Appropriate interpretation of HIV-1 RNA levels requires an understanding of differences in test results due to multiple factors, which include assay and biological variation as well as specimen-handling conditions. Multiple investigations with diverse patient populations and assays have suggested that the contributions of technical and biological variations to RNA levels were quite consistent and predictable and in the range of 0.3 to 0.6 log10 RNA copies/ml. To date, all of the studies that have assessed variations in the levels of HIV-1 RNA measured have been limited primarily to isolates of the B clade; thus, what is lacking is knowledge of the degree to which the clade subtype influences assay variation and whether the biological variation observed with the clade B subtype is consistent for other clades. The major finding from the workshop was the unexpected stability of the HIV-1 RNA collected and stored under a variety of specimen handling conditions. HIV-1 RNA was shown to be relatively stable in whole blood, plasma, and serum, with the greatest stability being in plasma. Separated plasma was found to have stable titers even after storage at room temperature for 24 to 48 h and repeated freeze-thaw cycles. Within the constraints of the studies described here, the potential differences in RNA levels due to various specimen- handling conditions were not large (10 to 20% due to the anticoagulant type used in the collection tube [30 to 80% if serum rather than plasma is used], 10 to 30% due to time at RT prior to processing within 24 h, 30 to 80% due to the use of a storage temperature of -20 or -80°C). Thus, the anticipated RNA levels for nonideally collected and processed plasma specimens may be only about 130% (0.11 log10) less than those for plasma specimens collected and processed ideally (assuming that these differences are additive). This 130% difference is relatively small compared to the potential total average standard deviation of up to about 400% or 0.6 log10 RNA copies/ml due to intra- and interassay (both 0.1 to 0.2 log10) and biological (0.1 to 0.2 log10) RNA copies/ml factors. On the basis of these findings, workshop participants concluded that retrospective studies, including those which have used sera or heparinized samples, should show biological comparability to studies performed under ideal conditions, and thus both retrospective and prospective studies are useful in providing an understanding of the role of HIV-1 RNA levels in blood in transmission and disease progression. However, for prospectively designed studies, workshop participants recommended that blood for quantitative HIV-1 RNA testing ideally be collected in tubes containing EDTA, processed within 6 h of collection (but up to 24 h is still acceptable), and then stored at -80°C until assayed. Novel methodological approaches which could be useful in diagnosing and quantitating viral load in developing countries were also described, i.e., the use of DPSs, or in other body fluids such as cervical-vaginal secretions, i.e., sno-strip wicks. Finally, workshop participants determined what laboratory evaluations, including assays of HIV-1 RNA levels, with blood samples should be a priority in pediatric cohort studies while acknowledging that this ultimately depends on the study question being asked. Recommendations concerning specimen handling were then developed for international and domestic studies that use assays for detection of HIV-1 RNA. The findings reported herein underscore the continued need for the exchange of information among investigation and industry with the aim of elucidating the technological parameters that influence the assays used to evaluate HIV-1 disease and therapeutic interventions. Only by understanding the factors that affect assay outcome can we appropriately discern their value and use in clinical studies and for patient management.