H5N1 False Negatives in Pakistan?
Recombinomics Commentary
December 24, 2007
“In their preliminary tests the WHO team excluded suspected human-to-human transmission, but we have sent the samples to Geneva for further confirmation,” health ministry spokesman Oriya Maqbool Jan told AFP.
The above comments suggest that the shipment of samples from patients in Pakistan to London was due to false negatives at the NAMRU-3 mobile lab. Media reports had indicated that the positive results obtained in October by labs in Pakistan were confirmed in Pakistan over a week ago. Moreover, the sustained transmission in H5N1 lab confirmed contacts strongly supports human-to-human transmission.
False negatives by the NAMRU-3 mobile lab could be due to a number of factors including sample degradation due to multiple tests, viral RNA depression due to collection of samples after the start of Tamiflu treatment, or primer mismatches due to sequence differences between the H5N1 in Pakistan compared to Egypt, lack of fine tuning in the testing by the mobile lab.
Unfortunately, there is precedent for all of the above. Media quotes from the Pakistan Health Minister indicated samples from patients were collected one to two after the start of Tamiflu treatment. Since Tamiflu inhibits the release of H5N1 from an infected cell, treatment can cause to drop in RNA levels, resulting in a false negative result. Sequence differences between the H5N1 in Pakistan and Egypt are likely because the isolates in Pakistan are likely to be related to the Uva Lake strain, which is currently widespread in Europe and was previously detected in Kuwait in early 2007, but has not been reported in human H5N1 sequences generated by NAMRU-3 in Egypt.
However, there is also evidence of detection failures at the regional centers also. A year ago, 21 patients were H5N1 confirmed by labs in Turkey. However, only 12 of the 21 were confirmed by Weybridge in England. These failures were almost certainly false negatives because the samples in Turkey formed clusters, and the likelihood of pairing up false negatives is remote. These failures extended to the fourth sibling from the index cluster. Moreover, sequences from only four patients have been released, suggesting that there were significant isolation failures at Weybridge. These failures extended to mixtures because the index case from Turkey had S227N, while the isolate from the sister of the index case did not. H5N1 false negatives in Egypt were also seen in the Gharbiya cluster. Although three family members died with bird flu symptoms, sequence data was not obtained for one of the three fatalities.
Some or all of these failures may also be linked to use of a Tamiflu blanket. In addition to failures in detecting clade 2.2 in the above examples, the Tamiflu blanket may have led to the lab confirmation failures for the recent suspect cluster in Indonesia.
Details on collections histories relative to the start of Tamiflu treatment would be useful.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings