Edited by Teri L. Hernandez, PhD, RN, FAAN
It is the most common instruction in gestational diabetes care. It is also the hardest one to follow up on inside a 15-minute visit.
The problem states itself in one line. There are three macronutrients. Take one away and something fills the gap. Protein only goes so far before it becomes intolerable, so in practice what fills the gap is fat.
Outside pregnancy that is a manageable trade. Inside pregnancy it is less obviously so, because maternal lipids are not bystanders in fetal growth.
The number is older than it looks
The conventional prescription of roughly 40% carbohydrate dates to 1990. Its logic traces back further, to the pre-insulin era, when a near-carbohydrate-free, hypocaloric diet was the only thing standing between a patient with type 1 diabetes and death. The Joslin diet of that period ran about 2% carbohydrate and 70 to 80% fat.
Nothing about that origin makes the prescription wrong. Plenty of good medicine started somewhere strange. But it does mean the number was never derived from pregnancy physiology, and it is worth asking what the evidence has shown since.
What the evidence shows is mostly an absence
The 2017 Cochrane review of dietary advice in gestational diabetes could not identify a superior macronutrient prescription, and rated the quality of the evidence as low. What does hold up is glycemic index, which is a question of carbohydrate quality rather than quantity.
The more interesting finding sits in a meta-analysis of 18 randomized trials covering 1,151 women: almost any improvement in dietary pattern made after gestational diabetes diagnosis produced lower fasting and postprandial glucose, lower birth weight, and a reduced need for adjunctive medication. The benefit tracked with improvement from baseline, not with which diet was assigned.
That is an uncomfortable finding for anyone attached to a specific prescription, and a reassuring one for anyone who has watched a patient fail to follow one.
The trial that controlled for the thing nobody controls for
Nutrition research in pregnancy has two chronic weaknesses. Investigators cannot verify what participants actually eat, and medication confounds the glycemic outcome.
A team at the University of Colorado removed both. In a randomized trial of 46 patients with gestational diabetes, every meal was prepared and provided for the final seven to eight weeks of pregnancy, and no participant received pharmacotherapy. One arm received the conventional plan at 40% carbohydrate, 45% fat, and 15% protein. The other received 60% carbohydrate, almost entirely complex and low to medium glycemic index, 25% fat, and 15% protein. The diets were eucaloric and matched for total sugars.
The primary outcome, newborn adiposity, did not differ between groups. Nor did birth weight, percent body fat, rates of large- or small-for-gestational age, or cord insulin and C-peptide.
Glycemia did not differ either. Both arms held roughly 90% time in range over 24 hours, on continuous glucose monitoring, at both 30 to 31 weeks and again at 36 to 37 weeks.
Three secondary findings ran against the conventional direction. Triglyceride area under the curve rose 14% on the higher-complex carbohydrate arm versus 20% on the conventional arm. Postprandial free fatty acid exposure was about 20% lower on the higher-complex carbohydrate arm, having risen on the conventional arm and fallen on the other. And glucose tolerance improved between 30 and 36 weeks only in the higher-complex carbohydrate group, a difference that was still present when participants were retested roughly two months postpartum.
Before anyone over-reads that
Forty-six patients is 46 patients. The trial was powered for a difference in newborn adiposity and found no difference, which is the honest headline.
Every meal was made and delivered. That is what makes the comparison clean and also what makes it difficult to generalize, because nobody’s patient is going to be fed for eight weeks. Real-world adherence to a 60% complex carbohydrate pattern of that quality is an open question, and the only other randomized attempt to separate carbohydrate targets, an Australian trial, reported 20% compliance in its lower-carbohydrate arm (~165 grams/day).
None of this says a lower-carbohydrate plan is harmful. It says superiority has not been demonstrated in either direction, which is a narrower and more defensible claim.
Four things that are specific to pregnancy
The maternal-fetal glucose gradient. The human fetus draws roughly 80% of its energy from maternal glucose, predominantly for brain growth, and transfer depends on facilitated diffusion down a concentration gradient. Observational data associate carbohydrate intake in the range of 47 to 70% of energy with normal fetal growth after adjustment for total energy, with lower birth weights and higher small-for-gestational-age rates at the lowest intakes.
Ketone exposure. Carbohydrate restriction raises the glucagon to insulin ratio and promotes ketonemia. In vitro, ketone exposure suppresses glucose uptake in first-trimester trophoblasts. The human data are thin and old, which is precisely the problem: the most cited study, from 1991, found third-trimester maternal ketones in women with pre-existing diabetes inversely correlated with offspring intelligence scores at age two, and nobody has been eager to run the trial that would settle it.
Micronutrient adequacy. Whole grains carry B vitamins, selenium, potassium and magnesium, and maternal micronutrient intake is already commonly below recommendation. The Australian trial found lower iron and iodine intake in its lower-carbohydrate arm, along with a smaller newborn head circumference.
Maternal triglycerides. This is the one that gets the least attention in clinic. Maternal triglycerides rise 40% or more across gestation by design, under placental estrogen, and a growing body of evidence implicates maternal lipids as predictors of fetal overgrowth at least as strong as glucose. A prescription that lowers carbohydrate raises fat by arithmetic necessity. We have not generally measured what that does.
The recommendation never counted the placenta
The pregnancy RDA for carbohydrate accounts for maternal brain and fetal brain. It does not account for the organ in between.
The placenta expresses the same glucose transporters as brain and relies on maternal glucose as its dominant fuel. Using published data from the four-vessel sampling model at cesarean delivery, Hernandez and Rozance calculated obligate placental glucose consumption at roughly 36 g per day. Added to the existing derivation, the figure moves the RDA for carbohydrate intake during pregnancy from 175 g to approximately 220 g per day.
Whether that becomes a revised RDA is for other people to decide. As a clinical observation it is hard to unsee: we may be prescribing below requirement before the patient restricts anything on her own.
What this changes, and what it does not
It does not change first-line therapy. Nutrition is still first-line, and a patient already doing well on her current plan should stay on it.
What it changes is the shape of the conversation. The useful question is not how many grams, it is what kind of carbohydrate, inside what total energy, with how much added sugar. That conversation is also more culturally portable: whole grains, beans, lentils, roots and tubers appear in essentially every regional diet pattern, which a gram ceiling does not.
The part the chart cannot show you
Here is what makes all of this hard to act on, and it has nothing to do with clinical knowledge.
You have twelve to fifteen minutes. You are managing her blood pressure, her growth scan, her anxiety about delivery, and the two other things she mentioned at the door. You give the instruction that fits in the time you have, which is cut the carbs, because it is the one she will remember.
Then she goes home and you do not see what happens next. You do not see the grocery store, or the Tuesday night when she is exhausted, or the family dinner she is now eating around instead of with. You see a number at her next visit, and the number tells you what happened but never why.
That is not a clinical knowledge gap. It is a visibility gap, and no amount of additional evidence closes it, because the missing data was never going to be in the chart.
Kari Cao, PA-C, is co-founder and CEO of RenewRx Health, which provides between-visit nutrition and glucose support for pregnant patients in Colorado.
This article draws on a lecture given by Teri L. Hernandez, PhD, RN, FAAN, University of Colorado Anschutz Medical Campus, to the Association of Diabetes Care and Education Specialists, Colorado, September 2026.
References
- Hernandez TL, Farabi SS, Fosdick BK, et al. Randomization to a provided higher-complex-carbohydrate versus conventional diet in gestational diabetes mellitus results in similar newborn adiposity. Diabetes Care. 2023;46(11):1931-1940. doi:10.2337/dc23-0617. https://doi.org/10.2337/dc23-0617
- Hernandez TL, Farabi SS, Fosdick BK, et al. A higher-complex carbohydrate compared with conventional low-carbohydrate diet in gestational diabetes mellitus results in lower postprandial free fatty acids: a randomized controlled trial. Am J Clin Nutr. 2026;124(5):101525. doi:10.1016/j.ajcnut.2026.101525. https://doi.org/10.1016/j.ajcnut.2026.101525
- Hernandez TL, Rozance PJ. Re-examination of the estimated average requirement for carbohydrate intake during pregnancy: addition of placental glucose consumption. Am J Clin Nutr. 2023;117(2):227-234. https://pmc.ncbi.nlm.nih.gov/articles/PMC10196558/
- Han S, Middleton P, Shepherd E, Van Ryswyk E, Crowther CA. Different types of dietary advice for women with gestational diabetes mellitus. Cochrane Database Syst Rev. 2017;2:CD009275.
- Yamamoto JM, Kellett JE, Balsells M, et al. Gestational diabetes mellitus and diet: a systematic review and meta-analysis of randomized controlled trials examining the impact of modified dietary interventions on maternal glucose control and neonatal birth weight. Diabetes Care. 2018;41(7):1346-1361.
- Mijatovic J, Louie JCY, Buso MEC, et al. Effects of a modestly lower carbohydrate diet in gestational diabetes: a randomized controlled trial. Am J Clin Nutr. 2020;112(2):284-292.
- Rizzo T, Metzger BE, Burns WJ, Burns K. Correlations between antepartum maternal metabolism and child intelligence. N Engl J Med. 1991;325(13):911-916.
- Barbour LA, Hernandez TL. Maternal lipids and fetal overgrowth: making fat from fat. Clin Ther. 2018;40(10):1638-1647.
- Hernandez TL, Barbour LA. A standard approach to continuous glucose monitoring and a less restrictive carbohydrate approach in gestational diabetes. Diabetes Res Clin Pract. 2018.
About the author
Kari Cao, PA-C, MHS, NFC
Kari Cao, PA-C, MHS, NFC, is a medical provider with 15 years of experience serving vulnerable populations in family medicine and obstetrics, and the founder and CEO of RenewRx Health, a public benefit corporation on a mission to guide more than 100 million lives to better health, starting with mom and never stopping there. After years of seeing how little of a patient’s care fits into a 15-minute visit, she trained as a chef with a nutritional focus and cofounded RenewRx in 2021. RenewRx combines technology and coaching to support patients between visits, with a focus on reaching families in maternal health deserts, and has supported more than 1,000 moms to date. RenewRx was named FemTech Product of the Year at MedTech World 2025 and has been selected for MedTech Innovator (top 4%, 2025) and Creative Destruction Lab. Kari has been featured in Bioscience Colorado magazine, was chosen for the Colorado BioScience Association’s mastermind group, and is a national AAPA keynote speaker on entrepreneurship as a PA.
About the editor
Teri L. Hernandez, PhD, RN, FAAN
Dr. Teri Hernandez is the Associate Dean for Research and Scholarship in the College of Nursing and a Professor of Nursing and Medicine at the University of Colorado Anschutz Medical Campus. Her program of research is focused on nutrition during pregnancy, glucose and lipid metabolism, infant growth, and gestational diabetes. She teaches biobehavioral science in the CU Nursing PhD program, and provides education nationally on pregnancy, nutrition, and lifestyle modification for cardiovascular disease and diabetes. At CU-Anschutz, she has cultivated Academic-Clinical nurse scientist partnerships between CU and University of Colorado Health and Children’s Hospital Colorado, leads the campus Perinatal Research Advisory and Facilitation Committee, and is a member of the Colorado Clinical and Translational Science Institute Executive Committee. Dr. Hernandez is a longstanding member of the Colorado Multiple Institutional Review Board, an educator in the Graduate School, mentors nurses, physicians, and PhD scientists widely, and has been a leader nationally through the National Pediatric Nurse Scientist Collaborative, the American Heart Association, the American Diabetes Association, and the Western Institute of Nursing. Dr. Hernandez is a Fellow of the American Academy of Nursing.