Influence of Physical Activity during Pregnancy on Birth Weight: Systematic Review and Meta-Analysis of Randomized Controlled Trials

Birth weight is a marker that is often referred to determine newborn health, potential growth trajectories and risk of future disease. Accordingly, interventions to promote appropriate and healthy birth weight have been extensively studied and implemented in pregnancy. In particular, physical activity in pregnancy is recommended to promote appropriate fetal development and newborn birth weight. This systematic review and meta-analyses aimed to summarize the effect of physical activity during pregnancy specifically from randomized controlled trials on the following outcomes: birth weight, macrosomia, low birth weight, being large for the gestational age, and being small for the gestational age (Registration No.: CRD42022370729). 63 studies (16,524 pregnant women) were included. There was a significant negative relationship between physical activity during pregnancy and macrosomia (z = 2.16; p = 0.03; RR = 0.79, 95% CI = 0.63, 0.98, I2 = 29%, Pheterogeneity = 0.09). No other significant relationships were found. Promoting physical activity during pregnancy may be an opportune time to reduce the risk of future chronic disease, such as obesity, through the prevention of macrosomia and the promotion of appropriate birth weights.


Introduction
Birth weight is an important and accessible factor to evaluate newborn health and predict growth trajectories and downstream risk of potential chronic disease such as obesity [1].The World Health Organization defines low birth weight as the weight of a newborn below 2500 g, typically representing the 10th percentile for its gestational age [2].Newborns that are small for their gestational age (less than the 10th percentile based on gestational age) are at greater risk of having a low birth weight, and associations have also been found with long-term cognitive deficits in childhood and adolescence [3,4].On the opposite end of the weight spectrum, macrosomia is defined arbitrarily as a birth weight exceeding 4000 g [5].Babies born with macrosomia have a higher likelihood of encountering adverse delivery outcomes, including shoulder dystocia, brachial plexus injury, clavicular fracture, birth asphyxia, and neonatal mortality [6][7][8].Moreover, macrosomia increases the risk of undergoing cesarean section and experiencing vaginal and perineal trauma, as well as postpartum hemorrhage [9,10].Similar complications are associated with newborns that are large for their gestational age (within the 90th percentile based on their gestational age), including prolonged delivery and the risk of injury at birth.Given the risk of complications at either end of the birth weight spectrum, interventions to promote appropriate birth weight are integral.Physical activity during pregnancy has been described as a modifiable and accessible health behavior to facilitate maternal and newborn health, including the promotion of an appropriate birth weight [11].International guidelines for prenatal physical activity suggest that all pregnant individuals without contraindications for being active should aim to accumulate 150 min of moderate-intensity activity per week [12].Contrary to popular social beliefs that physical activity in pregnancy could result in reduced fetal growth and development, no associations have been found with such factors [13].In fact, being active throughout pregnancy has been shown to advance placenta blood perfusion [14], which can improve nutrient transport and overall placental function and therefore facilitate fetal development [15].Furthermore, previous reviews have consistently advised that physical activity in pregnancy does not have adverse effects on fetal development [14,16,17].
However, associations between physical activity in pregnancy and birth weight have been inconclusive.For example, a recent systematic review of 32 studies on prenatal physical activity concluded that there is no association between engagement in physical activity throughout pregnancy and newborn body composition markers [18].Another study that included objective measures of prenatal physical activity via accelerometry suggested that physical activity was correlated with birth weight, and specifically, that aerobic activity reduced birth weight and increased the number of newborns weighing within the appropriate range [19].Contrarily, a recent meta-analysis of randomized controlled trials that included studies that had assessed gestational weight gain found no associations between prenatal physical activity and large-or small-for-gestational-age newborns [20].Despite the inconsistencies in the literature regarding birth weight, prenatal physical activity is still recommended, given the several health benefits for both the pregnant person and future child [16].In fact, a recent expert review emphasized that physical activity in pregnancy should be integrated into standard care, especially given the high quality and strong evidence base supporting its contribution to reduced perinatal complications and improved labor and delivery outcomes [16].
Rather than negating or minimizing the potential benefits of prenatal physical activity on birth weight due to inconsistent results, they indicate that it may be necessary to further examine this relationship via high-quality randomized controlled trials that specifically target these outcomes.Accordingly, the purpose of this systematic review and metaanalysis was to assess the effect of prenatal physical activity on markers of birth weight (i.e., total birth weight and incidences of being large for the gestational age, being small for the gestational age, low birth weight, and macrosomia) from randomized controlled trials.We hypothesized that the amalgamated findings from randomized controlled trials would be in favor of physical activity in pregnancy for reduced risk of being large for the gestational age, being small for the gestational age, having a low birth weight, and macrosomia, along with showing a significant association with reduced total birth weight.

Methods
A systematic review was developed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [21].The protocol was registered in the International Prospective Registry of Systematic Reviews (PROSPERO), Registration No. CRD42022370729.

Eligibility Criteria
The eligibility criteria for this systematic review and meta-analysis was guided by the PICOS framework: participants, interventions, comparisons, outcomes, and study design [20].The participants included pregnant women, the intervention was physical activity, the comparison was no physical activity/control, and the outcomes were birth weight, being small for the gestational age, being large for the gestational age, and macrosomia.

Population
The population of interest was pregnant individuals without contraindication to exercise or physical activity (following the most recent international clinical guideline about physical activity during pregnancy) [22,23].Absolute contraindications were characterized by conditions such as ruptured membranes, premature labor, persistent second-or thirdtrimester bleeding, and other similar factors.On the other hand, relative contraindications were characterized by a history of spontaneous abortion, mild/moderate cardiovascular or respiratory disease, etc. [22,23].

Intervention (Exposure)
We conducted a search to identify physical activity interventions during pregnancy that involved quantifiable forms of physical activity.The focus was on extracting information regarding the program's reporting of duration, intensity, type of activities, weekly frequency, session duration, participant adherence, and whether supervision was provided.

Comparison
The comparator was no exercise or physical activity (i.e., the control group of the selected studies), normally involving pregnant participants who followed a regular obstetrical follow-up in their health centers.

Outcome
The main outcomes of the study were birth weight, macrosomia, and low birth weight.Secondary outcomes were being large for the gestational age and being small for the gestational age.

Data Sources
A comprehensive search was carried out through the Universidad Politécnica de Madrid software in the following databases: Academic Search Premier, ERIC, MEDLINE, SPORTDiscus, OpenDissertations, Clinicaltrials.gov,Web of Science, Scopus, the Cochrane Database of Systematic Reviews, and the Physiotherapy Evidence Database (PEDro).To ensure equality in the selection process, the same article selection criteria were used for all databases, considering differences in controlled vocabulary and rules of selection syntax.The search terms used were:

Study Selection and Data Extraction
Only randomized controlled trials (RCTs) were selected.Also, systematic reviews previously published in the same field were searched to compare our results.Articles published between 2010 and 2023 written in English and Spanish were considered for the search.Reference lists of selected studies were retrieved to identify other studies that might have been missed by the electronic keyword search.
To ensure compliance with the inclusion criteria, two reviewers (MS and CS) conducted an independent screening of the titles and abstracts retrieved from the electronic searches.
The abstracts that met the initial screening were subjected to further analysis.The full texts were screened independently by two reviewers (JG and RB) to identify outcomes of interest for data extraction.
To identify any potential additional studies that were not captured by the electronic searches, the list of references from selected articles was screened.In cases where a study had multiple publications, the most recent or comprehensive publication was chosen as the primary source.However, relevant data from all the publications were extracted to ensure that no valuable information was overlooked.
For studies where one reviewer (DZ) recommended exclusion, both reviewers (CS and MS) tried to reach a consensus to make a final decision for exclusion or inclusion.In situations of absolute discrepancy, a third reviewer (RB) provided their expert opinion on whether the study should be included or excluded.The study selection process is detailed in Figure 1.
To ensure compliance with the inclusion criteria, two reviewers (MS a ducted an independent screening of the titles and abstracts retrieved from t searches.The abstracts that met the initial screening were subjected to furt The full texts were screened independently by two reviewers (JG and RB) to comes of interest for data extraction.
To identify any potential additional studies that were not captured by t searches, the list of references from selected articles was screened.In cases w had multiple publications, the most recent or comprehensive publication w the primary source.However, relevant data from all the publications were ensure that no valuable information was overlooked.
For studies where one reviewer (DZ) recommended exclusion, both re and MS) tried to reach a consensus to make a final decision for exclusion or situations of absolute discrepancy, a third reviewer (RB) provided their expe whether the study should be included or excluded.The study selection proce in Figure 1.Data extraction tables were created in an Excel sheet.One researcher e data, and then data extraction was independently verified by a content exper further analysis.Extracted data were study characteristics (i.e., author last na country), type of article (RCT), total sample size and group sample size, inte posure (exercise prescribed and/or measured), including: frequency, intens type, supervision of intervention, duration of intervention and adherence of primary and secondary outcomes (Table 1).Data extraction tables were created in an Excel sheet.One researcher extracted the data, and then data extraction was independently verified by a content expert to facilitate further analysis.Extracted data were study characteristics (i.e., author last name, year and country), type of article (RCT), total sample size and group sample size, intervention/exposure (exercise prescribed and/or measured), including: frequency, intensity, time and type, supervision of intervention, duration of intervention and adherence of intervention, primary and secondary outcomes (Table 1).Author: first author last name; Year: year of study; Country: country where the article was developed (usually in the method part); Type: type of article, if it is a randomized clinical (or controlled) trial, RCT is indicated.N: total number of women analyzed.Those of the GI and those of the CG have to coincide.EG: Number of women analyzed in the intervention group.CG: number of women analyzed in the control group.Freq: weekly frequency of exercise sessions (3 days a week, 2, etc.).Intensity: moderate, high...; Duration of program: program time.If the program lasted 10 weeks, or if it started in week 12 and ended in week 28, it is described as being 16 weeks long.Type of exercise: aerobic, muscle strengthening, etc. Superv.Classes: whether or not there was supervision.Duration of class: minutes of each session.Adh.: adherence of the participants to the intervention (%).This indicates how many women attended.Main variables analyzed: lists all the main variables of the study.This is usually in the method section in "outcomes", and they appear as "main outcomes".If main does not appear, they are the first.You will find it in several places.Secondary variables: the same as before, but secondary.In the same study, this may involve different types of exercises, varying durations for each exercise, and both supervised and unsupervised exercises.

Quality of Evidence and Risk of Bias Assessments
To evaluate the quality of evidence for each study design and outcome, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework was used.This framework provides a standardized and comprehensive approach to assess the strength of the evidence across multiple studies [87].
To evaluate the risk of bias, the Cochrane Handbook was utilized.The potential sources of bias evaluated are selection bias (inadequate randomization procedures for RCTs), performances bias (compliance with the intervention for RCTs), detection bias (flawed outcome measurement), attrition bias (incomplete follow-up and high loss to follow-up), and reporting bias (selective or incomplete outcome reporting) [88].

Statistical Analysis
Statistical analyses were performed with the software RevMan in its 5.3 version.For continuous variables, birth weight (grams), mean, and standard deviations were recorded.The overall confidence interval (CI) was calculated using the mean difference (MD) [89].All dichotomous outcomes, macrosomia, low birth weight, being large for the gestational age, and being small for the gestational age were expressed as categorical variables (Yes/No) to calculate the relative risk (RR) [90].Random effect models were applied.To establish the compensated average in both dichotomous and continuous analyses, a weight system was used that considered the sample size per group, and, generally, these were contributed by each study.To assess the variation in study results between studies (i.e., the degree of heterogeneity), the I 2 statistic was interpreted using established thresholds: 25% for low heterogeneity, 50% for moderate heterogeneity, and >75% for high heterogeneity [91].

Study Characteristics
In total, 63 studies that met the inclusion criteria were identified, involving 16,524 pregnant women across 23 countries on five continents.All of the studies were randomized control trials, including 59 exercise interventions only and 4 of exercise and dietary counselling.Studies varied in frequency from 2 to 7 days per week, with low to moderate intensities lasting 15 to 75 min per session.These interventions were carried out during the first, second, and third trimesters, and lasted from 3 to 30 weeks.The types of exercise included walking, stationary cycling, water aerobics, swimming, resistance training, stretching, Pilates, Yoga, pelvic floor muscle training, and combinations of various exercise types.Additional details about the studies can be found in the Table 1.The results of mean birth weight, macrosomia, low birth weights, being large for the gestational age, and being small for the gestational age are presented below.

Risk of Bias Assessment
Collectively, the quality of evidence varied from low to high.In some situations, the blinding of participants to the group (intervention or control group) was not feasible, and it is typically impossible to achieve due to the intervention characteristics (physical activity intervention), resulting in unclear or high risk of bias (performance bias), depending on how it was recorded.Other sources of bias in some cases were the impossibility of finding the published article protocol (to compare the planned and measured outcomes), but also a lack of reporting (or having an uncertain definition) of the randomization process.Overall, the majority of the studies presented a low risk of bias within the five types of bias assessed.The risk of bias analysis is reported in Figure 2.

Effect of Physical Activity during Pregnancy on Birth weight
There were a total of 61 studies that were incorporated in this analysis [24-66,68-83, 85,86].Regular exercise or physical activity during pregnancy did not have a significant relationship with birth weight (z = 0.11; p = 0.91) (Std.Mean Dif., Random, 95% CI = 0.00 (−0.04, 0.05) I 2 = 43%, P heterogeneity = 0.0003).The forest plot corresponding to the current meta-analysis is illustrated in Figure 3.

Discussion
This systematic review and meta-analysis demonstrated the positive effect of physical activity in pregnancy in reducing the risk of macrosomia by referring specifically to evidence from randomized controlled trials.No other significant associations were found with indices of birth weight, including birth weight as a continuous variable, risk of being small for the gestational age, and having a low birth weight.Birth weight is often used as an accessible marker of newborn health and as an assessment of potential growth trajectories and downstream risk of chronic disease [1,92].Physical activity during pregnancy may be a key factor in promoting appropriate birth weight, especially contributing to the prevention of macrosomia, and therefore is an important behavior that will facilitate both maternal and child health.
The only significant finding in relation to prenatal physical activity and birth weight was the reduced odds of developing macrosomia.This finding is consistent with previous reviews that have shown reduced odds of macrosomia with physical activity in pregnancy [13,93].In fact, one systematic review that conducted a sub-analysis of randomized controlled trials only found a 39% reduced risk of macrosomia [13], and our findings further underscored this, thus supporting the effectiveness of prenatal physical activity in the prevention of macrosomia.Macrosomia has shown strong associations with the risk of downstream childhood obesity, and there have been both physiological and environmental mechanisms that have been proposed [94,95].For example, it is theorized that macrosomia can be a proxy measure for potential adipose tissue function, including overactivity and therefore excess energy storage that can increase the risk for later-life obesity [96].Moreover, early food restriction practices have been associated with infants who are larger, potentially as an effort to bring their weight into expected trajectories; however, this may be a detriment to appetite regulation [97,98] later on in life.In line with the Developmental Origins of Health and Disease, uterine and early life environments can program downstream childhood obesity, and perhaps prenatal physical activity is a viable factor that can prevent this by reducing the risk of macrosomia [99].Notably, we also tested if birth weight was different only amongst the studies that reported on macrosomia, and this was

Discussion
This systematic review and meta-analysis demonstrated the positive effect of physical activity in pregnancy in reducing the risk of macrosomia by referring specifically to evidence from randomized controlled trials.No other significant associations were found with indices of birth weight, including birth weight as a continuous variable, risk of being small for the gestational age, and having a low birth weight.Birth weight is often used as an accessible marker of newborn health and as an assessment of potential growth trajectories and downstream risk of chronic disease [1,92].Physical activity during pregnancy may be a key factor in promoting appropriate birth weight, especially contributing to the prevention of macrosomia, and therefore is an important behavior that will facilitate both maternal and child health.
The only significant finding in relation to prenatal physical activity and birth weight was the reduced odds of developing macrosomia.This finding is consistent with previous reviews that have shown reduced odds of macrosomia with physical activity in pregnancy [13,93].In fact, one systematic review that conducted a sub-analysis of randomized controlled trials only found a 39% reduced risk of macrosomia [13], and our findings further underscored this, thus supporting the effectiveness of prenatal physical activity in the prevention of macrosomia.Macrosomia has shown strong associations with the risk of downstream childhood obesity, and there have been both physiological and environmental mechanisms that have been proposed [94,95].For example, it is theorized that macrosomia can be a proxy measure for potential adipose tissue function, including overactivity and therefore excess energy storage that can increase the risk for later-life obesity [96].Moreover, early food restriction practices have been associated with infants who are larger, potentially as an effort to bring their weight into expected trajectories; however, this may be a detriment to appetite regulation [97,98] later on in life.In line with the Developmental Origins of Health and Disease, uterine and early life environments can program downstream childhood obesity, and perhaps prenatal physical activity is a viable factor that can prevent this by reducing the risk of macrosomia [99].Notably, we also tested if birth weight was different only amongst the studies that reported on macrosomia, and this was not significant.We postulate that this may be due to the fact that macrosomia is defined in absolute values as >4000 g, whereas birth weight is continuous.It may also be due to the heterogeneity amongst included studies in the measurement of weight, as well as the types of exercises performed.Similar findings have been noted about nutrition and exercise interventions in pregnancy, where favorable prevalence outcomes have been found, such as the prevention of excessive gestational weight gain and macrosomia, but no differences were found when weight was measured continuously between intervention and control groups [24,57,70].This review did not find any significant relationship with prenatal physical activity and birth weight, large-or small-for-gestational-age newborns, and low birth weight.These null findings are consistent with previous systematic reviews, meta-analyses, and observational studies [13,17,93,100].Previous research has suggested that the relationship between physical activity during pregnancy and birth weight may have an inverted U-shape, such that higher frequencies and intensities of activity may be associated with being small for the gestational age or low birth weight, whereas lower frequencies and intensities of activity could increase risk of being large for the gestational age or macrosomia [101].Notably, though, engagement in regular moderate levels of physical activity does not increase the risk for small-for-gestational-age and low-birth-weight newborns [13].Therefore, although our results found no relationship with the lower end of the weight spectrum, it should be highlighted that engagement in physical activity throughout pregnancy does not increase the risk for smaller newborns.A common misconception is that physical activity in pregnancy could be unsafe for fetal development, as energy reserves would be diverted from the placenta, or there is a risk of physical harm [102].In order to improve knowledge on the safety of maternal physical activity, it is essential that public health messaging should debunk stereotypes or myths that suggest physical activity in pregnancy can deplete or divert energy reserves for fetal growth and development.
Physical activity throughout pregnancy elicits several benefits for maternal and newborn health, including the prevention of perinatal complications such as excessive gestational weight gain and gestational diabetes [103,104].In the present review, we assessed the direct relationship between engagement in a physical activity intervention and markers of birth weight; however, it should be acknowledged that benefits pertaining to birth weight could be moderated by improvements in other markers of perinatal health.For example, gestational weight gain and birth weight are positively correlated, and therefore, preventing excessive gestational weight gain may also affect the reduction in birth weight [105].Similarly, gestational diabetes increases the risk for large-for-gestational-age newborns [106].Physical activity in pregnancy reduces the risk for excessive gestational weight gain by 32% and gestational diabetes by 38% [102,103], and accordingly, may be attributed to also improving newborn birth weight.Physical activity also improves labor and delivery outcomes, including the reduced risk of cesarean section, thus preventing potential complications associated with higher-birth-weight newborns [107].Taken together, the benefits of physical activity extend beyond indices of weight and affect both the pregnant person and newborn.
Strengths of this review include the inclusion of both English and Spanish articles, expanding the scope of our search in comparison to previous reviews that were restricted to one language, and the inclusion specifically of randomized controlled trials, allowing for the assessment of the features of physical activity interventions that may not be captured through observational studies (e.g., frequency and type of activity).Moreover, randomized controlled trials are deemed to provide more high-quality evidence.However, these results should be interpreted with caution due to the inclusion of studies deemed to be of low quality, as well as the heterogeneity in the contents of the included interventions.Future research should aim to further extrapolate findings based on the intensity of the intervention and types of physical activity.Future comprehensive research should also expand on the inclusion of additional languages.

Conclusions
By referring to evidence from randomized controlled trials, this review identified that prenatal physical activity could reduce the risk for macrosomia.Prenatal physical activity did not have a significant effect on mean birth weight, small-or large-for-gestational-age newborns, or low birth weight.Importantly, though, prenatal physical activity does not also increase the risk for smaller newborns, and this is an important message that should be widely disseminated to debunk myths associated with reduced or diverted energy reserves for fetal growth and development with active pregnancies.Further research is needed to examine the effect of prenatal physical activity on birth weight, including specific recommended intensity and types of activity, and whether birth weight is moderated through the prevention of other perinatal complications by physical activity, such as excessive gestational weight gain and gestational diabetes.

Figure 1 .
Figure 1.Flow chart of the retrieved and analyzed articles.

Figure 1 .
Figure 1.Flow chart of the retrieved and analyzed articles.

Table 1 .
Characteristics of the studies analyzed.