Employment, exertion, and pregnancy outcome: Assessment by kilocalories expended each day

Everett E Magann, MD," Sharon E Evans, PhD, b and Jolm P. Newnham, MD b Jackson, Mississippi, and Subiaco, Western Australia, Australia
OBJECTIVE= Our purpose was to evaluate the influence of employment and physical exertion on pregnancy outcome as quantified by kilocalories expended each day. STUDY DESIGN= This prospective study assessed 2743 pregnant women who received prenatal care and were delivered at the major perinatal center in Western Australia between May 1989 and November 1991. All women completed an extensive questionnaire on their social, medical, psychosocial, and economic circumstances. The women were allocated to five groups on the basis of kilocalorie expenditure per day. RESULTS= The characteristics of women in each energy expenditure group were different, with those in the least-expenditure group being younger and shorter, more likely to be living in worse socioeconomic conditions, smoke cigarettes, be nulliparous, and to be of an ethnic group other than white. After confounding effects were adjusted, women in the medium energy expenditure group were delivered of babies of higher birth weight than were women in other groups. However, the differences in birth weight between the energy expenditure categories were small, and mean birth weights within each group were within the normal range. Women in the medium energy expenditure group also had fewer incidences of prelabor rupture of membranes and women in the lower energy expenditure category had increased risks of antepartum admission to the hospital and preterm birth. A variety of other differences were observed in pregnancy outcomes for women in each of the categories of energy expenditure, but most of these differences were explained by the characteristics of the women in each expenditure level rather than the exercise pattern itself. CONCLUSION= These results indicate that the effects of daily energy expenditure on pregnancy outcome are not great. Enthusiasm for counseling pregnant women of the benefits or hazards of extremes in daily activity should be tempered by the relative lack of an effect and the fact that most apparent differences are due to confounding variables rather than the exercise itself. (Am J Obstet Gynecol 1996;175:182-7.)

Key words: Employment, exertion, pregnancy outcome

The associations between employment, exercise, and pregnancy have become areas of important investigation now that working women are a large percentage of our workforce and many physically fit women continue to exercise regularly during pregnancy. Should employment be modified during pregnancy, and if so, at what gestational age should changes be made, and for which particular job characteristics? Can pregnant women continue to exercise and perform routine activity or must they modify their activities at particular gestational ages? Standing >8 hours each day has been associated with increased risks of preterm birth, preterm labor, and low

From the Department of Obstetrics and Gynecology, University of Mississippi Medical Centeg,a and the Foundation for Women's and Infants" Health, University Department of Obstetrics and Gynaecology, King Edward Memorial Hospital for Women.b Receivedfor publication August 21, 1995; revised September26, 1995; acceptedDecember28, 1995. Reprint requests:JohnP. Newnham, MD, Foundation for Women's and Infants' Health, University of Western Australia at King Edward Memorial Hospital for Women, 374 Bagot Road, Subiaco 6008, Western Australia. Copyright 1996 by Mosby-YearBook, Inc. 0002-9378/9655.00+ 0 6/1/71590
182

birth weight. 1-3 Moreover, reduction in physical exertion has been shown to decrease the incidence of preterm birth. ~6 The influence of exercise on pregnant physically fit women appears to either not have an adverse effect or to promote a beneficial outcome. In physically fit women exercise is not associated with early pregnancy loss. ~ Women continuing endurance exercise during pregnancy, when compared with physically fit women who do not continue exercises, have lower rates of operative abdominal and vaginal deliveries, a reduction in the duration of active phase labor, and less fetal distress, s Previous investigators have analyzed the effects of exercise and employment on pregnancy outcome primarily by use of patient surveys and questionnaires. Quantification of the exercise and employment in these studies was by terms such as mild, moderate, and heavy or sedentary and active. The divergence in conclusions from these studies p r o b a b l y results from failure to evaluate confounding variables, the retrospective design of many of the studies, and the absence of quantification of energy output of employment. 1' 9, lo To assess the effects of daily activity on pregnancy outcome, employment energy expended must be added to daily activity and should in-

Volume 175, Number 1 AmJ Obstet Oynecol

Magann, Evans, and Newnham 183

clude domestic work at home, leisure ventures, and athletic activity. The assessment of employment and activity levels in pregnancy, measured in kilocalorie usage and exertion, in a prospective inquiry has not previously been undertaken. The purpose of this investigation was to evaluate the influence on pregnancy outcome of exertion measured in kilocalorie usage. Material and methods This prospective study involved 2743 singleton pregnancies of women attending the prenatal clinic at the King Edward Memorial Hospital for Women, Subiaco, Western Australia, between May 1989 and November 1991. A full description of the database involving this cohort of women has been published previously.11Briefly, the inclusion criteria consisted of (1) maternal age _>18 years, (2) proficiency in English to understand the implications of participation, (3) an expectation of delivery at this hospital, and (4) intention to remain in Western Australia so that childhood follow-up was possible. The study was approved by the Institutional Ethics Committee. Women were enrolled between 16 and 18 weeks' gestation by a team of research midwives. These midwives assisted the women in the completion of their questionnaires and examined each questionnaire for completeness. Follow-up information was obtained by the nursemidwives during the pregnancy and postpartum period. Validation of the exercise component of these self-administered questionnaires has previously been confirmed. '2 This questionnaire evaluated the women's social, psychosocial, a n d economic circumstances; past medical and obstetric history; daily activity; and occupation, including a detailed job description and n u m b e r of hours worked per week so that daily kilocalorie use and leisure time physical activity could be calculated) ~ Maternal employm e n t was categorized by the descriptive title of the occupation and a brief explanation included on each questionnaire. All employed women were further evaluated by a daily diary with a detailed account of their activity at work and at leisure. Their kilocalorie use and activity levels were based on the combination of these judgments. Calculations were all performed by one of the authors (S.E.), who was blinded to pregnancy outcome. There were four peaks of energy expenditure, at 2100, 2350, 2500, and 2800 kcal/day. Consequently, the women were arbitrarily subdivided into five groups. Group 1 contained women whose daily kcal output was _<2300kcal/day; group 2, 2301 to 2500 kcal/day; group 3, 2501 to 2700 kcal/day; group 4, 2701 to 2900 kcal/day; and group 5, >2900 kcal/day. Some examples of energy expenditure follow: (1) <2300 kcal/day: a woman not working outside the home, with no children and no formal exercise other than 1 to 2 hours of general housework; (2) 2500 to 2700 kcal/day: a woman not working outside the home, with one child <5 years old, doing 2 hours of housework, 2

hours of general child care, and a further 2 hours of other mild activities such as walking or gardening; (3) >3000 kcal/day: a nurse working 8-hour shifts in a work environm e n t also doing 1 to 2 hours housework and a further 1 to 2 hours of aerobics or r u n n i n g each day. Outcomes measured included antenatal hospital admissions, antepartum hemorrhage, occurrence of hypertension (blood pressure >140/90 mm Hg or 30 m m increase in systolic or 15 mm increase in diastolic blood pressure after the twentieth week of pregnancy with proteinuria or edema) and diabetes (abnormal 3-hour oral glucose tolerance test result or pregnancy requiring insulin for glucose control), preterm deliveries (<37 weeks' gestation), prelabor rupture of membranes, intrauterine growth restriction according to the 3rd and 10th percentiles (based on nomograms for Western Australia that account for maternal height, parity, and fetal genderl~), abnormal fetal heart rate tracings in labol, duration of labor, mode of delivery, meconium-stained amniotic fluid at delivery, 1- and 5-minute Apgar scores, need for neonatal resuscitation, and admission to the neonatal intensive care unit. Gestational age was calculated from the date of the last menstrual period and an ultrasonographic examination at 18 weeks' gestation. The socioeconomic index, used to quantify socioeconomic status, was based on maternal and paternal levels of education, employment, family income, and marital status. High socioeconomic stares was inversely correlated to the socioeconomic score. The relatively low levels of this score and the narrow distribution reflects the middle-class nature of this population. The overall stillbirth rate was 0.8% and in the five groups included 4 (1.1%), 4 (0.7%), 7 (0.7%), 6 (1.0%), and 1 (0.6%) cases, respectively. These fetal loss rates were evenly distributed across the groups, and all subsequent analyses were based only on pregnancies resulting in live births. Univariate analyses on numeric variables were calculated by analysis-of-variance F tests with Bonferroni inequality for contrasts and Kruskal Wallis Z2 tests as appropriate. Group testing of categoric data was by Z2 test or MantelHaenszel Zz tests for linear association of ordinal variables. Individual cells within categoric tables were tested for significance with Z2 distribution with 1 degree of freedom. Logistic regression modeling was performed for binary outcome measures, and multivariate analysis of variance was used for parametric outcomes. Results Table I displays demographic and lifestyle data of the study population. The characteristics of women in each group were different, particularly those in group 1, who did the least exercise. These women were significantly younger and shorter, had a higher socioeconomic score

184 Magann, Evans, and Newnham

July 1996 AmJ Obstet Gynecol

Table I. D e m o g r a p h i c and lifestyle characteristics of study p o p u l a t i o n

Group 1 (<2300 kcal)

No. of patients Maternal age (yr) Maternal height (cm) Maternalweight (kg) Weightgain (18-34wk)(kg) Parity 0 1 >1 Race (%) White Asian Aboriginal Other Socioeconomic scale Smoking 0 0-10 ->10 Kilocalorie expenditure 359 23.5 162.8 59.2 9.0 84.7 10.9 4.5 77.2* 9.5 6.7 6.7 2.9 (0.9)* 63.2* 22.6 14.2 2124 (53) (6.0)* (6.7)* (12.5) (3.7)

Group 2 ] Group 3 Croup 4 (2301-2500kcal) I (2501-2700kcal) (2701-2900 kcal)

545 26.9 163.8 59.0 9.2 78.7 14.9 6.4 88.4 6.2 2.0 3.3 1.9 (1.0) 75.8 15.4 8.8 2359 (43) (5.4) (6.8) (11.7) (3.7) 1059 29.0 163.4 60.7 8.3 10.8" 47.3 41.9 87.8 3.7 2.5 6.0 2.3 (1.0) 74.9 11.6 13.5 2566 (34) (5.6)* (6.7) (13.4)* (3.7)* 608 27.0 163.9 59.5 9.0 64.6 20.7 14.6 92.9 3.1 0.7 3.3 1.9 (1.0) 73.9 17.3 8.9 2812 (26) (5.8) (6.3) (11.1) (3.6)

Group 5 (>2900kcal) I Significance

172 27.2 164.3 60.0 9.3 42.4 29.7 27.9 p = 0.0001 92.4 2.9 1.2 3.5 1.0 (0.9) 71.5 18.6 9.8 3112 (231) (5.8) (5.8) (10.2) (3.7) p= 0.0001 p= 0.045 p= 0.076 p = 0.008 p = 0.0001

p= 0.0001 p = 0.0001

p= 0.0001

Data are number, mean and SD, or percentage, as appropriate. *p < 0.05 by Bonferroni correction to Ftest or by cell X2 test.

( p o o r e r living conditions), smoked m o r e cigarettes, were m o r e likely nuUiparous, and were m o r e likely to be of an ethnic g r o u p o t h e r than white. Nearly 40% of the w o m e n were in g r o u p 3, which described a daily energy expenditure between 2501 and 2700 kcal. These w o m e n were older, heavier before pregnancy, had less weight gain d u r i n g pregnancy, and 90% were n o t having their first child. A n t e p a r t u m events are shown in Table II. A n t e p a r t u m admissions for p r e g n a n c y complications were significantly m o r e f r e q u e n t in g r o u p 1 c o m p a r e d with the o t h e r f o u r groups ( p = 0.027). Logistic regression m o d e l i n g controlling for m a t e r n a l age, p r e p r e g n a n c y weight, soc i o e c o n o m i c score, smoking, and parity c o n f i r m e d the w o m e n in g r o u p 1 were m o r e likely to have had an antenatal admission than those in g r o u p 3 (odds ratio 1.71, c o n f i d e n c e interval 1.19 to 2.46). T h e rates of admission in the o t h e r groups were n o t significantly different f r o m each other. T h e r e were no significant differences observed in the i n c i d e n c e of a n t e p a r t u m h e m o r r h a g e a m o n g the five groups. Prelabor r u p t u r e of m e m b r a n e s was seen less frequently in g r o u p 3 (p = 0.003) c o m p a r e d with the o t h e r f o u r groups; this r e m a i n e d a t r e n d even w h e n parity and s o c i o e c o n o m i c scores were controlled for (p=0.068). T h e r e was also a significant r e d u c t i o n in pregnancy-ind u c e d hypertension in g r o u p 3 (p = 0.001), a l t h o u g h this effect was entirely a c c o u n t e d for by differences in the w o m e n within this g r o u p in parity, smoking, prepregnancy weight, and age. T h e frequency of gestational diabetes was similar in each of the five groups. I n t r a p a r t u m events are shown in Table III. T h e r e was a

small statistically significant but clinically insignificant increase in gestational age at birth with increasing exercise level ( p = 0.009). T h e significance of this finding rem a i n e d even after multivariate m o d e l i n g that controlled for age, parity, an d smoking. T h e r e were no overall differences in the frequencies of p r e t e r m birth or p r e t e r m labor a m o n g the five groups. However, w h e n analyzed by logistic regression m o d e l i n g to control for socioecon o m i c score, parity, and weight gain, groups 1 and 2 had h i g h e r rates of p r e t e r m birth c o m p a r e d with the o t h e r three groups (p = 0.006, adjusted odds ratio 1.61 [1.15 to 2 . 2 6 ] ) and g r o u p 1 had an adjusted odds ratio for pret e r m labor of 2.66 (95% c o n f i d e n c e interval 1.20 to 5.90) c o m p a r e d with g r o u p 3. T h e r e were no differences between groups 3, 4, or 5. Breech presentation on admission to labor and delivery was m o r e f r e q u e n t in g r o u p 1 w o m e n c o m p a r e d with the o t h e r f o u r groups (p = 0.066); this effect was greater in multiparous w o m e n (9.1% vs 4.1%) than in nulliparous w o m e n (5.9% vs 4.1%). G r o u p 5 had a t r e n d to a h i g h e r rate of i n d u c t i o n of labor, b o t h for nulliparous and multiparous w o m e n (p = 0.035). T h e r e were no significant differences in the rates of spontaneous vaginal delivery between the five groups o t h e r than could be attributed to differences in parity. T h e p r o p o r t i o n of nonelective cesarean sections for cephalopelvic d i s p r o p o r t i o n was h i g h e r in g r o u p 5 w o m e n ( p = 0.054), although this i n c l u d e d 6 of only 15 w o m e n in this g r o u p who r e q u i r e d nonelective cesarean delivery. A b n o r m a l i n t r a p a r t u m fetal heart rate tracings that resulted in obstetric i n t e r v e n t i o n o c c u r r e d less frequently in g r o u p 3 w o m e n (p = 0.001), a l t h o u g h this ef-

Volume 175, Number 1 AmJ Obstet Gynecol

Magann, Evans, and Newnham 185

Table

H, A n t e p a r t u m c o m p l i c a t i o n s

Group 1 (<2300 kcal)

Antepartum admissions Anteparmm hemorrhage Prelabor rupture of membranes Pregnancy-induced hypertension Gestational diabetes Data are percentage. *p < 0.05 by cell Z2 comparison. 32.6* 3.3 17.6 27.9 1.7

Group 2 (2301-2500 kcal)

25.7 2.2 23.5 26.2 1.7

Group 3 ] Group 4 | Group5 (2501-2700 kcal) t (2701-2900kcal) [ (>2900kcal) Significance

26.3 3.0 14.4" 18.9" 2.7 23.9 1.8 20.7 23.7 1.3 28.5 2.9 18.0 24.4 1.2 p = 0,027 p = 0.453 p = 0.003 p = 0.001 p = 0,229

T a b l e III, I n t r a p a r t u m events

Group 1 Group 2 Group 3 Group 4 ] Group5 (<2300 kcal) (2301-2500kcal) (2501-2700kcal) (2701-2900kcal) ] (>2900kcal) Significance
Gestational age at delivery (wk) Preterm birth Preterm labor Breech presentation on admission in labor Onset of labor Nulliparous Spontaneous Induction Multiparous Spontaneous Induction Mode of delivery Nulliparous Spontaneous vaginal delivery Vaginal (assisted) Nonelective cesarean section Elective cesarean section Multiparous Spontaneous vaginal delivery Vaginal (assisted) Nonelective cesarean section Elective cesarean section Abnormal labor FHR influencing delivery 39.1 (2.2) 10.0" 5.0* 6.4" 39.3 (2.4) 10.3" 3.7 4.6 39.1 (2.0) 8.1 4.2 3.8 39.5 (1.9)* 8.1 2.3 4.6 39.4 (2.3)* 8.1 2.3 4.6 p = 0.009 p = 0.006 p = 0.269 p = 0.066 p = 0.035 60.5 32.2 50.9 30.9 60.2 23.4 9.2 7.2 70.9 3.6 7,3 18.2 7.5 58.5 34.7 53.5 33.6 50.8 29.6 12.8 6.8 64,7 12.1 10.3 12.9 7.3 60.5 35.1 53.3 31.2 58.8 24.6 12.3 4.4 71.3 8.0 5.2 15.5 3.2* 61.1 30.8 53.0 30.7 52.4 27.5 12.0 8.1 68.8 8.8 6.0 16.3 6.7 50.7 42.5* p = 0.287 49.5 38.4* p = 0.269 50.7 24.7 17.8" 6.9 p = 0.001 72.7 13.1 2.0 12.1 7.0

p = 0.001

Data are mean and SD or percentage as appropriate, FHR, fetal heart rate. *p < 0.05 by cell Z2 comparison.

fect was e n t i r e l y d u e to d i f f e r e n c e s in parity within this group. T h e d u r a t i o n s o f t h e first, second, a n d t h i r d stages o f l a b o r are s h o w n in Table IV, T h e r e was a t r e n d to l o n g e r d u r a t i o n o f t h e first stage o f l a b o r in n u l l i p a r o u s w o m e n in g r o u p 5 c o m p a r e d with t h e o t h e r f o u r g r o u p s ( p = 0.169); this effect was statistically significant w h e n m o d e l i n g a c c o u n t e d for d i f f e r e n c e s in m a t e r n a l age (p = 0.048), T h e d u r a t i o n o f tile s e c o n d stage o f l a b o r in n u l l i p a r o u s w o m e n was l o n g e r in g r o u p s 2 a n d 4; however, after m o d e l i n g to a c c o u n t f o r c o n f o u n d i n g variables, t h e r e were n o significant d i f f e r e n c e b e t w e e n t h e five groups, e i t h e r for n u l l i p a r o u s or m u l t i p a r o u s w o m e n , in t h e d u r a t i o n o f t h e s e c o n d a n d t h i r d stages o f labor. N e o n a t a l o u t c o m e is s h o w n in Table V. B i r t h w e i g h t was significantly associated with t h e kilocalorie g r o u p i n d e p e n d e n t o f c o n f o u n d i n g variables. T h e c o n f o u n d i n g

variables k n o w n to i n f l u e n c e b i r t h w e i g h t a n d t h a t were significant at a p value o f <0.1 in this analysis were gestational age, s m o k i n g status, u l t r a s o n o g r a p h i c g r o u p allocation, sex o f t h e n e o n a t e , m a t e r n a l h e i g h t , p r e p r e g n a n c y weight, a n d parity. After t h e effects o f these conf o u n d i n g variables were a d j u s t e d for in multivariate analyses o f variance, relative to b i r t h weights in g r o u p 3, t h e a d j u s t e d m e a n w e i g h t in g r o u p 1 was 73 g m less ( p = 0.011), in g r o u p 2 60 g m less ( p = 0.017), in g r o u p 4 23 g m less ( p = 0.33), a n d in g r o u p 5 22 g m less ( p = 0.515). By u n i v a r i a t e analysis t h e b i r t h weights o f babies o f n u l l i p a r o u s a n d m u h i p a r o u s w o m e n w i t h i n e a c h kilocalorie g r o u p were statistically similar; however, after m o d e l i n g t h e r e was a n effect o f a d d i n g to the b i r t h w e i g h t 70 g m if the w o m a n was p a r a 1 a n d 120 g m if t h e parity was greater. T h e p r o p o r t i o n s o f b i r t h w e i g h t <10th a n d <3rd p e r c e n t i l e s were similar in e a c h o f t h e five

186 Magann,Evans, and Newnham

July 1996 AmJ Obstet Gynecol

Table IV. Duration of labor

Group 1 (<2300 kcal)

First stage (hr) Nulliparous Multiparous Second stage (min) Nulliparous Multiparous Third stage (min) Nulliparous Multiparous 6.2 (4.0-8.8) 4.5 (2.4-7.0) 55.0 (31.5-94.5) 23.5 (14.5-38.0) 6 (5-9) 6 (5-9)

Group 2 (2301-2500 kcal)

6.4 (4.3-10.1) 3.9 (2.7-5.5) 80.0 (4i.5,133)* 22.0 (12.0-44.0) 6 (5-8) 6 (5-10)

Group 3 (2501-2700 kcal)

6.0 (3.%8.9) 4.0 (2.5-5.8) 52.0 (26.5-105) 19.0 (10.0-35.0) 5.5 (5-8.5) 6 (5-8)

Group 4 [ (2701-2900 kcal) I

6.5 (4.5-9.5) 3.5 (2.3-5.5) 71.0 (38.5-120)* 18.5 (10.0-35.0) 5 (4-8) 6 (5-9)

Group 5 (>2900kcal)
7.4 (5.3-10.3)* 3.5 (2.5-5.8) 57.0 (30-1152) 15.0 (10.0-38.5) 6 (5-9) 6 (5-9)

Significance
p=0.169 p=0.761 p=0.001 p=0.275 p=0.116 p=0.381

Data are median and interquartile range. *p < 0.005 by Wilcoxon rank-sum test comparison. Table V, N e o n a t a l o u t c o m e

Group 1 I Group 2 Group 3 Group 4 Group 5 (<2300 kcal) '2301-2500 kcal) (2501-2700kcal) (2701-2900kcal) (>2900kcal)
Birth weight (gm) Nulliparous Multiparous SGA <3rd percentile <10th percentile Meconinm at delivery No resuscitation at delivery Apgar score <7 1 min 5 min Admitted to special care nursery 3214 (546) 3142 (700) 4.2 12.8 17.6 36.5 22.3 3.1 7.0 3278 (599) 3269 (611) 3.7 11.6 19.6 38.0 20.2 2.0 5.7 3300 (608) 3387 (586) 3.6 11.7 11.7" 47.7* 17.0 2.2 5.0 3360 (510) 3386 (549) 2.1 10.0 16.0 39.6 20.1 2.1 4.9 3397 (611) 3362 (533) 3.5 10.5 14.0 40.7 19.8 0.6 5.2

Significance
p = 0.007 p = 0.013 p = 0.422 p= 0.717 p = 0.001 p = O.005 p = 0.067 p = 0.480 p = 0.663

Data are mean and SD or percentage as appropriate. SGA, Small for gestational age. *p < 0.05 by cell ) 2 comparison, groups. G r o u p 3 pregnancies involved a lesser ,rate of m e c o n i u m - s t a i n e d amniotic fluid ( p = 0.001), a greater p r o p o r t i o n n o t requiring n e o n a t a l resuscitation (p= 0.005), and a lower i n c i d e n c e of 1-minute Apgar scores <7 (p = 0.067) ; however, each of these three differences was entirely due to the h i g h e r parity of w o m e n in g r o u p 3. T h e r e were no i n d e p e n d e n t effects of exercise levels on these measures of n e o n a t a l welfare. T h e r e were no differences a m o n g the five categories in the rate of admission to the n e o n a t a l nursery. cies in the g r o u p 3 category (2501 to 2700 kcal/day) had statistically significant reductions in the rate o f m e c o n i u m stained amniotic fluid, the p r o p o r t i o n requiring neonatal resuscitation, and a lower incidence of 1-minute Apgar scores <7. Multivariate analyses have r e v e a l e d that the app a r e n t effects of exercise are due entirely to c o n f o u n d i n g variables and n o t to the level of activity itself. It is clear that we must i n t e r p r e t with caution the results of any studies that fail to account for these i m p o r t a n t c o n f o u n d i n g variables. An ideal daily kilocalorie e x p e n d i t u r e may exist to optimize fetal growth. G r o u p 3 w o m e n e x p e n d i n g 2501 to 2700 calories per day were delivered of n e o n a t e s with a m e a n weight of 3332 gm. After c o n f o u n d i n g variables are adjusted for, the p r e d i c t e d m e a n weights of n e o n a t e s in the two groups using <2500 kcal was significantly less. W o m e n in the two highest exercise groups were also delivered of babies of lower p r e d i c t e d birth weight than observed in g r o u p 3, but the differences were n o t statistically significant. These observations on the associations between birth weight and kilocalorie category are interesting, but further research on the relationships between m a t e r n a l energy e x p e n d i t u r e and fetal growth will be r e q u i r e d before a causal link can be stated with certainty.

Comment
T h e results of this study have revealed i m p o r t a n t differential features of w o m e n in each of the five exercise groups. These differences i n c l u d e d age, height, socioecon o m i c score, smoking practice, parity, and race. In part, the energy e x p e n d i t u r e category was i n f l u e n c e d by parity because the scoring algorithm allocated a high kilocalorie c o n s u m p t i o n to those w o m e n caring for one or m o r e y o u n g children. Nevertheless, our extensive database, collected prospectively, has allowed a full statistical exploration of the effects of these variables and has shown that m a n y a p p a r e n t features of exercise e x p e n d i t u r e are due purely to c o n f o u n d i n g variables. For example, pregnan-

Volume 175, Number 1

Magann, Evans, and Newnham

t87

AmJ Obstet Gynecol

Groups 1 and 2 (<2501 kcal/day) c o m p r i s e d w o m e n who were active <4 hours in each day. In these categories w o m e n were observed to have an increased n u m b e r o f a n t e p a r t u m hospital admissions with complications of p r e g n a n c y and an increased risk o f p r e t e r m birth. Birth weight <10th and <3rd percentiles were similar in each of the five groups. These observations conflict with the findings o f previous studies in which high activity levels were m o r e likely to be associated with p r e t e r m birth and p o o r fetal growth. ~6 T h e c o m p r e h e n s i v e i n f o r m a t i o n o b t a i n e d prospectively in the c u r r e n t study on 2743 w o m e n has p e r m i t t e d a m o r e extensive statistical analysis of conf o u n d i n g variables and has allowed a b r o a d e r assessment of the o t h e r factors that influence birth weight. O n admission to labor and delivery, b r e e c h presentation was m o r e c o m m o n in the m i n i m u m kilocalorie exp e n d i t u r e women, raising the possibility that daily activity may influence the presentation of the fetus. T h e effect was greater in parous w o m e n , suggesting that low activity and lax a b d o m i n a l musculature may increase the chance of b r e e c h presentation in such circumstances. Nonelective cesarean delivery for cephalopelvic disprop o r t i o n was m o r e f r e q u e n t in the high-activity women, alt h o u g h the n u m b e r o f cases was too small for a definite conclusion to be reached. This effect could result f r o m greater fetal size because the largest fetuses were in the highest exercise group, an effect caused by the c o n f o u n d ing effects within these women. It is also possible that w o m e n who have greater exercise d u r i n g p r e g n a n c y may have a b d o m i n a l musculature that influences labor and delivery or that such w o m e n may be m o r e likely to have an a n d r o i d pelvic shape. T h e effect of h i g h levels of exercise on p r e g n a n c y o u t c o m e has b e e n controversial for m a n y years. Previous investigations have r e p o r t e d in high-activity w o m e n shorter labors and less i n t r a p a r t u m fetal distress, s' 15 whereas o t h e r studies have observed no differences in the l e n g t h of labor or A p g a r scores 16 and l o n g e r delivery times. ~7 It would a p p e a r f r o m the c u r r e n t study that the effects of energy e x p e n d i t u r e on p r e g n a n c y o u t c o m e are n o t great and that m a n y of the previously r e p o r t e d effects may have b e e n the result of c o n f o u n d i n g variables. O u r study suggests that a m e d i u m energy e x p e n d i t u r e may be advantageous, with a h i g h e r birth weight and less risk of p r e l a b o r r u p t u r e of m e m b r a n e s . W o m e n in the low-activity g r o u p were m o r e likely to require antenatal admission, had an increased chance of p r e t e r m birth, and were m o r e likely to have b r e e c h presentation at t e r m gestation. Those in the high-activity g r o u p had a greater rate of i n d u c t i o n of labor, l o n g e r d u r a t i o n of the first stage o f labor, and a t r e n d for increased chance of cesarean sec-

tion for c e p h a l o p e M c disproportion. Nevertheless, n e o natal morbidity in each o f these categories was similar, and the differences in p r e g n a n c y o u t c o m e s between the five groups were n o t great and were of questionable clinical significance. In conclusion, the results of this study suggest a med i u m energy e x p e n d i t u r e may be of some advantage in pregnancy, but the chances of deleterious effects in o t h e r categories of e x p e n d i t u r e are small. _Altering m a t e r n a l lifestyle during pregnancy to the m e d i u m energy expenditure category may be warranted, but zealous counseling to the extremes is n o t justified.

REFERENCES

1. SimpsonJL. Are physical activity and employment related to preterm birth and low birth weight? Am J Obstet Gynecol 1993;168:1231-8. 2. KlebanoffMA, Shiono PH, Cm'eyJAC. The effect of physical activity during pregnancy on preterm delivery and birth weight. AmJ Obstet Gynecol 1990;163:1450-6. 3. Teitelman AM, Welch LS, Heilenbrand KG, Bracken MB. Effect of maternal work activity on preterm birth and low birthweight. AmJ Epidemiol 1990;131:104-13. 4. Mamelle N, Laumon B, Lazar P. Premamrity and occupational activity during pregnancy. Am J Epidemiol 1984;119: 309-22. 5. Mamelle N, Munoz E Occupational working conditions and preterm birth: a reliable scoring system. Am J Epidemiol 1987;126:150-2. 6. Manlelle N, Bermcat I, Munoz E Pregnant women at work: rest periods to prevent preterm birth? Paediatr Perinat Epidemiol 1989;3:19-27. 7. Clapp JE The effects of maternal exercise on early pregnancy outcome. A m J Obstet Gynecol 1989;161:145-7. 8. Clapp JF. The course of labor after endurance exercise during pregnancy. AmJ Obstet Gynecol 1990;163:1799-805. 9. Chamberlain GV. Work in pregnancy. Am J Indust Med 1993;23:559-75. 10. Katz VL, Jenkins T, Haley L, Bowes WAJr. Catecholamine levels in pregnant physicians and nurses: a pilot study of stress and pregnancy. Obstet Gyneeol 1991;77:338-42. 11. NewnhamJP, Evans SF, Michael CA, Stanley FJ, Landau LI. Effects of frequent ultrasound during pregnancy: a randomized controlled trial. Lancet 1993;342:887-91. 12. Eskenazi B, Pearson I<2 Validation of a self-administered questionnaire for assessing occupational and environmental exposures of pregnant women. Am J Epidemiol 1988;128: 1117-29. 13. Taylor HL, Jacobs DR, Schucker B, IGmdsen J, Leon AS, Debacker G. A questionnaire for the assessment of leisure time physical activities. J Chron Dis 1978;31:741-55. 14. Blair E, Stanley FJ. Birthweight charts from a Western Australian population for singleton Caucasian live births after 21 weeks gestation. Sydney: Australian Government Publishing Service, 1985. 15. Botkin C, Driscoll CE. Maternal aerobic exercise: newborn effects. Faro Pract ResJ 1991;11:387-93. 16. Lokey EA, Tran ZV, Wells CL, Mayers BC, Tran AC. Effects of physical exercise on pregnancy outcome: a meta-analytic review. Med Sci Sports Exert 1991;23:1234-9. 17. Rice PL, Fort IL. The relationship of maternal exercise on labor, delivery, and health of the newborn.J Sports Med Phys Exerc 1991;31:95-9.