In a case-control study looking at environmental and occupational factors for congenital malformations (Kurppa et al. 1986), 1,475 cases were identified from the Finnish Register of Congenital Malformations during the period between 1976 and 1982 (see table 1). A mother whose delivery immediately preceded a case, and was in the same district, served as a control for that case. Exposure to visual display units (VDUs) during the first trimester of pregnancy was assessed using face-to-face interviews conducted either at the clinic during a post-natal visit, or at home. The classification of probable or obvious VDU use was determined by occupational hygienists, blind to the pregnancy outcomes, using job titles and the responses to open-ended questions asking to describe the ordinary work day. There was no evidence of increased risk either among women who reported exposure to VDUs (OR 0.9; 95% CI 0.6 – 1.2), or among women whose job titles indicated possible exposure to VDUs (235 cases/255 controls).
A cohort of Swedish women from three occupational groups was identified through a linkage of occupational census and the Medical Birth Registry during 1980–1981 (Ericson and Källén 1986). A case-base study was conducted within that cohort: cases were 412 women hospitalized for spontaneous abortion and an additional 110 with other outcomes (such as perinatal death, congenital malformations and birthweight below 1500 g). Controls were 1,032 women of similar age who had infants without any of these characteristics, chosen from the same registry. Using crude odds ratios, there was an exposure–response relation between VDU exposure in estimated hours per week (divided into five-hour categories) and pregnancy outcomes (excluding spontaneous abortion). After controlling for smoking and stress, the effect of VDU use on all adverse pregnancy outcomes was not significant.
Focusing on one of three occupational groups identified from a previous study by Ericson a cohort study was conducted using 4,117 pregnancies among social security clerks in Sweden (Westerholm and Ericson 1986). Rates of hospitalized spontaneous abortion, low birthweight, perinatal mortality and congenital malformations in this cohort were compared to rates in the general population. The cohort was divided into five exposure groups defined by trade union and employer representatives. No excesses were found for any of the studied outcomes. The overall relative risk for spontaneous abortion, standardized for mothers’ age was 1.1 (95% CI 0.8 – 1.4).
A cohort study involving 1,820 births was conducted among women having ever worked at the Norwegian Postal Giro Centre between 1967–1984 (Bjerkedal and Egenaes 1986). The rates of stillbirth, first-week death, perinatal death, low and very low birthweight, preterm birth, multiple births and congenital malformations were estimated for pregnancies occurring during employment at the centre (990 pregnancies), and pregnancies occurring before or after employment at the centre (830 pregnancies). Rates of adverse pregnancy outcomes were also estimated for three six-year periods, (1967–1972), (1973–1978) and (1979–1984). Introduction of VDUs began in 1972, and were extensively used by 1980. The study concluded that there was no indication that introduction of VDUs in the centre had led to any increase in the rate of adverse pregnancy outcomes.
A cohort of 9,564 pregnancies was identified through logs of urine pregnancy tests from three California clinics in 1981–1982 (Goldhaber, Polen and Hiatt. 1988). Coverage by a Northern California medical plan was a requirement to be eligible for the study. Pregnancy outcomes were found for all but 391 identified pregnancies. From this cohort, 460 of 556 spontaneous abortion cases (<28 weeks), 137 of 156 congenital abnormality cases and 986 of 1,123 controls (corresponding to every fifth normal birth in the original cohort), responded to a retrospective postal questionnaire on chemical environmental exposures including pesticides and VDU use during pregnancy. Odds ratios for women with first trimester VDU use over 20 hours per week, adjusted for eleven variables including age, previous miscarriage or birth defect, smoking and alcohol, were 1.8 (95% CI 1.2 – 2.8) for spontaneous abortion and 1.4 (95% CI 0.7 – 2.9) for birth defects, when compared to working women who did not report using VDUs.
In a study conducted in 11 hospital maternity units in the Montreal area over a two-year period (1982–1984), 56,012 women were interviewed on occupational, personal and social factors after delivery (51,855) or treatment for spontaneous abortion (4,127) (McDonald et al. 1988).These women also provided information on 48,637 previous pregnancies. Adverse pregnancy outcomes (spontaneous abortion, stillbirth, congenital malformations and low birthweight) were recorded for both current and previous pregnancies. Ratios of observed to expected rates were calculated by employment group for current pregnancies and previous pregnancies. Expected rates for each employment group were based on the outcome in the whole sample, and adjusted for eight variables, including age, smoking and alcohol. No increase in risk was found among women exposed to VDUs.
A cohort study comparing rates of threatened abortion, length of gestation, birthweight, placental weight and pregnancy-induced hypertension between women who used VDUs and women who did not use VDUs was carried out among 1,475 women (Nurminen and Kurppa 1988).The cohort was defined as all non-cases from a previous case-control study of congenital malformations. Information about risk factors was collected using face-to-face interviews. The crude and adjusted rate ratios for the outcomes studied did not show statistically significant effects for working with VDUs.
A case-control study involving 344 cases of hospitalized spontaneous abortion occurring at three hospitals in Calgary, Canada, was conducted in 1984–1985 (Bryant and Love 1989). Up to two controls (314 prenatal and 333 postpartum) were chosen among women having delivered or susceptible of delivering at the study hospitals. The controls were matched to each case on the basis of age at last menstrual period, parity, and intended hospital of delivery. VDU use at home and at work, before and during pregnancy, was determined through interviews at the hospitals for postnatal controls and spontaneous abortion, and at home, work, or the study office for prenatal controls. The study controlled for socioeconomic and obstetric variables. VDU use was similar between the cases and both the prenatal controls (OR=1.14; p=0.47) and postnatal controls (OR=0.80; p=0.2).
A case-control study of 628 women with spontaneous abortion, identified through pathology specimen submissions, whose last menstrual period occurred in 1986, and 1,308 controls who had live births, was carried out in one county in California (Windham et al. 1990). The controls were randomly selected, in a two-to-one ratio, among women matched for date of last menstrual period and hospital. Activities during the first 20 weeks of pregnancy were identified through telephone interviews. The participants were also asked about VDU use at work during this period. Crude odds ratios for spontaneous abortion and VDU use less than 20 hours per week (1.2; 95% CI 0.88 – 1.6), and at least 20 hours per week (1.3; 95% CI 0.87 – 1.5), showed little change when adjusted for variables including employment group, maternal age, prior foetal loss, alcohol consumption and smoking. In a further analysis among the women in the control group, risks for low birthweight and intrauterine growth retardation were not significantly elevated.
A case-control study was conducted within a study base of 24,352 pregnancies occurring between 1982 and 1985 among 214,108 commercial and clerical employees in Denmark (Brandt and Nielsen 1990). The cases, 421 respondents among the 661 women who gave birth to children with congenital abnormalities and who were working at the time of pregnancy, were compared to 1,365 respondents among the 2,252 randomly selected pregnancies among working women. Pregnancies, and their outcomes, and employment were determined through a linkage of three databases. Information on VDU use (yes/no/hours per week), and job-related and personal factors such as stress, exposure to solvents, life-style and ergonomic factors were determined through a postal questionnaire. In this study, the use of VDUs during pregnancy was not associated with an increased risk of congenital abnormalities.
Using the same study base as in the previous study on congenital abnormalities (Brandt and Nielsen 1990) 1,371 of 2,248 women whose pregnancies ended in a hospitalized spontaneous abortion were compared to 1,699 randomly selected pregnancies (Nielsen and Brandt 1990). While the study was carried out among commercial and clerical workers, not all of the pregnancies corresponded to times when the women were gainfully employed as commercial or clerical workers. The measure of association used in the study was the ratio of the rate of VDU use among women with a spontaneous abortion to the rate of VDU use among the sample population (representing all pregnancies including those ending in spontaneous abortion). The adjusted rate ratio for any exposure to VDU and spontaneous abortion was 0.94 (95% CI 0.77 – 1.14).
A case-control study was carried out among 573 women who gave birth to children with cardiovascular malformations between 1982 and 1984 (Tikkanen and Heinonen 1991). The cases were identified through the Finnish register of congenital malformations. The control group consisted of 1,055 women, randomly selected among all hospital deliveries during the same time period. VDU use, recorded as never, regular or occasional, was assessed through an interview conducted 3 months after the delivery. No statistically significant association was found between VDU use, at work or at home, and cardiovascular malformations.
A cohort study was carried out among 730 married women who reported pregnancies between 1983 and 1986 (Schnorr et al. 1991). These women were employed as either directory assistance operators or as general telephone operators at two telephone companies in eight southeastern states in the United States. Only the directory assistance operators used VDUs at work. VDU use was determined through company records. Spontaneous abortion cases (foetal loss at 28 weeks’ of gestation or earlier) were identified through a telephone interview; birth certificates were later used to compare women’s reporting with pregnancy outcomes and when possible, physicians were consulted. Strengths of electric and magnetic fields were measured at very low and extremely low frequencies for a sample of the workstations. The VDU workstations showed higher field strengths than those not using VDUs. No excess risk was found for women who used VDUs during the first trimester of pregnancy (OR 0.93; 95% CI 0.63 – 1.38), and there was no apparent exposure–response relation when looking at time of VDU use per week.
A cohort of 1,365 Danish commercial and clerical workers who were gainfully employed at the time of pregnancy, and identified through a previous study (Brandt and Nielsen 1990; Nielsen and Brandt 1990), was used to study fecundability rates, in relation to VDU use (Brandt and Nielsen 1992). Fecundability was measured as time from stopping birth control use to time of conception, and was determined through a postal questionnaire. This study showed an increased relative risk for prolonged waiting to pregnancy for the subgroup with at least 21 weekly hours of VDU use. (RR 1.61; 95% CI 1.09 – 2.38).
A cohort of 1,699 Danish commercial and clerical workers, consisting of women employed and unemployed at the time of pregnancy, identified through the study reported on in the previous paragraph, was used to study low birthweight (434 cases), preterm birth (443 cases), small for gestational age (749 cases), and infant mortality (160 cases), in relation to VDU use patterns (Nielsen and Brandt 1992). The study failed to show any increased risk for these adverse pregnancy outcomes among women with VDU use.
In a case-control study, 150 nulliparous women with clinically diagnosed spontaneous abortion and 297 nulliparous working women attending a hospital in Reading, England for antenatal care between 1987 and 1989 were interviewed (Roman et al. 1992). The interviews were conducted face to face at the time of their first antenatal visit for the controls, and three weeks after the abortion for women with spontaneous abortion. For women who mentioned VDU use, estimates of time of exposure in hours per week, and calendar time of first exposure were assessed. Other factors such as overtime, physical activity at work, stress and physical comfort at work, age, alcohol consumption and previous miscarriage were also assessed. Women who worked with VDUs had an odds ratio for spontaneous abortion of 0.9 (95% CI 0.6 – 1.4), and there was no relation with the amount of time spent using VDUs. Adjusting for other factors such as maternal age, smoking, alcohol and previous spontaneous abortion did not alter the results.
From a study base of bank clerks and clerical workers in three companies in Finland, 191 cases of hospitalized spontaneous abortion and 394 controls (live births) were identified from Finnish medical registers for 1975 to 1985 (Lindbohm et al. 1992). Use of VDUs was defined using workers’ reports and company information. Magnetic field strengths were retrospectively assessed in a laboratory setting using a sample of the VDUs which had been used in the companies. The odds ratio for spontaneous abortion and working with VDUs was 1.1 (95% CI 0.7 – 1.6). When VDU users were separated in groups according to the field strengths for their VDU models, the odds ratio was 3.4 (95% CI 1.4 – 8.6) for workers who had used VDUs with a high magnetic field strength in the extremely low frequency bandwidth (0.9 μT), compared to those working with VDUs with field strength levels below the detection limits (0.4 μT). This odds ratio changed only slightly when adjusted for ergonomic and mental work-load factors. When comparing workers exposed to high magnetic field strengths to workers not exposed to VDUs, the odds ratio was no longer significant.
A study, looking at adverse pregnancy outcomes and fertility, was carried out among female civil servants working for the British Government tax offices (Bramwell and Davidson 1994). Of the 7,819 questionnaires mailed in the first stage of the study, 3,711 were returned. VDU use was determined through this first questionnaire. Exposure was assessed as hours per week of VDU use during pregnancy. One year later, a second questionnaire was sent out to assess the incidence of adverse pregnancy outcomes among these women; 2,022 of the original participants responded. Possible confounders included pregnancy history, ergonomic factors, job stressors, caffeine, alcohol, cigarette and tranquillizer consumption. There was no relationship between exposure as assessed one year previously and the incidence of adverse pregnancy outcomes.