Early-pregnancy diet quality tracks with fetal growth, by small amounts

Design

secondary analysis of a randomised clinical trial, adjusted linear regression

Population

238 Brazilian pregnant women with overweight; 192 mother–infant pairs at six months

Primary outcome

infant anthropometry at birth and at six months by early-pregnancy diet quality score

Effect

birth weight β = 12.311 g per point (95% CI 5.163 to 19.458); infant length at 6 months β = 0.061 cm (0.019 to 0.103)

A secondary analysis of a Brazilian randomised trial followed 238 pregnant women with overweight, with 192 mother–infant pairs at six months. Diet was captured by 24-hour recalls between 3 and 15 weeks and scored with a diet quality index adapted for pregnancy; models were adjusted for confounders including the original trial arm.

Higher diet quality was associated with higher birth weight (β = 12.311 g per index point, 95% CI 5.163 to 19.458), birth length (β = 0.050 cm, 0.015 to 0.084) and their Z-scores, and inversely with small-for-gestational-age and low birth weight. At six months the association persisted for length and length-for-age Z-score but not for weight.

The effects are small per point of score, and the exposure was measured once in early pregnancy. What it supports is the familiar message — diet quality in the first trimester matters for growth, and the signal looks more like linear growth than adiposity — not any particular dietary prescription.

• Ask about diet quality at the booking visit, not only about weight gain
• Frame counselling around overall dietary pattern rather than single nutrients
• In women with overweight, do not assume adequate intake — quality and quantity are different questions
• Follow infant length as well as weight where early-pregnancy nutrition was poor

Why it matters

It puts the nutrition conversation in the first trimester, where most antenatal care still puts it at 28 weeks.

Don't overread it

This was an observational analysis inside a trial — it cannot show that improving diet quality changes birth weight.

The statistics, in plain English

The betas are per unit of the index score, so 12 g of birth weight per point is not 12 g in total — the clinical size depends on how far apart two women's scores realistically are. The confidence intervals exclude zero but sit close to it, which is what a modest real association looks like in a few hundred pairs.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.