After careful review of the Avandia info, it appears the increase in cardiovascular events is caused by fluid retention. Although the the risk is about 40% higher, it's 40% higher of a very small #--out of the 25,000 patients in the meta analysis, the risk for events went from about 0.4% with non-Avandia to about 0.65% with Avandia. MI's were about 86 in the Avandia group with about 14,500 patients, and 72 in the non-Avandia group with about 11,500 patients. But about 2 weeks ago, 2 1/2 year data was released on the latest in-progress study slated to go for 4 years. Again, there were more CV events in the Avandia group, but was not statistically significant (however, is significant since the high-powered meta analysis showed that with enough numbers, it would be statistically significant). However, the big info from this study is that symptomatic congestive heart failure (CHF) had a relative risk of 2.6--big-time. Clearly, this is what fuels the increased CV events. In these studies they push to higher doses, and try to exclude CHF patients, but many especially elderly patients have what is called diastolic dysfunction--good pumping power, but problems with filling/disdending the left ventricle during relaxation/filling phase, which limits the total amount of blood that can be moved through the heart in response to fluid overload. Also, older patients often have leaky mitral valves which make some of the blood go backwards toward the lungs despite good ejection fraction. A large series of before and after echocardiograms with Avandia, 120 patients, showed that there was no negative effect on pumping power (ejection fraction) from Avandia--done as a poster 2 years ago at the ADA. I have been aware of this problem, and am very careful re signs of fluid overload, and use lower doses in older, more frail patients. Also, the mechanisms of fluid retention are augmentation of insulin's action at the kidneys to retain sodium (salt), but also, increased vasoepthelial growth factor (VEGF--there are several subtypes)--this increases capillary permeability causing increased leakage of fluid into the extravascular (outside blood vessels) space--this includes areas such as the lower legs, but also the lungs and pleural spaces. If the patient is frail and has low blood protein, especially albumin, this becomes very severe, since protein levels in the blood act like a vacuum cleaner and "suck" fluid back into the blood vessels. These effects are not reversable by diuretics such as Lasix. But, also know that TZD drugs are very protective and are the likely reason I haven't had an amputation, even of a toe, since the drugs came out. A small study, as well as my extensive clinical experience shows they stablize dementias ( a very large study is currently under way). The Dream trial in IGT patients showed a 60% reduction in progression to DM in about 5,000 patients over 4 years, "and that ain't chopped liva". The Accord study showed stablization of beta cell function, interferring in the usual progression of type 2 diabetes to requiring insulin injections. Bottom line, use lower doses in elderly, generally stay away from known CHF patients, and do not use in patients with low serum albumin/poor nutritional state.