Among 57,611 UK Biobank participants followed a median 5.1 years, people with type 2 diabetes and high systolic blood pressure variability faced a 2.80-fold higher risk of all-cause mortality (95% CI 2.12-3.69) than those with normal glucose and low variability. Dementia risk was nominally elevated (HR 2.46, 1.37-4.42). Absolute variability in every blood pressure measure tracked with mortality, while diastolic and mean arterial pressure variability tracked with stroke after false-discovery-rate correction. Adjusting for variability barely dented the risk tied to glycemic status.
The key message is that visit-to-visit blood pressure swings carry prognostic information beyond average blood pressure, and that this signal operates largely independently of glucose status rather than explaining it. That fits earlier work linking variability to stroke and cognitive decline, though this analysis adds the glycemic dimension. Practical takeaway: people with diabetes or prediabetes may warrant attention to blood pressure consistency, not only the mean reading, including medication adherence and measurement conditions. Important caveats apply. This is observational, so variability may mark frailty, arterial stiffness, or poor treatment adherence rather than cause harm. Follow-up was short, the dementia finding did not survive stringent correction, and UK Biobank skews healthier than the general population. Variability measured across clinic visits is also noisy. This is a preprint that has not been peer reviewed, so figures may shift. Overall the work is incremental and confirmatory rather than practice-changing, and no trial shows that lowering variability itself improves outcomes.