Dementia in the Middle East and North Africa is projected to rise 367% by 2050

Dementia prevalence in the Middle East and North Africa is projected to rise by 367% by 2050, according to a review that pooled estimates across the region [s1]. The projection is driven less by any single cause than by a stack of them: a population ageing quickly, and a risk profile in which several modifiable factors are more common than the global average [s1].

What the review found

The review, published in the Journal of Global Health on 17 July 2026 by authors at Weill Cornell Medicine Qatar, is a narrative synthesis rather than a new survey [s1]. It pooled prevalence estimates and examined population-attributable risk across the region [s1].

Pooled dementia prevalence was estimated at 1.84% among adults aged 50 and older, rising to 3.81% among those aged 60 and older, with substantial variation between countries [s1]. Those figures sit alongside the review's central projection — a 367% increase in cases by 2050 — which reflects the demographic transition underway across the region as populations that were young a generation ago move into older age [s1].

The risk structure

The review's more useful contribution is its account of why the trajectory is steep. It reports that several modifiable dementia risk factors are highly prevalent in the region: smoking above 30%, physical inactivity around 50%, and high burdens of diabetes, hypertension and obesity [s1]. These are the risk factors that global dementia-prevention frameworks identify as the most actionable, and their prevalence in MENA is what turns an ageing population into a projected surge [s1].

The diabetes burden is documented separately. An ecological analysis of Global Burden of Disease estimates for 21 MENA countries found that age-standardised diabetes incidence rose 92% between 1990 and 2023, from 251.7 to 482.5 per 100,000, and that prevalence more than doubled over the same period [s2]. A risk factor that is both common and rising compounds the dementia projection rather than simply adding to it.

The review also flags drivers that are more specific to the region: nutritional transitions, gut dysbiosis, consanguinity, and environmental exposures including air pollution and climate-related heat stress [s1]. It does not claim these are established causes on the level of the classic modifiable risks — they are described as additional, region-specific contributors that the evidence points to [s1].

The health-system side

The review pairs the epidemiology with a set of system constraints: limited research infrastructure, low health literacy, stigma, and inequitable access to care [s1]. These matter for the projection because they affect both prevention and detection. A region with high exposure to modifiable risks but limited capacity for early diagnosis is one where cases accumulate before they are counted, and where prevention programmes that work elsewhere have not been implemented at scale [s1].

The authors argue that strengthening prevention, early detection and equitable access should be regional health priorities [s1]. That framing is notable because it treats the projected rise as partly avoidable — the modifiable share of the risk is, by definition, the part that intervention could change.

The limits

This is a narrative review, not a meta-analysis with a formal pooling protocol, and the authors present it as such [s1]. Pooled prevalence estimates drawn from studies that used different methods and sampled different populations carry wide uncertainty, and the review notes substantial variability across countries [s1]. The 367% projection is a modelled figure that depends on demographic and incidence assumptions, and it should be read as a scenario, not a forecast [s1]. The diabetes figures come from Global Burden of Disease modelling, which itself carries uncertainty intervals [s2].

What the review establishes is not a precise count but a structure: an ageing region carrying an unusually heavy load of the risk factors most linked to dementia, with health systems not yet built to detect or prevent it at scale [s1].