Sleep that swings night to night, and your heart
How long people slept on average, in this study (Tran, 2026). That result doesn't make short sleep safe. In a separate pooled study of 322,721 East Asian adults, the risk of dying was lowest at 7 hours of sleep. It rose with both shorter and longer sleep (Svensson, 2021). In the Fitbit study, the estimate for average sleep length was a 4% lower risk of heart disease or stroke. The likely range runs from 15% lower to 8% higher (Tran, 2026). Sleep length that swung from night to night was linked to more heart disease or stroke. For every extra 29 minutes that sleep length swung from night to night, the risk of heart disease or stroke was 24% higher. The likely range is 9% to 41% higher. This was in 10,362 US adults who wore a Fitbit for about 6 months, followed typically for 3.7 years (Tran, 2026). The Fitbit study can't show that steadier sleep would lower your risk of heart disease or stroke. Every number links to its source.
The short version
- Sleep length, bedtime and wake-up time that swung more from night to night each came with a higher risk of heart disease or stroke, 24%, 23% and 20% higher, with overlapping likely ranges (Tran, 2026). The abstract gives no direct comparison showing which link is strongest.
- Each 29 minutes more that sleep length varied from night to night came with a 24% higher risk of heart disease or stroke (likely range 9% to 41% higher) (Tran, 2026). In groups of the same size, that works out to about 5 cases for every 4 among people with steadier sleep length, or about 10 for every 8, likely 9 to 11.
- How long people slept on average wasn't clearly linked to heart disease or stroke in this study: the estimate was 4% lower risk, with a likely range from 15% lower to 8% higher, which could be chance, and the abstract doesn't say for how much longer sleep that estimate is (Tran, 2026). That doesn't make short sleep safe. In a separate pooled study of 322,721 East Asian adults, the risk of dying was lowest at 7 hours of sleep and rose with both shorter and longer sleep, most at 10 hours or more (Svensson, 2021); the sleep length guide has that study.
- Sleep length that swung more from night to night was also linked to heart failure, high blood pressure and type 2 diabetes (Tran, 2026).
- The study watched what people already did, so it can't show night-to-night swings in sleep length caused the cases. 230 people got heart disease or had a stroke (Tran, 2026).
Which parts of your sleep pattern were linked to heart disease or stroke
Each row compares people whose sleep varied more from night to night with people whose sleep varied less, for heart disease or stroke, after allowing for demographic, socioeconomic, lifestyle and health factors. For sleep length, one step is 29 minutes. The paper's abstract doesn't give the step in minutes for the other rows (Tran, 2026).
| What varied from night to night | Risk of heart disease or stroke, per step more variation |
|---|---|
| How long you sleep | 24% higher (likely range 9% to 41% higher) (Tran, 2026) |
| Bedtime | 23% higher (likely range 10% to 37% higher) (Tran, 2026) |
| Wake-up time | 20% higher (likely range 9% to 32% higher) (Tran, 2026) |
| Midpoint of sleep | 21% higher (likely range 9% to 35% higher) (Tran, 2026) |
The percentages compare risks between groups. They aren't percentage points, and they aren't what any one person would gain or lose by changing their sleep. The abstract reports risk rising with each step more variation and gives no cut-off for how steady is steady enough.
Did the hours you sleep matter at all?
Not clearly, in this study. Average sleep length itself, as distinct from how much it varied, wasn't clearly linked to heart disease or stroke: the estimate was 4% lower risk, with a likely range from 15% lower to 8% higher, which could be chance. The abstract doesn't say how big a difference in average length that estimate is for (Tran, 2026). That doesn't make short sleep safe. In a separate pooled study of 322,721 East Asian adults, the risk of dying was lowest at 7 hours of sleep and rose with both shorter and longer sleep, most at 10 hours or more (Svensson, 2021); the sleep length guide has that study.
What else night-to-night swings in sleep length were linked to
Each figure compares people whose sleep length varied 29 minutes more from night to night with people whose sleep length varied less (Tran, 2026).
- Heart failure: 32% higher risk (likely range 13% to 55% higher) (Tran, 2026).
- High blood pressure: 16% higher risk (likely range 8% to 25% higher) (Tran, 2026).
- Type 2 diabetes: 20% higher risk (likely range 7% to 35% higher) (Tran, 2026).
The paper also reports, as an exploratory finding in 7,199 people, that more sleep cycles a night, estimated from the wearable's data, came with a 28% lower risk of atrial arrhythmia, an irregular rhythm in the heart's upper chambers (likely range 11% to 42% lower), and a 42% lower risk (likely range 23% to 56% lower) once sleep length, regularity and timing were allowed for. The authors call it a new finding that needs confirming (Tran, 2026).
The study of 10,362 people
Tran and Nguyen used the US All of Us Research Program, which links participants' Fitbit data to their electronic health records from 2009 to 2023. They took 10,362 adults, median age 53 and 70% women, and worked out 11 sleep measures over a 180-day baseline with at least 90 valid nights. New diagnoses were counted only after a 180-day washout, to limit the chance that early illness was changing sleep rather than the other way round. Over a median follow-up of 3.7 years, 230 people had a heart attack, heart failure, cardiomyopathy (heart-muscle disease) or an ischemic stroke (a stroke from a blocked blood vessel), the study's combined outcome, about 2 in 100 (Tran, 2026).
- The models allowed for demographic, socioeconomic, lifestyle and comorbidity factors (Tran, 2026).
- Sleep length that varied more was linked to the combined outcome, to heart failure, to high blood pressure and to type 2 diabetes. Average sleep length was not linked to the combined outcome (Tran, 2026).
Where to start
- In this study, sleep with less night-to-night swing in its length and timing came with less heart disease or stroke. The study can't tell whether making your own sleep steadier would change your risk.
Get the supplement tier list
30 popular supplements, ranked S to F by what trials found. Leave your email and the tier list opens straight away. I'll only write when there's a study worth your time.
Every email will have an unsubscribe link.
How far to trust these numbers
- This was one follow-up study that watched what people already did. It can't show night-to-night swings in sleep length caused heart disease or strokes (Tran, 2026).
- 230 people got heart disease or had a stroke, over a median of 3.7 years (Tran, 2026). That is a small number of cases, so the exact size of that link is uncertain. The abstract gives no case counts for the other outcomes.
- Everyone wore a Fitbit and shared its data with a research programme, so they may differ from other people, and sleep was measured as the Fitbit recorded it over the 180-day baseline (Tran, 2026).
- The study allowed for demographic, socioeconomic, lifestyle and health factors, but other differences between steadier and less steady sleepers may remain.
- Only the paper's abstract is open to read, so this guide is built on the abstract. The authors call the sleep-cycle finding exploratory and say it needs confirming. The other cautions here are this guide's reading of the study's design.
Sources
- Tran AK, Nguyen MTV. Sleep and cardiometabolic disease: a wearable-electronic health record study in the US All of Us Research Program. Circ Popul Health Outcomes, 2026.
- Svensson T et al. Association of sleep duration with all- and major-cause mortality among adults in Japan, China, Singapore, and Korea. JAMA Netw Open, 2021.