Physical activity in pregnancy: a Norwegian-Swedish mother-child birth cohort study

BACKGROUND Physical activity during pregnancy is important for maternal and offspring health. Optimal conditions during pregnancy may help reduce the burden of noncommunicable diseases. National and international guidelines recommend at least 150 minutes of physical activity of at least moderate intensity per week. To optimize physical activity in pregnant women, it is important to identify factors associated with higher levels of physical activity. OBJECTIVE This study aimed to explore types and levels of physical activity in midpregnancy in Norway and Sweden and to identify factors associated with higher levels of physical activity. MATERIALS AND METHODS From the population-based mother-child cohort Preventing Atopic Dermatitis and Allergies in Children study recruiting 2697 women in Norway and Sweden from 2014 to 2016, we included 2349 women who answered an electronic questionnaire at enrollment in midpregnancy. Women were asked about regular physical activity in the last 2 weeks of pregnancy and afterward for types and levels of physical activity in pregnancy and before pregnancy and socioeconomic status, lifestyle, and maternal health. Logistic regression analyses were used to identify factors associated with higher levels of physical activity in pregnancy, defined as >30 minutes per session of ≥2 times per week of moderate- or high-intensity brisk walking, strength training, jogging, and bicycling. RESULTS No regular physical activity during the last 2 weeks before answering the questionnaire at midpregnancy was reported by 689 women (29%). In this study, 1787 women (76%) reported weekly strolling during pregnancy. Regular physical activity at least twice weekly in the first half of pregnancy was reported as brisk walking by 839 women (36%), bicycling by 361 women (15%), strength training by 322 women (14%), and other activities by <10% of women. Among the 1430 women with regular moderate- or high-intensity physical activity, the estimated median duration per week was 120 minutes. Higher physical activity levels were achieved in 553 women (23.5%) by brisk walking, 287 women (12.2%) by strength training, 263 women (11.2%) by bicycling, and 114 women (4.9%) by jogging. Higher physical activity levels were positively associated with regular physical activity before pregnancy, dog ownership, and atopic dermatitis and negatively associated with higher body mass index, study location in Østfold, previous pregnancy or pregnancies, non-Nordic origin, suburban living, and sick leave. CONCLUSION At midpregnancy, 29% of women were inactive, and less than 50% of women had at least 2 hours of moderate-intensity physical activity weekly. Awareness of physical activity in pregnancy should be discussed at pregnancy follow-up visits, particularly among women with higher body mass index, sick leave, previous pregnancy or pregnancies, and non-Nordic origin.

RESULTS: No regular physical activity during the last 2 weeks before answering the questionnaire at midpregnancy was reported by 689 women (29%). In this study, 1787 women (76%) reported weekly strolling during pregnancy. Regular physical activity at least twice weekly in the first half of pregnancy was reported as brisk walking by 839 women (36%), bicycling by 361 women (15%), strength training by 322 women (14%), and other activities by <10% of women. Among the 1430 women with regular moderate-or high-intensity physical activity, the estimated median duration per week was 120 minutes. Higher physical activity levels were achieved in 553 women (23.5%) by brisk walking, 287 women (12.2%) by strength training, 263 women (11.2%) by bicycling, and 114 women (4.9%) by jogging. Higher physical activity levels were positively associated with regular physical activity before pregnancy, dog ownership, and atopic dermatitis and negatively associated with higher body mass index, study location in Østfold, previous pregnancy or pregnancies, non-Nordic origin, suburban living, and sick leave. CONCLUSION: At midpregnancy, 29% of women were inactive, and less than 50% of women had at least 2 hours of moderate-intensity physical activity weekly. Awareness of physical activity in pregnancy should be discussed at pregnancy follow-up visits, particularly among women with higher body mass index, sick leave, previous pregnancy or pregnancies, and non-Nordic origin.
Key words: bicycling, brisk walking, maternal exercise, maternal health, mother-child birth cohort, physical activity, pregnancy, risk factors, strength training O ptimal conditions during pregnancy may reduce the burden of noncommunicable diseases in the offspring, 1−3 and modifiable lifestyle factors during pregnancy have been associated with maternal and offspring health. Physical activity (PA) in pregnancy is considered beneficial and safe for the mother and fetus 3−6 and seems to reduce the risk of cesarean delivery, 7,8 hyperemesis gravidarum, 9 gestational diabetes mellitus, hypertensive disorders of pregnancy, excessive gestational weight gain, lumbopelvic pain, and preterm birth. 5,10 Avoiding these adverse pregnancy outcomes seemed to have benefits in women's future health. 11−15 Patterns of PA among pregnant women have been previously studied 16−19 ; however, this knowledge needs to be regularly updated to continually enhance maternity care. Brisk walking and swimming were the most commonly performed exercises in the Avon Longitudinal Study of Parents and Children (ALSPAC) 16 ; however, a recent Danish cross-sectional study found that bicycling was the most prevalent activity, followed by brisk walking. 17 Recommendations for PA during pregnancy are available in many countries. 20 To improve health-related outcomes, pregnant women should perform at least 150 minutes of moderate-to high-intensity exercise per week. 3,5 These recommendations were implemented in Norway in 2019. 21 Nonpregnant adults have similar recommendations, with the addition that ≥150 minutes of moderate-intensity PA may be replaced with ≥75 minutes of high-intensity PA. 22 To optimize PA in pregnancy, it is important to identify factors associated with lower and higher levels of PA. Primiparity and normal weight have been shown to increase the likelihood of higher PA levels in Scandinavian studies, 17,18 whereas in United Kingdom studies, dog ownership has been associated with more regular PA in pregnancy. 23 Several other factors have previously been explored, such as maternal age, education, regular PA before pregnancy, body mass index (BMI), and smoking, with differing results. 19 In this study, the primary aim was to explore the types and levels of PA reported in midpregnancy in Norway and Sweden, and the secondary aim was to identify factors associated with higher levels of the most commonly performed physical activities in pregnancy.

Study design
Data from the Preventing Atopic Dermatitis and Allergies in Children (Pre-ventADALL) study, 24 a Scandinavian general population-based mother-child birth cohort, enrolling 2697 women from December 2014 to October 2016, were used in this substudy on PA in pregnancy.
Pregnant women were recruited during the routinely offered 18 weeks' gestation ultrasound examination and enrolled in the PreventADALL study at AJOG MFM at a Glance Why was this study conducted? Physical activity (PA) in pregnancy is important for maternal and offspring health. This study aimed to explore types and levels of midpregnancy PA and factors associated with higher PA levels in women participating in a Scandinavian mother-child birth cohort.

Key findings
Strolling, brisk walking, strength training, and bicycling were most commonly performed at midpregnancy; furthermore, 29% reported no PA in the last 2 weeks before inclusion. Less than 50% reported 120 minutes or more of at least moderate-intensity PA per week. Higher levels of PA were significantly associated (Pvalue <.05) with regular PA before pregnancy and dog ownership and negatively associated with higher body mass index, previous pregnancy or pregnancies, non-Nordic origin, suburban living, and sick leave.

What does this add to what is known?
In this recently established mother-child birth cohort, most women performed less than 2 hours of moderate-intensity PA weekly. Our study points to the need to discuss PA during pregnancy follow-up visits.
Original Research ajog.org Oslo University Hospital, Østfold Hospital Trust, Norway, and Karolinska University Hospital, Stockholm, Sweden. All women attending the 18 weeks' gestation routine ultrasound examination at 1 of the participating facilities were invited to participate by letter of invitation attached to the appointment letter and information about the study by the midwife or the study personnel at the maternity clinic. After the ultrasound examination, women were invited to the study facility where they received further information from the study team before enrollment.
The inclusion criteria for the Preven-tADALL study were gestational age (GA) of 16 to 22 weeks at the time of the routine ultrasound examination, singleton or twin pregnancy, no severe fetal disease, and proficiency in the Scandinavian language.
At the enrollment visit, all women signed informed consent forms, followed by a brief interview; measurements of weight, height, and blood pressure; recording of ultrasound examination data; and information about study participation. The women were asked to complete a detailed electronic questionnaire (e-questionnaire) shortly after enrollment, 24 which provided the basis for this study. The e-questionnaire was sent by email, followed by 1 reminder the following week if there was no response, ensuring 1 response only.

Study population
In this study, we included 2349 women (87%) who returned the e-questionnaire associated with enrollment. The 351 nonresponding women (13%) were similar to those included in the study in age, parity, and BMI (Table 1).

Physical activity
The women were first asked if they had been regularly physically active during the last 2 weeks of pregnancy before answering the questionnaire. All subsequent questions were related to PA typically performed during the pregnancy and the average frequency for each of the following activities: strolling, brisk walking, jogging, bicycling, strength training, aerobics, skiing, ballgames, swimming, horse riding, yoga or pilates, and other types of PA. The frequency alternatives were rarely or never, 1 to 3 times a month, once a week, 2 to 3 times a week, 4 to 5 times a week, 5 to 6 times a week, every day, and more than once per day. Regular PA before pregnancy was defined as 1 or more PAs per week with a duration of at least 20 minutes. The women were asked to compare their current level of PA during the pregnancy with their PA level before pregnancy.
The exercise intensity was recorded with the question, "How intensively do you usually exercise (so far in pregnancy)?" with the following mutually exclusive categories: no sweating or shortness of breath (low intensity), sweaty and some shortness of breath (moderate intensity), or very sweaty and very heavy breathing (high intensity).
The duration of a typical PA session was reported as <30 minutes, 30 to 60 minutes, 1 to 2 hours, or more than 2 hours. The questions on intensity and duration were based on validated questions from the Akershus Birth Cohort study, 18 a Norwegian cohort study by Haakstad et al 25 and the Norwegian Mother and Child Cohort study. 26 The questions were later somewhat modified by our research team during the development of the questionnaire.
Prepregnancy weight was self-reported at the enrollment visit, where current weight was measured and recorded as kilograms with 1 decimal point. Height was measured using a standardized stadiometer. Pregnancy in gestational week was estimated on the basis of fetal femur length, as previously reported. 24 Outcomes, definitions, and explanatory variables For the primary aim, the outcomes were frequency, duration, and intensity of the reported types of PA. The general activity level for each woman was estimated among women who reported activity of at least moderate intensity by adding the numbers of reported PA sessions per week and multiplying by exercise duration in minutes. Strolling, being a low-intensity activity, was excluded. Because the frequency and duration of PA was reported with a range, we calculated both the minimum and maximum numbers of minutes of PA per week. The Supplemental Information section provides further details.
Higher PA levels used in the secondary aim required PA at least 2 to 3 times a week, performed with a duration of ≥30 minutes at moderate or high intensity. Women were categorized into higher PA level for each of the 4 activities most commonly reported at least twice a week: brisk walking, bicycling, strength training, and jogging.
For each of the 4 higher level PAs, we included the following possible explanatory variables in the regression model: age, prepregnancy weight, BMI and weight gain at 18 weeks of pregnancy, marital status (cohabitant and married combined into 1 category), previous pregnancy or pregnancies, education, family income, country of origin, living environment, regular PA before pregnancy, current dog and/or cat ownership, current sick leave, smoking and/or snus use in pregnancy, doctor-diagnosed asthma, doctor-diagnosed atopic dermatitis (AD), and/or doctor-diagnosed allergic rhinitis (AR).

Statistical analysis
The descriptive results were given as percentages of women reporting the respective activities; number (n) was listed for each activity. For univariate and multivariate analyses, missing data were set to 0, assuming that missing response reflected lack of performing the relevant activity.
To identify factors associated with higher levels of the 4 most commonly performed PAs in pregnancy, we performed univariate logistic regression analysis for potential covariates, retaining all variables with global P value of ≤.05 and categorical P values of ≤.2 in the final multivariate logistic regression model. The significance level was set to 5%. We

Background characteristics
The mean age of the 2349 women included in this study was 32.4 years, the mean BMI at inclusion was 25, most women had higher education, and approximately half of the women were nulliparous (Table 1).

Physical activity levels
At midpregnancy, 689 women (29%) did not report any PA during the last 2 weeks of pregnancy before answering the questionnaire. General activity level was estimated for 1430 women (60.9%) reporting moderate-or high-intensity activity. Based on the minimum number of minutes of PA per week (Supplemental Table 1), the median number of active minutes was 120 minutes, with an estimated 386 women (27.0%) performing PA for ≥150 minutes per week (Supplemental Figure 2). Using the maximum estimates (Supplemental Table 1), 711 women (49.7%) were estimated to perform PA of at least 150 minutes per week.

Secondary aim
For each of the 4 activities most commonly performed at least twice weekly, we calculated that higher PA levels were achieved in 553 women (23.5%) by brisk walking, 287 women (12.2%) by strength training, 263 women (11.2%) by bicycling, and 114 women (4.9%) by jogging.
The results of the univariate analyses of factors associated with higher levels of brisk walking, bicycling, strength training, or jogging are shown in Supplemental

Principal findings
At midpregnancy, 29% of women reported no regular PA. Furthermore, apart from strolling, the most common PA reported at least twice weekly during pregnancy was brisk walking (36% of women), followed by strength training, bicycling, and jogging. The most commonly performed PAs with higher levels in intensity and duration were brisk walking, bicycling, strength training, and jogging. Women reporting regular prepregnancy PA and dog ownership had higher levels of PA, whereas higher BMI, previous pregnancy or pregnancies, non-Nordic origin, living in Østfold county, and being on sick leave were negatively associated with higher levels of PA in pregnancy.

Results
The frequency of the different activities performed during pregnancy through enrollment in our study is partly in line with other studies. In the Danish National Birth Cohort, including 88,000 pregnancies, approximately one-third of the women reported some type of exercise during early pregnancy or midpregnancy, most often as low-impact activities, such as swimming or bicycling. 27 Geographic, topographic, and cultural differences may partly explain the differences in preferred PA. For example, bicycling may be favored in countries with flat topography, favorable climatic factors, and traffic facilitation.
Our finding that 13% of women reported strength training at least twice per week was higher than the 8% of Original Research ajog.org women reported in a Danish cross-sectional study of almost 8000 pregnant women in their first trimester of pregnancy 17 and the 0.3% of women in the ALSPAC study reported at 18 weeks' gestation. 16 Jogging at least twice weekly was less common in the PreventADALL study (5%) compared with the Danish study (10%) 17 but was more common than in the ALSPAC study (0.3%). 16 Less than 10% of the women in our study reported other PAs performed at least twice weekly, in line with both the Danish and ALSPAC studies. 16,17 The higher PA levels observed among 5% (jogging) to 24% (brisk walking) of the women in our study were not directly comparable with studies assessing the proportion of women reaching national recommendations. A Danish study reported that 38% of pregnant  16 Our estimates suggested that at least 27% and no more than 50% of women performed 150 minutes or more of PA with moderate or high intensity in our study. This is higher than the 15% of women at 32 weeks' gestation who performed ≥20 minutes of moderate-intensity activity at least 3 times per week in the Norwegian Akershus Birth Cohort study, which included 3482 women. 18 This may in part be explained by differences in data collection methods and response categories and, more importantly, by the different duration of pregnancy. Gjestland et al 18 reported that primiparity, higher education (college or university), and prepregnancy BMI of <30 were associated with increased probability of meeting the national guidelines of 20 minutes of moderateintensity PA ≥3 times per week. Women in the PreventADALL study who were physically active before pregnancy were more likely to have higher levels of PA (≥30 minutes ≥2 times per week of at least moderate intensity) during pregnancy, in line with previous studies. 17,19 Dog ownership more than doubled the likelihood of higher levels of brisk walking in our study, supported by 50% increased likelihood in the ALSPAC study 23 and a recent multinational crosssectional study showing that dog owners walked more and spent more time in outdoor environments. 28 The Danish cross-sectional study conducted in 2012−2014 identified the following risk factors for not meeting the recommendations of daily PA of 30 minutes at moderate intensity during pregnancy: lack of exercise before pregnancy, being overweight, <4 years of higher education, not being proficient in the Danish language, multiparity, a previous miscarriage, smoking before pregnancy, and becoming pregnant after assisted reproductive technology. 17 In line with previous studies, higher pregnancy BMI 16−18 and previous pregnancy or pregnancies 16−19 were associated with decreased likelihood of higher levels of PA. The reduced likelihood of higher PA levels by women of non-Nordic origin agrees with the Danish findings. 17 Being on sick leave was associated with reduced likelihood of higher PA levels by strength training and bicycling.
In contrast to other studies, [16][17][18][19]27 neither education nor age was significantly associated with higher levels of the 4 most commonly performed PAs. However, these findings were supported by a Portuguese study, including 133 women during the first 2 trimesters of pregnancy. 29 Our cohort was somewhat biased in terms of education, as more than 50% of the women had ≥4 years of higher education; however, the women in the previous Norwegian, 18 Danish, 17 and ALSPAC 16 studies had similar educational levels.
Higher levels of PA were not significantly associated with doctor-diagnosed asthma or AR. This may suggest that mild or well-regulated allergic disease does not limit PA. However, we did find that doctor-diagnosed AD was positively associated with higher levels of jogging. To the best of our knowledge, this is a novel finding with unclear implications. A recent study from the United States 30 found that AD was associated with less PA in US adults, whereas a systematic review from 2016 31 found insufficient evidence to conclude whether AD was associated with more or less PA. In addition, we are not aware of any previous studies reporting higher levels of jogging in women with AD.

Strengths and limitations
This study offered study participation to all pregnant women at 16 to 22 weeks' gestation who attended the national routine fetal ultrasound screening in their midtrimester of pregnancy. Unintentionally, the enrolled study population had higher education attainment, had slightly higher age than the national average, and was predominantly of Norwegian and Swedish origins but is relatively representative of city populations. 32,33 Furthermore, our population matched that of other similar cohort studies in terms of age, parity, education, and income levels. 17,18 The skewing of our population toward higher education may have bearings on the generalizability of our results, with PA observed in our study possibly overestimating that of the general population.
However, if our finding that 29% of the women had been inactive at midpregnancy is an overrepresentation, the general population may be even less active than the population in the Pre-ventADALL study.
A limitation in the PreventADALL study was that women without sufficient Norwegian or Swedish language skills were excluded from participation. Therefore, our study is not generalizable for some minority populations. Furthermore, the questionnaire was not appropriate for direct comparison with the current Norwegian guidelines published in April 2019. The information on PA in pregnancy included questions in line with those reported by Haakstad et al 25 in a Norwegian pregnancy cohort but were modified to fit our electronic questionnaire. Because of the study design, it was not feasible to include accelerometer or other objective measures of PA; therefore, the data presented in this article were exclusively self-reported. Contraindications for PA were not explored in this study, as the study population consisted of relatively healthy women, pregnant with 1 or 2 fetuses.

Clinical implications
Despite the acknowledged benefits to maternal and offspring health by regular PA in pregnancy, our data showed that less than 50% of the women were regularly active at a high level in midpregnancy. This pointed to a need to address the importance of PA during pregnancy follow-up visits.

Research implications
The potential benefits of high levels of PA in pregnancy for the mother and her offspring in terms of noncommunicable disease development need further investigations, as do the potential effects of suboptimal levels of PA in pregnant women.

Conclusion
At midpregnancy, almost one-third of women reported no regular PA in the last 2 weeks before answering the ajog.org Original Research questionnaire, whereas less than 50% of women had 2 hours or more of regular moderate-intensity PA per week during pregnancy. The most common activities performed at least twice weekly were brisk walking, bicycling, and strength training. Being physically active before pregnancy, owning a dog, and having AD were associated with higher levels of 1 or more of the most commonly performed physical activities. Awareness of PA in pregnancy should be discussed at pregnancy followup visits, particularly among women with higher BMI, on sick leave, with previous pregnancy or pregnancies, and of non-Nordic origin-groups who often do not reach higher levels of PA. &