Can I keep doing vigorous sessions late in pregnancy?
In low-risk pregnancies (12 studies, 8,006 babies), vigorous work into the third trimester showed no difference in birth weight (+8 g, −57 to 74) or small babies. Mostly <90% HRmax.
Main caveat: Mostly observational evidence. Most pooled women come from cohort and case-control studies with mixed control of confounding (four studies adjusted for none).
Beetham and colleagues asked whether vigorous exercise during pregnancy, and in particular in the third trimester, when fetal demands are highest, harms the infant or the mother.
They searched six databases to November 2018 for RCTs, quasi-experimental, cohort and case-control studies in which pregnant women did vigorous exercise against lower-intensity exercise or standard care, and reviewed each study's methods against the Exercise and Sports Science Australia intensity terms (vigorous: at least 70% of maximum heart rate, or unable to hold a conversation; high intensity: 90% or more).
Fifteen studies were included: 5 RCTs (623 women) and 10 observational studies (32,080), of which 13 went into random-effects multi-level meta-analyses. Planned trimester comparisons were not possible: the data allowed only exercise continued into the third trimester.
Birth weight did not differ between infants of women who did and did not exercise vigorously (mean difference 8.06 g, 95% CI −57.44 to 73.55, 12 studies, 8,006 infants), with no difference by study design or comparison group.
Small-for-gestational-age and low birth weight showed no significant increase, maternal weight gain did not differ (−0.46 kg, −2.05 to 1.12), gestation was 0.21 weeks longer (0.15 to 0.27; 39+4 vs 39+3 weeks) and prematurity was slightly lower. The RCTs reported no excess adverse events, but their programmes started late in the first trimester or in the second.
Most studies stayed below 90% of maximum heart rate; the authors conclude that vigorous exercise into the third trimester "appears to be safe for most healthy pregnancies", and advise women to avoid exertion relative to 90% of maximum heart rate or more until research confirms its safety.
Our reading: this is the base's pooled answer for vigorous work late in pregnancy in the general population, reassuring on growth and gestation; the ratios reported as risk ratios are log risk ratios (the forest plots are labelled so), which turns the small-for-gestational-age estimate into about 1.16 (0.94 to 1.42) and the prematurity estimate into about 0.82 (0.70 to 0.97) on the usual scale (our arithmetic); and the advice to stay under 90% is a precaution resting on a 6-athlete study (REF-1400), not a tested limit.
Protocol
Design: systematic review and meta-analysis to PRISMA, PROSPERO CRD42018102109. Databases: PubMed, Medline, EMBASE, Cochrane Library, Web of Science, CINAHL, to November 2018; English only.
Eligibility: RCTs, quasi-experimental, cohort and case-control studies; pregnant women of any age doing vigorous exercise in any trimester, compared with lower-intensity exercise or standard care; at least one of birth weight, small for gestational age (SGA, below the 10th percentile), low birth weight (LBW, under 2500 g), gestational age at delivery, preterm birth (under 37 weeks) or gestational weight gain. Studies of women with gestational diabetes were excluded.
Intensity: each study's methods were judged against the Exercise and Sports Science Australia terms, whatever the authors called the intensity; three studies labelled moderate were re-classified as vigorous (70% of heart rate reserve, 70% of maximum heart rate, and "getting puffed").
Included: n = 32703 women in 15 studies (5 RCTs, n = 623; 10 observational, n = 32080; our sum). Two studies (one that stopped vigorous exercise after the second trimester, one that recorded it in the first 15 weeks only) were reviewed but not pooled, leaving 8 observational studies (n = 7225) and 5 RCTs (n = 623) in the analyses.
RCTs: Canada (walking at 70% of heart rate reserve vs 30%, 3-4 times a week, 25-40 min), Canada (supervised aerobic and resistance training at 70% of maximum heart rate, 3 times a week, 60 min), Brazil (water aerobics at 70% of maximum, 3 times a week, 50 min), New Zealand (home cycle ergometer at 65% VO2max, up to 5 times a week, 40 min), China (supervised cycle ergometer, 30 min with 3-5 intervals of 30 s at 75-85% of maximum, 3 times a week); randomised at 10 weeks, before 14 weeks, or at 16-20 weeks.
Observational studies: 1983-2018, from the USA, Australia, Denmark, the UK and an international cohort; exposures from "vigorous walking" and "working up a sweat" to 60-80% and 85% of maximum heart rate, ≥150 beats/min, and running.
Analysis: random-effects multi-level meta-analysis (metafor in R) with cluster-robust standard errors; mean differences for continuous outcomes, ratios for dichotomous ones (presented on the log scale in the forest plots); moderation by design and comparison group; I² with confidence intervals.
Risk of bias: RCTs mixed (Cochrane tool; attrition poorly reported); cohorts scored 4 to 9 of 9 and the two case-control studies 2 and 3 of 9 on the Newcastle-Ottawa Scale; four studies adjusted for no confounder.
Funding and interests: funding to the first author from the Faculty of Health Sciences at Australian Catholic University to pay for help with the search, data extraction and manuscript; "The funding source had no role in any aspect of the study." The authors declare no competing interests. No device maker, federation or data platform is named.
Key Results
| Outcome, vigorous exercise into the third trimester vs control | Studies (infants or women) | As printed (95% CI) | On the usual ratio scale (our arithmetic) |
|---|---|---|---|
| Birth weight | 12 (8,006) | +8.06 g (−57.44 to 73.55), p = 0.79, I² 53.92 [0, 85.03] | — |
| Small for gestational age | 7 (4,504) | RR as printed 0.15 (−0.06 to 0.35), p = 0.13 | RR about 1.16 (0.94 to 1.42) |
| Low birth weight (<2500 g) | 4 (2,454) | RR as printed 0.44 (−0.83 to 1.7), p = 0.35 | RR about 1.55 (0.44 to 5.47) |
| Gestational age at delivery | 7 (4,281) | +0.21 weeks (0.15 to 0.27), p < 0.001; 39+4 vs 39+3 weeks | about 1.5 days |
| Prematurity | 4 (3,025) | RR as printed −0.20 (−0.36 to −0.03), p = 0.03 | RR about 0.82 (0.70 to 0.97) |
| Prematurity, the two RCTs only | 2 (312) | RR as printed −0.41 (−1.64 to 0.82) | RR about 0.66 (0.19 to 2.27) |
| Maternal weight gain | 7 (1,834) | −0.46 kg (−2.05 to 1.12), p = 0.5, I² 68.94 [0, 95.2] | — |
Reading the nulls: for birth weight the interval runs from about 57 g lighter to 74 g heavier, so a clinically large difference is unlikely; for SGA the interval on the ratio scale runs from about 6% lower to 42% higher risk, and for LBW from 56% lower to over five-fold higher (our arithmetic) — no evidence of a difference, but LBW is imprecise.
Moderators: birth weight results held across retrospective, prospective and RCT designs and against moderate-or-less and light-or-less comparison groups; the gestational-age gain was similar in size but not significant within each design; the prematurity reduction did not replicate in the two RCTs alone.
RCTs vs observational studies: the RCTs showed a slightly lower, non-significant birth weight than controls, not seen in the cohorts; SGA risk was still not raised.
Weight gain: no overall difference, but the two RCTs in overweight or obese women reduced gestational weight gain; in one Canadian trial 53% of non-exercising controls had excessive weekly weight gain vs 31% with vigorous and 35% with light walking.
Adverse events: all studies were in low-risk women; the RCTs showed no increase in adverse events (in the Chinese trial, 4 miscarriages with exercise vs 3 miscarriages and 1 fetal death with standard care), but the programmes often started only late in the first trimester or early in the second (randomisation from 10 weeks), so first-trimester miscarriage is not addressed.
Not pooled: an international prospective cohort of 3,513 first-time mothers found three-fold the risk of an SGA infant with daily vigorous exercise in the first 15 weeks; it did not record exercise later in pregnancy.
Intensity reached: most included studies were vigorous (under 90% of maximum heart rate), not high intensity (90% or more).
Mechanisms
Measured (pooled outcomes only): no change in birth weight, SGA or LBW; slightly longer gestation and fewer preterm births with vigorous exercise into the third trimester. The review measured no blood flow, fetal heart rate or placental function.
Authors' interpretation: vigorous exercise transiently redirects blood from the uterus to working muscle; brief hypoxia in the first and second trimesters may even stimulate placental growth, while uterine blood flow rises from about 50 mL/min in the first trimester to 500 mL/min in the third and the fetus has compensatory responses to brief reductions.The lower birth weight seen in some earlier studies is thought to reflect less fetal fat, not lean mass. The fall in prematurity more likely means that women who exercise vigorously, or join exercise studies, are healthier with lower-risk pregnancies than that exercise prevents preterm birth.
Inference: birth outcomes integrate months of growth; they cannot show whether a single very hard effort briefly stresses the fetus. That acute question is answered, in small samples, by REF-1400 (fetal bradycardia above about 90% HRmax in Olympic athletes) and REF-1401 (none after 1-min intervals peaking near 96% HRmax).
Practical Application
Can I keep doing vigorous sessions late in pregnancy? In low-risk pregnancies (12 studies, 8,006 babies), vigorous work into the third trimester showed no difference in birth weight (+8 g, −57 to 74) or small babies. Mostly <90% HRmax.
Vigorous work (about 70-90% of maximum heart rate) into the third trimester: no sign of harm to growth (A-, mostly observational, low-risk women). This supports, for the general population, what governing bodies say for previously vigorous exercisers (REF-567) and goes beyond the 2019 Canadian guideline's moderate-intensity floor (REF-568). It is not clearance: the obstetric team rules out contraindications first (REF-567).
Above 90% is untested, not shown harmful: the authors advise avoiding exertion relative to 90% or more of maximum heart rate "until further research can confirm its safety" — a precaution resting on a study of Olympic athletes (REF-1400) and one other small study, not a limit any trial tested.In 15 women, 1-min reps peaking near 96% HRmax caused no fetal bradycardia (REF-1401).The authors also recommend perceived exertion over heart rate, because maximum heart rate is hard to measure in pregnancy; the only heart-rate range in the paper, 142 to 169 beats/min for vigorous work, is one the authors attribute to the 2019 Canadian guideline, which the base's entry (REF-568) does not carry and has not checked.Not from this paper; the base's own practice: perceived effort and the talk test govern, and any number comes from the obstetric team.
If the baby is measuring small: the authors suggest moderate rather than vigorous exercise in the third trimester (their caution from a non-significant lower birth weight in the RCTs, not a tested rule).
Early pregnancy is the thin part: no pooled data exist for vigorous exercise in the first or second trimester alone, and one large cohort (not pooled) linked daily vigorous exercise in the first 15 weeks with three-fold SGA risk; first-trimester miscarriage was not covered by any RCT.
Weight gain: vigorous exercise did not change maternal weight gain overall; the two trials in overweight or obese women did reduce it.
What it is good for as a source: the pooled general-population answer on birth weight, SGA, LBW, gestation and weight gain for vigorous exercise continued into the third trimester. What it does not show: safety above 90% of maximum heart rate or of sustained near-maximal efforts; effects of elite training volumes; first-trimester effects; fetal responses during a session; outcomes such as pelvic floor or diastasis after heavy resistance training (the authors list these as future work).
Limitations
Mostly observational evidence. Most pooled women come from cohort and case-control studies with mixed control of confounding (four studies adjusted for none); women who choose vigorous exercise differ from those who do not.
Intensity is loosely defined. Studies described intensity differently (from "vigorous walking" to 85% of maximum heart rate); the reviewers re-classified three, and most stayed under 90%.
Only the third trimester could be analysed; the RCT programmes often did not start until late in the first trimester or early in the second, so first-trimester effects and miscarriage are not answered.
Reporting. The values labelled risk ratios for SGA, LBW and prematurity are log risk ratios (negative values and the forest-plot labels show it); the abstract's risk ratio of 0.15 for SGA reads like an 85% reduction but corresponds to about 1.16 (our arithmetic).The birth-weight analysis lists 8,006 infants although the text says 7,225 observational and 623 RCT participants entered the analyses (7,848; our arithmetic). The review describes the elite-athlete study it cites (REF-1400) as seven athletes, two of whom exceeded 90% of maximum heart rate; that study's abstract says six.
Wide heterogeneity intervals. I² point estimates were modest, but with few studies their confidence intervals were very wide (up to 0 to 95.2), so unexplained heterogeneity cannot be excluded.
No certainty grading (no GRADE); English-language studies only.
Funding and interests: funded by the first author's faculty at Australian Catholic University, with no role in the study; no competing interests declared.
How this entry was checked and changed
Why this level
a registered systematic review with meta-analysis; held below A because most pooled women come from observational studies (8 cohort or case-control studies vs 5 RCTs with 623 women), vigorous intensity was defined inconsistently across studies and re-classified by the reviewers, the confidence intervals of I² are very wide (e.g. 0 to 85.03 for birth weight), no GRADE or other certainty rating was made, and the ratios are reported on the log scale while labelled risk ratios. Read from the full text (Europe PMC, CC BY 4.0); the three additional files (search terms, intensity definitions, moderation tables) not read
Full text (checked 2026-10-28)
Ten cohort studies (n = 32,080) and five randomized control trials (n = 623) were included in the systematic review, with 13 studies included in the meta-analysis (full text)
Vigorous and high-intensity exercise is defined as being at least 70% of maximum heart rate (MHR), or an activity in which a conversation generally cannot be maintained; most studies described vigorous intensity (< 90% MHR) and not high-intensity (≥90% MHR) (full text)
Birth weight: mean difference 8.06 g, 95% CI −57.44 to 73.55, p = 0.79, n = 8006, k = 12, I^2 = 53.92 [0, 85.03] (full text)
Small for gestational age: RR = 0.15, 95% CI −0.06 to 0.35, p = 0.13, n = 4504, k = 7; low birth weight: RR = 0.44, 95% CI −0.83 to 1.7, p = 0.35, n = 2454, k = 4; the forest plots are labelled "Log risk ratio" (full text)
Gestational age at delivery: mean difference = 0.21 weeks; 95% CI 0.15 to 0.27, n = 4281, k = 7; 39 + 4 weeks vs. 39 + 3 weeks (full text)
Prematurity: RR = −0.20; 95% CI −0.36 to −0.03, p = 0.03, n = 3025, k = 4; the two RCTs alone: RR = −0.41; 95% CI −1.64 to 0.82, n = 312 (full text)
Maternal weight gain: mean difference = −0.46 kg, 95% CI −2.05 to 1.12, p = 0.5, n = 1834, k = 7, I^2 = 68.94 [0, 95.2] (full text)
A three-fold risk of delivering a SGA infant in a prospective study of 3513 primiparous mothers who reported daily vigorous intensity exercise in the first 15 weeks of pregnancy; not included in the meta-analysis (full text)
On the cited elite-athlete study: "Only two of the seven athletes reached an intensity greater than 90% MHR" (full text)
"Pregnant women should avoid exercising at a perceived exertion relative to ≥90% MHR, until further research can confirm its safety" (full text)
The Canadian guideline is cited for a target heart rate of between 142 and 169 bpm for vigorous work, depending on age (full text)
Funding was provided by the Faculty of Health Sciences at Australian Catholic University; "The funding source had no role in any aspect of the study"; the authors declare that they have no competing interests (full text)