Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Lifestyle Interventions to Prevent Diabetes

Systematic review
Results
Plasma glucose
Diabetes incidence
Comment

Bandolier readers are familiar with the associations of weight, sedentary lifestyle, and poor diet with type 2 diabetes in older adults, and increasingly with younger adults and even children. Most readers would instantly prescribe diet and exercise, but would probably not have the numbers immediately to hand to press home any arguments about whether diet and exercise worked or not. A new systematic review and meta-analysis [1] provides them.

Systematic review


The review set out to answer the question whether a lifestyle education programme compared with conventional education improved glucose levels or incidence of diabetes in people at high risk of developing type 2 diabetes. Two outcome measures were sought. The change in glucose values two hours after a 75 g oral glucose load at baseline, and more than six months later. The other outcome was diabetes incidence. Subjects had to have impaired glucose tolerance, impaired fasting glucose, or borderline values, according to standard criteria.

Trials had to be randomised, and searching was limited to English language papers published to November 2004.

Results


Nine studies were included, eight with information on plasma glucose, and five on diabetes incidence. Seven involved diet and exercise, and two diet alone; controls were generally conventional counselling or advice. Trials enrolled 80 to 3,000 subjects, with follow up from six months to six years; five larger trials had follow up of three years or longer.

The average age of patients in the trials ranged from 39 to 57 years, and the average BMI from 24 to 31 kg/sq m. Baseline two hour plasma glucose after oral load averaged 8-10 mmol/L in individual studies.

Plasma glucose


The prime end point was the difference between the two hour plasma glucose from baseline to one year or more. Results for individual studies are in Figure 1. Overall, there was a reduction in two hour plasma glucose after an oral load by 0.8 mmol/L (95% confidence interval 0.4 to 1.3 mmol/L).


Figure 1: Change in plasma glucose at 2 hours after oral load from baseline to one year






Diabetes incidence


The incidence of diabetes reported in five studies lasting one to six years is shown in Figure 2. The weighted mean duration was 3.6 years. The incidence of diabetes was 28% with control (7.8%, or 1 in 13 per year), but only 15% with diet and exercise (4.2%, or 1 in 24 per year). The relative risk was 0.6 (0.5 to 0.7), and the number needed to treat was 7.4 (6.2 to 9.4) to prevent one person developing type 2 diabetes.


Figure 2: Incidence of diabetes over one to six years (mean 3.6 years)






Comment


This is good evidence, with consistent reductions both in glucose levels and in diabetes incidence over reasonably long periods of time. The reduction of diabetes incidence by half may even understate the case. While the average duration was commendably long, at 3.6 years, continued use of a healthy lifestyle should continue to generate benefits. A 10-year NNT can be calculated at 2.6, for instance.

Put another way, if people in the studies continued to develop diabetes at the same rate, after 10 years 8 out of 10 in the control group would have developed diabetes, compared with only 4 out of 10 with the intervention. And of course, we have no information about adherence, so we only know about the effects of treatment on average. It is likely that individuals who really changed their lifestyle and stuck to it could reduce their individual risk of developing diabetes to much lower levels. The message is, yet again, that the best way of avoiding seeing your doctor, except socially, is to eat fruit and vegetables, have a good walk, and perhaps enjoy a nice glass of wine when you get back.

Reference:


  1. K Yamaoka, T Tango. Efficacy of lifestyle education to prevent type 2 diabetes. A meta-analysis of randomized controlled trials. Diabetes Care 2005 28: 2780-2786.

previous or next story