HTB South

Point of care HIV PCR test for infant diagnosis: good performance but poorer results in youngest age group

Polly Clayden, HIV i-Base

CROI logo

A point of care (POC) assay performed well in infant diagnosis in a large laboratory study. But its error rate of 6% has implications for implementation and its newborn performance requires further evaluation.

Nei-Yuan Hsiao from the University of Cape Town presented these findings at CROI 2015.

The study evaluated the Alere q HIV1/2, a qualitative POC, nucleic acid test for HIV designed for use in the clinic that does not require a laboratory or extensive training. The test is real-time PCR, targeting unspliced HIV RNA, cartridge based, requiring a 25 uL blood sample with turnaround time of approximately 55 minutes.

Samples were routinely collected from HIV-exposed infants in the Western Cape from December 2013 to August 2014 from: primary care facilities (infants); maternity services (newborns < 7 days) and tertiary hospitals (infants and newborns).

The performance of the POC assay was compared with the local standard-of-care (SOC), Roche CAP/CTM HIV-1 V1 qualitative PCR in a laboratory setting and to the final infant HIV status (positive on two different PCR assays or two positive PCR assays on two different samples).

The investigators tested 1131 samples from 1098 infants with a median age of 47 days (IQR: 42 –117); 90 were newborns. The majority of samples were from primary facilities (74%), the remainder were from tertiary paediatric services (19%) and maternity services (7%).

There were error results in 60 samples (6%) with the first test of which, 70% were resolved with a repeat test, and >90% were assay or sample related. Errors are more common with newborns.

Excluding retest or errors, sensitivity of Alere q was 95.5% (95% CI: 91.7 to 97.9) and specificity was 99.8% (95% CI: 99.1 to 100%). See Table 1 for results.

Table 1: POC performance against SOC assay
Alere q first test Roche CAP/CTM HIV-1 PCR
Positive Negative Total
Positive 192 2 194
Negative 9 832 841
Error 11 52 63
Total 212 886 1098

Dr Hsiao noted that with re-test, sensitivity improved but specificity remained the same.

False negative samples had a median cycle threshold (CT) value of 33 on the SOC assay, higher than the true positives and errors, both p<0.001. This was possibly due to lower levels of circulating virus during early infection.

Performance against the final infant HIV test result, excluding indeterminate status and Alere q errors: sensitivity 96.9% (95% CI: 93.4 to 98.9) and specificity 100% (lower 95% CI: 99.6)

Stratified by age group, the investigators found the performance was worse in younger infants: 6-10 weeks, sensitivity 96%, specificity 100%, error rate, 4%; newborns <7 days, sensitivity 93%, specificity 100%, error rate 10%.


Hsiao N et al. Evaluation of the Alere q for point-of-care early infant HIV diagnosis in South Africa. CROI 2015. Seattle, Washington. 23-26 February 2015. Oral abstract 34.