite high heterogeneity among the studies. This heterogeneity may be attributed to variations in the types of birth defects, methods of exposure assessment, methodology, and study populations.

The assessment of SGA reveals diverse findings. Some studies report a significant association between residential exposure to HF sites and increased odds of SGA [13], while others, such as Cushing et al. [20], did not find a significant association between exposure to low and high flaring from UOGD within 5 km and SGA. Stacy et al. [61] identified a higher risk of SGA associated with higher quartiles of UGD exposure. Additionally, Tran et al. [15] found an increased risk of SGA with exposure to HF in rural mothers but not in urban mothers. These inconsistent results highlight the challenges in elucidating the precise impact of environmental exposures on pregnancy outcomes. However, the pooled effect of exposure to oil and gas extraction activity from these four studies [13, 15, 20, 61] indicated a significant association with a 23% increase in the likelihood of SGA, with substantial heterogeneity. This significant heterogeneity among studies can be attributed to the different phases of oil and gas development and population characteristics.

The reviewed studies present controversial findings on the association between maternal exposure to oil and gas extraction activities and the risk of LBW [14, 15, 18, 21, 57, 62, 64]. Both Oliveira et al. [64] and Yang et al. [57] did not find significant differences in LBW rates when comparing exposed regions to reference regions. In contrast, McKenzie et al. [18] identified a negative association between mild exposure to NGD wells and LBW, indicating that lower levels of exposure might be protective. In contrast, Ha et al. [21] and Harville et al. [14] reported increased LBW risks associated with residential exposure to active oil and gas plants and oil spills, respectively. Tran et al. [15] highlighted the disparity in LBW risk between rural and urban mothers exposed to HF, with increased risks observed in rural areas only. The meta-analysis integrating six studies [14, 15, 18, 21, 57, 64] found no significant overall association between maternal exposure to oil and gas extraction activities and LBW. However, the substantial heterogeneity observed among the studies indicates variability in the results, likely due to differences in study populations, geographic locations, definitions, and measurements of exposure.

In terms of BW, Stacy et al. [61] and Casey et al. [17] both reported significant reductions in mean BW for women living in areas with higher densities of UGD compared to those in lower-density areas. Cushing et al. [20] found no significant relationship between flaring levels from OGD and BW. Tran et al. [15] observed a significant reduction in BW with prenatal exposure to HF among rural mothers. A meta-analysis of three studies [15, 17, 20] confirmed a significant inverse association between maternal exposure to oil and gas extraction activities and BW, indicating an overall reduction in BW. This meta-analysis showed no significant heterogeneity, suggesting consistent findings across these studies.

Exposure assessment

The heterogeneity observed among the included studies examining the association between maternal exposure to oil and gas extraction activities and adverse birth outcomes can be attributed to differences in exposure assessment methodologies [6, 71, 72]. Various studies employed distinct approaches to measure exposure, including proximity to oil and gas sites, duration of exposure, and specific temporal windows related to different phases of oil and gas development, such as drilling [65], hydraulic fracturing (fracking) [13, 15], and production [19]. Additionally, the type of oil and gas development varied, encompassing offshore oil operations [69], petrochemical plants [19, 21, 56, 57, 60, 62, 64, 66–68], natural gas development (NGD) wells [17, 18, 65], unconventional natural gas development (UNGD), unconventional oil and gas development (UOGD) [20], unconventional gas drilling (UGD) [61, 63, 70]. Some studies focused on residential exposure [13–15, 17–21, 56–65], while others assessed occupational exposure [16, 66–70], contributing to the variability in findings. Furthermore, different windows of vulnerability, such as specific trimesters of gestation [19], were considered, adding another layer of complexity. The lack of standardized exposure assessment methods and the diverse environmental and industrial contexts across study locations and populations likely contributed to the substantial heterogeneity in the results, impacting the overall effect estimates.

Study location and population

Another critical source of heterogeneity is the variation in study locations and populations. The included studies were conducted in different countries and regions, each with unique industry practices, regulatory environments, formation characteristics, and time periods of study. These geographical and demographic differences can affect the consistency of the associations observed between maternal exposure to oil and gas extraction process and birth outcomes. For example, variations in industry regulations and practices can lead to different levels of exposure and consequently different health outcomes. Furthermore, differences in population characteristics such as socioeconomic status, access to healthcare, and baseline health conditions can also influence the observed associations. These factors underscore the importance of considering contextual differences when interpreting the results of such studies.

Biological mechanisms

The biological mechanisms underlying the association between oil and gas extraction-related environmental exposures and adverse birth outcomes are multifaceted [13, 19, 72]. Exposure to pollutants from oil and gas extraction activities, such as volatile organic compounds (VOCs), particulate matter (PM), and heavy metals, can lead to systemic inflammation, oxidative stress, and endocrine disruption [73, 74]. These physiological changes can negatively impact fetal development by impairing placental function, reducing oxygen and nutrient delivery to the fetus, and causing direct toxic effects on fetal tissues [75, 76]. For example, exposure to VOCs and PM has been associated with increased levels of inflammatory markers and oxidative stress, which can contribute to preterm labor and restricted fetal growth [77]. Recent studies further support the role of maternal immune activation in response to environmental exposures as a key mediator of reduced fetal growth and lower birth weight [78]. Additionally, environmental contaminants have been shown to interfere with hormonal regulation and nutrient transport across the placenta, contributing to intrauterine growth restriction [79]. Moreover, endocrine-disrupting chemicals can interfere with hormonal regulation critical for maintaining pregnancy and supporting fetal development [80]. These biological pathways highlight the plausibility of the observed associations and underscore the potential health risks posed by environmental exposures related to oil and gas extraction activities.

Strengths and limitations

Our research has several notable strengths, including a comprehensive search strategy and a rigorous evaluation of study quality using the NOS and GRADE methods. The inclusion of a diverse range of study designs and populations enhances the generalizability of the results. However, several limitations must be acknowledged. Firstly, the included studies varied in design, population characteristics, exposure assessment methods, and outcome measures, which may affect the overall findings and limit the ability to conduct a consistent meta-analysis. Secondly, most studies were observational, making it challenging to establish causal relationships between maternal exposure and adverse birth outcomes. Thirdly, despite an exhaustive search strategy, publication bias cannot be entirely ruled out. Fourthly, the majority of studies were conducted in high-income countries, limiting the generalizability of the findings to populations with different exposure profiles and socio-economic backgrounds.

An additional key limitation concerns the lack of detailed reporting on the timing of exposure during pregnancy. Most studies did not specify whether exposure was assessed during a particular trimester or throughout gestation, instead broadly defining exposure as occurring “during pregnancy.” This limited our ability to investigate potential differences in effect estimates based on the developmental window of exposure—a critical factor in fetal susceptibility to environmental stressors. As such, we were unable to perform subgroup analyses by specific gestational exposure intervals, which could have provided insight into critical periods of vulnerability and contributed to explaining heterogeneity across studies.

In our meta-analysis, the limited number of available studies necessitated the combination of studies examining different phases of oil and gas development and various types of oil and gas activities. To address the impact of heterogeneity, we employed random-effects models, which account for variability between studies and provide more conservative effect estimates. Despite conducting subgroup analyses based on outcome, year of publication and meta-regression models to explore sources of heterogeneity, high between-study heterogeneity remained in some pooled estimates (e.g., PTB and miscarriage). While factors such as study design and exposure route were examined, further stratification by geographic region, exposure assessment approach, and type of industrial activity (e.g., petrochemical vs. unconventional gas development) was limited by incomplete and inconsistent data reporting across studies. While combining studies increased sample size and statistical power, it also introduced potential heterogeneity that must be carefully considered. This underscores the need for standardized exposure characterization in future observational studies, including geographic metrics, exposure quantification, and industrial process classification, which would enhance the capacity of future meta-analyses to disentangle contextual sources of heterogeneity.

Another important limitation relates to outcome classification for birth defects. Due to insufficient stratified reporting across primary studies, we were unable to perform separate meta-analyses by defect type (e.g., CHDs, NTDs, oral clefts). As a result, pooling heterogeneous congenital anomalies into a single category (“birth defects”) may obscure specific associations or introduce outcome misclassification. We encourage future studies to report birth defect subtypes explicitly and separately to enable more refined analyses.

Research gaps and future directions

Our findings highlight potential links between maternal exposure to oil and gas extraction processes and adverse birth outcomes; however, several research gaps remain. Most studies used observational designs and ecological or proximity-based exposure assessments, limiting causal inference. Future research should prioritize prospective cohort studies with individual-level exposure data, including biomonitoring and air quality measurements, and incorporate GIS-based exposure modeling to better define exposure windows during pregnancy. Additionally, there is a lack of standardized methods for assessing exposure across different phases of oil and gas development. Studies also often aggregate diverse birth defects into a single category, preventing meaningful subtype analysis. Future work should report specific defect types (e.g., neural tube defects, congenital heart defects) for more accurate risk assessment. Most evidence comes from high-income countries; therefore, research in low- and middle-income settings, where exposure levels and vulnerability may be higher, is urgently needed. Finally, integrating biomarkers of inflammation, oxidative stress, and endocrine disruption can help clarify the biological mechanisms underlying these associations. Addressing these gaps will require interdisciplinary collaboration and improved methodological consistency across studies.

Conclusion

This meta-analysis of 24 studies suggests that maternal exposure to oil and gas extraction processes may be associated with increased risks of PTB, miscarriage, small for SGA, and reduced BW. Specifically, we found a 7% increase in the likelihood of PTB, a doubling of the risk of miscarriage, a 23% increase in the likelihood of SGA, and an average reduction in BW of 30.36 g. However, no significant associations were identified for stillbirth, birth defects, or LBW. Despite these findings, several limitations must be acknowledged. There was substantial heterogeneity across studies, inconsistencies in exposure assessment methods, and some evidence of publication bias. Additionally, many of the confidence intervals were wide, indicating uncertainty around some estimates. These findings highlight the need for further investigation using standardized methodologies, improved exposure classification, and longitudinal data from larger and more diverse populations. Policymakers should consider these potential risks when regulating emissions and siting industrial activities near residential areas, particularly those with vulnerable populations such as pregnant women. Strengthened environmental monitoring, implementation of buffer zones, and enhanced public health surveillance are essential to mitigate potential harms. Continued research and policy action are necessary to reduce environmental health disparities and protect maternal and infant health.

Abbreviations

PTB

Preterm Birth

SGA

Small for Gestational Age

LBW

Low Birth Weight

BW

Birth Weight

BHC

Birth Head Circumference

Authors’ contributions

Ali Mohammad Latifi: conceptualization, methodology, writing—review & editing; Fatemeh Abdi: methodology, data curation, writing—review & editing; Mohammad Miri: methodology, literature search writing—review & editing, writing-original draft; Sara Ashtari: literature search, data extraction, data curation, risk of bias, writing—review & editing, writing-original draft; Seyedeh Noushin Ghalandarpoor-attar: writing-review & editing; Milad Mohamadzadeh: writing-review & editing; Abbas Ali Imani-Fooladi: methodology, writing—review & editing; Shahab Uddin: conceptualization, methodology, writing—review & editing; Amir Vahedian-azimi: supervisor, literature search, data extraction, data curation, risk of bias, writing—review & editing, writing-original draft.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data is provided within the supplementary information files.

Declarations

Ethics approval and consent to participate

Research ethics confirmation was received from the Baqiyatallah University of Medical Sciences under the code: IR.BMSU.REC.1402.098, and is registered with the International Prospective Register of Systematic Reviews (CRD42024555447).

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.