Preterm birth, defined as a delivery occurring at less than 37 weeks’ gestation, 1 is the leading cause of perinatal mortality in North America and Europe and a major predictor of neonatal and infant morbidity. 2 The rate of preterm birth remains higher in the United States than in many other industrialized countries. 2,3 This burden underscores the importance of understanding factors related to preterm delivery. There are potential biological consequences of vigorous activity during pregnancy that could increase the risk of preterm birth or other complications. Exercise increases catecholamines, which could produce uterine vasoconstriction leading to decreased uterine blood flow, and may trigger uterine contractions. 4 Vigorous exercise also increases blood flow to exercising muscles and may divert needed blood away from the placenta and baby. 5 However, several studies have found leisure activity during pregnancy was associated with a reduced risk of preterm birth 2,6–9 and may provide other potential benefits related to preterm birth. 4
Postterm birth, defined as delivery occurring at or beyond 42 weeks’ gestation, 1 is associated with increased infant mortality and morbidity, and increased frequency of surgical delivery or induced labor. 10 The incidence of postterm birth ranges from 2%–14% depending on the study population and methodology used. 10 Risk factors for postterm birth have received little attention and epidemiologic investigation could help elucidate mechanisms of normal pregnancy. 10 We are aware of only one published study examining the relation between physical activity and postterm birth. 8 The authors found that vigorous leisure activity was weakly associated with a reduced risk of late delivery. To further investigate these issues, this study examined the risk of preterm and postterm birth for women who participated in vigorous leisure activity before pregnancy and during the first or second trimester of their pregnancy.
The Pregnancy, Infection, and Nutrition Study was designed to investigate risk factors for preterm birth. The study recruited predominantly lower- to middle-income women at 24–29 weeks’ gestation from four prenatal care clinics. 11 Recruitment began in August 1995; the present analyses include women who started their pregnancy between January 1995 and June 1998. During this time, 4136 women were eligible for the study, 2,493 were successfully recruited and provide informed consent, and 2,359 delivered at the participating hospitals. Patterns of study participation were examined and have been reported elsewhere. 11 The procedures followed for this study were in accordance with the ethical standards of the Institutional Review Board of the University of North Carolina School of Medicine.
Vigorous Leisure Activity Measurement
Women participated in a telephone interview in the 2 weeks after enrollment (26–31 weeks’ gestation). Women were asked, “Thinking back to 3 months before you got pregnant until now, have there been times when you have done any regular exercise or strenuous activity, like aerobic exercise or jogging, at least twice a week?” If the woman answered affirmatively, she was asked about participation in selected activities at any of three time periods (3 months before pregnancy, the first 3 months of pregnancy, and the second 3 months of pregnancy) and how many hours of participation per week in each period. The selected activities were identified from the National Maternal and Infant Health Survey as some of the most common vigorous activities, 12 including swimming laps, jogging at a moderate to fast pace, aerobics or aerobic dance, other fast dancing, and moderate to fast bicycling. Women could also report other kinds of exercise or strenuous activity. A dichotomous variable was created to indicate whether the woman participated in any of these activities, separately for each of the three time periods. This variable was defined as participation in any vigorous leisure activity, because all of the predefined activities as well as most of the “other” activities required an intensity of at least six metabolic equivalents (METS). 13 Vigorous leisure activity was also examined in categories based on hours of activity per week (0, 0.1–2.9, or 3+).
Other Covariate Measurement
During the recruitment visit, women were asked whether or not they currently smoked. Women were asked during the telephone interview about their race, marital status, education, parity (live plus still births), and bedrest. Bedrest was determined by the question, “Did a doctor or nurse tell you that you should stay at home in bed rather than follow your regular schedule?” Information concerning dietary intake during the second trimester was collected using a modified National Cancer Institute Block Questionnaire;14 details of the modifications are described elsewhere. 15 Quartiles of average daily energy intake were used for these analyses. Information obtained from the medical record included self-reported weight and measured height for the determination of prepregnancy body mass index (BMI; weight [kg] per height squared [m2]). The mother’s age at time of conception was determined from the medical record and interview.
Pregnancy Outcome Measurements
We obtained pregnancy outcome information from hospital delivery logs. For the 2359 participants, we estimated gestational age at the time of delivery based on an algorithm that combined last menstrual period with ultrasound dating. 16 If both measures were available and they agreed within 14 days, then the last menstrual period dating was assigned (N = 1603). If the disagreement was greater than 14 days, the ultrasound dating was assigned (N = 313). If a reliable last menstrual period was not available, information from the earliest available ultrasound was used (N = 244). Similarly, if an ultrasound was not recorded, then the last menstrual period was used for calculations (N = 199).
Preterm birth was defined as delivery before the completion of 37 weeks’ gestation. 1 One of three obstetricians examined the charts of all preterm deliveries to determine whether the delivery was spontaneous or medically induced, as well as the clinical presentation. Because the underlying mechanisms may differ, 17 these outcomes were also examined separately. We defined idiopathic preterm labor as labor starting without apparent reason before rupture of the membranes, resulting in preterm delivery. Premature rupture of the membranes (PROM) was defined as labor or membrane rupture that precipitated a sequence of events culminating in delivery. Idiopathic preterm labor and PROM were combined for analysis and defined as spontaneous preterm birth. Medically indicated delivery was defined as an indication of labor or caesarian section before 37 weeks based on the health of the mother or fetus. Preterm deliveries were also divided into early (<34 weeks) and late (34–36 weeks). Postterm birth was defined as gestational age >41 weeks. 1
We compared prevalence of vigorous leisure activity by birth outcome. Unadjusted prevalence ratios were calculated by dividing the prevalence of vigorous leisure activity in one group by the prevalence of vigorous leisure activity among term births (referent). We generated unadjusted Kaplan-Meier curves showing time to delivery for vigorous leisure activity at the three time periods, and compared these using Gehan’s Wilcoxon’s test. Unadjusted and adjusted unconditional logistic regression models were used to examine the relation between vigorous leisure activity (separately for each time period) and birth outcomes, with term birth as the referent group. Because hazard ratios may be more reflective of events that influence timing of birth, we used Cox proportional hazard models to examine the prediction of days of gestation for women who delivered preterm (N = 193).
Selection of potential confounders was determined from findings, reported in the literature, on factors that may be related to both vigorous leisure activity and birth outcomes (eg, maternal age, marital status, race, education, pregravid BMI, smoking, parity, bedrest, and energy intake). 2,10 There were 2359 women eligible for analysis. After we removed women with second pregnancies (N = 45), those whose preterm etiology could not be assessed (N = 2), those who did not complete the physical activity portion of the interview (N = 221), and those who were missing covariate information (N = 394), 1699 women remained for analysis. The odds ratios for preterm birth were also examined, after stratification by whether or not women had a prior preterm birth.
Prevalence of Vigorous Leisure Activity
Among the participants, the overall prevalence of vigorous leisure activity was 22% before pregnancy, 14% during the first trimester, and 8% during the second trimester. The median number of hours per week of vigorous leisure activity among those reporting activity was 4 hours per week before pregnancy, 3 hours per week in first trimester, and 3 hours per week in second trimester.
Vigorous Leisure Activity and Preterm Birth
Participation in vigorous leisure activity before pregnancy was not different between women who delivered preterm and women who delivered term (Table 1). However, vigorous leisure activity during first and second trimester was less common among women who delivered preterm as compared with women who delivered term.
Logistic regression models predicting preterm birth, adjusting for confounding factors, indicated that vigorous leisure activity 3 months before pregnancy was not associated with preterm birth (Table 2). According to the multivariable odds ratios, vigorous leisure activity during first trimester, and more strongly during second trimester, was associated with a decreased risk of preterm birth. The odds ratios (ORs) indicated a reduced risk of spontaneous preterm birth for women who were vigorously active during leisure for the first and second trimester. Across all three time periods, vigorous leisure activity was more strongly associated with a reduced risk for early preterm births (<34 weeks) as compared with later preterm births (34–36 weeks). In most models, the adjustment for confounders attenuated the odds ratios only minimally.
We reexamined these associations, excluding women who may have been at risk for a poor birth outcome (based on a prior preterm birth), because such women may be especially likely to curtail activity for fear of exacerbating their risk. The crude prevalence of preterm birth was 10% (140 of 1391) among women without a prior preterm birth and 24% (53 of 224) among women with a prior preterm birth, excluding postterm births from the denominator. Among low-risk women, the adjusted odds ratio predicting preterm birth for participation in vigorous leisure activity was as follows: 0.96 (95% confidence interval [CI] = 0.60–1.53) before pregnancy, 0.80 (95% CI = 0.44–1.45) during the first trimester, and 0.50 (95% CI = 0.21–1.22) during the second trimester. Among higher-risk women, the adjusted odds ratios for participating in vigorous leisure activity were: 1.61 (95% CI = 0.70–3.69) before pregnancy, 0.94 (95% CI = 0.29–3.04) during first trimester, and 0.69 (95% CI = 0.14–3.51) during second trimester. This suggested some attenuation of the association, although this was difficult to evaluate because of imprecision.
To examine further the influence of vigorous leisure activity on preterm birth, multivariable adjusted Cox proportional hazard models (controlling for factors listed in Table 2) were used to examine the prediction of days of gestation for women who delivered preterm (N = 193). The adjusted hazard ratios for preterm birth were: 0.93 (95% CI = 0.63–1.38) with vigorous leisure activity before pregnancy, 0.86 (95% CI = 0.50–1.50) with vigorous leisure activity in the first trimester, and 1.46 (95% CI = 0.62–3.43) with vigorous leisure activity in the second trimester. These associations suggest that women participating in vigorous leisure activity before pregnancy and during first trimester may be slightly less likely to have an early preterm birth as compared with a later preterm birth, but women participating during second trimester are more likely to have an early preterm birth as compared with a later preterm birth. However, differences were modest overall, and imprecise.
Using multivariable odds ratios (controlling for factors listed in Table 2), we did not detect differences in preterm birth risk when comparing number of hours per week of vigorous leisure activity during the first trimester (0 hours, referent; 0.1–2.9 hours, 0.75 [95% CI = 0.36–1.56]; 3+ hours, 0.85 [95% CI = 0.44–1.66]). We also did not detect differences in preterm birth risk when comparing total number of hours per week of vigorous leisure activity during second trimester (0 hours, referent; 0.1–2.9 hours, 0.41 [95% CI = 0.12–1.38]; 3+ hours, 0.61 [95% CI = 0.23–1.57]).
Vigorous Leisure Activity and Postterm Birth
In unadjusted analyses, there was no association between participation in vigorous leisure activity 3 months before or during the first and second trimesters of pregnancy comparing postterm births with term births (Table 1). Similarly, in multivariable analyses, vigorous leisure activity was not associated with postterm birth (Table 2). During first or second trimester, we did not find a pattern of decreasing risk with greater duration of activity when considering the number of hours per week of participation (0 hours compared with 0.1–2.9 hours and 3+ hours; data not shown). However, the estimates were imprecise.
Summary Kaplan-Meier Graphs
Unadjusted Kaplan-Meier analyses of time to delivery compared time to delivery by participation in vigorous leisure activity before pregnancy, during first trimester, and during second trimester (Figure 1). There were no differences in delivery time when comparing those who did and did not participate in vigorous leisure activity 3 months before pregnancy or during the first trimester (Wilcoxon test P for trend = 0.64 and P for trend = 0.41, respectively). However, vigorous leisure activity during the second trimester was related to later delivery among women delivering between approximately 31–40 weeks’ gestation (Wilcoxon test P = 0.13). Differences were not detected for postterm births.
In a 1993 review of risk factors for preterm birth, participation in leisure-time physical activity was classified as having “insufficient data”. 2 Since the review was published, several studies have indicated that leisure activity is associated with a reduced risk of preterm delivery, 6–9 although one smaller study reported only a weak association (OR = 0.7; 95% CI = 0.3–1.6). 18
In this study, participation in vigorous leisure activity during the first and especially the second trimester of pregnancy was associated with a reduced risk of preterm birth. This finding is consistent with other studies of general obstetric populations. 6,8,19 However, vigorous leisure activity in the 3 months before pregnancy was not related to preterm birth. The 1994 American College of Obstetrics and Gynecology guidelines for physical activity during pregnancy 20 indicate that women with normal pregnancies can engage in exercise with almost no restriction without compromising fetal growth and development or complicating pregnancy, labor, or delivery. Our results support a portion of this statement: vigorous leisure activity 3 months before pregnancy and during the first and second trimesters clearly did not increase the risk of preterm birth.
This study also examined the categories of preterm birth, and found that across all three time periods (before pregnancy, first trimester, and second trimester), vigorous leisure activity was more strongly associated with a reduced risk for early preterm births (<34 weeks) as compared with later preterm births (34–36 weeks). One other study examined the difference in association between regular physical activity and very premature (<33 weeks), premature (33–35 weeks), and borderline term delivery (36–37 weeks). 9 Regular physical activity was associated with a reduced risk of very premature delivery (OR = 0.45; 95% CI = 0.16–1.26) and premature delivery (OR = 0.30; CI = 0.11–0.83) as compared with borderline term delivery, controlling for multiple confounders.
We did not find a pattern of decreasing risk with greater duration of activity. However, our measure was imprecise and future studies should attempt to examine this further. There may be a threshold effect above which increasing duration of activity during pregnancy no longer becomes beneficial. When considering intensity of activity, the Hatch et al. study 8 indicated that moderate activity was not related to preterm birth, whereas vigorous activity was associated with a lower risk. Our study collected information only on vigorous leisure activity and was not able to make this useful comparison. Future studies should seek to measure both moderate and vigorous intensity activities, because many of the activities women are doing are moderate rather than vigorous in nature. 21
To isolate women who were less likely to reduce their physical activity related to medical risks, we conducted stratified analyses by risk status for an adverse delivery outcome, based on whether women reported a prior preterm birth. This was intended to isolate the effect of physical activity from its antecedents, namely that well-being may be necessary for some women to choose to be vigorously active during leisure. However, risk estimates for preterm birth were similar between lower- and higher-risk women participating in vigorous leisure activity during first or second trimester, subject to imprecision.
In this study, vigorous leisure activity before pregnancy and during first or second trimester was not associated with postterm birth. Evidence from one other study 8 indicated that vigorous leisure activity was weakly associated with a reduced risk of late delivery. This area of research has been largely ignored and merits further exploration to determine whether moderate or vigorous activity is related to postterm birth in other populations and, if so, what the dose-response effect might be. However, the routine use of medical intervention to truncate pregnancies deemed to be too long remains a limitation in this research.
Several limitations of this study should be noted. Despite efforts to correctly ascertain preterm birth, some amount of misclassification is inevitable. 22,23 Furthermore, the significance of the preterm birth classifications (eg, spontaneous or medically indicated) is of uncertain etiologic significance. 24 The statistical power to detect differences was reduced when we conducted stratified analyses. Because vigorous leisure activity was assessed before delivery, any misclassification is likely to be random with respect to preterm delivery. However, second trimester activity might be better recalled than first trimester and prior pregnancy activity, because all three time periods were assessed at the same time, during the telephone interview after recruitment, at 24–29 weeks’ gestation. Additionally, this study examined participation only in vigorous leisure activities. The women may have participated in other leisure activities, such as walking, that require light- or moderate-intensity effort, but were not assessed here. The questionnaire also did not adequately collect information on irregular activity.
Women who exercised during pregnancy may have better health behaviors or higher socioeconomic status. We attempted to control for these in the statistical modeling, although the possibility of residual confounding remains. Minimal confounding by measured factors was identified in our data and the scarcity of other strong risk factors for preterm birth makes it unlikely that adjustment for additional confounders would have a major impact on results. However, self-selection for engaging in vigorous activity could be strong and difficult to capture. Lastly, the study recruitment period excluded women who obtained prenatal care late or not at all, excluding some higher-risk women and limiting the generalizability of the findings.
These data from a large prospective cohort study of an obstetric population suggest that participation in vigorous leisure activity before pregnancy and during the first and second trimester did not increase the risk of preterm or postterm birth. In fact, vigorous leisure activity during first and even more so during the second trimester was related to a decreased risk of preterm birth. The favorable results obtained might be attributable to self-selection for participation in vigorous leisure activity among the healthiest women rather than a direct biological benefit. 25 Women who feel better may choose to be more active during leisure, whereas women with less healthy pregnancies might choose not to be active during leisure. To address the etiologic role of activity on pregnancy outcome and to overcome self-selection, a randomized clinical trial would be needed.
The Pregnancy, Infection, and Nutrition Study is a joint effort of many investigators and staff members whose work we gratefully acknowledge.
1. World Health Organization. WHO: recommended definitions, terminology, and format for statistical tables related to the perinatal period and use of a new certificate for cause of perinatal deaths: modifications recommended by FIGO as amended October 14, 1976. Acta Obstet Gynaecol Scand 1977; 56: 247–253.
2. Berkowitz G, Papiernik E. Epidemiology of preterm birth. Epidemiol Rev 1993; 15: 414–443.
3. Ventura S, Martin J, Curtin S, Mathews T. Births: final data for 1997. Natl Vital Stat Rep 1999; 47: 1–96.
4. Clapp J III. Pregnancy
outcome: Physical activities inside versus outside the workplace. Semin Perinatol 1996; 20: 70–76.
5. Clapp J III, Rokey R, Treadway J, Carpenter M, Artal R, Warrnes C. Exercise
. Med Sci Sports Exerc 1992; 24 (suppl 6): S294–S300.
6. Magann E, Evans S, Newnham J. Employment, exertion, and pregnancy
outcome: assessment by kilocalories expended each day. Am J Obstet Gynecol 1996; 175: 182–187.
7. Misra D, Strobino D, Stashinko E, Nagey D, Nanda J. Effects of physical activity on preterm birth. Am J Epidemiol 1998; 147: 628–635.
8. Hatch M, Levin B, Shu X, Susser M. Maternal leisure-time exercise
and timely delivery. Am J Public Health 1998; 88: 1528–1533.
9. Petridou E, Salvanos H, Skalkidou A, Dessypris N, Moustaki M, Trichopoulos D. Are there common triggers of preterm deliveries? Br J Obstet Gynecol 2001; 108: 598–604.
10. Shea K, Wilcox A, Little R. Postterm delivery: a challenge for epidemiologic research. Epidemiology 1998; 9: 199–204.
11. Savitz D, Dole N, Williams J, et al
. Determinants of participation in an epidemiologic study of preterm delivery. Paediatr Perinat Epidemiol 1999; 13: 114–125.
12. Zhang J, Savitz D. Exercise
among U.S. women. Ann Epidemiol 1996; 6: 53–59.
13. Ainsworth B, Haskell W, Whitt M, et al
. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 2000; 32 (suppl 9): S498–S516.
14. Block G, Woods P. Validation of a self-administered diet history questionnaire using multiple diet records. J Clin Epidemiol 1990; 43: 1327–1335.
15. Siega-Riz A, Herrmann T, Savitz D, Thorp J. Frequency of eating during pregnancy
and its effect on preterm delivery. Am J Epidemiol 2001; 153: 647–652.
16. Taipale P, Hiilesmaa V. Predicting delivery date by ultrasound and last menstrual period in early gestation. Obstet Gynecol 2001; 97: 189–194.
17. Savitz D, Blackmore C, Thorp J. Epidemiologic characteristics of preterm delivery: etiologic heterogeneity. Am J Obstet Gyn 1991; 164: 467–471.
18. Alderman B, Zhao H, Holt V, Heather Watts D, Beresford S. Maternal physical activity in pregnancy
size for gestational age. Ann Epidemiol 1998; 8: 513–519.
19. Berkowitz G, Kelsey J, Holford T, Berkowitz R. Physical activity and risk of spontaneous preterm delivery. J Reprod Med 1983; 28: 581–588.
20. American College of Obstetricians and Gynecologists. Exercise
and the postpartum period. ACOG Technical Bulletin Number 189–February 1994. Int J Gynaecol Obstet 1994; 45: 65–70.
21. Masse L, Ainsworth B, Tortolero S, et al
. Measuring physical activity in midlife, older, and minority women: issues from an expert panel. J Womens Health 1998; 7: 57–67.
22. Kramer M, McLean F, Boyd M, Usher R. The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. JAMA 1988; 260: 3306–3308.
23. Gjessing H, Skjaerven R, Wilcox A. Errors in gestational age: evidence of bleeding early in pregnancy
. Am J Public Health 1999; 89: 213–218.
24. Klebanoff M, Shiono P. Top down, bottom up, and inside out: reflections on preterm birth. Paediatr Perinat Epidemiol 1995; 9: 125–129.
25. Hatch M, Shu X, McLean D, et al
. Maternal exercise
, physical fitness, and fetal growth. Am J Epidemiol 1993; 137: 1105–1114.
26. Institute of Medicine. Nutrition During Pregnancy
: Part I Weight Gain; Part II Nutrient Supplements. Washington DC: Committee on Nutritional Status During Pregnancy
and Lactation, Food and Nutrition Board, National Academy Press, 1990.