Accuracy of plasma HIV RNA quantification: a multicentre study of variability.
Inaccurate quantification of plasma HIV RNA concentration may be detrimental to patient care, yet little is known about how reproducible results are within and between laboratories. Each week between January and April 1998 a different laboratory represented at the Public Health Laboratory Service HIV Diagnosis Forum sent aliquots of the same anti-HIV positive plasma specimen by First Class Mail to the other 12 laboratories and to itself. Aliquots were frozen on receipt and examined in the next assay run. At the end of the 13 week period each laboratory reported their findings and provided further information about the specimen that they had dispatched. The correlation of results between laboratories and between the four different assay kits used was generally satisfactory. HIV RNA concentrations determined by the Roche Monitor and AcuGen kits were higher, and by the Chiron Quantiplex v 2.0 kit lower, than average. The Chiron Quantiplex gave the most reproducible concentrations. Nine 'below detection limit' results occurred, associated with three specimens. One specimen gave a below detection limit result in every one of six laboratories using the Organon Teknika Nuclisens HIV-1 QT kit, and was found to contain viral RNA of HIV-1 clade G. Another below detection limit result was probably due to technical error, and the remaining two to assay insensitivity. The findings suggested that an unsustained change in HIV RNA of <log(10)1.00 may not be a reliable basis for modifying treatment. This study confirms the need for performance assessment and for conservative interpretation of isolated changes in the estimate of HIV RNA concentration. Parallel testing of consecutive specimens from individual patients, though expensive, probably offers the best indication of significant change in the HIV RNA concentration.