Table 2. Quality assurance testing.

	Tanzania	Zimbabwe		Soweto South Africa	Vulindlela South Africa		Total
Quality assurance testing of HIV NEG samples a
Tested with HIV Combo (%)b	1,932		1,999	2,900		2,910			9,741
Combo reactive	15 (0.78%)		7 (0.35%)	65 (2.24%)		23 (0.79%)			110 (1.13%)
HIV POS and HIV DISC samples censored (among those originally classified as HIV POS or HIV DISC)
Excluded due to contamination	0		0	292		0			292
Excluded for other reasonsc	0		1	23		4			28
Classification of HIV DISC samples a
Samples remaining after exclusions	187		19	20		46			272
HIV uninfected	184		18	18		45			265
HIV infected	3		1	2		1			7
Acute infection	1		0	0		0			1
Early infection	0		0	0		0			0
Established infection	2d		1d	2e		1e			6
Classification of HIV POS samples a
Samples remaining after exclusions	542	1,547		1,672		3,645		7,406
HIV uninfected	9	17		22		4		52
HIV infected	533	1,530		1,650		3,641		7,354
Acute infection	0	0		0		3		3
Early infection	1	0		0		2		3
Established infection	532	1,530		1,650		3,636		7,348

Abbreviations: WB: Western blot; HIV Combo: ARCHITECT HIV Ag/Ab Combo assay.

^aThe HIV status of study participants was initially characterized based on the results of the two HIV rapid tests performed in-country (see Methods): HIV POS: two reactive HIV rapid tests. HIV DISC: one reactive and one non-reactive HIV rapid test. HIV NEG: two non-reactive HIV rapid tests. The testing algorithm used to confirm the HIV status of HIV NEG and HIV DISC samples (quality assurance testing) are shown in Figure 1. Quality assurance testing was only performed for HIV POS samples if results from the avidity assay suggested absent or extremely low levels of anti-HIV antibodies.

^bThis indicates the number of samples that had reactive results using the HIV Combo assay (signal/cutoff >1). According to the package insert, specimens that are initially reactive with HIV Combo should be retested in duplicate and only repeatedly reactive specimens are considered reactive. In this study, samples were analyzed only once using the HIV Combo assay.

^c28 samples were excluded for reasons other than contamination, including: no CD4 cell count obtained (N = 5); insufficient quantity of plasma stored for testing (N = 2); failure of sample tracking (N = 17); protocol violation (N = 4).

^dThese three samples were subsequently classified as MAA positive.

^eThese three samples were subsequently classified as MAA negative.