Good news on dementia prevalence—we can make a difference

government. We must develop the evidence base on child health, particularly for early intervention, including the Healthy Child Programme, and promote resilience in its widest sense. Further recommendations seek to translate the evidence presented in the Annual Report into practical actions, which include commissioning age-appropriate training for health professionals, translating the evidence on children and young people’s views of health care into a “Health Deal”(panel 2), and developing tool kits for schools to improve educational attain ment through Personal Social Health and Economic (PSHE) education. We badly need independent monitoring of pro gress against outcome indicators. 11

The fi ndings of the Cognitive Functioning and Ageing Study (CFAS) I and II are unequivocally good news. New data, reported in The Lancet, 1 suggest that the prevalence of dementia in the UK in 2011 was signifi cantly lower than would have been expected based on the estimated prevalence in 1991. For CFAS I, data were taken from three geographical areas-Cambridgeshire, Newcastle, and Nottingham-to generate an estimate of the prevalence of dementia in the UK in 1991. This was based on a randomly selected sample of 7635 people aged 65 years and older interviewed in these areas, with 1457 being diagnostically assessed for dementia. An algorithmic diagnostic approach was used so that the resulting estimate, which was standardised to the 1991 population structure, could be compared with future results. For CFAS II, investigators used an identical diagnostic method to estimate the prevalence of dementia in the UK in 2011. On the basis of the age and sex specifi c prevalence estimates from CFAS I, 664 000 individuals were estimated to have dementia in 1991. After applying the eff ects of population Good news on dementia prevalence-we can make a diff erence government. We must develop the evidence base on child health, particularly for early intervention, including the Healthy Child Programme, and promote resilience in its widest sense. Further recommendations seek to translate the evidence presented in the Annual Report into practical actions, which include commissioning age-appropriate training for health professionals, translating the evidence on children and young people's views of health care into a "Health Deal"(panel 2), and developing tool kits for schools to improve educational attain ment through Personal Social Health and Economic (PSHE) education. We badly need independent monitoring of pro gress against outcome indicators. 11 This report is not unprecedented. Much has been written about the state of child health in the UK, [12][13][14] and children are championed by front-line staff and many groups in civil society, as well as the Children's Commissioner. The experts who contributed to this report drew together the contemporary scientifi c evidence, and we combine this with new economic analyses to make a powerful case for focusing anew everyone's eff orts on children and young people. Economic realities alongside the rising health-care demands for an ageing population make improvement in this area challenging. But this report shows that prevention can pay. Our children deserve better. ageing to this estimate, the number of people with dementia was projected to be 884 000 in 2011. However, extrapolation of the results of CFAS II suggests that the number of people with dementia in 2011 was 670 000, a decrease in prevalence from 8·3% in 1991 to 6·5% in 2011 (OR for CFAS I vs CFAS II 0·7; 95% CI 0·6-0·9). Importantly, the study's greatest strengths are in its relative estimates of changes between phases rather than its absolute estimates of numbers. As with any research criteria, the diagnostic system used in this study has limitations, as acknowledged by the authors. The low response rate (56% vs 80% for the fi rst phase) for the second phase might also be a source of error. Even with the adjustments and modelling made, it remains entirely possible that there is a substantial response bias, which might have a substantive eff ect on the prevalence estimates.
However, these data do suggest that things that the population has done have decreased the age-specifi c incidence of dementia, and therefore the number of people with dementia is lower than it would have been without our making these changes. The adage asserted in England's National Dementia Strategy 2 that "what's good for your heart is good for your head" therefore seems to be supported by the new evidence presented. It is plausible that changes in health behaviour and provision, including smoking cessation and improved management of cardiovascular risk factors such as hypertension, have prevented or delayed the onset of dementia at a population level. The next questions must be: how much further can we go in pursuit of this preventive agenda? How many more cases can be prevented? What do we need to do to have the greatest eff ect? These questions need empirical investigation followed by purposeful strategy formulation and implementation.
A powerful message from these data 1 is that what we as individuals and services do matters in terms of dementia. The CFAS data point to substantial added value from exist ing healthy lifestyle messages. They suggest that life style changes-eg, in diet, exercise, and smokingmight reduce the risk of dementia and promote more general health and wellbeing. This notion should be incorporated into health promotion messaging. Inclusion of the potential benefi t of dementia prevention in communications could drive greater adoption of healthy lifestyles with resulting benefi ts for individuals and society. This is a message of empowerment, but it comes with a warning. As with all investments, in this investment in health, positive past performance does not always predict future gains. Thus, a need exists for caution in prediction of the future numbers with dementia. If positive changes in health behaviour can decrease prevalence of dementia, then negative lifestyle choices might promote, rather than prevent, dementia. The cohorts of people who have been developing dementia in the past 30 years lived through periods of austerity during which diet was often controlled and this might have protected them in the balance of risk. This was not the case for present cohorts entering the period of risk for dementia (ie, those aged >60 years). It is plausible that the present epidemic of morbid obesity, with consequent cardiovascular disorders, stroke, and diabetes, might act to increase the proportion of people with dementia in future cohorts.
What these data do not mean is that dementia should be any less of a priority. Even with the changes described by Fiona Matthews and colleagues, 1 dementia remains very common, very expensive, and profoundly negative in its eff ects on people with the disorder and their families. 2 Even with a small decrease in incidence and prevalence, population ageing will still double the numbers with dementia worldwide in the next generation. 3 Dementia remains one of the greatest challenges faced by healthcare and social-care systems worldwide, in low-income and middle-income countries as well as in more developed economies. 4 Dementia is one of the very few health disorders that in itself has a macroeconomic eff ect, 5 driven by the contribution of dementia to longterm care costs, with at least three quarters of people in care homes now having dementia, as shown by the CFAS estimates, 1 and by people with dementia being over-represented in general hospital and emergency populations. The existing management of dementia at a population and an individual level is improving with the execution of national plans and improved individual care. 6 But there remain public and professional misconceptions, a low level of diagnosis with more than half of those with dementia never identifi ed as such, and discontinuities and poor quality care from diagnosis to end of life for many.
Dementia is a powerful example of the complexity and long-term nature of the disorders that are now the major outstanding challenges for health-care systems. Those with dementia are generally an old and frail population with multimorbidity; data from the Scottish School of Primary Care 7 suggest that only 17% of people with dementia have no other long-term disorder. If we In The Lancet, Anthony Barnett and colleagues 1 provide evidence that linagliptin, a dipeptidyl peptidase-4 (DPP4) inhibitor, eff ectively lowered glycated haemoglobin (HbA 1c ) in a cohort of 241 older patients (mean age 75 years) with type 2 diabetes poorly controlled with usual treatments. This randomised, placebocon trolled trial was done in a 24-week period; mean HbA 1c was 7·8% (SD 0·8) at baseline, and at week 24, placebo-adjusted mean change in HbA 1c with linagliptin was -0·64% (95% CI -0·81 to -0·48, p<0·0001). The investigators acknowledge the scarcity of specifi c studies of glucose-lowering treatments in older patients with diabetes and of the DPP4-inhibitor class in particular. They pro vide some potential reasons why this class can off er some advantages in treatment of older patients, such as a reduced risk of hypoglycaemia and no appreciable weight gain. Additionally, they argue that it is important to consider the special issues of medical comorbidities and frailty in older patients in trials of this kind. The Lancet, perhaps in recognition of this shortfall in the medical literature, also published a similar study of the DPP4 inhibitor vildagliptin in a group of older patients with type 2 diabetes. 2 The study by Barnett and colleagues 1 was well designed and powered to show a clinically meaningful, signifi cant diff erence in HbA 1c in the groups studied, without an increased risk of hypoglycaemia. In general, the patients studied had levels of duration of diabetes and renal impairment commonly noted in older patients with this disorder. The investigators accept the limitations of this type of pharma-directed, short duration study. Longer-term monitoring data are needed to ensure safety of patients.
This study, however, missed crucial opportunities to provide increased insight into the management of

Frailty and diabetes
can get services right for dementia, then we will be a long way towards getting them right for all individuals with complex and long-term disorders. The CFAS results suggest that prevention is possible and that we can have agency in this most complex of disorders. These fi ndings should spur us on, to go further and faster in secondary and tertiary prevention as well as primary prevention in dementia, for the benefi t of all. This study shows that we can all make a diff erence.

Sube Banerjee
Brighton and Sus sex Medical School, University of Sussex, Brighton BN1 9RY, UK s.banerjee@bsms.ac.uk I declare that I have no confl icts of interest.