Obesity and weight gain have consistently been shown to be the one of the strongest modifiable risk factors for diabetes [140–142]. The ratio of a person’s weight in kilograms divided by the square of their height in meters called the body mass index (BMI) has been used in numerous studies as a surrogate for obesity. In a representative sample of the US population, each unit increase in BMI was associated with a 12% increased risk of T2DM . Compared to people with BMI <22 kg/m2 those with BMI of 25–27 kg/m2 had 2.75 times the risk of diabetes, and each kilogram increase in body weight over 10 years was associated with a 4.5% increase in diabetes risk. Numerous studies have found similar results in different populations [144,145], but the magnitude of risk associated with a given BMI may differ across populations.The distribution of weight and weight gain are also important risk factor for diabetes. Central obesity, that is, deposition of fat in the trunk and abdominal areas, has been shown to be a strong risk factors for diabetes. The surrogate measures of central obesity include circumference of waist, waist-to-hip ratio, and waist-to-thigh ratio. More recent technological advances like dual-energy X-ray absorptiometry (DEXA) and magnetic resonance imaging (MRI), and computer tomography have made it possible to measure subcutaneous and intra-abdominal fat. Intra-abdominal fat has been shown to increase the risk of insulin resistance and diabetes in a number of studies in different populations [145–148] and this effect may be independent of the effects of total body obesity [149,150]. Obesity always occurs when energy intake exceeds energy expenditure.
Physical activity has been consistently reported to be inversely related to future risk of diabetes in most populations. Higher levels of physical activity are associated with a lower risk of diabetes in observational studies . Increased physical activity reduces the risk of obesity but its effect on diabetes risk has been shown to be independent of its effect on body weight. Numerous studies have shown that exercise is related to acute and long-term improvements in insulin sensitivity and reduction in insulin concentrations [152,153]. Several epidemiologic studies in diverse populations [154–161] confirm a dose–response relationship between levels of physical activity and diabetes incidence. One study among Finnish men  found a strong gradient of diabetes risk associated with intensity of physical activity regardless of duration, but most evidence indicates that the total amount (i.e., number of days or minutes) of physical activity per week is a more important determinant.
Sedentary lifestyle has also been shown to be a risk factor for both diabetes and obesity. Prolonged television (TV) watching as a surrogate of a sedentary lifestyle has also been reported to be a risk factor for diabetes [162,163]. Compared to men spending 0–1 h per week watching TV, those spending 2–10 h had 66% higher risk of diabetes, and the risk increased progressively to 187% higher incidence among those spending >40 h per week . In women, every 2 h per day increment in TV watching was associated with a 23% increase in risk of obesity and 14% increase in risk of diabetes . Furthermore, each 2-h increment in sitting at work was associated with a 5% increase in diabetes risk. It is estimated that a relatively active lifestyle (<10 h per week of TV watching and ≥30 min per day of brisk walking) can prevent 43% of new cases of diabetes .
Evidence from intervention studies
Results from the first diabetes prevention trials have clarified the role of physical activity as a risk factor for diabetes. Results from the Malmö trial in Sweden , Da Qing trial in China , Diabetes Prevention Study (DPS) in Finland , and the Diabetes Prevention Program in the Unites States  have all shown that moderate physical activity reduces the progression of IGT to diabetes by 30–58%. In these studies moderate physical activity was included in a lifestyle intervention arm of the trials with dietary changes that include reduction in caloric intake and in the percentage of dietary fat and increase in fiber intake (Table 3.4). In most of these studies it is difficult to discern the individual effect of the dietary intervention as it was combined with modifications in physical activity.
Diet is a phenomenon that is made up of complex interactions between numerous foods and nutrients that are highly correlated and are determined by personal preference, cultural heritage, and socioeconomic factors. Diet is traditionally measured using questionnaires and food diaries. A number of studies have examined the relationship between diabetes and different aspects of diet including absolute intake of nutrients, nutrient intake as a percentage of total energy, dietary patterns, and bioavailability characteristics of foods like glycemic index.
Nutrients and diabetes risk
Animal studies and clinical human studies suggest several plausible mechanisms relating diet to the etiology of T2DM. High-fat diets have been associated with obesity , increased body fat for a given weight , and altered fat distribution . Furthermore, alterations in cell membrane composition induced by the composition of dietary fat could alter membrane fluidity and/or insulin-mediated signal transduction as well as subsequent insulin action .
The results from epidemiologic studies on the relationship of total fat and T2DM have been controversial. In general, most of the migrant [168,169] and retrospective [170,171] studies show a positive relationship between high-fat, low-carbohydrate diets and T2DM, whereas results from prospective studies from different populations have been mixed [172–174] and not as consistent.
The effect of different subtypes of fat such as saturated, polyunsaturated, monounsaturated fat [175,176] and ω-3 fatty acids [177–179] on diabetes risk has also been investigated.
An extensive review  concluded that while neither total fat nor total carbohydrate as proportions of total energy play a major part in the development of T2DM, specific types of fat and carbohydrate are important. For example, they concluded that (1) higher intake of polyunsaturated fat and long-chain n-3 fatty acids (fish oil) may be beneficial, and (2) higher intake of saturated fat and trans-fatty acids may be deleterious.
The possibility of the risk of diabetes being related to carbohydrate consumption has been frequently raised. This hypothesis was generated because of the stronger, more direct, and more immediate challenge that ingestion of carbohydrate presents to the β cells, compared with ingestion of proteins or fats. Results from studies on carbohydrate and diabetes risk have, however, been inconclusive [169,181,182]. Other aspects of carbohydrate composition and metabolism including glycemic index and fiber content have also been examined.
The glycemic index  of food is a measure of the postprandial excursion of glucose as a result of the ingestion of a fixed amount of food. Foods with high glycemic index cause greater excursions of glucose.The postprandial excursions are dependent on the rate of absorption of glucose, which in turn is dependent on a number of factors including the type of carbohydrate (simple or complex) and the amount of fiber. Food with lower glycemic index has been associated with a lower risk of developing diabetes in some studies [182,184], but others have failed to show any association between the glycemic index of foods and the risk of developing diabetes [185,186].
High-fiber intakes have been related to a reduction in the risk of diabetes in some populations [182,184,187–190]. The mechanism by which high fiber reduces the risk of diabetes is unknown. High-fiber diets usually have a lower glycemic index but some studies that have shown a relationship between fiber intake and diabetes have failed to demonstrate a beneficial effect of the glycemic index of the diet . Thiamine and vegetable protein are found in high concentrations in foods that are rich in fiber and have also been shown to reduce 2-h glucose concentration in women and may explain some of the association between fiber intake and diabetes [191,192].
Vegetable and fruit intake
Higher intakes of vegetables have been shown to be associated with a reduction in the risk of diabetes [172,173,193]. The component of vegetables that confers lower risk has not been clearly identified but may include antioxidants such as carotenoids and tocopherols , higher fiber intake, and vitamins [172,195,196]. The effect of fruit consumption on diabetes risk has been inconclusive [173,193].
Dietary patterns and scores
Studies have examined the effect of the whole composition and quality of diet on diabetes risk using various scores and dietary patterns identified by principal component and factor analyses [197,198] on diabetes risk. The results from these studies have shown that a diet that has relatively higher intakes of vegetables and fruits and lower intakes of fat-rich foods has been associated with a reduction in diabetes risk [197,198].
Self-rated dietary patterns can be used to discriminate differences in the nutrient composition in diet. Pima Indians who self-reported a diet more similar to their traditional Pima diet had a lower risk of developing diabetes over a 12-year period  when compared to a more Westernized dietary preference. The traditional Pima diet had a higher fiber and lower fat intake when compared to the more Westernized diet. The traditional diets of most indigenous populations have a higher intake of fiber and lower glycemic index [200,201] than the more recent Westernized diets that most of these populations have adopted.
Nontraditional modifiable risk factors for type 2 diabetes
Although insulin resistance and relative insulin deficiency represent the main characteristics of T2DM, the underlying mechanism responsible for these abnormalities remains largely unknown. On the basis of hypothesis that both T2DM and atherosclerosis may generate from a “common soil” [202,203], more recently a growing body of evidence points out that inflammation may constitute the common factor that leads to the development of both these diseases. One of the most used markers of subclinical inflammation is the C-reactive protein (CRP). A number of cross-sectional studies have shown that increased concentrations of CRP are associated with abnormalities characterizing the metabolic syndrome, including obesity, insulin resistance, low HDL cholesterol, and hypertriglyceridemia [204–206]. In addition, prospective studies have demonstrated that high levels of CRP increase the risk of developing T2DM [207–212]. These findings have opened new avenues for understanding the pathogenesis of T2DM and, eventually, for preventing the disease. Indeed, if, as these data have indicated, subclinical inflammation is an important determinant of T2DM, then the use of anti-inflammatory drugs may prevent diabetes. In this regard, it is noteworthy that in patients with T2DM high-dose aspirin reduces insulin resistance and improves glucose tolerance .
Numerous prospective studies have indicated that smoking is associated with the development of diabetes. In the US Nurses’ Health Study, women who smoked at least 25 cigarettes per day compared to those who never smoked had a relative risk of developing diabetes of 1.42, even after controlling for known risk factors . Similar results were shown in men .
These findings were confirmed by Will et al. in a prospective study involving over 275,000 men and 434,000 women from the US Cancer Prevention Study . Among those who smoked at least two packs per day at baseline, men had a 45% higher diabetes rate than men who had never smoked; in women the comparable increase was 74%. More important, quitting smoking reduced the rate of diabetes to that of nonsmokers after 5 years in women and after 10 years in men. In support of these findings are cross-sectional data on the association between cigarette smoking and hemoglobin AlC (HbAlC) from the East Anglian component of the European Prospective Investigation into Cancer (EPIC-Norfolk) . In this study, current smokers had highest mean HbAlC concentrations, lowest levels were observed in never smokers, and intermediate in former smokers. HbAlC levels also correlated in a dose–response manner both with the number of cigarettes smoked per day and with total amount of smoking as measured by pack-years. This association persisted even after adjusting for potential confounders including BMI, waist-to-hip ratio, physical activity, and dietary variables. However, to better assess the role of smoking as a determinant of T2DM we need cohort studies to further confirm that this association exists in different populations.
There is evidence that smoking is associated with insulin resistance [217,218]. In a study by Facchini et al.  chronic smokers compared to nonsmokers had significantly higher plasma triglycerides and lower HDL-cholesterol levels and higher insulin concentrations after a 75-g oral glucose challenge. Well-designed clinical studies of the effects of acute and chronic smoking on insulin resistance are needed to elucidate the mechanism by which smoking induces insulin resistance. It is plausible that oxidative stress caused by smoking may induce endothelial dysfunction, resulting in insulin resistance into muscle and liver.
The possibility that smoking may play a causal role in the development of T2DM has important implications for prevention. Both diabetes and smoking are conditions common enough that even a small increase in the risk of diabetes associated with smoking may have an important public health impact.