High-protein, low-GI pregnancy diet linked to 11% higher child LDL at age 9

A high-protein, low-glycemic-index diet during pregnancy did not produce better metabolic markers in the children studied at age 9. In a randomized trial, children whose mothers were assigned to that diet had total cholesterol 0.30 mmol/L higher and LDL cholesterol 0.24 mmol/L higher than children in the comparison group. Those differences were 7.4% and 11.3%, respectively. 1

That is a signal to stop treating "high protein plus low GI" as an automatic pregnancy optimization strategy. It is not proof that all higher-protein diets or all low-GI foods are harmful. The trial tested a combined diet in a specific group of pregnant women, and nearly half of the children were lost before the 9-year assessment.

The APPROACH trial enrolled pregnant women with a pre-pregnancy BMI of 28 to 45 kg/m2. Women were cluster-randomized in groups of 4 to 8 at about gestational week 15. The intervention group was guided toward a diet providing 25% of energy from protein and a glycemic index below 55. The comparison group followed a diet based on the Nordic Nutrition Recommendations, with about 18% of energy from protein and no glycemic-index restriction. Both groups could eat freely within their assigned pattern and received up to nine counseling sessions. 2

The diets separated as planned during pregnancy. Mothers in the high-protein, low-GI group consumed 25% of energy from protein, compared with 18.9% in the comparison group. Their estimated dietary GI was about 45 versus 54. The intervention also reduced maternal gestational weight gain by 1.21 kg. That short-term maternal result helps explain why the dietary pattern might look attractive, but it does not tell us whether the pattern benefits the child years later. 1

The child follow-up began with 209 births. One stillbirth in the comparison arm left 104 children in each group at birth. At age 9, 62 children remained in the intervention group and 52 in the comparison group, a 45.5% dropout rate from birth to follow-up. The researchers measured fasting blood lipids and glucose, up to 14 days of blinded continuous glucose monitoring, body composition with DXA, and abdominal and liver fat with MRI. 1

The lipid differences favored the comparison diet, while most other measured metabolic outcomes did not differ.

The LDL-to-HDL ratio was also 0.20 higher in the intervention group, while triglycerides, glucose regulation, blood pressure, pulse, fat mass, visceral fat, subcutaneous fat, and liver fat showed no statistically significant between-group difference. Bone mineral density was 0.03 g/cm2 higher in the intervention group, a secondary finding that does not establish a health advantage. 1

The time pattern matters. The authors report that LDL differences became more apparent with age, while the HDL pattern moved in the opposite direction. But this is still one follow-up of one trial. It does not show that the children will develop cardiovascular disease, and it does not provide a lifetime risk estimate.

Random allocation gives this result more weight than an observational link between maternal diet and child cholesterol. The study also used an objective protein-intake biomarker during pregnancy and assessed children with fasting laboratory tests and imaging. Those are useful safeguards against relying only on dietary recall.

The design cannot separate the effects of protein from the effects of lowering GI, because the intervention changed both at once. The original trial was powered to detect a difference in maternal gestational weight gain, not differences in offspring outcomes. The authors also could not explain the biological mechanism behind the cholesterol result. Missing dietary data at age 9 reduced the precision of comparisons, and parent-reported pubertal status may have been misclassified. Most importantly, the high loss to follow-up may have introduced selection bias and limits how confidently the result can be generalized. 1

The paper reports no competing interests. It also states that the named funders did not take part in the study design, data collection, analysis, interpretation, or writing. 1

If you are pregnant and considering a high-protein, low-GI plan mainly to control weight or improve your baby's future metabolic health, do not start by chasing the trial's targets of 25% protein and a GI below 55 on your own. Take the proposed menu to your obstetric clinician or a registered dietitian, especially if you have overweight, obesity, diabetes risk, or a history of pregnancy complications.

For a same-day default, keep the diet varied: include protein-rich foods each day and choose higher-fiber starchy foods such as wholegrain bread, brown rice, oats, or potatoes with their skins. That matches general pregnancy guidance without turning one study into a universal restriction. The NHS also advises eating a variety of foods, plenty of fruit and vegetables, and protein foods daily rather than following a special diet. 3

The practical conclusion is narrow but useful: a lower GI and higher protein target may help a mother's short-term weight trajectory, yet this randomized follow-up did not show a corresponding benefit for the child's glucose, body fat, or liver fat, and it found higher LDL and total cholesterol at age 9. Until longer follow-up and component-specific trials clarify the result, pregnancy nutrition should be personalized around adequacy, variety, and clinical needs, not a high-protein, low-GI target pursued as a guarantee of better offspring metabolism.

Read the full open-access paper and the trial registration.