Designing AD trials well, drumming up the funding, and coordinating sites across the country and the world—all that would seem to be challenge enough for clinicians in the Alzheimer disease field. But even once that’s done, they don’t enjoy smooth sailing. A particularly vexing problem down the line is that even the best trials are hemorrhaging participants, and doing so in surprisingly large numbers. What’s more, participants drop out unevenly from study arms that were well balanced at trial outset, draining the remaining sample of precious power and leaving it biased.
Why are people quitting? Jennifer Emond, a statistician at the University of California, San Diego, tackled this question. She presented her results at the 10th International Conference on Alzheimer’s Disease and Related Disorders, held July 15 to 20 in Madrid. Emond chose as her example a secondary prevention trial that was praised for its careful design and seen as a forerunner for future trials of the earliest stage of AD. Led by Ronald Peterson, Leon Thal, and other members of the Alzheimer Disease Cooperative Study (ADCS), this publicly funded trial randomized 769 people with amnestic MCI into a vitamin E arm, a donepezil arm, or a placebo arm. It aimed to treat them for 3 years. The results have received detailed discussion on Alzforum (Petersen et al., 2005). For her part, Emond took the trial database to address the issue of how many people dropped out and why. If those reasons were better understood, future trials could focus specific training and retention efforts on people at risk for early withdrawal.
Emond found that, overall, 30 percent of enrollees did not finish the trial. The donepezil group lost the most subjects at 36 percent. That was no surprise, as side effects of the study medication often cause a disproportionate dropout from the treatment group. But there is more going on than side effects: The placebo group also lost 25 percent, and Emond noticed that the incidence of side effects from donepezil was actually the same among people who stayed and those who dropped out. The biggest other factors influencing early dropout that emerged from her multivariate analysis included race/ethnicity (non-white and Hispanic participants dropped out at twice the rate than white participants) and what type of center conducted the study. Of the 69 participating sites, non-academic non-ADRC (Alzheimer Disease Research Center) sites were about twice as likely to lose study participants than were academic ADRCs. By contrast, how severely a subject’s cognition was impaired at baseline had nothing to do with dropout.
Moreover, Emond discovered that, frequently, it was the caregiver more than the study subject who was unwilling or unable to continue the study. The influence of the caregiver seems to have been great, Emond said. At the univariate level, more unmarried people than married people left prematurely, as did more women than men, raising the question of whether these two factors might be related to the caregiver’s motivation, as well. People who discontinued the study also cited as reasons their doubt about the effectiveness of the study medication, and some noted that they or their caregiver had chosen an alternate treatment. Finally, the burden of having to travel to the center for assessments every six months played a role, as well.
This initial study cannot offer a definitive answer to the question, and more large trial samples should be analyzed for dropout reasons. All the same, Emond said one safe recommendation even now might be that the staff at non-academic sites could use more training in actively supporting patients, and particularly caregivers, through the course of the trial.—Gabrielle Strobel.
- Petersen RC, Thomas RG, Grundman M, Bennett D, Doody R, Ferris S, Galasko D, Jin S, Kaye J, Levey A, Pfeiffer E, Sano M, van Dyck CH, Thal LJ, . Vitamin E and donepezil for the treatment of mild cognitive impairment. N Engl J Med. 2005 Jun 9;352(23) Epub 2005 Apr 13 PubMed.