Etiology and Epidemiology

Insulin-Dependent Diabetes Mellitus (Type I)

Insulin-dependent DM is characterized by clinically acute onset, usually at an early age, lymphoid infil­tration of the islets of Langerhans, reduction in the functioning and production of their betacells, reduc­tion in the production and excretion of insulin, in­creases in islet cell antibodies, weight loss, thirst, frequent urination, and high levels of blood sugar.

Typical type I diabetes is uncommon. The preva­lence of insulin-dependent diabetes mellitus is less than 0.5 percent of the world’s population. The re­ported prevalence of type I diabetes ranges from 0.1 to nearly 4.0 per 1,000 children under 20 years of age. There are 18 registries evaluating geographic patterns of type I diabetes, many in northern Europe and North America, and as much as a 35-fold differ­ence in prevalence has been reported among popula­tions around the world.

Genetic Markers. The etiology of type I diabetes is multifactorial. There is strong evidence for a genetic susceptibility associated with the antigens of the major histocompatibility complex, the human leuko­cyte antigens (HLA). Among individuals in whom DM occurs before the age of 16 years, 90 to 96 per­cent carry either HLA DR3 or HLA DR4. These are present in about 22 percent and 25 percent of the white population, respectively. The presence of one of these antigens increases the risk of the occurrence of DM 3- to 4-fold. Since only 30 to 50 percent of even these individuals develop DM, environmental fac­tors also appear to have an etiologic role.

Infectious Diseases and Immune Responses. The role of infectious diseases, particularly viruses, cou­pled with the appropriate HLA type, has been inves­tigated in a number of populations. Although there is ample experimental and clinical evidence to impli­cate the role of Coxsackie B4, mumps, rubella, and other viruses in the destruction of the betacells of the pancreas, the epidemiological data are not defini­tive. There were early reports of mumps-induced pancreatitis. In 1927 E. Gunderson observed that mortality rates from diabetes increased 2 to 4 years after epidemics of mumps. More recently, other vi­ruses have been under investigation. For example, investigators in Poland discovered increases in anti­bodies against enteroviral meningitis and Coxsackie A and B viruses in newly diagnosed diabetics. The meningitis epidemics regularly preceded periods of increased incidence of diabetes by 4 to 6 months.

Many studies have attempted to link infectious diseases to seasonal patterns in the onset of diabe­tes. The results have not demonstrated a consistent picture. Furthermore, seasonal variation may be linked to factors other than infectious diseases. Worldwide, there is apparently no “peak season” for the appearance of type I diabetes.

Recent work indicates that there is an auto­immune factor in the destruction of the beta cells of the pancreas. Circulating immune complexes have been found in 30 to 90 percent of all newly diagnosed type I diabetics. Antibodies against islet cells are the most common, but antibodies against gastric pari­etal cells, thyroid, and adrenal tissue have also been detected in some diabetics. A strong association ex­ists between the presence of autoimmune antibodies and HLA DR3.

Other Population Genetic Factors. Genetic factors at the population level may play a role in the etiol­ogy of type I diabetes. The relative risk of HLA DR3 is lowest for African blacks and is approximately the same for Caucasians, African-Americans, and Japa­nese (relative risk: 3.3-3.5).

Low rates of type I DM are found among New World Asian-derived native populations (Eskimos, Aleuts, and Amerindians), as well as Old World Asian populations, including the Japanese and Chi­nese. Prevalence rates vary considerably on the In­dian subcontinent. Rates are lower in the southern part of India (i.e., Bombay) than in northern India (i.e., New Delhi). K. M. West (1978) cites a number of studies indicating an increased prevalence of type I diabetes among populations that previously showed low rates. Increases are particularly marked among Japanese, black Africans, and black Americans. In general, these increases are associated with a “West­ernization” in life-style since World War ∏.

Biological Factors. Sex and age also play roles in the onset of type I diabetes. In most populations, the number of boys and girls afflicted with the disease is nearly equal. However, in children under 5 years of age, the incidence is slightly higher for boys. Age of onset of type I diabetes peaks at the beginning of puberty and during the adolescent growth spurt for both girls and boys.

Nutritional Factors. Some nutritional factors have been suspected of being involved in the etiology of type I diabetes, although again there is no consistent epidemiological picture. Excess caloric intake does not seem to be the important etiologic factor it is in type II diabetes. A Danish study has reported a nega­tive correlation between breast-feeding and type I DM, and finally T. Helgason and M. R. Jonasson (1981) suspect that ΛΓ-nitroso compounds from smoked mutton may be responsible for the seasonal incidence of type I diabetes found in Iceland, where this meat is traditionally consumed at Christmas time. The range of nutritional factors implicated in type II diabetes is much greater and is discussed below.

Other Factors. A further risk factor in type I DM is urbanization with accompanying changes in life­style and a greater risk of infection due to increased population density.

Non-Insulin-Dependent Diabetes (Type II) Between 90 and 95 percent of all individuals with diabetes have type II diabetes and are over the age of 35 years. In contrast to type I diabetes cases, there are many cases (up to 50 percent) that remain undi­agnosed. Type II diabetics produce insulin but may require more of it in order to manage their glucose levels. The majority of all patients are treated with dietary modifications, often with caloric reductions for weight loss, and with oral hypoglycemic tablets. Type II diabetics are ketosis resistant but may suffer from the macrovascular and microvascular complica­tions ofketosis (see Table VIII.34.1).

Many factors have been implicated in the etiol­ogy and pathogenesis of type II diabetes. Not only are they numerous but also their interactions are complex.

Genetic Factors. The numerous problems in discov­ering a genetic mechanism responsible for diabetes have been outlined in works by W Creutzfeldt, J. Kobberling, and J. V. Neel (e.g., 1976), and K. M. West (1978). Unlike type I DM, type ∏ DM (the more common non-insulin-dependent DM) has no associ­ated genetic markers. However, genetic mechanisms Table VIII.34.1. Causes and suspected causes of, and risk factors for, diabetes

Generally accepted

Obesity

Caloric excess, indolence

Heredity

Destruction or damage of beta cells

Cancer

Pancreatectomy

Pancreatitis from many causes

Viral infections

Coxsackie viruses, measles

Beta cytotoxic drugs

Diabetogenic hormones (exogenous or endogenous) Growth hormone

Epinephrine

ACTH or glucocorticoids

Hemochromatosis

Disorders of insulin receptors

Factors of immunity and autoimmunity

Human leukocyte antigens (HLA)

Widely suspected of causing or precipitating diabetes

Pregnancy

Excessive serum iron

Cassava (manioc or tapioca) consumption

Some drugs

Cirrhosis

Potassium deficiency

Certain brain lesions

Affluence in some circumstances and poverty in others Myotonic dystrophy

Suspected by some scientists but doubted by others

Dietary fat

Dietary sugars

Refined carbohydrates

Insufficient dietary fiber

Contraceptive steroids

Deficiency states

Chromium, iron, zinc, pyroxidine

Severe protein malnutrition

Psychological stress

Stress of prolonged or severe illness

Temperate climate

Hypertriglyceridemia

Specific racial susceptibility

(e.g., Jews, Indians, American Indians, Polynesians)

Source: Adapted from K.

interacting in complex ways with environmental fac­tors are involved in the risk for type II diabetes.

Type II DM occurs more frequently in families in which one or more members have DM than in the general population. Yet family studies have not sup­ported a simple mode of inheritance. A 40-fold risk of DM occurs for children of fathers who had an age of onset of diabetes under 20 years. Risks are some­what less for children of type I mothers, and are increased only one to three times for offspring of parents who had an age of diagnosis of diabetes over the age of 40.

Biological Factors. Both age and sex are risk fac­tors for type II diabetes. In most affluent societies, the rate of diagnosed type II diabetes increases steadily from 30 to 60 years of age. The decline in the rate in old age may be the result of the smaller number of individuals at risk. The extent to which physiological aging is a risk factor is unknown. How­ever, H. Silwer (1978) showed that for Swedish popu­lations the proportion of cases that were “severe” declined sharply in old age, from 28 percent in the sixth decade to 4 percent in the ninth decade.

Many observations have been made on sex differ­ences in the frequency of type II diabetes. Before 1900, diabetes was observed to be more frequent in men in both Europe and the United States. Indeed R. Lepine cited seven different studies from Europe in which women constituted 17 to 43 percent of the cases, whereas Emerson and Larimore found that in England and Wales the male/female ratios were typi­cally 2:1 before 1910. Since 1930, however, clinicians in both America and Europe have repeatedly ob­served a greater frequency in female diabetics.

Nonetheless, many developing nations show a male predominance. Populations with a high male/ female ratio include rural Africans, Hong Kong Chi­nese, and populations in Iraq, Jordan, Japan, Korea, India, and Pakistan. Populations with a predomi­nance of females include the United States, the Ca­ribbean countries, Sweden, Belgium, the former So­viet Union, Thailand, and South Pacific countries.

Parity has also been considered a risk factor. In 1924, Emerson and Larimore reported that married women in New York City had a rate of mortality from diabetes about four times higher than that of single women. Joslin and his associates, however, found that married diabetic women weighed on the average about 20 pounds more than single diabetic women, and in well-controlled studies no relation­ship was found between parity and diabetes in Pima Indians. On the other hand, it is true that pregnancy can have a physiological effect that is diabetogenic, and women who suffer from gestational diabetes are at greater risk for type II diabetes in their nonpreg­nant state. In summary, it appears that parity is not a major risk factor for diabetes, and that the associa­tion is confounded by both obesity and age.

Exercise. Exercise would seem to be a potent protec­tive factor in diabetes. It enhances both carbohy­drate metabolism and the efficient use of insulin. As early as 1893, B.C. Sen noted that active men in India had a lower rate of diabetes than their more sedentary counterparts. Overall, a decrease in en­ergy expenditure has been related to changes in life­style, a relationship that has been particularly well- documented in the South Pacific. Unfortunately, quantitative data on exercise are limited, and the short- and long-term effects of exercise on carbohy­drate metabolism are incompletely understood.

Obesity. There is abundant worldwide evidence as­sociating obesity with type II diabetes, and the 1980 World Health Organization Expert Committee on Diabetes has concluded that it is the most powerful risk factor for non-insulin-dependent DM. Many studies have shown that between two-thirds and nine-tenths of individuals may be classified as obese at the time of diagnosis of DM.

This association between obesity and diabetes is an old one, first noted by Indian physicians. Modern concern with obesity began with Joslin’s landmark paper in 1921, in which he assembled overwhelming data showing obesity as the major risk factor for type II diabetes. Diabetes rates were 6 to 12 times greater in obese than in lean individuals. More re­cently, West (1978) documented high frequencies of both obesity and diabetes among a number of Amer­indian tribes in Oklahoma. Later works on native American groups have corroborated these findings. The positive relationship of diabetes with obesity is also found in Latin American populations. Yet it does not hold true for all peoples; in some popula­tions with high levels of both conditions - for exam­ple, the Pima Indians and Mexican-Americans - for specific age and sex groups diabetics are not more obese than nondiabetics.

The distribution of body fat presents an additional risk factor. A centripetal distribution of fat around the chest, waist, and abdomen presents a greater risk than an increase in fat in the lower portion of the torso, hips, and thighs. This central or android distribution of fat has been correlated with risk for diabetes in a number of ethnic and racial groups. The duration of the diabetes and other hormonal factors may be additional risks related to the amount of fat tissue or degree of obesity.

Dietary Factors. Calories are units of energy de­rived from food, and high caloric intake and low caloric expenditure are both related to obesity. Diabe­tes rates appear to have increased markedly in a number of countries wher caloric consumption per capita has also increased. These countries include Japan, Taiwan, Haiti (among the wealthy), New Guinea, and many parts of Africa. A number of au­thors have noted that during World War I and II in Europe and during World War ∏ in Japan, when caloric intake was markedly decreased, the rate of obesity and diabetes both declined. On the other hand, recent work in 1975 by J. H. Medalie and his colleagues indicate no significant differences in calo­ric intake among diabetic and nondiabetic Israeli men.

Many dietary components have been investigated as potential risk factors for type II diabetes. Most researchers agree that there is no convincing evi­dence that a single dietary component increases the risk of diabetes. For example, West reported 21 stud­ies indicating that table sugar or sucrose consump­tion was a risk factor for diabetes and another 22 studies suggesting that it was not. Furthermore, the distinction between simple and complex carbohy­drates has been called into question in recent years. A number of researchers focusing on previously des­ignated complex carbohydrates (starches such as po­tato and wheat bread) have shown that these starches can lead to the production of high levels of insulin and glucose in nondiabetics and type II dia­betics. In fact, the glucose response to white potatoes and dextrose sugar was approximately the same.

H. C. Trowell, D. P. Burkitt, and colleagues sug­gest that dietary fiber favorably alters absorption patterns to decrease the risk for diabetes, and cur­rently many different forms of dietary fiber are be­ing investigated in this connection. In most of the epidemiological work, however, it is difficult to disen­tangle the effects of decreased fiber consumption from increased sugar and refined carbohydrate in­take, increased total calories, total fat, decreased caloric expenditure, and stresses associated with rapid dietary change and modernization. Thus, there remain many unanswered questions concern­ing the role of refined and unprocessed carbohy­drates and dietary fiber in the risk and treatment of DM.

Dietary fats have also been viewed as dietary risk factors, although studies investigating fat intake are often confounded because of the higher caloric density of fat compared to carbohydrates and pro­teins (ratio = 9:4:4). Furthermore, high fat intake is generally associated with low fiber intake. Histori­cally, one of the more important works on the matter is that ofHarold P. Himsworth (1935-6), in which he presented evidence positively linking fat and nega­tively linking carbohydrate consumption to diabe­tes. Yet other epidemiological observations have not supported the relationship between high fat con­sumption and high diabetes rates when diet is con­trolled for caloric intake. Furthermore, evidence that challenges any such relationship is found among Eskimos living a traditional life-style with diets high in fat and protein yet suffering little from DM. Alternatively, protein deficiency in the tropics has been implicated in certain types of diabetes, particularly those secondary to pancreatic disease.

Finally, deficiencies in micronutrients - in particu­lar, chromium — have been postulated as potential factors for diabetes. However, there are few epidemi­ological data and the clinical data are conflicting.

Other Factors. Other life-style factors such as ur­banization and psychosocial stress have been impli­cated in the etiology of type II diabetes. As popula­tions become more “developed,” there is an increase in diabetes. The populations of the South Pacific, for example, provide a natural laboratory in which the degree of modernization has been correlated with increases in chronic diseases including diabetes. Many other factors associated with diabetes are listed in Table VIII.34.1.

Tropical Diabetes. A third type of diabetes found primarily in developing countries in tropical areas has clinical characteristics of both type I and type II diabetes. Tropical diabetes has been described in many areas of the world including the South Pacific, the West Indies, Africa, Southeast Asia, and India. Thus, the disease has been referred to variously as phasic insulin-dependent diabetes, J-type diabetes for Jamaica, K-type diabetes for Kenya, malnutri­tion diabetes, and tropical pancreatic diabetes. The clinical profile involves the following: (1) a different genetic pattern of diabetes than in temperate re­gions; (2) a low prevalence rate of type I DM; (3) a younger age on onset of type II; (4) a sex ratio with male predominance in India and Africa, but female predominance in the West Indies; (5) an association of low calorie and protein intake during childhood and adulthood with the presentation of lean or under­weight diabetic individuals in Old World tropical areas, but overweight diabetic individuals in the Western Hemisphere; (6) the predominance of diabe­tes in urban areas in most countries, with the excep­tion of rural sugar-cane farming populations in the West Indies; (7) ketosis resistance; and (8) intermit­tent need for insulin therapy.

Information is relatively sparse on the genetics of diabetes in tropical countries. Recent studies on the human leukocyte antigens (HLA) and properdin (fac­tor B, BF) system have shown that there is great population variability in the association Ofparticular HLA haplotypes and increased susceptibility to diabe­tes. Genetic studies of Indian populations suggest a stronger familial or genetic factor among Indian dia­betics compared to diabetics in other populations.

Environmental risk factors for tropical diabetes involve unique dietary items. In 1979, D. E. Mc­Millan and P. J. Geevarghese suggested that the cyanide-containing glycosides of some types of cas­sava (manioc) may be toxic to the beta cells of the pancreas and thus produce pancreatic damage. Ab­normal changes in pancreatic tissue give support to the cassava-cyanide hypothesis in the pathogenesis of diabetes. However, many tropical areas show high rates of diabetes in populations that do not consume cassava, and conversely, some populations have high cassava consumption and low rates of diabetes. In Kenyan (K type) diabetes, a local alcohol called changaa is implicated in causing the disease. Fi­nally, in most tropical areas carbohydrates consti­tute 70 to 80 percent of the total calories of the diet, which is considerably higher than the percentage of carbohydrates in diets of many developing countries. Indeed, such a diet is implicated in classic “malnutri­tion diabetes” because of the relatively low nutrient density and high fiber content.

Gestational Diabetes. In 1882, J. M. Duncan noted a type of diabetes present only during pregnancy. However, it was not until the 1940s that the term “gestational diabetes” appeared in the English lan­guage medical literature. D. R. Haddon found that from 1975 to 1984, 165 studies on gestational diabe­tes were published, representing investigations in 25 countries. The majority (69 percent) were from the United States and Europe, although 16 were reported from Australia. One important difficulty in evaluating the incidence of gestational diabetes is that the frequency of type II diabetes among women of childbearing age is unknown for many countries. Therefore, a woman could have diabetes prior to pregnancy, but not have it diagnosed until preg­nancy. Babies bom to diabetic mothers usually are large (over 9 pounds), but may have organ systems that are not mature, in which case they may not survive.

The prevalence of gestational diabetes varied from 0.15 percent in Newcastle, United Kingdom, to 12.2 percent in Los Angeles, California. In general, cities in the United States report a higher prevalence of gestational diabetes then do European cities. The one African report from Nairobi, Kenya, indicates a prevalence of 1.8 percent. The highest reported rate of gestational diabetes occurs among the Pima Indi­ans of Arizona, who also have the highest prevalence of type II diabetes of any known population.