Dear FDA Dockets Manager:
AABB is an international association dedicated to advancing transfusion and cellular therapies worldwide. Our members include more than 1,800 hospital and community blood centers and transfusion and transplantation services as well as approximately 8,000 individuals involved in activities related to transfusion, cellular therapies and transplantation medicine. For over 50 years, AABB has established voluntary standards for, and accredited institutions involved in, these activities. AABB is focused on improving health through the advancement of science and the practice of transfusion medicine and related biological therapies, and developing and delivering programs and services to optimize patient and donor care and safety.
AABB appreciates the opportunity to comment on this draft guidance document. On behalf of the AABB West Nile Virus Task Force, the following comments to the draft “Guidance for Industry: Use of Nucleic Acid Tests to Reduce the Risk of Transmission of West Nile Virus from Donors of Whole Blood and Blood Components Intended for Transfusion and Donors of Human Cells, Tissues, and Cellular and Tissue-Based Products (HCT/Ps)” are submitted.
AABB plans to submit data for the 2008 WNV season to support the recommendations contained in this communication, and in response to the FDA request for data as published in the July 7, 2008 Federal Register.
The comments are arranged in the following format:
Section – language from draft guidance reprinted with page # and other identifying information.
Recommendation – recommendation with rationale.
Background – information supporting the recommendation.
Section – II. BACKGROUND, A. Whole Blood and Blood Components (page 3): Although year-round ID-NAT testing of all blood and blood components may not be currently feasible, we believe that ID-NAT on a limited basis during periods of high WNV activity to maximize the benefit to the public health is more practicable.
Section – III. RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, A.1. (page 4): We recommend that you screen year-round for WNV using a licensed NAT on donor samples of Whole Blood and blood components intended for transfusion.
Section – IV. RECOMMENDATION FOR TESTING OF HCT/P DONORS (page 11): We recommend that blood specimens from all HCT/P donors be tested year-round for WNV by ID-NAT using a licensed screening NAT test.
Recommendation – We believe FDA should allow the option to discontinue WNV NAT (as has been done in Canada (HemaQuebec)) during low incidence months (December 1 through April 30), after blood establishment computer systems have been validated to provide appropriate cGMP compliance when turning testing on and off. With regard to HCT/P donors we believe the appropriate language for the guidance document would be "using a licensed screening NAT test in accordance with manufacturers' instructions."
Background – We now have experience with year-round testing since the summer of 2003. The guidance language is explicit that FDA considers year-round testing appropriate and implies that year-round ID-NAT is a reasonable goal. We do not agree that either year-round ID-NAT or even year-round MP-NAT is appropriate. Targeted ID-NAT has already been adopted and refined voluntarily by the blood community and there have been only 2 transmissions from a single infected donor in the past 2 seasons that, noted in the guidance document, were associated with suboptimal communication of regional WNV activity.
Viremic donations identified between December 1 and May 1 are exceedingly rare and there are no reports of transfusion-transmission from missed donors during those annual intervals. For the 2006-2007 transmission seasons, there was a single confirmed donor the week of Dec. 2, 2007 (http://www.aabb.org/Content/Programs_and_Services/Data_Center/West_Nile_Virus/, accessed 24 June 2008). Clinical onsets of WNV neuroinvasive disease (NID) reported to CDC during the December to May interval constitute only 0.2% of recognized cases (personal communication Lyle Petersen, DVBID, CDC); of note there are on average 4 WNV NID cases for every viremic donation detected (Busch et al., Journal of Infectious Diseases, 2007), and consequently rare cases of NID reported between December and May should not override the absence of observed viremic donations during this period from over 5 years of continuous NAT screening.
The issues regarding or involving testing donors of HCT/Ps are more complex given that many of their recipients are highly immunosuppressed. We believe the risk from late fall to early spring is so low that this should be publicly discussed in the context of the US epidemiology of WNV and of acceptable safety. Obviously, when testing is done, it must be according to the manufacturer’s package insert. Current package inserts require ID-NAT, however this may change in the future.
Section – III, RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, A.2.b. (page 5): Prior to notifying the donor of his or her deferral and counseling the donor, we recommend that you perform additional testing on the specimen from the index donation using the same ID-NAT or using an alternate NAT with sensitivity equal to or greater than that of the screening assay. We also encourage you to test the specimen using a cleared test for antibodies to WNV.
If the repeat ID-NAT is reactive, we recommend that you notify the donor of his or her deferral and that you counsel the donor that he or she tests positive for WNV infection.
If the repeat ID-NAT is non-reactive but the test for antibodies to WNV is reactive, we recommend that you notify the donor of his or her deferral and that you counsel the donor that he or she tested positive for WNV infection.
Table 2 – Recommendation on Additional Testing if Blood and Blood Components (page 10): ***Due to the potential for false negative test results, encourage that donor return after at least 30 days for follow-up testing on a new specimen using ID-NAT and a cleared test for antibodies to WNV. Counsel the donor about their WNV status.
Recommendation – We would suggest that serology recommendations in the guidance specifically state IgM testing when needed (as noted below) and not the use of IgG testing. IgM is useful only (and should be performed) in the specific circumstance where the index sample does not repeat as NAT reactive.
We recommend, based on the availability in real time of repeat NAT, and the data as referenced and submitted below, that all initially reactive samples be repeated at least once by NAT, and if repeat reactive, the specimen be designated confirmed-positive. We further recommend that NAT follow-up sampling be optional based on the high sensitivity of index confirmation, and if follow-up sampling is performed, that no time frames be specified in the Guidance.
Background – A WNV confirmed-positive donor is defined as having repeat NAT reactivity using the same or alternate NAT assay, or reactivity in an IgM antibody assay; in both cases reactivity may occur in an independent sample from the index donation or from a follow-up sample (Stramer et al., New England Journal of Medicine, 2004; Busch et al., New England Journal of Medicine, 2006; Stramer et al., 2006 AABB abstract; Busch et al., Journal of Infectious Diseases, accepted 2008). Thus, this definition was included in the 2008 AABB Association Bulletin (#08-03, West Nile Virus – Revised Recommendations for Triggering Individual Donation Nucleic Acid Testing and Use of Communication Plans).
Together the use of repeat NAT and IgM will confirm 99% of all WNV confirmed-positive donors. According to data generated from 2003-2005 in a combined US blood donor study involving the American Red Cross (ARC), Blood Systems Inc (BSI) and the Roche users (Stramer et al., Transfusion, 2006, vol. 46, Supplement 2A) using the definition of confirmation that has been described (see above) and including only those donations with follow-up sampling (1559) to confirm further their WNV status, 1019 were confirmed positive. Of the 1019 confirmed positives, 1009 (99%) were confirmed by index sample results. This included 920 based on repeat NAT reactivity at index (sensitivity of 90%) and 239 having IgM reactivity at index (sensitivity of 23.5%). Within both groups, there were 156 donations that had both repeat NAT and IgM reactivity at index and a further 10 (1%) of the 1019 that required follow-up sampling for confirmation. Similarly, a recently published Canadian study (Tilley et al., Journal of Infectious Diseases, 2006, 206:193) reported on 276 persons in Alberta having a laboratory-confirmed case of WNV infection of which greater than 80% were West Nile fever cases. Of the 276 submitted samples, 191 were tested by both NAT and serology (defined as IgM antibody testing) with 69 (36%) detected by only NAT, 94 (49%) detected by only IgM and 17 (9%) detected by both NAT and IgM; 11 (5.8%) required a follow-up sample for confirmation. Therefore at index sampling, 86 (69+17; 45%) were detected by NAT, 111 (94+17; 58%) by IgM and 180 (69+17+94; 94.2%) by a combined algorithm of both NAT and serology. The number of individuals confirmed by NAT, and ultimately the total confirmed at index, was higher in the blood donor study (Stramer et al. AABB abstract 2006) than reported by the Canadian investigators (Tilley et al., Journal of Infectious Diseases, 2006, 206:193) likely due to the earlier sampling, prior to the onset of symptoms, that occurs in blood donors.
The IgG diagnostic test that is used in the US has a published false-positive rate in blood donors of 3% (Hogrefe et al., J Clin Micro, 2004). Donors with isolated low-level IgG reactivity have been shown to be uninfected based on the absence of other WNV reactivity at index or follow-up sampling including repeat NAT, using the same or alternate NAT assay, appearance of IgM, absence of evolution of IgG and absence of WNV-specific neutralizing antibodies by plaque reduction neutralization testing (PRNT). Donors undergoing WNV seroconversion demonstrate IgM reactivity within 4 days of peak viremia and then IgG seroconversion within another 3.4 days (Busch et al., JID accepted 2008). In this study of 245 followed donors, IgM reactivity was robust shortly after infection (up to 50 days) followed by IgG seroconversion with signal strengths greater than a signal-to-cutoff (S/CO) ratio of 3 far before the decline of IgM. Furthermore, with the increased use of ID NAT, the number of donors having initial, nonrepeatable NAT results with isolated IgG reactivity has increased. Data generated by the ARC (unpublished) demonstrated that from 2003-2008 there were 1003 donors confirmed as WNV infected using a combination of repeat NAT and IgM as discussed above. There were an additional 31 WNV initially reactive donors that were IgG-only reactive of which 20 (65%) occurred in 2007 when the greatest volume of ID NAT was performed. Although all the IgG-only-reactive donors were NAT initially reactive, 27 of 31 (87%) had low index NAT S/CO ratios (between 1-3) and all samples at index were repeat NAT nonreactive, IgM negative, and of 23 (74%) that were followed, none had any further WNV NAT reactivity or antibody evolution including 5 with IgG S/CO values above 3. All 20 IgG-only donations detected in 2007 were PRNT negative.
If follow-up sampling occurs, it should occur promptly since WNV RNA reactivity disappears rapidly (within 13-14 days; 99% followed donors cleared viremia within 50 days; Busch et al., JID, accepted 2008). In addition, since WNV symptoms develop soon after the appearance of viremia, if they develop at all, follow-up data would be most useful as soon as possible. As already discussed, between 94-99% of confirmation of WNV infection will occur by a combination of repeat NAT and IgM testing at index. A confirmatory algorithm requiring follow-up testing will never have 100% sensitivity in practice because not all donors will participate in follow up. In addition, the few true positive donors who would not be classified as confirmed positive based on index testing would already have been counseled for possible WNV infection, been deferred for 120 days, and components from their donations would have been quarantined; thus, there is no adverse impact on blood safety by not requiring follow-up testing.
Section – III, RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, A.2.b.iii. Note (page 5): In the event that the NAT screening assay does not discriminate between WNV and other Flaviviruses that belong to the JE serogroup (namely, Saint Louis Encephalitis virus, Japanese Encephalitis virus, Murray Valley Encephalitis virus and Kunjin virus), we encourage you to use a WNV-specific discriminatory NAT assay to assist donor counseling.
Recommendation – We disagree with the recommendation that a WNV-specific discriminatory NAT assay is necessary for donor counseling.
If there is utility in discriminating samples from donors detected by NATs with cross reactivity among these agents, it is a research and surveillance activity that is not appropriate to prescribe in FDA guidance intended to reduce the risk of transmission of WNV from donors.
Background – We do not understand what benefit “encouraging” this activity adds to donor counseling. First, only St. Louis Encephalitis Virus is present in the US, so the probability that we will confuse WNV with Kunjin, Murray Valley or JE is de minimis. The differentiation would be useful only during known or suspected simultaneous outbreaks. Over the past 7 years, public health data have not demonstrated simultaneous WNV and St. Louis Encephalitis Virus activity. If surveillance programs identify such events, at that time discrimination might be a useful epidemiological tool.
Treatment would not be offered to an asymptomatic donor under any circumstance. Treatment of a donor who becomes symptomatic after donation remains supportive at this time, so there would be no change in clinical management if discriminatory information becomes available. Furthermore, it is very unlikely that discriminatory results will be available in real time under any circumstance, so donor counseling would not be materially affected.
Section – III, RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, B.1. (page 6): We recommend that you convert from MP-NAT to ID-NAT if there is one (1) WNV NAT-reactive individual donation(s) from your region.
Section – III, RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, B.4. (page 7): If you wish to revert to MP-NAT testing, we recommend that you do so when the high WNV activity in the defined geographic area has subsided (for example, when a minimum of 7 days has passed without a single WNV ID-NAT-reactive donation).
Recommendation – We recommend that the guidance state that the trigger should be 2 presumed viremic donations (PVDs) in areas that lack any indicators of on-going WNV activity but should be 1 PVD in any geographic area having local WNV activity. Additionally, we suggest that triggering and de-triggering decisions be based on PVDs instead of initially reactive donations.
Background – The AABB Association Bulletin (#08-03) recommends that triggering and de-triggering decisions be based on PVDs – a classification that can be made prior to obtaining additional confirmatory results - rather than on initially reactive results. The use of PVDs as a rapid indicator of the confirmatory status of an initially reactive donation was initiated by the CDC in reporting WNV donor activity to ArboNET, and has been maintained as such from 2003 to the present. As defined by AABB (Association Bulletin #08-03) and by Stramer et al. (2006 AABB abstract), a PVD is an initially reactive donation that either repeats as reactive on the original donation sample or one that has a S/CO ratio > 17 (the latter may be applied to the Chiron Procleix™ WNV Assay; samples having a S/CO ratio < 17, or initially reactive using the Roche cobas TaqScreen WNV Test, must be repeated to determine if they are PVDs). We recommend that an initially reactive sample that is not a PVD not be counted toward triggering and de-triggering decisions; however, all such initially reactive donations should still be submitted for confirmatory testing.
Field experience with WNV donor screening using licensed lots of WNV reagent (Chiron Procleix™ WNV Assay) at the ARC (unpublished) indicate that the specificity of NAT reagent lots does not support the use of initially reactive results to make triggering and de-triggering decisions. Since licensure, the ARC has used 5 lots of WNV reagents on both the eSAS (semi-automated) and TIGRIS (automated) platforms. An indicator of assay specificity is the initial reactive rate of minipools (MPs) of 16 donations that do not resolve to a reactive individual donation, which the ARC tracks by lot, by laboratory and by platform used. The unresolved MP (false positive) rate using the eSAS platform has ranged between 0.00 and 0.42% by lot through May 31, 2008; the number of MPs of 16 donations tested per lot ranges between 2139 and 106,670 with the vast majority of testing having been performed using 2 lots having 0.14-0.19% unresolved (false positive) rates. The unresolved rates using the TIGRIS platform range between 0.01-0.15% with a minimum number of tests of 40,000 MPs performed on one lot with the remainder having MPs tested in the range of 108,000-172,000. These rates are consistent with those reported in the Gen-Probe package insert (500630, Rev A) for ID NAT at 0.09% (95% confidence interval of 0.04-0.14%). These false-positive rates strongly argue against continuing to perform ID NAT based solely on initially reactive results; regions that collect 200-1000 donations per day would be expected to generate at least one initially reactive, false-positive result in a 7-day interval, thus the use of an initial reactive criterion would lock them into prolonged ID NAT in the absence of any true yield from this procedure. Although the specificity of MP NAT is higher than that of ID NAT, with a reported false-reactive rate of 0.05% (95% confidence interval of 0.02-0.10% for the Chiron Procleix™ WNV Assay), this rate also indicates that there are likely to be false-positive triggering decisions. Therefore, we suggest that triggering and de-triggering decisions be based on PVDs instead of initially reactive donations. Data defining the sensitivity and positive predictive value (PPV) of the use of PVDs has been generated by the ARC (unpublished). From 2003 through the 2007 season, the PVD sensitivity for the 1003 confirmed-positive donors identified has ranged by year from 89-97%, increasing to 93-97% over the last 3 years with the use of a consistent IgM antibody confirmatory test. Similarly, the PVD PPV by year has ranged from 84-99%, and has increased over the last three years to 95-99%. Therefore the use of the PVD definition is both sensitive and highly predictive of a confirmed-positive test result. Of the 1003 confirmed-positive donors identified by the ARC, only 78 (7.8%) were not PVDs, and of those, 46 (59%) occurred in 2003 prior to the use of uniform triggering criteria at the ARC. Of the remaining 32 non-PVD, confirmed-positive donations in 2004-2007, at index all had nonrepeatable NAT reactivity, 29 of the 32 (91%) were IgM reactive and 27 of the 32 (84%) had IgG reactivity thus indicating that these donations had a low probability of causing infection in a transfusion recipient.
Triggering criteria according to AABB recommendations, contained in Association Bulletin #07-02, (West Nile Virus – Recommendations for Triggering Individual Donation Nucleic Acid Testing and Developing a Communication Plan), were to convert from MP to ID NAT when 2 PVDs and a rate of 1:1000 PVDs occurred for a defined geographic area over a 7-day rolling period, and that regions in overlapping and adjacent areas combine their data to establish “multi-site” triggers. Triggering using these approaches has been shown to be effective. Whereas 7 cases of WNV transfusion transmission were documented in 2003-2004 before the use of uniform triggering criteria by most of US blood centers, there were no transmissions in 2005 and only one transmitting donor (who transmitted to 2 recipients) in 2006. Following the use of multi-site triggers in 2007 facilitated by direct communication among neighboring blood centers and the use of the AABB Biovigilance Network, no further WNV transfusion-transmitted cases have been reported. We believe that the low rate of documented transfusion transmission is a direct result of the success of the volunteer triggering policies that have been in place in the US. However, to examine the sensitivity of the triggering policy in place in 2007 the ARC did a prospective study to determine the number of donations detected using a trigger of 2 PVDs and a 1:1000 rate as compared to a trigger of 2 PVDs or a trigger of 1 PVD (neither with a rate function) (Stramer et al., AABB abstract accepted 2008). Six regions located in WNV-endemic areas and ranging in collection volumes from approximately 200 to greater than 1000 donations per day were selected for study. Each region converted from MP to ID NAT based on a trigger of 1 PVD. During the time period of study, 136,388 donations were subjected to ID NAT with 30 WNV confirmed-positive blood donors identified whose donation samples required ID NAT for detection (i.e., MP NAT nonreactive). Of these 30 ID NAT-only reactive donations, 7 were IgM-antibody negative and likely infectious to recipients. The data show that, despite the absence of transfusion-transmitted WNV cases in 2007, 25 (83%) of the 30 WNV-infected donors (MP NAT nonreactive) who were detected by 1 PVD would have been undetected using a trigger of 2 PVDs and a 1:1000 rate; of these 25 donors, 5 were IgM-antibody negative. Use of 2 PVDs without a rate criterion would have detected 15 of the 25 donors including 4 of the 5 IgM-antibody-negative donors. Thus, adding a rate function to the triggering criterion was shown to decrease sensitivity (i.e., detection of only 5 of the 30 infected donors including only 2 of the 7 IgM-negative donors). During the period of ID NAT, an additional 56 donations were identified as false positive which accounted for 98% of all false positives for those regions during the 2007 mosquito season. Therefore, the use of 1 PVD as compared to 2 PVDs could increase sensitivity marginally but at a significant loss in specificity. Therefore, we recommend that the Guidance state that the trigger should be 2 PVDs in areas that lack any indicators of on-going WNV activity but should be 1 PVD in any geographic area having local WNV activity. We agree that once a defined region has triggered, ID NAT should be continued for a minimum of 7 days even in the absence of a PVD. Further, ID NAT should be continued beyond 7 days in areas with ongoing WNV activity: in these circumstances, continuing ID-NAT for 14 days should be considered. Our findings are supported by unpublished modeling data from the CDC (Petersen and Biggerstaff, manual script in preparation) who found that triggering on one PVD was more efficient than triggering on 2 and increasing the duration of ID NAT testing from 7 to 14 days increased effectiveness and efficiency. In this study, effectiveness was defined as the proportion of breakthrough infections prevented relative to MP NAT screening alone and efficiency was defined as the absolute reduction of breakthrough infections as a function of the number of tests performed. Their findings of differential sensitivities of various triggers were especially important in epidemics having an outbreak rate of 0.5-2 per thousand (during larger outbreaks, all triggering strategies were effective and during very small outbreaks, the use of seasonal triggers were more effective). It was also noted that effectiveness increased as blood center size increased beyond 560 donations per day, with the recommendation for small blood centers in overlapping geographic regions to communicate both numerator and denominator data. We believe that blood center communication is occurring effectively via AABB recommendations contained in Association Bulletin #08-03.
Section – III, RECOMMENDATIONS FOR DONATIONS OF WHOLE BLOOD AND BLOOD COMPONENTS, B.3. (page 7): If you obtain WNV NAT-reactive test result(s) for individual donation(s) more than 24 hours after collection of the individual donation(s), we recommend that you consider retrospective ID-NAT testing of retained samples from donations collected between the date of collection of those donations whose test results caused the threshold for implementing ID-NAT to be met or exceeded and the date of actual ID-NAT implementation.
Recommendation – We recommend changing the MP to ID NAT conversion period from 24-hours to 48-hours.
Background – The AABB Association Bulletin (#07-02) recommended conversion from MP to ID NAT within 24 hours of obtaining the test results that were the threshold for triggering using the same verbiage as now is included in the draft Guidance. If the 24-hour period from conversion to ID NAT was exceeded, AABB recommended retrospective testing to the time of the triggering event (collection date). Based on feedback from the industry regarding the logistics of sample receipt especially from remote areas, time required for initial and repeat testing, and the impact of retrospective testing on components that have been released for transfusion (some of which may have been transfused for which there is no treatment other than supportive), AABB recommended that 48-hours be used as the time period for the conversion from MP to ID NAT for the 2008 season (Association Bulletin #08-03). In consultation with the CDC (Petersen and Biggerstaff) and using models that have been developed (unpublished), the risk of not detecting a WNV PVD in this 24-48 hour interval is less than 1 per season for the entire US.
Questions concerning these comments may be directed to Joseph L. Giglio, Deputy Director, Regulatory Affairs, AABB firstname.lastname@example.org
Roger Y. Dodd, PhD
Chair, AABB West Nile Virus Task Force