We examined the effect of light intensity resistance exercise training performed during the second and third trimester of pregnancy on the newborn's birth size. We also studied the association between maternal body weight prior to pregnancy and newborn's birth size.
Randomised controlled trial.
We randomly assigned 160 sedentary gravidae to either a training (n=80) or a control (n=80) group. The training programme focused on light resistance and toning exercises (three times per week, 35–40 min per session). We recorded the Apgar score, birth weight, birth length, and head circumference of the newborn, as well as gestational age at time of delivery from hospital perinatal records. We also measured maternal weight and height before parity and gestational weight gain.
Maternal characteristics neither differed between groups (all P>0.1) nor newborn characteristics (all P>0.1). Maternal body weight was positively and significantly associated with newborn's birth weight and length only in the control group (β=19.20 and 0.065, respectively, P<0.01).
Light intensity resistance training performed over the second and third trimester of pregnancy does not have a negative impact on the newborn's body size or overall health. Exercise interventions might attenuate the adverse consequences of maternal body weight before pregnancy on the newborn's birth size.
Body size at birth is a marker of intrauterine environment. Foetal adaptation to an adverse intrauterine environment involves programming of metabolic pathways that might predispose to cardiovascular disease in later life. This adverse environment may change gene expression, leading to physiological phenotypes associated with morbidity and mortality.1, 2 Newborns with a low birth weight have increased risk for perinatal as well as adult morbidity and mortality,2 whereas newborns with a high birth weight are at increased risk for several complications, such as shoulder dystocia, operative delivery, and birth canal lacerations.3 High birth weight is also associated with increased risk during adulthood of type 2 diabetes and some types of cancer (for example, endometrial, breast, or prostate).1, 2, 4, 5
The effect of exercise training during the course of pregnancy on the newborn's birth weight is unclear. Two meta-analyses reported that exercise during pregnancy normally does not affect birth weight,6, 7 but a more recent Cochrane meta-analysis concluded that the available data are insufficient to infer important risks or benefits for the mother or infant.8 The authors of all three meta-analyses nonetheless agreed that randomised controlled trials on large population samples are needed to assess the effect of regular physical activity during pregnancy on pregnancy outcomes such as birth weight when accounting for potential confounding factors, such as age, smoking habits, alcohol intake, number of hours standing, and gestational weight gain.
The role of muscle strength in the performance of exercise activities of daily living, as well as in preventing disease, has become increasingly recognised.9, 10 Resistance exercise training (for example, weightlifting) increases muscle strength and is recommended by major health organisations for improving public health.11, 12, 13 In the case of pregnant women, light resistance exercise training might be easier and better tolerated than aerobic training, owing to lower cardiovascular stress and caloric expenditure. To better understand how maternal resistance exercise training during pregnancy affects offspring's birth weight is, therefore, of medical relevance.
The primary aim of the present randomised controlled trial was to examine the effect of light intensity resistance and toning exercise training performed during the second and third trimester of pregnancy by earlier sedentary and healthy gravidae on the newborn's birth weight and length. As pre-pregnancy body weight is associated with pregnancy outcome,14 we also studied whether the association between maternal body weight before pregnancy and newborn's birth weight and length was affected by exercise training during pregnancy.
Materials and methods
A complete description of design and methods has been published elsewhere.15 In brief, the study was a randomised controlled trial (ClinicalTrials.gov ID NCT00813657). The research protocol was reviewed and approved by the Hospital Severo Ochoa (Madrid, Spain). The study was performed between January 2000 and March 2002, following the ethical guidelines of the Declaration of Helsinki, last modified in 2000.
We contacted a total of 480 Spanish pregnant women (Caucasian descent for three or more generations) of low-to-medium socio-economic class from a primary care medical centre (Centro de Salud María Montesori, Leganés, Madrid, Spain) (Figure 1). Criteria for belonging to a ‘low-middle class’ socio-economic class was having a family or individual mean income below or equal to the Spanish mean income between 2000 and 2002 (that is, family or individual income 23 000 and 16 000€ per year, respectively). After they provided written informed consent, we randomly assigned 160 healthy gravidae (aged 25–35 years) to either a training (n=80) or a control (n=80) group. They were sedentary, that is exercising >20 min on >3 days per week, with singleton pregnancy and not at high risk for preterm delivery (no more than one earlier preterm deliveries). Women not planning to give birth in the same obstetrics hospital department (Hospital Severo Ochoa, Madrid, Spain), and not being under medical follow-up throughout the entire pregnancy period, were not included in the study. Women with any serious medical condition preventing them from exercising safely were not included either.16
Exercising women participated in three sessions per week for about 26 weeks. The training intensity was carefully and individually controlled and was kept to light to moderate with relatively low cardiovascular stress (that is, heart rate 80% of age-predicted maximum heart rate value, calculated as 220 minus women's age). The exercise training programme started the second trimester (week 12–13) and was continued until the end of pregnancy (week 38–39). We originally planned an average of about 80 training sessions for each participant in the event of no preterm delivery. Details of the exercise training protocol have been described by Barakat et al.15 In brief, each session consisted of 35–40 min of exercise divided into a light intensity (60% maximal heart rate) warm-up period (∼8 min), followed by toning and very light resistance exercises (∼20 min), and finishing with a light intensity cool-down (∼8 min) period.
The core portion consisted of toning and joint mobilisation exercises involving major muscle and joint groups. Toning and joint mobilisation exercises included shoulder shrugs and rotations, arm elevations, leg lateral elevations, pelvic tilts, and rocks. Resistance exercises included one set of 10–12 repetitions of abdominal curls, biceps curls, arm extensions, arm side lifts, shoulder elevations, seated bench press, seated lateral row, lateral leg elevations, leg circles, knee extensions, knee (hamstring) curls, and ankle flexion and extensions. The women used barbells (3 kg per exercise) or low-to-medium resistance bands (Therabands). Exercises involving extreme stretching and joint overextension, ballistic movements, jumps, and those types of exercises performed on the back were specifically avoided. Supine postures were also avoided. All participants wore a heart rate monitor (Accurex Plus, Polar Electro OY, Finland) during the training sessions, so heart rate was continuously monitored. To further minimise cardiovascular stress, we specifically instructed participants to avoid the Valsalva manoeuvre. This was, however, a minor concern because of the overall low intensity of the exercises.
All resistance exercise training sessions were performed under observation and supervision in an exercise room. We used the exercise training facilities from the primary care medical centre in which the participants were monitored throughout the pregnancy.
Women in the non-exercise control group were asked to maintain their level of activity during the study period. All participants were followed up throughout the entire pregnancy period.
We recorded birth weight, birth length, and head circumference of the newborn, as well as gestational age at time of delivery (in weeks, days) from hospital perinatal records. Newborns were classified as having macrosomia when birth weight was >4000 g. We obtained Apgar scores from the reports of delivery room personnel (midwife) at 1 and 5 min after the complete birth of the baby. We calculated birth weight and length Z-score by using reference standards described for the Spanish population according to sex and gestational age.17 We also computed the ponderal index as kg m–3. Maternal weight gain was calculated as body weight before delivery minus body weight before pregnancy.
Physical activity and other measures
We assessed the occupational activities and other daily activities, such as number of hours standing, with the Minnesota leisure-time physical activity questionnaire.18 We measured weight and height of the mother by standard procedures at the start of the study, before delivery, and eventual preterm deliveries, which is before 37 completed weeks of gestation. Pre-pregnancy body mass index (BMI) was calculated as weight (in kilograms) divided by height (in meters) squared and categorised as follows: underweight (<18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), and obese (30.0 kg/m2). We also recorded smoking habits and alcohol intake at the start of the study and before delivery through an interview.
Gestational diabetes mellitus (GDM) was diagnosed when plasma glucose concentration—2 h after a 75-g oral glucose tolerance test performed with overnight fasting (abbreviated 2 h-glucose) in the 24–25th week—was 140 mg per 100 ml (7.8 mmol l–1). Blood glucose was measured with an automated analyser (Hitachi System 717; Roche Diagnostics, Basel, Switzerland). Treatment of women who developed GDM included:19 (i) nutrition counselling and (ii) daily self-monitoring of blood glucose to ensure adequate postprandrial glycaemia (that is, maximum acceptable limit of 140 mg per 100 ml and ideal levels <120 mg per 100 ml). Those GDM gravidae unable to control glycaemia (n=3) received regular insulin therapy.
Participant retention and adherence
To reduce participant drop out and to maintain adherence to the training programme, all sessions were accompanied by music and were performed in an airy, well-lighted exercise room. A qualified fitness specialist carefully supervised every training session and worked with groups of 10–12 women.
Blinding and randomisation
The participant randomisation assignment followed an allocation concealment process.20, 21 The researcher in charge of randomly assigning participants did not know in advance which treatment the next person would receive and did not participate in assessment. Assessment staff were blinded to participant randomisation assignment and participants were reminded not to discuss their randomisation assignments with assessment staff.
We used a conservative approach to sample size estimation. We made power calculations for the primary outcome measures of gestational age, Apgar score, birth weight, and length. We determined that adequate power (>0.80) would be achieved with 70 pregnant women in the training group and with 70 pregnant women in the control group. All power computations assumed that comparisons would be tested at a 5% significance level. All power computations allowed for 10% dropouts over 26 weeks.
We presented maternal characteristics of the study sample by group (training and control) as means and standard deviations, unless otherwise stated. For group comparisons, we analysed continuous and nominal data with a t-test for unpaired data and χ2 tests, respectively. We compared Apgar scores between groups using the non-parametric Mann–Whitney's U-test. We adjusted multiple comparisons for mass significance as described elsewhere.22, 23
After a bivariate correlation analysis, we used multiple regressions to study the association of maternal body weight with birth weight and length of the newborn. As there was a significant (P<0.05) interaction effect between body weight and group (exercise and training), we performed all analyses separately for the exercise and control groups.
We also performed regression analysis to study the association between 2 h-glucose and birth weight separately in control and training groups. Binary logistic regression was used to study the relationship between GDM and macrosomia (birth weight >4000 g).
We conducted all analyses using the intent-to-treat principle. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS, version 16.0 for WINDOWS; SPSS Inc., Chicago, IL, USA) and the level of significance was set to 0.05.
Adherence to training and possible adverse effects
Eight women from the training group discontinued the intervention because of diagnosed risk for premature labour (n=1), pregnancy-induced hypertension (n=1), persistent bleeding (n=1), or personal reasons (n=5). Five participants from the control group were excluded from the study because of pregnancy-induced hypertension (n=2), molar pregnancy (n=1), and threat of premature delivery (n=2). Five participants from the same group were also excluded because they decided to give birth in a different hospital. The final number of participants we included as valid study pregnant women was 72 in the training group and 70 in the control group. There were no exercise-related injuries experienced during pregnancy. Adherence to training in the experimental group was >90%. No women changed from the control group to the intervention group or vice versa, and there were no protocol deviations from study as planned.
We noted no major adverse effects and no major health problems in the participants except for two preterm deliveries in the training group and three preterm deliveries in controls. Of the two women in the training group with a preterm delivery, one (gestational age: 36 weeks, 2 days) had earlier history of preterm delivery (n=1) and the other one was a primigravida (gestational age: 35 weeks, 6 days). Both participants finished the training programme at the end of week 35 and the health status of the newborn was normal (Apgar score at 5 min 9). Of the three controls with preterm delivery (gestational age: 36 weeks, 2 days; 36 weeks, 4 days; and 36 weeks, 5 days) none had earlier history of preterm delivery and two were primigravidae.
Table 1 shows maternal characteristics in the training and control group. Both groups were similar in all the variables studied. Maternal body weight before pregnancy did not differ significantly between the two groups. Weight gain during pregnancy by group and BMI categories is presented in Table 2. We did not observe significant differences between groups (all P>0.1).
Newborn characteristics in the training and control group are presented in Table 3. We did not observe significant differences between groups in mean birth weight, mean birth length, head circumference, and gestational age. The prevalence of offspring with birth weight >4000 g was 10% (n=7) in the control vs 1.4% (n=1) in the training group (P>0.1). Apgar scores at 1 and 5 min did not differ (P>0.1) between groups and ranged within the upper scores indicative of an excellent prognosis for the newborn.24 All individual Apgar score values were 9 at 5 min.
Maternal body weight prior pregnancy and newborn's body weight and length
Figure 2 shows bivariate relationships between maternal body weight before pregnancy and birth weight and length of the newborn. The results were similar when we repeated the analyses using birth weight Z-score (r=0.201, P=0.091 and r=0.337, P=0.004 for training and control group, respectively) and length Z-score (r=0.150, P=0.207 and r=0.233, P=0.053 for training and control group, respectively). The results of the regression models with birth weight and length of the newborn as the outcome variables and maternal body weight before pregnancy as the predictor variable are shown in Table 4. Maternal body weight was positively and significantly associated with both birth weight and length of the newborn in the control group, whereas there was no significant association in the training group. The association between maternal body weight and birth length of the newborn did not persist after further controlling for maternal body height. Likewise, the associations were attenuated when we used the ponderal index as a marker of birth size (β=0.048, 95% CI: −0.018 to 0.113, P=0.149).
The results did not materially change when we repeated the analysis using birth weight and length Z-score. Maternal body weight was associated with birth weight Z-score of the newborn in controls (β=0.037, 95% CI: 0.012–0.061, P=0.004), but not in the training group (β=0.015, 95% CI: −0.02 to 0.033, P=0.091). Maternal body weight was associated with newborn's length Z-score in the control group (β=0.023, 95% CI: 0.01–0.048, P=0.053), but not in the training group (β=0.014, 95% CI: −0.008 to 0.035, P=0.207).
The results were also similar when we used maternal BMI instead of maternal body weight. Maternal BMI was associated with birth weight of the newborn in controls (β=44.184, 95% CI: 7.142–81.227, P=0.02), but not in the training group (β=3.649, 95% CI: −20.431 to 27.369, P=0.773). The association between maternal BMI and newborn's length approached statistical significance in controls (β=0.123, 95% CI: −0.017 to 0.264, P=0.083), whereas we did not observe an association in the training group (β=0.003, 95% CI: −0.104 to 0.110, P=0.995).
Two-hour glucose was significantly and positively associated with birth weight in the control group (β=5.774, 95% CI: 2.130–9.417, P=0.002), but not in the training group (β=0.592, 95% CI: −2.670 to 3.855, P=0.718). In the control group, the risk of having a macrosomic (>4000 g) newborn was higher in GDM than in non-GDM gravidae (odds ratio: 4.683, 95% CI: 1.579–13.851, P=0.005). No GDM gravida from the training group had an offspring with macrosomia. The figures were strengthened when the analyses were repeated after considering macrosomia if the newborn’s birth weight was >90th percentile (>3845 g). In the control group, the risk of having a macrosomic newborn (>3845 g) was higher in GDM than in non-GDM gravidae (odds ratio: 8.571; 95% CI: 1.998–36.767, P=0.004). Three women from the training group had an offspring heavier than 3845 g.
Our main finding was that supervised, light intensity resistance and toning exercise training performed over the second and third trimester of pregnancy in previously sedentary women does not affect newborn's birth size. We also observed that maternal body weight before pregnancy in the non-exercise group was positively and significantly associated with body weight of the newborn, whereas such association did not exist in the training group. Furthermore, GDM increased the odds ratio of having an offspring with macrosomia in the control group, but not in the training group. Taken together, our results suggest that regular, light exercise during pregnancy might counteract the negative impact that excess maternal body weight before pregnancy has on the newborn's birth size, while also attenuating the effect of GDM on the risk of macrosomia.
Whether maternal exercise during pregnancy affects offspring's birth size is not clear. Earlier research focused on the potential negative effects of maternal exercise during pregnancy on birth weight,6, 7, 8 based on the hypothesis that redistribution of blood flow to exercising muscles might reduce placental circulation, thereby theoretically compromising foetal growth.25 Different patterns of maternal exercise training such as type, intensity, or frequency of exercise as well as timing of pregnancy seem to have different effects on the newborn's birth size.26, 27, 28, 29 Clapp and Dickstein28, 30 reported that women who exercised 6 days per week for at least 60 min per session throughout the course of the pregnancy had lighter babies (−300 to −500 g) than those who stopped before the 28th week, whereas Hatch and Stein29 reported an apparently opposite finding. They studied >800 pregnant women who were categorised as non-exercisers, light–moderate exercisers (<1000 kcal per week), or heavy exercisers (mean of ∼2000 kcal per week). Women in the latter group delivered significantly heavier babies (+276 g) compared with non-exercisers. To note is that in these studies exercise information was collected either through questionnaire28, 30 or telephone interview.29 In this study, we did not find significant difference in birth weight between exercising and non-exercising women (142 g, P>0.1). Differences in study design, mode of exercise training, and exercise intensity makes comparisons between studies difficult.
We also observed that the prevalence of offspring with birth weight >4000 g was 10% (n=7) in the control vs 1.4% (n=1) in the training group. This is of relevance as newborns with high birth weight are more likely to become obese children.31, 32 Over the long term, childhood overweight–obesity is strongly associated with adult obesity, that is fivefold increase in risk for being overweight in early adulthood.33, 34, 35 Adults who were overweight in childhood are also at increased risk for cardiovascular disease compared with adults who were thin as children.33, 34, 36, 37 These facts might give support to begin prevention early.
Another appealing finding raised from this study is that overweight (BMI: 25–29.9 kg/m2) non-exercising women gained 1300 g more than those in the training group (12.2 vs 10.9 kg, respectively, P>0.1). Women from the control group gained ∼800 g more than the 1990 Institute of Medicine recommendations for gestational weight gain.38 Gestational weight gain above recommendations is associated with multiple adverse newborn outcomes, that is low 5-min Apgar score, seizure hypoglycaemia, polycythemia, and meconium aspiration syndrome.39 The American Dietetic Association states that weight gain during pregnancy should be carefully controlled to improve pregnancy outcome, avoiding excessive maternal postpartum weight retention, and reducing the risk of later chronic disease for the child.14 These observations might be in close connection to the fact that maternal body weight prior parity in our control group was positively associated with the body size of the newborn, whereas this was not the case in the training group. The results from a recent prospective study showed that newborn’s birth weight was associated with mother’s pre-pregnancy body weight.40 Our findings suggest that exercise training during the second and third trimester of pregnancy might attenuate the adverse effect of high maternal body weight before pregnancy on the newborn's birth size, as well as prevent the gravida from excessive gestational weight gain. Several studies showed that maternal excessive weight gain during pregnancy as well as maternal obesity are associated with a high risk for birth defects.41, 42, 43
That GDM increases the odds ratio of having an offspring with macrosomia concurs with the findings by Hillier et al.44 in a population of >80 000 mothers and newborns. They reported a higher risk of foetal macrosomia with increasing maternal blood glucose levels. We did not find such an association in those women who underwent the exercise programme over the second and third trimester of pregnancy. Our results support the hypothesis that physical training can contribute to reduce the risk of macrosomia in GDM gravidae. Regular exercise can reduce maternal glycaemia through increased insulin sensitivity and contraction-induced uptake of blood glucose into skeletal muscle fibres.45 This in turn can reduce blood glucose transfer to the foetus. The high prevalence of GDM observed in this study is of concern, and, therefore, these results should be taken with caution.
Our study is unique in that several potential confounding variables that might affect pregnancy outcome, such as age, earlier parity history, smoking habits, alcohol intake, number of hours standing, gestational weight gain, and obesity, were appropriately taken into account, as there were no differences between the training and the control group. Exercise training followed in this study consisted mainly of light resistance and toning exercises, whereas earlier exercise intervention trials in pregnant women focused mainly on aerobic exercises such as stationary cycling,46, 47, 48, 49, 50, 51 walking,52, 53, 54 or other exercise modalities such as stretching,53 and only one non-randomised study included weight lifting and strength conditioning.55
There is compelling evidence that improved muscular strength has beneficial effects in the prevention of chronic diseases, as well as in the ability to cope with daily living activities in both healthy and diseased people.9, 10 Regular resistance-type physical activities, such as the ones performed by our training group, are the main determinants of muscular strength11, 56 and are recommended for improving public health by major medical organisations.11, 12, 13 In the case of pregnant women, increased muscle strength may attenuate the cardiovascular response to any given load during physical activities of daily living because the load now represents a lower percentage of the maximal muscle contraction.57 Other potential benefits of resistance training include better posture, prevention of gestational low back pain and diastasis recti, and strengthening of the pelvic floor.58, 59
Limitations of this study include the fact that we did not record detailed information about dietary habits. However, it is unlikely that this had a major influence on the results because both groups of women were well nourished, ate to appetite, and had similar weight gain during pregnancy, and also similar BMI before pregnancy and before delivery. Whether this exercise programme is better than other programmes, such as walking 60 min every day, is not known.
In conclusion, the results of this study suggest that supervised, light intensity resistance exercise training performed over the second and third trimester of pregnancy does not have a negative impact on the newborn's body size. The fact that maternal body weight was positively and significantly associated with both birth weight and length of the newborn in the control group, but not in the training group, suggests that exercise interventions may attenuate the adverse consequences of excess maternal body weight before pregnancy on the newborn's birth size. The findings presented in this report provide additional evidence of the benefits of regular exercise during pregnancy for the maternal–foetal unit. We propose that medical interventions aimed at combatting GDM-related macrosomia should also include prescription of regular exercise during pregnancy.
Conflict of interest
The authors declare no conflict of interest.
This work was partially supported by the programme I3 2006 and by the post doctoral research programme (EX-2007-1124), Ministerio de Educación y Ciencia, Spain, and Swedish Council for Working Life and Social Research (FAS). We thank the technical assistance of the Gynaecology and Obstetric Service of ‘Severo Ochoa’ Hospital of Madrid.
About this article
Evaluation of a birth preparation program on lumbopelvic pain, urinary incontinence, anxiety and exercise: a randomized controlled trial
BMC Pregnancy and Childbirth (2013)