Commentary
Human H7N2 Sequence Confusion and Delays Raise Concerns
Recombinomics Commentary 14:34
May 28, 2008
Five months after the patient checked into Westchester Medical Center complaining of fever and cough, no one can say how he was infected with avian influenza. The man recovered and went home after a few weeks, but it was not until a month ago that the federal Centers for Disease Control and Prevention suspected an avian virus had caused his illness, and only last Friday that the centers confirmed that diagnosis.
"We can't figure out how he was exposed and why he's an isolated case," said Nancy J. Cox, an influenza expert at C.D.C. "We need to understand how he got infected."
Officials said the man was infected with Type A influenza, Strain H7N2, the same one that hit chicken farms in New Jersey, Maryland and Delaware this year.
The Westchester patient, a Caribbean immigrant, lives in Yonkers with his wife and children, officials said. (Hospitals and health officials do not reveal the names of patients in cases involving public health issues.) He entered the hospital in November suffering from other serious ailments that weakened his immune system and that might have masked the symptoms of avian flu. One official said the patient had symptoms of a respiratory illness, including coughing and an abnormal chest X-ray. Doctors at first suspected tuberculosis.
"We knew it was something weird, but we didn't know what it was," said Claire Palermo Flower, spokeswoman for the hospital. "They did an elaborate culture and asked the lab to do more than the usual tests."
It was not until February that C.D.C. tested the sample, when scientists there found that the virus was not from the H1 group, Dr. Cox said. A subsequent test ruled out another family of flu viruses, Type B. Further testing showed that it was Type A, but not the H1, H3 or H5 subtypes.
Finally, on March 17, scientists using other tests identified the virus as H7N2. The next day, Dr. Cox said, C.D.C. notified health officials in New York that they had a suspect human case of avian flu. To be certain that the sample had not been contaminated in a laboratory, they did further tests.
Doctors asked the patient for another blood sample, to compare antibody levels in it with another sample kept from the initial phase of his illness. Last week, the tests confirmed a recent infection with H7N2, and the C.D.C. alerted state and local officials in a conference call on Friday.
The above comments from a 2003 New York Times article provide some of the background for the H7N2 case from 2002 that was diagnosed in 2003. The case is back in the news this week because of a PNAS paper which demonstrated increased binding for human alpha 2,6 receptors, coupled with evidence of enhanced transmission in a ferret model.
H7 infections in humans has been a cause for concern for some time. Following the infection in the above case, there was a major H7N7 outbreak in the Netherlands in 2003. 89 cullers had evidence of H7 infections and a veterinarian died. Most of the cases were mild and involved eye infections. However, a follow-up studied identified H7 antibodies in over 1000 contacts, indicating the H7 was readily transmitted.
That outbreak was followed by several H7 outbreaks in Europe and North America involving H7N3 and H7N2. Most of these outbreaks were linked to human cases, and an H7N2 infection In England involved respiratory disease. Like the New York infection in 2002, the H7N2 infections in humans were hard to confirm. Although there were more suspect human cases than confirmed poultry cases, most infections were not confirmed, although flu-like conditions were present in contacts, including owners of affected holdings.
These confirmation failures raise serous concerns because most cases would be mild and not easily distinguished from seasonal flu. Moreover, the recent results on receptor binding domain studies and transmission in a ferret model raise concerns of widespread undetected human infections.
These concerns were increased by the publication of the sequences from the isolate from the New York patient, A/New York/107/2003(H7N2). The sequences represented a clear reassortant, with three avian genes (H7, N2, NP), and four human H3N2 genes (PB1, PA, MP, NS). The sequence of the PB2 gene was not released.
The four year delay in the publication of these sequences, as well as the lack of any discussion of the reassortant nature of the constellation of genes, is cause for concern. These concerns were increased by the removal of the sequences from the public database shortly after submission and acceptance of the PNAS paper.
Although the sequences were labeled with a “removal” message, the seven gene segments are still available, and the HA sequence matches the description in two peer reviewed publications, including the PNAS paper which became public on Monday. The sequence has an 8 amino acid deletion proximal to the receptor binding domain, coupled with an adjacent non-synonymous change, S213N, which has been conserved in more recent H7N2 poultry isolates from New York, alomg with three basic amino acids at the HA cleavage site.
The delay in diagnosis in 2002/2003, coupled with the delay in releasing the sequences and the lack of discussion of the sequences, are causes of concern, as was the recent removal of the sequences.
At this time it is unclear if the sequences are in error, or if they were removed because of publication issues.
An explanation and clarification are long overdue.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings