In this issue of The Lancet Infectious Diseases, Francisco Marty and colleagues1 report a very important but negative study. The study was a multicentre, multinational, randomised, placebo-controlled, doubleblind, phase 3 trial to establish the efficacy of maribavir in prevention of cytomegalovirus infection and disease in recipients of stem-cell transplantation. The trial failed to show superiority of prophylaxis with maribavir compared with placebo (with subsequent pre-emptive therapy with valganciclovir or ganciclovir once cytomegalovirus viraemia was detected). Frequency of cytomegalovirus disease was almost identical in the maribavir (4· 4%) and placebo (pre-emptive; 4· 8%) groups. Viral detection by PCR showed similar results for both groups too. We need more anticytomegalovirus drugs since ganciclovir resistance seems to be increasing2 and the options for treatment of infections caused by these isolates are scarce. 3 Furthermore, ganciclovir-associated bone marrow toxic effects have generally limited the use of the drug for cytomegalovirus prophylaxis in stem-cell transplants. 4 With a potent in-vitro antiviral profile, and a pilot trial5 that established the potential for maribavir to prevent cytomegalovirus replication and disease, why did this trial fail? And what are the implications for future development of anticytomegalovirus drugs? The chosen dose of maribavir might have been inadequate. The study design included the lowest dose that showed biological effectiveness and minimised the most common toxic effects (dysgeusia and nausea) in the phase 2 trial. 5 The pilot trial indicated no dose response for cytomegalovirus infection or disease prevention in the comparison of three doses: 100 mg orally twice a day, 400 mg orally once a day, or 400 mg orally twice a day. 5 On the basis of these data, the investigators chose the dose of 100 mg twice a day. 1 Inspection of the plasma concentration profiles for varying doses of maribavir in