Results of the dexamethasone suppression test (DST) in demented nondepressed patients have varied considerably, and there is evidence that these depend on the type of dementia, ie, Alzheimer’s disease (AD) or multi-infarct dementia (MID). AD is characterized by decreased central acetylcholine (Ach) activity, while MID is thought to be vascular in etiology. Pirecetam normalizes cerebral blood flow, along with other positive hemorheological properties. It may also increase central Ach levels. We performed the DST twice, 1 week apart, on 96 stable, demented geriatric inpatients. In 33 (21 AD, 11 MID), piracetam was started at an oral dose of 4.8 g/d 4-6 months before the first DST. The remaining 64 (39 AD, 25 MID) were untreated. Among untreated patients, those with MID were less suppressible (P < 0.001 and P= 0.01) and had more week-to-week variability (P = 0.0006) in the DST than AD patients. This is consistent with our previous findings. However, treated patients showed no difference in either postdexamethasone cortisol levels or reproducibility of DST results. We interpret these findings to mean that piracetam, by increasing central Ach, makes AD patients less suppressible on the DST, and by maintaining more constant limbic blood flow, makes MID patients more DST stable.