Level 1: Likely reliable evidence

Impaired glucose tolerance (IGT) has been defined as a two-hour, 75-g oral glucose tolerance test result of 140-199 mg/dL (Diabetes Care. 2003;26:3160-3167; full text available online without charge at: http://care.diabetesjournals.org/cgi/content/full/26/11/3160. Accessed June 12, 2007). In patients with IGT, the effect of interventions on reducing the risk of progression to type 2 diabetes mellitus was assessed in a systematic review of 21 randomized trials (BMJ. 2007;334:299; full text available online without charge at: www.bmj.com/cgi/rapidpdf. Accessed June 12, 2007).

Seven trials evaluated lifestyle interventions (diet plus exercise in five trials, diet in one trial, addition of exercise to diet in one trial); nine trials evaluated pharmacologic interventions (oral diabetes drugs in six trials, orlistat [Xenical] in two trials, Chinese herb in one trial); five trials evaluated lifestyle plus pharmacologic interventions. Trials varied in interventions; in patient ethnicity, weight, and age; and in definitions of type 2 diabetes and of IGT.

An overall meta-analysis included 17 trials with 8,084 patients, and each of three meta-analyses (lifestyle interventions, oral diabetes drugs, orlistat) remained significant in sensitivity analyses employing different definitions and methodologic quality criteria.

Using a baseline five-year cumulative incidence of diabetes of 37.1%, the estimated benefit was NNT 7 for lifestyle interventions, based on 12 trials. The effect was significant for diet plus exercise, diet alone, and exercise alone; diet alone was less effective than exercise. Oral antidiabetic drugs had an NNT of 11, based on nine trials. The effect was significant for acarbose (Precose) and for metformin (Glucophage, others). Troglitazone (Rezulin) was excluded from meta-analysis because of withdrawal from the market. There was an increased rate of GI side effects with antidiabetes medications. Orlistat had an NNT of 6, based on two trials of obese patients.

There was also an increased rate of GI side effects with orlistat. There was no significant effect with jiangtang bushen recipe (Chinese herb) in one trial.