Pentoxifylline in Cerebrovascular Dementia

OBJECTIVE: To test the effect of pentoxifylline, a hemorheologic agent used to treat intermittent claudication, on the course of vascular dementia. DESIGN: Randomized, double-blind, placebo-controlled, parallel group trial. SETTING: Outpatient tertiary care center. PATIENTS: 64 patients meeting DSM-III criteria for multi-infarct dementia with modified Hachinski ischemic scores greater than or equal to 6, 38 of whom completed the trial. INTERVENTION: Pentoxifylline (Trental) 400 milligram tablets three times daily vs placebo for 36 weeks. MAIN OUTCOME MEASURE: Alzheimer's Disease Assessment Scale (ADAS). RESULTS: Baseline demographic values and psychometric variables were similar in the placebo and control groups; endpoint statistical analysis was used to allow the use of data from all patients in this clinically high-risk group. For the total group, the slowing of deterioration did not reach statistical significance (by 2-tailed t test), as measured by scores on the total ADAS (P = 0.058) or on the cognitive (ADAS items 1-11; P = 0.064) or non-cognitive subscales (ADAS items 12-21; P = 0.234), although it was significant on the cognitive subscales excluding memory (ADAS items 2-6, 8-10; P = 0.036). For the subgroup of 40 patients who had CT and/or MRI evidence of stroke as well as meeting the other inclusion criteria, treatment with pentoxifylline was associated with significantly slower deterioration, as measured by the total ADAS (P = 0.023) and cognitive subscores (P = 0.020) but not non-cognitive subscores (P = 0.118). For the subgroup of 37 patients who had at least one discrete clinical stroke, treatment with pentoxifylline was associated with significantly less deterioration on the total ADAS (P = 0.002) and both the cognitive (P = 0.001) and non-cognitive (P = 0.017) subscores. CONCLUSION: Treatment with pentoxifylline may slow the progression of dementia in patients who meet DSM-III criteria for "multi-infarct dementia" and who also have clinical and neuroradiological evidence of cerebrovascular disease.