Long-term lifestyle intervention lowers the incidence of stroke in Japanese patients with type 2 diabetes: a nationwide multicentre randomised controlled trial (the Japan Diabetes Complications Study)

Participants in the study were previously diagnosed patients with type 2 diabetes aged 40–70 years whose HbA _1c levels were ≥6.5%. From outpatient clinics in 59 university and general hospitals nationwide that specialise in diabetes care, a total of 2,205 patients (mean age 58.6 years; 47% women) were initially registered from January 1995 to March 1996. Excluded were patients with a history of angina pectoris, myocardial infarction, stroke, peripheral arterial disease, familial hypercholesterolaemia, type III hyperlipidaemia, non-diabetic nephropathy, nephrotic syndrome, pre-proliferative and proliferative retinopathy, intra-ocular surgeries, serum creatinine levels >120 μmol/l, and mean values of two spot urine examinations for an albumin excretion rate of <150 mg/g creatinine. Diabetes mellitus and IGT were diagnosed according to the Report of the Committee of the Japan Diabetes Society on the Classification and Diagnostic Criteria of Diabetes Mellitus, which is almost identical in terms of cut-off values for glucose levels to those of the WHO. The protocol for the study, which is in accordance with the Declaration of Helsinki and the Ethical Guidelines for Clinical/Epidemiological Studies of the Japanese Ministry of Health, Labour and Welfare, received ethical approval from the institutional review boards of all of the participating institutes (RCT registration number was C000000222 in www.​umin.​ac.​jp). Written informed consent was obtained from all patients enrolled.

Before April 1996, when the intervention began, patients who did not meet the eligibility criteria were excluded. Finally, a total of 2,033 patients aged 58.5 ± 6.9 years and who had diabetes for a duration of 10.9 ± 7.2 years (both mean ± SD) were included from the present analysis. Figure  1 is a flow diagram of the JDCS. Patients were allocated randomly into either a lifestyle intervention (INT) group or a conventional treatment (CON) group. Randomisation and all analyses were done by a central computer at our database centre. This study was open-labelled and the interventions for the INT group were continued until March 2003.

As basal therapeutic management of all patients in both the CON and INT groups, regular specialists’ care was provided throughout the study period and patients were treated as they were before the study started. This included dietary advice by an administrative dietitian, using the ‘Food Exchange Lists Dietary Guidance for Persons with Diabetes’ [ 24].

In addition to this routine conventional treatment, education of patients in the INT group was given through individual counselling on dietary habits, physical activities and adherence to treatment, including taking medicine properly. Counselling was provided by physicians, nurses, dietitians and other co-medical staff during each outpatient clinic visit. Patients in the INT group had a typically 5–10 min longer interview than the patients in the CON group at each clinic visit for a discussion on possible causes of any changes in HbA _1c levels, weight and other control variables from the previous visit, with emphasis on lifestyle changes. For example, when it was revealed that control of glycaemic and other variables had worsened, that dietary intake, including quantity and content, and alcohol intake had changed, that patterns of physical activity had changed or that patients tended to forget to take their medicine, possible strategies for improving lifestyle and habits were discussed. Furthermore, patients in the INT group also received additional advice regarding one or two particular topic(s) at each visit and were given educational material consisting of 23 brochures that discussed various aspects of diabetes care with an emphasis on the importance of lifestyle and behavioural changes such as ‘Why am I obese even if I do not eat so much?’, ‘Tips for continuing exercise’, ‘What kinds of stress affect the control of diabetes’ or ‘Is your triglyceride level OK?’.

Patients in the INT group also received 15 min telephone counselling sessions at least once every 2 weeks by nurses, dietitians and psychotherapists who were trained in diabetes education. These telephone sessions were performed based on a structured and uniform format. Additional counselling sessions were encouraged at any convenient time, depending on the needs of patients in the INT group. A diary to record the progress of laboratory and other data was distributed to the INT patients to provide better feedback on therapeutic results. A pedometer was also distributed to INT patients for objective exercise assessment.

Goals were set for patients in the INT group and their physicians: i.e. HbA _1c level <6.5%; BMI <22 kg/m ²; BP <140/85 mmHg; serum cholesterol level <5.72 mmol/l; serum triacylglycerol level <1.65 mmol/l; serum HDL-cholesterol >1.04 mmol/l; WHR <0.9 for men and <0.8 for women; smoking cessation; and abstinence from alcohol. Goals regarding BP and serum cholesterol levels were updated in accord with the revision of guidelines made by the Japan Diabetes Society, which were <130/80 mmHg and <5.17 mmol/l, respectively.

During the study period, patients in the INT group with poor control (HbA _1c >8.0%, serum total cholesterol level >5.72 mmol/l, serum triacylglycerol level >1.65 mmol/l or BMI >22 kg/m ²) were identified and sent additional educational material. At the same time, their physicians were encouraged to strengthen the intervention through increasing time for education by telephone or at clinic visits or recommending hospital admissions for education. Changes in medication including insulin and oral antihypertension/dyslipidaemia agents were not restricted in either group and were made based on therapeutic necessity.

Extensive lifestyle surveys were performed at baseline and 5 years after the start of the intervention in both groups. We used detailed questionnaires for patients to determine dietary (including alcohol drinking) content, amount of exercise and smoking rate. The dietary survey comprised food records and a food frequency questionnaire, results of which were analysed by an administrative dietitian using standardised software for population-based surveys and nutrition counselling in Japan (EIYO-KUN v.4.5, manufactured at Shikoku University Nutrition Database) [ 25] based on the Standard Tables of Food Composition in Japan [ 26] edited by the Japanese Ministry of Education, Culture, Sports, Science and Technology. Physical activity was determined by a questionnaire that inquired about types of exercise (walking, jogging, tennis, etc) and average time (min) spent exercising per day at baseline and the Baecke’s Total Physical Activity Index [ 27] at the fifth year.

Mean values of at least two measurements each year were obtained for the following variables: HbA _1c, fasting plasma glucose/insulin/C-peptide, serum lipids/creatinine/urea nitrogen, and urine analysis. All other measurements including those for body weight, BP and WHR and a neurological/ophthalmological examination were done at least once yearly, with a mean value obtained if measurements were done twice a year. HbA _1c assays were standardised by the Laboratory Test Committee of the Japan Diabetes Society, with 5.8% as the upper normal limit. All other laboratory tests were done by standard methods in each clinic. Electrocardiograms and chest x-rays were performed annually. Patients were assessed yearly after the baseline evaluation.

Primary outcomes of the study were development and progression of microangiopathy and occurrence of macrovascular complication events. Microangiopathy endpoints consisted of those related to retinopathies and nephropathies. Retinopathy was evaluated by qualified ophthalmologists using the following classification designed for this research: Stage 0, no retinopathy; Stage 1, haemorrhage and hard exudates; Stage 2, soft exudates; Stage 3, intraretinal microvascular abnormalities and venous changes including beading, loop and duplication; Stage 4, new vessels, vitreous haemorrhage, fibrous proliferation and retinal detachment. The retinopathy endpoint was (1) development of retinopathy (from Stage 0 to any other stage confirmed in two continuous years), and (2) progression from Stage 1 to Stage 3 or 4. The nephropathy endpoint was defined as development of overt nephropathy (spot urinary albumin excretion >300 mg/g creatinine in two consecutive samples).

Macroangiopathy endpoints included the occurrence of definite CHD (angina pectoris or myocardial infarction) or stroke. Diagnosis of angina pectoris and myocardial infarction was according to criteria defined by the WHO/MONICA (Multinational Monitoring of Trends and Determinants in Cardiovascular Disease) project and diagnosis of stroke was according to guidelines defined by the Ministry of Health, Labour and Welfare of Japan [ 21]. Information regarding primary outcome and other clinical variables for each patient was collected through an annual report from each physician. Adjudication of endpoints was made by central committees comprising experts in each complication based on additional data such as CT or MRI of the brain or sequential changes in electrocardiograms.

The sample size required to compare the net change in HbA _1c level (level at the third year point minus the baseline level) between the INT and CON groups is based on a consideration of power. It is assumed that the type I error (α) is 0.05 and the type II error ( β) is 0.1; therefore, in order to detect a difference of 0.2 in net change in HbA _1c level, with an SD of 1.0 among patients, a total sample size of 1,025 is required. In addition, if allowance is made for an up to 20% dropout rate, for 20% of the INT group being unable to complete the intervention and for 10% of the CON group to change treatment method within the follow-up period, the required sample size increases to 2,616. Thus, in terms of the feasibility of the study, it was necessary to recruit more than 2,000 patients. All statistical analyses and data management were conducted at a central data centre. The Wilcoxon’s rank sum test, Fisher’s exact test and Mantel’s test were used for comparison of numerical and categorical variables between groups.

The study endpoints were analysed as time-to-event variables, i.e. clinical data on patients who were lost during follow-up were used for the period for which they could be followed. Survival curves for the diabetic complications were estimated by the Kaplan–Meier method, and the logrank test was also conducted. Cox regression analysis was used to calculate the unadjusted and adjusted HRs and 95% CIs for group and risk factors. In multivariate Cox analysis, all significant variables selected for the univariate analysis were used with the criterion of p < 0.1. All values are presented as means ± SD unless otherwise stated. All p values are two-sided and the significance level is 0.05. All statistical analyses were conducted using SAS packages ver. 9.1 (SAS Institute, Cary, NC, USA).