The minimum inhibitory concentration (MIC) of adenine arabinoside (ara-A) in rabbit kidney microtiter tissue cultures (RK-13) to a prototype strain of herpes simplex virus, type 1 (E115) based upon inhibition of cytopathic effects is 1.5 mug/ml. In this system, the MIC of arabinosylhypoxanthine (ara-Hx), the major in vivo metabolic derivative of ara-A, is 75 mug/ml. Inhibition of cytopathic effects of herpes simplex virus, type 1 (HSV-1) in microtiter wells of RK-13 cells varies directly with the concentrations of ara-A or ara-Hx, and inversely with residual HSV-1. The MIC of ara-A for HSV-1 in RK-13 cells is 5-20 times lower than similar measures with vero renal, mouse embryo, or human foreskin cultures. With RK-13 tissue cultures in microtiter plates, an assay for "ara-A equivalents" in human body fluids was developed which compares in sensitivity with high pressure liquid chromatography and has the advantage of simultaneously measuring combined antiherpesvirus effects of ara-A and its major metabolic derivative, ara-Hx. In vitro checkerboard studies in RK-13 cells confirmed that ara-A and ara-Hx in combination had antiviral effects which are synergistic. The total of the fractional MIC of ara-A plus ara-Hx in combination also varies inversely with residual HSV-1 in microtiter wells. Because virus adsorption is complete at 2 h before specimens to be tested are added in this assay, and because human interferon is not measured in rabbit cells, the antiviral assay is not affected by the presence of type-specific antiherpesvirus antibody or human interferon.Antiviral activity (AVA) was assayed as ara-A equivalents in sera and urines from 10 patients with serious herpesvirus infections who received 2.5-20 mg/kg daily of ara-A by intramuscular or intravenous routes. When a dosage schedule of 10 mg/kg per day or more was used, sustained concentrations of AVA that ranged from 0.8 to 14.4 mug/ml were found. When an inhibitor of adenosine deaminase (covidarabine) was not added to the specimens, mean serum concentrations were congruent with3.0 mug/ml (10 mg/kg per day, i.v.), and 4.1 mug/ml (20 mg/kg per day). However, in a single patient given 20 mg/kg of ara-A daily with covidarabine immediately added to the sera, the mean concentration of AVA was 12.9 mug/ml. Urines contained even higher AVA. Assays of 19 sera were performed both by microbiologic assay for AVA and by high pressure liquid chromatography for ara-A and ara-Hx. AVA was greater by microbiologic assay, and was greater than that which could be accounted for by stoichiometric chromatographic measures of ara-A and ara-Hx. These results with sera of treated patients are consistent both with the in vitro synergy of ara-A and ara-Hx found by checkerboard titrations, and with the beneficial responses to ara-A of patients with herpesvirus infections reported here and elsewhere.