Failure of itraconazole to prevent Enterocytozoon bieneusi infection.

Firstly, when T/2 values are as little as 1-5-2-5 days, the time between baseline venesection and first therapy must be recorded accurately as delays of initial dose by only a few hours may produce 10-20% errors in any calculation. Secondly, assumptions made in these models do not allow for sudden biological changes in viral load at or near the time of onset of treatment, and we find such events do occur in up to 10% of patients studied and will totally invalidate the subsequent analysis. Nevertheless, we believe there is a place for this simple model and have already used it to assist in the early prediction (within 1 month) of the relative efficacies of different antiretroviral therapies in HIV-1 disease.9 Such "screening" procedures in small patient groups (10-20 patients/group), receiving different combination therapies, using virological endpoints will allow the selection of superior drug combinations that can go forward into longer, more expensive clinical trials. C LOVEDAY Department of Virology, University College London Medical School, London WIP 6DB, UK