Introduction

Female infertility has increased in recent years [1]. It was estimated that this condition affects 1 in seven couples in developed countries [2]. Most cases of female infertility are related to specific disorders, namely, ovulatory disorders, endometriosis, chromosomal abnormalities and male factors [3,4,5,6,7]. There is also evidence that air pollutant could play a role in the pathogenesis of female infertility [8,9,10]. Air pollution appears to be a cause of concern for human health. For instance, it has been associated with an increased risk of cancer [11], and cardiovascular [12] and respiratory disorders in adults and children [13, 14]. In addition, air pollutants have been associated with adverse perinatal outcomes [15, 16].

Anthropogenic activities, namely traffic, industrial facilities and combustion of fossil fuels, which are particularly intense in large cities and in proximity of farms, are the main sources of health-related air pollutants. Air pollutants are in four main categories: gaseous pollutants (sulfur dioxide [SO2], nitrate oxide [NO2] and carbon monoxide [CO]), organic compounds (organic solvents or dioxins), heavy metals (lead and copper) and particulate matter (PM10 PM2.5-10and PM2.5) [17]. Ingestion and inhalation are the most common routes of exposure [17]. Ingestion is also facilitated by the fact that air pollution contributes to the contamination of food and water [18]. Some air pollutants, namely Cu, Pb and diesel exhaust seem to exert endocrine activity [19] that could affect female reproduction. Moreover, these endocrine “disruptors” exert estrogenic, antiestrogenic and antiandrogenic activity and some could interfere with the thyroid axis and influence metabolic disorders, such as insulin resistance and obesity, which are strictly related to infertility [20,21,22]. The increase in female infertility seems to parallel the increase in toxic emissions, which suggests that the impact of air pollution on human health could increase in the next years [23, 24]. In an attempt to summarize current evidence, we carried out a systematic review of studies devoted to the impact of air pollutions on female infertility.

Material and methods

Protocol and eligibility criteria

The present study was exempt from institutional and ethics board approval because it did not involve human intervention. We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [25]. The selection criteria are described according to PICO (Patients, Intervention, Comparison, and Outcomes). In detail, we evaluated fertility outcomes in women on reproductive age (in the general and IVF populations) in relation to exposure to air pollutants (Additional file 1: Table S1).

Search strategy

We conducted a systematic search using MEDLINE (PubMed) and SCOPUS databases to identify all relevant studies published before October 2017. Combinations of the following keywords and MESH search terms were used: “air pollutants” AND (“miscarriage” OR “embryo” OR “pregnancy” OR “IVF OR “fecundability” OR “infertility” OR “menstrual disorders”). No time or language restrictions were adopted, and queries were limited to human studies. We also hand-searched reference lists of relevant studies to ensure we did not miss pertinent studies.

Selection of studies

Four reviewers (G.C., M.M., G.CO and P.D.) independently evaluated titles and abstracts. Duplications were removed using Endnote online software and manually. Disagreements were resolved by discussion with a third authors (A.C. and C.D.), and if required, with the involvement of the most experienced authors (R.P.,C.A., G.D.). Articles were included only if they appeared in peer-reviewed journals. Case series, case reports, book chapters, congress abstracts and grey literature [26], which includes a range of documents not controlled by commercial publishing organization, were not included.

Data extraction

Data were extracted independently by four reviewers (G.C., M.M., G.CO and P.D.) using predefined data fields, including study quality indicators. Discrepancies were resolved by discussion with the senior authors (R.P., C.A. and G.D.).

Risk of bias, summary measures and synthesis of the results

The risk of bias and quality assessment of the included studies were performed adopting the Newcastle-Ottawa Scale (NOS) [27]. Four authors (A.C, C.D., G.C. and P.D.) independently assessed the risk bias for each study. The senior authors (R.P., C.A. and G.D.) resolved conflicts. The NOS score was used to evaluate the studies included, and judgment on each one was passed according to three issues: selection of the study group, comparability between groups, and ascertainment of exposed/not exposed cohorts. Primary outcomes were conception rate after spontaneous intercourse and live birth rate after IVF procedures. Secondary outcomes were first trimester miscarriage, stillbirths, infertility, number of oocytes retrieved and embryos transferred.

Results

Study selection and characteristics

A total of 4687 items were identified (Pubmed 2834 and Scopus 1853). A total of 2013 duplicates were removed manually and using the EndNote online library. The titles and abstracts of 2674 papers were scrutinized and 21 full papers were assessed for eligibility. Ten papers were excluded because they did not fulfill the inclusion criteria. Eleven articles were included in the analysis (Fig. 1). The characteristics of the studies included in the present study are reported in Table 1.

Fig. 1
figure 1

Flow chart

Full size image
Table 1 Characteristics, findings and risk of bias of included studies
Full size table

Risk of bias

The risk of bias was evaluated with the NOS score and is reported in Table 1.

Summary of results

We summarized our findings considering per each pollutant both IVF women and reproductive age women in general population (Table 2).

Table 2 Synthesis of results
Full size table

NO2

IVF cycles

Increases in NO2 concentrations were significantly associated with a lower live birth rate especially from embryo transfer to pregnancy test (OR 0.76, 95% CI 0.66–0.86, per 0.01 ppm increase) [28]. No effect on the number of oocytes retrieved or embryo transferred was observed [28].

General population

In a cross-sectional study involving women of reproductive age between 15 and 40 years, the fertility rate was not significantly associated with NO2 exposure (OR 0.97, 95% CI 0.94–1.003) [9]. In contrast, another retrospective cohort study, showed that there was a significant decreased fecundability ratio per each increase of 10 μg/m3 NO2 exposure (OR 0.72, 95% CI 0.53–0.97) [29]. Miscarriage rate was significantly increased in women exposed to NO2 compared to not exposed group (OR 1.16, 95% CI 1.01–1.28, per each 10-ppb increase in NO2 concentration) [30].

CO

General population

Exposure to CO was significantly associated with stillbirth in the second (OR = 1.14, 95% CI: 1.01, 1.28) and third trimester (OR = 1.14, 95% CI: 1.06, 1.24) [30]. No significant association with first trimester miscarriage was reported (OR = 1.14, 95% CI 0.98, 1.32) [30].

O3

IVF-cycles

A detrimental effect was observed in terms of live birth rate in women exposed to O3 from embryo transfer to date of live birth (OR 0.62, 95% CI 0.48–0.81, per 0.02 ppm increase) [28]. No effect on the number of oocytes retrieved or embryo transferred was observed [28].

General population

Only one study assessed the fecundability rate in the general population but no difference was reported between exposed and unexposed group [29].

PM2.5

IVF cycles

Exposure to PM2.5 during embryo culture was associated with a decreased conception rate (OR 0.90, 95% CI 0.82–0.99, per 8 μg/m3 increase) but not with live birth rates [28]. No effect on the number of oocytes retrieved or embryo transferred was observed [28].

General population

Multivariate hazard ratio (HR) analysis did not reveal any association with infertility considering 2 years average exposure (HR 1.09, 95% CI 0.77–1.55), 4 years average exposure (HR 0.91, 95% CI 0.78–1.05) and cumulative average exposure (HR 1.05, 95% CI 0.93–1.20) [8]. Consistently, in another trial multivariate analysis did not reveal any association with fertility rate [9]. On the other hand, The adjusted fecundability ratio was significantly decreased with each increase of 10 units (0.78, 95% CI 0.65–0.94) [29]. No statistically significant difference was observed in terms of late (second and third trimester) or early miscarriage (first trimester) [30].

PM2.5–10

General population

Multivariate HR analysis did not reveal any association between infertility and PM2.5–10 considering 2-year average exposure (HR 1.10, 95% CI 0.98–1.23), 4 year average exposure (HR 1.05, 95% CI 0.93–1.19) and cumulative exposure (HR 1.10, 95% CI 0.99–1.22) [8]. Conversely, another study reported a significant reduction of spontaneous fertility rate in women exposed to PM2.5–10 (incidence risk ratio: 0.88, 95% CI 0.84, 0.94) [9].

PM10

IVF cycles

No significant effect was observed in terms of live birth rate, number of oocytes retrieved or embryos transferred in exposed women undergoing their first IVF cycle [28]. Furthermore, no significant effect was observed in the amount of gonadotropin used, number of oocytes retrieved, number of MII oocytes, embryo quality, clinical and live birth rate [26, 31]. A higher risk of miscarriage was observed in women with a higher exposure to PM10 (> 56.72 μg/m3) comparing with those exposed to lower amount of PM10 (≤ 56.72 μg/m3) (OR 5.05 95% CI 1.04–25-51) [31].

General population

Multivariate adjusted HR analysis per year did not reveal any association with infertility considering 2 years average exposure (HR 1.04, 95% CI 0.96–1.11), 4 years average exposure (HR 0.99, 95% CI 0.91–1.08) and cumulative average exposure (HR 1.06, 95% CI 0.99–1.13) infertility [8]. Multivariate incidence risk (IRR) ratio adjusted did not reveal any association between PM10 exposure and fertility rate (IRR 0.99, 95% CI 0.96–1.02) [9]. A significant association with early miscarriage was observed in women exposed to over 56.72 µg/m3. [32].

SO2

IVF cycles

Exposure to SO2 did not significantly affect birth rate, number of oocytes retrieved or embryos transferred in women undergoing their first IVF cycle [28].

General population

No differences in terms of adjusted fecundability rate was observed per an increase of 10 units in the SO2 pollutant levels [29]. Conversely, in another study, the fecundability in the first unprotected menstrual cycle was significantly reduced only in couples exposed in the second month before conception to the following SO2 levels: 40–80 μg/m3 (OR 0.57, 95% CI 0.37–0.88); ≥ 80 μg/m3 (OR 0.49, 95% CI 0.29–0.81) [33]. The adjusted odds of miscarriage were significantly associated to SO2 exposure (OR 1.13, 95% CI 1.01–1.28 per each 3 ppb increase in concentration) [30].

Organic solvents

General population

Female exposure to air contaminated with organic solvents (hexane and hexane isomers, toluene, methyl ethyl ketone, acetone, ethyl acetate, isopropyl alcohol and dichloromethane, n-hexane, hexane isomers and toluene) was associated with reduced fecundability density ratio (FDR = 0.55, 95% CI 0.40–0.74) for low exposure (exposure index 0.01–0.14), and for high exposure (exposure index > 0.14), (FDR = 0.70, 95% CI 0.52.0.94). Moreover, exposure for less than 6 years was more strongly associated with reduced FDR in both low (FDR = 0.50, 95% Cl 0.30 to 0.83) and high exposure groups (FDR = 0.50, 95% CI 0.28–0.90) [34].

Traffic pollutants

General population

In a large cohort study involving 4979 women, traffic pollutants were associated with an increased but not with significant risk of miscarriage rate among women exposed to a maximum annual average of traffic pollutants within 50 m (AOR 1.18 95%, CI 0.87–1.60). A significant association was observed in a subgroup analysis involving African Americans (AOR = 3.11; 95% CI, 1.26–7.66) and nonsmokers (AOR = 1.47; 95% CI, 1.07–2.04) [35]. In another large cohort study, women living closer to a major road had a higher risk of infertility than did women living far from a major road (HR, 1.11 95% CI: 1.02–1.20) [8].

Coal combustion pollutants

General population

In a small prospective study of 260 women, the miscarriage rate was higher, albeit not significantly, in women exposed to coal combustion pollutants than in non-exposed women (OR 2.99, 95% CI 0.91–9.80) [36] .

Discussion

Only 11 studies have evaluated the potential effect of air pollutants on female reproduction.. In the IVF context, NO2 and O3 were associated with impaired live-birth rates. In addition, exposure to high levels of PM10 (> 56.72 μg/m3) resulted in an increased miscarriage rate after IVF procedures. Consistently, no study reported a significant effect on other quantitative (i.e. number of oocytes retrieved, number of embryos transferred, and consumption of gonadotropin) and qualitative (embryo quality, and number of MII oocytes) IVF outcomes [28, 31, 32]. In natural conception, reduced fecundability was associated with solvents and SO2 [33, 34]. Notably while abortion rate was associated with traffic pollutants [8, 35], and in particular SO2 and NO2 [30], no clear relation to coal combustion pollutants emerged [36]. Contrasting findings between infertility and PM2.5–10 were reported [8, 9].

Only three retrospective studies evaluated the effects of air pollution on IVF [28, 31, 32]. Although Legro and colleagues studied a large IVF population, the heterogeneity of IVF protocols and the lack of information about male partners represent two important limitation factors [28]. Moreover, the two studies conducted by Perin et al., are limited by the fact that only one pollutant was investigated and by the low number of cases enrolled [31, 32].

Eight studies have been conducted on the general population. Of the three prospective studies, the one by Mahalingaiah et al. is the largest (more than 36,000 patients) and has the highest qualitative NOS score [8]. The quality of evidence was lowest in the study by Mohorovic et al. as was the number of observations, and the authors did not report effect size for each air pollutant separately [36]. The same weakness emerges in the Green et al. paper, which however analyzed such important factors as work exposure, residential history and employment status of the population studied [35]. Of the five retrospective studies conducted to-date, the quality of evidence is highest in two large studies conducted by Faiz and colleagues [30] and by Slama and colleagues [29] demonstrating that air pollutants significantly affect fertility and stillbirths rates. The remaining three retrospective studies have several limitations, namely a paucity of data regarding the population studied [9], a low number of pollutant analyzed [33] and the methods adopted to assess exposure [34].

The relationship between air pollutants and spontaneous fertility was first observed in an animal model [37]. In detail, Mohallem et al. observed an increased implantation failure rate and a significant reduction of births in mice exposed to polluted city air compared to non-exposed mice [38]. Similarly, Veras et al. found significantly fewer antral follicles and a lower fertility index in mice exposed to traffic pollutants versus non-exposed mice [39].

The effect of air pollutants on human spermatogenesis has also been investigated [40,41,42,43]. The largest study, conducted by Hammoud et al., reported that PM2.5 exposure negatively correlated with sperm morphology and motility [40]. The negative effect of particulate matter was confirmed in a recent prospective cohort study that identified a significant association between PM10 and PM2.5 and sperm chromosomal abnormalities (i.e. disomy Y and disomy chromosome 21) [44].

The mechanism underlying the effect of air pollutants on female fertility is still a matter of debate. Several pathogenetic mechanisms have been proposed. Firstly, it was hypothesized that air pollutants could mimic the effect of androgens and estrogens in humans [45]. These endocrine-disrupting properties could exert their effect by interacting with nuclear receptor, the estrogen or androgen repertory or by interacting with specific targets in cytosol thus resulting in activation of the /Ras/Erk pathway [46]. Others have suggested that air pollutants could promote oxidative stress and inflammatory processes [17]. In this sense, we recently demonstrated that the addition of anti-oxidant factors to ovarian stimulation could improve reproductive outcome in women with polycystic ovarian syndrome [47]. However, whether antioxidant products could mitigate the effect of air pollutants on IVF outcomes remains to be determined. Finally, it has been suggested that air pollutants could exert a genotoxic effect. For instance, increased sperm DNA fragmentation was associated with exposure to elevated air pollution levels (at or above the upper limit of US air quality standards) [48]. Furthermore, DNA methylation seems to be significantly influenced by air pollutants [49]. Indeed, in a recent study of 777 men, an increase in air pollutant concentrations was significantly associated with F3, ICAM-1, and TLR-2 hypomethylation, and IFN-γ and IL-6 hypermethylation [50].

Our review has several limitations. First, most of the studies included in our analysis are observational and retrospective, and hence more prone to bias. Second, exposure ascertainment was heterogeneous among studies. Most of the trials assessed air quality using a specific air monitoring station, others estimated exposure according to proximity to the potential source [8, 35, 36]. In addition, the reference levels of each pollutant varied significantly among studies. Lastly, the populations investigated as well as the definitions used to assess infertility and miscarriage were also heterogeneous. These factors render a meta-analytic and quantitative approach to this issue challenging.

In conclusion, our meta-analysis suggests there is a close association between female infertility and air pollution. However, a more robust meta-analytic approach is required before any definitive conclusion can be reached.