Real-world HIV Diagnostic Testing Patterns in the United States

, , , , , ,
The American Journal of Managed Care, February 2022, Volume 28, Issue 2

This retrospective study evaluated real-world implementation of the updated CDC HIV algorithm in a large US laboratory.

ABSTRACT

Objectives: To understand real-world implementation of the updated CDC HIV diagnostic testing algorithm.

Study Design: Retrospective database analysis.

Methods: Using data from Quest Diagnostics, we identified patients with at least 1 HIV-1/HIV-2 antibody differentiation test (BioRad Geenius HIV 1/2 Supplemental Assay [Geenius]) between January 1 and December 31, 2017. Study measures included Health Insurance Portability and Accountability Act–compliant patient demographics, test results, test frequency, and sequence relative to the CDC HIV diagnostic algorithm, including HIV-1 RNA Qualitative Assay (Aptima) or HIV-2 nucleic acid test (NAT).

Results: A total of 26,319 patients were identified (mean [SD] age, 40.7 [14.3] years; 66.4% male), with 28,954 Geenius tests, 7234 Aptima tests, and 298 HIV-2 NATs. In 26.4% of test sequences, the Geenius results were indeterminate or negative and required subsequent confirmatory NATs. A total of 8.5% of patients had more than 1 Geenius test in 2017, and 11.2% of the time, results of the first and second tests differed. A total of 74.2% of test sequences matched the CDC-recommended algorithm.

Conclusions: Our study findings suggest that the CDC HIV diagnostic algorithm is complex and may pose suboptimal testing efficiency. Opportunities to improve diagnostic efficiency by reducing indeterminate results and repeat tests are warranted.

Am J Manag Care. 2022;28(2):e42-e48. https://doi.org/10.37765/ajmc.2022.88826

_____

Takeaway Points

Is the updated (2014) CDC HIV algorithm optimal for diagnosis of HIV-1 and HIV-2?

  • The CDC HIV diagnostic algorithm is complex.
  • Findings of this study suggest that inefficiencies in the testing algorithm remain, as indicated by the number of discrepant test results, indeterminate results, and repeat tests.
  • Future opportunities to improve diagnostic efficiency are warranted.

_____

In 2014, the CDC and the Association of Public Health Laboratories (APHL) recommended an updated HIV diagnostic testing algorithm with a 3-step sequence of tests for detection, differentiation, and confirmation of HIV-1 and HIV-2. The algorithm begins with a fourth-generation HIV-1/HIV-2 antigen/antibody (Ag/Ab) combination immunoassay. If the initial immunoassay is nonreactive, no further testing is required. However, if the initial immunoassay is reactive, a supplemental HIV-1/HIV-2 antibody differentiation test is recommended. If the initial immunoassay is reactive and the antibody differentiation test is nonreactive or indeterminate, then a confirmatory qualitative HIV-1 RNA test is suggested to resolve discrepant results. Further testing with an HIV-2 nucleic acid test (NAT) may be requested if clinically indicated.1 The CDC released a technical update to the HIV diagnostic guideline in 2016; the update is intended to be used in conjunction with the 2014 recommendations.2

Relative to previous testing standards—HIV-1 Western blot or HIV-1 indirect immunofluorescence assay—the updated algorithm yields more accurate diagnoses of acute HIV-1 and HIV-2 infections, fewer indeterminate results, and faster turnaround times.3 However, recent studies have suggested that inefficiencies remain, with many attributed to the Bio-Rad Geenius HIV 1/2 Supplemental Assay (Geenius), which was historically the only FDA-approved test for HIV-1/HIV-2 differentiation and confirmation. (As of August 12, 2020, the cobasHIV-1/HIV-2 Qualitative Test for use on the cobas 6800/8800 Systems, a nucleic acid amplification test, was FDA approved for the detection, differentiation, and confirmation of HIV-1 and HIV-2.) Findings of 2 studies comparing the performance of the Geenius with that of the Bio-Rad HIV-1/2 Multispot Rapid Test suggested that use of Geenius may lead to additional testing because of the need to resolve indeterminate results, especially HIV-2 indeterminate results, to confirm infections.4,5

In 2017, more than 3 million HIV tests were conducted in the United States by CDC-funded health departments and community-based organizations, with millions more regularly performed in other laboratory settings.6 Given the clinical and public health implications of inaccurate or delayed test results on patient management decisions and linkage to care,7 it is important to understand how the HIV testing algorithm is implemented in the real world to identify opportunities to optimize diagnostic efficiency. The objective of this study was to evaluate real-world implementation of the updated CDC HIV diagnostic testing algorithm and to describe subsequent testing patterns in a large US clinical laboratory.

METHODS

This retrospective study utilized the Quest Diagnostics (Quest) database for entries dated January 1, 2017, through December 31, 2017. The database consists of more than a decade of laboratory data for approximately 150 million patients and is nationally representative of the US population. Key information in the Quest database includes patient demographics (age, sex, and geography), diagnostic tests identified using Logical Observation Identifiers Names and Codes and laboratory test names, and test results.

For this analysis, we started with the second step of the algorithm to limit the size of the patient cohort and contain the volume of data. We included patients who received at least 1 supplemental HIV-1/HIV-2 antibody differentiation test (Geenius) with valid test results from January 1, 2017, through December 31, 2017. These patients were assumed to have had a reactive immunoassay in the first screening step and thus proceeded to the second step for HIV-1/HIV-2 differentiation. Test result entries of “not indicated” or “not performed” were considered invalid. The earliest Geenius test date for each patient was designated the index date. All data accessible for this study were deidentified with no identifiable protected health information elements included. Patient demographic characteristics (age, sex, and geographic region) were reported. Geographic region was determined from 3-digit zip code and categorized by US Census regions (Northeast, Midwest, South, West). Testing-related measures were reported for the index Geenius test (software version 1.1) and any repeated Geenius tests or subsequent confirmatory NATs, such as the Aptima HIV-1 RNA Qualitative Assay (Aptima) or HIV-2 NATs, that occurred on or after the index date. Testing-related measures included the frequency of tests in 2017, the distribution of test results, and concordance of Geenius test results among patients with more than 1 Geenius test in 2017 (ie, repeat testing). Concordance of repeat tests was examined only among patients whose first 2 Geenius tests could be defined (ie, patients with > 2 tests on the index date or > 1 test on the second test date were excluded from concordance analyses due to the inability to determine the order of testing). Geenius test results were determined from the HIV-1 and HIV-2 antibody results in the Quest data and interpreted according to the package insert.8 In exploratory analyses, the test date of service and the account identifier fields were used to identify when repeat Geenius tests were conducted and whether they were ordered by the same health care provider or practice. The frequency of acute HIV-1 infection, defined as a negative or indeterminate HIV-1 antibody result on the earliest Geenius test followed by a positive HIV-1 RNA result on Aptima, was also evaluated.

Additionally, HIV diagnostic test sequence was described, which was defined as a series of a maximum of 3 tests starting with Geenius, followed by Aptima, and ending with HIV-2 NAT, in this specific order. Tests in this sequence could occur chronologically on the same day or on different days. If tests did not follow this sequence, then the value “not performed” was substituted in place of the missing piece of the sequence. Patients with more than 1 of the same HIV tests on the same specimen collection date were excluded from analyses of testing sequence because it was not possible to distinguish which test occurred first because of a lack of specific test time in the database. All descriptive analyses were conducted using SAS version 9.4 (SAS Institute Inc).

RESULTS

This study included 26,319 patients with at least 1 Geenius test in 2017, with 28,954 total test results. The mean (SD) age was 40.7 (14.3) years, with 27.2% (n = 7153) of patients aged between 25 and 34 years. Two-thirds (66.4%; n = 17,475) were male, and most patients were located in the South (42.5%; n = 11,190) and West (29.3%; n = 7722) (Table 1).

HIV-1/HIV-2 Antibody Differentiation Test Results

Test results from 28,954 Geenius tests conducted in 2017 are shown in the Figure. Among the test results, 73.3% (n = 21,226) were positive for HIV-1, 0.03% (n = 10) were positive for HIV-2, 0.01% (n = 2) were HIV-2 positive with HIV-1 cross-reactivity (considered HIV-2 positive), 0.16% (n = 45) were HIV positive untypable (undifferentiated; indicates the possibility of dual infection with HIV-1 and HIV-2), and 24.3% (n = 7038) were HIV negative. Additionally, 2.2% (n = 630) of results were indeterminate (including 1.4% [n = 407] HIV-1 indeterminate, 0.7% [n = 192] HIV-2 indeterminate, and 0.1% [n = 31] HIV indeterminate).

Repeat Antibody Differentiation Testing

Concordance of the first and second Geenius test results on serially collected specimens was assessed for 8.5% (n = 2231) of patients who received more than 1 Geenius test in 2017. Of the test results among all patients with Geenius tests, 4.7% (n = 1353) were HIV-1 positive/HIV-2 indeterminate. Per the 2016 technical update, these results would warrant retesting of the same specimen with the Geenius. The specimen is reported as negative if the repeat test result is negative. If the specimen is repeatedly HIV-2 indeterminate, a confirmatory HIV-1 NAT is recommended.2 Based on this recommendation, the majority of patients with repeat Geenius tests would be expected to have results of HIV-1 positive/HIV-2 indeterminate. Instead, patients with HIV-1 positive/HIV-2 negative results or HIV-1 negative/HIV-2 negative results on their first Geenius test accounted for most repeat tests (58.0% and 33.7%, respectively). These repeat tests were not warranted per the updated algorithm; the reason for repeating the Geenius test for these specific patients could not be determined. Of the repeat Geenius tests, 11.2% (n = 250) had discordant results between the first and second tests, 35.2% (n = 88) of the 250 had either an indeterminate HIV-1 or HIV-2 antibody test result on the first test, and 23.2% (n = 58) of the 250 had an HIV-negative test result on their first test.

In exploratory analyses of time between first and second Geenius tests, about 50% of the 2231 patients with repeat Geenius tests had the second test within 30 days of the first (results not shown); for 60% (n = 1340) of patients, the 2 serial tests were ordered by the same provider and/or medical practice.

Confirmatory NAT

Per the CDC algorithm, discordant results for specimens reactive on the initial Ag/Ab immunoassay and nonreactive or indeterminate on the HIV-1/HIV-2 antibody differentiation immunoassay should proceed to the third step of the algorithm for a confirmatory HIV-1 NAT. Table 2 [A] describes the use of the Aptima HIV-1 NAT after the Geenius by test result. Of the 26,319 patients in the study population, 24.2% (n = 6377) received an Aptima HIV-1 NAT and 1.1% (n = 285) received an HIV-2 NAT on the same day as or after the index Geenius test (Figure). Approximately 9% of Aptima HIV-1 test results were positive (n = 647). There were 432 patients with a positive result on Aptima and a negative or indeterminate HIV-1 antibody test result on their index Geenius test, suggesting that 7.0% of patients in this study had acute HIV-1 infections. Lastly, of the 298 HIV-2 NATs, only 1 test (0.3%) was positive.

Testing Frequency

In total, 91.2% of patients (n = 24,008) received the Geenius test once, 7.8% (n = 2049) received the test twice, and 1.0% (n = 262) received the test 3 or more times. Of patients who received an Aptima HIV-1 test, 88.8% (n = 5662), 9.4% (n = 601), and 1.8% (n = 114) received Aptima 1, 2, and 3 or more times, respectively. Lastly, of patients who received HIV-2 NAT, 95.4% (n = 272) and 4.6% (n = 13) received HIV-2 NAT once and twice, respectively.

Testing Sequence

Testing sequence was examined in 26,225 patients (tests from 94 patients with > 1 of the same test on the same collection date were excluded), resulting in a total of 28,869 test sequences. Overall, 74.2% (n = 21,408) of test sequences were consistent with the updated CDC diagnostic algorithm (including HIV-2 NAT as part of the algorithm), assuming a positive HIV-1/HIV-2 Ag/Ab immunoassay.

Of the 21,128 test sequences with an HIV-positive result on Geenius, 98.9% (n = 20,898) of these did not include a confirmatory NAT; the final diagnosis was likely made based on the concordance of reactivity between the Ag/Ab screening and antibody differentiation immunoassay results. In total, 26.4% (n = 7612) of test sequences had negative (n = 6983) or indeterminate (n = 629) Geenius results, warranting confirmatory NAT. However, only 89.6% (n = 6818) of these sequences included a confirmatory Aptima HIV-1 NAT, of which 93.5% (n = 6374) of these Aptima test results were negative (Table 2 [A]). If clinically indicated, these 6374 tests would be followed by HIV-2 NAT. However, this occurred in only 2.5% (n = 160) of those test sequences, of which all results were negative (Table 2 [B]).

DISCUSSION

In this study, we retrospectively evaluated a large US laboratory database to better understand HIV testing patterns and real-world implementation of the updated CDC/APHL HIV diagnostic testing algorithm. We identified the analytic cohort as those who received at least 1 supplemental HIV-1/HIV-2 antibody differentiation test (Geenius) during the specified time frame. Starting with the second step of the algorithm helped to contain the size of the cohort for data analysis, limiting it to only those assumed to have had a reactive Ag/Ab immunoassay screening test in the first step of the algorithm and therefore proceeded to the second step of the algorithm. In total, 26,319 patients with at least 1 Geenius test in 2017 were identified in the Quest database. These patients had a mean age of 40.7 years, with 40% of patients aged between 13 and 34 years, and they were predominantly (66%) male. A CDC report on newly diagnosed cases of HIV in 2017 at CDC-funded testing sites similarly showed that 44% of new diagnoses were among persons aged 20 to 29 years, although males made up a larger proportion of new HIV diagnoses (84%).6 Such an observed difference may be expected given that the current study population is not patients with a definitive diagnosis of HIV but instead patients who were assumed to have a positive Ag/Ab screening test.

Juxtaposed to the CDC-recommended algorithm, our study indicates that about three-fourths of the HIV diagnostic testing sequences among the cohort were consistent with the recommended algorithm after a positive HIV-1/HIV-2 Ag/Ab immunoassay. Discordant results between Geenius and the initial screening assay occurred in 26.4% of test sequences and required subsequent confirmatory testing. Moreover, repeat Geenius testing was observed in about 10% of patients; many of these repeats were not warranted per the updated algorithm. Lastly, 7.0% of patients in this study appeared to have acute HIV-1 infections; similar rates of acute HIV-1 infection were observed in patients from targeted testing programs in San Francisco (6%-10%),9 but lower rates were observed in a statewide screening program in North Carolina (3.4%).10

We acknowledge that an updated Geenius software version 1.3 was introduced in 2017 to reduce the occurrence of HIV-2 indeterminate results.5 In our study, HIV-2 indeterminate results made up a small proportion (0.66%) of those that necessitated confirmatory testing. The majority of confirmatory tests were potentially needed because of the HIV-negative results on the Geenius test and the diagnostic uncertainty likely caused by discrepant screening (first step) and differentiation (second step) test results.

In evaluating test sequences, we observed that HIV-2 NAT was infrequently performed (included in 1.0% of all test sequences) even though it is recommended as a fourth step for patients with negative HIV-1 NAT results, if clinically indicated. In the study cohort, of potentially 6587 patients who may be assumed to be clinically indicated, only 285 patients received HIV-2 NAT. Although HIV-2 is suggested to be rare in the United States,11,12 the prevalence may be underestimated because of unrecognized infections from differences in clinical presentation between HIV-1 and HIV-2. A growing body of evidence shows the prevalence of HIV-2 increasing over time around the world because of migration.13-15 A recent study conducted by the New York State Department of Health Wadsworth Center reported that the number of patients identified with HIV-2 positive viral load has increased annually from 2011 to 2018.16 Based on this, low reported prevalence of HIV-2 could be secondary to infrequent use of HIV-2 NAT (as observed in this study) to confirm or rule out HIV-2 and also previous lack of an FDA-approved test with the ability to distinguish HIV positive untypable results or to diagnose HIV-2 acute infections. Because HIV-2 is resistant to all nonnucleoside reverse transcriptase inhibitors and the fusion inhibitor enfuvirtide and has shown variable resistance to protease inhibitors,17 differentiation of HIV-1 from HIV-2 is important to guide therapy selection. Alternative treatment strategies would be employed in HIV-2 monoinfected or HIV-1 and HIV-2 dually infected patients compared with HIV-1 infected patients to prevent treatment failure.18

In discussions about reducing wasteful medical spending, focus has largely been on treatment costs or expensive hospital procedures. Less attention has been on the costs and outcomes associated with laboratory medicine, despite the fact that “laboratory testing is the single highest-volume medical activity with an estimated 13 billion tests performed in the [United States] each year.”19 A laboratory diagnosis is the starting point of patients’ journeys in the health care system; it has been estimated that approximately 70% of downstream medical decisions are based on laboratory test results.19 As health care systems seek to contain costs and improve the quality of care, diagnostic tests associated with higher rates of indeterminate results, repeat tests, and/or complex diagnostic testing algorithms with frequent discordance that may require 3 to 6 different tests prior to a confirmed diagnosis should be reevaluated. Our study presents results from 1 large US clinical laboratory. If these results are generalized across laboratories in the United States, there could be a significant number of HIV diagnostic tests performed in excess. To that end, 2 recent studies have suggested that HIV-1 NAT, rather than Geenius, be used after the initial HIV Ag/Ab immunoassay, to more accurately identify both acute and established infections and to reduce the overall number of tests required for diagnosis.20,21

Limitations

A limitation of retrospective real-world data analyses is missing data for which the impact is unknown. For example, if patients in this study were nonadherent to testing orders, the data would not be available and would be considered missing (invalid). The analysis is limited to the data available in the Quest database and does not provide insights about the downstream clinical outcomes associated with the diagnostic testing sequences. Additionally, our insight into observed testing patterns, including repeat testing and adherence to the testing algorithm, is limited by the lack of transparency into the patient and provider decision-making process. For example, in our analyses of repeat testing and test result concordance, the second test may have been part of routine care for patients at high risk of HIV infection and it is possible that they may have been infected with HIV later in 2017 and, accordingly, the Geenius antibody result would change from a negative to a positive result. However, given that more than half of patients who had repeat testing had an initial HIV positive screening result and about 50% of all patients with more than 1 Geenius test had the second test within 30 days of the first, typically ordered by the same provider or practice, repeat testing may pose an unnecessary testing burden. Lastly, the results of the patients’ HIV screening tests were not available for this study to assess the rate of false positives or provide a complete view of the testing algorithm, but it is unlikely that patients would have proceeded to the second step of the algorithm if they had initially screened negative. Despite these limitations, this study provides visibility into real-world HIV diagnostic testing patterns using a large, nationally representative laboratory database and provides insights into challenges and opportunities to improve HIV diagnostic testing efficiency.

CONCLUSIONS

Although the updated CDC/APHL testing algorithm was implemented because of its improved ability to accurately diagnose HIV with fewer indeterminate results, the findings of this study suggest that inefficiencies in the testing algorithm remain, as indicated by the number of discrepant test results and repeated Geenius tests. Future clinical diagnostic algorithms should consider potential inefficiencies that may be attributed to complex multistep testing protocols and seek opportunities to streamline laboratory workflows to maximize efficient disease diagnosis.

Author Affiliations: Real World Solutions, IQVIA (CB, AMN, JT, JF), Durham, NC; Medical and Scientific Affairs, Roche Diagnostics Corporation (PLR), Indianapolis, IN; Global Access and Health Economics, Roche Molecular Systems (JKK, MMC), Pleasanton, CA.

Source of Funding: This study was funded by Roche Molecular Systems, Inc, manufacturer of the cobas HIV-1/HIV-2 Qualitative Test and the cobas 6800/8800 Systems. IQVIA received payments from Roche Molecular Systems, Inc, to design and conduct the study, interpret study findings, and draft the manuscript.

Author Disclosures: Dr Burudpakdee and Ms Faccone were employees of IQVIA, which was paid a consultancy fee to analyze data and interpret results, at the time of the study. Ms Near and Ms Tse are employed by IQVIA. Dr Rodriguez is employed by Roche Diagnostics, which sells HIV testing solutions. Dr Karichu is employed by Roche Molecular Systems. Dr Cheng was employed by Roche Molecular Systems at the time of the study and holds stock in Roche.

Authorship Information: Concept and design (CB, AMN, JT, PLR, JKK, MMC); acquisition of data (CB, JF); analysis and interpretation of data (CB, AMN, JT, JF, PLR, MMC); drafting of the manuscript (AMN, JT, PLR, MMC); critical revision of the manuscript for important intellectual content (CB, AMN, JT, PLR, JKK, MMC); statistical analysis (JF); obtaining funding (CB); administrative, technical, or logistic support (JKK, MMC); and supervision (AMN).

Address Correspondence to: James K. Karichu, PhD, Roche Molecular Systems, Inc, 4300 Hacienda Dr, Pleasanton, CA 94588. Email: james.karichu@roche.com.

REFERENCES

1. 2018 quick reference guide: recommended laboratory HIV testing algorithm for serum or plasma specimens. CDC. January 2018. Accessed May 1, 2019. https://stacks.cdc.gov/view/cdc/50872

2. Technical update on HIV-1/2 differentiation assays. CDC. August 12, 2016. Accessed April 2, 2021. https://stacks.cdc.gov/view/cdc/40790

3. Branson BM, Owen SM, Wesolowski LG, et al. Laboratory testing for the diagnosis of HIV infection: updated recommendations. CDC. June 27, 2014. Accessed October 31, 2018. https://stacks.cdc.gov/view/cdc/23447

4. Delaney KP, Ethridge S, Wesolowski LG, Owen M, Branson BM. Performance of the Geenius HIV-1/HIV-2 assay in the CDC HIV testing algorithm. Abstract presented at: Conference on Retroviruses and Opportunistic Infections; February 23-26, 2015; Seattle, WA. Abstract 621.

5. Luo W, Sullivan V, Smith T, et al. Performance evaluation of the Bio-Rad Geenius HIV 1/2 supplemental assay. J Clin Virol. 2019;111:24-28. doi:10.1016/j.jcv.2018.12.006

6. CDC-funded HIV testing: United States, Puerto Rico, and the U.S. Virgin Islands, 2017. CDC. 2018. Accessed May 21, 2019. https://www.cdc.gov/hiv/pdf/library/reports/cdc-hiv-funded-hiv-testing-report-2017.pdf

7. Fauci AS, Redfield RR, Sigounas G, Weahkee MD, Giroir BP. Ending the HIV epidemic: a plan for the United States. JAMA. 2019;321(9):844-845. doi:10.1001/jama.2019.1343

8. Bio-Rad. Geenius HIV 1/2 Supplemental Assay instructions for use. FDA. 2014. Accessed November 13, 2018. https://www.fda.gov/downloads/BiologicsBloodVaccines/BloodBloodProducts/ApprovedProducts/PremarketApprovalsPMAs/UCM420735.pdf

9. Pilcher CD, Louie B, Facente S, et al. Performance of rapid point-of-care and laboratory tests for acute and established HIV infection in San Francisco. PLoS One. 2013;8(12):e80629. doi:10.1371/journal.pone.0080629

10. Kuruc JD, Cope AB, Sampson LA, et al. Ten years of screening and testing for acute HIV infection in North Carolina. J Acquir Immune Defic Syndr. 2016;71(1):111-119. doi:10.1097/QAI.0000000000000818

11. Wesolowski LG, Chavez PR, Cárdenas AM, et al. Routine HIV test results in 6 US clinical laboratories using the recommended laboratory HIV testing algorithm with Geenius HIV 1/2 Supplemental Assay. Sex Transm Dis. 2020;47(5S suppl 1):S13-S17. doi:10.1097/OLQ.0000000000001102

12. Peruski AH, Wesolowski LG, Delaney KP, et al. Trends in HIV-2 diagnoses and use of the HIV-1/HIV-2 differentiation test — United States, 2010-2017. MMWR Morb Mortal Wkly Rep. 2020;69(3):63-66.
doi:10.15585/mmwr.mm6903a2

13. Spach DH. Epidemiology of HIV. National HIV Curriculum. Updated June 29, 2021. Accessed May 2, 2019. https://www.hiv.uw.edu/go/screening-diagnosis/epidemiology/core-concept/all

14. Valadas E, França L, Sousa S, Antunes F. 20 years of HIV-2 infection in Portugal: trends and changes in epidemiology. Clin Infect Dis. 2009;48(8):1166-1167. doi:10.1086/597504

15. de Mendoza C, Cabezas T, Caballero E, et al. HIV type 2 epidemic in Spain: challenges and missing opportunities. AIDS. 2017;31(10):1353-1364. doi:10.1097/QAD.0000000000001485

16. Styer L, Parker M. Clinical HIV-2 viral load testing of a large population of HIV-2 infected individuals. Presented at: 2019 HIV Diagnostics Conference; March 25-28, 2019; Atlanta, GA.

17. Witvrouw M, Pannecouque C, Switzer WM, et al. Susceptibility of HIV-2, SIV and SHIV to various anti-HIV-1 compounds: implications for treatment and postexposure prophylaxis. Antivir Ther. 2004;9(1):57-65.

18. Campbell-Yesufu OT, Gandhi RT. Update on human immunodeficiency virus (HIV)-2 infection. Clin Infect Dis. 2011;52(6):780-787. doi:10.1093/cid/ciq248

19. Dickerson JA, Fletcher AH, Procop G, et al. Transforming laboratory utilization review into laboratory stewardship: guidelines by the PLUGS National Committee for Laboratory Stewardship. J Appl Lab Med. 2017;2(2):259-268. doi:10.1373/jalm.2017.023606

20. Pitasi MA, Patel SN, Wesolowski LG, et al. Performance of an alternative laboratory-based HIV diagnostic testing algorithm using HIV-1 RNA viral load. Sex Transm Dis. 2020;47(5S suppl 1):S18-S25. doi:10.1097/OLQ.0000000000001124

21. Masciotra S, Luo W, Rossetti R, et al. Could HIV-1 RNA be an option as the second step in the HIV diagnostic algorithm? Sex Transm Dis. 2020;47(5S suppl 1):S26-S31. doi:10.1097/OLQ.0000000000001137