By Michael Smith, North American Correspondent, MedPage Today
Published: January 23, 2012
Reviewed by Robert Jasmer, MD; Associate Clinical Professor of Medicine, University of California, San Francisco.
A rapid HIV test gives slightly less exact results when used with oral samples compared with blood samples, researchers reported.
In pooled results from a systematic review and meta-analysis, the sensitivity of the Oraquick rapid test was 98.03% when oral samples were used, compared with 99.68% using blood, according to Nitika Pant Pai, MD, of McGill University Health Centre in Montreal, and colleagues.
On the other hand, the specificity of the test was similar regardless of the type of samples at 99.74% for oral samples and 99.91% for blood, Pant Pai and colleagues reported online in The Lancet Infectious Diseases.
The Oraquick test, originally approved as a point-of-care HIV test for use with blood, received approval from the FDA in 2004 for use with oral mucosal transudate, the researchers noted. The test has become especially popular in developing countries.
Given the increased emphasis on HIV testing, data on the performance of the Oraquick test using the different sampling approaches will be important to help clinicians and patients interpret results, they argued.
Pant Pai and colleagues analyzed data from 45 studies, including 24 that could be used to understand the comparative accuracy of the test using oral or blood samples, and 21 that could be used to analyze positive predictive values. Because they depend on prevalence, positive and negative predictive values can only be derived from cross-sectional or other population-based studies.
Seven of the 24 accuracy studies contained head-to-head comparisons, while six only gave results for oral samples and 11 only looked at whole blood samples.
Pooling the analyses showed that the sensitivity of the test was about 2% lower when oral samples were used, but specificity was about the same, the researchers reported. In other words, using oral samples was likely to yield more false-positive results.
The researchers calculated positive predictive values using the 21 studies that permitted the analysis. They found that the test's value was similar regardless of sample type in high-prevalence regions, but differed if prevalence was low.
In high-prevalence groups, defined as more than 1% of the population infected with HIV, the positive predictive value of the test was 98.5% using blood samples and 98.65% using oral samples.
But in low-prevalence populations, the value for blood samples was 97.65% compared with 88.55% for oral samples.
Put another way, if the prevalence is low, a positive result using oral samples is less likely to be correct than if blood were used.
The researchers cautioned that although the test's sensitivity seems to be lower with oral versus blood specimens, the estimates "were at the extreme upper end of the range and there is a great deal of overlap" in confidence intervals.
It's possible that the "clinical significance of this difference might also be overshadowed by intrinsic variability in host status and time of testing relative to exposure," they added.
Much of the data comes from high-income settings and the rest from well controlled studies in developing settings, they noted, so the results may not reflect routine testing services in poorer areas.
Finally, the results only apply to the Oraquick test, they wrote.
The Oraquick test is attractive because it is convenient and non-invasive, commented Chi Chiu Leung, MBBS, of the Hong Kong Department of Health, and Shui Shan Lee, MD, of the Chinese University of Hong Kong.
Those characteristics might help increase the rate of HIV testing, they argued in an accompanying comment article, but they appear to come "at the cost of a substantial false-positive rate."
That false-positive rate has to be taken into consideration by clinicians as the rapid test becomes more widely used, they concluded, especially in areas where the prevalence of HIV is low.
Confirmatory testing will remain important, given that the diagnosis of HIV has such profound implications, they wrote.
The study was supported by the Canadian Institutes for Health Research. The authors said they had no conflicts of interest.
The comment authors said they had no conflicts of interest.
Primary source: The Lancet Infectious Diseases
Pant Pai N, et al "Head-to-head comparison of accuracy of a rapid point-of-care HIV test with oral versus whole-blood specimens: a systematic review and meta-analysis" Lancet Infect Dis 2012.
Additional source: The Lancet Infectious Diseases
Leung C, Lee S "Rapid HIV tests: from meta-analysis to field application" Lancet Infect Dis 2012.