H5N1 Human to Human Tranmission in Pakistan
Recombinomics Commentary
December 18, 2007
[Another version of events with the added information that the 2 deceased brothers and an H5N1-positive man and his H5N1-positive niece all lived in the same town (Abbottabad), and all worked on the same farm (previously located in Mansera). Circumstances consistent with infection from a common source rather than by person-to-person transmission of virus. - Mod.CP]
The above Promed commentary on the recent update of the H5N1 cases in Pakistan, is among the most curious to date. The chief data point that distinguishes a cluster due to a common source verses human-to-human (H2H) transmission is disease onset date. Clusters of clusters do not reduce the likelihood of H2H transmission. Instead they increase the likelihood, especially for an agent like H5N1, which historically is not efficiently passed to humans. The presence of two familial clusters in the same geographical area signals a more efficient transmission of H5N1 to humans, and the clustering within families raises the possibility of H2H, but the key determinate is the time between disease onset dates.
The above commentary, like the earlier comments from Pakistan suggesting that sequence data would support a common source, fails convey the key parameter in a transmission chain, which is a gap in disease onset dates. The gaps are due to the incubation time between an infection and symptoms. For H5N1, which is not efficiently transmitted, there is an additional period required for H5N1 titers to rise and for exposure of the contact.
Most of the H2H transmission to date for H5N1 has involved family members who were infected while caring for a relative, or a health care worker, while caring for a patient. This close contact creates a transmission opportunity because contact is over an extended time period, and H5N1 can be found in several bodily fluids.
Most H2H H5N1 clusters are small and involve only two family members. There is no absolute proof for a small cluster of two, but when there is a 5-10 day gap between disease onset dates, H2H is likely. These likely transmissions are not absolute, because there frequently is a potential common source. However, infections due to common sources usually do not have a large gap between onset dates because most patients show symptoms 2-4 days after exposure.
When there are multiple transmissions involving three or more patients, or when the contact is a health care worker, H2H is more likely. In the past there have been multiple clusters of three or more cases, which were relatives or health care workers, but the chains have not clearly extended beyond H2H2H.
In the Pakistan clusters, disease onset dates have not been released. However, the index case for the larger cluster was likely infected during a cull on October 21-23. Media reports indicated his disease onset date was October 25. Four of his brothers were either H5N1 confirmed or died with H5N1 symptoms, strongly suggesting that he infected them because the two that died were students and did not participate in the cull, and another H5N1 confirmed brother was a teacher.
Moreover, the dates of death have been reported as November 19 and 29. so the gap between the dates of death, as well as the extended time after the disease onset date for the index case support H2H or H2H2H. Moreover, reports of lab confirmation in a health care worker raises the possibility of H2H2H2H, which would be the longest transmission chain reported to date. Verifying this chain however, requires disease onset dates, contact dates, and confirmation of infections, although the absence of a sample from one of the deceased brother should not exclude him from the chain.
The details on the laboratory confirmed cases would be useful. Evidence reported to date strongly supports some H2H transmission, clusters of clusters, and sequence data, not withstanding.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings