Although the prevalence of dementia worldwide has been projected to increase due to population aging, a growing number of studies in recent years are reporting a decreasing trend in the dementia prevalence (Table 1). In 2005, Manton et al. (7), analyzing data from the National Long-Term Care Survey (NLTCS), reported a decline in the prevalence of dementia in the US elderly aged 65 yr and older. The prevalence of severe cognitive impairment (SCI), defined using a cognitive screening test, decreased from 5.7% in 1982 to 2.9% in 1999, amounting to 310,000 fewer cases. The decline in prevalence was more pronounced in men than women. Using diagnosis data from the medical records, a significant decline in mixed dementias but not in Alzheimer's dementia was also observed.

The Zaragoza Study (8) examined the prevalence of dementia in a representative sample of both institutionalized and non-institutionalized older people in Spain. Dementia cases were identified using a standardized clinical interview schedule according to Diagnostic and Statistical Manual of Mental Disorders (DSM) IV criteria. Among urban residents aged 65 yr and older, the dementia prevalence was found to be quite stable between two periods, 1988-1989 (5.2%; 95% confidence interval [CI], 3.9-6.6) and 1994-1996 (3.9%; 95% CI, 3.3-4.5), with a prevalence ratio (PR) of 0.75 (95% CI, 0.56-1.02). The prevalence, however, significantly decreased among men (PR, 0.40; 95% CI, 0.25-0.65), particularly those in the 70 to 84 yr age group, but not in women (PR, 1.02; 95% CI, 0.69-1.51).

Langa et al. (9) and Rocca et al. (10) used the Health and Retirement Survey (HRS), a national survey in the US, to examine temporal trends in the prevalence of cognitive impairment between 1993 and 2002. A 35-point cognitive function scale was used to define cognitive impairment consistent with dementia (CI-D). CI-D showed a significant decrease from 12.2% to 8.7% over the decade. The unadjusted odds ratio (OR) in the prevalence between 2002 and 1993 was 0.68 (95% CI, 0.60-0.77). Even adjusting for demographic and lifestyle factors and chronic conditions, the trend toward decreasing prevalence remained significant (OR, 0.72; 95% CI, 0.62-0.83). A significant decline in mortality from cognitive impairment during the same decade suggested compression of cognitive morbidity. In contrast, analysis using population-based data from the US states of Minnesota (1975-1994), Illinois (1997-2008), and Indiana (1992 vs. 2001) did not support a decrease in the prevalence of dementia (10).

The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) reported a decreased prevalence of dementia during a 20-yr period among residents of Cambridgeshire, Newcastle, and Nottingham in the UK (11). CFAS I (1989-1994) and CFAS II (2008-2011) employed an identical sampling design and methods. Based on the estimated 664,000 with dementia in 1991, 884,000 were expected to have dementia in 2011 due to population aging. However, the actual observed number in 2011 was 670,000, demonstrating a decrease of 24%. This translated to a decrease of 1.8 percentage points, from 8.3% to 6.5%, in the predicted prevalence (OR, 0.7; 95% CI, 0.6-0.9). The prevalence of dementia in residential and nursing homes increased from 56.4% to 69.6% during the same period, but with a coinciding reduction in the proportion of older people in care settings from 5% to 3%.

In two Danish cohorts of nonagenarians, cognitive functioning was found to be significantly better in the 1915 birth cohort in 1998 than in the 1905 birth cohort in 2010 (12). The proportion with cognitive impairment, measured on the Mini-Mental State Examination (MMSE) scale, was estimated to be 47% among the 1905 birth cohort and 40% in the 1915 birth cohort. The proportion of respondents exhibiting a maximum score range of 28-30, the highest level of cognitive functioning, on the MMSE was higher in the 1915 birth cohort (23%) than in the 1905 birth cohort (13%). Moreover, cognitive performance, assessed by a battery of tests involving fluency tasks, forward and backward digit spans, and immediate and delayed recall, was significantly better in the later cohort (P<0.001).

Two additional studies have reported a decrease in the incidence of dementia. In the Rotterdam Study (13), a decline in dementia incidence was observed during a 20-yr period. Among the 1990 cohort of 25,696 person-years of follow-up, 286 people developed dementia, whereas in the 2000 cohort of 8,384 person-years of follow-up, 49 developed dementia, with an incidence rate ratio (IRR) of 0.75 (95% CI, 0.56-1.02). Although the IRR was not statistically significant, the reduction in risk was consistent across age- and gender-specific strata. In addition, a significant decline in mortality rates (rate ratio, 0.63; 95% CI, 0.52-0.77) indicated a decrease in dementia incidence, considering the competing risk of vascular risk factors for both dementia and mortality. The diagnostic criteria used in the two cohorts were the same. The authors further suggested that an increased awareness of dementia by physicians, with an increased case identification over the years, would have resulted in an underestimation of the difference in the incidence rates between the two cohorts. Moreover, brain imaging showed a higher total brain volume and fewer white matter lesions over time, indicating less brain atrophy and cerebral small vessel disease.

In Sweden, trends in the prevalence of dementia were assessed over 20 yr in two population-based surveys in the Kungsholmen area of central Stockholm (14). The DSM-III-R was used to diagnose dementia in both surveys. Among the study participants 75 yr and older, the dementia prevalence was stable at 17.5% during 1987-1989 and 17.9% in 2001-2004, controlling for age, sex, and education (OR, 1.17; 95% CI, 0.95-1.46). Mortality, however, decreased during this period (hazard ratio, 0.71; 95% CI, 0.57-0.88). Although the prevalence of dementia was stable between the late 1980s and the early 2000s, the increased survival during the interval suggested a decrease in the incidence of dementia.

Notwithstanding potential sources of bias in these studies, such as sample selection, attrition, and methods of case identification and diagnosis, and the need for further verification of the findings, there appears to be accumulating evidence of a declining prevalence or incidence of dementia in the US and several European countries. The cause for the decline in dementia prevalence has been attributed to the reduction and control of risk factors. Examples include an increased level of education in older people and intellectual stimulation (7, 9, 10, 11, 12, 13), improved general living, work environment, and social welfare (12, 14), enhanced healthy behaviors such as avoidance of smoking, increased participation in physical and social activities, and nutrition (8, 10, 14), a low burden of infectious diseases (12), control of vascular risk factors such as diabetes mellitus and cardiovascular diseases (8, 10, 11, 13, 14), a decline in stroke rates and consequent decline in post-stroke dementias (7, 13), and an increased use of prophylactic medications such as statins, nootropics, estrogen, and anti-inflammatory and anti-oxidative agents (7, 9, 13).

Because no effective treatment for dementia is currently available, prevention is key to control its projected rise in prevalence worldwide. Preventive measures that delay the onset of dementia would be effective in reducing the number of future dementia cases. Alternatively, therapeutic interventions delaying the progression of the disease would reduce the proportion of late-stage or severe dementia.

In the US, the number with Alzheimer's disease was estimated to be 2.32 million in 1997, and estimated to increase to 8.64 million in the next 50 yr (15). Delaying the onset of dementia by 5 yr is expected to reduce the number of new cases by 4 million. A two-year delay would reduce the number by 1.94 million, and delaying the onset even by one year would reduce it by 770,000 incident cases.

The potential impact of dementia prevention is even greater on a global scale. In 2006, the prevalence of Alzheimer's disease was estimated to be 26.6 million worldwide and predicted to grow to 106.8 million by 2050 (16). If the onset of dementia could be delayed by 2 yr, there would be 22.8 million fewer cases. Even a year's delay would decrease the prevalence by 11.8 million. In contrast, delaying the progression of dementia by one year would actually increase the prevalence by 2.8 million. However, a combination of a year's delay in both the onset and progression would result in a decrease of 9.2 million cases, with 8.7 million fewer late-stage dementia cases.

Because cognitive decline and dementia are attributed to many modifiable lifestyle and health-related factors, the changing trend in the major risk factors of dementia may foretell the future prevalence of dementia. In Korea, however, the reported trends in the risk factors do not support an optimistic forecast. According to the Korea National Health and Nutrition Examination Survey, in adults aged 65 yr and older, the prevalence of hypertension in 2005 was 52.4% in men and 59.0% in women, an increase from 47.0% for men and 57.0% for women in 1998 (17). Type 2 diabetes also demonstrated an increased prevalence in both men (13.2% to 16.2%) and women (11.4% to 16.0%) during this period. Those overweight or obese showed an increase from 11.2% to 28.1% in men and 33.3% to 36.6% in women. The percentage of current smokers decreased from 51.8% to 36.0% in men, but increased from 14.9% to 18.1% in women. The education level among older people has improved over the years, with the proportion having received no formal education decreasing from 52.9% in 1998 to 31.6% in 2011 (18, 19). However, many older people still suffer from low income and poor living conditions, with 45.1% living in poverty, defined as the proportion of the elderly living in households below 50% of median disposable income (20).

Among adults aged 19 yr and older, the smoking prevalence in men steadily decreased from 67% in 1998 to 45.1% in 2007, but has not changed much since, and was estimated to be 46% in 2011 (21). In women, smoking rates were stable at 6.6% in 1998 and 6.5% in 2011. The prevalence of overweight or obesity between 1998 and 2011 has increased from 29% to 34% among those aged 30 yr or older. Participation in moderate intensity physical activity has declined during 2005-2011, from 71.1% to 50.6% in men and from 65.7% to 42.6% in women. Among the vascular risk factors, the prevalence of hypertension has not changed much in men (31.1% in 1998 to 33.9% in 2011) and women (27.0% in 1998 to 27.8% in 2011). However, the age-adjusted prevalence has shown a 0.2%-0.3% annual decrease. There was also an increase in treatment rates from 16% in 1990 to 61% in 2011, and in control rates from 1% in 1990 to 43% in 2011. The prevalence of hypercholesterolemia, in comparison, has increased from 9.9% in 1998 to 14.5% in 2011. The treatment and control rates for hypercholesterolemia were low at 33% and 26%, respectively, in the two periods.

In light of this trend in the risk factors for dementia, it is unlikely that a decline in the prevalence of dementia in Korea will be evident in the near future. In Japan, several regional studies between 1992 and 2008 have shown an increase in dementia prevalence (22). This has been attributed to an increase in the prevalence of type 2 diabetes and the metabolic syndrome, and a plateau since the late 1990s in the previously declining stroke incidence.

Despite the dire outlook for an emerging epidemic of dementia, prevention efforts may help to mitigate the projected trend in dementia prevalence. It has been estimated that about half of Alzheimer's disease cases worldwide (17.2 million) are attributable to potentially modifiable risk factors, such as diabetes mellitus, midlife hypertension, midlife obesity, depression, physical inactivity, smoking, and low education (population attributable risk, 50.7%) (23). Reduction in the prevalence of these seven risk factors by 10%-25% would result in 1.1 to 3 million fewer cases worldwide. By applying effective preventive interventions to reduce the risk of dementia, it may be possible to stem the rising tide of the dementia outbreak.

Primary prevention needs to be at the core of the strategy to control dementia. Increasing the level of awareness about dementia and recognizing the importance of prevention by both the public and professionals would be a first step in the right direction. Evidence of the potential benefits of a healthy lifestyle is accumulating, with prospective cohort studies indicating that physical activity, social activity, cognitive activity, avoiding smoking, moderate alcohol intake, maintenance of normal body weight, and healthy dietary patterns and nutrition reduce the risk of cognitive decline and dementia (24, 25). Recent interventional research further supports the beneficial effects of physical activity on cognitive functioning (26). More recently, multicomponent lifestyle interventions, including a combination of exercise, nutrition, cognitive training, and management of vascular risk factors, are being conducted to examine the effects of multiple health behavior change on cognitive impairment and dementia in older adults (27). Also, considering that lifestyle risks are prevalent throughout the lifespan, it is important to adopt prevention strategies that target specific stages of the life course. Improving education and living conditions early in life, reducing midlife vascular risk factors such as obesity and hypertension, and encouraging participation in physical, social, and mental activities in late life may help to maintain cognitive health through different phases of life.

The government and the healthcare system must also be prepared to address this problem. At the national level, policies and programs that support dementia prevention efforts are urgently needed. To this end, the Dementia Management Act, enacted in 2011, is anticipated to help direct resources to dementia prevention projects. Healthcare financing, with its currently disproportionate compensation for administering therapeutic interventions, needs restructuring to incentivize providers to deliver effective health promotion and preventive services. Priority should be assigned to support prevention research and program development that informs policy and practice.