Two neurologists can look at the same person, the same MRI, the same halting account of a year in which the car keys kept moving and the bank statements stopped making sense, and arrive at different diagnoses. Not because one is careless, but because they are reading from different rulebooks. That is the uncomfortable fact sitting at the centre of the World Stroke Organization's first vascular dementia fact sheet, written by Yuan Cai and Vincent Chung Tong Mok of the Chinese University of Hong Kong with Hugh Markus of the University of Cambridge, published online in the International Journal of Stroke on 1 January 2026 and carried in the February print issue. It was reviewed and approved by the WSO executive, which makes it the closest thing the stroke world has to an official position on a disease it has long treated as somebody else's problem.
Start with the arithmetic, because the fact sheet does. The Global Burden of Disease study counted 56.9 million people living with dementia in 2021, a figure the authors project to exceed 137 million by 2050. To work out how much of that is vascular, they went to the only evidence that settles the question, brains examined after death. A random-effects meta-analysis of population-based autopsy studies found that pure vascular dementia accounted for 15 percent of dementia cases (seven studies, 95 percent confidence interval 9 to 24 percent), with mixed vascular and degenerative pathology adding another 16 percent (four studies, 95 percent confidence interval 7 to 31 percent). Applied to the global count, that is roughly 8.5 million people with pure vascular dementia and 9.1 million with mixed dementia today, and, if the proportions hold, 42.7 million with vascular dementia of one kind or another by 2050. Those last figures are modelling, not a head count: the authors used linear regression on log-transformed GBD prevalence data and then allocated 31 percent of the total to vascular causes. The confidence intervals are wide and heterogeneity between the autopsy studies was high, above 90 percent by the I-squared measure. Only one of the studies came from a low- or middle-income country, Brazil, which is precisely where the burden is growing fastest.
The more consequential section is the one on diagnosis, and it is bracing. The fact sheet reports that applying different accepted criteria to the same population can produce vascular dementia rates ranging from under 10 percent to over 90 percent. The ADDTC criteria pick up more cases than the NINDS-AIREN criteria; agreement between the various systems is poor; and the authors conclude that different criteria systematically select different groups of people and cannot be treated as interchangeable. For a family, that is not an academic footnote. It means the word written on a clinic letter partly reflects which framework the clinic uses, and that a second opinion which lands somewhere else is not necessarily anyone's error. An attempt at a fix already exists: the revised VasCog-2-WSO criteria, built through a multi-round Delphi consensus and published in JAMA Neurology on 16 September 2025 by Perminder Sachdev, Adam Bentvelzen and colleagues, are intended as an international standard. Whether clinics adopt them is another matter.
Underneath the labels sits the mechanism, and here the fact sheet is careful in a way that is worth copying. Pure vascular dementia, as defined in the Rush Religious Orders Study and Memory and Aging Project, means dementia in a brain showing cerebrovascular damage alone: large infarcts, microinfarcts, atherosclerosis, arteriolosclerosis, without significant neurodegenerative pathology alongside. Small vessel disease is the quiet engine of much of this, damaging the deep arterioles that supply white matter long before anything looks like a stroke. But the authors are explicit that vascular pathology very often coexists with Alzheimer's and other neurodegenerative changes, and that its presence increases the risk that those other pathologies express themselves as clinical dementia. Read that sentence slowly. It does not say blocked vessels cause Alzheimer's disease. It says a brain already carrying amyloid and tau may cope until the blood supply falters, and then it does not cope. That is why the tidy either-or of the diagnostic clinic keeps failing at the autopsy table.
There is one genuinely hopeful thread, and it comes from an odd direction. Two long-running autopsy cohorts, ROS-MAP in the United States and the Hisayama study in Japan, both show the prevalence of vascular pathology and autopsy-confirmed vascular dementia falling over time, while Alzheimer's pathology stays flat or rises. Both are in high-income countries where stroke incidence has also declined, and the authors are careful to say the trend may not hold in low- and middle-income countries. Still, it suggests that decades of blood pressure control, smoking reduction and stroke care have left a mark on the brain, not just on the heart. The 2024 Lancet Commission, cited in the fact sheet, identified 14 modifiable risk factors that together account for around 45 percent of dementia risk, hypertension and diabetes and smoking and physical inactivity among them. In a companion review by the same three authors in the same journal in 2024, the SPRINT-MIND trial is described as showing that lowering systolic blood pressure to 120 mmHg in people with hypertension and no stroke reduced mild cognitive impairment and the combined endpoint of MCI plus dementia, and slowed the growth of white matter lesions on MRI. That is trial evidence about groups, not a target for any individual reader, and nothing here should change what anyone takes or stops taking without their own clinician in the room.
What the fact sheet does, in the end, is name the gap. Vascular dementia is the second most common cause of dementia and among the least researched. It has no approved disease-modifying drug, no single agreed definition, and a global burden estimated largely by extrapolating from a handful of autopsy series in wealthy countries. Putting the WSO's name on that admission is the point of the exercise.