Quality Control Assessment of Human Immunodeficiency Virus Type 2 (HIV-2) Viral Load Quantification Assays: Results from an International Collaboration on HIV-2 Infection in 2006

Abstract

Human immunodeficiency virus type 2 (HIV-2) RNA quantification assays used in nine laboratories of the ACHI_EV_2E (A Collaboration on HIV-2 Infection) study group were evaluated. In a blinded experimental design, laboratories quantified three series of aliquots of an HIV-2 subtype A strain, each at a different theoretical viral load. Quantification varied between laboratories, and international standardization of quantification assays is strongly needed.

The human immunodeficiency virus type 2 (HIV-2) epidemic remains essentially confined to West Africa (4). In Europe, most of the patients live in France, Portugal, and Spain, although patients are also found in The Netherlands, Belgium, Luxemburg, Germany, Switzerland, Italy, the United Kingdom, and Sweden (5, 10-13). Many questions remain unanswered concerning the evolution of virological, immunological, and clinical parameters under treatment. Answering these questions requires the largest possible number of observations and standardization of outcome measurements such as that of the HIV-2 load in plasma. There is currently no available commercial assay for the quantification of plasma HIV-2 RNA. In this study, we evaluated the validity of assays used for the quantification of HIV-2 RNA among nine centers in the ACHI_EV_2E (A Collaboration on HIV-2 Infection) study group, in Belgium, France, Gambia, Germany, Portugal, Spain, Sweden, Switzerland, and the United Kingdom, in order to identify appropriate actions to improve the quality of these assays. Each of the nine participating laboratories performed a different HIV-2 RNA quantification assay (2, 3, 7-9). Laboratories used mainly different “in-house” real-time PCR assays with primers and probes located in various regions of the HIV-2 genome, and they employed different HIV-2 standards (Table 1).

TABLE 1.

Characteristics of the different quantification assays assessed by the ACHI_EV_2E Collaboration in 2006

Laboratory	System	Primer(s) and probe localization	RNA extraction	Standard used	Threshold (log10 copies/ml)	Subtypes tested	Primer supplier(s)	Probe supplier
1	LightCycler 2.0	LTRa region (SYBR green detection)	Nuclisens-MiniMag	External, synthetic RNA (ROD seq)	1.7	A, B	Eurogentec	None used
2	LightCycler	Gag gene	Magnapure	External, NIHZ quantified by electron microscopy	2.0	A, B, H	MWG, Proligo, and Applied	Proligo
3	In-house PCR + ELONA	LTR	Boom	CBL23 + internal control	2.0	A, B	MWG	MWG
4	LightCycler	Gag gene	Qiagen, viral RNA	External, NIHZ quantified by electronic microscopy	2.7	A, B	TIB MolBiol	TIB MolBiol
5	In-house QcRT-PCRb	Env gene and biotin/digoxigenin	easyMag Biomérieux	Internal control	2.3	A, B	Invitrogen	Invitrogen
6	Nuclisens EasyQ v1.1	Gag gene	Nuclisens EasyMAG	Internal	2.3	A, B	bioMérieux (Nuclisens EasyQ kit)	bioMérieux (Nuclisens EasyQ kit)
7	ExaVir Load Analyzer v1.61	Exavir RT test from Cavidi	Exavir RT test from Cavidi	Exavir RT test from Cavidi	2.6	A, SIVsm
8	TaqMan	Gag leader	HIV Monitor	External, ST isolate, only relative quantification	NTc	A, B	Microsynth	Microsynth
9	ABI Prism 7000 TaqMan probe	LTR	Qiagen viral RNA	CBL22 external + BMV internal	2.0	A, B	MWG	MWG

^aLTR, long terminal repeat.

^bQcRT-PCR, quality control reverse transcription-PCR.

^cNT, not tested.

The sample panel was prepared in a single virology laboratory (Bichat-Claude Bernard Hospital, Paris, France) by performing serial dilution of the NIHz HIV-2 subtype A supernatant, quantified by electron microscopy (ABI Technologies), in HIV-negative human plasma to obtain 10 1-ml aliquots each with final HIV-2 RNA concentrations of 1.7, 2.3, and 3.0 log₁₀ copies/ml.

The accuracy and reproducibility of assays were estimated for the three theoretical concentrations. A quantification assay was defined as accurate if at least 9 of the 10 measurements fell within a clinically acceptable interval based on the NIHz theoretical concentrations. The interval was defined as one-third of the theoretical viral load to three times the theoretical viral load for theoretical viral loads of 3.0 and 2.3 log₁₀ copies/ml and as 0 to three times the theoretical viral load for a theoretical viral load of 1.7 log₁₀ copies/ml.

Reproducibility of assays was evaluated by using the intralaboratory coherence coefficient (ILCC) and the coefficient of variation (CV) at each of the three theoretical concentrations (1). A quantification heterogeneity was defined as an ILCC greater than the theoretical value L, extracted from a reference table for p laboratories and n repetitions (1).

Each of the participating laboratories quantified 35 coded and randomized aliquots. All laboratories reported an undetectable viral load for the five HIV-negative aliquots.

At a theoretical level of 3.0 log₁₀ copies/ml, median quantifications varied from 2.9 log₁₀ copies/ml in laboratory 6 to 4.2 log₁₀ copies/ml in laboratory 8 (Fig. 1A). Five laboratories reported accurate measurements (i.e., 9 of the 10 quantifications fell in the interval from 2.52 to 3.48 log₁₀ copies/ml). The four other assays overquantified viral loads (laboratories 1, 3, 7, and 8) (Fig. 1A). Seven laboratories reported reproducible measurements (CV = 3 to 6%; ILCC = 0.5 to 0.9; L = 1.35) (laboratories 1, 2, 3, 6, 7, 8, and 9). Laboratories 2, 6, and 9 reported both accurate and reproducible measurements. At a theoretical level of 2.3 log₁₀ copies/ml, the accuracy and reproducibility of the tests at laboratories 4 and 5 could not be evaluated since 10 and 9 negative results were reported, respectively. Median quantifications varied from 2.4 log₁₀ copies/ml in laboratory 9 to 3.4 log₁₀ copies/ml in laboratory 8 (Fig. 1B). Three laboratories reported accurate measurements (i.e., 9 of the 10 quantifications fell in the interval from 1.82 to 2.78 log₁₀ copies/ml), and four laboratories overquantified viral loads (laboratories 1, 3, 7, and 8) (Fig. 1B). Six laboratories reported reproducible measurements (CV = 5 to 8%; ILCC = 0.5 to 1.1; L = 1.35). Only two laboratories, 2 and 9, reported both accurate and reproducible measurements.

FIG. 1.

Accuracy of quantification assays evaluated by the ACHI_EV_2E Collaboration in 2006. Quantification results are reported for each participating laboratory. The accuracy interval is represented by the white area for each of the three theoretical viral loads used. Panels: A, theoretical viral load = 3.0 log₁₀ copies/ml; B, theoretical viral load = 2.3 log₁₀ copies/ml; C, theoretical viral load = 1.7 log₁₀ copies/ml.

At a theoretical level of 1.7 log₁₀ copies/ml, the accuracy and reproducibility of the assays used by laboratories 4, 5, 6, and 7 could not be evaluated since 9, 10, 7, and 9 negative results were reported, respectively. The median quantifications varied from 1.4 log₁₀ copies/ml in laboratory 2 to 2.9 log₁₀ copies/ml in laboratory 8 (Fig. 1C). Two laboratories, 2 and 9, reported accurate measurements (i.e., 9 of the 10 quantifications were below 2.18 log₁₀ copies/ml), and three laboratories overquantified viral loads (laboratories 1, 3, and 8) (Fig. 1C). These three laboratories reported reproducible measurements (CV = 5 to 11%; ILCC = 0.4 to 0.8; L = 1.35). At this level, no laboratories reported both accurate and reproducible measurements.

Our study is the first international evaluation of the validity of HIV-2 RNA quantification assays. We found a fair homogeneity of results of assays evaluated at a theoretical level of 3.0 log₁₀ copies/ml and a decreasing accuracy and reproducibility of assays with decreasing viral loads, as previously shown in a quality control study of HIV-1 RNA quantification assays (6). Overall, four laboratories systematically overquantified viral loads, no. 1, 3, 7, and 8. Two laboratories, 4 and 5, reported 18 negative results out of 20 quantifications at theoretical viral loads of both 2.3 and 1.7 log₁₀ copies/ml. The thresholds of these two assays had been evaluated at 2.7 and 2.3 log₁₀ copies/ml, respectively. The cutoff values of the methods used by laboratories 5 and 7 may also explain their results for the 1.7-log₁₀-copy/ml aliquots. One laboratory (no. 6) reported both accurate and reproducible results at a theoretical viral load of 3.0 log₁₀ copies/ml but heterogeneous quantifications at 2.3 log₁₀ copies/ml and negative results at 1.7 log₁₀ copies/ml. Two laboratories, 2 and 9, reported both accurate and reproducible results at theoretical viral loads of 3.0 and 2.3 log₁₀ copies/ml and accurate but relatively poorly reproducible quantifications at 1.7 log₁₀ copies/ml. The aliquot were made from the NIHz HIV-2 supernatant, which is the standard used for the calibration of the laboratory 2 assay. That may have improved the results reported by this laboratory and decreased those observed in other laboratories such as no. 8, which has not yet been evaluated with calibrated specimens. However, laboratory 9, which did not use NIHz as a standard, reported results very similar to those obtained by laboratory 2.

This study provides an overview of the performance of different HIV-2 RNA quantification assays. If the HIV-2 supernatant counted by electron microscopy is accepted as a standard, only two laboratories reported accurate and reproducible measurements at both 3 and 2.3 log₁₀ copies/ml. This heterogeneity may yield difficulties in the comparison of results between the different cohorts of the ACHI_EV_2E Collaboration. The ACHI_EV_2E network is working toward the standardization of quantification assays to improve the interpretation of results for case management and collaborative clinical trials.