Racial/ethnic differences in the burden of type 2 diabetes over the life course: a focus on the USA and India
Type 2 diabetes is a common disease worldwide, but its prevalence varies widely by geographical region and by race/ethnicity. This review summarises differences in the frequencies of type 2 diabetes according to race, ethnicity, socioeconomic position, area of residence and environmental toxins. Type 2 diabetes susceptibility often begins early in life, starting with genetic susceptibility at conception and continuing in later life, via in utero, childhood and adult exposures. Early-life factors may lead to overt type 2 diabetes in childhood or in later life, supporting the concept of developmental origins of health and disease. The causes of the racial/ethnic differences in incidence of type 2 diabetes are not well understood. Specifically, the relative contributions of genetic and environmental factors to such differences are largely unknown. With a few exceptions in isolated populations, there is little evidence that differences in frequencies of known type 2 diabetes susceptibility genetic alleles account for racial/ethnic differences, although the search for genetic susceptibility has not been uniform among the world’s racial/ethnic groups. In the USA, race/ethnicity is associated with many other risk factors for type 2 diabetes, including being overweight/obese, diet and socioeconomic status. Some studies suggest that some of these factors may account for the race/ethnic differences in prevalence of type 2 diabetes, although there is inadequate research in this area. A better understanding of the impact of these factors on type 2 diabetes risk should lead to more effective prevention and treatment of this disease. This has not yet been achieved but should be a goal for future research.
KeywordsLife course development Race/ethnicity Review Type 2 diabetes mellitus
Gestational diabetes mellitus
National Health Interview Survey
Type 2 diabetes is considered a complex disease, in that many known genetic, environmental and personal behavioural risk factors affect disease susceptibility and few, if any, variables are completely determinative. Other, yet undiscovered, risk factors also likely contribute to this disease. In this article, we review these complex causes and discuss the extent to which they may explain the racial/ethnic differences in type 2 diabetes prevalence. We focus on research from the USA and India, regarding factors influencing the development of type 2 diabetes, rather than those influencing management and consequences of the disease, such as the development of acute metabolic and chronic vascular complications. Because the incidence of complications is strongly related to duration of diabetes, factors leading to younger age of onset of type 2 diabetes are also associated with greater prevalence of complications over the lifetime, but whether differences in the prevalence or incidence of complications are due to factors other than those leading to type 2 diabetes itself is beyond the scope of this review. While not a meta-analysis or systematic review, this paper includes findings that we consider representative and relevant to racial/ethnic differences in type 2 diabetes susceptibility.
Although previously considered primarily a disease only of adults (and formerly called ‘adult-onset diabetes’), what is now called type 2 diabetes also occurs in children. Evidence suggests that susceptibility often begins early in life, starting at conception, owing to genetics, and continues in later life via in utero, childhood and adult exposures. Early-life factors may lead to overt type 2 diabetes in childhood or later, confirming the concept of developmental origins of health and disease.
Although race and ethnicity are associated with health and disease incidence [2, 3, 4], these terms have not been used consistently in the medical literature. In the USA, ‘race’ indicates continent or region of ancestral origin and ‘ethnicity’ (Hispanic or not) indicates cultural identity. Hispanic ethnicity refers to people who identify with a Spanish-speaking culture. Any racial group may also identify as Hispanic or non-Hispanic. Thus, in the USA ‘non-Hispanic black’ refers to people of African descent who do not identify as Hispanic; ‘non-Hispanic white’ refers to those with ancestry from Europe or the Mediterranean who do not identify as Hispanic; Asian-American refers to individuals of East Asian, South Asian, South East Asian and Pacific Island ancestry; and American Indian/Alaska Native to descendants of the original inhabitants of the Americas [5, 6]. These categories are based on self-report and include heterogeneous groups. Analogous terms are not used uniformly in other parts of the world. In this review, when studies are discussed that use more specific terms in defining racial/ethnic subgroups, we will use them accordingly.
Genetics and differences in diabetes prevalence
Given the many genetic differences among racial/ethnic groups and the well-recognised genetic influences on type 2 diabetes, genetic factors probably account, in part, for the differences in diabetes prevalence. However, few studies have quantified the contributions of genetic factors to these racial/ethnic differences. In Mexican-Americans, a higher proportion of American Indian ancestry, determined by skin reflectance, was associated with higher prevalence of diabetes ; similar findings were seen in Pima Indians, with American Indian ancestry determined by self-report . Studies that used ‘traditional’ genetic markers, such as blood groups, to estimate American Indian heritage also found positive associations with diabetes in Pima and Mexican-American populations [9, 10, 11]. With the advent of modern genetic techniques, large numbers of ancestry-informative genetic markers can be typed, resulting in more robust estimates of genetic ancestry. Studies using these techniques in Hispanic populations have also found that a higher level of American Indian ancestry is associated with higher prevalence of diabetes [12, 13, 14, 15]. However, American Indian ancestry was also associated with socioeconomic status (SES), and adjustment for SES greatly attenuated the risk associated with genetic admixture. In studies of African-American individuals, a higher proportion of African ancestry was associated with increased prevalence of diabetes and this finding persisted with adjustment for SES . Given the imprecision in assessing environmental sociocultural factors shared within ancestry groups, genetic ancestry associations may be confounded by environmental factors. One might expect comparisons of genetic ancestry estimates within sibships to be robust to such confounding, but such studies require large sample sizes and large numbers of genetic markers to confidently assess the modest differences in genetic admixture among siblings.
Early-life risk factors for type 2 diabetes
Some evidence of early-life determinants of type 2 diabetes comes from India and is discussed here as a model of diabetes susceptibility in economically deprived societies. With rapid economic growth, India has had an alarming rise in type 2 diabetes prevalence out of proportion to its increase in affluence . The rise has been most pronounced in the states that have suffered financial, geographical or sociopolitical difficulties. This points towards a role for previous deprivation in the evolution of the diabetes epidemic. Traditionally, type 2 diabetes susceptibility is ascribed to a ‘thrifty genotype’ that evolved over millennia. However, the ‘thrifty phenotype’ hypothesis  suggests a more recent establishment of susceptibility through epigenetic programming due to enhanced survival in the setting of intergenerational undernutrition, especially during the first 1000 days of life (intrauterine and first 2 years). Thus, intergenerational influences seem to operate both genetically and epigenetically.
Adults with type 2 diabetes in the Indian subcontinent differ in phenotype from the ‘textbook’ description of European individuals . The differences may be explained, at least in part, by early-life exposures. Indian individuals with type 2 diabetes are diagnosed at least a decade earlier and have a lower BMI but higher central obesity at diagnosis . Body fat percentage in Indian populations is higher than that in European individuals at a given BMI [28, 29]; this ‘thin-fat’ phenotype is associated with higher insulin resistance and diabetic dyslipidaemia. Recent investigations also revealed that Indian people with type 2 diabetes have reduced beta cell function compared with Europeans . This phenotype is clearly not attributable to adult lifestyle factors only; a comparison of newborn Indian and English babies revealed that it originates during intrauterine development , supporting the ‘thrifty phenotype’ hypothesis [25, 32]. This ‘thin-fat’ and impaired beta cell function phenotype suggests a new explanation for the high susceptibility of Indian individuals to type 2 diabetes since more than a quarter of Indian babies have low birthweight (<2.5 kg), multigenerational undernutrition with small maternal size is common, and Indian babies are among the smallest in the world. A small and thin Indian newborn (mean weight 2.7 kg) has comparable subscapular skinfold thickness to an English baby (mean weight 3.5 kg)  and higher subcutaneous and visceral fat as assessed by MRI . The cord blood of Indian babies also has higher concentrations of leptin and insulin, but lower concentrations of adiponectin, confirming a high-risk profile for future type 2 diabetes . The contribution of genetics and epigenetics to this phenotype is currently being actively investigated [35, 36].
The Pune Children’s study demonstrated that children born small had higher glucose and insulin concentrations during a glucose tolerance test . Babies born small who grew to have high fat mass later in childhood had the highest levels of risk factors for type 2 diabetes and cardiovascular disease (insulin resistance, high adiposity, lipid abnormalities and high blood pressure) . The Delhi cohort expanded these findings in young individuals with glucose intolerance: glucose intolerant individuals were born smaller compared with glucose tolerant participants, grew poorly in the first 2 years of life, and then progressively increased in weight and BMI until adult age. Though not obese by international criteria, individuals who had greater weight gain in relation to their early years were at higher risk of type 2 diabetes .
This life course trajectory exemplifies the risk of ‘double burden’ of malnutrition on lifetime risk for type 2 diabetes, a common situation in the developing populations of the world. The ‘dual teratogenesis’ construct  describes the progressive evolution of risk factors and type 2 diabetes in the intergenerationally undernourished who tend to be ‘thin-fat’, insulin resistant and beta cell deficient in rural environments, but become obese and glucose intolerant when they live in more urbanised conditions. Urbanisation produces an unmanageable load of overnutrition and physical inactivity on a multigenerationally low capacity system [41, 42]. This is supported by the finding that urban middle-class Indians were more insulin resistant than urban slum dwellers and individuals living in rural areas . In addition, beta cell responses in South Asian individuals were incommensurate with the degree of insulin resistance, signifying beta cell dysfunction  and suggesting that pancreatic development is suboptimal in early life.
Diet probably plays a major role in influencing intergenerational susceptibility to type 2 diabetes. In Indian and other populations in which vegetarianism is common, there is a substantial prevalence of vitamin B12 deficiency. Maternal vitamin B12 deficiency and hyperhomocysteinaemia are related to poor fetal growth, and subsequent insulin resistance and adiposity in childhood . High folate status worsens the situation, suggesting that a balance between vitamin B12 and folate is essential. Moreover, Indian diets are associated with macronutrient imbalances; they tend to be high in energy and carbohydrates and have a high glycaemic index . In addition to diet, South Asians, as a group, seem to be less physically active than their European counterparts . All these factors combine to bring about the higher, earlier susceptibility to type 2 diabetes in South Asian populations.
Prevalence and incidence of type 2 diabetes in youth
Prevalence of diagnosed type 2 diabetes in youth in the USA by race/ethnicity
We hypothesise that the effects of low birthweight, early nutritional deficiencies, and in utero exposure to maternal type 2 diabetes are not limited to any racial/ethnic group, but are magnified in certain groups with high risk of young-onset type 2 diabetes, especially American Indians. The vicious cycle of transgenerational inheritance of diabetes risk probably exists in all populations in which type 2 diabetes occurs before reproductive age, but it may be evident only in those racial/ethnic groups in which type 2 diabetes frequently occurs at young ages. This likely explains the more rapid rise in incidence of type 2 diabetes from 2003–2012 in American Indians and non-Hispanic black individuals vs other racial/ethnic groups in the USA, in which rates were lower at the start of the study (Fig. 3) . Similarly, the effects of early nutritional deficiencies and developmental delays are likely to have the greatest population-level impact and, thus, be most evident in racial/ethnic groups that have higher type 2 diabetes incidence rates at young ages.
Prevalence and incidence of type 2 diabetes in adults
Diabetes in adults: prevalence
Age-adjusted prevalence of diagnosed type 2 diabetes in the USA by race/ethnicity in adults ≥18 years of age 
Age-adjusted prevalence (%)
American Indian/Alaska Native
Age-adjusted prevalence of diagnosed type 2 diabetes in adults aged 20–79 years, in different regions of the world in 2017 
Age-adjusted prevalence (%)
Middle East and North Africa
North America and the Caribbean
South and Central America
South East Asia
The Indian subcontinent is an epicentre of diabetes burden [10, 65]. India is second only to China in the number of people with diabetes. There is great population diversity within India, and data are available by state, urban/rural residence, and socioeconomic position, but not by ‘race/ethnicity’ as defined in the USA. The South Asian region is expected to have the world’s largest number of individuals living with diabetes by 2030 . The prevalence varies by country within the South Asian region, with India having the largest burden.
There have been a large number of epidemiological reports of diabetes prevalence in India over the last 30 years, most from South India. The recently reported Indian Council of Medical Research–India Diabetes (ICMR–INDIAB) study was the largest population-based study, carried out over 15 states, representing half of India’s vast population . The overall prevalence of diabetes was estimated at 7.3%. Highlighting the heterogeneity of the population of India, there was wide variation in the prevalence of type 2 diabetes in different states, which paralleled the per capita gross domestic product (GDP). Prevalence was higher in urban vs rural areas. While diabetes was higher among the affluent in rural areas, in urban areas it was more common in those with lower socioeconomic scores .
The rise in diabetes prevalence in India parallels a worldwide rise from 4.7% in 1980 to 8.5% in 2014 . In a recent comparison between two surveys in the southern Indian state of Tamil Nadu, using similar methods carried out 10 years apart (2006 and 2016) , diabetes prevalence increased significantly in a city (from 18.6% to 21.9%), a town (16.4% to 20.3%), and in peri-urban villages (9.2% to 13.4%). At both time points, there was an urban–village gradient, with prevalences being highest in the city (defined as having a population ≥4 million) and lowest in villages, where occupations were primarily agricultural. In addition to age and family history of diabetes, central obesity was associated with increases in prevalence.
Diabetes prevalence and migration
Studies that compare prevalence of diabetes in migrants to that of individuals in their country of origin can yield information about the importance of environmental factors, under the assumption that the two groups are genetically similar. In general, migrants to more urbanised, industrialised countries have a higher prevalence of diabetes than individuals in their countries of origin . For example, populations of African origin living in the USA or the UK have higher prevalence of diabetes than those living in African countries, whereas prevalence is intermediate in individuals of African ancestry living in Jamaica [69, 70]. Similarly, migrants from the Polynesian atoll of Tokelau living in New Zealand had a higher risk of diabetes than those who remained in Tokelau , and second-generation Indian immigrants in Singapore had a higher prevalence of diabetes than first-generation migrants . On the other hand, in a comparison between Indians living in the USA and those living in urban India (Chennai), those in India had higher prevalence of diabetes than the Indian-Americans . In addition, the prevalence of diabetes in urban Ghana was much higher than in rural Ghana and comparable with that of African-origin populations in urban Europe . Moreover, among Peruvians who migrated from rural to urban areas, the risk of developing diabetes was much higher than among those who remained in the rural regions . These studies implicate environmental factors associated with urbanisation as contributing to the high risk of diabetes in migrant populations.
Diabetes in adults: incidence
In adults from the USA, aged ≥18 years, there were approximately 1.5 million new cases of diabetes according to the National Health Interview Survey (NHIS) in 2015 . The annual age-adjusted incidence of diagnosed diabetes was higher in non-Hispanic black individuals (9.0 per 1000 people) and individuals of Hispanic origin (8.4 per 1000 people) compared with non-Hispanic white individuals (5.7 per 1000 people) and Asians (6.0 per 1000 person-years). Overall diabetes incidence in the NHIS increased rapidly during 1990–2008, followed by a levelling off and then a decrease from 2008–2012; however, this primarily occurred in non-Hispanic white individuals, while the incidence continued to rise in non-Hispanic black people and Hispanic individuals during this same time period . We were unable to identify regional data on worldwide diabetes incidence.
Other factors that may contribute to racial/ethnic differences in type 2 diabetes
Genetic variation, as discussed above, is an obvious candidate that may explain racial/ethnic variation in type 2 diabetes. A limitation is that most research on genetics of diabetes has been conducted in the USA, Europe or eastern Asia, but most areas of the world have not been well represented. Therefore, genetic variants that may have large effects but are found only in some populations may not have been discovered. Factors in the external environment must also be considered. Arsenic is a naturally occurring toxin in many parts of the world. Exposure to it has been associated with the prevalence of type 2 diabetes in several different countries, including Bangladesh [76, 77], Mexico  and the USA [79, 80]. Environmental exposure to persistent organic pollutants and several heavy metals has also been associated with diabetes [81, 82, 83]. There is also evidence that infection with the hepatitis C virus, and perhaps other viruses, may increase diabetes risk . Such environmental factors vary geographically, so they may increase diabetes incidence in areas that are inhabited predominantly by certain racial/ethnic groups. Thus, some racial/ethnic differences in disease susceptibility could be caused by environmental exposures that differ by race/ethnicity. If the environmental exposures are not known or considered, these differences could be erroneously attributed to genetics or other factors.
Race/ethnicity in preventing type 2 diabetes
Despite the large differences in diabetes incidence rates among different populations around the world, the progression from high-risk states (such as being overweight or obese with impaired glucose regulation, or so-called ‘prediabetes’) to overt diabetes has been shown to occur at similar rates and responded similarly to preventive interventions among different racial/ethnic groups in the USA Diabetes Prevention Program . Most behavioural or medical interventions for diabetes prevention have worked in racially/ethnically diverse populations around the world . The achieved reductions in rates of type 2 diabetes incidence differed among clinical trials conducted in different countries, but the differences may be due to differences in the types or intensities of interventions among the studies rather than race/ethnicity. To our knowledge, only the USA Diabetes Prevention Program compared effects of the same interventions given to different racial/ethnic groups; the effects of the metformin and lifestyle interventions on reducing progression to type 2 diabetes did not differ significantly by race/ethnicity . Thus, despite the racial/ethnic differences in diabetes prevalence, current methods for diabetes prevention are the same across racial and ethnic categories.
This review summarises differences in the frequencies of type 2 diabetes according to race, ethnicity, socioeconomic position, area of residence and environmental toxins. The causes of these differences are not well understood. Specifically, the relative contributions of genetic and environmental factors to racial/ethnic differences are largely unknown. With a few exceptions in isolated populations, there is little evidence that differences in frequencies of known type 2 diabetes susceptibility genetic alleles account for racial/ethnic differences in type 2 diabetes, although the search for genetic susceptibility has not been uniform among the world’s racial/ethnic groups. In the USA, race/ethnicity is associated with many other risk factors for type 2 diabetes, including being overweight/obese, diet and SES. Some studies suggest that some of these factors may account for the race/ethnic differences in prevalence of type 2 diabetes, although there is inadequate research in this area. Better understanding of these factors should lead to more effective prevention and treatment. This has not yet been achieved to a great extent but should be a goal for the future.
CY thanks R. Wagh (Diabetes Unit, KEM Hospital and Research Center, Pune, Maharashtra, India) for assistance.
All authors were responsible for drafting the article and revising it critically for important intellectual content. All authors approved the version to be published.
In some of the studies mentioned in the review, CY has been supported by grants from: The Wellcome Trust, UK; the Medical Research Council of the UK; the Department of Biotechnology of the Government of India; and the Indian Council of Medical Research. RLH and WCK are supported by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases, USA.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
- 1.Cowie CC, Casagrande SS, Geiss LS (2018) Prevalence and incidence of type 2 diabetes and prediabetes. In: Cowie CC, Casagrande SS, Menke A et al (eds) Diabetes in America, 3rd edn. National Institutes of Health, Bethesda, MD, pp 3–1–3–32Google Scholar
- 8.Knowler WC, Williams RC, Pettitt DJ, Steinberg AG (1988) Gm 3;5,13,14 and type 2 diabetes mellitus: an association in American Indians with genetic admixture. Am J Hum Genet 43(4):520–526Google Scholar
- 10.Hanis CL, Chakraborty R, Ferrell RE, Schull WJ (1986) Individual admixture estimates: disease associations and individual risk of diabetes and gallbladder disease among Mexican-Americans in Starr County, Texas. Am J Phys Anthropol 70(4):433–441. https://doi.org/10.1002/ajpa.1330700404 Google Scholar
- 11.Chakraborty R, Ferrell RE, Stern MP, Haffner SM, Hazuda HP, Rosenthal M (1986) Relationship of prevalence of non-insulin-dependent diabetes mellitus to Amerindian admixture in the Mexican Americans of San Antonio, Texas. Genet Epidemiol 3(6):435–454. https://doi.org/10.1002/gepi.1370030608 Google Scholar
- 20.Klimentidis YC, Abrams M, Wang JL, Fernandez JR, Allison DB (2011) Natural selection at genomic regions associated with obesity and type-2 diabetes: East Asians and sub-Saharan Africans exhibit high levels of differentiation at type-2 diabetes regions. Hum Genet 129(4):407–418. https://doi.org/10.1007/s00439-010-0935-z Google Scholar
- 27.Yajnik CS (2001) The insulin resistance epidemic in India: fetal origins, later lifestyle, or both? Nutr Rev 59(1 Pt 1):1–9. https://doi.org/10.1111/j.1753-4887.2001.tb01898.x Google Scholar
- 33.Modi N, Thomas EL, Uthaya SN, Umranikar S, Bell JD, Yajnik C (2009) Whole body magnetic resonance imaging of healthy newborn infants demonstrates increased central adiposity in Asian Indians. Pediatr Res 65(5):584–587. https://doi.org/10.1203/01.pdr.0000350364.10602.33 Google Scholar
- 37.Yajnik CS, Fall CH, Vaidya U et al (1995) Fetal growth and glucose and insulin metabolism in four-year-old Indian children. Diabet Med 12(4):330–336. https://doi.org/10.1111/j.1464-5491.1995.tb00487.x Google Scholar
- 55.Centers for Disease Control and Prevention (2017) National Diabetes Statistics Report, 2017. Available from: www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf. Accessed 8 April 2019
- 59.Kitzmiller JL, Ferrara A, Peng T, Cissel MA, Kim C (2018) Preexisting diabetes and pregnancy. In: Cowie CC, Casagrande SS, Menke A et al (eds) Diabetes in America, 3rd edn. National Institutes of Health, Bethesda, MD, pp 5–1–5–106Google Scholar
- 61.Link CL, McKinlay JB (2009) Disparities in the prevalence of diabetes: is it race/ethnicity or socioeconomic status? Results from the Boston Area Community Health (BACH) survey. Ethn Dis 19(3):288–292Google Scholar
- 64.International Diabetes Federation (2017) IDF Diabetes atlas, 8th edn. Available from: https://diabetesatlas.org/resources/2017-atlas.html. Accessed 8 April 2019
- 69.Rotimi CN, Cooper RS, Okosun IS et al (1999) Prevalence of diabetes and impaired glucose tolerance in Nigerians, Jamaicans and US blacks. Ethn Dis 9(2):190–200Google Scholar
- 71.Ostbye T, Welby TJ, Prior IA, Salmond CE, Stokes YM (1989) Type 2 (non-insulin-dependent) diabetes mellitus, migration and westernisation: the Tokelau Island migrant study. Diabetologia 32(8):585–590Google Scholar
- 76.Rahman M, Tondel M, Ahmad SA, Axelson O (1998) Diabetes mellitus associated with arsenic exposure in Bangladesh. Am J Epidemiol 148(2):198–203. https://doi.org/10.1093/oxfordjournals.aje.a009624 Google Scholar
- 81.Lee DH, Jacobs DR (2006) A strong dose-response relation between serum concentrations of persistent organic pollutants and diabetes: results from the National Health and Nutrition Examination Survey 1999-2002. Response to Porta. Diabetes Care 29(11):2568. https://doi.org/10.2337/dc06-1720 Google Scholar
- 86.Knowler WC, Crandall JP, Chiasson J-L, Nathan DM (2018) Prevention of type 2 diabetes. In: Cowie CC, Casagrande SS, Menke A et al (eds) Diabetes in America, 3rd edn. National Institutes of Health, Bethesda, MD, pp 38–1–38–21Google Scholar