REF-1782 · Oxnard et al., 2026 · Journal of Science and Medicine in Sport, · Paper · PubMed

Exercising more than 900 MET-min a week in pregnancy may go with lighter babies than exercising less

In a systematic review of six studies, the four that compared high with lower volumes found birthweights lower by 44 to 522 g (a range across studies, not pooled; very low certainty by GRADE), while comparisons with women who did not exercise were inconsistent (−206.9 to +120 g); associations between exposure groups, not a tested effect, from the abstract

Main caveat: The authors' own: few studies per outcome, very low certainty for birthweight, clinical significance unclear.
moderate evidenceMeasured on non-athletes to trained athletesRead from the abstract only

Oxnard and colleagues asked whether exceeding the usual antenatal exercise volume recommendations is safe. They searched six databases in May 2023, updated in November 2025, for studies reporting maternal or neonatal outcomes in people who exercised more than 900 MET-min a week in pregnancy, compared with people who exercised less or not at all, and included six studies.

Four studies, rated very low certainty by GRADE, suggest that high-volume exercisers may give birth to lighter infants than lower-volume exercisers (mean differences ranging from −522 to −44 g); comparisons with non-exercisers were inconsistent (−206.9 to +120 g).

Gestational age at birth (two studies), gestational weight gain (three), low or high birthweight (one), small or large for gestational age (one) and preterm birth (three) were reported too rarely to draw conclusions.

The authors state that the clinical significance of the possible lower birthweight, and the links between high-volume exercise and other pregnancy outcomes, remain unclear and need more research.

Participants

Six studies of pregnant people exercising above 900 MET-min a week, with comparison groups exercising less or not exercising.
Number of participants, countries, ages, sports, training or competition level, and study designs: not in the abstract.

Protocol

Search: CINAHL, MEDLINE Complete, Embase, Web of Science, SPORTDiscus, Cochrane Library; May 2023, updated November 2025; PROSPERO CRD42023427569.
Inclusion: maternal or neonatal outcomes reported for participants exercising more than 900 MET-min a week in pregnancy, compared with lower-volume exercise or no exercise.
Certainty: GRADE.
Synthesis: results reported as ranges of mean differences; whether any meta-analysis was attempted is not in the abstract.
Not in the abstract: how exercise volume was measured (questionnaire or device), which trimester it covered, the risk-of-bias tool and its results, and adjustment for confounders.

Key Results

OutcomeStudiesFinding
Birthweight, high vs lower volume4lighter infants with high volume; mean differences −522 to −44 g; very low certainty
Birthweight, high volume vs no exercisenot statedinconsistent; mean differences −206.9 to +120 g
Gestational age at birth2reported; no direction in the abstract
Gestational weight gain3reported; no direction in the abstract
Preterm birth3reported; no direction in the abstract
Low or high birthweight1reported; no direction in the abstract
Small or large for gestational age1reported; no direction in the abstract
Maternal outcomes other than weight gain (for example delivery mode or complications) are not named in the abstract.
No pooled estimates, confidence intervals or heterogeneity statistics are given.

Practical Application

For "how much is too much", the evidence on high training volumes in pregnancy is thin: six studies, a possible lower birthweight against lower-volume exercisers at very low certainty, and too few studies to say anything about preterm birth, gestational age or growth categories. It does not set an upper volume, and the authors call the clinical significance of the birthweight difference unclear.
Where it sits against [SYNTH-95]: the guide keeps a cleared athlete training and notes that vigorous work into the third trimester did not change birthweight (REF-1396, A-, about intensity). This review is about volume, a different exposure, and its signal runs the other way for birthweight against lower-volume exercisers, at very low certainty with no pooled estimate.It does not overturn the guide's rule; it marks high weekly volume as the least-studied part — six studies, very low certainty, no known upper limit — alongside the guide's existing gap on elite training volumes.
For scale (the base's arithmetic, not from the paper; the abstract gives no hours): 900 MET-min a week is, for example, 150 min of an activity rated 6 METs or 90 min of one rated 10 METs, so a woman who keeps training for sport may well be above this threshold (inference); the abstract does not say how far above it the studied groups were.
Inference, not tested here: a woman who trains at high volume in pregnancy tells her obstetric provider her weekly volume, and growth scans or a baby measuring small go to that provider, who sets any limit; the assistant states that a high-volume ceiling is unknown rather than inventing one.
Not treatment advice. Training volume in pregnancy and any change to it are decided with the obstetric provider (SYNTH-95).

Limitations

The authors' own: few studies per outcome, very low certainty for birthweight, clinical significance unclear.
Not pooled: results are ranges across studies with no single estimate or interval, so the size of any birthweight difference is unknown; the −522 g end of the range comes from a study not identified in the abstract.
Exposure: self-selected exercise volume, so women who exercise a lot may differ in other ways (an inference: study designs and confounder adjustment are not in the abstract).
Not in the abstract: number of participants, study designs, how volume was measured, the risk-of-bias tool and its results, sports and training level, whether any study included athletes, which outcomes favoured or disfavoured high volume beyond birthweight, and funding.
Transfer: a volume threshold in MET-min, not intensity, heart rate or sport-specific load; no first-trimester or per-trimester breakdown in the abstract.
How this entry was checked and changed
Why this level

the abstract reports the GRADE certainty for birthweight (very low), which the rubric caps at A- for a review; held at B+, as the base does for reviews of very-low-certainty comparisons of self-selected exercise (REF-1402, REF-1395), because the result is a range across four studies with no pooled estimate or interval, the study designs and the risk-of-bias tool and its results are not in the abstract, and the other outcomes rest on one to three studies