HIV Antibody Testing Algorithm
Send a Question to the Network Laboratory
Answers to Frequently Asked Questions about HIV Antibody Testing
HPTN Requirements
HPTN research sites performing HIV-1 antibody testing for HPTN protocols must validate each HIV antibody testing algorithm that they intend to use for any HPTN study in which incident HIV infection is the primary endpoint. FDA approved HIV antibody tests are recommended especially for HPTN protocols conducted under an IND, but regardless of which HIV antibody tests are used, each site needs to validate the sensitivity and specificity of the testing algorithm they will use for HPTN protocols.This validation is for the algorithm, so if the same algorithm is used for multiple protocols, the validation need not be repeated for each study. If different algorithms will be used for different protocols, each algorithm needs to be validated.
The validation process requires testing specimens from 100 confirmed HIV-infected individuals and 100 confirmed HIV-uninfected individuals using an FDA approved kit along with the kit (or kits) planned for use in a study. If serum or whole blood (venous) will be used at the site for future testing then serum or whole blood specimens must be tested. If whole blood finger prick specimens will be tested, then the assays will need to be performed using finger prick specimens obtained from confirmed HIV infected and uninfected subjects. The person testing the specimens for validation should be blinded as to the HIV status of the specimens.
For example, if both the Abbott Determine and Ora-Quick tests will be used on whole blood with resolution of discrepant results by HIV-1 Western blot, then at least 100 HIV positive samples and 100 HIV negative samples should be obtained, for which the positives have been confirmed positive by Western blot or IFA that is approved by the FDA or your country’s regulatory authority. The negatives should be confirmed negative by an FDA approved HIV EIA or by two FDA approved rapid HIV tests. The final result of the testing algorithm should then be compared to the positive or negative result of the reference test (i.e., Western blot if positive or EIA if negative).
If testing reveals one or fewer false-positive or false-negative results, then the testing algorithm has been validated and the testing algorithm is approved for use in HPTN protocols. If testing results in more than one false-negative or false-positive sample, then more specimens will need to be tested.
If it is later decided to use a different algorithm for a study at the site, the validation process should be repeated, as described, for that algorithm.
Each site should send a validation report to Estelle Piwowar-Manning (epiwowa@jhmi.edu) describing the validation process used and the results of the testing in terms of sensitivity and specificity. Upon review of this documentation, the NL will provide written concurrence that the algorithm has been approved for use in HPTN studies.
There may be unique circumstances at each site that require clarification or modification of this validation process. Sites are encouraged to contact the NL for further guidance and to provide the NL with the plan for completing this requirement in advance of implementation to ensure that the process to be undertaken is adequate If subjects have given consent to be tested for HIV, additional consent should not be necessary. Since this is a quality assurance activity, not a research activity, IRB review is not required, at least in the United States.
