Analytical and Clinical Performance Evaluation of the Elecsys HIV Combi PT Assay on the Cobas e 602 Analyzer for the Diagnosis of Human Immunodeficiency Virus

Denise L. Uettwiller-Geiger, PhD; Marvin Lessig, DO; Jie An, PhD; Tara Barsch; Susan Smith; Sharan Walker; Alexander Riedel, PhD; Yi Hao, MSc; Amin A. Mohammad, PhD


Am J Clin Pathol. 2019;151(4):377-385. 

In This Article

Materials and Methods

Study Design

Three geographically diverse laboratories in the United States (John T. Mather Memorial Hospital, Port Jefferson, NY; Baylor Scott & White Health, Temple, TX; and Nationwide Laboratory Services, Boca Raton, FL) tested a total of 9,899 different samples with the Elecsys HIV combi PT assay on the cobas e 602 analyzer between July 2014 and November 2015. Details of the samples tested by each laboratory are shown in Table 1 . Two additional laboratories in the United States (Quest Diagnostics, Valencia, CA, and ZeptoMetrix, Buffalo, NY) performed reference and confirmatory testing. Serum/plasma samples were collected at 18 clinical sites or purchased from vendors between July 2013 and October 2015. The clinical performance, analytical performance, and seroconversion detection sensitivity of the Elecsys HIV combi PT assay were evaluated. Inclusion criteria for the cohorts tested are provided in Supplementary Table 1 (all supplemental materials can be found at American Journal of Clinical Pathology online). Analytical specificity and drug interference assessments were performed at Roche Diagnostics (Penzberg, Germany).

Samples were collected according to common/local ethical principles and were deidentified before testing. The study was conducted in accordance with applicable regulations, the principles of the Declaration of Helsinki, and the Convention of the Council of Europe. All sites received institutional review board approval from John T. Mather Memorial Hospital Institutional Review Board, Baylor Scott & White Health Institutional Review Board, Western Institutional Review Board (for Nationwide Laboratory Services), and Copernicus Group (for Quest Diagnostics) to conduct the study.


Primary objectives were to determine the clinical performance of the Elecsys HIV combi PT assay on the cobas e 602 analyzer relative to final diagnosis (determined by the recommended confirmation algorithm[9,10] and independent of the Elecsys HIV combi PT assay result) and the positive percent agreement (PPA) and negative percent agreement (NPA) between the Elecsys HIV combi PT assay and reference assay (ARCHITECT HIV Ag/Ab Combo assay, fourth generation; Abbott Laboratories). Secondary objectives were analytical performance, seroconversion detection sensitivity (relative to the reference assay), and analytical specificity (in populations with other viral diseases and confirmed HIV-negative pregnant women) of the Elecsys HIV combi PT assay on the cobas e 602 analyzer.


The Elecsys HIV combi PT assay is a qualitative serologic, three-incubation step sandwich assay (total assay time 27 minutes; further details on the assay test principle are in the Supplementary Appendix). Results are determined automatically by the cobas e 602 analyzer by comparing the electrochemiluminescence signal from the sample with the cutoff value determined by calibration. Samples with a cutoff index (COI) less than 1.0 are nonreactive and were considered negative for HIV-1 p24 antigen and antibodies to HIV-1 and HIV-2; no further testing is required. Samples with a COI of 1.0 or more were considered reactive. These initially reactive samples were retested in duplicate with the Elecsys HIV combi PT assay; any samples with a COI of 1.0 or more in either retest were considered repeatedly reactive. These samples were confirmed according to the CDC-recommended confirmatory algorithm.[9,10] See Figure 1 for the HIV testing algorithm used in this study. All known HIV-positive samples were confirmed with a reactive HIV-1/2 Western blot and compared with the final diagnosis determined by the HIV testing algorithm.

Figure 1.

HIV testing algorithm. Algorithm for the method comparison of the Elecsys HIV combi PT assay with comparator HIV testing methods. Ab, antibody; Ag, antigen; EIA, enzyme immunoassay; HIV, human immunodeficiency virus. aUS Food and Drug Administration–approved HIV Ag/Ab reference assay. bBio-Rad Genetic System. cAbbott Real Time. dBio-Rad Genetic Systems. eZeptoMetrix. fImmunetics QualiCode.

Manufacturer instructions and results reporting procedures were followed for both the Elecsys HIV combi PT and ARCHITECT HIV Ag/Ab Combo assays.

Clinical Performance

The Elecsys HIV combi PT assay was evaluated on the cobas e 602 analyzer at three clinical laboratories using four reagent lots. Samples (United States and outside the United States) included individuals at low and high risk of HIV (adults and children/adolescents), pregnant women, and individuals confirmed positive for specific HIV groups (including groups O and M, as well as HIV-2). Samples were also evaluated on the reference assay and additional assays/methods used in the testing algorithm (Figure 1).

Fresh samples (low-risk cohorts) were tested on the Elecsys HIV combi PT assay and reference assay within 72 hours of collection. Frozen samples (−20°C or lower) were randomized and distributed approximately equally among testing sites, which were blinded to sample/donor information. Samples that were reactive on the Elecsys HIV combi PT assay and had a final diagnosis of HIV negative underwent confirmatory analysis, including individual testing against HIV-1, HIV-2, and p24 antigen. Samples nonreactive for HIV-1, HIV-2, and p24 antigen individually but reactive on the Elecsys HIV combi PT assay were deemed false positives. It should be noted that a reactive result on the Elecsys HIV combi PT assay does not differentiate between reactivity of the HIV-1 p24 antigen or of antibodies to HIV-1 and HIV-2.

Analytical Performance

Repeatability (within-run precision), intermediate precision (within-laboratory precision), and reproducibility (between-laboratory precision) were evaluated according to Clinical and Laboratory Standards Institute (CLSI) EP05-A3 guidelines.[11] Experiments were performed at three sites using cobas e 602 analyzers with three reagent lots and one lot each of PreciControl HIV Gen II (levels 1–3) and PreciControl HIV; HIV-2+ GrpO (levels 4–5). Sample pools included PreciControl HIV levels 1 to 5, eight spiked human serum pools, and dummy samples (laboratory deidentified, leftover samples). Samples were analyzed in random order during two runs per day (two replicates of each sample per run) for 21 days.

Seroconversion Detection Sensitivity

Twenty seroconversion panels, obtained from SeraCare Life Sciences and ZeptoMetrix, comprising samples taken from patients at multiple time points before and after they tested positive for HIV, were tested with the Elecsys HIV combi PT assay and reference assay.

Analytical Sensitivity/Specificity

The Elecsys HIV combi PT assay was designed to have an analytical sensitivity of 2 IU/mL or less using the First International Standard HIV-1 p24 Antigen (National Institute for Biological Standards and Control code 90/636). In an internal study, the standard was diluted with HIV-negative serum. Using three lots of reagent, six dilution steps for each standard were prepared and measured in duplicate.

Analytical specificity of the Elecsys HIV combi PT assay was tested using samples from individuals with other viral diseases/medical conditions, which were selected on the basis that they are common HIV coinfections or conditions that may potentially interfere with the assay (further details are in the Supplementary Appendix). Testing was conducted neat (unspiked; specificity testing) and with aliquots individually spiked with p24 antigens and antibodies to HIV-1 and HIV-2 (sensitivity testing). All experiments were performed on a cobas e 602 analyzer. Confirmatory testing was not performed.

Potential Interfering Analytes, Endogenous Substances, and Therapeutic Drugs

The Elecsys HIV combi PT assay was evaluated for potential cross-reactivity using samples from individuals known to be infected with hepatitis A (acute, recovered, or chronic), B, and C; rubella virus; cytomegalovirus; Epstein-Barr virus; or herpes simplex virus. Ten samples from each viral disease state were spiked to near-cutoff values with p24 antigen and antibodies to HIV-1 and HIV-2. All samples were randomized and tested both spiked and neat (unspiked).

Interference from endogenous substances (including biotin) and 18 therapeutic drugs (at concentrations 3–10 times the maximum clinically indicated dosage, per CLSI EP07-A2 guidelines[12]) was evaluated (see Supplementary Table 2). Each drug was spiked into a negative, anti-HIV antibody-positive (signal/cutoff ratio 2–4) and HIV antigen-positive (signal/cutoff ratio 2–4) sample (three replicates per drug per sample), respectively, and evaluated with the assay.

Statistical Analyses

Data from the Elecsys HIV combi PT assay were captured by Windows-based computer-aided evaluation software (WinCAEv; Roche Diagnostics). Data from reference and confirmatory assays were entered offline into WinCAEv. Clinical information was entered into an electronic database (Medrio version 9.0) at sample collection sites (prospectively collected) or at Roche (commercially acquired samples). Statistical analyses were performed using R software (R Foundation for Statistical Computing) and SAS (SAS Institute).

Imprecision was assessed for positive and negative pools by calculating coefficient of variation (CV) and SD. Acceptance criteria for positive pools were repeatability (CV ≤6%), intermediate precision (CV ≤10%), and reproducibility (CV ≤25%). Acceptance criteria for negative pools were repeatability (COI <.6, SD <.06), intermediate precision (COI <.6, SD <.09), and reproducibility (COI <.9, SD <.15). Acceptance criteria for the entire imprecision study were 95% or more of all precision results within the acceptance criteria.

To assess clinical performance, PPA and NPA were calculated by comparing Elecsys HIV combi PT assay results with the reference assay; sensitivity and specificity were calculated by comparing Elecsys HIV combi PT assay results with the final diagnosis. Data were presented as point estimates and 95% confidence intervals (CIs). Further details on the statistical analyses are in the Supplementary Appendix, including Supplementary Table 3 and Supplementary Table 4.

A list of all study cohorts, pooled data sets, and subgroups is shown in Supplementary Table 5.