All topics › Pregnancy preparing for birth
How much exercise is safe and beneficial in pregnancy, and what does the evidence say about miscarriage, preterm birth, and diabetes risk?
Pregnancy comes with a folk fear that exertion harms the baby — and a medical history of prescribing rest. The modern evidence points the other way for uncomplicated pregnancies. So: how much should you move, what actually improves, and where are the real limits?
The American College of Obstetricians and Gynecologists (ACOG, Committee Opinion 804, 2020) recommends that women with uncomplicated pregnancies aim for at least 150 minutes a week of moderate-intensity aerobic activity, spread across the week — for example 20–30 minutes on most days [1]. Those who were already active can generally continue, including vigorous exercise. The headline from the evidence: regular exercise in uncomplicated pregnancy lowers the risk of gestational diabetes and hypertensive disorders, and the long-feared harms — miscarriage, poor fetal growth, preterm birth — have not been substantiated [1][2].
The caveats matter: most benefit trials enrolled women with uncomplicated, singleton pregnancies; ACOG lists absolute contraindications (such as placenta previa after 26 weeks, ruptured membranes, or pre-eclampsia) where exercise is off the table [1]; and the evidence is thinnest on exactly how hard you can push, especially competitive-level training.
The target comes from the US Physical Activity Guidelines (2018), endorsed by ACOG: at least 150 minutes of moderate-intensity aerobic activity per week during pregnancy and postpartum, spread through the week [1]. "Moderate" is a feel thing, not a heart-rate zone — ACOG suggests the talk test (you can hold a conversation while exercising) and a perceived-exertion rating of 13–14 on the Borg scale ("somewhat hard"), because heart-rate responses in pregnancy are unreliable [1]. Previously sedentary women should build up gradually; previously active women can generally continue what they were doing [1].
Gestational diabetes affects roughly 7% of pregnancies [3]. Across 27 pooled trials in a 2018 systematic review, 271 of 3,505 exercisers (77 per 1,000) developed it versus 380 of 3,429 controls (111 per 1,000) — 34 fewer cases per 1,000; pooled OR 0.62 (95% CI 0.52–0.75) [2]. An umbrella review of 23 reviews found the benefit was most consistent with supervised, low-to-moderate exercise begun in the first trimester, and its updated meta-analysis of 35 trials found the same pattern (OR 0.61, 95% CI 0.51–0.74) [4]. Not every review agrees — of the umbrella's 20 gestational-diabetes reviews, 13 found a reduction and 7 didn't — but the pooled trial data point the same way [4]. A 2017 trial of 300 overweight or obese women found that cycling 30 minutes, 3 times a week from the first trimester significantly cut gestational diabetes [1].
For gestational hypertension, the 2018 review's RCT data give actual event counts: 61 of 2,627 exercisers (23 per 1,000) versus 105 of 2,689 controls (39 per 1,000) — 16 fewer per 1,000; pooled OR 0.61 (95% CI 0.43–0.85) [2]. A 2023 update across 22 trials found the same direction (OR 0.53, 95% CI 0.40–0.71) [4]. For pre-eclampsia — dangerously high blood pressure in pregnancy, affecting roughly 3–8 in 100 women who give birth worldwide [5] — the 2018 review found 34 of 1,719 exercisers (20 per 1,000) versus 49 of 1,603 controls (31 per 1,000); OR 0.59 (95% CI 0.37–0.94) [2]. But a larger 2023 update pooling 20 trials found no overall reduction (OR 0.81, 95% CI 0.61–1.07), with a reduction only when exercise was started in the first trimester (OR 0.34) [4]. Treat any pre-eclampsia benefit as uncertain.
ACOG's summary is direct: concerns that regular activity causes miscarriage, poor fetal growth, musculoskeletal injury, or premature delivery "have not been substantiated for women with uncomplicated pregnancies" [1]. Three meta-analyses found birthweight differences between exercisers and controls were minimal to none — with one nuance: women who kept up vigorous exercise into the third trimester delivered babies 200–400 g lighter on average, without any increased risk of fetal growth restriction [1].
The main conflicting evidence is a Danish cohort of 92,671 pregnancies (Madsen et al., 2007) that reported a 3.7× higher early-miscarriage rate with more than 7 hours a week of exercise versus none (HR 3.7, 95% CI 2.9–4.7), and with high-impact activity — a relative estimate; the paper doesn't give absolute rates in a form that converts reliably [7]. The association was absent after 18 weeks of gestation. The authors themselves warned the data were collected retrospectively and could explain the association — it's a signal to discuss, not a finding to act on, and it stands against the meta-analytic and guideline consensus [1][7].
Relatedly: bed rest has no credible evidence as a way to prevent preterm labour and should not be routinely prescribed — prolonged inactivity brings its own harms (blood clots, bone loss, deconditioning, and negative psychosocial effects) [1].
ACOG's absolute contraindications — where aerobic exercise should not be done — include haemodynamically significant heart disease, restrictive lung disease, incompetent cervix or cerclage, multiple gestation at risk of premature labour, persistent second- or third-trimester bleeding, placenta previa after 26 weeks, premature labour in the current pregnancy, ruptured membranes, pre-eclampsia or pregnancy-induced hypertension, and severe anaemia [1]. Activities to avoid: contact sports, activities with a high risk of falling (e.g. gymnastics, downhill skiing), scuba diving, skydiving, and hot yoga or hot Pilates [1]. Well-studied safe options include walking, stationary cycling, dancing, resistance exercises, stretching, and water aerobics [1].
Stop exercising and seek advice if you experience: vaginal bleeding, abdominal pain, regular painful contractions, amniotic fluid leakage, breathlessness before exertion, dizziness, headache, chest pain, muscle weakness affecting balance, or calf pain or swelling [1].
This is the rare pregnancy topic where the parent's own body is the primary beneficiary.
Mood and mental health. The clearest parental outcome in the trials is lower prenatal depressive symptoms with exercise [6]. Pregnancy is a high-risk window for depression, and a daily walk is one of the few interventions with RCT support behind it — no prescription needed. The evidence doesn't clearly extend to anxiety or to preventing postpartum depression [6], so exercise is a support, not a treatment plan; low mood still deserves a conversation with your midwife or GP.
Back pain. More than 60% of pregnant women experience low back pain [1], and a Cochrane review of 34 trials (5,121 women) found exercise improved low-back pain and disability — though the evidence quality was low and the interventions varied widely [8]. Water exercise is ACOG's suggested alternative when land-based exercise hurts [1]. If your back is the thing that's stopping you moving, that's exactly the symptom exercise is most likely to help.
Sleep, energy, and recovery. Exercise trials in pregnancy rarely measure sleep or daytime energy well — that's an evidence gap — but postpartum, women who exercised tend to recover faster [1], and pelvic-floor and abdominal exercises started soon after birth reduce diastasis recti (abdominal separation) [1]. Lactating women can exercise without affecting milk production, composition, or infant growth [1].
Partners: the evidence base has nothing to say about non-pregnant partners' exercise. If a partner wants a role, joining the walk is a fine one — but that's common sense, not science.
| Intervention / comparison | Outcome | Effect in absolute terms | Evidence rating |
|---|---|---|---|
| Exercise-only RCT interventions vs control | Gestational diabetes | 271 of 3,505 exercisers (77 per 1,000) vs 380 of 3,429 controls (111 per 1,000) — 34 fewer per 1,000; pooled OR 0.62 (95% CI 0.52–0.75) [2]. A 2023 update of 35 trials found the same pattern (OR 0.61, 95% CI 0.51–0.74) [4] | B |
| Exercise vs control (RCT event counts) | Gestational hypertension | 39 → 23 per 1,000 (16 fewer per 1,000); OR 0.61 (95% CI 0.43–0.85) [2]. A 2023 update across 22 trials found the same direction (OR 0.53, 95% CI 0.40–0.71) [4] | B |
| Exercise vs control (RCT event counts) | Pre-eclampsia | 2018 review: 34 of 1,719 (20 per 1,000) vs 49 of 1,603 controls (31 per 1,000); OR 0.59 (95% CI 0.37–0.94) [2]. But a larger 2023 update pooling 20 trials found no significant reduction (OR 0.81, 95% CI 0.61–1.07) [4]. Effect uncertain | C |
| Aerobic exercise (overweight/obese subgroup) | Preterm birth | RR 0.62 (95% CI 0.41–0.95) across 9 trials — subgroup finding; no reliable absolute conversion [9]. In normal-weight women: 4.5% vs 4.4%, RR 1.01 (95% CI 0.68–1.50) — no association either way [10] | C |
| Exercise vs control | Caesarean birth | Lower in observational studies and some trials; less consistent than GDM/hypertension effects [1]. In normal-weight women: 17.9% vs 22% (RR 0.82) [10] | C |
| Exercise vs control (RCTs) | Prenatal depressive symptoms | Reduced symptoms and depression odds; no verified absolute rates [6] | C |
| Exercise vs control | Low back pain / disability | Improved, but low-quality evidence, varied interventions [8] | C |
| Vigorous exercise into 3rd trimester | Birth weight | 200–400 g lighter on average; no increased fetal growth restriction [1] | B (for the null on harm) |
| >7 h/week high-impact exercise (observational) | Early miscarriage | HR 3.7 (95% CI 2.9–4.7) vs none — relative estimate, no reliable absolute conversion; no association after 18 weeks; retrospective data, authors cautioned [7] | D (conflicting) |