Physical Activity, Body Mass Index, and Risk of Type 2 Diabetes in Patients With Normal or Impaired Glucose Regulation FREE

Epidemiological evidence suggests that higher levels of leisure time physical activity are associated with a reduced risk of type 2 diabetes.^1- 11 Our previous study also reported that moderate and high occupational physical activity, as well as daily walking or cycling to and from work for more than 30 minutes, independently and significantly reduced risk of type 2 diabetes.¹¹ Evidence from several studies indicates that obesity and weight gain are associated with an increased risk of type 2 diabetes.^12- 14 A number of studies have concluded that individuals with impaired glucose regulation are at markedly higher risk for development of type 2 diabetes than individuals with normal glucose levels.^15- 18 Recent clinical trials in China, Finland, and the United States have demonstrated that lifestyle intervention (dietary modification and enhanced physical activity) reduced the risk of progressing from impaired glucose tolerance to type 2 diabetes.^19- 21 These studies, however, did not examine the joint association between physical activity, body mass index (BMI), glucose levels, and the risk of type 2 diabetes in the general population. The aim of this study was to examine the single and joint association of physical activity, BMI, and glucose levels with risk of type 2 diabetes in a prospective cohort study.

We carried out baseline surveys in 2 eastern Finnish provinces, North Karelia and Kuopio, and in the Turku-Loimaa region of southwestern Finland in 1987 and 1992. The survey was expanded to the Helsinki capital area in 1992. In the 2 surveys, the sample included subjects aged 45 to 64 years. The 1987 and 1992 cohorts were combined in this analysis. The original random sample was stratified from the national population. The sampling schemes and survey procedures have been described in detail elsewhere.²² The surveys were conducted according to the ethical rules of the National Public Health Institute.

A self-administered questionnaire was mailed to the participants. It included questions on medical history, socioeconomic factors, physical activity, and smoking habits. Education level, measured as the total number of school years, was divided into birth cohort–specific tertiles. The participants were classified into 2 smoking categories, current smokers and nonsmokers.

Physical activity included occupational, commuting, and leisure time physical activity. A detailed description of the questions is presented elsewhere,^23- 24 and the way the questions were constructed was evaluated previously.²⁵ The subjects reported their occupational physical activity as "light" when it was physically very easy (eg, the desk work of a secretary), or as "moderate or active" when it included standing, walking, walking and lifting, or heavy manual labor (eg, the work of a store assistant or of an industrial or farm worker). The subjects were asked whether they walked, rode a bicycle, or used motorized transportation to and from work and about the daily duration of this activity. The daily commuting journey was divided into 2 categories: using motorized transportation or walking and/or bicycling less than 30 minutes, and walking and/or bicycling 30 minutes or longer. Self-reported leisure time physical activity was divided into 2 categories: "low" when subjects were almost completely inactive (eg, they read, watched TV, or engaged in a physical activity not of moderate or high level), and "moderate or high" when they engaged more than 4 hours per week in a physical activity such as walking, bicycling, light gardening, fishing, or hunting (but excluding travel to work), or performed vigorous physical activity more than 3 hours per week such as running, jogging, skiing, swimming, any types of team ball games, heavy gardening, or regular exercise such as competitive sports several times per week. Since our previous study has shown that moderate and high occupational, commuting, or leisure time physical activity independently and significantly reduces risk of type 2 diabetes,¹¹ physical activity categories were merged and regrouped into 3 categories: "low" when subjects reported light levels of occupational, commuting (<30 minutes), and leisure time physical activity; "moderate" when subjects reported only 1 of the 3 types of "moderate to high" physical activity; and "high" when subjects reported 2 or 3 types of "moderate to high" physical activity.

Physical measurements including height, weight, and blood pressure were taken, and a 2-hour 75-g oral glucose tolerance test was taken for measurement of glucose levels by specially trained nurses. Blood pressure was measured with a standard sphygmomanometer from the right arm of the participants who were seated for 5 minutes before the measurement. Height was measured to the nearest 0.5 cm and weight (with light indoor clothing and no shoes) with a precision of 0.1 kg. Body mass index was calculated as weight in kilograms divided by the square of the height in meters, and the subjects were defined as obese if their BMI was 30.0 or greater.

Subsamples of 3404 subjects in 1987 and 2540 subjects in 1992 were invited to receive the standard oral glucose tolerance test at baseline. The participation rate was 81% in 1987 and 76% in 1992. Whole blood samples were collected after subjects fasted for at least 4 hours in 1987, and plasma samples were obtained after an overnight fast in 1992. Then, the 2-hour oral glucose tolerance test was administered in both 1987 and 1992. Blood samples were mailed to a central laboratory where glucose concentration was determined by the hexokinase method.

Whole blood glucose concentration was converted to plasma glucose concentration before data analysis using the following equation: plasma glucose (millimoles per liter) = 0.558 + 1.119 × whole blood glucose (millimoles per liter). The equation is based on 294 paired samples of whole blood and plasma glucose concentrations drawn in the Diabetes and Genetic Epidemiology Unit, National Public Health Institute, Helsinki, Finland, from a standard 75-g oral glucose tolerance test in 74 individuals at 0, 30, 60, and 120 minutes. The relationships between glucose concentrations as measured by the different methods were estimated using the mixed model with random effects of individual and sample developed at the Steno Diabetes Centre in Denmark (this model was personally communicated to J.T. by B. Carstensen, PhD, in 2002).

Diabetes at baseline was defined either as previously diagnosed, according to a history of antidiabetic medication,²⁶ or as undiagnosed, according to a fasting plasma glucose level of 126 mg/dL (7.0 mmol/L) or greater and/or a 2-hour plasma glucose level of 200 mg/dL (11.1 mmol/L) or greater.²⁷ The final sample consisted of 2017 men and 2352 women and excluded subjects with previously diagnosed (n = 204) and newly diagnosed (n = 154) diabetes at baseline. Classification of impaired glucose regulation was made based on a fasting plasma glucose level of 110 to 126 mg/dL (6.1-6.99 mmol/L) and/or a 2-hour plasma glucose level of 140 to 299 mg/dL (7.8-11.09 mmol/L). Normal glucose level was defined as a fasting plasma glucose level less than 110 mg/dL (6.1 mmol/L) and a 2-hour plasma glucose level less than 140 mg/dL (7.8 mmol/L).

We ascertained incident cases of diabetes from the National Hospital Discharge Register and the National Social Insurance Institution's Register. The National Hospital Discharge Register data were linked to the risk factor survey data with the personal identification numbers assigned to every resident of Finland. Antidiabetic drugs prescribed by a physician are free of charge in Finland subject to approval by a physician who reviews each case history. The physician confirms the diagnosis of diabetes on the basis of the World Health Organization criteria. All patients receiving free medication (either oral antidiabetic agents or insulin) are entered into a register maintained by the Social Insurance Institution. The National Hospital Discharge Register has reported the codes for type 1 and type 2 diabetes in Finland since 1987. Follow-up of each participant in our present analysis continued through the end of 1998 or until death. All new cases identified during the follow-up of this cohort were type 2 diabetes.

Statistical package SPSS for Windows, version 11.0 (SPSS Inc, Chicago, Ill), was used for statistical analysis. Differences in risk factors between normal and impaired glucose regulation were tested using univariate analysis of variance or logistic regression analysis after adjustment for age, sex, and study year.

The Cox proportional hazards model was used to estimate the single or joint effect of different levels of physical activity, BMI, and plasma glucose on the risk of type 2 diabetes incidence. All analyses were adjusted for the following covariates: age, sex, study year, education, systolic blood pressure, and smoking.

There were 120 cases of type 2 diabetes identified during a mean follow-up period of 9.4 years. The age-, sex- and study year–adjusted major risk factors of subjects with normal glucose levels and impaired glucose regulation are shown in Table 1. In general, subjects with impaired glucose regulation were older, had a lower education level, and had significantly higher mean values of BMI, blood pressure, and prevalence of obesity than subjects with normal glucose levels.

Table 2 shows the single relative risk of type 2 diabetes for different levels of physical activity, BMI, and plasma glucose. Multivariate relative risks of type 2 diabetes adjusted for age, sex, study year, systolic blood pressure, smoking, and education decreased significantly with increasing physical activity. The hazard ratios for risk of diabetes associated with light, moderate, and high physical activity were 1.00, 0.85, and 0.43, respectively (P = .004 for trend) after further adjustment for BMI. Multivariate-adjusted relative risks of type 2 diabetes based on 3 ranges of BMI values (<25, 25-29.9, and ≥30) were 1.00, 1.79, and 6.25 (P<.001 for trend). This positive association was still significant after further adjustment for physical activity (P<.001 for trend). Subjects with impaired glucose regulation showed approximatively a 5-fold risk of developing type 2 diabetes than subjects with normal glucose level after adjustment for all covariates.

The inverse association between physical activity and risk of type 2 diabetes was persistent in subgroup analyses based on BMI (<30 and ≥30) and glucose levels (normal glucose level and impaired glucose regulation) (Figure 1A and B). The positive association between BMI and risk of type 2 diabetes was confirmed in subjects with either normal glucose levels or impaired glucose regulation (Figure 1C).

The joint association of different levels of physical activity, BMI, and plasma glucose with the risk of type 2 diabetes is shown in Figure 2. The positive association of BMI and the protective effect of physical activity were found among both normoglycemic subjects and subjects with impaired glucose regulation. In comparison with nonobese subjects who reported a high level of physical activity and had a normal glucose level, obese subjects who reported a low level of physical activity and had impaired glucose regulation showed a 30-fold risk of developing type 2 diabetes. The relative effect of obesity and physical activity in diabetes risk was larger in subjects with normal glucose regulation, but the absolute effect was larger among subjects with impaired glucose regulation.

Our analysis addressed the joint association of physical activity, BMI, and glucose levels with regard to the risk of type 2 diabetes. Increasing physical activity was associated with a significantly reduced risk for type 2 diabetes, and this inverse association was consistent in subjects with any joint levels of BMI and glucose. The positive association between BMI and the risk of type 2 diabetes was similar in subjects with different levels of physical activity and glucose. Furthermore, subjects who were inactive, obese, and had a normal glucose level were associated with increased risk of type 2 diabetes compared with subjects who were physically active, nonobese, and had impaired glucose regulation.

A number of prospective studies among specific populations of college alumni,¹ nurses,^2,8 and male physicians^4,10 in the United States or who resided in England^6,9 and Finland^7,11 have indicated that regular leisure time physical activity is associated with decreased risk of type 2 diabetes. Two of these studies among American nurses and male physicians also found that the inverse association between walking and the risk of type 2 diabetes was similar to that between vigorous leisure time physical activity and type 2 diabetes.^8,10 Our previous study reported that moderate and high occupational physical activity, and daily walking or cycling to and from work for more than 30 minutes, independently and significantly reduced risk of type 2 diabetes.¹¹ The present study indicated an inverse association between physical activity and risk of type 2 diabetes not only in nonobese subjects with normal glucose regulation, but also in subjects with either obesity or impaired glucose regulation, or in subjects with both obesity and impaired glucose regulation. An exercise program should be prescribed or monitored by a qualified health care professional.

Overweight, obesity, or weight gain has been shown to be an important risk factor for development of type 2 diabetes.^12- 14 Overweight or obese persons are less active than people who have a normal weight. We found that obese and inactive subjects with normal glucose levels had an increased risk of type 2 diabetes compared with nonobese and active subjects with impaired glucose regulation. Obesity and inactivity might be major risk factors for development of type 2 diabetes not only in persons with impaired glucose regulation but also in persons with normal glucose levels. Recent clinical trials have demonstrated that changes in lifestyle are effective in preventing both diabetes and obesity in high-risk adults with impaired glucose tolerance.^19- 21 Increasing physical activity and improving dietary habits might reduce the risk of type 2 diabetes indirectly through decreased body weight or improved body fat distribution. Physical activity facilitates weight loss and weight maintenance.²⁹ Furthermore, physical training has been shown to influence visceral adiposity more than BMI.³⁰ The metabolic syndrome, which plays an important role in the development of type 2 diabetes,^31- 32 describes a cluster of metabolic abnormalities (insulin resistance, hypertension, dyslipidemia, and impaired glucose regulation) associated with obesity and its central distribution. Abdominal obesity is related to increased release of nonesterified fatty acids from adipose tissue.³³ Increased serum levels of nonesterified fatty acids are probably an integral component of the metabolic syndrome.³³

In accordance with several other studies,^15- 18 impaired glucose regulation seem to be a better predictor of future diabetes. The present study found that the risk of diabetes could be reduced by increased physical activity in persons with or without obesity but with impaired glucose regulation, which is consistent with results from previous clinical trials.^19- 21 Regular physical activity could have an independent effect on glucose disposal by improving insulin sensitivity and reducing other components of the insulin resistance syndrome.^9,34 A recent analysis from the Insulin Resistance Atherosclerosis Study showed that both vigorous and nonvigorous activities were associated with significantly higher insulin sensitivity among 1467 men and women aged 40 to 69 years.³⁴ The mechanism by which physical activity improves insulin sensitivity may involve increased insulin delivery because of improved blood flow to muscle, increased level and translocation of muscle glucose transporter (GLUT-4), and increased glycogen synthase activity, which results in increased glycogen synthesis and glucose disposal.^34- 36 In addition, physical activity also improves insulin sensitivity by reducing the adipose tissue mass.²⁹ The British Regional Study examined the role of serum insulin level and components of the insulin resistance syndrome in the relationship between physical activity and the incidence of type 2 diabetes among 5159 men aged 40 to 59 years.⁹ It showed that physical activity (including at light and moderate levels) was significantly and inversely associated with serum insulin levels and with many of its components, and serum insulin levels and components of the insulin resistance syndrome appeared as a mediating factor in the relationship between physical activity and the incidence of type 2 diabetes.

The strength of our study was that we measured fasting and 2-hour oral glucose tolerance test at baseline, and previously diagnosed and undiagnosed diabetic subjects were excluded from present analyses. In some previous studies, the diagnosed criteria for diabetes using medical and treatment history,^{2,4,7- 8,10- 11} or only fasting glucose,^6,9 or only 1-hour glucose load,⁵ may have included some cases of asymptomatic diabetes at baseline. The misclassification of baseline diabetes would weaken the direct association between physical activity, BMI, and risk of type 2 diabetes. One limitation of our study was the self-report of physical activity. Using a questionnaire to assess habitual physical activity is crude and imprecise. Misclassification is inevitable and usually results in a biased estimate of the association of physical activity and risk of type 2 diabetes.

In conclusion, the present study confirms that increasing physical activity can reduce the risk of type 2 diabetes. This inverse association was consistent in subjects with any joint levels of BMI and glucose. Physical activity and weight control are critical factors in diabetes prevention in persons with either normal or impaired glucose regulation.

Corresponding author: Gang Hu, PhD, Diabetes and Genetic Epidemiology Unit, Department of Epidemiology and Health Promotion, National Public Health Institute, Mannerheimintie 166, FIN-00300 Helsinki, Finland (e-mail: hu.gang@ktl.fi).

Accepted for publication May 30, 2003.

This study was supported by grants 38387, 46558, 52342, 53585, 76502, 77618, 204274, and 205657 from the Finnish Academy.