Historically, pregnant women have been encouraged to reduce physical activity and stop working, especially during the latter stages of pregnancy. Despite possible negative effects, increasing numbers of women are engaging in regular exercise and working throughout pregnancy. What effect this is having on the pregnancy outcomes of these women is not clear.

The present study examines the relations of exercise, employment, and other daily activities with pregnancy outcome using the Missouri Maternal and Infant Health Survey1 (MMIHS). The MMIHS is a population-based case-control study developed to examine factors related to the prevention and subsequent health care needs of very-low-birth-weight infants as well as stillbirths. It is a joint project of the National Institute of Child Health and Human Development and the Missouri Department of Health.

The MMIHS is a study of all Missouri resident very-low- birth-weight (VLBW, <1,500 g) infants born between December 1, 1989, and March 31, 1991, with similar numbers of moderately low-birth-weight (MLBW, 1,500-2,499 g) infants and normal-birth-weight (NBW, >2,500 g) infants serving as control for singletons. The study also covered all Missouri resident fetal deaths (20 weeks gestation or more) for the same time period.

The overall response rate for the maternal survey was 75.6 percent (3,102 of 4,104). Maternal surveys were available for the following singleton birth categories, 450 fetal deaths, 782 VLBW, 802 MLBW, and 794 normal-birth-weight births.

For the exercise question, we asked how frequently the mothers engaged in exercise activities during the 3 months before pregnancy, during each trimester, and during the 3 months after pregnancy. We gave examples of what we meant by vigorous exercise activity (at least 20 minutes at a time) and asked them to check one of four boxes for each time period requested: never, less than once a week, one to two times a week, and three or more times a week. We also asked them to describe the activity they did most often.

Unconditional logistic regression was used to calculate adjusted odds ratios of exercising three or more times per week (exercise) for each trimester between the various outcome groups: VLBW, MLBW, NBW, and fetal deaths. Covariates used in these models included the following variables: mother's perception of general health during pregnancy (4-point scale from excellent to poor), physician advised mother to exercise more or less, age, education level and prepregnancy weight of mother, bed rest prescribed for 24 hours per day, mother smoked during pregnancy, race (black, non-black), and source of survey (hospital or mail).

As seen in Table 1, the exercise frequency percentages changed throughout the pregnancy by outcome (i.e., VLBW, MLBW, NBW and fetal deaths). Before pregnancy, the exercise distributions by outcome were very similar with approximately one third of each group not exercising, slightly less than one third exercising three or more times per week. Slightly less than 60 percent of those who exercised walked while the rest were involved in more vigorous activities such as swimming, doing aerobics, or playing tennis.

As their pregnancies progressed, more women discontinued exercise as the percentage of those exercising three or more times per week decreased to approximately 16 percent for the total study sample in the third trimester whereas only 7 percent of the VLBW mothers exercised three or more times in this trimester. Of those who did exercise, 80 percent of the NBW mothers and 85 percent of the VLBW mothers primarily walked.

In the 3 months immediately after delivery, the no exercise level returned to approximately the first trimester level and the proportion exercising three or more times per week was about the same as in the second trimester. There was little difference between the outcome groups during this period after pregnancy.

Table 2 reflects the decreased exercise crude odds ratio between VLBW and NBW mothers as their pregnancies progressed. From a non-statistically significant ratio of 0.95 (95 percent confidence interval (CI) 0.76-1.18) before pregnancy, the exercise crude odds ratio decreased to 0.71 (95 percent CI 0.55-0.90) in the first trimester, 0.52 (95 percent CI 0.39-0.68) in the second trimester, and 0.32 (95 percent CI 0.20-0.49) in the third trimester. During the 3 months after pregnancy, the ratio returned to a nonsignificant 0.82 (95 percent CI 0.57-1.16).

After adjustment for the confounding variables, the exercise odds ratios changed only slightly, as also shown in Table 2. The ratio in the second trimester was still 0.54 to 1, and the exercise odds ratio was 0.33 to 1 in the third trimester between VLBW and NBW mothers.

The relationship of exercise by trimester between other pregnancy outcomes is shown in Table 3. No significant relation was found between mothers of moderately low-birth- weight infants and those having normal-birth-weight babies. VLBW mothers were significantly less likely to exercise in the first, second, and third trimesters than MLBW mothers; and VLBW mothers were also significantly less likely to exercise in the second or third trimesters than mothers having fetal deaths.

Next we examined other daily activities during pregnancy. There was no relation between working during pregnancy and outcome of pregnancy inasmuch as approximately 60 percent in each group work. There also was no relation between hours worked per week and pregnancy outcome.

No significant associations were found for other daily activities and adverse pregnancy outcome. The activities tested included climbing stairs, standing for 3 or more hours, not being able to rest when tired, carrying loads, and engaging in strenuous household activities. One exception was the VLBW and MLBW mothers were slightly less likely to take care of preschool children during pregnancy than NBW mothers (odds ratios = 0.74 and 0.81, respectively). This is another indication of VLBW mothers being less active during pregnancy than other mothers.

DISCUSSION

The major finding of this study is that mothers of VLBW infants are much less likely to exercise during pregnancy than mothers having other pregnancy outcomes. The exercise odds ratios between VLBW mothers and other mothers decreased significantly in the second or third trimesters. No significantly increased risks were found between employment during pregnancy, types of work activities and conditions, or other daily activities and adverse pregnancy outcome.

Our exercise findings are consistent with the recent relaxation of American College of Obstetricians and Gynecologists guidelines (2) regarding exercise in pregnancy. They are also consistent with the Lokey et al. (3) and Hatch et al. (4) studies that found no association with adverse outcome for low-risk women exercising during pregnancy.

The relations between VLBW and lack of maternal exercise in the second and third trimesters may be due to a number of factors. It may be because the VLBW mothers are in poorer physical health during pregnancy. Medical risk indicators during pregnancy may lead physicians to encourage VLBW mothers to restrict activity. We did, however, control for various questions related to poor maternal health and physician's advice regarding activity, which had little effect on the exercise odds ratios between pregnancy outcomes. A stratification analysis by pregnancy symptoms also did not substantially affect odds ratios.

The relations may represent some artifact in the data. The exercise questions were self-reported and were answered by mothers after their pregnancies, involving possible recall bias for the VLBW mothers in particular. The exercise odds ratios may have been deflated, particularly in the third trimester, because the VLBW mothers had shorter pregnancies and fewer weeks to exercise. The lack of a stronger relationship between exercise and some of the covariates may be an indication of inaccuracy in maternal responses to the survey questions.

It is possible that lack of exercise does contribute to very-low-birth-weight outcomes. To confirm this hypothesis, it would be necessary to do a large prospective study in which pregnant women's exercise programs would be very closely supervised. The conclusion of this Missouri study provides no indication that exercising regularly, working, or doing other normal activities during pregnancy contributes to adverse pregnancy outcomes.

References: