To the Editor: In their article, ‘‘Outcome of Alzheimer’s Disease: Potential Impact of Cholinesterase Inhibitors,’’ Gillette-Guyonnet and colleagues investigated the impact of cholinesterase inhibitors (CEIs) on the natural history of Alzheimer’s disease (AD) using rapid deterioration of cognitive function, institutionalization, and weight loss as outcomes (1). They found that the risk of all three outcomes was significantly decreased after one year’s use of CEIs. These results support the role of CEIs in improving AD outcomes. The authors excluded patients whose follow-up was discontinued for any reason, including death, and no information is provided about whether mortality rates varied significantly between treated and untreated groups. Previous studies of mortality in dementia have found that it is reduced (2) or unchanged (3) by CEI use. Trials investigating the use of galantamine in mild cognitive impairment were terminated due to a nonstatistically significant higher mortality rate in the treated arm (4). Given these varied findings, mortality is an important outcome to consider in the postmarket analysis of CEIs. We have conducted a retrospective mortality review of all patients seen at our Memory Clinics since their commencement in October 1999 to May 2005. Charts were systematically reviewed to obtain the cognitive diagnosis, the treatment given and its duration, past medical history, medication use including over-the-counter supplements, physical examination findings, and imaging results. Patient status (dead or alive) was assessed by direct contact with the patients, their families, their family doctors, and by searches of local obituary columns. Of the 440 patients seen, 248 were eligible for treatment with CEIs, based on a diagnosis of Alzheimer’s disease, Lewy body dementia, or mixed dementia. Of these, 22 patients were lost to follow-up. The remaining 226 were divided into two groups based on whether they had been treated with any CEI for a period longer than 3 months, with 26 who had declined treatment (group 1) and 200 who had received it (group 2). Group 1 (untreated) patients were significantly older at referral than those in group 2 (81.4 6 4.7 vs 75.9 6 8.0, t 5 3.049, p , .05), and group 1 contained a higher proportion of female participants than group 2 (p , .05). There were no differences in MiniMental State Exam score or comorbidities at baseline. There were 8/26 (30.8%) deaths in group 1, and 35/200 (17.6%) in group 2, a difference that was not significant. Within group 1, only high blood pressure was significantly more frequent among dead patients (p , .05). Group 2 was further subdivided based on the last medication taken: donepezil (n 5 97), rivastigmine (n 5 43), or galantamine (n 5 60). The mortality rates between the three subgroups were not significantly different. The size of our sample does not allow us to analyze the extent to which the age or sex difference between groups 1 and 2 accounted for our results. Schaufele (5) cites previous estimates of survival time in dementia from 5 to 10 years. The average time of follow-up in our population (from diagnosis to last in-clinic contact) was 25.7 months in group 1 and 26.9 in group 2. Our analysis shows no increased mortality in patients treated with CEIs compared with those who were not, or between patients receiving different CEIs. However, the number of untreated patients was small and prevents us from drawing firm conclusions. It is important to continue the postmarket evaluation of these drugs. Similar research in a larger population would be desirable to clarify the possible impact of CEIs on mortality in patients with dementia.