In the Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOPNIDDM), ¹⁰ acarbose, a blocker of disaccharide digestion, not only reduced the incidence of diabetes, but also demonstrated a reduced rate of increase in carotid intima-media thickness compared with placebo. In the same study, acarbose treatment was associated with a 49% reduction in composite CVD events. However, the studies have not been confirmed. In addition, acarbose is taken before each meal and has gastrointestinal side effects that limit its widespread use. As mentioned above, in the DPP study ⁹ metformin 850mg twice a day was one of the treatment arms. There was a 31% reduction in the development of diabetes compared with the control group. Unfortunately, this was a lesser reduction than the lifestylemodification arm and was associated with a drop-out rate of 17% due to gastrointestinal side effects. After discontinuing metformin for an average of 11 days, the percentage reduction in those developing diabetes fell to 24.9%. Clearly, in some patients metformin was reducing the incidence of diabetes by treating, not delaying, the development of diabetes. Lastly, the blood pressure response to metformin was not different from placebo.

The thiazolidinediones (TZDs) that are currently marketed include rosiglitazone and pioglitazone. These are antihyperglycemics that bind to the peroxisomeactivated receptor (PPAR-?) in the nucleus of the cell. They sensitize muscle, fat, and the liver to insulin, thereby reducing insulin levels. In addition, they lower blood pressure, increase HDL cholesterol, decrease triglycerides (predominately pioglitazone), reduce microalbuminuria, and decrease inflammation. With these actions, they would seem ideal for the delay of the development of diabetes and reduction of CVD. The Troglitazone in Prevention of Diabetes (TRIPOD)¹¹ study evaluated the ability of this TZD to reduce the incidence of diabetes in a group of Mexican-American women with a previous history of gestational diabetes (a high-risk group). At 400mg per day, troglitazone reduced the incidence of diabetes by around 58% by the fifth year compared with placebo at five years. Interestingly, those that were destined to not develop diabetes with this TZD were those whose insulin levels fell upon treatment. This would suggest that the earlier the treatment is initiated, the better. Contrary to the experience with metformin, once the TZD was discontinued, there was no immediate increase in the incidence of diabetes. This study has been confirmed with pioglitazone, in the Pioglitazone in Prevention of Diabetes (PIPOD) study,¹² and has shown that the beta cell preservation may be a class effect of the TZDs. These studies did not address cardiovascular risk factors or CVD events. The Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM)13 study evaluated the reduction in development of diabetes in subjects with prediabetes. Confirming the class effect of TZDs, the incidence of diabetes was again reduced to around 60%. Although there were more CVD events in the rosiglitazone-treated group compared with placebo, it was not significant due to the small number of subjects. Especially interesting is the TZD effect on CVD surrogate markers. Carotid intima-media thickening appears to be a effect of TZDs—at least troglitazone and pioglitazone.^11,12 For CVD events, the Prospective Pioglitazone Clinical Trial In Macro Vascular Events (PROACTIVE) study15 evaluated the effects of the addition of pioglitazone 45mg per day to existing therapy for glucose, lipids, and blood pressure in subjects with type 2 diabetes and demonstrated vascular disease. Although the primary composite outcome of secondary events was lower but not significantly different from the control group, the secondary outcome established before the study was conducted was significant. The latter outcome revealed all-cause mortality, nonfatal myocardial infarction (excluding silent myocardial infection), and stroke was reduced by 16% (p=0.027).

Dampening the enthusiasm for TZDs, however, are the costs and the side effects. Weight gain and edema can be troublesome. In the DREAM and PROACTIVE studies, the incidence of congestive heart failure was increased in the treated groups. Although the overall incidence of new heart failure was low, in the PROACTIVE study incidence of new heart failure was increased by 44%. It should be noted that the mortality rate in those with heart failure was not higher than in the control group.

Conclusion

Diabetes is commonly associated with CVD. Early detection of diabetes is important to identify those patients at risk for the progression to diabetes and the development of CVD. Treatment of the risk factors, such as hypertension and dyslipidemia, is paramount. Diet and increased activity is still the recognized mainstay for the treatment of the prediabetic patient. Although TZDs are tantalizing to consider in treating prediabetes, the cost and side effects deter the thoughtful clinician.