Background:  Early diagnosis of HIV infection in exposed infants is critical to prompt access to care and treatment. Dried blood spot (DBS) -based DNA polymerase chain reaction (PCR) is the method of choice in resource-limited settings. Nigeria embarked on a demonstration project from February to August 2007 to evaluate testing algorithms and quality assurance procedures.

Methods:  DBS samples were collected from HIV-exposed children, aged 6 weeks to 18 months, and tested in a central laboratory by Roche Amplicor DNA PCR v. 1.5 using 6-mm disks. Initial positive samples were retested using second disks from the same DBS samples. If both tests were positive, the sample was reported positive. If the second test was negative, the sample was further investigated:  10% of negative, all positive, and all discordant samples were further tested in a secondary lab for additional quality assurance. Both laboratories were enrolled in an external proficiency program.

Results:  Of the 561 samples tested, 123 (21.9%) were initially positive and confirmed by second tests from the same DBS cards in the primary laboratory. Of the initial positive samples, 31 were found to be negative due to laboratory errors. Of these, 18 errors were a result of a run involving a new technical staff member. However, because of the retesting requirement for positive samples, these errors were recognized prior to the release of test results. For additional quality assurance/quality control, 97 positive samples were examined by a quality assurance laboratory; 96 (99%) were found positive and 1 was found negative. This negative sample was investigated and found to yield 4 negative, 1 indeterminate, and 2 positive results indicating that its HIV DNA content was at the level of detection threshold. We retested 40 negative samples in the quality assurance laboratory:  39 were found to be negative and 1 was indeterminate.

Conclusions:  Establishing a public health program of PCR-based infant diagnostic testing in resource-limited settings requires close monitoring and quality assurance. Performing DNA PCR on a second disk from the same DBS sample confirms the initial results and reduces errors. Laboratory staff training and strengthening of the quality assurance system are critical to producing reliable results. Countries currently limited to testing just one DBS spot may consider adopting the approach outlined here. The cost-benefit of this approach merits further evaluation and comparisons with collecting and testing a second DBS sample from infants before performing another test.