Exercise During Pregnancy: Increased Oxygen Uptake and Metabolic Demand

Abstract

INTRODUCTION:

Current guidelines for exercise during pregnancy are limited in their precision on exercise intensity as the established compendium for exercise modes and intensities is based on estimates in non-gravid adults. To address this gap, we used indirect calorimetry to measure prenatal exercise energy demand during standard aerobic exercises in order to compare intensity to published standards.

METHODS:

Participants (n = 7; 30.7 ± 3.3 yrs of age) completed 5 minutes each of exercise at low and moderate intensity for walking and cycling, on the same testing day, separated < 1 minute. For the first 5-minute bout, the participant performed an external work rate that elicited low intensity exercise, i.e., 40 - 59% peak aerobic capacity (VO₂ peak). This was followed by 5 minutes of the same exercise mode at moderate intensity, i.e., 40 - 59% VO₂ peak. This process was repeated on a cycle ergometer. Measured VO₂ was compared with ACSM-estimated VO₂ at each trimester. Reported results were presented separately for low and high exercise intensities.

RESULTS:

Measured absolute and relative VO₂ during walking and cycling exercises were significantly higher (p < .001) in all trimesters compared with predicted values, which corresponded with a 35% higher measured MET value.

CONCLUSIONS:

These findings highlight a critical underestimation in the quantification of effort during prenatal exercise, which is essential to address to safely prescribe exercise for this unique population.

Graphical Abstract

New & Noteworthy:

We measured metabolic equivalents (METs) of prenatal aerobic exercise with indirect calorimetry at light and moderate intensities, then compared to estimates from common equations. Our findings indicate higher oxygen uptake and METs across all trimesters of pregnancy during walk and cycle exercise. We further show an increase in RPE at the same work rate in the 2^nd and 3^rd trimesters. We conclude that higher metabolic demand must be accounted for in prenatal exercise prescriptions.

INTRODUCTION

The American College of Sports Medicine (ACSM), the American Congress of Obstetricians and Gynecologists (ACOG), the Society of Obstetricians and Gynecologists of Canada (SOGC), the Canadian Society of Exercise Physiologists (CSEP) and others have published guidelines for exercise during pregnancy and the postpartum period (1-4). The guidelines establish general metabolic equivalents (METs) for differing physical activity modes and intensities (5, 6), but follow a few assumptions. These guidelines provide similar recommendations for moderate intensity exercise throughout pregnancy and include guidance for women with normal weight as well as women with overweight or obesity. Indeed, recent works have found regular maternal exercise within the recommended intensity range to be beneficial to many indicators of cardiometabolic health of mother and offspring (7-17). Still, evidence suggests that the accurate determination of intensity may be missing from prenatal exercise programs (18, 19).

Changes in metabolic demand and substrate utilization have not been comprehensively reported in pregnant women performing submaximal steady-state exercise, and the potential findings could result in discrepancies between predicted and actual energy expenditure during prenatal exercise. Pregnancy-related physiological and metabolic adaptations such as oxygen utilization in the placenta and fetus could increase the oxygen cost of exercise (20). Additionally, as gestation advances, increases in body mass and shifts in the center of gravity and gait could influence exercise economy (21), particularly during weight-bearing activities such as walking. Direct analysis has not yet compared measured intensity during pregnancy with published intensity used for exercise programming. Furthermore, the accuracy of predictive equations provided by ACSM to estimate energy expenditure has not been tested during prenatal exercise. Through indirect calorimetry assessment and evaluation of the MET during prenatal exercise, we aimed to address this important gap in scientific and practical applications.

The specific objectives of this study were to determine the direct measure of energy cost, intensity, and economy during specific exercises in each trimester. We hypothesized that published METs would underestimate the oxygen consumption, reflecting increased energy cost, intensity, and economy of acute bouts of each exercise in each trimester.

METHODS

Study Participants

Participants with healthy singleton pregnancies were recruited before 16 weeks’ gestation. A total of 7 participants were enrolled, and all participants completed the exercise testing sessions in every trimester. The primary focus of this investigation was to examine the metabolic cost of exercise during the first, second, and third trimesters of pregnancy. Specifically, we were interested in assessing whether the volume of oxygen uptake (VO₂) and METs estimated from exercise work rate were accurate and appropriate for use with exercising pregnant women. Female participants enrolled in this study met the following criteria: clearance from a health care provider to participate in physical activity; between 18 and 40 years of age; pre-pregnancy body mass index (BMI) >18.5 kg▪m⁻²; singleton pregnancy; ≤16 weeks of gestation; and no current alcohol or tobacco use. Criteria for exclusion included smoking, known pre-existing conditions (i.e., diabetes mellitus, hypertension, cardiovascular disease, and comorbidities, systemic lupus erythematosus), and/or medications known to affect fetal growth and development. Participants self-identified race and ethnicity. All women enrolled went on to have healthy full-term deliveries.

Ethics Statement

Approval for this study was obtained from the East Carolina University (ECU) Institutional Review Board (UMCIRB 23-001366). Written informed consent was obtained from each participant upon enrollment. All experimental procedures were conducted at ECU.

Peak Oxygen Consumption via Indirect Calorimetry: Submaximal Treadmill Test

After study enrollment, participants completed a submaximal treadmill test to determine aerobic capacity and calculate the target heart rate (THR) range for low- and moderate-intensity training. Before the bouts of exercise, the participant was fitted with a wearable HR monitor. This test was completed before 16 weeks of gestation, and participants were given a minimum 48-hour washout period before further exercise testing. Participants arrived to the laboratory reporting no nutritional intake besides water in the previous 4 hours, and no strenuous exercise in the preceding 24 hours. Peak oxygen uptake (VO₂ peak) was estimated via the modified Balke protocol previously validated for pregnant females (22). THR zones for the submaximal exercise components corresponded to maternal HR at 20 – 40% (low intensity) and 40 – 59% (moderate intensity) of peak oxygen uptake.

Oxygen Consumption via Indirect Calorimetry: Steady State Exercise Test

Steady state exercise was performed in a fixed order on the cycle ergometer and treadmill during each trimester of pregnancy. During exercise, VO₂ was measured via open-circuit spirometry using a mobile indirect calorimeter (COSMED K5 Portable Metabolic System, COSMED, Rome, Italy). Before the bouts of exercise, the participant was fitted with a wearable HR monitor and a sealed, silicone mask covering their nose and mouth with a one-way air flow breathing tube that funnels exhaled air into a metabolic gas analyzer. The indirect calorimeter performed a breath-by-breath analysis that measured exhaled air to calculate O₂ consumption and carbon dioxide (CO₂) production. Additional physiological variables were assessed as secondary outcomes, including ventilation rate (i.e., breathing frequency) and respiratory exchange ratio (i.e., ratio of carbon dioxide production to oxygen consumption; RER), which provided a ratio of carbohydrate to fat utilization during exercise (i.e., RER of 1.0 indicates carbohydrate use, RER of 0.7 indicates fat use). After donning the mask, the participant remained seated for 3 minutes, after which HR was recorded and BP measured. The participants began exercise with a 2 - 3 minute warm-up at a self-selected walk pace.

Participants completed 5 minutes of low-intensity exercise and then 5 minutes of moderate-intensity exercise on the treadmill, followed by 5 minutes of low-intensity exercise and then 5 minutes of moderate-intensity exercise on the cycle ergometer. For the first 5-minute bout, the participant performed an external work rate that elicited an HR within their THR zone for low-intensity exercise. Following the low-intensity bout of exercise, the participant performed an external work rate that elicited an HR within their THR zone for moderate intensity, maintaining this intensity of exercise for 5 minutes. Subject rating of perceived exertion (RPE) was reported using a Borg 6-20 scale during the last minute of each session (23). Immediately following treadmill exercises, the subjects were escorted to the cycle ergometer. This break in data collection did not last > 60 seconds. Steady state VO₂ from the last two minutes of each bout was exported using COSMED Omnia Software (version 2.5.2; Rome, Italy), then compared with ACSM-estimated VO₂ using the equations: treadmill walking VO₂ = (0.1 x speed) + (1.8 x speed x grade) + 3.5; and cycle VO₂ = (1.8 x 673 kgm/min) ÷ kg body mass + 7, where 1 watt x 6.12 = kgm/min. In addition to O₂ uptake, the relative O₂ uptake to the external work rate and speed (exercise economy) was calculated based on the O₂ requirement for each unit of power (watts) and speed (treadmill only). Reported results were presented separately for low and high exercise intensities.

Statistical Analysis

To assess the effect of pregnancy on the accuracy of O₂ uptake and MET equations, unpaired, two-tailed independent samples t-tests were calculated between estimated and measured values, with an α of p = .05 considered to be statistically significant. Based on previous work (24), for 80% at α = .05, a sample size of at least 5 participants to assess the presence or absence of differences in the compendium vs. actual (measured) METs in pregnant women. One-way analysis of variance (ANOVA) was performed to test the effect of trimester (1, 2, 3) on O₂ uptake and METs. One-way ANOVA tested only the effect of trimester on exercise economy. Finally, the O₂ uptake mean and standard deviation were plotted continuously throughout each session. For figures 1 and 2, statistical analyses were completed using SPSS software (RRID: SCR_002865, version 28.0.1.1, SPSS Inc., IBM Corp., Chicago, IL). For the presentation of VO₂ in Figure 3, data curation and formal analysis of the multisensor dataset were executed using a custom Python pipeline. The raw COSMED K5 time-series data were systematically ingested, scrubbed of missing values, and standardized. The disparate datasets were then computationally grouped by subject ID across all three trimesters to generate comparative time-series visualizations (e.g., mapping absolute Oxygen Uptake / VO₂ against elapsed time)

Figure 1. Exercise VO₂ & METs During Pregnancy.

A: Absolute VO₂ was higher than estimated for treadmill and cycle low-intensity exercise (n = 14 data points), leading to a large discrepancy in estimated and measured relative VO₂ and METs. B: Absolute VO₂ was higher than estimated for moderate-intensity exercise (n = 14), thus furthering the discrepancy between estimated and measured relative VO₂ and METs. Despite the influence of body mass in 2^nd and 3^rd trimesters, the higher relative VO₂ and METs seen in 1^st trimester (C) continued through 2^nd and 3^rd, revealing misestimation of METs through all trimesters (n = 14 each). Significance determined by Independent samples T-tests estimated vs. measured within each trimester.

Figure 2. Perceived Exertion Across Trimester.

Subject RPE at the same work rate was higher in trimester 2 and 3 compared with trimester 1 (n = 7 each intensity and trimester). No significant rise in RPE was seen between trimester 2 and 3. Significance determined by ANOVA with Tukey post-hoc between each trimester.

Figure 3. Variation of VO₂ in Prenatal Exercisers.

The time axis for each test was normalized to percentage of test completion (n = 7; 0–100%). The plot shows three mean curves, one for each trimester (T1, T2, and T3), with shaded regions including ±1 standard deviation from the mean.

RESULTS

Women (n = 7) were on average 30 ± 3 years of age, with pre-pregnancy BMI 24.7 ± 3.4 kg/m², and were evenly representative of white and BIPOC races in the study sample (Table 1). Compared with ACSM estimations (Table 1), measured absolute and relative VO₂ during prenatal exercise were significantly higher (p < .001) during each exercise session and in each trimester (Figure 1). Accordingly, this corresponded with a 35% higher measured MET value than estimated (4.6 measured vs. 3.4 estimated; p < .001). This led to a 15-kilocalorie discrepancy in 10 minutes of exercise, which could equate to larger discrepancies in longer-duration bouts of walking exercise. Measured VO₂ was 42% and 30% higher in low and moderate intensity exercises, respectively, leading to misestimation of METs at both intensity levels (Figure 1A & 1B). Higher VO₂ than estimated was observed equally through each trimester (Figure 1C), i.e., VO₂ did not significantly increase in the second or third trimester compared with the first trimester.

Table 1.

Trimester Exercise Metrics

			Trimester
Intensity	Mode	Metric	1	2	3
		Body Mass (kg)	65.7 ± 8.7	71.8 ± 7.7	77.6 ± 9.1
		Maternal age (yrs)	30.3 ± 3.2
		pP BMI (kg/m2)	24.7 ± 3.4
		Gravida/Parity	2(1,3) / 1(0,2)
		% BIPOC	43
		VO2peak (ml/kg/min)	25.9 ± 3.6
		METs at VO2peak	8.7 ± 1.2
Low (1 - 3 METs)	Treadmill	Speed (mph)	3.1 ± 0.4	3.0 ± 0.4	2.9 ± 0.5
		Speed (m/min)	82.3 ± 11.2	81.2 ± 10.3	78.5 ± 13.3
		% Grade	0	0	0
		ACSM-estimated VO2 (ml/kg/min)	11.9	11.8	11.6
		RPE	8.5 ± 0.5	11.6 ± 1.7	11.9 ± 1.1
	Cycle	Watts	35 ± 6	35 ± 6	32 ± 5
		ACSM-estimated VO2 (ml/kg/min)	12.9	12.4	11.6
		RPE	8.8 ± 0.4	11.4 ± 2.1	11.9 ± 0.8
Moderate (3 – 6 METs)	Treadmill	Speed (mph)	3.4 ± 0.4	3.4 ± 0.4	3.3 ± 0.4
		Speed (m/min)	90.7 ± 11.3	91.1 ± 10.7	88.1 ± 11.5
		% Grade	2	2	1
		ACSM-estimated VO2 (ml/kg/min)	16.5	16.6	14.6
		RPE	11.5 ± 2.1	13.4 ± 1.0	13.1 ± 0.8
	Cycle	Watts	57 ± 10	59 ± 10	54 ± 11
		ACSM-estimated VO2 (ml/kg/min)	16.6	16.0	14.7
		RPE	11.8 ± 1.5	13.2 ± 1.2	13.1 ± 1.1

Data are mean ± SD from 10 min aerobic exercise during pregnancy. pP prepregnancy; BMI Body Mass Index; BIPOC Black or Indigenous People of Color. Non-parametric data - median (min,max). METs metabolic equivalents; ACSM American College of Sports Medicine - VO₂ equations: treadmill walking VO₂ = (0.1 x speed) + (1.8 x speed x grade) + 3.5; cycle VO₂ = (1.8 x 673 kgm/min) ÷ kg body mass + 7, where 1 watt x 6.12 = kgm/min. RPE Borg Rating of Perceived Exertion (6-20).

While the same exercise intensity was performed at each trimester, the METs that were reported differed between trimesters. Additionally, subject RPE rose 25% (p < .001) in the 2^nd and 3^rd trimester across all exercise modes and intensities, despite no increase in external work rate (Figure 2). In trimester 3, the external work rate of treadmill walking at any percent gradient appears to increase more than the associated VO₂. However, when observing the trend in VO₂ per unit of speed, exercise economy was highest in trimester 3 (1 & 2: 14 vs. 3: 15 (mL/min)/(m/min) p = .03), reflecting the increased metabolic cost of walking exercise with no gradient.

Concerning the significant increase in exercise economy in trimester 3 of pregnancy, we would also like to highlight the variability in VO₂ between subjects (Figure 3), despite determining these steady state work rates from the existing ACSM and ACOG recommended METs. At the beginning of exercise (0 – 10% progress), VO₂ remains low, yet as exercise progresses (20 – 40% progress), a rise in O₂ consumption occurs, with a large variation observed between subjects in trimester 3. Certain segments of the exercise, such as 40 – 50% progress, as well as 70 – 80%, remain at an elevated VO₂ steady state, indicating higher metabolic cost at the same relative work rate.

DISCUSSION

The central hypothesis of this study was that the current estimates for metabolic equivalents during exercise would underestimate energy cost, intensity, and economy of acute bouts of each exercise in each trimester. Our findings show higher actual VO₂ compared to the ACSM-estimated values across both walking and cycling exercise, and throughout each trimester of pregnancy. The ACSM equations consistently underestimated relative VO₂ and METs by ~35% for both low and moderate-intensity exercises. Though absolute VO₂ remained elevated across all trimesters, trimester-specific analyses revealed a significant increase in exercise economy in the 3^rd trimester, indicating that physiological adaptations seen in this gestational period could influence metabolic cost independent of the external work rate during exercise.

The current finding of increased metabolic demand of exercise reflect previous works, which show the significant cardiovascular and metabolic adaptations to pregnancy (25). These findings include but are not limited to increases in metabolic rate at rest, as well as higher plasma and red cell volume, and amplified cardiac output during exercise, all of which are known contributors to O₂ delivery and uptake during exercise (26-28). It is possible that these exacerbations in baseline physiological parameters influence O₂ uptake during exercise in pregnancy such that VO₂ and metabolic cost are increased. Our finding of increased RPE at a set work rate in the 2^nd and 3^rd trimester for both treadmill and cycle exercise occurred independent of an increase in oxygen consumption or metabolic demand. Indeed, previous works have shown difficulty in estimating METs from this population (19). Still, much of the available literature has focused on two polarities of metabolic demand: 1) at rest (29) or 2) during maximal exercise testing (30). Limited attention has been given to studying the actual energetic demand at steady state during prenatal exercise. Thus, exercise guidance during pregnancy continues to require better precision regarding the energy costs and fueling considerations in different modalities and intensities (7, 31, 32).

Building on the misestimation of O₂ uptake and METs during both walking and cycling exercises, we observed an interesting novel relationship between modality and exercise economy. While VO₂ and METs were elevated equally across all trimesters in cycling and walking, exercise economy was significantly altered only in the 3^rd trimester. Essentially, the increase in metabolic demand and O₂ uptake was high enough to increase exercise economy per unit of speed during treadmill walking in the 3^rd trimester, which again was reflected by increased RPE in the 2^nd and 3^rd trimesters. When comparing the economy at different watt levels from incline treadmill walking exercise, no significant differences were observed. This change in 3^rd trimester walking economy per unit of speed suggests that alterations in gait (33), metabolic work of ventilation and other exercise responses (34), or postural support (35) could all potentially lead to increased energy expenditure during 3^rd trimester exercise. Importantly, this explains the mild exercise intolerance (external work rate ~ 5% lower) seen in the 3^rd trimester, and previous reports of higher exhaustion and perceived exertion (36-38). Together, these findings validate increases in perceived exertion in the 3^rd trimester, while also emphasizing that changes in economy might not be present equally across all modalities.

Strengths & Limitations

This study provides the first in-depth analysis of energy expenditure through indirect calorimetry during prenatal exercise at multiple intensities, through 2 different modalities, across each trimester of pregnancy. This also provides the first assessment of modality-specific exercise economy during pregnancy, to our knowledge, and our findings reveal a potential explanation for changes in perceived exertion and mild exercise intolerance in the 3^rd trimester. Still, several limitations should be considered. As our study population was restricted to healthy pregnancies free from gestational conditions, the findings cannot be generalized to all such as those that develop gestational diabetes, hypertension and/or pre-eclampsia, or other adverse pregnancy outcomes. The basis of exercise intensity in the 2^nd and 3^rd trimesters from a 1^st trimester peak exercise test could be a limitation, yet we emphasize that absolute VO₂ was not altered, and exercise at these work rates was well-tolerated late in pregnancy. Additionally, our use of a fixed exercise order helped us control the metabolic demand profile across each trimester, yet exposes the possibility that RPE was influenced by the duration of the trial. Nonetheless, RPEs were noted as higher early in the 2^nd and 3^rd trimester sessions – both during walking (performed first) and cycling. While our study sheds light on the metabolic discrepancies of exercise during pregnancy, future work is needed to explore the influence of other metrics on metabolic demand in prenatal exercise.

CONCLUSION

In conclusion, the current study presents a first-ever analysis of energy expenditure and economy of exercise across healthy pregnancy in each trimester. Our findings demonstrate that commonly employed estimations of exercise energetics derived from non-pregnant populations underestimate the uptake of oxygen to supply bioenergetic processes during aerobic exercise at different intensities of cycling and walking exercise in pregnancy. In particular, walking exercise in the 3^rd trimester induces a notable increase in energy expenditure, which is not directly involved in maintaining the external work rate. Collectively, our findings highlight a critical limitation in the quantification of effort during prenatal exercise, which is an essential component of exercise guidance offered to this population. Specifically, future work should aim to expand this investigation across larger, diverse cohorts, while including a broader range of exercise modalities and intensities to help guide this effort.

Data Availability:

Data generated/analyzed during the current study are available upon request from the corresponding/senior author.