Metabolic health and type 2 diabetes

Insulin is released by the pancreas after a meal and signals muscle, liver and fat tissue to take up and store glucose. In insulin resistance these tissues respond weakly, so the pancreas must release more insulin to keep glucose normal. Insulin resistance can persist for years before glucose rises into the prediabetic or diabetic range, and it tends to occur together with raised blood pressure, abnormal blood fats and excess abdominal fat, all of which also raise the risk of cardiovascular disease. The amount and location of body fat are therefore central to metabolic health, and are described in more detail under body weight.

Fat stored around the waist carries more metabolic risk than the same weight carried elsewhere. In the EPIC-InterAct study of 340,234 Europeans, waist circumference predicted type 2 diabetes independently of body mass index (BMI, weight in kilograms divided by the square of height in metres), and overweight adults with a large waist had a 10-year diabetes incidence comparable to that of adults with obesity [7]. A 2022 review of 216 studies with 26 million participants found that each 10 cm larger waist was associated with 61% higher risk of type 2 diabetes, and that central fat was linked to diabetes independently of overall fat [8]. In 650,386 adults, a larger waist was linked to higher mortality at every BMI level from 20 to 50 [9].

Skeletal muscle is a major site of glucose uptake after a meal, and muscle strength and activity are associated with better metabolic outcomes. A review of 16 prospective studies found that muscle-strengthening activities were associated with 10 to 17% lower risk of death from any cause, cardiovascular disease, cancer and diabetes, with the largest reduction in diabetes risk at up to 60 minutes a week [10]. In the PURE study of 139,691 adults in 17 countries, lower grip strength predicted death from any cause and from cardiovascular disease more strongly than systolic blood pressure did [11]. Short-term trials show that breaking up prolonged sitting also changes glucose handling: in a pooled analysis of seven one-day crossover trials, frequent short bouts of light walking lowered the rise in glucose and insulin after meals compared with continuous sitting, and did so more than standing breaks [12]. Related evidence is described under sedentary time and walking and daily steps.

Diet affects metabolic health mainly through energy intake and food quality. In an inpatient trial at the US National Institutes of Health, 20 adults ate 508 more calories a day on an ultra-processed diet than on an unprocessed diet matched for presented calories, sugar, fat and fibre, and gained 0.9 kg in two weeks [13]. In the French NutriNet-Santé study of 105,159 adults, each 10-point higher share of ultra-processed food in the diet was associated with 12% higher risk of cardiovascular disease [14]. A series of reviews partly funded by the World Health Organization, covering 185 prospective studies and 58 trials, found 15 to 30% lower mortality and lower incidence of type 2 diabetes, heart disease and stroke in people with the highest fibre intake, with the largest risk reductions at 25 to 29 g a day [15]. Overall diets are described under dietary patterns and the timing of meals under meal timing.

Full deep dive: Metabolic health.

Type 2 diabetes accounts for most cases of diabetes in adults. Persistently high glucose damages small and large blood vessels, and diabetes is the leading cause of kidney failure and of new cases of blindness among US adults; it is also associated with cardiovascular disease and fatty liver disease, and was estimated to be the seventh leading cause of death in the United States in 2017 [1]. Prevalence differs by ancestry: in US data it was 14.7% in American Indian and Alaska Native adults, 12.5% in Hispanic adults, 11.7% in non-Hispanic Black adults, 9.2% in Asian adults and 7.5% in non-Hispanic White adults [1].

The Emerging Risk Factors Collaboration analysed 123,205 deaths among 820,900 people. After adjustment for age, sex, smoking and BMI, diabetes was associated with a hazard ratio of 1.80 (95% CI 1.71 to 1.90) for death from any cause, 2.32 for death from vascular causes and 1.25 for death from cancer, including cancers of the liver, pancreas, ovary, colon and rectum, lung, bladder and breast. Diabetes was also linked to deaths from kidney disease, liver disease, pneumonia and other infections, and self-harm. Fasting glucose above 100 mg/dL (5.6 mmol/L) was associated with higher mortality, while levels of 70 to 100 mg/dL were not. About 40% of the 6-year difference in survival at age 50 came from deaths not caused by vascular disease [2].

Body fat is the strongest modifiable risk factor. In 18 prospective studies, adults with obesity had 7.19 times and overweight adults 2.99 times the risk of diabetes of normal-weight adults [16], and each 5-unit higher BMI was associated with 72% higher risk in a later review [8]. Family history is also strong. In the Framingham Offspring Study, diabetes in the mother or in the father each raised the odds of type 2 diabetes about 3.5-fold and diabetes in both parents about 6-fold [5]. In EPIC-InterAct, a family history of diabetes was associated with a hazard ratio of 2.72, which fell only to 2.44 after adjustment for BMI and waist size, and a genetic score of 35 known variants explained only 2% of the excess risk [17]. Diabetes during pregnancy (gestational diabetes) is one of the strongest single predictors [4].

Sleep and smoking are also associated with risk. A review of 482,502 participants found a U-shaped relationship between sleep duration and type 2 diabetes, with the lowest risk at 7 to 8 hours a night; risk rose 9% for each hour below 7 hours and 14% for each hour above [18]. An earlier review reported 1.57 times the risk in people with difficulty falling asleep and 1.84 times with difficulty staying asleep [19]. Across 88 studies of 5.9 million people, current smoking was associated with 37% higher risk, rising from 21% in light smokers to 57% in heavy smokers; risk was 54% higher than in never smokers in the first 5 years after quitting and 11% higher after 10 or more years [6]. In three US cohorts, the temporary rise in diabetes risk after quitting was proportional to weight gain and did not offset the lower cardiovascular and all-cause mortality of quitters [20]. Smoking itself is described under smoking.

Full deep dive: Type 2 diabetes risk factors.

Prediabetes and type 2 diabetes are detected with blood tests. HbA1c (glycated haemoglobin, the share of haemoglobin in red blood cells with glucose attached) measures long-term blood glucose and is not affected by short-term changes caused by stress or illness. Under US criteria, an HbA1c of 6.5% or higher, a fasting plasma glucose of 126 mg/dL or higher, or a glucose of 200 mg/dL or higher two hours after a 75 g glucose drink is consistent with diabetes, and a diagnosis should be confirmed by repeat testing. An HbA1c of 5.7 to 6.4%, a fasting glucose of 100 to 125 mg/dL or a two-hour glucose of 140 to 199 mg/dL is consistent with prediabetes [1].

The US Preventive Services Task Force recommends screening adults aged 35 to 70 who have overweight (BMI 25 or more) or obesity (BMI 30 or more). It advises considering earlier screening in groups with a high prevalence of diabetes, and screening Asian American adults from a BMI of 23, because diabetes occurs at lower body weights in this group. For adults with a normal result, repeating the test every 3 years was judged a reasonable interval [1].

The Diabetes Prevention Program randomly assigned 3,234 US adults with raised fasting and after-meal glucose (mean age 51, mean BMI 34.0) to placebo, metformin or a lifestyle programme with goals of at least 7% weight loss and at least 150 minutes of physical activity a week. Over an average of 2.8 years, new diabetes occurred at 11.0, 7.8 and 4.8 cases per 100 person-years in the three groups. The lifestyle programme reduced incidence by 58% (95% CI 48 to 66%) and metformin by 31%, and about 7 people needed to take part in the lifestyle programme for three years to prevent one case [3].

The longest follow-up comes from the Da Qing Diabetes Prevention Study in China, which compared a 6-year lifestyle intervention of diet, exercise or both with no intervention. At 30 years, all-cause mortality was 45.7% in the combined intervention groups and 56.3% in the control group (hazard ratio 0.74, 95% CI 0.61 to 0.89), although the difference was not seen at 20 years and the groups differed at baseline [1].

Most evidence on individual risk factors is observational (it comes from tracking people rather than assigning them to an exposure), so the associations may partly reflect other differences between groups. The Framingham Offspring Study drew mainly on white participants at relatively low risk of diabetes [5], and the effect of family history is not explained by measured body size, lifestyle or known genetic variants [17]. Both short and long sleep were associated with higher risk, a pattern that makes the direction of cause and effect difficult to establish [18, 19].

The prevention trials measured new diagnoses of diabetes over a few years rather than death or life expectancy [3]. Evidence that preventing diabetes lowers mortality comes mainly from the long follow-up of the Da Qing study, which was limited by baseline differences between groups [1].