• Newly presenting patients aged 25 to 65 years
• Fasting plasma glucose of greater than 6.0mmol per L on two occasions.

• Ketonuria
• History of myocardial infarction in the previous year
• Angina
• Heart failure
• More than one major vascular episode
• Serum creatinine greater than 175mcmol per L
• Severe retinopathy requiring photocoagulation
• Accelerated hypertension
• Uncorrected endocrine abnormality
• Occupation that would not allow randomization to insulin therapy (e.g., heavy goods vehicle driver)
• Severe pre-existing illness likely to limit life (e.g., cancer) or requiring extensive systemic treatment (e.g., ulcerative colitis)
• Inadequate comprehension to allow cooperation
• If reasonable, diuretics and estrogen therapy were stopped before assessment.

• Recruitment: In 15 centers, newly presenting patients meeting inclusion criteria were referred for inclusion in the study
• Design: Randomized controlled trial
• Blinding used: Implied with measurements.

Intervention

• All patients were advised to follow the British Diabetic Association recommendations of a diet containing 50% carbohydrate, 30% fat and 20% protein, with energy restricted according to the patient's obesity and activity
• The mean prescribed diet was 1,361kcal:
  • Mean 1,672kcal for those patients less than 110% ideal body weight
  • Mean 1,361kcal for patients 110% to 129% ideal body weight
  • Mean 1,217kcal for patients 130% to 149% ideal body weight
  • Mean 1,098kcal for patients greater than 150% ideal body weight.
• After three months, if the mean fasting glucose was greater than 6.0mmol per L, the patient was entered into the mean randomization and allocated to either sulphonylurea, metformin or insulin therapy or to continue on diet alone. If the mean fasting glucose was less than 6.0mmol per L, patients were regarded as diet satisfactory and remained on diet alone. If the mean fasting glucose remained greater than 15.0mmol per L or they had symptoms of hyperglycaemia such as polydipsia, polyuria or excessive tiredness, the patients were regarded as primary diet failures and were randomly allocated to sulphonylurea, metformin or insulin therapy.
• Centers had different availability of dietetic advice. Before the results were available, three centers had been categorized as above average in this respect, as they had additional, specific dietetic personnel for the patients in the study, 10 centers as average with advice from routine National Health Service dietitians and two centers below average, sometimes having lack of continuity of care from a single dietitian.

Statistical Analysis

• Weight loss achieved over three months and the decrease in fasting plasma glucose was assessed for each center
• Data analysis was performed using SPSS
• Regression analyses were performed after checking that the data was normally distributed.

Timing of Measurements

• At each clinic visit, a fasting plasma glucose was taken and the patient weighed
• Patients were seen at one-month intervals for three months.

Dependent Variables

• Fasting plasma glucose measured by glucose-oxidase method
• Weight.

Independent Variables

• All patients were advised to follow the British Diabetic Association recommendations of a diet containing 50% carbohydrate, 30% fat and 20% protein with energy restricted according to the patient's obesity and activity.
• The mean prescribed diet was 1,361kcal:
  • Mean 1,672kcal for those patients less than 110% ideal body weight
  • Mean 1,361kcal for patients 110-129% ideal body weight
  • Mean 1,217kcal for patients 130-149% ideal body weight
  • Mean 1,098kcal for patients greater than 150% ideal body weight.

Initial N

• 3,044 patients
• 59% male
• 41% female.

Attrition (Final N)

• 2,597 patients who were on a diet for three months before therapy decision
• 447 (15%) primary diet failures were excluded
• The change in fasting plasma glucose and body weight in 823 patients who continued on diet alone for a further year after therapy decision were assessed, which included the 482 diet satisfactory patients who had a fasting plasma glucose less than 6.0mmol per L and were continued on diet and the 341 patients who had a value greater than 6.0mmol per L and were allocated to diet in the main randomization.

Age
Five to 65 years; mean, 52±8 years.

Ethnicity
Not reported.

Other Relevant Demographics
Not reported.

Anthropometrics
Mean body weight, 130±26% ideal body weight.

Location
United Kingdom.

Key Findings

Other Findings

• Weight decreased from 130±26% to 123±24% of ideal body weight and fasting plasma glucose decreased from 11.4±3.3mmol to 8.1±1.8mmol per L
• The body weight decrease was most in the first month and in the heaviest patients, who were usually prescribed the most restricted diets
• The initial fasting plasma glucose had no effect on the degree of weight loss
• The decrease in fasting plasma glucose was also greatest in the first month after presentation, with a continued reduction over the next two months
• The reduction in the fasting plasma glucose was greater in those presenting with high initial fasting plasma glucose (R=0.76, P<0.001) and in those who lost more weight (partial correlation coefficient R=0.19, P<0.001), but was not related to the degree of obesity
• By stepwise regression analysis, 31% of the variation of the decrease in fasting plasma glucose over the first three months was accounted for by the initial fasting plasma glucose, a further 8% by the weight loss over three months, with no contribution from the initial weight
• Weight loss in patients in the individual centers was proportional to the degree of availability of dietetic advice and was also associated with the degree to which the fasting plasma glucose decreased over three months
• There was considerable variation in the response to dieting, but on average, patients presenting with a fasting plasma glucose of 10mmol to 12mmol per L needed to lose 28% ideal body weight (18kg) to attain a fasting plasma glucose less than 6.0mmol per L
• 16% of all patients achieved a near-normal (less than 6.0mmol per L) fasting plasma glucose after three months of dieting, ranging from 50% of those presenting with fasting plasma glucose of 6.0mmol to 8.0mmol per L to 10% of those with fasting plasma glucose of 16mmol to 22mmol per L
• In those who achieved less than 6.0mmol per L, in the second three months the fasting plasma glucose increased by a mean of 0.4mmol per L even though there was a further mean weight reduction of 2.1% ideal body weight (1.4kg) in addition to their loss of 11.6% ideal body weight (8 kg) in the initial three months, confirming the decrease in fasting plasma glucose is determined more by the restriction of energy intake than by the body weight
• Those who maintained their fasting plasma glucose at less than 6.0mmol per L in the year following the initial three-month dietary period lost a further 3% (2.0kg) ideal body weight
• On average, to achieve a fasting plasma glucose of less than 6.0mmol per L, a weight loss of 16% ideal body weight was needed if the initial fasting plasma glucose was in the range of 6.0 to 8.0mmol per L, a weight loss of 21% ideal body weight for an initial fasting plasma glucose of 8.0mmol to 10mmol per L, 28% ideal body weight loss for 10mmol to 12mmol per L, 35% ideal body weight loss for 12mmol to 14mmol per L and 41% ideal body weight loss for an initial fasting plasma glucose over 14.0mmol per L
• 14% of all patients who presented maintained a fasting plasma glucose less than 6.0 mmol per L after the subsequent year's diet (15 months of dieting in all). On average, they had slightly lower fasting plasma glucose at presentation and had lost more weight than other patients. Of the main randomization patients who had a fasting plasma glucose more than 6.0mmol per L after three months on diet, only 8% were able to maintain less than 6.0mmol per L after a subsequent year's diet. They were patients who had fasting plasma glucose values just above 6.0mmol per L after the three months' dieting and then continued to lose, on average, a further 4% ideal body weight.

• The degree of obesity on presentation made little difference to the fasting plasma glucose response to diet, which was greater in those patients presenting with high fasting plasma glucose and in those who lost most weight on dieting
• This study shows that considerable dietary restriction and weight loss is needed in most patients if the fasting plasma glucose is to be reduced to less than 6.0mmol per L
• The data confirm the value of dieting, but in view of the large weight loss and equivalent large reduction in energy intake required in most patients, it is not surprising that few patients achieve near-normal fasting plasma glucose concentrations by diet alone.