The number of people with dementia will rise by more than half to 1.2 million people in England and Wales within 24 years, a new study found.
However the numbers will not be as great as once feared because new cases are declining year by year thanks to to advances in medical care and people adopting healthy lifestyles.
But because people are living longer and fewer are dying of other causes such as heart disease, the numbers living with the condition will still increase.
A new mathematical model predicts numbers will increase by 57 per cent from 792,000 in 2016 to to 1.2 million by 2040.
Alzheimer's Society UK had predicted if age specific prevalence of dementia remained
constant, there will be over 1.7 million people with dementia in the UK by 2050.
But the new model warned if efforts to tackle new cases fail and the risk of developing dementia does not continue to decline, it is estimated it could hit 1.9 million by 2040 increasing the financial cost on health and care systems
It currently costs the UK economy an estimated £23bn annually to deal with dementia.
Yet accurate projections for the burden of dementia is a key step for planning to meet future need and existing forecasts may not be precise as they assume that rates of dementia remain constant.
So researchers at University College London, University of Liverpool, University of Edinburgh along with Polish, French and Finnish colleagues set out to predict the future burden of dementia with more certainty by developing the mathematical model (IMPACT - Better Ageing Model).
This model takes account of disease trends and death rates alongside the effects of increasing life expectancy.
Their results showed although the incidence or number of newly diagnosed cases of dementia is falling, the overall prevalence or number of people living with the condition is set to increase substantially.
The study published in the journal BMJ used data from 18,000 men and women from the English Longitudinal Study of Ageing (ELSA).
This began tracking the health of a representative sample of the population in England aged 50 and older from 2002.
Participants were selected randomly in six waves from 2002 to 2013.
At each wave, tests were carried out to assess memory, verbal fluency and numeracy function, and basic activities of daily living for example getting in or out of bed, dressing and eating.
Dementia was identified by these assessments, complemented by interviews with carers or by diagnosis by a doctor.
The decline in dementia incidence each year in 2002-13 was steeper than that
observed in previous studies.
After accounting for the effect of dropout from the study, the team found the rate of dementia incidence goes down by 2.7% per year between 2002 and 2013.
Despite the decrease in incidence and age specific prevalence, the number of people with dementia is projected to increase to 872,000, 1,092,000, and 1,205,000 in 2020, 2030, and 2040, respectively.
Lead author Research Associate Dr Sara Ahmadi-Abhari from University College London, said: "The risk of developing dementia at any given age is going down over time, shifting dementia to later years in life.
"This decline is mainly because of improvements in healthcare and adopting healthier lifestyles.
"Our estimate of 1.2 million people with dementia by 2040 is based on the assumption that the decline in risk of developing dementia continues to the future.
"If public health efforts fail and the risk of developing dementia does not continue to decline, the growth in numbers of people living with dementia will be much larger, reaching 1.9 million by 2040."
Senior investigator Professor Eric Brunner also at UCL added: "Our novel prediction model integrates recent downward trends in dementia and cardiovascular disease incidence with declining mortality rates in England and Wales.
"If these trends continue then the number of people with dementia will more than likely increase, from 792,000 in 2017 to more than 1.2 million in 2040.
"The projected increase in the burden of dementia, despite the substantial downward trend in age specific incidence, results largely from improvements in life expectancy.
"The results have important policy implications in terms of care needs and public spending.
"The findings of our prediction model act as a benchmark to measure the impact of possible dementia prevention initiatives."
In a linked editorial, Professor Emiliano Albanese at the University of Geneva said: "Because the most prominent risk factor for dementia is ageing, the urgent need to address the public health challenges of dementia is heightened by the "greying" of societies worldwide."
He said the results "should be interpreted cautiously " as dementia diagnosis requires demonstration of cognitive decline and other factors such as depression and delirium may play a role and not be accounted for in the study.
But he concluded: "However, the results from the ELSA study do confirm that the
absolute number of people with dementia will increase substantially in the coming years in England and Wales, as a result of an ageing population, and it is these numbers that matter most to policy makers planning future care and services."