Blood Screening for Influenza

Influenza viruses, including highly pathogenic avian influenza virus (H5N1), could threaten blood safety. We analyzed 10,272 blood donor samples with a minipool nucleic acid amplication technique. Analytical sensitivity of the method was 804 geq/mL and 444 geq/mL for generic influenza primers and influenza (H5N1) subtype–specific primers. This study demonstrates that such screening for influenza viruses is feasible.

I n the 20th century, 3 infl uenza-related pandemics occurred (1918( Spanish infl uenza, 1957 Asian infl uenza, and 1968 Hong Kong infl uenza) (1), which are now known to represent 3 different antigenic subtypes of the infl uenza A virus: H1N1, H2N2, and H3N2. Major infl uenza epidemics show neither periodicity nor a predictable pattern, and all differ from one another. Evidence suggests that true pandemics involving changes in hemagglutinin subtypes are caused by genetic reassortment in animal infl uenza A viruses. Since 2003, the World Health Organization has reported the infection of ≈218 persons and 124 deaths (56.9%; as of May 23, 2006) caused by the (H5N1) subtype in 10 different countries; a probable person-to-person transmission of the avian infl uenza virus was suggested (2). Most countries predicted death rates of 14-1,685 persons per 100,000 population in the event of a pandemic and estimated that up to 2,707 persons per 100,000 population would become infected (3).
Our study demonstrates that screening donor blood for infl uenza A (H5N1) subtype or for infl uenza viruses in general by minipool nucleic acid amplifi cation technique (NAT) is feasible. To ensure the safety of blood products, this screening technique could be introduced into the blood-screening procedure without delay in the case of a pandemic.

The Study
To increase blood safety, we introduced minipool NAT screening in our blood donor service in 1997 for hepatitis B virus (HBV), hepatitis C virus (HCV), and HIV-1 and in 2000 for hepatitis A virus (HAV) and parvovirus B19 (4). For these purposes, 100-μL aliquots of up to 96 blood samples were pooled. The complete pool of up to 9.6 mL was centrifuged at 58,000× g for 60 min at 4°C. Viruses were extracted by using spin columns, and nucleic acid was eluted in a total volume of 75 μL. Only 60 μL of extract is needed for routine NAT screening. A residual volume of 15 μL can then be used for additional NAT testing (5) for infl uenza viruses.
The real-time quantitative amplifi cation of infl uenza/ H5 was performed according to the manufacturer's instructions (Artus Infl uenza/H5 LC RT-PCR Kit, QIAGEN, Hamburg, Germany) by using a thermocycler (LightCycler; Roche Applied Science, Mannheim, Germany). The test consists of 2 individual amplifi cation reactions. In the fi rst step, a generic infl uenza PCR is performed. The specifi city of this reaction was demonstrated for all subtypes of infl uenza A (H1-H15, N1-N9) and all subtypes of infl uenza B. Samples with a positive test result in the fi rst PCR were analyzed in a second PCR with infl uenza (H5N1)-specifi c primers and probes. Therefore, the assay allows differentiation between avian infl uenza (H5N1) and other infl uenza virus strains.
To mimic a situation like an H5-positive donation, a purifi ed culture supernatant of Vero cells infected with infl uenza (H5N1) (strain A/Thailand/1 (KAN-1)/2004) (6) was used as an external quantifi cation standard. Virion integrity in this preparation was confi rmed by electron microscopy. The viral RNA concentration was determined in an external laboratory by multiple quantitative real-time PCR determinations (7). Different dilutions of the external infl uenza (H5N1) subtype quantifi cation standard (0.0, 0.91, 1.96, 3.91, 7.81, 15.63, 31.25, 62.5, 125, and 250 PFU/mL) were prepared, and 100 μL of each dilution was spiked into 9.5-mL negative plasma pools. Each dilution was repeatedly spiked and tested in 8 minipools. Five microliters of the extract was analyzed with the generic infl uenza NAT as well as with the specifi c infl uenza (H5N1) NAT. Results are shown in Tables 1 and 2. Probit analysis of these data yielded a detection probability of >95% in parallel tests when an average of at least 13.4 PFU/mL (95% confi dence interval [CI] 8.3-184 PFU/mL) and 7.4 PFU/mL (95% CI 5.2-14.7 PFU/mL) for infl uenza generic assay and for the infl uenza (H5N1)-specifi c test, respectively, were present in individual plasma samples before pooling.
A total of 117 routine minipools, representing 10,272 blood donor samples, containing an average of 88 ± 8 samples per pool, had previously been tested for HIV-1, HBV, HCV, HAV, and parvovirus B19. All pools were negative for infl uenza virus when tested with the generic infl uenza PCR and the infl uenza (H5N1)-specifi c PCR. One pool had invalid results (failed amplifi cation of internal control RNA, representing 0.01% of all analyzed runs).  Table 2, the infl uenza (H5N1) subtype was detected by the generic infl uenza primers as well as by the infl uenza (H5N1)-specifi c primers when our routine minipool screening procedure was used. Sensitivity was expressed as PFU/mL and can be converted into viral genome copy number according the calculation of Yoshikawa et al. (7). Therefore, the analytical sensitivity was ≈804 geq/mL and 444 geq/mL for a generic infl uenza and for the infl uenza (H5N1) subtype, respectively.
After screening 10,272 samples by minipool NAT, none of the samples were found to be infected by infl uenza, which corresponds with the low EISS Index (European Infl uenza Surveillance Scheme index) of <20 during the study period (February-April, 2006) (10). An EISS index >80 is expected during an infl uenza epidemic, as was seen in 2005. Therefore, blood screening should be repeated during the next acute infl uenza season.
Accepted incubation periods for infl uenza range from 2 to 10 days (11,12). As with other viruses, a viremic phase of infection can be assumed to precede clinical symptoms such as fever (13,14). Recently Chutinimitkul et al. (15) detected infl uenza (H5N1) virus (3,080 copies/mL) in the plasma of a 5-year-old boy, which indicates a viremic phase of infl uenza (H5N1) infection. Those donors may be infective, especially to immunosuppressed patients. In addition to quarantine of infected patients, treatment with antiviral drugs, and development of avian infl uenza vaccines, blood donors should be tested during a pandemic to avoid transfusion-transmitted infections. Our study demonstrates that NAT screening could be incorporated into blood testing without delay and that the infl uenza virus could be suffi ciently enriched by centrifugation. Sensitivity of our infl uenza-screening method would have been suffi cient to detect recently reported virus concentrations in plasma of infected persons (15). However, as with all minipool methods, infections can be transmitted to transfusion recipients on rare occasions because the viremia level in the donor is below the analytical sensitivity of the screening assay.
To reduce this risk, a selective infectious dose NAT strategy (e.g., triggering of infectious dose NAT testing when at least 1 viremic donation is collected per week with the standard minipool screening algorithm), as performed for West Nile virus (WNV) screening in the United States might be necessary. Implementation of WNV-NAT in the United States in 2003 interdicted well over 1,000 donations from persons infected with WNV and is a good example of successful implementation of NAT screening for emerging viruses.
The collective fi ght against new viruses such as severe acute respiratory syndrome virus, WNV, or infl uenza (H5N1) presents an immense challenge for the whole community, but new molecular-biologic methods offer opportunities to overcome this challenge. NAT screening tests are now available soon after the sequencing of new viruses. In the absence of a general pathogen inactivation method for all blood products (erythrocytes, platelets, and plasma), the NAT screening procedure allows testing for new viruses to ensure blood safety.   9 3/8 37.5 0.0 0/8 0 *Influenza (H5N1) standard was extracted from 9.6 mL of 96 pooled donor samples after centrifugation. Five microliters of 75-μL nucleic acid extract was analyzed. The 95% detection limit was 7.4 PFU/mL; the 50% detection limit was 2.5 PFU/mL.