Wearables may help detect gestational diabetes weeks before standard screening methods begin.
Researchers are testing a new way to detect gestational diabetes long before the usual screening window, using heart data as a predictor. The approach uses changes in a pregnant woman’s heart rate while she sleeps. These changes can be measured with a small device worn at home, like a watch or wristband. When compared with current methods that rely on blood tests or medical history, this new system could catch early warning signs weeks sooner. It could also help doctors suggest healthier habits earlier in the pregnancy, when changes have more time to make a difference.
Gestational diabetes happens when the body can’t manage blood sugar properly during pregnancy. It affects around one in seven pregnancies worldwide and increases the risk of complications for both mother and baby. Typically, it’s diagnosed between 24 and 28 weeks using a blood test that measures how well the body handles sugar. But researchers say that problems in fetal growth, especially in babies of older or overweight mothers, may begin much earlier. That’s why there’s growing interest in finding ways to predict risk during the first trimester.
This new study involved nearly 2,750 women in the United States. All were in their first pregnancy and underwent sleep testing at home early in gestation. The researchers looked at heart rate variability—tiny changes in the time between heartbeats that can reflect how the nervous system is working. In people who are stressed or whose bodies are under strain, the pattern of heartbeats tends to shift. These patterns are linked to how the body handles sugar and other metabolic processes.
Heart rate variability, or HRV, may show signs of stress in the body’s automatic systems. During pregnancy, blood volume and heart rate both increase, putting new demands on the heart. HRV tends to drop when the sympathetic nervous system—the part responsible for the fight-or-flight response—becomes more active. At the same time, insulin resistance can rise, which is a key factor in gestational diabetes. The theory is that a more active stress response could be an early signal of trouble ahead.
The researchers built a computer model using machine learning to predict who would develop gestational diabetes. One model used only standard health factors like age, weight, and family history. Another used just heart rate variability data. A third model combined both types of information. That combined model turned out to be the most accurate, especially for younger or lower-weight women. It predicted gestational diabetes more reliably than older methods, including those used by major health organizations.
The results suggest that heart data from home devices could help fill the gap left by less reliable screening tools. Notably, the current guidelines often flag women as high risk based on a single factor, even if their actual risk is low. That leads to false alarms and unnecessary stress. The machine learning model, in contrast, can weigh each factor differently, leading to a more balanced prediction.
One key heart measurement—the average heart rate during sleep—stood out as the strongest clue to future risk. Higher average rates were more common in women who later developed diabetes. This supports the idea that the body’s automatic systems may shift in ways that increase vulnerability even before blood sugar levels rise.
The study does have limits. Not all women had the same test for gestational diabetes, and the heart data came from only one point in time. Also, the model worked better for some groups than others. For example, it was more accurate in younger and slimmer women. Older or heavier women, who already face higher risks, saw smaller gains in accuracy. Still, the findings are promising and suggest that noninvasive tracking tools may one day help catch problems earlier.
If validated in broader studies, this method could help bring earlier screening into routine prenatal care. While cost, access, and device training remain issues, the potential for wearable tech to detect changes early—and at home—offers a new direction for managing pregnancy-related health risks.
Sources:
Predicting diabetes risk in pregnancy via sleep HRV patterns
Join the conversation!