Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jan 1.
Published in final edited form as: Int Arch Occup Environ Health. 2017 Sep 16;91(1):47–56. doi: 10.1007/s00420-017-1257-4

The Gulf oil spill, miscarriage, and infertility: The GROWH Study

Emily W Harville 1,*, Arti Shankar 2, Leah Zilversmit 1, Pierre Buekens 1
PMCID: PMC5754231  NIHMSID: NIHMS906758  PMID: 28918439

Abstract

Purpose

To examine whether reported exposure to the Gulf oil spill (2010) was related to reproductive reported miscarriage or infertility.

Methods

1524 women aged 18–45 recruited through prenatal and Women, Infant, and Children (WIC) clinics, and community events were interviewed about their experience of the oil spill and reproductive history. 1434 women had information on outcomes of at least one pregnancy, and 633 on a pregnancy both before and after the spill. Generalized estimating equations were used to examine the relationship between contact with oil and economic and social consequences of the spill with postponement of pregnancy, miscarriage, and infertility (time to pregnancy >12 months or reported fertility issues), with adjustment for age, race, BMI, smoking, and socioeconomic status. Results were compared for pregnancies occurring prior to and after the oil spill.

Results

77 (5.1%) women reported postponing pregnancy due to the oil spill, which was more common in those with high contact with oil or overall high exposure (aOR 2.92, 95% CI 1.31–6.51). An increased risk of miscarriage was found with any exposure to the oil spill (aOR, 1.54, 95% CI 1.17–2.02). Fertility issues were more common in the overall most highly exposed women (aOR 1.88, 1.19–2.95), when the data were limited to those with pregnancies before and after. However, no particular aspect of oil spill exposure was strongly associated with the outcomes, and effects were almost as strong for pregnancies prior to the oil spill.

Conclusions

The oil spill appears to have affected reproductive decision-making. The evidence is not strong that exposure to the oil spill was associated with miscarriage or infertility.

Keywords: petroleum, abortion, spontaneous, fertility, pregnancy

Introduction

The alarming reports that dolphins in Barataria Bay are failing to calve in the aftermath of the 2010 Gulf oil spill (Lane et al. 2015) added to the already-expressed concerns of community members that the oil spill could have had adverse effects on pregnant women (Goldstein et al. 2011). Whether such concern is plausible is not completely clear. Although to our knowledge no studies have examined the relationship between oil spills and miscarriage or infertility specifically, petroleum production has been a source of concern as a possible danger to perinatal health under other circumstances, such as in petrochemical plants and near fracking sites (Colorado Department of Public Health and Environment 2014; Schlanger 2014; Xu et al. 1998). The environmental toxicants that were most likely to have been released by the oil spill are volatile organic compounds (VOCs), heavy metals, and polyaromatic hydrocarbons (PAHs); evidence for effects of these pollutants on reproductive failure, represented hereby infertility and miscarriage, in humans is mixed, with occasional studies finding substantially increased risk, many studies finding no effect, consistent concerns about recall bias and confounding, and reviews generally concluding “limited or inadequate evidence” (Bukowski 2001; Dechanet et al. 2011; Duong et al. 2011; Hertz-Picciotto 2000; Pineles et al. 2014; Wigle et al. 2008).

Although chemical exposures are perhaps the most immediate concern, social effects should be considered as well. Many people lost employment and income due to the fishing bans and drilling moratorium (Aldy 2014). Others found the spill (and associated images, such as oil-soaked pelicans) to be distressing (Osofsky et al. 2011). Stressful life events in the first trimester have been linked to spontaneous abortion (Neugebauer et al. 1996), as have economic downturns (Bruckner et al. 2016). More general collective trauma studies have sometimes found an increase in miscarriage after floods or tsunami (Fujimori et al. 2014; Neuberg et al. 2001), but the difficulties of assembling a reasonable control group makes interpretation of these studies difficult. In this study, we explored whether reported exposure to the oil spill – either direct contact with oil or related stressors – was related to reported miscarriage or infertility.

Methods

Participants

The GROWH (Gulf Resilience on Women’s Health) began in 2011; this analysis contains data collected through March 2016. Women were recruited from prenatal, health, and WIC clinics; day care centers; and community events and gathering places in southeastern Louisiana (targeting Lafourche, Plaquemines, St. Bernard, Terrebonne, and the West Bank of Jefferson and Orleans Parishes). Eligibility criteria include: aged 18–45, living in the Gulf area during the oil spill, and, if pregnant, carrying a singleton gestation. Women were interviewed, completed a questionnaire (usually on the spot, although taking it home and returning it by mail was allowed), and provided saliva and blood samples. 1620 filled out at least one questionnaire or interview, including 443 women who were pregnant at the time of the interview.

Measures

Exposures

Women were interviewed about their experience of the oil spill using measures from several sources, including questions about: (1) a participant’s involvement in work on the clean-up and contact with oil, taken from the Gulf Workers’ Study (National Institutes of Health); (2) direct exposure to the oil spill, taken from studies performed after the Exxon Valdez spill (Palinkas et al. 1993); (3) the social and economic effects of the oil spill, from a previous study (GUMBO, R03 NR012052), and (4) involvement in litigation, after Exxon Valdez studies (Palinkas et al. 1993; Picou et al. 2004). Confirmatory factor analysis was used to see if the patterns of grouping of similar response questions matched the underlying latent constructs: financial/income consequences; direct contact with oil (both dichotomized as any/none and none/some/a lot); oil spill-related trauma (damage to people or own property); loss of use of the coast (damage to areas where one or one’s family fishes, boats, or goes to the coast or beach). In addition, separate variables for any exposure to the oil spill (0 versus 1) and total exposure to the oil spill (sum of the above individual experiences – money, direct contact, trauma, loss of use, and litigation, weighted equally; theoretical range was 0 to 10; range in this sample was 0 to 9) were created.

Outcomes

As part of a list of possible behavior changes due to the oil spill, women were asked, “As a result of the oil spill, have you postponed getting pregnant?” A second question included in the questionnaire was “Was there ever a time when you wanted to get pregnant, but weren’t able to?”, and, if so, how old was she when this happened. Based on her birthdate, this was translated into having occurred before or after the oil spill. Timings that were within six months of the oil spill and thus could not be precisely calculated as occurring before or after were omitted from analysis. All women were asked these questions, regardless of pregnancy history or status.

Each woman was also asked for a reproductive history including up to 8 pregnancies. Questions for participants included date of/age at each pregnancy and its outcome. The woman was also asked for how long she was having unprotected sex before she got pregnant (time to pregnancy [TTP]).

The start date of each pregnancy was determined to occur before (prior to April 20, 2010) or during or after the oil spill (on or after April 20, 2010). If the precise date of the start or end of pregnancy was not known (as was often the case for a miscarriage) and it was estimated to have occurred (based on age or year of occurrence) within 6 months of the oil spill, it was omitted from the analysis.

Sample

Reproductive decision-making

1524 women had data on at least one oil spill exposure and postponement of pregnancy (women lost to follow-up were less likely to be black (p<0.01) and older (mean difference 3.3 years, p=0.02); there were no differences in parity, income, BMI, pregnancy status, or parish of residence). Oil spill exposure was examined as a predictor of postponing pregnancy using logistic regression, first unadjusted, then with control for age at interview (continuous), BMI (continuous), race, income (ordinal), education (ordinal), smoking (recent/not), year of interview, and age at first pregnancy (set to the mean for those with no pregnancies).

Miscarriage and infertility

1434 women reported on the outcome of at least one pregnancy, and for 1419 women (3504 pregnancies) these could be dated. 633 of these women had outcome data on a pregnancy both before and after the spill; compared to the women with pregnancies in only one time period, these women were more likely to be older, more parous, and from Jefferson or St. Bernard Parish (no differences in BMI, smoking status, income, or race).

The infertility analysis draws on two sources of information: reported time periods when a woman wished to become pregnant but was not able to, and TTP for each pregnancy. For the first analysis, fertility issues were defined as attempting pregnancy for ≥12 months, either with a “yes” response to the question “Was there ever a time when you wanted to become pregnant, but were not able to?” and answer to the follow-up question of “one year or more”, or a reported TTP ≥ 12 months for any pregnancy. The analysis was limited to those who were 18 years or older at the time of the oil spill, and women with untested fertility (operationalized as “no” to the attempting pregnancy question and no reported pregnancies; n=41) were excluded as well. This left 1164 total women with post-oil spill data and 1046 women with pre- and post-oil spill data.

Statistical analysis

To examine whether any relationships seen could be attributed to correlated errors or reporting, we examined pregnancies both before and after the oil spill. As the oil spill could not be the cause of pregnancy outcomes that occurred prior to the spill, this provides a check as to whether certain women are systematically likely to report both greater exposure to the oil spill and greater risk of the outcomes (or vice versa). A subanalysis was limited to women with data about pregnancies or infertility both before and after the spill, to allow for a repeated-measures analysis.

First, the woman’s whole history was considered. For miscarriage, all pregnancies were examined. For infertility, pregnancies and the time period of attempting pregnancy were categorized as occurring prior to or after the pregnancy. Pregnancies/time periods prior to the oil spill were categorized as unexposed, while pregnancies after the oil spill categorized as exposed or unexposed, depending on the particular indicator under study. Generalized estimating equations, logistic models with an autoregressive correlation matrix, were used to control for correlation within woman (up to 8 pregnancies/woman for miscarriage, 2 observations/woman for infertility). Both unadjusted and adjusted logistic regression were used, Models of miscarriage were adjusted for age at pregnancy (continuous), gravidity at that pregnancy, BMI (continuous), income (ordinal), education (ordinal), smoking (dichotomous), and weight gain during pregnancy (continuous) (details of categories in table 1). Multiple imputation was used to deal with missing covariate data; most frequently missing was income (4%).

Table 1.

Characteristics of women participating the GROWH study, south Louisiana, 2011–2016.

data on oil spill exposure and postponing pregnancy (n=1524) any pregnancy with outcome data (n=1434) both a pre- and post-oil spill pregnancy (n=633)
N % N % N %
Participant characteristics
Age
 18–25 467 31.8 407 29.9 101 16.6
 >25–30 418 28.4 393 28.8 226 37.2
 >30–35 302 20.5 290 21.3 172 28.3
 >35 283 19.3 273 20.0 109 17.9
Partnered
 yes 621 39.7 561 40.3 264 42.4
 no 944 60.3 832 59.7 358 57.6
Race
 white 439 28.3 398 28.0 181 28.8
 black 949 61.2 894 62.9 392 62.3
 other 164 10.6 130 9.1 56 8.9
Income
 <$15K 677 46.0 638 46.4 293 47.5
 $15K–35K 495 33.7 463 33.7 210 34.0
 >=$35K 299 20.3 275 20.0 114 18.5
Smoker (any smoking, last two years)
 Yes 415 27.3 404 28.4 191 30.4
 no 1104 72.7 1021 71.7 438 69.6
BMI
 <=20 103 6.8 81 6.0 37 6.1
 >20–25 350 23.0 306 22.5 122 20.2
 >25 378 24.9 341 25.0 159 26.3
 >30 690 45.4 634 46.6 286 47.4
Parish of residence
 Jefferson 582 40.4 563 40.7 203 44.4
 Orleans 312 20.4 275 19.9 74 16.2
 Plaquemines 145 10.2 138 10.0 35 7.7
 St. Bernard 128 7.4 125 9.0 47 10.3
 Lafourche 122 8.6 117 8.5 46 10.1
 Terrebonne 137 9.9 130 9.4 40 8.8
 Other 42 3.1 34 2.5 12 2.5
Outcome of first pregnancy among those with a pregnancy
 livebirth 1117 78.2 1117 78.2 508 80.6
 stillbirth 31 2.2 31 2.2 18 2.9
 miscarriage 145 10.2 145 10.2 77 12.2
 abortion 33 2.3 33 2.3 23 3.7
 molar/ectopic/etc. 2 0.1 2 0.1 2 0.3
 still pregnant 100 7.0 100 7.0 2 0.3
Exposure to the oil spill
any exposure 796 51.7 747 52.3 326 51.7
Income loss due to oil spill 448 29.1 420 29.4 169 26.7
Trauma/property damage 88 5.8 85 6.0 43 6.9
Loss of use of the coast due to the oil spill 615 41.3 579 41.8 236 38.5
litigation 373 24.5 352 24.6 155 24.5
Indicators of contact with oil
 0 1256 81.3 1160 81.2 520 82.4
 1 173 11.2 161 11.3 71 11.3
 2+ 116 7.5 107 7.5 40 6.3
Indicators of total oil spill exposure
 0 685 44.2 623 43 290 45.8
 1–2 460 29.7 432 432 187 29.5
 3–4 261 16.8 239 239 97 15.3
 5+ 145 9.4 140 140 59 9.3
Open in a new tab

Infertility models were adjusted for BMI, race, income, education, smoking, year of interview, and age at interview. A second infertility analysis analyzed just the TTP pregnancy data (up to 8 pregnancies/woman), with 857 total women and 428 women with TTP data both before and after the oil spill, with adjustment for age at pregnancy, gravidity at pregnancy, BMI, race, income, education, and smoking). A sensitivity analysis examined TTP ≥ 6 months as the cut-off instead of 12 months.

Second, a similar GEE model was run, limited to women who had both a pre- and post-oil spill pregnancy (for miscarriage) or information about the time frames pre- and post-oil spill (for infertility).

Sensitivity analyses limited the analysis to those whose pregnancies occurred within two years of the oil spill (n=707 overall and 437 with pregnancies both before and after). The effect of removing those who were pregnant at the time of the interview from the “non-miscarriage” group was also assessed, but most women were interviewed late enough in pregnancy that a miscarriage would have been unlikely.

Finally, an analysis was conducted to examine whether reporting bias or confounding was a likely explanation for observed associations. Models were run examining oil spill exposure as a fixed exposure, predicting pregnancy outcomes both before and after the oil spill, with an interaction term to test whether the estimates were statistically different. As the oil spill could not logically cause events prior to its occurrence, we examined whether the effect estimates were equivalent in the two time periods. The interaction between timing of the pregnancy (pre-/post-oil spill) and reported exposure was examined, and a stratified analysis performed for pre- and post-oil spill pregnancies.

The study methods were approved by the Institutional Review Boards of Tulane University, Ochsner, and WIC, and all participants provided written informed consent.

Results

Women were predominantly low-income and black, with a high BMI. A large majority had at least one child (Table 1). 145 (10.2%) reported miscarriage as the outcome of the first pregnancy. 77 (5.1%) women reported postponing pregnancy due to the oil spill, which was more common in those exposed to the spill (table 2). 19.4% of women had at least one indicator of infertility pre-oil spill, and 17.6% post-oil spill.

Table 2.

Oil spill exposure and reported postponement of pregnancy after the Gulf oil spill (2010) in a cohort of Southern Louisiana women, 2011–2016.

unadjusted adjusteda
N with outcome % OR 95% CI OR 95% CI
Any oil spill exposure 1.78 1.11–2.88 1.92 1.17–3.15
 Yes 50 6.4
 No 27 3.7
Income loss due to oil spill 1.42 0.88–2.29 1.55 0.95–2.55
 Yes 28 6.3
 No 49 4.5
Trauma/property damage 1.43 0.60–3.38 2.00 0.82–4.89
 Yes 6 6.9
 No 70 4.9
Loss of use of the coast due to the oil spill 1.50 0.93–2.39 1.58 0.97–2.57
 Yes 37 6.1
 No 36 4.2
Litigation 2.89 1.82–4.59 2.72 1.70–4.35
 Yes 36 9.7
 No 41 3.6
contact with oil
 low 56 4.5 1.00 1.00
 medium 10 6.0 1.34 0.67–2.69 1.38 0.68–2.80
 high 10 8.8 2.03 1.01–4.10 2.13 1.04–4.39
total oil spill exposure
 low 20 3.0 1.00 1.00
 medium-low 27 6.0 2.06 1.14–3.71 2.08 1.14–3.77
 medium-high 20 7.8 2.75 1.46–5.21 2.86 1.49–5.48
 high 10 6.9 2.43 1.11–5.30 2.92 1.31–6.51
Open in a new tab
a

adjusted for age at interview, BMI, race, income, education, smoking, year of interview, and age at first pregnancy. In all cases, reference group is those without the specified type of exposure; those unexposed in one analysis might be exposed in another.

Overall risk of miscarriage was lower for pregnancies after the spill, although this was likely due partly to the relatively large number of currently pregnant women – when women with ongoing pregnancy were excluded, the protective effect disappeared (aOR 0.89, 0.66–1.20). An increased risk of miscarriage was found with any reported exposure to the oil spill (aOR 1.54, 1.17–2.02), an effect that was similar or stronger when limited to those with pregnancies before and after the spill (aOR 1.79, 1.25–2.55), or which occurred within two years of the spill (table S1, aOR 1.87, 1.12–3.12 and 2.56, 1.37–4.79 for those with both a pre- and post-oil spill pregnancy). No specific aspect of oil spill exposure (income, contact, trauma, etc.) could be identified as a stronger predictor than any other. However, reported exposure to the oil spill was also associated with miscarriage in pregnancies prior to the oil spill (aOR 1.64, 1.06–2.54; table 4). No significant interactions indicating stronger effects on pregnancies after the oil spill.

Table 4.

Reported oil spill exposure and miscarriage as outcome of pregnancy, limited to those with both a pre- and post-oil spill pregnancy, reported oil spill exposure as a predictor of pregnancy outcomes occurring prior to and after the oil spill (n=633 women, 2011 pregnancies)

unadjusted adjusteda
Pregnancy occurred pre-oil spill Pregnancy occurred post-oil spill Pregnancy occurred pre-oil spill Pregnancy occurred post-oil spill
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
any exposure 2.00 1.29–3.10 2.15 1.27–3.64 1.64 1.06–2.54 1.98 1.15–3.41
income 2.01 1.33–3.04 1.92 1.14–3.23 1.80 1.18–2.74 1.65 0.98–2.76
contact with oil 1.41 0.87–2.30 1.40 0.73–2.66 1.29 0.79–2.12 1.10 0.56–2.16
trauma 0.93 0.42–2.05 1.36 0.57–3.23 0.76 0.35–1.66 1.27 0.55–2.92
coast 2.08 1.37–3.15 2.06 1.21–3.50 1.81 1.19–2.75 1.73 0.98–3.03
litigation 1.40 0.89–2.21 1.34 0.74–2.44 1.58 1.01–2.47 1.40 0.76–2.56
contact with oil
 low 1.00 1.00 1.00 1.00
 medium 1.26 0.69–2.30 1.60 0.77–3.30 1.20 0.68–2.12 1.25 0.57–2.72
 high 1.70 0.84–3.42 1.05 0.32–3.45 1.45 0.67–3.13 0.85 0.25–2.89
total oil spill exposure
 low 1.00 1.00 1.00 1.00
 medium-low 1.80 1.08–3.00 1.60 0.82–3.11 1.72 1.02–2.90 1.46 0.73–2.92
 medium-high 2.67 1.53–4.68 3.08 1.65–5.75 2.70 1.56–4.68 2.56 1.38–4.74
 high 2.04 1.04–4.01 1.68 0.67–4.19 1.65 0.83–3.31 1.32 0.54–3.24
Open in a new tab
a

adjusted for age, gravidity, weight gain during pregnancy, BMI, race, income, education. In all cases, reference group is those without the specified type of exposure; those unexposed in one analysis might be exposed in another.

Fertility issues were equally commonly reported before and after the spill, and were more common with the highest overall oil spill exposure (Table 5). Long TTP was also more common in those reporting income loss (aOR 1.44, 1.01–2.07; Table S2). Results were similar if 6 months instead of 12 months was used as the cut-off (Table S2). No significant interactions indicated stronger effects on pregnancies after the oil spill.

Table 5.

Oil spill exposure and self-reported issues becoming pregnant in a cohort of southern Louisiana women, 2011–2016

all with data (n=1164) All with data on pre- and post-oil spill time frames (n=1046)
N with outcome % unadjusted adjusteda unadjusted adjusteda
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
post-oil spill 0.90 0.74–1.10 0.89 0.73–1.09 0.99 0.80–1.22 0.99 0.80–1.22
 Yes 192 17.6
 No 216 19.4
any exposure 0.91 0.72–1.15 0.87 0.68–1.10 0.99 0.78–1.27 0.97 0.75–1.25
 Yes 101 17.5
 No 307 18.9
Income 1.20 0.90–1.59 1.16 0.87–1.55 1.32 0.98–1.78 1.33 0.98–1.79
 Yes 69 21.4
 No 339 18.0
Trauma 0.96 0.53–1.75 0.95 0.51–1.76 1.01 0.54–1.88 1.04 0.55–1.96
 Yes 13 18.6
 No 393 18.5
Coast 0.97 0.75–1.26 0.92 0.71–1.20 1.06 0.80–1.38 1.03 0.78–1.36
 Yes 81 18.4
 No 321 18.5
litigation 1.15 0.85–1.56 1.17 0.85–1.61 1.26 0.92–1.74 1.27 0.91–1.77
 Yes 55 20.1
 No 353 18.3
contact with oil
 low 366 18.2 1.0 1.0 1.0 1.0
 medium 27 22.9 1.28 0.83–1.99 1.30 0.84–2.01 1.40 0.89–2.22 1.46 0.93–2.29
 high 15 19.2 1.01 0.58–1.77 1.01 0.58–1.78 1.09 0.61–1.97 1.14 0.64–2.04
total oil spill exposure
 low 292 18.4 1.0 1.0 1.0 1.0
 medium-low 56 16.9 0.91 0.68–1.23 0.90 0.66–1.23 0.98 0.71–1.35 0.97 0.70–1.35
 medium-high 32 17.3 0.96 0.65–1.40 0.89 0.60–1.33 1.01 0.67–1.52 0.96 0.63–0.47
 high 28 26.9 1.52 0.97–2.38 1.50 0.96–1.36 1.80 1.14–2.85 1.88 1.19–2.95
Open in a new tab
a

adjusted for age, gravidity, BMI, income, education, smoking, year of interview, weight gain during pregnancy. In all cases, reference group is those without the specified type of exposure; those unexposed in one analysis might be exposed in another. Self-reported issues during pregnancy=time to pregnancy≥12 months or time of attempting pregnancy without success of 12 months or more. “All with data” refers to all women who provided TTP or pregnancy attempt information; data on the pre- and post-oil spill time frame refers to those who provided TTP or pregnancy attempt information for the time frames before and after the Gulf oil spill (2010).

Discussion

In this study of the 2010 Gulf oil spill, few aspects of oil spill exposure were associated with miscarriage or infertility. In a few cases, general indicators of exposure were associated with these outcomes; this could be a chance finding given that multiple comparisons were conducted (given approximately 10 indicators of exposure, at least one or two chance associations would be expected across the miscarriage and infertility analyses). Still, the size of the association is consistent with the size of the effects seen in previous studies indicating associations between psychological or economic stress and miscarriage or infertility (Bruckner et al. 2016; Lynch et al. 2014). Stress could affect conception and miscarriages through behavioral pathways, such as reduced frequency of intercourse or increased smoking by a woman (controlled for, though residual confounding is a possibility) or her partner, and hormonal or immunological changes (Ferin 1999; Kwak-Kim et al. 2014; Whirledge and Cidlowski 2013). Overall, however, the effect of the oil spill on risk of miscarriage was similar in pregnancies pre- and post-oil spill, suggesting minimal effects.

No associations were found with reported direct contact with oil. We are not aware of any studies that have examined the relationship between oil spills and miscarriage or infertility specifically, although petroleum production has been a source of concern under other circumstances (Colorado Department of Public Health and Environment 2014; Schlanger 2014; Xu et al. 1998) Evidence is inconsistent for effects on miscarriage or fertility for the three major types of environmental pollutants associated with the spill: VOCs, metals, and PAHs (Bukowski 2001; Duong et al. 2011; Wigle et al. 2008)(Detmar et al. 2006; Hertz-Picciotto 2000; Hombach-Klonisch et al. 2005; Wu et al. 2010). There is little evidence on the human reproductive effects of dispersants, a major public concern; animal studies are mixed in their conclusions (Pollino and Holdway 2002; Rowe et al. 2009; Van Scoy et al. 2012; Wooten et al. 2012). Overall, however, the lack of effect was not surprising given the fairly low levels of exposure and the mixed evidence for effects of related toxicants on the outcomes studied.

We examined all reported pregnancies, and also the subset of women with a pregnancy both before and after the spill. As the oil spill could not logically have caused outcomes that happened prior to its occurrence, this provides some check on the possibility of correlated over- or under-reporting, as each woman serves as her own control. However, women included in this analysis are necessarily older and higher gravidity at the later pregnancy, which means the earlier pregnancies are an imperfect control and residual confounding is possible. If the oil spill caused women to postpone pregnancy, they would be older at time of pregnancy and thus could have increased risk for infertility or miscarriage, another possible behavioral pathway. If women postponed pregnancy completely, the most highly exposed could be selected out of the miscarriage analysis altogether, although given a fairly small proportion of women who reported postponing pregnancy, the number of women this would affect is most likely small relative to the number included in the analysis and is unlikely to be a source of major bias. It is possible that women might have become more vigilant in noticing or reporting miscarriages and delayed fertility in the aftermath of the oil spill, but pregnancies after the spill were not more commonly reported to have ended in miscarriage, nor was reported infertility more common. Nonetheless, this cannot be completely ruled out as a cause of associations seen. A true effect could have been masked if the most exposed people moved to another region or state and thus could not be included in the sample. An overrepresentation of women who were both highly-exposed and who experienced or were at high risk for pregnancy complications could produce a spurious association, but the demographics, geographical distribution, and medical outcomes of the sample do not suggest this is likely. Compared to the population of births in the parishes we were studying (based on vital statistics), our sample is more likely to be black, somewhat older, and less likely to be married, most likely due to the extensive recruitment in WIC clinics and the retrospective report of pregnancy history. The proportion of women reporting a miscarriage is within the range that would be expected, given a young population (Lang and Nuevo-Chiquero 2012), and the reported prevalence of lifetime infertility is similarly comparable with national studies (Schmidt 2009).

Studying reproductive failure in humans presents challenges. Early miscarriages frequently are not reported to medical providers, and may not even be noticed. Many women meet the clinical definition of infertility (>12 months of unprotected sex without becoming pregnant) without considering themselves infertile. Many others desire pregnancy but do not seek medical treatments. One strength of this study is the unselected nature of the population with respect to fecundability; many of the women reported having a TTP longer than a year but not infertility, and women were asked about their pregnancy postponement and issues getting pregnant regardless of pregnancy history. For these reasons, self-report often provides a fuller picture of these outcomes, though the outcomes cannot always be independently verified. Validation studies of self-report find reasonable if not excellent agreement with medical records and prospective data (Joffe et al. 1995; Kristensen and Irgens 2000); late miscarriages and long TTPs (dichotomized) are particularly well-reported (Cooney et al. 2009; Wilcox and Horney 1984). Similarly, no biomarker exists that could independently verify exposure to the oil spill, especially during the time frame of this study (2011–2016; the oil spill occurred in 2010). It is possible that detailed biomarker measurements taken close in time to the spill would demonstrate stronger associations.

This study provides some moderate evidence for a combined effect of behavioral, social, and economic effects of the oil spill on miscarriage or infertility, but no particular evidence for effects of chemical exposure. Future studies should focus on identifying chemical signatures that can be directly related to oil spill exposures, and attempting to understand mechanisms by which social or economic stressors might affect reproductive success.

Supplementary Material

supplement

Table 3.

Reported oil spill exposure and miscarriage in a cohort of southern Louisiana women, 2011–2016

any pregnancy with outcome data (n=1419 women/3504 pregnancies) both a pre- and post-oil spill pregnancy, outcome data on both pregnancies (n=633/n=2011)
N with outcome % unadjusted adjusteda unadjusted adjusteda
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
pregnancy after spill 0.76 0.59–0.98 0.68 0.51–0.92 0.86 0.73–1.00 0.66 0.41–1.06
 yes 120 8.0
 no 199 10.0
any exposure 1.69 1.29–2.20 1.54 1.17–2.02 2.06 1.46–2.91 1.79 1.25–2.55
 yes 205 11.1
 no 113 6.9
income 0.97 0.66–1.44 0.85 0.58–1.25 1.23 0.78–1.94 1.24 0.76–2.03
 yes 35 9.1
 no 284 9.1
trauma 1.07 0.50–2.30 0.95 0.45–1.99 1.08 0.47–2.50 1.05 0.44–2.50
 yes 9 10.5
 no 308 9.1
coast 1.13 0.79–1.62 1.05 0.70–1.57 1.16 0.76–1.76 1.18 0.72–1.91
 yes 57 10.3
 no 255 8.8
litigation 0.88 0.56–1.40 0.88 0.55–1.42 0.99 0.57–1.70 1.20 0.67–2.15
 yes 30 8.3
 no 289 9.2
contact with oil
 low 290 8.9 1.0 1.0 1.0 1.0
 medium 16 10.3 1.13 0.63–2.03 0.98 0.55–1.77 1.20 0.59–2.44 1.06 0.49–2.29
 high 13 14.4 1.65 0.74–3.66 1.55 0.63–3.78 0.75 0.22–2.57 0.68 0.19–2.41
total oil spill exposure
 low 242 9.0 1.0 1.0 1.0 1.0
 medium-low 39 8.4 0.86 0.56–1.32 0.87 0.56–1.36 0.84 0.48–1.45 0.97 0.53–1.76
 medium-high 28 12.1 1.33 0.82–3.15 1.21 0.71–2.04 1.58 0.94–2.63 1.63 0.94–2.84
 high 10 8.3 0.92 0.46–1.86 0.75 0.36–1.54 0.88 0.38–2.07 0.82 0.34–2.00
Open in a new tab
a

adjusted for age, gravidity, weight gain during pregnancy, BMI, race, income, education. In all cases, reference group is those without the specified type of exposure; those unexposed in one analysis might be exposed in another. Generalized estimating equations used to predict outcome of pregnancies with oil spill exposure as a time-varying exposure. First columns include all women with at least one post-oil spill (2010) pregnancy. Second columns include only those women with a pregnancy before and after the oil spill.

Acknowledgments

Funding sources and human subjects

This research was supported by NIH grant U19 ES020677 and the Baton Rouge Area Foundation.

The study methods were approved by the Institutional Review Boards of Tulane University, Ochsner, and WIC, and all participants provided written informed consent.

Abbreviations

WIC

Women, Infants, and Children (supplemental nutrition program)

BMI

body mass index

TTP

time to pregnancy

VOC

volatile organic compounds

PAH

polycyclic aromatic hydrocarbons

OR

odds ratios

CI

confidence intervals

Footnotes

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. Aldy JE. Labor market impacts of the 2010 deepwater horizon oil spill and offshore drilling momentum. Harvard Kennedy School Taubman Center for State and Local Government; 2014. [Google Scholar]
  2. Bruckner TA, Mortensen LH, Catalano RA. Spontaneous pregnancy loss in denmark following economic downturns. Am J Epidemiol. 2016 doi: 10.1093/aje/kww003. [DOI] [PubMed] [Google Scholar]
  3. Bukowski JA. Review of the epidemiological evidence relating toluene to reproductive outcomes. Regul Toxicol Pharmacol. 2001;33:147–156. doi: 10.1006/rtph.2000.1448. [DOI] [PubMed] [Google Scholar]
  4. Colorado Department of Public Health and Environment. Glenwood springs prenatal report. 2014. [Google Scholar]
  5. Cooney MA, Buck Louis GM, Sundaram R, McGuiness BM, Lynch CD. Validity of self-reported time to pregnancy. Epidemiology. 2009;20:56–59. doi: 10.1097/EDE.0b013e31818ef47e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Dechanet C, Anahory T, Mathieu Daude JC, Quantin X, Reyftmann L, Hamamah S, et al. Effects of cigarette smoking on reproduction. Hum Reprod Update. 2011;17:76–95. doi: 10.1093/humupd/dmq033. [DOI] [PubMed] [Google Scholar]
  7. Detmar J, Rabaglino T, Taniuchi Y, Oh J, Acton BM, Benito A, et al. Embryonic loss due to exposure to polycyclic aromatic hydrocarbons is mediated by bax. Apoptosis. 2006;11:1413–1425. doi: 10.1007/s10495-006-8442-3. [DOI] [PubMed] [Google Scholar]
  8. Duong A, Steinmaus C, McHale CM, Vaughan CP, Zhang L. Reproductive and developmental toxicity of formaldehyde: A systematic review. Mutat Res. 2011;728:118–138. doi: 10.1016/j.mrrev.2011.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ferin M. Clinical review 105: Stress and the reproductive cycle. J Clin Endocrinol Metab. 1999;84:1768–1774. doi: 10.1210/jcem.84.6.5367. [DOI] [PubMed] [Google Scholar]
  10. Fujimori K, Kyozuka H, Yasuda S, Goto A, Yasumura S, Ota M, et al. Pregnancy and birth survey after the great east japan earthquake and fukushima daiichi nuclear power plant accident in fukushima prefecture. Fukushima J Med Sci. 2014;60:75–81. doi: 10.5387/fms.2014-9. [DOI] [PubMed] [Google Scholar]
  11. Goldstein BD, Osofsky HJ, Lichtveld MY. The gulf oil spill. N Engl J Med. 2011;364:1334–1348. doi: 10.1056/NEJMra1007197. [DOI] [PubMed] [Google Scholar]
  12. Hertz-Picciotto I. The evidence that lead increases the risk for spontaneous abortion. Am J Ind Med. 2000;38:300–309. doi: 10.1002/1097-0274(200009)38:3<300::aid-ajim9>3.0.co;2-c. [DOI] [PubMed] [Google Scholar]
  13. Hombach-Klonisch S, Pocar P, Kietz S, Klonisch T. Molecular actions of polyhalogenated arylhydrocarbons (pahs) in female reproduction. Curr Med Chem. 2005;12:599–616. doi: 10.2174/0929867310504050599. [DOI] [PubMed] [Google Scholar]
  14. Joffe M, Villard L, Li Z, Plowman R, Vessey M. A time to pregnancy questionnaire designed for long term recall: Validity in oxford, england. J Epidemiol Community Health. 1995;49:314–319. doi: 10.1136/jech.49.3.314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kristensen P, Irgens LM. Maternal reproductive history: A registry based comparison of previous pregnancy data derived from maternal recall and data obtained during the actual pregnancy. Acta Obstet Gynecol Scand. 2000;79:471–477. [PubMed] [Google Scholar]
  16. Kwak-Kim J, Bao S, Lee SK, Kim JW, Gilman-Sachs A. Immunological modes of pregnancy loss: Inflammation, immune effectors, and stress. Am J Reprod Immunol. 2014;72:129–140. doi: 10.1111/aji.12234. [DOI] [PubMed] [Google Scholar]
  17. Lane SM, Smith CR, Mitchell J, Balmer BC, Barry KP, McDonald T, et al. Reproductive outcome and survival of common bottlenose dolphins sampled in barataria bay, louisiana, USA, following the deepwater horizon oil spill. Proc Biol Sci. 2015;282:20151944. doi: 10.1098/rspb.2015.1944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lang K, Nuevo-Chiquero A. Trends in self-reported spontaneous abortions: 1970–2000. Demography. 2012;49:989–1009. doi: 10.1007/s13524-012-0113-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lynch CD, Sundaram R, Maisog JM, Sweeney AM, Buck Louis GM. Preconception stress increases the risk of infertility: Results from a couple-based prospective cohort study–the life study. Hum Reprod. 2014;29:1067–1075. doi: 10.1093/humrep/deu032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. National Institutes of Health Gulf study. Available https://gulfstudy.nih.gov/en/index.html [accessed September 1 2015]
  21. Neuberg M, Jakubowska-Szwed B, Neuberg J. reproductive behavior after the flood disaster in klodzko region–july 1997. Ginekol Pol. 2001;72:1037–1041. [PubMed] [Google Scholar]
  22. Neugebauer R, Kline J, Stein Z, Shrout P, Warburton D, Susser M. Association of stressful life events with chromosomally normal spontaneous abortion. Am J Epidemiol. 1996;143:588–596. doi: 10.1093/oxfordjournals.aje.a008789. [DOI] [PubMed] [Google Scholar]
  23. Osofsky HJ, Osofsky JD, Hansel TC. Deepwater horizon oil spill: Mental health effects on residents in heavily affected areas. Disaster medicine and public health preparedness. 2011;5:280–286. doi: 10.1001/dmp.2011.85. [DOI] [PubMed] [Google Scholar]
  24. Palinkas LA, Downs MA, Petterson JS, Russell J. Social, cultural, and psychological impacts of the exxon valdez oil spill. Human Organization. 1993;52:1–13. [Google Scholar]
  25. Picou JS, Marshall BK, Gill DA. Disaster, litigation, and the corrosive community. Social Forces. 2004;82:1493–1522. [Google Scholar]
  26. Pineles BL, Park E, Samet JM. Systematic review and meta-analysis of miscarriage and maternal exposure to tobacco smoke during pregnancy. Am J Epidemiol. 2014;179:807–823. doi: 10.1093/aje/kwt334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Pollino CA, Holdway DA. Reproductive potential of crimson-spotted rainbowfish (melanotaenia fluviatilis) following short-term exposure to bass strait crude oil and dispersed crude oil. Environ Toxicol. 2002;17:138–145. doi: 10.1002/tox.10043. [DOI] [PubMed] [Google Scholar]
  28. Rowe CL, Mitchelmore CL, Baker JE. Lack of biological effects of water accommodated fractions of chemically- and physically-dispersed oil on molecular, physiological, and behavioral traits of juvenile snapping turtles following embryonic exposure. Sci Total Environ. 2009;407:5344–5355. doi: 10.1016/j.scitotenv.2009.06.036. [DOI] [PubMed] [Google Scholar]
  29. Schlanger Z. In utah boom town, a spike in infant deaths raises questions. Newsweek. 2014 May 30; 2014. [Google Scholar]
  30. Schmidt L. Social and psychological consequences of infertility and assisted reproduction - what are the research priorities? Human fertility (Cambridge, England) 2009;12:14–20. doi: 10.1080/14647270802331487. [DOI] [PubMed] [Google Scholar]
  31. Van Scoy AR, Anderson BS, Philips BM, Voorhees J, McCann M, De Haro H, et al. Nmr-based characterization of the acute metabolic effects of weathered crude and dispersed oil in spawning topsmelt and their embryos. Ecotoxicol Environ Saf. 2012;78:99–109. doi: 10.1016/j.ecoenv.2011.11.009. [DOI] [PubMed] [Google Scholar]
  32. Whirledge S, Cidlowski JA. A role for glucocorticoids in stress-impaired reproduction: Beyond the hypothalamus and pituitary. Endocrinology. 2013;154:4450–4468. doi: 10.1210/en.2013-1652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wigle DT, Arbuckle TE, Turner MC, Berube A, Yang Q, Liu S, et al. Epidemiologic evidence of relationships between reproductive and child health outcomes and environmental chemical contaminants. J Toxicol Environ Health B Crit Rev. 2008;11:373–517. doi: 10.1080/10937400801921320. [DOI] [PubMed] [Google Scholar]
  34. Wilcox AJ, Horney LF. Accuracy of spontaneous abortion recall. Am J Epidemiol. 1984;120:727–733. doi: 10.1093/oxfordjournals.aje.a113940. [DOI] [PubMed] [Google Scholar]
  35. Wooten KJ, Finch BE, Smith PN. Embryotoxicity of corexit 9500 in mallard ducks (anas platyrhynchos) Ecotoxicology. 2012;21:662–666. doi: 10.1007/s10646-011-0822-y. [DOI] [PubMed] [Google Scholar]
  36. Wu J, Hou H, Ritz B, Chen Y. Exposure to polycyclic aromatic hydrocarbons and missed abortion in early pregnancy in a chinese population. Sci Total Environ. 2010;408:2312–2318. doi: 10.1016/j.scitotenv.2010.02.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Xu X, Cho SI, Sammel M, You L, Cui S, Huang Y, et al. Association of petrochemical exposure with spontaneous abortion. Occup Environ Med. 1998;55:31–36. doi: 10.1136/oem.55.1.31. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS906758-supplement.docx (38.9KB, docx)

RESOURCES