Infant human immunodeficiency virus diagnosis in resource-limited settings: issues, technologies, and country experiences

Identifying infants for testing 

The most obvious and accessible population requiring HIV testing is hospitalized and ill outpatient infants, who will have a high prevalence of HIV in any highly HIV-affected country. When available, PCR testing should be applied in hospital settings to confirm the HIV status of infants and direct their long-term medical management. In hospital settings in high-prevalence areas, routine HIV screening for all people with symptoms suggestive of HIV, quick access to HIV test results, and/or policies to begin prophylaxis for opportunistic infections and HIV treatment before definitive diagnoses are made can save many lives. Immediate HIV diagnosis for children older than 18 months can be provided through routine rapid HIV testing of all ill children. Rapid antibody tests can also play a significant role in testing of infants younger than 18 months old. Many can be determined to be HIV negative through the use of an antibody test. Those who test antibody positive and have signs and symptoms compatible with HIV may need to initiate highly active antiretroviral therapy (ARV) therapy without a definitive diagnosis if PCR is not available.15 Wherever possible, virologic tests should be used to make the definitive diagnosis, but children who access ARV therapy without virologic testing can have their HIV status confirmed by an antibody test at 18 months.

All infants known to have HIV-positive mothers, whether ill or well, should also receive diagnostic HIV testing. Successful testing of HIV-exposed infants requires a system for record keeping that facilitates their identification in settings in which children receive care. Follow-up care and HIV testing for exposed infants should take place within existing child health systems, such as immunization and growth-monitoring clinics, which are well attended in many countries and provide an excellent opportunity for provision of prophylactic cotrimoxazole, HIV testing, and referral for other services. Ensuring that each child’s HIV exposure status is documented on his or her immunization and growth records is an essential step for countries seeking to expand access to infant HIV testing. Several suggested methods of improving the care of HIV-affected children through modifications to routine maternal-child health services are provided in Table 1.

TABLE 1. Suggested additions to standard maternal-child health systems to improve care, tracking, and testing of HIV-exposed infants in settings with high HIV prevalence

	Setting	Suggested additions to standard practices
		Care	Documentation
	Antenatal care	Include routine HIV testing, screening for antiretroviral therapy, and interventions to prevent mother-to-child transmission of HIV in package of routine antenatal services.	Modify clinic log books and patient antenatal cards to include space for HIV counseling received, HIV test results, clinical status of mother (by CD4 or exam), referrals for or provision of prophylactic or therapeutic ARV drugs, and infant feeding counseling.
	Delivery care	Include routine HIV testing for women in labor, and single-dose nevirapine for HIV-infected women, in package of routine delivery services.	Modify delivery log books and antenatal cards to include HIV test results, provision of antiretroviral drugs during labor and to infants, and infant feeding counseling.
	Well-child or immunization clinics	Include follow-up of HIV-exposed children in routine well-child or immunization visits. Follow-up includes routine physical examinations, nutritional status evaluations, cotrimoxazole prophylaxis, HIV testing, and referral for antiretroviral therapy if needed.	Ensure that maternal HIV status is transferred to child health cards either at delivery or through sending antenatal care cards with mothers to child health visits. Modify child health cards and logs to include space for maternal HIV status, PMTCT interventions received by mother and infant, feeding method and nutrition counseling, infant HIV test method and results, and referrals to ARV therapy if needed.
	Ill-child care or hospital settings	Include screening for HIV exposure and HIV infection in standard intake procedures for ill children.	Modify hospital logs and charts to include space for maternal and infant HIV status, clinical status of mother and infant, and provision of or referral for ARV therapy for both.

Child health clinics can also be used to identify infants with unknown exposure status. HIV testing for mothers can be offered in these settings, and either mothers or infants can be tested to determine the need for ongoing HIV care.16 In settings with high HIV incidence, repeat HIV testing for women who were HIV negative during pregnancy may also be offered during routine child health care because women who acquire HIV during breastfeeding are at high risk of transmitting HIV to their infants.17

The role of physical examination in HIV diagnosis 

Physical examination of infants may reveal signs of AIDS, including failure to thrive, lymphadenopathy, hepatosplenomegaly, chronic dermatitis, oral candidiasis, and recurrent pneumonia or diarrhea. If the infant is known to be HIV exposed by either maternal history or a positive antibody test, these findings are highly specific (89% in 1 study in South Africa)5 in identifying infants with AIDS. However, many infants do not manifest any sign of a failing immune system until a serious infection occurs. A clinical examination by an experienced doctor had 53% sensitivity at 6 weeks and 93% sensitivity at 12 months for detecting HIV infection in the same study. Other clinical case definitions have been evaluated; 1 algorithm from Kenya had 80% sensitivity in children younger than 24 months of age.18 Criteria used by the Integrated Management of Young Child Illness system allowed identification of only 19% of HIV-infected children at 6 weeks and 53% at 12 months.19 Based on these data, clinical assessment has a limited but essential role in HIV diagnosis. Even in the presence of advanced testing technologies, both clerical and technological errors may cause incorrect test results to be delivered to clinicians, and clinical evaluations must not be overlooked.

HIV antibody testing 

Despite its limitations, HIV antibody testing is a vital part of pediatric HIV testing programs, and antibody tests are inexpensive ($1-2 for ELISA or rapid tests). Even in a very young infant, an antibody test may have an important role if the mother’s HIV status is unknown because the infant’s test result can indicate whether or not he or she was exposed to HIV. Because infants lose maternal antibodies at different rates, the use of antibody tests to identify infants who have lost maternal antibody (seroreverted) is a reasonable and low-cost approach to ruling out HIV infection in children younger than 18 months old. Most published data on time to seroreversion are relatively old, and studies using new rapid tests to redefine the time to seroreversion are ongoing. Preliminary results indicate that a large proportion of uninfected infants are antibody negative by 9 months of age using rapid tests and that seroreversion sometimes occurs as early as age 4 months.20 Errors have been found with some rapid test kits producing false-negative results in extremely ill infants, possibly because of hypogammaglobulinemia accompanying severe illness (Gayle Sherman, personal communication, May 2006). This reinforces the need for clinical judgment and physical examination in HIV diagnosis but does not negate the usefulness of negative antibody tests.

The testing of saliva for HIV antibody may be particularly useful in infants because oral fluid contains lower concentrations of all antibodies in comparison with blood. Waning maternal HIV antibodies in HIV-exposed but uninfected infants likely become undetectable earlier in oral fluids than in blood. Laboratory-based or rapid HIV tests performed on oral fluid can potentially exclude HIV infection earlier in life, and sample collection is less traumatic for infant and caregiver. Further validation of oral fluid assays is needed to establish the youngest age at which seroreversion can be detected and to determine the sensitivity and specificity of the test at different ages.21

HIV deoxyribonucleic acid (DNA) PCR 

HIV DNA PCR is the standard method for virologic diagnosis of HIV in infants in the developed world. It has been used for many years, is the diagnostic test of choice recommended by the American Academy of Pediatrics and the WHO, and has excellent sensitivity and specificity.22, 23, 24 HIV infection can often be detected at birth, and essentially all perinatal infections are detectable by 4 weeks of age.25 Infections acquired postpartum (ie, through breast-feeding) can be detected by 4-6 weeks after the last exposure.

A variety of commercial and in-house processing methods for DNA PCR testing exist worldwide, and not all tests are equally accurate with all HIV subtypes. One commercially available PCR test, Amplicor HIV-1 DNA PCR (version 1.5, Roche Molecular Systems, Branchburg, NJ), is highly accurate in detecting the multiple HIV-1 subtypes circulating in Africa, is standardized and supported for use in Africa, and has been used by researchers and pilot infant diagnosis programs in several countries.26, 27, 28 The cost for each PCR test in developing countries currently varies between $8 and $18. An earlier version of this assay, the Roche Amplicor version 1.0, is slightly less sensitive in detecting non-B subtypes. The choice of assay should be dictated by the prevailing subtypes in the country.

Ribonucleic acid (RNA) PCR 

An HIV RNA PCR, quantitative or qualitative, is also an accurate method of diagnosing HIV in young infants, with sensitivity and specificity comparable with DNA PCR testing.9, 29 However, this test is more expensive and requires the use of plasma, which is difficult to obtain from infants and transport intact, and for these reasons it has not yet had a role in large-scale infant diagnostics.

Real-time PCR 

Real-time PCR allows the technician to view the increase in the amount of DNA or RNA because it is amplified. Real-time PCR as a new approach is gaining acceptability because of its improved rapidity, sensitivity, reproducibility, and the reduced risk of carry-over contamination, and it may reduce the cost of nucleic acid testing.30, 31 This method is in use in numerous research settings and performs very well. However, at present the only commercial kits available are for quantitative and not qualitative detection of HIV, and large-scale use of these assays for public health programs has not been attempted.

Ultrasensitive (US) p24 antigen assay 

The US p24 antigen assay is slightly less sensitive than HIV PCR in identifying HIV infection in infants across various subtypes and has a specificity similar to that of HIV PCR.21, 32, 33 This quantitative viral protein detection assay utilizes simpler technology than is required for detection of viral nucleic acids, but it is still relatively complex with multiple processing steps. US p24 is not in general use because studies validating it for infant diagnosis are recent, and achieving valid results in field settings has been challenging. US p24 may provide a useful alternative where PCR is not available.

Dried blood spots (DBS) 

The collection of infant blood on DBS is expanding opportunities for infant diagnosis in settings around the world. Because infant blood for testing can be taken by simply pricking a heel, toe, or finger and dried cards are stable for relatively long periods without refrigeration, many logistical barriers to infant testing can be overcome using this simple technique. PCR performed on DBS is as accurate as PCR performed on whole blood but has higher reagent cost and some increase in processing time. DNA and RNA PCR, both standard and real time, have been successfully performed on DBS in a wide variety of settings and HIV subtypes with no loss of accuracy.9, 25, 26, 27 US p24 assays have been performed successfully on DBS; however, further validation is required before DBS can be recommended for routine use with this assay.33 Routine collection of DBS for early infant diagnosis is being implemented in many settings, including some described in Table 2.

TABLE 2. Current infant HIV diagnosis practices in resource-limited settings

	Country	HIV prevalence in pregnant women	Number of HIV-exposed infants born each year	PMTCT interventions provided by national program	HIV testing schedule for HIV-exposed infants	Infant testing technology, cost, number of laboratories	Progress toward scale-up
	Botswana	33.4% (2005 surveillance)	15,000	Mothers: CD4 <200: ARV therapy; CD4 >200: 12 weeks AZT, SD NVP Infants: 4 weeks AZT, SD NVP, formula for 12 months	PCR at 6-week immunization visit, repeat 6 weeks after weaning if breast-fed	Roche Amplicor 1.5 DNA PCR on DBS (manual extraction) and whole blood in 1 national reference laboratory, cost $18/test.⁎	Approximately one third of HIV-exposed children tested nationwide in 2005-2006, rollout of DBS testing to all clinics providing care for infants completed in 2007
	South Africa	30.2% (2005 surveillance)	300,000	Mothers: SD NVP Limited access to ARV if CD4 <200 Infants: SD NVP	PCR at 6-week immunization visit, repeat 6 weeks after weaning if breast-fed	Roche Amplicor 1.5 DNA PCR on DBS and whole blood in 6 national laboratories, charged at ∼$50/test	Approximately one third of required lab capacity available, ∼32% of HIV-exposed children tested in September 2006
	Uganda	6.5% (2004-2005 surveillance)	78,000	Mothers: 8-12 weeks AZT or ZDV/3TC 6-8 weeks; SD NVP Infants: 1 week AZT, SD NVP	PCR at 6 weeks and when clinically indicated, serology at 18 months	Roche Amplicor 15 DNA PCR on DBS. Cost varies by lab from $8 to $25. Currently at least 12 labs do PCR, although not all do on DBS	PMTCT services delivered to ∼10% of HIV+ pregnant women nationally in 2005, early infant testing available in capital. Serology available in rural areas.
	Rwanda	4.9% (2005 surveillance)	12,000	Mothers: CD4 <350: ARV therapy; First ANC visit before 28 weeks: AZT 12 weeks, SD NVP, AZT/ 3TC 1 week after delivery; First ANC visit after 28 weeks: ARV if available, otherwise above regimen. Infants: SD NVP	In pilot sites: PCR at 6-week immunization visit, repeat 6 weeks after weaning if breast-fed In other sites: rapid test at 15 months	Roche Amplicor 1.5 DNA PCR on DBS (manual extraction) in 1 national reference lab (second lab is planned) Cost $12-15 (U.S.) per test	Pilot sites provide PCR to >600 infants. Plan to roll out routine early infant testing at rural sites in 2007.

ARV, highly active antiretroviral therapy; AZT, zidovudine; SD NVP, single-dose nevirapine.

⁎ Estimated price to purchase test kits and process test in laboratory.

WHO infant HIV diagnosis guidelines 

The most recent WHO infant testing guidelines recommend using DNA PCR on dried blood spots for routine screening of exposed infants as early as 6 weeks old.15 Confirmation of a positive test result is recommended where possible, but because of the high accuracy of the DNA PCR test, it is recommended that antiretroviral therapy be initiated before confirmation based on a single positive PCR if therapy is indicated.28 In settings in which infants are tested later, screening with antibody tests is recommended before performing PCR on infants older than 9 months.15 These guidelines recommend HIV testing technologies but do not address the complex issues of test timing and leave country programs to determine an optimal algorithm based on the local context.

Timing of infant HIV testing 

In countries in which pediatric ARV therapy is readily available, infants of HIV-positive mothers do not breast-feed, and resources permit multiple tests, infants of HIV-positive mothers are tested at birth, 1-2 months old, and 2-4 months old, with confirmation of seroreversion at 12-18 months.22 This approach is expensive, and each repeat test is less cost effective because fewer new infections are detected.34, 35 In resource-limited settings, funding may dictate that an infant can have only 1 virologic test. For non–breast-fed infants, this test should probably be done at 4-6 weeks of age when perinatal infections are detectable and infants are entering the child health care system for immunizations (Figure 1).

In settings in which the use of infant formula is not acceptable, feasible, affordable, sustainable, and safe (AFASS), most HIV-positive women breast-feed, and some new infections occur throughout infancy, with slightly less than 1% of infants acquiring HIV for each month of breastfeeding after 2 months of age.36, 37, 38 In these settings, determining the optimal timing for infant HIV testing is more complex. Early testing at 4-6 weeks detects most infections: those acquired during pregnancy, birth, and early breastfeeding. This approach allows early treatment of the 20% of perinatally-infected infants who may die before 6 months of age without antiretroviral therapy.2 Some programs have noted, however, that an HIV-negative infant test result may prompt HIV-infected mothers to stop breastfeeding very early. This may dramatically increase mortality of otherwise healthy infants in settings in which replacement feeding is not AFASS.39, 40 Programmatic data from Uganda in 2006 revealed 19% mortality during 2 years of follow-up among HIV-negative infants of women receiving ARV therapy who were tested at 6 weeks, with short duration of breast-feeding (median 3 months) being the most significant predictor of mortality.41

Programs with low perinatal HIV transmission may find that testing early compromises infant survival unless intensive patient education and infant feeding counseling can ensure that infants are not weaned too early. Testing early also necessitates a second test after weaning for infants who tested negative on their first test, thus increasing laboratory costs. In some settings in which infant formula use is not AFASS and risks of malnutrition and other morbidity are higher than the risk of HIV transmission, the WHO recommends that infants of HIV-infected mothers should breast-feed for 6 months or longer.42 If these infants are tested 6 weeks after their last breast milk, between 8 and 12 months of age, an HIV antibody test could be used for screening, reserving PCR for those infants who remain seropositive (Figure 2).

Testing breastfed infants only after weaning may promote a more appropriate length of breastfeeding and requires only 1 test per infant, but a number of infants, especially in programs with high HIV transmission rates, will become ill and die from AIDS without an opportunity for early treatment. Identification of HIV-exposed infants at routine follow-up visits and provision of high-quality basic medical care (including prophylactic cotrimoxazole) can play a significant role in reducing mortality in young infants.43 In areas in which pediatric ARV therapy is not available at all, testing at 6 weeks may confer little benefit, and a single test after weaning may be more appropriate. Individual programs must decide which approach promotes greater child survival in their context; ensure appropriate infant feeding practices are followed after testing; and be prepared to adapt their approach to changing circumstances, especially changing transmission rates, feeding practices, costs, and ARV treatment availability.

Examples of country approaches to infant diagnosis 

Table 2 describes the current state of infant HIV diagnosis efforts in several resource-limited settings. An increasing number of programs are testing young infants, using increased resources available for PMTCT and pediatric HIV care. Small-scale projects have been successful in many settings, and expansion is ongoing. Programs initiating early HIV testing work closely with maternal and child health programs and are increasingly utilizing DBS for ease of sample collection and transport to referral laboratories performing PCR testing.