Early infant diagnosis

Polly Clayden, HIV i-Base

Several guidelines now recommend universal treatment for HIV-infected infants. However, in resource-limited settings early infant diagnosis (EID) is frequently an obstacle to early initiation of antiretrovirals.

A survey by World Health Organization (WHO) asked, “What is available for early infant diagnosis?” and found the number of laboratories in several countries mismatched to the estimated number of HIV-exposed infants and necessary tests. This assessment of national capacity was conducted to inform revisions to their guidelines for infant diagnosis and treatment. [1]

For this survey, a questionnaire on clinical and laboratory capacity was sent to HIV experts in 34 high-burden countries and data were collected between February and April 2008. Replies were received from 18 of the 34 selected countries: 12 African, two South American, two Asian and one Middle Eastern.

This revealed huge variation in the number of children assessed per laboratory (range 7 – 190 000 during the study period). When virological tests were offered, the entry points were usually inpatient/outpatient services, prevention of mother-to-child transmission (PMTCT) or antiretroviral therapy (ART) sites, and laboratories were centralised and usually located in capital cities. Six countries surveyed implement HIV DNA polymerase chain reaction (PCR), 5 RNA PCR and 7 both. Ten countries used filter paper with dried blood spots (DBS) to transport samples. All the countries that responded had capacity to measure CD4% and absolute CD4 cell counts.

Although the survey confirmed that several high-burden countries are building capacity for EID, it showed that at present in many countries capacity does not reflect estimated need.

In many resource-limited countries it is only possible to use a single diagnostic test. The optimal time to perform this is unclear, however, particularly when children are breastfed. The WHO researchers used a model to calculate the number of children becoming infected and being diagnosed at different time points from birth in order to estimate the optimal time to diagnose the maximum number of children but at the same time minimise mortality. [2]

This modelling showed a decreasing trend of infant survival at 6 months, depending on the time the test was performed. The investigators suggested that 4 – 6 weeks of age is the optimal time for infant testing in a breastfeeding population.

With greater laboratory capacity and newer technology, testing earlier than 6 weeks could mean earlier initiation of treatment. But the sensitivity of viral detection tests before 6 weeks of age is unknown, particularly when performed on infants with antiretroviral exposure for PMTCT.

A South African study looked at the sensitivity of assays at earlier time points in infants born to HIV-positive women at Rahima Moosa Hospital, Johannesburg.3 Blood was sampled at birth and at 2, 4, 6 and 10 weeks, and stored. HIV-exposed infants were routinely tested at 6 weeks with HIV DNA PCR using a liquid blood sample.

Stored DBS samples from each time point were tested with HIV DNA PCR (Amplicor v1.5), TaqMan HIV-1 (CAP/CTM) and APTIMA HIV-1 (GEN-PROBE) assays. The investigators used samples from two age-matched, PCR-negative infants as controls.

Mothers received a range of PMTCT interventions: no antiretrovirals, single-dose nevirapine (NVP), single-dose NVP plus zidovudine (AZT) or HAART.

At 9 months of the study, 253/373 (68%) infants had 6-week PCR results; the remaining 120 (32%) did not return for testing. Eighteen (7.1%) were HIV infected at 6 weeks despite the majority receiving formula milk exclusively and all receiving NVP and AZT PMTCT prophylaxis.

Of the 17 infected infants with complete results, both CAP/CTM and APTIMA assays were positive in 11/17, 13/13 and 14/14 birth, 4- and 6-week samples, respectively.

The quantitative CAP/CTM assay showed lower viral load results at 2 weeks of age (the only time point when false negatives occurred). The investigators noted that this was probably due to PMTCT prophylaxis increasing the proportional number of infants infected in utero who can therefore be diagnosed at birth.

Both assays were more sensitive for earlier HIV detection than HIV DNA PCR, which detected 9/17 birth samples. CAP/CTM had the highest specificity (100%) and HIV DNA PCR the lowest (95%).

Although this is a small sample, newer technologies appear to be more sensitive than standard PCR. These initial results suggest that the majority of in utero and perinatal infections can be detected by using either CAP/CTM or APTIMA assays if they are available.

There were also reports from programmes using DBS.

A sub-study of the PMTCT Keso Bora trial conducted in Burkina Faso used a quantitative HIV RNA assay (Biocentric) and assessed DBS samples compared with paired plasma samples obtained from HIV-exposed infants aged up to 6 weeks, 3 – 6 months and 9 – 18 months.[4] All measurements were performed locally.

The study investigators reported 100% sensitivity (102/102) and specificity (105/105) (95% confidence interval (CI) 97.2 – 100%, correlation 0.906) using DBS. Of note, Biocentric is the homebrew ANRS assay, so they would have to develop their own probes, reagents, etc.

A Cambodian study assessed the feasibility of very early diagnosis (0 – 3 days of age) using heel-prick samples on DBS and a real time DNA assay (Bicentric). [5] A second DBS was performed at week 6. Infants with positive results at 0 – 3 days or 6 weeks were followed up with HIV RNA quantification as soon as possible.

At 0-3 days, 3/370 (0.8%) infants had positive results (1 infant died before week 6). 327/333 were confirmed negative at 6 weeks and 6 were DNA positive (1.8%) and subsequently confirmed RNA positive.

The investigators suggested that these preliminary results demonstrate the feasibility of a minimally invasive very early diagnosis using DBS.

This article first appeared in issue 36 of the Journal of HIV Medicine, the journal of the Southern African Clinicians Society.


  1. Unless otherwise stated, all references are to the programme and abstracts of the 5th IAS Conference on HIV Pathogenesis, Treatment and Prevention, 19 – 22 July 2009, Cape Town.
  2. Penazzato M, Crowley S. What is available for early infant diagnosis?: results from WHO survey 2008. Abstract WEPEB269.
  3. Penazzato M, Crowley S. Early infant diagnosis in resource limited settings: determining the optimum timing in a breastfeeding population. Abstract WEPEB270.
  4. Sherman G et al. Earlier diagnosis of HIV infection in infants in low resource settings. Abstract WEPEB267.
  5. Gampini SE et al. Early diagnosis of paediatric HIV-1 infection among West-African breast-fed children using dried blood spots and a quantitative HIV-1 RNA assay. Abstract WEPEB264.
  6. Ngin S et al. Very early diagnosis of HIV infection in newborn at day 0-3 on DBS in Cambodia. Abstract MOPEB009.

Links to other websites are current at date of posting but not maintained.