In populations with HIV prevalence greater than 5%, a two-test serial testing strategy is recommended to diagnose HIV infection [1–4]. In resource-limited settings, generally two positive HIV antibody tests, including either rapid antibody tests or ELISA antibody tests, are used to diagnose HIV infection in adults and children over 18 months of age. Ongoing risk of HIV transmission through breastfeeding, the passive transfer during pregnancy of maternal HIV antibodies from infected mothers to their infants, and the persistence of those antibodies until as late as 18 months of age, further complicates HIV testing in young infants. To confirm HIV infection in this group, virologic testing with HIV DNA PCR is recommended for children less than 18 months of age [1–4].
Although PCR testing using dried blood spots (DBS) is becoming more widely available globally, it is still unavailable in some high HIV prevalence areas. Where this is the case, the WHO recommends that clinical diagnosis of presumptive severe HIV disease guide the use of antiretroviral therapy (ART) in children less than 18 months of age. In resource-limited settings where PCR is available, expense or other concerns restrict the use of a second confirmatory PCR to verify a first positive result. To confirm the diagnosis in both of these circumstances, the WHO 2006 guidelines recommend HIV antibody testing after 18 months of age . Despite this recommendation, there is little published data on post-18-month HIV testing, and assessing the antibody response at 18 months in infants starting ART immediately has been identified by WHO as an issue for further research . The purpose of this article is to highlight the challenges of such confirmatory testing by describing the experience with post-18-month confirmatory HIV testing in children in an HIV clinic in Lesotho.
Patient population and management
The sub-Saharan African country of Lesotho has a population of 1.8 million and at 23.6% the world's third highest adult prevalence of HIV. It is estimated that 28 000 children less than 15 years of age are living with HIV and 14 000 children are born each year to HIV-infected mothers [5,6]. The Baylor College of Medicine Bristol-Myers Squibb Children's Clinical Center of Excellence (COE) is a pediatric and family HIV clinic in Maseru, Lesotho. Since 1 December 2005, the COE has offered comprehensive care to HIV-exposed and HIV-infected children, most of whom sought care directly or were referred from clinics in the urban area of Maseru or surrounding districts.
All HIV testing at the COE follows Lesotho Ministry of Health and Social Welfare (MOHSW) guidelines. A national program for early infant diagnosis of HIV using DBS HIV DNA PCR was introduced in April 2006. Lesotho MOHSW guidelines allowed for a single positive PCR test to be done for children less than 18 months of age, and recommended serologic testing after 18 months of age to confirm the diagnosis. During this period at the COE, when children within this group had two positive post-18-month confirmatory rapid antibody, no further testing was done. For children with discordant or negative parallel post-18-month confirmatory rapid antibody tests, hereafter referred to as non-positive rapid antibody tests, additional testing with PCR and/or ELISA was done, depending on availability of ELISA and at the discretion of the clinician. During the study period, ART initiation in this age group followed WHO guidance and was based on WHO clinical stage and CD4 percentage (CD4%), although toward the end of this time practice shifted to ART initiation in all infants less than 12 months with a positive PCR.
HIV rapid antibody tests were performed with the Determine HIV-1/2 (Inverness Medical Innovations Inc./Abbot Laboratories, Abbot Park, Illinois, USA) and Double-Check Gold HIV1&2 (Inverness Medical Innovations Inc./Orgenics Ltd, Waltham, Massachusetts, USA) according to printed test instructions, onsite at the COE by layperson staff trained in HIV testing according to Lesotho national guidelines. All specimens for PCR were collected, stored, and packaged by COE nurses trained in DBS specimen collection according to national guidelines. All PCRs were processed through the Lesotho national program for early infant diagnosis, which at the time sent specimens to the National Institute for Communicable Diseases (NICD) laboratory in South Africa for testing using the Amplicor HIV-1 Test (Roche, Basel, Switzerland). The NICD laboratory system supports the Department of Health in South Africa, and is accredited by the South African National Accreditation System (SANAS), an independent organization that ensures laboratory compliance with international standards. ELISA antibody testing was done at the Lesotho National Laboratory at Queen Elizabeth II Hospital or at a private laboratory in South Africa.
Study design and analysis
This report is a description of three illustrative cases and a retrospective chart review of routine program data. The COE electronic medical record (EMR) and patient files were reviewed for children who met the following inclusion criteria: enrolled at the COE between 1 December 2005 and 1 February 2009; documented positive HIV DNA PCR at less than 18 months of age; and documented HIV rapid antibody tests after 18 months of age. Data from patients meeting these criteria were extracted from the above sources and analyzed using Stata 10.1 Special Edition software (2009 edition; StataCorp, College Station, Texas, USA).
Baseline characteristics were compared between PCR-positive infants meeting inclusion criteria and those excluded, using χ2-test and Wilcoxon rank-sum test. The results of post-18-month rapid antibody tests and post-18-month PCRs and ELISAs are described. Using binary logistic regression, age at ART initiation, sex, presence of severe immunosuppression and/or severe wasting [weight-for-height Z-score (WHZ) <−3.0], and use of antiretrovirals for the prevention of mother to child transmission (PMTCT) were analyzed to determine if any factors were associated with non-positive confirmatory rapid antibody tests. Missing values for PMTCT and level of immunosuppression were imputed using multiple imputation by chained equations, which generated 25 imputed datasets [7,8]. Values were imputed using logistic regression that included PMTCT, confirmatory rapid antibody test, severe immunosuppression, sex, severe wasting, and age at ART initiation. The imputed datasets were combined and analyzed using binary logistic regression. A P-value of less than 0.05 was considered statistically significant.
The distribution of age at ART initiation was skewed and, therefore, analyzed dichotomously: less than 9 months versus at least 9 months. Severe immunosuppression was defined using the WHO immunologic classification for established HIV infection based on age and CD4% and was analyzed dichotomously as severe or nonsevere, with the lowest CD4% measure before confirmatory rapid antibody test used in the analysis. If a patient on ART did not have a pre-ART CD4 measurement, a designation of severe immunosuppression was not assigned. Patients were considered to have received PMTCT if the mother and/or the child received any antiretrovirals for PMTCT. Anthropometric measures were extracted from the EMR and WHZs generated from the 1977 National Center for Health Statistics/WHO growth references, which were in clinical use in Lesotho at the time, using Epi Info version 3.3.2 (Centers for Disease Control and Prevention, Atlanta, Georgia, USA) . Severe wasting was defined as WHZ less than −3.0 and was analyzed dichotomously. The lowest WHZ documented before the confirmatory rapid antibody test was used in the analysis.
For the purposes of determining a definitive HIV status, children were considered to be definitively positive when they had two positive PCRs, or one positive PCR with either a positive post-18-month ELISA antibody test or two positive post-18-month rapid antibody tests. Children with one positive PCR, discordant post-18-month rapid antibody tests, and no further testing were considered to be probably positive. Children were considered definitively negative when they did not meet any of the above criteria, and after review of all test results and clinical details were determined by the authors to be negative. Children not meeting any of the above definitions were considered undetermined.
Ethical approval for the study was obtained from the Baylor College of Medicine Bristol Myers-Squibb Children's Clinical Center of Excellence – Lesotho Institutional Review Board #1 (Maseru, Lesotho), the Lesotho MOHSW Research and Ethics Committee (Maseru, Lesotho), and the Baylor College of Medicine Institutional Review Board (Houston, Texas, USA).
Table 1 summarizes the details of the three patients described below.
A 1.6-month-old female born to a known HIV-infected mother presented for initial evaluation and was found to be asymptomatic with an unremarkable physical examination. The PMTCT status of mother and child were unknown. Baseline laboratory results included a positive HIV DNA PCR and a CD4 cell count of 1342 cells/μl (19%), so she started ART at 2.6 months of age with zidovudine (ZDV), lamivudine (3TC), and nevirapine (NVP). She was followed closely by the clinic and had a good immunologic and clinical response to ART. The child stopped breastfeeding at 6 months of age. Confirmatory testing at 18.9 months of age demonstrated negative parallel rapid antibody tests. She continued ART, but because of these results a HIV DNA PCR was sent. Six weeks later parallel rapid antibody tests were again negative. She never returned for follow-up, and because the confirmatory PCR result subsequently returned negative, the initial PCR was presumed to be a false-positive result and the child was believed to be definitively HIV-negative, and she was not traced. The mother brought the child back to the clinic 3 months later at 23 months of age because of draining ears, and bilateral suppurative otitis media and generalized lymphadenopathy were noted. The rapid antibody tests were redrawn that day and were positive. An HIV DNA PCR and HIV-ELISA were sent and both returned positive. There were no identifiable HIV exposures between her clinic discharge and return to care. In the interim, the CD4 cell count had dropped to 1633 cells/μl (15%) and she was reinitiated on ART.
An 8.3-month-old male presented to our clinic for evaluation. His mother was deceased and her HIV status was unknown. The child had discordant rapid antibody tests at presentation and his initial HIV DNA PCR was positive. He was WHO clinical stage 4 (severe malnutrition) and diagnosed with pulmonary tuberculosis, and at 8.8 months of age initiated ART with ZDV–3TC–NVP. The patient had a good clinical and immunologic response to treatment, with resolution of malnutrition, completion of tuberculosis treatment, and improved CD4%. Confirmatory rapid antibody tests done at 27.6 months of age and a repeat PCR sent on the same day were negative. On the basis of these negative test results and the presumption that the initial PCR was a false-positive result, ART was stopped and the patient was discharged from clinic. Because of the previous case, this patient was traced and returned to clinic 3 months later at 32 months of age. Repeat rapid antibody tests were discordant and a repeat PCR was positive. There were no identifiable exposures to HIV during the time between clinic discharge and his return to care. As he was clinically well with a stable CD4 cell count, ART reinitiation was postponed as per country guidelines at the time.
An asymptomatic 1.6-month-old female born to a known HIV-infected mother presented to the clinic for evaluation. The mother had received single-dose NVP for PMTCT. A PCR drawn at her first visit was positive, and a baseline CD4 cell count was 824 cells/μl (19%). She initiated ART at 2.4 months of age with ZDV–3TC–NVP, and had a good immunologic response. Confirmatory rapid antibody tests and a follow-up PCR done at 19 months of age were negative. Rapid antibody tests done 6 weeks later were again negative, and based on these results the child was diagnosed as definitively HIV-negative, ART stopped, and the patient discharged from clinic. Because of the previous cases, she was traced and returned to the clinic 6 months later at 26 months of age. She was clinically well with only a mild rash. Repeat rapid antibody tests were positive. There were no identifiable HIV exposures during the time between discharge from clinic and return to care. In the interim, the CD4 cell count had dropped to 1041 cells/μl (20%), so she was reinitiated on ART.
During the study period, 1775 patients less than 18 months of age were enrolled at the COE, with 397 having a documented positive HIV DNA PCR. Of these, 288 had no documentation of post-18-month rapid antibody tests, leaving a total of 109 patients meeting the study inclusion criteria (Fig. 1). Baseline demographic and clinical characteristics are detailed in Table 2.
Confirmatory HIV testing
Table 3 shows the post-18-month confirmatory HIV test results. Twenty-two patients (20.2%) had negative results and 27 patients (24.8%) discordant post-18-month confirmatory rapid antibody tests. For the 20 patients on ART with negative confirmatory rapid antibody tests, the mean age at ART initiation was 8.0 months (SD 3.6, range 2.4–15.7 months). In univariate analysis, only age less than 9 months at time of ART initiation [odds ratio (OR) 4.25, P = 0.002] was associated with non-positive rapid antibody tests (Table 4). When controlling for PMTCT, severe immunosuppression, and sex, age at ART initiation was still the only covariate associated with a non-positive confirmatory rapid antibody test [OR 3.33, 95% confidence interval (CI) 1.26–8.81, P = 0.016].
Of the 22 patients with negative confirmatory rapid antibody tests, nine of 15 (60%) follow-up PCRs were negative (Table 3). Of those nine patients with both negative confirmatory rapid antibody tests and PCRs, four had HIV antibodies tested by ELISA and all were found to be positive. Of the 27 patients with discordant confirmatory rapid antibody tests, two of nine (22.2%) had negative follow-up PCRs. Of those two, one had no additional testing documented and the other had a positive PCR and positive ELISA drawn 1 month later.
Definitive HIV status
On the basis of chart review and all available HIV test data, an attempt was made to determine the definitive HIV status of all 49 children with non-positive confirmatory rapid antibody tests. Of these 49, 29 (59.2%) were definitively positive, 17 (34.7%) were probably positive, two (4.1%) were definitively negative, and one (2%) had undetermined status at the time of the analysis (Table 3).
Initiation of potent ART during acute HIV infection has been associated with failure to seroconvert and seroreversion in a small percentage of adults [10,11]. This phenomenon is more common in children starting ART early in life [12–15], with one report describing negative ELISA and western blot in 15 of 16 infants with undetectable viral loads after early initiation of ART . Previous reports postulated that these results were explained by decreased antigenic stimulation from suppression of viral replication during a period in early childhood when immune responses, particularly HIV-specific responses, were impaired [16–18]. Using a post-18-month confirmatory testing strategy as recommended in WHO 2006 guidelines, we also found a considerable number of false-negative rapid antibody tests in children. This was significantly associated with starting ART at less than 9 months of age, although the age range at ART initiation in patients with false-negative rapid antibody tests was wide, suggesting this is a problem not just for the youngest infants. Very few false-negative rapid antibody test results would be expected for routine diagnosis of children in our population, with negative predictive values approaching 100%, although rates in children on long-term ART are not known . The rapid antibody tests reported here were done by laypersons trained in HIV testing, and reports have described decreased sensitivity of rapid antibody tests in field settings when performed by nonlaboratory personnel [19,20]. Although it cannot be excluded that incorrect testing procedures may have contributed to our results, it is unlikely this alone explains such a high rate of false-negative results. Our data adds to the known phenomenon of seroreversion in children on effective ART, but in the context of confirmatory testing in a resource-limited setting. Additionally, this provides field data on some HIV tests after 18 months in children initiated early on ART.
It was previously reported that the Determine (Abbott GmbH, Wiesbaden, Germany) and Capillus (Trinity Biotech plc., Bray, Co. Wicklow, Ireland) rapid HIV tests become negative sooner than a third-generation enzyme immunoassay when tested with serial dilutions of an HIV-positive specimen, suggesting that rapid antibody tests may be less sensitive at lower antibody levels . The different sensitivities of HIV ELISA and rapid antibody tests at low levels of antibody explain why HIV ELISA was positive in almost all cases when utilized after negative rapid antibody test and even after negative PCR. Reliance on HIV ELISA for confirmation is problematic though, as access is limited in many places due to the widespread adoption of rapid antibody tests. Additionally, loss of detectable antibody occurs for both HIV ELISA and western blot testing in young children on effective ART [13–16].
Two of 109 patients (1.8%) had false-positive initial PCRs. The percentage of false-positive PCRs in our cohort with limited PMTCT is similar to the expected number for a 30% rate of MTCT of HIV, calculated to be 1.4% by a group in South Africa and 2.3% in WHO guidelines [4,30]. Despite laboratory quality assurance, potential remains for false-positive PCRs related to the inherent test characteristics as well as to errors in transcription and labeling, among others, at the many steps between specimen collection and return of results. Our finding underscores the importance of confirming an initial positive PCR with a second virologic test, particularly as access to PMTCT improves and the expected percentage of false-positive results will increase.
There are limitations to our study. The utilization of post-18-month testing was not standardized, and due to timing of follow-up visits and variation in clinical practice, some testing was done well after 18 months of age. Missing post-18-month PCR and ELISAs limit the ability to draw robust conclusions regarding those tests. As viral load was not routinely available in Lesotho, it was not possible to definitively demonstrate that false-negative results occurred in patients on ART with suppressed viral replication. Each HIV antibody testing brand has its own specific characteristics and, therefore, it is possible other brands used for post-18-month confirmatory testing may perform differently than the tests we describe here. We do not describe the breastfeeding characteristics of our participants, although it is unlikely that many children were still breastfeeding at 18 months of age or that breastfeeding would be related to the key finding of false-negative HIV tests.
False-negative HIV tests are confusing and anxiety provoking for caregivers and healthcare providers, and may result in the interruption of treatment of truly HIV-infected children, with consequences including disease progression and development of antiretroviral drug resistance. Existing guidelines make no mention of the possibility of false-negative results when recommending post-18-month confirmatory testing. Where a definitive diagnosis based on two virologic tests is impossible, our data should not be interpreted to suggest that ART should be postponed until after 18 months because of challenges confirming the diagnosis. However, healthcare providers caring for HIV-infected children must be aware of this possibility so that results can be accurately interpreted. In settings where PCR is available, new 2010 WHO recommendations call for the immediate confirmation of a first positive PCR with a second separately collected virologic test without delaying ART initiation . Establishing a definitive diagnosis with two virologic tests prior to prolonged ART will most effectively avoid post-18-month confirmatory testing, and our report lends support to this recommendation. For children already more than 18 months old, on ART, and without a confirmed diagnosis, rapid antibody tests will establish the definitive diagnosis for a percentage, although where available ELISA may be a better test. PCR testing may provide a diagnosis in some additional children, although negative results must be interpreted with caution. If all confirmatory tests are negative, children may need to have ART interrupted cautiously with close follow-up and repeat testing to establish a definitive diagnosis. There is an urgent need for point of care HIV tests for use in children less than 18 months of age, which would increase access to early infant diagnosis, limiting the need for a presumptive clinical diagnosis, and make initial and confirmatory testing simpler and quicker. Great strides have been made in increasing access to early infant diagnosis globally, and it should continue to be a priority to ensure that all children receive accurate HIV tests for such a life-changing diagnosis.
The authors would like to acknowledge the support of the leadership and staff of the Baylor Children's Clinical Center of Excellence Lesotho, and thank all the patients and families attending the clinic.
All authors participated in the study design; A.G.P., H.R.D., and J.E.S. participated in data collection; H.R.D. performed the statistical analyses; A.G.P. and H.R.D. wrote the first draft of the manuscript; all authors participated in critical review of the manuscript.
Funding source: none.
Conflicts of interest
All authors declare that they have no conflicts of interest.
1. World Health Organization. Antiretroviral therapy of HIV
infection in infants and children in resource-limited settings: toward universal access. Recommendations for a public health approach.
Geneva: WHO; 2006. http://whqlibdoc.who.int/publications/2007/9789241594691_eng.pdf. [Accessed 15 August 2006]
2. Working Group on Antiretroviral Therapy and Medical Management of HIV
-Infected Children. Guidelines for the use of antiretroviral agents in pediatric HIV infection.
2009. pp. 1–139. http://aidsinfo.nih.gov/ContentFiles/PediatricGuidelines.pdf. [Accessed 18 July 2010]
3. PENTA Steering Committee. PENTA 2009 guidelines for the use of antiretroviral therapy in paediatric HIV infection.HIV Med
4. World Health Organization. WHO recommendations on the diagnosis of HIV infection in infants and children.
Geneva: WHO; 2010. http://whqlibdoc.who.int/publications/2010/9789241599085_eng.pdf. [Accessed 27 August 2010]
5. Ministry of Health and Social Welfare (MOHSW) (Lesotho), Bureau of Statistics (BOS) (Lesotho), ORC Macro 2005. Lesotho demographic and health survey 2004.
Calverton, Maryland: MOH, BOS, and ORC Macro.
6. United Nations Children's Fund. Children and AIDS: fifth stocktaking report, 2010. UNICEF; 2010. http://http://www.unicef.org
/publications/files/Children_and_AIDS-Fifth_Stocktaking_Report_2010_EN.pdf. [Accessed 10 November 2011]
7. Lee KJ, Carlin JB. Multiple imputation for missing data: fully conditional specification versus multivariate normal imputation
. Am J Epidemiol
8. White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice
. Stat Med
9. Hamill PV, Drizd TA, Johnson CL, Reed RB, Roche AF. NCHS growth curves for children birth: 18 years, United States.
DHEW Publ. No (PHS) 78-1650. Vital Health Stat
 1977; (165)
10. Hare CB, Pappalardo BL, Busch MP, Karlsson AC, Phelps BH, Alexander SS, et al. Seroreversion in subjects receiving antiretroviral therapy during acute/early HIV infection
. Clin Infect Dis
11. Kassutto S, Johnston MN, Rosenberg ES. Incomplete HIV type 1 antibody evolution and seroreversion in acutely infected individuals treated with early antiretroviral therapy
. Clin Infect Dis
12. Zanchetta M, Anselmi A, Vendrame D, Rampon O, Giaguinto C, Mazza A, et al. Early therapy in HIV-1-infected children: effect on HIV-1 dynamics and HIV-1-specific immune response
. Antivir Ther
13. Hainut M, Peltier CA, Goetghebuer T, Van der Linden D, Marissens D, Zissis G, et al. Seroreversion in children infected with HIV type 1 who are treated in the first months of life is not a rare event [letter]
. Clin Infect Dis
14. Luzuriaga K, Bryson Y, Krogstad P, Robinson J, Stechneberg B, Lamson M, et al. Combination treatment with zidovudine, didanosine, and nevirapine in infants with human immunodeficiency virus type 1 infection
. N Engl J Med
15. Claassen M, van Zyl GU, Korsman SN, Smit L, Cotton MF, Preiser W. Pitfalls with rapid HIV antibody testing in HIV-infected children in the Western Cape, South Africa
. J Clin Virol
16. Luzuriaga K, McManus M, Catalina M. Early therapy of vertical human immunodeficiency virus type 1 (HIV-1) infection: control of viral replication and absence of persistent HIV-1-specific immune responses
. J Virol
17. Luzuriaga K, Holmes D, Hereema A, Wong J, Panicali DL, Sullivan JL. HIV-1-specific cytotoxic T lymphocyte responses in the first year of life
. J Immunol
18. Luzuriaga K, Koup RA, Pikora CA, Brettler DB, Sullivan JL. Deficient human immunodeficiency virus type 1-specific cytotoxic T cell responses in vertically infected children
. J Pediatr
19. Black V, von Mollendorf C, Moyes J, Scott L, Puren A, Stevens W. Poor sensitivity of field rapid HIV testing: implications for mother-to-child transmission programme
20. Moodley D, Moodley P, Ndabandaba T, Esterhuizen T. Reliability of HIV rapid tests is user dependent
. S Afr Med J
21. Saitoh A, Hsia K, Fenton T, Powell CA, Christopherson C, Fletcher CV, et al. Persistence of human immunodeficiency virus (HIV) type 1 DNA in peripheral blood despite prolonged suppression of plasma HIV-1 RNA in children
. J Infect Dis
22. Andreoni M, Parisi SG, Sarmati L, Nicastri E, Ercoli L, Mancino G, et al. Cellular proviral HIV-DNA decline and viral isolation in naive subjects with <5000 copies/ml of HIV-RNA and >500 x 10(6)/l CD4 cells treated with highly active antiretroviral therapy
23. Garrigue I, Pellegrin I, Hoen B, Dumon B, Harzic M, Schrive MH, et al. Cell-associated HIV-1-DNA quantitation after highly active antiretroviral therapy-treated primary infection in patients with persistently undetectable plasma HIV-1 RNA
24. Aleman S, Visco-Comandini U, Loré K, Sönnerborg A. Long-term effects of antiretroviral combination therapy on HIV type 1 DNA levels
. AIDS Res Hum Retroviruses
25. Desai N, Mathur M, Abu-Lawi K. HIV-seronegativity in a child with proved perinatal HIV-infection on HAART
. Sex Transm Infect
26. Hainaut M, Peltier CA, Gérard M, Marrisens D, Zissis G, Levy J. Effectiveness of antiretroviral therapy initiated before the age of 2 months in infants vertically infected with human immunodeficiency virus type 1
. Eur J Pediatr
27. Barlow KL, Tosswill JH, Parry JV, Clewley JP. Performance of the Amplicor human immunodeficiency virus type 1 PCR and analysis of specimens with false-negative results
. J Clin Microbiol
28. Zazzi M, Romano L, Catucci M, De Milito A, Almi P, Gonnelli A, et al. Low human immunodeficiency virus type 1 (HIV-1) DNA burden as a major cause for failure to detect HIV-1 DNA in clinical specimens by PCR
. J Clin Microbiol
29. Mitchell C, Jennings C, Brambilla D, Aldrovandi G, Amedee AM, Beck I, et al. Diminished human immunodeficiency virus type 1 DNA yield from dried blood spots after storage in a humid incubator at 37 degrees C compared to -20 degrees C
. J Clin Microbiol
30. Feucht U, Forsyth B, Kruger M. False-positive HIV DNA PCR testing of infants: implications in a changing epidemic
. S Afr Med J
Keywords:© 2012 Lippincott Williams & Wilkins, Inc.
diagnosis; false-negative; HIV antibodies; HIV-1 DNA; HIV; pediatric; PCR