Materials and Methods

The Nurses’ Health Study II was established in 1989 as a prospective cohort study of 116,430 U.S. nurses, aged 25 to 42, in 14 states. Participants completed mailed questionnaires regarding their medical and reproductive history at baseline and were sent follow-up questionnaires every 2 years. In the 2001 questionnaire, participants were asked if they had experienced at least 1 pregnancy since 1993 and had worked as a nurse during the most recent of those pregnancies. If so, participants were asked whether they would be willing to complete a mailed supplemental questionnaire regarding occupational activities during their most recent pregnancy.

Among the 101,681 respondents to the 2001 biennial questionnaire, 11,177 (11%) had a pregnancy since 1993 during which they worked as a nurse. Willingness to complete the supplemental survey was indicated by 9547 (85%) participants; 8461 (89%) of these women completed and returned the supplemental survey they were mailed, resulting in an overall participation rate of 76% for the supplement. Because multiple pregnancies per woman are not independent events, we asked only about the most recent pregnancy.

Pregnancies ending in an induced abortion (n = 147), ectopic pregnancy (n = 57), molar pregnancy (n = 13), or multiple pregnancy (n = 235), as well as 7 pregnancies with a missing pregnancy outcome, were excluded. Stillbirths, defined as a pregnancy loss after 20 weeks’ gestation, were also excluded (n = 42). Pregnancies were excluded if the participant reported that the pregnancy was not confirmed by a clinical or over-the-counter pregnancy test (n = 130), as well as pregnancies without data on the year the pregnancy ended (n = 31) or the length of the pregnancy (n = 22). Women who reported working less than 1 hour per week as a nurse during the first trimester of pregnancy (n = 65), or who did not provide information on shift work (n = 24), anesthetic gases (n = 19), sterilizing agents (n = 91), antineoplastic drugs (n = 14), antiviral drugs (n = 33), or X-rays (n = 49) were also excluded. In total, 979 (11.6%) women were excluded, leaving data on 7482 women in the analysis.

For each trimester of pregnancy, the following questions were asked: “On average how many hours per day did you work with the following agents:” and the following categories, with examples, were given “a. Anesthetic gases (eg, nitrous oxide, halothane, enflurane, isoflurane),” “b. Anticancer drugs (eg, Cytoxan, Fluroplex, Adrucil, Etoposide, 5-FU),” “c. Antiviral drugs (eg, gancyclovir or the interferons),” “d. Sterilizing agents (eg, ethylene oxide, formaldehyde, glutaraldehyde),” and “e. X-ray radiation.” Data on work schedule, night work, average hours worked per week, frequency of lifting 25 pounds or more at work, hours per day of standing or walking at work, smoking, alcohol consumption, and caffeine consumption were also collected for each trimester of pregnancy. From the main biennial questionnaires, data were also available on age, race/ethnicity, body mass index (BMI), medication use, parity, and prior spontaneous abortion.

Participants reported the outcome of the index pregnancy as a single live birth, stillbirth, twins, triplets+, induced abortion, tubal/ectopic pregnancy, miscarriage, or molar pregnancy. Categorical information on pregnancy duration was reported on the supplemental survey as weeks since last menstrual period (<8, 8-11, 12-19, 20-23, 24-27, 28-31, 32-36, 37-41 [full-term], and ≥42 weeks). Pregnancies ending involuntarily before 20 weeks’ gestation were classified as spontaneous abortions.

Completion of the self-administered questionnaire was considered implied informed consent. The study and informed consent procedure were approved by the institutional review board of the Brigham and Women’s Hospital.

Descriptive statistics (frequency, range, and age-adjusted means) were calculated for selected maternal characteristics. The relationship between spontaneous abortion and first-trimester exposure to anesthetic gases, antineoplastic agents, sterilizing agents, antiviral drugs, and X-rays was examined in univariable and multivariable analyses. We modeled age as a continuous variable. Because the risk of spontaneous abortion rose exponentially with age, we also included a quadratic age term (age-squared) in the models. Exposure categories were dichotomized as 1+ hour per day vs <1 hour per day for all reported exposures.

For univariable and multivariable analyses, we used logistic regression to compute the odds ratio (OR) using SAS software (SAS Institute, Cary, NC). Covariates that changed the estimate by 10% or more were retained in the final multivariable model.

We first considered the associations of these individual exposures with risk of spontaneous abortion, adjusted for age. Our full multivariable model included all 5 work exposures of interest, age, parity, and work schedule. The addition of lifestyle and other factors, such as cigarette smoking, caffeine consumption, alcohol consumption, BMI, race/ethnicity, lifting at work, and standing at work did not change the estimated effects by more than 10%. Because previous occurrence of spontaneous abortion could have been due to the same environmental or occupational exposures studied in the index spontaneous abortion, and to avoid possible bias, we did not adjust for previous spontaneous abortion in these analyses. Our primary analysis looked at all spontaneous abortions as the outcome. In our secondary analysis, we stratified by the timing of the spontaneous abortion, because early and late spontaneous abortion may be controlled by different mechanisms, and the percentage of spontaneous abortions caused by chromosomal abnormalities decreases as gestational age increases. For this analysis, early spontaneous abortion is modeled as <12 weeks’ gestation, and late spontaneous abortion as 12-20 weeks. For the analysis of late spontaneous abortions, early spontaneous abortions were excluded (n = 575). To assess statistically significant differences between early and late models, we calculated P values for common effects with a χ ² test statistic using the maximum likelihood estimates from the logistic regression. We also assessed interactions between parity and each exposure by modeling a cross-product interaction term in a model containing the main effects of parity and the exposure, as well as other covariates.

Results

Among 7482 eligible participants, the pregnancies of 775 (10%) ended in spontaneous abortion. Seventy-four percent of those ended before the 12th week of pregnancy (n = 575). The year of the pregnancies ranged from 1993-2002 (the mean year was 1996), with 82% occurring between 1993 and 1998. The spontaneous abortion rate varied by specialty area; the lowest rates were for medical/surgical and critical care (8.4% and 8.8%, respectively), and the highest rates for home health/community and oncology (13.1% for each). However, 32% of nurses specified “other” as their specialty area (11.0% spontaneous abortion rate).

Table 1 shows age-adjusted prevalence of selected characteristics of eligible participants by pregnancy outcome. Women whose pregnancies ended in spontaneous abortion were older and less likely to be parous than those with live births. Prior spontaneous abortion, higher consumption of caffeinated beverages and alcohol, and cigarette smoking were also more common among pregnancies ending in spontaneous abortion. Occupational exposures were reported more often in spontaneously aborted pregnancies, particularly exposure to antineoplastic drugs, which was reported nearly twice as often compared with live births ( Table 2 ). As previously reported, women whose pregnancies failed were more likely to have worked the night shift and to have worked long hours during the first trimester than women with live births.

TABLE 1

Demographic and lifestyle factors ^a by pregnancy outcome

Study subject characteristics	Spontaneous abortion <20 weeks’ gestation n = 775 (10.4%)		Live birth n = 6707 (89.6%)
Maternal prepregnancy BMI b Mean, range, (SD)	25.3, 16.1–50.0 (5.7)		24.3, 15.0–50.0 (4.9)
Maternal age (y) Mean, range, (SD)	39.5, 30–51 (3.8)		36.4, 29–50 (3.4)
Maternal age category	n	%	n	%
≤30	6	0.8	175	2.6
31-35	112	14.5	2609	38.9
36-40 (referent)	344	44.4	3113	46.4
41+	313	40.4	810	12.1
Race
African American c	5	0.4	42	0.7
Asian	12	1.1	119	1.9
White	719	92.5	6229	92.8
Hispanic	11	1.3	101	1.5
Other	15	2.2	100	1.5
Missing	13	2.5	116	1.7
Parous	535	68.2	5698	84.6
Previous spontaneous abortion	394	44.9	2296	34.9
First-trimester lifestyle factors
Caffeinated coffee servings d
None (reference)	412	55.1	4151	61.6
≤1 cup per day	232	28.8	1874	28.1
≥2 cups per day	130	15.8	676	10.2
Missing	1	0.4	6	0.1
Caffeinated soda/tea servings d
None (reference)	379	45.7	3191	47.9
≤1 serving per day	284	38.4	2781	41.2
≥2 servings per day	112	15.9	720	10.7
Missing	0	0.0	15	0.2
Alcoholic beverage servings d
None (reference)	584	76.8	5459	81.2
<1 drink per week	120	14.8	926	13.9
≥1 drink per week	70	8.3	315	4.8
Missing	1	0.1	7	0.1
Smoked cigarettes per day
None (reference)	710	92.2	6316	94.2
≥1 cigarette per day	65	7.6	389	5.8
Missing	0	0.0	2	0.0

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Autism risk in small- and large-for-gestational-age infants
2. Factors associated with screening for sexually transmitted infections
3. Hospital-based lateral episiotomy and obstetric anal sphincter injury rates: a retrospective population-based register study
4. Early detection of placental inflammation by MRI enabling protection by clinically relevant IL-1Ra administration

Stay updated, free articles. Join our Telegram channel

Join