Infertility Risk and Employment History Among US Women: Findings From National Health and Nutrition Examination Survey 2013–2014

ABSTRACT

Introduction

It is well established that exposure to chemical, radiological, and biological hazards in the workplace are infertility risk factors. Although workplace‐specific infertility risks have been documented, the associations between employment history and infertility risk remain largely unexplored. The cross‐sectional, nationally representative US National Health and Nutrition Examination Survey (NHANES) provided an opportunity to explore the relationship between employment history and infertility risk.

Methods

Employment history and reproductive measures are available in the NHANES 2013–2014 cycle. Infertility was assessed using self‐reported, binary indicators: unsuccessful conception attempts and sought fertility consultation; combined to create a third, aggregate infertility measure. Associations between infertility risk and employment history, categorized by industry group and compared to respondents with no work history, were examined using three weighted binary logistic regression models while adjusting for sociodemographic, reproductive health, and relevant behavioral covariates.

Results

The sample consisted of 1126 women, aged 20 to 50 years; 1082 (96.1%) with a history of employment across 10 industry groups. Infertility was reported by 177 (17.2%), who reported unsuccessful conception attempts or fertility consultations. Across industry sectors, employment in the manufacturing of durable goods industry was significantly associated with unsuccessful conception attempts (adjusted odds ratio [AOR]: 4.15; 95% CI: 1.22, 14.08), fertility consultations (AOR: 6.94; 95% CI: 1.44, 25.25), and the aggregate infertility measure (AOR: 4.43; 95% CI: 1.32, 14.81).

Conclusion

Employment in durable goods manufacturing is associated with multiple self‐reported infertility measures, compared to women with no work history. In the absence of individualized workplace risk assessments, industry‐level analyses may better inform workplace health promotion, with emphasis on reproductive health protections.

1. Introduction

Over three‐quarters of American women between 25 and 54 years of age were employed in 2024 [1]. Most female workers are employed in management, professional, and related occupations; however, increasingly, women work in skilled trades, manufacturing, and other industrial sectors [2, 3]. These fields, along with agricultural, healthcare, and personal services, are the primary workplace settings for exposures to environmental hazards [4, 5]. Exposure to workplace hazards—chemical, ergonomic, physical, psychological—can increase the risk of infertility [6, 7, 8]. Occupational hazard classes pose significant risks to female‐factor fertility; these include radiation, heavy metals, teratogenic and carcinogenic chemicals such as pesticides, organic solvents, and antineoplastic agents, as well as ergonomic stressors. Infertility risk has been characterized among women employed in the plastic [9], shoe [10], and semiconductor [11, 12] manufacturing industries, as well as among woodworkers [13], cosmetologists and hairdressers [14, 15, 16], and those employed in the healthcare [17, 18, 19, 20], printing [21], and agriculture [22, 23, 24] industries.

Infertility results from reproductive tract disease, injury or other impairments that prevent conception or successful pregnancy despite regular sexual intercourse [25]. The etiology of these impairments may be male‐factor, female‐factor, combined, or unexplained. Although occupational risks to male fertility are well‐established [7, 26, 27, 28, 29], the continued engagement of women in the workforce requires consideration of the effects of occupational hazard exposures on female‐factor infertility. Characterization of occupational reproductive risks among men are not directly applicable to women due to differences in assigned work responsibilities, biological mechanisms of exposure, and the quality of acquired personal protective equipment (PPE) [30, 31].

Despite the existing body of evidence linking occupational and environmental hazards to female‐factor fertility, infertility risk characterization of industry sectors reflects the growing engagement of women in these fields, and supports workplace risk assessments. The National Health and Nutrition Examination Survey (NHANES) is a nationally representative cross‐sectional survey on noninstitutionalized civilians in the United States (US) [32]. Therefore, this study aims to characterize employment history, classified by industry sector, and infertility prevalence using NHANES.

2. Materials and Methods

2.1. Data Source

The data source was the 2013–2014 NHANES—the only survey cycle that included both infertility indicators and coded industry groups [32]. The 2013–2014 cycle consisted of an initial sample of 10,175 questionnaire respondents. We excluded participants' data if they were male (n = 5003), less than 20 years or greater than 50 years of age (n = 3558), members of the Armed Forces, or missing infertility, industry, and covariate responses (n = 488), which resulted in a final sample of 1126 participants (Figure 1). We recognize that individuals of diverse genders with female reproductive organs face occupational health risks. In this analysis, as the gender variable in the NHANES 2013–2014 dataset was binary (male/female), we refer to participants as “female” or “women.” NHANES 2013–2014 was approved by the National Center for Health Statistics (NCHS) Ethics Review Board (Continuation of Protocol #2011‐17) with written informed consent obtained from all participants [33]. The age range of 20–50 years was selected to capture occupational histories during the reproductive lifespan, ensuring that participants had sufficient opportunity for purposive conception attempts while minimizing the risk that their recollections reflect age‐related subfertility or perimenopausal changes rather than infertility‐related experiences.

Figure 1.

Flowchart of the participant selection process: NHANES 2013–2014.

2.2. Classification of Employment History

The NHANES 2013–2014 occupation questionnaire incorporated 22 coded industry groups that were classified according to the 2010 US Bureau of the Census Industrial & Occupational (I&O) Classification coding system, which relate broadly to the North American Industry Classification System used in the United States, Canada, and Mexico [34, 35]. Respondents' participation in industry sectors was identified from responses to the employment history question OCD390G: “Thinking of all the paid jobs or businesses you ever had, what kind of work were you doing the longest?”, including instances where the longest‐held job corresponded to respondents' current occupation, with responses collected earlier in the survey (OCD231: “What kind of business or industry is this?”) [34]. Employment history was represented using the original NHANES/2010 Census I&O codes, with select small‐sample categories merged to ensure adequate sample sizes for analysis, based on conceptual similarity and operational characteristics (Figure 2). Respondents with no history of employment were designated as the reference population for statistical analyses.

Figure 2.

Classification of participants' employment history. Response options to OCD390G: “Thinking of all the paid jobs or businesses you ever had, what kind of work were you doing the longest?” included: (1) Enter occupation (OCD31‐Industry group code: longest job); (2) Same as current occupation (asked earlier in the survey as OCD231‐Industry group code: current job); (3) Armed Forces; (4) Never worked. Members of the Armed Forces were excluded (see Figure 1). Employment history was represented by the original NHANES/2010 US Bureau of the Census Industrial & Occupational Classification groups, with select small‐sample groups merged for statistical analyses. Respondents with no history of employment (never worked) designated as the reference population for statistical analyses.

2.3. Outcome Variables

Infertility measures were restricted to respondents' experiences, with no available data related to timing, duration, or partner's experiences with infertility. Three measures were used to assess infertility. These included unsuccessful conception (≥ 12 months) (RHQ074: “Have you ever attempted to become pregnant over a period of at least a year without becoming pregnant?” Response = Yes), and fertility consultation (RHQ076: “Have you ever been to a doctor or other medical provider because you have been unable to become pregnant?” Response = Yes) [36]. A third composite infertility measure was derived by aggregating the unsuccessful conception (≥ 12 months) and fertility consultation responses [37].

2.4. Covariates

Covariates used in the analyses were selected a priori based on findings from previous studies [38, 39, 40, 41, 42, 43] and consisted of demographic and related‐risk factors from the NHANES 2013–2014 questionnaire [36, 44, 45, 46, 47, 48, 49]. Demographic information included age (RIDAGEYR) at the time of data collection, ethnicity (RIDRETH1), educational attainment (DMDEDUC2), and poverty‐to‐income ratio (PIR, INDFMPIR) [44]. PIR, an income measure related to the US poverty threshold, is used in this paper as an indicator of socioeconomic status, as it is not impacted by family size or inflation, unlike household income [43]. Related risk‐factors in the analyses included pelvic inflammatory disease history (PID, RHQ078) [36], history of sexually transmitted infection (STI) diagnosis [48], body mass index (BMI, BMXBMI) [45], smoking status [49], alcohol consumption status [46], and work‐related functional impairment (WFI, PFQ051) [47]. Age, PIR, and BMI were continuous variables. The history of PID, STI diagnosis, and WFI were modeled as dichotomous (yes/no). Ethnicity (Mexican American, Other Hispanic, non‐Hispanic White, non‐Hispanic Black, and other race including multiracial) and educational attainment (less than high school graduate, high school graduate, some college or associate's degree, and college graduate or above) were modeled as categorical. History of STI diagnosis was obtained by merging responses to SXQ270: “Has the doctor ever told you that you had gonorrhea?” and SXQ272: “Has the doctor ever told you that you had chlamydia?” [38]. Parity (nulliparous and parity > 1) was classified based on responses to RHQ131: “Have you ever been pregnant?” and RHQ171: “How many of your deliveries resulted in a live birth?” [36, 50]. Smoking status was classified as never smoker, former smoker, current smoker based on responses to SMQ020: “Have you smoked at least 100 cigarettes in your entire life?” and SMQ040: “Do you now smoke cigarettes?” [49, 51]. Alcohol consumption status was classified as never drinker, former drinker, current drinker based on responses to the following measures: ALQ101: “In any 1 year, have you had at least 12 drinks of any type of alcoholic beverage?” (categorized as current drinkers) and ALQ110: “In your entire life, have you had at least 12 drinks of any type of alcoholic beverage?” (if yes to ALQ110 but no to ALQ101, categorized as former drinkers) [46, 51].

2.5. Statistical Analyses

All analyses were performed on R (RStudio Integrated Development Environment 2020. Open Source Edition. Version 4.3.3) [52]. Survey weights were applied to account for NHANES sampling designs and to ensure the generalizability of findings among noninstitutionalized US civilians [53]. Preliminary descriptive analyses were conducted to identify participant characteristics stratified by employment history. All three infertility measures were subsequently cross‐tabulated with employment history and all covariates. Categorical variables were described using weighted proportions and unweighted counts and compared using Pearson chi‐square tests with the Rao–Scott correction [54]. Continuous variables were summarized using weighted means and standard deviations, and group differences assessed by design‐adjusted independent samples t‐tests. Following these bivariate analyses, three weighted binary logistic regression models were conducted to assess the association between each infertility measure and employment history. Respondents with no history of employment were selected as the reference group to reflect baseline exposures to reproductive hazards independent of occupational experiences. This group enabled the evaluation of female‐factor fertility‐risks across industry sectors relative to respondents unexposed to chemical, biological or physical hazards, including ergonomic strain and shiftwork in the workplace.

Variance inflation factors (VIF) for all variables ranged from 1.0 to 1.3 indicating no multicollinearity (VIF < 5). Logistic models accounted for demographic variables and established related‐risk factors. Following the main analyses, sensitivity analyses were performed by generating five imputed datasets using multiple imputation by chained equations (MICE) to address missing reproductive‐health‐related covariates (history of PID or STI diagnosis, parity), adding 135 respondents previously excluded. The models used to conduct sensitivity analyses were adjusted with the same predictors used in the main analysis. These sensitivity analyses incorporated results that were pooled using Rubin's rules [55]. For all analyses, p < 0.05 was considered significant.

3. Results

3.1. Participant Demographics Stratified by Employment History

A total unweighted sample of 1126 female participants were included in the study: 1082 with a history of employment and 44 with no history of employment (Table 1, Supporting Information: Table S1). Survey‐weighted percentages are used throughout, unless otherwise stated. Participants with employment histories were mostly nonracialized (non‐Hispanic white: 62.2% vs. 35.4%; p = 0.004) and had higher educational attainment (college or associate's degree: 37.5% vs. 19.1%; p < 0.001), compared to participants with no employment history. There was a modest difference in BMI (p = 0.047) between groups, however those with a history of employment were more likely to be current alcohol consumers (75.4%) compared to the reference group (42.2%; p < 0.001).

Table 1.

Participant demographics stratified by employment history.

Variable	No history of employment (n = 44)	History of employment (n = 1082)	p value
Employment historya
No history of employment	44 (100%)	—	—
Manual	—	40 (3.5%)
Manufacturing durable goods	—	41 (3.1%)
Manufacturing nondurable goods	—	38 (3.1%)
Trade	—	156 (13.8%)
Professional services	—	180 (18.1%)
Educational services	—	96 (9.4%)
Healthcare and social assistance	—	270 (24.6%)
Leisure and hospitality	—	172 (15.4%)
“Other” services	—	55 (5.8%)
Public administration	—	34 (3.2%)
WFIa			0.70
Yes	5 (15.3%)	143 (12.7%)
No	39 (84.7%)	939 (87.3%)
Age (years), mean ± SDb	34.6 ± 9.5	35.4 ± 9.2	0.74
PIR, mean ± SDb	2.1 ± 1.9	2.8 ± 1.7	0.25
Ethnicitya			0.004
Mexican American	14 (27.6%)	156 (10.6%)
Other Hispanic	7 (14.6%)	104 (7.0%)
Non‐Hispanic White	8 (35.4%)	461 (62.2%)
Non‐Hispanic Black	3 (5.0%)	208 (12.4%)
Other race	12 (17.6%)	153 (7.8%)
Educational attainmenta			< 0.001
Less than high school	19 (38.4%)	166 (12.0%)
High School graduate	5 (8.2%)	206 (18.4%)
Some college or associate's degree	13 (34.2%)	398 (37.5%)
College graduate or above	7 (19.1%)	312 (32.2%)
BMI (kg/m2), mean ± SDb	31.1 ± 7.4	29.4 ± 8.1	0.05
Smoking statusa			0.13
Never smoker	37 (84.9%)	721 (64.5%)
Former smoker	1 (5.6%)	132 (13.6%)
Current smoker	6 (9.5%)	229 (21.9%)
Alcohol consumption statusa			< 0.001
Never drinker	19 (34.0%)	157 (12.3%)
Former drinker	8 (23.8%)	168 (12.3%)
Current drinker	17 (42.2%)	757 (75.4%)
STI Historya			0.62
Yes	0 (0%)	22 (2.1%)
No history	44 (100%)	1060 (97.9%)
PID Historya			0.29
Yes	0 (0%)	56 (5.0%)
No history	44 (100%)	1026 (95.0%)
Paritya			0.15
Parity > 1	39 (83.9%)	824 (73.7%)
Nulliparous	5 (16.1%)	258 (26.3%)

Abbreviations: BMI, body mass index; PID, pelvic inflammatory disease; PIR, poverty income ratio, SD, standard deviation; STI, sexually transmitted infection; WFI, work‐related functional impairment.

^a Categorical variables expressed as unweighted counts (n) with weighted proportions (%). Statistical comparisons conducted by the Pearson chi‐square test with the Rao‐Scott correction.

^b Continuous variables expressed as weighted mean ± weighted SD. Statistical analyses conducted by weighted independent samples t‐test.

3.2. Characteristics of Participants Across Infertility Measures

Across the sample, 156 (15.3%) reported unsuccessful conception (≥ 12 months), and 110 (11.2%) reported fertility consultation (Table 2). A greater proportion of employed participants (174, 97.6%) were categorized under the aggregate infertility measure, defined as reporting either or both experiences, compared to participants with no employment history (3, 2.4%; p = 0.02). Similarly, chi‐square analysis demonstrated that employment history across all sectors compared with participants with no employment history was significantly associated with unsuccessful conception attempts ≥ 12 months (p = 0.01) and fertility medical consultation (p = 0.001; Table 2, Supporting Information: Table S2). Among covariates, participants who responded “yes” to any of the infertility measures had significantly higher mean age (p ≤ 0.001) and mean BMI (p ≤ 0.002). Similarly, experience of WFI (p ≤ 0.004) and history of PID (p ≤ 0.02) were also associated with all infertility measures. PIR (p < 0.001) and educational attainment (p = 0.01), both measures of socioeconomic status, were significantly higher among respondents who reported fertility consultations. Finally, current smokers were less likely to seek fertility consultation (10.4%) compared with never smokers (70.8%) or former smokers (18.8%; p = 0.002).

Table 2.

Characteristics of participants across infertility measures.

Variable	Unsuccessful conception attempt (≥ 12 months)			Fertility medical consultation			Aggregate infertility measurea
	Yes	No	p value	Yes	No	p value	Yes	No	p value
Unweighted n	156 (15.3%)	970 (84.7%)		110 (11.2%)	1016 (88.8%)		177 (17.2%)	949 (82.8%)
Employment histb
No history of employment	3 (2.7%)	41 (3.2%)	0.01	2 (3.2%)	42 (3.1%)	0.001	3 (2.4%)	41 (3.3%)	0.02
Manual	4 (1.5%)	36 (3.7%)		2 (1.0%)	38 (3.7%)		4 (1.3%)	36 (3.8%)
Mfg: Durable goods	13 (6.2%)	28 (2.5%)		12 (8.1%)	29 (2.4%)		13 (5.5%)	28 (2.5%)
Mfg: Nondurable goods	3 (1.2%)	35 (3.3%)		4 (3.0%)	34 (3.0%)		4 (1.9%)	34 (3.2%)
Healthcare/social assist.	31 (19.6%)	239 (24.6%)		17 (14.8%)	253 (25.0%)		36 (19.1%)	234 (24.8%)
“Other” services	11 (8.2%)	44 (5.2%)		8 (8.9%)	47 (5.2%)		12 (8.2%)	43 (5.1%)
Professional services	33 (19.9%)	147 (17.1%)		24 (21.4%)	156 (17.1%)		39 (22.0%)	141 (16.6%)
Leisure and hospitality	21 (13.2%)	151 (15.3%)		13 (11.0%)	159 (15.4%)		23 (12.4%)	149 (15.5%)
Educational services	15 (11.0%)	81 (8.8%)		14 (13.5%)	82 (8.6%)		17 (10.5%)	79 (8.8%)
Public administration	7 (6.6%)	27 (2.5%)		7 (7.8%)	27 (2.5%)		8 (6.1%)	26 (2.5%)
Trade	15 (9.9%)	141 (14.0%)		7 (7.5%)	149 (14.1%)		18 (10.6%)	138 (14.0%)
WFIb			< 0.001			0.004			< 0.001
No	118 (74.7%)	860 (89.5%)		83 (76.6%)	895 (88.5%)		136 (76.2%)	842 (89.5%)
Yes	38 (25.3%)	110 (10.5%)		27 (23.4%)	121 (11.5%)		41 (23.8%)	107 (10.5%)
Age, mean ± SDc	39.1 ± 8.5	34.7 ± 9.2	0.001	39.7 ± 7.7	34.8 ± 9.2	< 0.001	38.8 ± 8.6	34.7 ± 9.2	< 0.001
Ethnicityb
Non‐Hispanic white	74 (68.1%)	395 (60.1%)		55 (70.5%)	414 (60.2%)		83 (67.8%)	386 (60.0%)
Mexican American	16 (7.4%)	154 (11.8%)		10 (6.7%)	160 (11.7%)		18 (7.5%)	152 (11.9%)
Other Hispanic	14 (5.6%)	97 (7.5%)		11 (6.0%)	100 (7.4%)		15 (5.2%)	96 (7.7%)
Non‐Hispanic Black	35 (12.9%)	176 (12.0%)		15 (7.8%)	196 (12.7%)		38 (12.6%)	173 (12.1%)
Other Race	17 (6.0%)	148 (8.5%)		19 (9.2%)	146 (8.0%)		23 (6.9%)	142 (8.4%)
Educationb						0.01
Less than high school	20 (9.9%)	165 (13.4%)		8 (3.7%)	177 (14.0%)		21 (9.0%)	164 (13.6%)
High school graduate	27 (14.9%)	184 (18.6%)		17 (13.6%)	194 (18.6%)		29 (14.8%)	182 (18.7%)
Some college/assoc.deg.	61 (42.3%)	350 (36.5%)		41 (42.2%)	370 (36.8%)		69 (42.4%)	342 (36.4%)
≥ College graduate	48 (32.9%)	271 (31.5%)		44 (40.4%)	275 (30.7%)		58 (33.9%)	261 (31.3%)
PIR, mean ± SDc	3.0 ± 1.6	2.7 ± 1.7		3.4 ± 1.5	2.7 ± 1.7	< 0.001	3.1 ± 1.6	2.7 ± 1.7	0.02
Smoking historyb						0.002
Never smoker	108 (70.0%)	650 (64.2%)		75 (70.8%)	683 (64.4%)		121 (69.8%)	637 (64.1%)
Former smoker	22 (13.7%)	111 (13.3%)		23 (18.8%)	110 (12.7%)		29 (15.2%)	104 (13.0%)
Current smoker	26 (16.3%)	209 (22.5%)		12 (10.4%)	222 (23.0%)		27 (15.0%)	208 (22.9%)
Alcohol useb
Never drinker	22 (11.4%)	154 (13.3%)		17 (13.4%)	159 (12.9%)		24 (11.4%)	152 (13.3%)
Former drinker	20 (11.5%)	156 (12.9%)		18 (14.1%)	158 (12.5%)		24 (11.8%)	152 (12.9%)
Current drinker	114 (77.1%)	660 (73.9%)		75 (72.5%)	699 (74.6%)		129 (76.9%)	645 (73.8%)
BMI, mean ± SDc	32.7 ± 9.5	28.8 ± 7.7	0.002	31.8 ± 9.3	29.1 ± 7.9	0.002	32.7 ± 9.6	28.7 ± 7.6	0.001
STI historyb
No history	151 (96.6%)	953 (98.2%)		106 (96.7%)	998 (98.1%)		170 (96.0%)	934 (98.4%)
Yes	5 (3.4%)	17 (1.8%)		4 (3.4%)	18 (1.9%)		7 (4.0%)	15 (1.6%)
PID historyb			0.02			0.01			0.003
No history	143 (90.3%)	927 (96.0%)		100 (90.4%)	970 (95.7%)		162 (90.1%)	908 (96.2%)
Yes	13 (9.8%)	43 (4.0%)		10 (9.6%)	46 (4.3%)		15 (9.9%)	41 (3.8%)
Parityb			0.02
Nulliparous	29 (17.3%)	234 (27.6%)		21 (20.9%)	242 (26.7%)		35 (19.0%)	228 (27.5%)
Parity > 1	127 (82.8%)	736 (72.4%)		89 (79.1%)	774 (73.3%)		142 (81.1%)	721 (72.5%)

Abbreviations: Assist., assistance; Assoc, associate; BMI, body mass index; Deg., degree; Hist, history; Mfg, manufacturing; PID, pelvic inflammatory disease; PIR, poverty income ratio; SD, standard deviation; STI, sexually transmitted infection; WFI, work‐related functional impairment.

^a Aggregate infertility measure—Unsuccessful conception attempts (at least 12 months) and/or fertility medical consultation.

^b Categorical variables expressed as unweighted counts (n) with weighted proportions (%). Statistical comparisons conducted by the Pearson chi‐square test with the Rao‐Scott correction.

^c Continuous variables expressed as weighted mean ± weighted SD. Statistical analyses conducted by weighted independent samples t‐test.

3.3. Associations Between Employment History and Infertility Measures

Three independent logistic regression models were used to analyze the associations between employment history and our three infertility measures. Initial logistic regression models (crude odds ratios) of employment history (reference: no history of employment) and our three infertility outcome measures were not significant (Table 3). Regression models were subsequently adjusted for demographic variables and related risk factors (Table 4). Relative to those with no history of employment, employment in the manufacturing of durable goods industry was strongly associated with all three infertility measures‐unsuccessful conception (AOR: 4.15; 95% CI: 1.22–14.08; p = 0.026), fertility consultation (AOR: 6.04; 95% CI: 1.44–25.25; p = 0.017), and the aggregate infertility measure (AOR: 4.43; 95% CI: 1.32–14.81; p = 0.019). We found no association between any other industry group and the reference population across all three infertility measures. Sensitivity analyses supported these results (Supporting Information: Tables S3–S5).

Table 3.

Unadjusted model‐association between employment history and infertility.

Variable	Unsuccessful conception attempt (≥ 12 months)b	Fertility medical consultationb	Aggregate infertility measurea,b
	COR (95% CI)	COR (95% CI)	COR (95% CI)
Employment history
No history of employment (Reference category)
Manual	0.49 (0.08, 3.02)	0.27 (0.04, 2.09)	0.49 (0.08, 3.02)
Manufacturing: Durable goods	3.01 (0.46, 19.67)	3.35 (0.40, 27.71)	3.01 (0.46, 19.67)
Manufacturing: Nondurable goods	0.44 (0.06, 3.08)	0.99 (0.11, 8.65)	0.83 (0.11, 6.34)
Trade	0.84 (0.26, 2.75)	0.53 (0.12, 2.34)	1.04 (0.37, 2.93)
Professional services	1.39 (0.45, 4.31)	1.25 (0.23, 6.82)	1.82 (0.50, 6.54)
Educational services	1.50 (0.35, 6.34)	1.56 (0.30, 8.04)	1.63 (0.39, 6.88)
Healthcare and social	0.95 (0.30, 3.06)	0.59 (0.16, 2.24)	1.06 (0.34, 3.31)
Leisure and hospitality	1.04 (0.25, 4.32)	0.71 (0.16, 3.19)	1.10 (0.26, 4.57)
“Other” services	1.90 (0.34, 10.72)	1.69 (0.23, 12.26)	2.22 (0.37, 13.28)
Public administration	3.16 (0.71, 14.05)	3.09 (0.48, 20.02)	3.40 (0.75, 15.36)

Abbreviations: CI, confidence interval; COR, crude odds ratio.

^a Aggregate infertility measure—Unsuccessful conception attempts (at least 12 months) and/or fertility medical consultation.

^b Reference category (No).

Table 4.

Adjusted model‐association between employment history and infertility.

Variable	Unsuccessful conception attempt (≥ 12 months)b	Fertility medical monsultationb	Aggregate infertility measure a,b
	AOR (95% CI)	AOR (95% CI)	AOR (95% CI)
Employment history
No history of employment (Reference)
Manual	0.40 (0.07, 2.37)	0.29 (0.04, 1.96)	0.41 (0.07, 2.35)
Manufacturing: Durable goods	4.15 (1.22, 14.08)*	6.04 (1.44, 25.25)*	4.43 (1.32, 14.81)*
Manufacturing: Nondurable goods	0.66 (0.11, 4.00)	1.88 (0.25, 14.19)	1.34 (0.19, 9.75)
Healthcare and social assistance	0.82 (0.29, 2.32)	0.59 (0.20, 1.77)	0.94 (0.36, 2.44)
“Other” services	1.69 (0.40, 7.07)	1.68 (0.34, 8.32)	2.02 (0.46, 8.97)
Professional services	1.26 (0.52, 3.03)	1.05 (0.31, 3.60)	1.70 (0.69, 4.19)
Leisure and hospitality	1.63 (0.54, 4.94)	1.43 (0.41, 5.04)	1.82 (0.57, 5.76)
Educational services	1.38 (0.43, 4.36)	1.56 (0.46, 5.31)	1.57 (0.54, 4.55)
Public administration	2.77 (0.75, 10.25)	2.93 (0.40, 21.19)	3.14 (0.83, 11.80)
Trade	0.93 (0.32, 2.70)	0.62 (0.17, 2.24)	1.19 (0.47, 2.96)
WFI
No (Reference)
Yes	2.20 (1.31, 3.70)**	2.14 (1.25, 3.66)**	2.04 (1.17, 3.56)*
Age	1.04 (1.00, 1.08)	1.04 (1.01, 1.07)**	1.04 (1.00, 1.07)
Ethnicity
Non‐Hispanic White (Reference)
Mexican American	0.54 (0.21, 1.39)	0.59 (0.13, 2.66)	0.53 (0.18, 1.53)
Other Hispanic	0.89 (0.51, 1.55)	1.01 (0.53, 1.93)	0.79 (0.43, 1.43)
Non‐Hispanic Black	0.88 (0.41, 1.89)	0.50 (0.15, 1.63)	0.85 (0.40, 1.81)
Other Race	0.65 (0.27, 1.59)	0.92 (0.33, 2.51)	0.77 (0.31, 1.91)
Educational attainment
High School Graduate (Reference)
Less than high school	1.34 (0.63, 2.88)	0.60 (0.19, 1.94)	1.34 (0.72, 2.49)
Some college or AA	1.60 (0.87, 2.94)	1.73 (0.74, 4.09)	1.64 (1.04, 2.58)*
College graduate or above	1.25 (0.72, 2.17)	1.40 (0.54, 3.64)	1.26 (0.82, 1.95)
PIR	1.18 (1.04, 1.33)*	1.34 (1.08, 1.67)*	1.20 (1.05, 1.37)*
Smoking history
Never smoker (Reference)
Current smoker	0.45 (0.24, 0.84)*	0.37 (0.24, 0.60)***	0.42 (0.24, 0.73)**
Former smoker	0.60 (0.24, 1.47)	1.01 (0.50, 2.01)	0.69 (0.31, 1.54)
Alcohol consumption history
Never drinker (Reference)
Former drinker	0.89 (0.45, 1.75)	0.93 (0.37, 2.36)	0.91 (0.41, 2.02)
Current drinker	1.17 (0.62, 2.22)	0.70 (0.27, 1.85)	1.11 (0.50, 2.46)
BMI	1.06 (1.03, 1.09)***	1.05 (1.03, 1.07)***	1.07 (1.04, 1.09)***
STI History
No history (Reference)
Yes	4.85 (0.80, 29.39)	6.50 (1.48, 28.60)*	6.19 (0.89, 43.31)
PID History
No history (Reference)
Yes	2.26 (0.82, 6.28)	2.45 (1.13, 5.30)*	2.50 (1.11, 5.66)*
Parity
Parity > 1 (Reference)
Nulliparous	0.59 (0.31, 1.12)	0.76 (0.30, 1.95)	0.65 (0.31, 1.37)

Abbreviations: AA, associate‐level degree; AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; PID, pelvic inflammatory disease; PIR, poverty income ratio; STI, sexually transmitted infection; WFI, work‐related functional impairment.

^a Aggregate infertility measure—Unsuccessful conception attempts (at least 12 months) and/or fertility medical consultation.

^b Reference category (No).

^* p < 0.05;

^** p < 0.01;

^*** p < 0.001.

Consistent with preliminary analyses, WFI was associated with unsuccessful conception (AOR: 2.20; 95% CI: 1.31–3.70; p = 0.005), fertility consultation (AOR: 2.14; 95% CI: 1.25–3.66; p = 0.009), and the aggregate infertility measure (AOR: 2.04; 95% CI: 1.17–3.56; p = 0.016). BMI (AOR range 1.05–1.07; p < 0.001) and PIR (AOR range 1.18–1.34; p < 0.05) were significantly associated with all three infertility measures, whereas age was positively associated only with fertility consultation (AOR: 1.04; 95% CI: 1.01–1.07; p = 0.009). Recognized reproductive risk factors such as PID (AOR: 2.45; 95% CI: 1.13–5.30; p = 0.026) and STI (AOR: 6.50; 95% CI: 1.48–28.60; p = 0.017) history were each highly associated with fertility consultation, as expected. Relative to participants who never smoked, current smokers had lower odds of experiencing unsuccessful conception (AOR: 0.45; 95% CI: 0.24–0.84; p = 0.016), fertility consultation (AOR: 0.37; 95% CI: 0.24–0.60; p < 0.001), and infertility (AOR: 0.42; 95% CI: 0.24–0.73; p = 0.005).

4. Discussion

Women now constitute about 29% of the American manufacturing sector, with labor participation still in recovery due to the Great Recession in 2008 and the COVID‐19 pandemic [56, 57]. The manufacturing sector is steadily becoming more attractive as a career option for girls and women, with diversification of manufacturing jobs requiring STEM training, reflecting the industry's technological modernization, including greater use of automation [57]. Our analysis of employment history and infertility risk using NHANES 2013–2014 demonstrated a significant association between employment in durable goods manufacturing and infertility measures among female workers aged 20–50 years, after adjusting for demographic and related‐risk factors.

4.1. Manufacturing and Infertility

Durable goods manufacturing involves the use of machines to chemically, physically, and mechanically transform materials into products [58]. This industry sector includes the manufacture of wood, nonmetallic minerals, primary metal, fabricated metals, machinery, computer and electronic products, electrical equipment, appliances, transportation equipment, furniture, and medical equipment [58]. Although several occupational studies report increased infertility risk in female Taiwanese semiconductor manufacturing workers [11], Portuguese shoe manufacturers [10], and in Finnish wood processors [13], we found no comparable NHANES studies.

Durable goods manufacturing was significantly associated with our infertility measures only when adjusted for respondent characteristics and risk factors, and not in the unadjusted (crude) logistic regression model. Compared to respondents with no employment history, durable goods manufacturing workers tended to be older, with lower BMI and slightly more experiences of WFIs and PID history. Sociodemographic and reproductive risk factors may be exacerbated by occupational factors and contribute to adverse health. Indeed, several NHANES studies report combined effects of age or BMI with toxicant exposures. Age, BMI, and lead exposure are intersecting risks to infertility, reported in a study of NHANES respondents [59]. Similarly, BMI interacts with volatile organic compounds (VOC) [60] or lead [61] to increase infertility risk among overweight female NHANES participants. In contrast, exposures to mixed metals and in particular, cadmium, increased infertility risk among nonoverweight NHANES women [61]. Advanced age is a well‐established risk factor for female infertility, with several NHANES studies identifying older women as a subpopulation who are more likely to experience adverse reproductive health effects, such as infertility and uterine fibroids, from hazard exposures such as mixed metals [61] or mercury [62]. These risk relationships suggest that within the heterogeneous population of workers, the interactions of individual characteristics such as age and BMI contribute to infertility risk.

4.2. Toxicant Body Burdens and Infertility

Infertility risk may be related to occupational hazards, with reports of demonstrated exposures to several reproductive toxicants including organic solvents [4, 63, 64], heavy metals [5, 61, 65], plastics [9], and endocrine‐disrupting chemicals [66] among female manufacturing workers. Adverse reproductive health effects associated with these occupational hazards may be caused by endocrine disruption, inflammation, and mechanisms of cellular toxicity such as oxidative stress, epigenetic modifications or genotoxicity [7]. Although we did not evaluate NHANES respondent body burdens, there are several studies which support our investigation. A large analysis of multiple NHANES cycles concluded that blue‐collar occupations, such as durable goods manufacturing, are associated with higher worker body burdens of a range of occupational hazards, although the findings were not disaggregated by sex [67]. Durable goods manufacturing body burdens of heavy metals such as cadmium and lead, VOCs like benzene and xylene, phthalates including Di(2‐ethylhexyl) phthalate (DEHP) metabolites, acrylamide, and other toxicants were higher compared to white‐collar public administration workers [67]. Another NHANES study also reported that blue‐collar workers, including those in manufacturing occupations, exhibited higher blood cadmium levels compared with other occupations, with levels higher in women compared with men [68]. Many of these occupational hazards associated with manufacturing are also reproductive toxicants. Lead [69, 70], cadmium [59, 70], arsenic [59], heavy metal mixtures [61, 69], and volatile organic compounds [60, 71] have all been associated with self‐reported infertility in female NHANES respondents, although none of these studies explored potential associations with occupation.

Given that workers in durable goods manufacturing have high body burdens of phthalates [67], plasticizers added to products to increase flexibility and durability, microplastic accumulation is likely. Gendered differences in clothing, facial hair, and use of cosmetics were associated with microplastic exposure in male and female workers in plastic products manufacturing [72]. Emerging evidence suggests that microplastics are persistent in the environment with ubiquitous exposure and have been detected in most human tissues [73, 74, 75]. As with many occupational hazards, microplastics increase oxidative stress, immune dysregulation, apoptosis, and disrupt endocrine signaling [73, 74]. The role of microplastics exposure and infertility is still unclear; however, detection of microplastics in the placenta, testes, semen, follicular fluid, and breast milk [73, 74, 75] and a growing body of evidence from animal and in vitro studies suggest a role in both male‐ and female‐factor infertility [75].

4.3. Occupational Hazard Risk Mitigation

The contribution of the occupational environment to infertility is challenging to investigate. Occupational classifications, including the strategy used here, are notably heterogeneous. Although prevalence of adverse reproductive health outcomes is often observed among women employed in healthcare [14, 20, 64], we found no similar associations with infertility, possibly due to the 2010 Census I&O coding, which merges healthcare with social assistance workers. Even though our sample included two manufacturing sectors, durable and nondurable goods, the latter was not associated with our infertility measures, despite similar occupational exposures [67]. Given the lack of employment history, lifestyle and biological information for respondents' spouses or partners, we considered demographic and reproductive characteristics to help contextualize differences between these industry sectors. Descriptively, respondents with a history of employment in nondurable goods manufacturing were younger, more racialized, nulliparous, and had lower WFI compared to those from durable goods manufacturing.

Even with more specific industry classifications, adherence to local health and safety legislation, utilization of the hierarchy of controls to reduce hazard exposures, and employer and worker compliance with safety protocols, provisions to ensure safe conception and pregnancy will vary across workplaces. The US Occupational Safety and Health Administration (OSHA) [76] grants American workers rights to training, information about workplace hazards, and education related to the importance and appropriate use of PPE. Together with state legislation and workplace policies, OSHA also provides protections for workers who may experience impairments to their health or functional capacity due to regular exposure to workplace hazards, including reproductive toxicants [77].

Globally, US workplace health and safety protections for pregnancy align with the United Nations International Labor Organization [78], however less robust protections are evident for fertility. Although differentiated protection policies [79, 80] are widely used to address risks specific to pregnant workers, these policies rely on workers to disclose their reproductive status to employers, with fertility challenges less likely to be discussed in the workplace. The US Pregnant Workers Fairness Act (PWFA) [81, 82]; provides for workplace accommodations for pregnancy, childbirth and related “medical conditions”—explicitly defined to include infertility and fertility treatments. As differentiated protections are often associated with workplace discrimination and barriers to career advancement [79, 80], a comprehensive strategy that ensures protections for both reproductive health and fetal development would ensure equal opportunity. Implementation of a universal, inclusive framework for workplace protections would reduce the need for personal, medical disclosures by recognizing that all workers are biologically vulnerable to reproductive and developmental hazards.

4.4. Strengths and Limitations

To our knowledge, this is the first study to explore the association between employment history and infertility risk using NHANES. Our findings suggest that female workers employed in the durable goods manufacturing sector may have an increased risk of infertility. However, there are several limitations worth noting. We used industry sectors here as proxies for exposures to reproductive hazards; however, actual body burden measurements of exposure are preferable. Industry sector classification was based on longest job worked; however, NHANES did not collect a full employment history, resulting in data gaps for our sample regarding the total number of different industries worked, periods of unemployment, and the timing of employment relative to infertility experiences. A causal association between employment history and infertility risk cannot be established due to the cross‐sectional nature of NHANES, and the lack of temporal information to determine whether employment preceded infertility experiences. The 2010 Census I&O codes, combined with our merged industry groups, prevent the identification of specific industries that may contribute to increased infertility risk due to heterogeneous classification schemes and the varied exposures within each sector. Several industry groups remained relatively small despite category merging. However, sensitivity analyses using MICE produced results consistent with the main analyses, indicating that limited sample size did not substantially influence our findings.

Our analyses did not show an association between infertility outcomes and several high‐hazard exposure industry sectors, such as healthcare (merged with social assistance in NHANES/2010 Census I&O code) and nondurable goods manufacturing. This may be explained by difference in respondent's demographics, lifestyle, and reproductive characteristics across sectors. Nevertheless, we lacked data on respondents' parenting intentions, age at the time of infertility, and information about their spouses or partners–factors that are highly relevant in studies of infertility. Similarly, the NHANES reproductive health questionnaire did not capture information related to respondents' specific reproductive pathologies, or characterize infertility as male‐factor, female‐factor, combined, or unexplained. The NHANES questionnaire data relies on self‐report, with responses subject to recall bias and social desirability bias, particularly relevant for identification of reproductive‐risk factors.

Our reference group of participants with no employment history was small and differed slightly in demographics from employed respondents. We did not select a low‐occupational exposure sector as the reference population, given the heterogeneity of 2010 Census I&O codes and our merged categories. Sensitivity analyses using multiple imputation to include participants with selected missing reproductive covariates were consistent with our main analyses, suggesting that our choice of reference group is unlikely to have biased our findings.

5. Conclusion

To our knowledge, this is the first study that used NHANES to investigate the association between infertility risk and employment history by considering a range of industry sectors as proxy measures for exposures. The findings reveal an increased risk of infertility among women whose primary employment was in the manufacturing of durable goods industry, compared with women without a history of employment. Our exploratory study supports a more focused examination of infertility risk in this sector, considering both female‐ and male‐factor infertility. Expanding both the occupational and infertility questionnaires in future NHANES cycles to collect more complete employment and reproductive histories would enhance the assessment of infertility risk across industry sectors.

Author Contributions

Cerine Benomar: conceptualization, methodology, data analysis, writing – original draft, review. Anne T. M. Konkle: conceptualization, methodology, writing – review, editing, supervision. Karen P. Phillips: conceptualization, methodology, writing – review, editing, revision to original draft, supervision.

Funding

The authors received no specific funding for this work.

Ethics Statement

NHANES 2013–2014 was approved by the NCHS Ethics Review Board (Continuation of Protocol #2011‐17) with written informed consent obtained from all participants. Our study is a secondary data analysis of the NHANES dataset, publicly available, anonymized to ensure the confidentiality of participants.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Table S1: Detailed participant demographics stratified by employment history. Supplementary Table S2: Bivariate analyses of employment history across infertility measures. Supplementary Table S3: Demographic characteristics: Complete‐case sample and excluded participants through MICE. Supplementary Table S4: Unadjusted model ‐ Association between employment history and infertility after multiple imputation (MICE). Supplementary Table S5: Adjusted model – Association between employment history and infertility after multiple imputation (MICE).

Data Availability Statement

The data that support the findings of this study are available in NHANES 2013–2014 at https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/default.aspx?BeginYear=2013.