Researchers studied 220 women with PCOS (a common hormone condition) who tried a ketogenic diet to see why some developed sleep problems. They found that anxiety, depression, blood sugar changes, and cholesterol levels were the main reasons sleep got worse. The scientists created a tool that can predict which patients might have trouble sleeping before it happens. This matters because sleep problems are common in PCOS patients, and understanding what causes them could help doctors prevent these issues when recommending keto diets.
The Quick Take
- What they studied: Why some women with PCOS develop sleep problems when following a ketogenic (very low-carb) diet, and whether doctors could predict who might have this problem
- Who participated: 220 women with PCOS from a hospital in Beijing, China, between 2021 and 2023. About 56 of them developed sleep disorders after starting the keto diet
- Key finding: Four main things predicted sleep problems: anxiety levels, depression symptoms, how well blood sugar was controlled, and LDL cholesterol (the ‘bad’ cholesterol). The prediction tool was about 93% accurate at spotting who would have sleep issues
- What it means for you: If you have PCOS and are considering a keto diet, talk to your doctor about your anxiety and depression levels first. Doctors might be able to identify if you’re at risk for sleep problems before they start, and take steps to prevent them
The Research Details
This was a retrospective study, meaning researchers looked back at medical records from women who had already tried the keto diet. They divided 220 women into two groups: 132 women were used to identify patterns (the modeling group), and 88 women were used to test if those patterns worked (the validation group). Within the modeling group, they compared 56 women who developed sleep problems to 76 women who slept fine.
The researchers used statistical tools to identify which factors were most important for predicting sleep problems. They then created a mathematical formula (a prediction model) that combines four key factors to estimate someone’s risk of developing sleep issues. Finally, they tested this formula on the validation group to make sure it actually worked in real life.
Using past medical records is practical and cost-effective, though it can’t prove cause-and-effect like a controlled experiment could. The two-group approach (modeling and validation) is important because it shows the prediction tool works on new patients, not just the original group. This makes the findings more trustworthy for real-world use by doctors.
Strengths: The study used a reasonable sample size (220 patients), tested the prediction tool on a separate group of patients, and used multiple statistical methods to verify accuracy. The prediction tool showed very high accuracy (93% in the original group, 84% in the test group). Limitations: This was a single-hospital study in China, so results might differ in other populations. The study looked backward at records rather than following patients forward, which limits what we can conclude about cause-and-effect. No information about how the keto diet was monitored or how strictly patients followed it.
What the Results Show
The study identified four main factors that predicted sleep problems in PCOS patients on keto diets: anxiety (the strongest predictor), depression, blood sugar control, and LDL cholesterol levels. Anxiety was the most important factor—patients with higher anxiety scores were much more likely to develop sleep problems. Depression was also significant, suggesting that mood disorders play a big role in sleep quality during keto diets.
Blood sugar control mattered too. Patients whose blood sugar wasn’t well-controlled while on the keto diet were more likely to have sleep issues. This is interesting because keto diets are supposed to help control blood sugar, but apparently this doesn’t happen equally for everyone. Finally, higher LDL cholesterol (bad cholesterol) was linked to more sleep problems, though researchers aren’t sure if this is a direct cause or just happens together.
The prediction model combining all four factors was very accurate. When tested on the original group of 132 patients, it correctly identified who would have sleep problems 93% of the time. When tested on a new group of 88 patients, it was still quite accurate at 84%, showing it works in real-world situations.
The study showed that the prediction model had good practical value for doctors. The calibration curves indicated that when the model predicted a 50% risk of sleep problems, about 50% of patients actually developed them—meaning the predictions were reliable and not over- or under-estimating risk. The decision curve analysis showed that using this model would help doctors make better treatment decisions compared to just guessing or treating everyone the same way.
Previous research has shown that PCOS patients commonly experience sleep problems, and that anxiety and depression are common in PCOS. This study is novel because it specifically looks at what happens to sleep when PCOS patients try a keto diet, and it creates a tool to predict problems before they happen. Most previous studies just described the problem; this one offers a way to prevent it. The findings about anxiety and depression being important align with existing research, but the specific connection to keto diet response appears to be new.
This study only included patients from one hospital in Beijing, so results might not apply to other countries or populations with different genetics or healthcare systems. The study looked backward at medical records rather than following patients forward, so we can’t be completely sure these factors cause sleep problems versus just appearing together. The researchers didn’t report how strictly patients followed the keto diet or how it was monitored, which could affect results. We don’t know if the findings apply to men with PCOS or to people trying keto for other reasons. The study didn’t measure other factors that might affect sleep, like exercise, caffeine use, or sleep environment.
The Bottom Line
If you have PCOS and are considering a ketogenic diet, discuss your mental health (especially anxiety and depression) with your doctor beforehand. Ask your doctor to monitor your sleep quality, blood sugar control, and cholesterol levels during the diet. If you have high anxiety or depression, your doctor might recommend treating these first or being extra careful with the keto diet. These recommendations are based on solid research (the prediction model was 84-93% accurate), but they should be personalized to your situation. This is not a reason to avoid keto if it’s right for you—just a reason to monitor these specific factors.
PCOS patients considering or currently on a keto diet should definitely pay attention to this research. Doctors treating PCOS patients should consider using this information when recommending diets. Women with PCOS who also have anxiety or depression should be especially aware. People without PCOS probably won’t find this directly relevant. This research doesn’t apply to other types of diets, only ketogenic diets.
Sleep problems typically developed within weeks to months of starting the keto diet in this study (though exact timing wasn’t specified). If you’re going to experience sleep issues from keto, they’d likely appear relatively soon after starting. Benefits from addressing the underlying factors (anxiety, depression, blood sugar control) might take 4-12 weeks to show up in improved sleep.
Want to Apply This Research?
- Track sleep quality (rate 1-10 each morning), anxiety level (rate 1-10 daily), and mood/depression symptoms (rate 1-10 daily). Also log blood sugar readings if you monitor them. Over 4 weeks, look for patterns between anxiety/mood and sleep quality.
- If the app shows your anxiety is high on days you sleep poorly, use the app to set reminders for anxiety-reducing activities (deep breathing, 5-minute meditation, short walk) at the same time each day. Log which activities help most. Share this data with your doctor at your next visit.
- Set up weekly check-ins where you review your average sleep score, anxiety score, and mood score. If sleep drops below 6/10 for more than 3 days while anxiety is above 6/10, flag this for your doctor. Track cholesterol and blood sugar results when available and correlate with sleep patterns over months.
This research describes factors associated with sleep problems in PCOS patients on keto diets but does not prove these factors directly cause sleep issues. These findings are from one hospital in China and may not apply to all populations. This information is educational and should not replace personalized medical advice from your doctor. Before starting any new diet, including ketogenic diets, consult with your healthcare provider, especially if you have PCOS, anxiety, depression, or other health conditions. Do not use the prediction model described in this study to self-diagnose or self-treat. If you experience sleep problems while on a keto diet, contact your doctor rather than making diet changes on your own.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.