The prevalence of diabetes has been estimated by applying the WHO 1999 criteria  for 13 European and 11 Asian cohorts participating in the DECODE and the DECODA studies. In Europe, the age-specific prevalence of diabetes rose with age up to 70s and 80s in both men and women  (Figure 3.3). In most of the studies, the prevalence was less than 10% in subjects younger than 60 years and between 10 and 20% at 60–79 years of age. They were higher in Malta than in other populations. The prevalence of isolated postload hyperglycemia (2-h glucose ≥11.1mmol L−1 and fasting glucose <7.0mmol L−1) increased more with age than did isolated fasting hyperglycemia (fasting glucose ≥7.0mmol L−1 and 2-h glucose <11.1mmol L−1), particularly in women.
A recent German study using WHO 1999 criteria  showed that the prevalence of diabetes in Germany in 2000 was 16.7% in men and 8.6% in women at 55–59 years of age and 23.1% in men and 17.0% in women at 70–74 years of age ; that is, rates are within the variation reported for the DECODE study: In the recent Turkish Diabetes Epidemiology Study (TURDEP-II) undertaken in 2010, the prevalence of diabetes was 16.5% and increased with age, reaching a peak of 37.7% in urban men at 70–74 years and 43.6% in urban women at 75–79 years of age . These rates are higher than those reported by an earlier Turkish study in 1997–1998 .
Compared with most of the other racial and ethnic groups worldwide, where age- and sex-specific prevalence of diabetes has been reported, Europeans have a moderate to low prevalence of diabetes .
In the United States, the prevalence of diabetes varies considerably among different ethnic groups. The prevalence was 1.9 times greater in Latino Americans and 1.6 times in African Americans than in Whites of the same age in the Third National Health and Nutrition Examination Survey (NHANES III) . In NHANES III, where a single fasting plasma glucose≥7.0mmol L−1 was applied for the diagnosis of diabetes, the prevalence of diabetes in US Whites in 1988–1994 was 5.9% in men and 4.8% in women at 40–49 years of age and reached a peak of 19.2% in men and 16.6% in women at 75 years or older . The prevalence of undiagnosed diabetes according to the same fasting glucose criteria at a comparable age range of 40–59 years was higher in US Whites than in most of the female and in half of the male European populations participating in the DECODE study . The rates were higher in Mexican Americans than in US Whites, and were higher than seen in any of the populations included in the DECODE study.
The Pima Indians have the highest prevalence of diabetes in the world, being 50% at 30–64 years of age, estimated using 2-h postload glucose test alone . The prevalence of diabetes in Native Hawaiians (Polynesian population of Hawaii) in two rural communities was estimated using WHO 1985 criteria . At 30–39 years of age, the prevalence was 6.5% in women and 10.7% in men, reaching a peak of 34.6% in women and 40.0% in men at 70 years or older. Other Native American tribes also have a higher prevalence of diabetes than Caucasoids do.
Central and South America
The age-standardized prevalence of diabetes using 2-h glucose criteria alone was investigated in a Brazilian population in Sao Paulo in 1987 and in a Colombian population in Bogota in 1988–1989 . The prevalences in both populations were similar, around 7% for men and 9% for women. The age-specific prevalence of diabetes in a Mexican population in Mexico City was 4.2% in men and 3.2% in women at 35–39 years of age and reached 23.1% in men and 41.7% in women at 60–64 years of age .
The Australia Diabetes, Obesity and Lifestyle Study (AusDiab) took place in 1999–2000 applying WHO 1999 criteria mainly in Caucasoids.The prevalence of diabetes increased with age, from 2.7% in men and 2.2% in women at 35–44 years of age to 23.5% and 22.7% at 75 years or older . These rates were higher than those in most of the DECODE populations .
Asia and Pacific Islands
Asia The prevalence of diabetes varies markedly among Asian populations. In those participating in the DECODA study, it rises with age up to 70s and 80s in Chinese and Japanese, and in Indian men and women with age up to 60s and then declines  (Figure 3.4a,b). The age- and sex-specific prevalence and the peak prevalence of diabetes were higher in cohorts from India and Singapore than in most of the Chinese and Japanese cohorts. In Chinese and Japanese, the prevalence was less than 10% at 30–49 years of age; the peak prevalence was less than 20% in most of the cohorts and none exceeded 30%. In contrast, in India and Singapore the prevalence was over 10% among people aged 40–49 years, and over 30% among those aged 50–69 years for most of the cohorts. The urban Chinese and Japanese had significantly higher prevalences of diabetes than their rural counterparts at 40–69 years of age in men and at 50–79 years of age in women (Figure 3.4).The prevalence of diabetes in Korea was within the range observed in China and Japan .
Type 2 diabetes was found at a relatively younger age in Pakistan and the prevalence reached the peak in the age group 55–64 years [21–23]; the prevalence pattern was similar to that in India. In a rural community the prevalence was over 13% at 35–44 years of age, with the highest prevalence of 30% in men at 65–74 years of age , indicating diabetes has already become a major health threat in Pakistan as in India.
In the late 1990s, King et al. carried out a series of surveys on diabetes in Uzbekistan and Mongolia using WHO 1985 criteria . In Uzbekistan, the prevalence of T2DM was relatively low in people younger than 45 years, around 1%, and 10–15% at 65 years or older . It was relatively rare in people younger than 45 years, around 1%, and 10–15% at 65 years or older. The prevalence in Mongolia  was relatively low with a peak of less than 5% at the age of 65 years or older, comparable to the prevalence reported from China in 1994 . Taking into account the high positive cutoff value of 7.8mmol L−1 for fasting glucose for diabetes in the WHO 1985 criteria, the prevalence reported in these studies would be somewhat higher if the current cutoff value of 7.0mmol L−1 were used. Because different diagnostic criteria have been applied, the results from some of these studies cannot be compared directly with those from the DECODA study .
Compared with the European populations included in the DECODE study , the age-specific prevalence of diabetes in urban Chinese and Japanese was slightly higher than that in Europeans at 30–69 years of age, but was lower than that in Indians. In the elderly population, however, the peak prevalence was higher in a few European populations than in Indians, such as in Maltese, in Finnish women in Oulu, and in women living in the Canary Islands, Spain.The age at which the peak prevalence of diabetes was reached was similar for Europeans, Chinese, and Japanese (over 70 years), while Indians had their highest prevalence at the age up to 60s and then started to decline. These differences in prevalence in the elderly (women, in particular) are probably due to selective mortality associated with diabetes. The recent survey in China showed a marked increase in the prevalence of T2DM; it was 9.7%, representing an estimated 92.4 million adults in China with diabetes in 2007 .
In all Asian populations included in the DECODA study, the prevalence of isolated fasting hyperglycemia (fasting plasma glucose ≥7.0mmol L−1 and 2-h plasma glucose <11.1mmol L−1) did not increase with age (Figure 3.4).The prevalence of isolated 2-h hyperglycemia (2-h plasma glucose ≥11.1mmol L−1 and fasting plasma glucose <7.0mmol L−1) tended to increase with age in Chinese and Japanese but not in Asian Indians .
There are remarkable differences in the prevalence of diabetes among the Melanesia, Micronesian, and Polynesian populations of the Pacific Islands. According to the 2-h postload glucose criteria alone , an age- standardized diabetes prevalence of more than 40% was revealed in the Micronesia population of Nauru in the 1980s [28,29]. The prevalence of diabetes in the Melanesia population of Papua New Guinea had been reported close to 0% in highland populations , whereas in the urbanized Koki people the age-standardized rate exceeds 40%, approaching that of Nauru , exhibiting an extreme urban–rural gradient. Intermediate rates are seen in other Pacific Island populations. In the Polynesia population of the Western Samoa, the crude prevalence rates were 3.4 and 8.7% in rural and urban populations, respectively. By 1991, these rates had risen to 6.5 and 9.0% in two rural communities and to 16% in the urban settings of Apia . A recent study in the Polynesia population of Tonga in 1998–2000 using 2-h OGTT showed that the peak prevalence of diabetes reached 20% in men and 40% in women aged 60 years or older .
The prevalences of diabetes in Arabian countries, calculated according to the WHO 1985 criteria , have been reported to be high [33–35]. It is relatively low before the age of 30 years and starts to increase during the 40s, with the highest in the oldest age group (Figure 3.5). In a rural Palestinian village, in 1996 the prevalence of diabetes was less than 4.0% in people younger than 40 years but increased to 11.0% at 40–49 years of age, with a peak of 21.7% in men and 31.6% in women at 60–65 years of age . The prevalence in Palestine was similar to that in rural Saudi Arabia, but both were lower than that in urban Saudi Arabia . The prevalence of T2DM in 1995–1996 in Kuwait was recalculated using ADA 1997 criteria . It was lower than 3% at age 20–29 years, around 9% at age 30–39 years and higher than 15% at age 40–49 . Diabetes is prevalent in all Arab countries despite the differences in economic status among these countries, indicating that genetic susceptibility and cultural factors may play an important role in the development of the disease.
The age-specific rate ranged from 8% at 40–44 years of age to 25% at 60–64 years of age in Israeli Jews .
In subjects aged 30–64 years, the age-standardized prevalence of diabetes using 2-h glucose alone has been reported to be higher in Hindu and Muslim Indians living in Mauritius  and Tanzania , around 10% in Tanzania and 13–18% in Mauritius. The age-standardized rate was very low in Bantu in Tanzania, less than 1% in women and 0.9–3.3% in men . The prevalence was 8% in Tunisia . It is interesting to note that the age-standardized prevalence was much higher in Chinese living in Mauritius than that in Da Qing in China in the mid 1980s [17,37], indicating the importance of impact of the environmental factors.
In Native People: Mapuche and Aymara in Chile and in Siberia in Russian Federation
The native people who still practice their traditional lifestyle and undertake considerable physical activity have extremely low prevalences of diabetes despite their high prevalence of obesity [40,41]. Among Aymara the prevalence of diabetes in 1997 was almost undetectable despite the fact that 13% of the men and 24% of the women had a body mass index (BMI) higher than 30 kg m−2 . Similar findings have been reported among native Mapuche [41,42]. The indigenous groups in northern Siberia also showed a very low prevalence, being less than 1% in 1994 . This suggests that a healthy lifestyle with much physical activity provides protection from the development of diabetes. A recent report from Mapuche  showed that the prevalence in 1998 was higher than that reported 15 years ago, as was the prevalence of obesity , suggesting a possible impact of lifestyle changes on the trends in prevalence of diabetes.
Click here to view Chapter 3 references.