|Home||Founder||What's New||In The News||Consulting|
Paradigm Shift Intervention Monitoring
Unsubtypable CDC Reporting In FluView
Four human infections with novel influenza A viruses were detected in children from two states (Indiana (1) and Pennsylvania (3)). All four children were infected with swine-origin influenza A (H3N2) viruses; two were hospitalized; all four have since recovered from their illness. The case in Indiana did not report exposure to pigs prior to illness onset, although an epidemiologic investigation concluded human to human transmission was likely as a close contact reported direct contact with pigs prior to the child’s illness onset. The three cases in Pennsylvania were associated with attendance at a local fair where swine were exhibited.
The above comments from the first CDC FluView (week 40) for the 2011/2012 season note indirect linkage of the four trH3N2 cases with swine, but acknowledge that there is no direct linkage for the Indiana case, and is silent on direct contact for the three Pennsylvania cases. Although the three cases in Pennsylvania represent the largest number of confirmed human trH3N2 from a single location, and the flu gene constellation matches the Indiana case, the reporting of these cases in the sub-typing figure and underlying data has been inconsistent and confusing, raising serious questions about the status and reporting of other trH3N2 cases.
The inconsistencies in the 2011 trH3N2 cases follow delays in reporting 2010 cases. The clustering of 2010 cases began a year ago and these clusters were announced in a WHO pager alert for two cases, represented by A/Wisconsin/12/2010 and A/Pennsylvania/14/2010 (PA/14/10). However, at the time of the alert there was a second Pennsylvania case, A/Pennsylvania/40/2010 (PA/40/10) under review, who was infected less than a week prior to the Wisconsin case. However, PA/40/10 was initially reported as seasonal H3N2, and the report of this case as trH3N2 was delayed 5 months and reported in week 4 of 2011. Similarly, an Minnesota case (represented by A/Minnesota/11/2010, was reported shortly after the pager alert, and the index case was linked to additional symptomatic family members. The daughter of the index case subsequently serologically lab confirmed to have been trH3N2, presumably by her father or other family members since she had no reported linkage or contact with swine. However, reporting of this case was delayed 6 months, and testing of symptomatic family members was reported as inconclusive. Moreover, the CDC has used the isolate from the index case to develop a pandemic trH3N2 vaccine.
Concern due to the clustering of the 2010 trH3N2 cases was increased by the clustering of the 2011 trH3N2 cases. All four of the 2011 cases had the same constellation of flu genes. 5 of these genes were closely related to the dominant trH3N2 sequences. The PB1 was related to a prior triple reassortant outbreak at the Huron County fair and represented by two human isolates, A/Ohio/01/2007 and A/Ohio/02/2007. The NA gene in the 2011 isolates was closely related to the gene found in PA/14/10, while the MP gene matched the sequences in pandemic H1N1. The acquisition of the pandemic H1N1 MP gene increased concerns because the MP gene was deemed critical for the jump of pandemic H1N1 from swine to humans.
The presence of this rare constellation of flu genes in all four human 2011 isolates, and absence of this constellation in any swine isolates, strongly suggested that the 2011 trH3N2 was evolving and transmitting in humans. The CDC claim of no sustained transmission was blunted by the failure to identify a source for any of the four confirmed cases. The Indian case had no contact with swine, and although his caretaker did have contact, neither the caretaker no the associated swine had flu symptoms, and no evidence of swine influenza infections have been presented. Similarly, the three Washington County attended the same agricultural fair where swine was exhibited, but no symptoms have been reported for the swine at the fair and the sequences of the cases indicates the isolate from the Schuylkill resident, A/Pennsylvania/09/2011, were distinct from the other two cases, signaling two source linked to the fair, but the sequences from the two latter cases (A/Pennsylvania/10/2011 and A/Pennsylvania/11/2011) were virtually identical to the Indiana case, A/Indiana/08/2011.
The sequence data strongly suggests that the number of trH3N2 cases at the fair was markedly higher than the three confirmed cases, but detection of these cases is problematic because the H3 gene traces back to a seasonal H3N2 infection of North American swine in the 1990’s. This origin led to the delay in reporting the trH3N2 case, PA/40/10, which was initially reported as sesasonal H3N2. However, the 2011 H3 sequences have evolved further, and the initial Washington County fair cases were reported as unsubtypable, indicating this additional evolution had reduced the sensitivity of the human H3 sub-typing reagents for the 2011 trH3N2 cases. Thus, samples identified as influenza A positive, H3 negative would be likely trH3N2, which could be confirmed by the CDC with a PCR test specific for trH3N2. This scenario is similar to initial testing of pandemic H1n1, which were influenza A positive but H1 and H3 negative, which were then sent to the CDC for further testing because the states did not have the reagents for PCR testing for trH1N1. When the volume of samples to the CDC overwhelmed the testing, PCR kits for trH1N1 were then distributed to the states.
Currently, the PCR tests for trH3N2 have not been distributed and detection is dependent on states sending suspect cases to the CDC for further analysis. However, the recent reports in FluView have raised serious question about the state or CDC testing. The reports for PA/09/11 have been inconsistent. All sequences from this case have been made public at GISAID and Genbank, and the case was reported in the early release MMWR as well as the associated text in week 34 FluView, the unsubtypable categorization first appeared in the weekend following publication in these two source in the table of underlying data for the sub-typing figure. The week 33 collection was then reported in the table and figure in week 35 and week 36 reports, but was then absent in the week 37 and week 38 reports. It reappeared in the week 39 report, but as seen in the figure below,
it again disappeared in the week 40 report, which had data for the first week of the 2011/2012 seasonal as well as the data for weeks 21-39, which included the collection dates for all four 2011 samples positive for trH3N2.
The reporting of PA/10/11 was even more bizarre and inconsistent. The unsubtypable (sample collected in week 34) was first reported in the week 35 report and was again displayed in the week 36 report. However, it disappeared from the figure and table in week 37 and week 38 reports, and then reappeared in the week 39 table, but not in the week 39 figure, which is based on the week 39 table.
Moreover, the other two 2011 trH3N2 cases, from samples collected in week 30 and week 34, have not been represented in any of the tables or figures (the week 30 data has no reported unsubtypables or H3N2 cases).
The inconsistent and bizarre reporting of the unsubtypables raises concerns that these patterns are linked to the detection of additional trH3N2 cases or suspect cases, and associate changes in criteria for the representation of these cases in the sub-typing figures and tables in FluView.
These changes have been made without explanation, and a clear definition of criteria for reporting of unsubtypable / trH3N2 in the sub-typing figure and table is long overdue.