Reducing False-Positive HIV Diagnosis in Niger

Abstract

The entry into both HIV care and secondary prevention is first through the knowledge of one’s own HIV status. Testing for HIV remains challenging in countries where clinicians rely on rapid testing algorithms because the routine use of confirmatory Western blot technology is unavailable. In this case report, we describe the case of a pregnant woman in Niger, who was falsely labeled as HIV positive during prenatal visits. We also describe our clinical algorithm that was developed to facilitate retesting in patients who initially tested HIV positive or indeterminant with rapid diagnostic tests. Vigilance is necessary to ensure that appropriate identification and treatment of HIV is provided to reduce mother-to-child transmission of HIV, to appropriately allocate resources, and to avoid falsely labeling patients with HIV.

Introduction

In low- and middle-income countries where the routine use of confirmatory Western blot technology is unavailable, implementation of routine testing for HIV remains a challenge. Clinicians utilize 2 to 3 rapid testing kits, following an algorithm to diagnose HIV infection. ¹ Rapid diagnostic tests (RDTs) for HIV are inexpensive, convenient point-of-care tests that do not require advanced laboratory services. Results and concordant counseling are provided to patients on the same day. Testing in prenatal clinics is essential for ensuring the mother’s own health and for the prevention of mother-to-child transmission (PMTCT) of HIV. Annually, an estimated 240 000 new perinatal HIV infections occur in children globally, most of which occur in sub-Saharan Africa. ² With adequate intervention, mother-to-child transmission (MTCT) of HIV can be reduced to less than 1%. However, worldwide, only about 50% of diagnosed HIV-infected pregnant women received some form of antiretroviral (ARV) therapy in 2012. ³ These rates remain regionally variable but are rapidly improving as the international community has recognized the importance of initiating ARV therapy. ⁴ A 2012 meta-analysis indicated that up to 70% of pregnant women received some form of ARV medication to prevent MTCT of HIV, which included over 75 000 pregnant women from 15 countries in sub-Saharan Africa. ⁵ These numbers suggest that real progress is being made in treating HIV-positive pregnant women with ARV medications for PMTCT.

Although these statistics are promising, clinicians must realize that increased testing may also translate into more false-positive results. Rapid diagnostic tests may result in falsely labeling a patient as HIV positive because RDTs do not have 100% positive predictive values. Additionally, the interpretation of RDTs can challenge the providers. False-positive HIV diagnoses disproportionately affect women, since pregnant patients are routinely screened for HIV as part of (PTMCT) programs. If providers falsely classify a pregnant woman with HIV, there can be devastating personal consequences. Furthermore, if the misidentified pregnant woman is treated with ARV medications, an unnecessary risk is imposed on her and her fetus. Finally, the unintentional misuse of ARV medications can further financially encumber HIV treatment programs with limited resources.

To further highlight the importance of this topic, we report on a patient who was falsely diagnosed with HIV during her routine prenatal testing using the national HIV RDT protocol in Niger. In order to reduce the recurrence of this scenario, we developed a testing protocol to identify and retest patients who were likely to be falsely labeled as HIV positive, in Niger, where HIV confirmatory Western blot testing is not feasible.

Case Presentation

Fatchima (not her real name) is a 29-year-old pregnant woman who presented for HIV testing after she missed 2 menstrual cycles. We informed her that she tested positive using 2 RDTs per our routine testing protocol in Niger. Fatchima, a mother of 3 healthy children, refused to believe the test result. She was like many of the pregnant women who we test for HIV in our setting. As we have scaled up HIV testing, offering it to all the women who seek antenatal care at our clinic, we have had the unfortunate task of telling hundreds of women that they are HIV positive. Our trained counselors spend time in a private setting, reviewing the disease, transmission, and treatment. The counselors perform a careful explanation of the testing protocol, but many patients still do not believe the results. Fatchima was informed and more insistent than most of the patients we counsel. Over a period of minutes, she explained to our counselor how her only partner, her husband, had been tested negative last month. She did not see how she could have gotten the virus if not from him. When her medical and social history were reviewed, we concluded that she was at high risk to be falsely labeled as HIV positive by our RDTs. Accordingly, she returned in 2 weeks for repeat testing, and when we delivered the good news to Fatchima, she smiled knowingly.

Discussion

In Niger, the HIV prevalence rate is 0.5%. ⁶ At SIM Galmi Hospital in Niger (Hôpital de la SIM Galmi at www.galmi.org), the national protocol for HIV testing has been implemented, which includes using the Determine HIV 1-2 (Abbott Laboratories, Abbott Park, Illinois, USA) and ImmunoComb II BiSpot HIV 1 and 2 (PBS Orgenics, Courbevoie, France) RDTs in a serial approach. Determine HIV 1-2 and Immunocomb II BiSpot HIV 1 and 2 both measure reactivity of immunoglobulin G (IgG) antibodies to HIV 1 and 2. They are considered to be second-generation enzyme immunoassays. ^7,8 Currently, in Niger, a patient is classified as HIV positive if both Determine HIV 1-2 and Immunocomb II BiSpot HIV 1 & 2 yield a positive result. If the results are discordant, the result is labeled as indeterminate and the HIV status of the individual remains unknown. The official country-wide protocol recommends that only 2 RDTs be used because a novel, third assay is not available.

As at many African centers, we are testing increasingly low-risk patients (see Table 1). This protocol yields an expected sensitivity of 100% but a specificity of 99.5% to 99.7% under ideal conditions, according to package inserts. ⁷ Performance of RDTs in resource-limited settings has been found to vary. ^9,10 In 2004, the specificity of RDTs ranged from 82.5% to 100% in Cameroon. ¹¹ In 2009, the 2-step protocol using RDTs as in Niger yielded a specificity of only 91.5%. ¹² The low specificity could result in the inclusion of up to 12% of HIV-negative people in antiretroviral therapy programs. ¹²

Table 1.

Incidence of Positive HIV Testing among Pregnant Women Attending SIM Galmi Hospital Prenatal Clinic, Niger 2009 to 2013.

	2013	2012	2011	2010	2009
Number of pregnant women tested/HIV+ results	7069/86	7930/110	3592/67	2971/63	3415/79
Percentage of HIV+ pregnant women tested	1.2%	1.4%	1.9%	2.1%	2.3%

Abbreviation: HIV+, HIV positive.

Low specificity is not the only problem that plagues the use of RDTs in low-resource settings. Interpretation of RDT results can be confusing when results show ambiguous bands or “weak bands” in the sample windows. ^13,14 The package insert for Determine HIV 1-2, for example, advises that any band, regardless of whether it is weak or strong, is a positive result. However, literature has suggested that the results of weak bands are actually a result of antigen–antibody cross-reactivity, which can lead to false-positive results, and that weak bands should not be interpreted as positive but as indeterminant. ^{13 –15}

According to the World Health Organization (WHO), in areas of low prevalence (ie, HIV prevalence rates below 5%), clinicians should perform 2 RDTs in succession if the first rapid test is positive. ¹⁶ If the second RDT is negative, the patient is deemed HIV negative. However, if the second RDT is positive, the WHO recommends that a third, novel assay be used as a tiebreaker to confirm the diagnosis of HIV because the positive predictive values of 2 RDTs are low. ¹⁷ If a third RDT is available, its platform should be different from previous platforms used because the chance of cross-reactivity with test antigens still exists. ¹⁸ The cross-reactive antibodies react to test antigens in variable ways, therefore, the outcome is determined essentially by chance if the third RDT platform is the same as the previous RDTs. ¹⁹ Novel platforms, such as third-generation RDTs, which detect both IgG and immunoglobulin M (IgM) to HIV 1 and 2, are just beginning to be used in Niger. Fourth-generation tests detect HIV IgG and IgM as well as the p24 antigen and are not widely available yet in resource-poor settings. ²⁰ Fourth-generation HIV RDTs are promising tools in detecting HIV earlier than second- and third-generation platforms. Fourth-generation RDTs could function as the tiebreakers in cases of indeterminate test results. However, extensive field testing in diverse settings is warranted before widespread implementation. ^9,21

To help more accurately diagnose patients with HIV, clinicians at SIM Galmi Hospital developed a protocol based on CD4 counts performed after positive or indeterminant HIV RDTs to retest patients who were likely to be falsely labeled as HIV positive in our testing sites (see Figure 1). CD4 counts are widely available in resource-poor settings because they are performed after HIV infection is established by RDTs to help guide therapy. ²² The protocol calls for retesting the “positive patients” if they never had an HIV-related illness; never had a CD4 count below 500; and (1) tested positive on routine prenatal screening, (2) had 2 CD4 counts of greater than 700 in more than 6 months apart, or (3) had a single CD4 count of greater than 1200. We chose these cutoffs as ranges of CD4 counts have been described in various African countries and were comparable to those reported for the United States, with a reported CD4 count reference values of 713 cells/mL in men versus 824 cells/mL in women in Mozambique. ²³ The patients will be retested with either Determine HIV 1 or 2 or HIV viral load testing. Blood will be sent to the national reference laboratory for viral load testing, or when the point-of-care viral load testing becomes available in Niger, the HIV viral load can be done as a rapid confirmatory test. As with others, we will use the cutoff of 2000 copies/mL for the indication of HIV infection. ¹³ This protocol is to be applied to the database of 3957 HIV-positive patients who receive care at SIM Galmi Hospital in Niger. A total of 127 patients have been found to be appropriate for retesting using our CD4 count cutoffs. Using this protocol, so far, the clinicians in Niger have identified 3 female patients who were falsely labeled as HIV positive. Our goal is to retest the remainder of the 127 patients we identified through this protocol and report our findings.

Figure 1.

Niger algorithm for retesting patients.

A recent article describes the retesting of clients in 3 African countries at Medecins Sans Frontiers Clinics. Two programs retested patients who had CD4 counts more than 400 or 500 cells, and 1 program retested all clients. They identified a total of 47 individuals who were falsely labeled as HIV positive for an average of 484 days. The longest a patient lived as being falsely labeled with HIV was 4 years, which can have a devastating negative impact on the patients and their families. ²⁴ From a financial standpoint, since pregnant women are routinely started on ARV medications under the national PMTCT protocol (at an estimated cost of US$144/drugs per person per year), our hospital may annually draw US$4957 for inappropriate ARV drug costs alone. Retesting entire cohorts of patients may not be feasible in resource-limited settings, but the false diagnosis of HIV is still an important consideration to patients and HIV care programs. Because women are routinely tested during prenatal care, false-positive HIV diagnoses disproportionately affect pregnant women. While testing and treating HIV is important, falsely labeling a patient with HIV has many negative consequences. ²⁴ As a result of an HIV diagnosis, a Nigerienne woman is susceptible to psychological (depression), physical (ARV side effects), professional (reduced employment opportunities), social (stigma), sexual, and financial (divorce from a seronegative husband) challenges. Furthermore, a child’s nutritional status may be compromised, as children of HIV-infected women may be weaned from breast-feeding more rapidly than necessary. From the perspective of the community, there may be a loss of trust in the medical establishment because of the perceived questionable accuracy of RDTs.

Conclusion

Although both testing and treating HIV are crucial to the control of HIV, accurate diagnoses must be made. In a low-resource setting like ours, the accessibility of RDTs, along with the critical importance of scaling up HIV testing, especially among high-impact populations such as pregnant women, makes RDTs invaluable tools in the global community’s armamentarium. However, we must recognize the downsides of this testing modality and develop ways that can ameliorate these drawbacks. Staff training in the interpretation of rapid point-of-care test results and subsequent retesting of patients with indeterminate results are pivotal to accurately diagnosing patients with HIV using RDTs. CD4 counts performed after RDTs can also guide clinicians, per our protocol, to retest patients at risk of false labeling of HIV. We hope that our protocol can be discussed, refined, and adopted widely to help prevent falsely labeling patients with HIV.