News · Brain & Mental Health
Spouses of people with dementia had 1.7 times the risk
A JAMA Network Open study of 955,105 married adults in Taiwan found that people whose partner developed dementia went on to develop it more often themselves, though the extra risk over five years was about three percentage points.
Based on a peer-reviewed nationwide cohort study in JAMA Network Open
- 955,105 married adults in Taiwan, with each exposed spouse matched to four unexposed controls.
- Women whose partner had dementia had 1.74 times the risk (95% CI, 1.70-1.78); men, 1.69.
- In absolute terms that is 2.75 percentage points over five years for women, 3.49 for men.
- The gap narrowed among people with higher income or more children.
- Claims data with no education or lifestyle measures, so shared circumstances cannot be ruled out.
A dementia diagnosis arrives for one person and settles on two. The healthier partner becomes the one who remembers the appointments, manages the money and sleeps badly. Writing in the journal JAMA Network Open, researchers working with Taiwan’s national health records have now measured what happens to that second person over the years that follow, and the answer is that their own risk goes up.
Spouses of individuals with dementia had a higher risk of all-cause dementia than unexposed spouses. For women the risk ran about 1.7 times higher, and for men slightly less.
The link itself was not a surprise. Smaller studies had pointed at it for years. What was missing, as the authors put it, is that evidence regarding subgroup variation and absolute risk remains limited. Knowing a risk is elevated tells you almost nothing useful until you know by how much, for whom, and starting from what baseline.
How much a partner’s diagnosis moved the numbers
The study drew on Taiwan’s National Health Insurance Research Database from 1998 to 2022, which covers essentially the whole population and records diagnoses as they are billed.
Spouses of people newly diagnosed with dementia were matched 1 to 4 to unexposed controls by birth year, sex, income, and urbanicity. The matched cohort included 955,105 individuals.
That matching matters. Comparing a 70-year-old widow in Taipei against the general population would mostly measure age and address. Comparing her against four women born the same year, earning similar money in similar places, isolates something closer to the thing being studied.
The relative increase was consistent across sexes. The absolute increase was small: a 5-year adjusted risk difference of 2.75 percentage points for women, and 3.49 percentage points for men.
Those two framings describe the same result and land completely differently. Around three extra cases per hundred people over five years is a real signal at national scale and a modest one for any individual couple.
Age split the two apart. Relative risks were more prevalent at younger ages, whereas absolute risk differences increased with age. A younger spouse sees a bigger proportional jump from a very small starting risk. An older spouse sees a smaller proportional jump from a much larger one, which produces more actual cases.
Why a partner’s dementia might raise your own risk
Nothing is passing between the two people. The authors set out three overlapping explanations, and the first begins long before any diagnosis.
Assortative mating may contribute to spousal concordance in dementia risk. People tend to partner with people like themselves, including in education, income and health behavior, all of which track with dementia risk. Two people who were always going to have similar risk profiles will tend to develop similar diseases.
Then comes the marriage itself. Shared environmental and lifestyle exposures accumulated over the life course means decades of the same meals, the same neighborhood air, the same sedentary evenings.
Only third does caregiving appear, through caregiver distress, depression, sleep disturbance, social isolation and related pathways. It is the explanation that feels most intuitive and the one the study can say least about, because a billing record captures a consultation and nothing of the exhaustion behind it.
The gap that narrowed with income and children
The most suggestive finding is where the effect weakened. The association was attenuated among individuals with higher income or more children.
Read carefully, that looks like a clue about who absorbs the work. A household with more money and more adult children has more capacity to soak up the demands of care, so less of it lands on one aging person.
The authors decline to overclaim it. Number of children and income should be interpreted as proxy measures of family structure and socioeconomic status. Neither one records how much help arrived.
What claims data cannot show about couples
The study’s scale is also its constraint. Dementia outcomes were identified from administrative claims and may be misclassified, because a billing code is a record of a consultation rather than a diagnosis confirmed by biomarkers.
Two absences matter more. Educational attainment was unavailable, which is awkward when education is one of the strongest known influences on cognitive reserve. And lifestyle and social factors, including physical activity, diet, smoking intensity, and social engagement, were also unavailable. Every one of those could sit behind the association without any spousal effect existing at all.
Marriage records were imperfect too. Couples who divorced during the study still counted as exposed, which would blur the comparison rather than manufacture it.
The authors also flag reach: findings may be most generalizable to populations with similar health care systems and cultural contexts.
What it changes
Age remains the dominant factor here. Aging is the biggest risk factor for dementia, and no household arrangement rivals it. What this study adds is that dementia is not a normal part of aging and not, apparently, an entirely individual event either.
None of which makes a partner’s diagnosis a prognosis. What it does is put the well spouse back in frame as a patient in their own right, with risks worth asking a doctor about.
People also ask
Is dementia contagious between partners?
No. Nothing passes between spouses. The researchers offer three overlapping explanations, none of which involve transmission. Couples tend to resemble each other in education, income and health habits before they ever meet, a pattern called assortative mating. They then share decades of the same food, air, housing and routine. And caring for a partner with dementia brings distress, broken sleep and isolation, each of which is separately linked to cognitive decline.
How big is the extra risk in real terms?
Smaller than the headline ratio suggests. Over five years the difference was 2.75 percentage points for women and 3.49 for men. So out of 100 people whose partner had dementia, roughly three more developed it than would otherwise be expected. The relative figure of about 1.7 times sounds alarming precisely because the underlying risk is low to begin with.
Why did the risk look different for richer families?
The absolute gap shrank as income and the number of children rose. The authors are careful about what that means. Income and family size were available in the data as proxies for family structure and socioeconomic position, not as direct measures of how much practical help anyone actually received. The pattern is suggestive rather than demonstrated.
Does this mean carers should be screened for dementia?
The study does not say that, and it is not medical advice. What it argues is narrower: that household context is worth considering when assessing risk. Anyone worried about their own memory or a partner's should raise it with a doctor rather than act on a population statistic.
Would this hold outside Taiwan?
Unknown. The authors note their findings may be most generalizable to populations with similar health care systems and cultural contexts. Taiwan has near-universal single-payer coverage and different norms around multigenerational households, both of which shape who provides care and how much strain falls on a spouse.