WHO/BS/10.2137 ENGLISH ONLY
EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION Geneva, 18 to 22 October 2010
Report of a WHO collaborative study to assess the suitability of a candidate International Standard for antibody to pandemic H1N1 influenza virus
John Wood and Alan Heath
National Institute for Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Herts EN6 3QG, UK
© World Health Organization 2010
All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail:
bookorders@who.int). Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e- mail: permissions@who.int).
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned.
Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. The named authors alone are responsible for the views expressed in this publication.
Abstract
Haemagglutination-inhibition (HI) and virus neutralisation (VN) serology assays are being used to assess immunogenicity of many pandemic H1N1 (H1N1pdm) vaccines in clinical trial throughout the world, yet the assays are poorly standardised and previous studies have shown that results of such assays to be variable between different laboratories. A freeze dried candidate International Standard (IS) for H1N1pdm antibody was prepared from pooled sera of subjects who had either recovered from recent pandemic influenza infection or had received an
adjuvanted subunit pandemic H1N1 vaccine prepared from the reassortant virus NYMC X-179A (derived from A/California/2009 virus). Ten laboratories from seven countries tested the
candidate IS 09/194 and a panel of human sera from convalescent subjects and recipients of A/California/7/2009 pandemic vaccine; ten using HI assays and nine using VN assays. A negative serum was also included in the panel. The tests were performed on at least three occasions using the A/California/2009 virus (6 laboratories) and/or the reassortant virus NYMC X-179A virus (ten laboratories). For tests of antibody to NYMC X-179A, the % geometric coefficient of variation (%GCV) for 09/194 between laboratories was 83% for HI and 192% for VN. For tests of all sera, the median % GCV ranged from 95-345% for HI and 204-383% for VN, but for the titres relative to 09/194 the median % GCV was much reduced (HI 34-231%;
VN 44-214%). For tests of antibody to the A/California/2009 wild type virus there were similar reductions in % GCV when 09/194 was used.
The results demonstrate that the candidate IS 09/194 will be of use in assays of antibody to A/California/2009 virus vaccines and it is proposed that 09/194 be established as the first IS for antibody to A/California/2009 virus with an assigned potency of 1300 International Units per ampoule ie 1300 IU/ml when reconstituted as directed with 0.5 ml dH2O.
Introduction
There are many pandemic H1N1 vaccine clinical trials in progress or completed in the world and the various vaccines will be evaluated by two serological techniques: virus neutralisation (VN) and haemagglutination-inhibition (HI). Recent WHO collaborative studies to evaluate influenza H3N2 and H5N1 serological tests have found a high degree of variability between VN and HI results from different laboratories. However the use of a standard serum reduced the variability significantly [1,2]. Difficulties in comparing VN and HI titres include the lack of standard protocols, differences in sensitivity and specificity of reagents and assay variability, all of which will create an obstacle for global assessment of H1N1pdm vaccine immunogenicity. It is likely that an International Standard (IS) for H1N1pdm antibody will be of great benefit.
A proposal by NIBSC to establish an IS was made in July 2009 The aims of the collaborative study are:
• To evaluate the suitability of a freeze-dried serum pool to serve as an IS for antibody to H1N1pdm virus with an assigned potency in International Units per ampoule for use in HI and VN assays. There is no international conventional reference measurement procedure and this measurand is not traceable to International System of Units (SI) of quantity.
• To assess the candidate IS in terms of reactivity in the conventional HI and VN assays used to detect H1N1pdm antibody.
• To assess commutability ie to establish the extent to which the IS is suitable to serve as a standard for the variety of different samples being assayed.
Materials
Candidate International Standard code 09/194
This is a batch of ampoules containing a freeze-dried pool of serum obtained from:
• 22 human subjects who had received Novartis MDCK cell grown inactivated subunit influenza vaccine containing the H1N1pdm strain NYMC X-179A (A/California/7/2009) in the presence of MF-59 adjuvant. This trial took place at the Leicester Royal
Infirmary, UK under the direction of Dr Iain Stephenson.
• 9 NIBSC staff and their partners who had recently recovered from pandemic influenza infection.
Subjects gave informed consent for the use of their serum, which was approved by the ethical committees of Leicester Royal infirmary and NIBSC respectively.
The donations were tested and found negative for HBsAg, antibodies to HIV-1, 2 and HCV RNA by nucleic acid amplification tests. After being tested and found positive for antibody to NYMC X-179A virus by HI and VN, the donations were pooled and the pooled serum (volume 568ml) was filled in ampoules and freeze-fried at NIBSC on 17 September 2009 following documented procedures. At the end of processing, the ampoules were back-filled with Nitrogen. This was a 0.5g fill weight with a mean filled weight of 0.5179g (contents of 82 ampoules weighed) and mean dry weight of 0.0455g (contents of 6 ampoules weighed). The coefficient of variation (CV) for filled weights was 1.034%. 1256 ampoules were filled and 1135 were available for issue.
Residual moisture on 12 samples was 0.208µg with a CV of 10.65% and oxygen headspace measured in 12 ampoules of 0.30% with a CV of 48.84%. The ampoules were stored at NIBSC at -20oC. Currently (16 July 2010) there are only 331 ampoules of 09/194 available for use since it has been issued following announcement by WHO on 19 October 2009 [3]. There are plans to prepare a replacement candidate IS.
Additional coded study samples
In order to demonstrate that the candidate IS is suitable for use in a variety of situations; the following antiserum preparations were included as coded samples along with the candidate IS:
Sample code Description
D Human serum negative for antibody to H1N1pdm A H1N1pdm convalescent human serum - low titre
B Post vaccination serum (Novartis MDCK subunit H1N1pdm vaccine +MF59) - medium titre
F Post vaccination serum (Novartis MDCK subunit H1N1pdm vaccine +MF59) – high titre
C Pool of human serum before freeze drying as 09/194
E Duplicate of B
Viruses supplied
If needed, participants were supplied with the following H1N1pdm viruses:
• A/California/7/2009
• NYMC X-179A reassortant derived from A/California/7/2009 (used for most pandemic H1N1 vaccine production)
Participants were expected to grow their own stocks of virus in readiness for serology tests
Design of study
Participants were requested to store sera at -20oC and viruses at -70oC until use. Participants were asked to reconstitute the candidate IS 09/194 with 0.5 ml distilled water and then to test 09/194 and the coded study samples for antibody to the two H1N1pdm viruses concurrently, on three separate occasions. It was recommended that wherever possible, a new ampoule of 09/194 was used for each assay. An excel spreadsheet was supplied for reporting of the raw data from each test.
Statistical methods
The results for the HI and VN tests were provided by participants as absolute titres (reciprocal end-point serum dilutions). Within each laboratory, replicate titres were combined as geometric means. Where a titre was negative, or below the starting dilution, a value of half the starting dilution was assigned (e.g. a titre of <10 was taken as 5) to allow calculations of geometric means. Similarly, for values reported as greater than a final dilution, a value of twice the final dilution was assigned (e.g. >1280 was taken as 2560). Overall mean titres were calculated as the geometric mean of the individual laboratory means. Variation between laboratories was
expressed as the percentage geometric coefficient of variation (%GCV).
To assess within laboratory reproducibility, a comparison was made of replicate tests in each laboratory and calculating the percentage of tests giving greater than 2-fold or greater than 4-fold variation in titres for all study samples. Each laboratory was requested to test the 7 sera using the two viruses, giving 14 sets of replicate titres. The calculations were also performed for the human sera A, B, F, C and E and 09/194 alone.
To assess the effect of using a common serum standard, titres were expressed relative to the candidate standard 09/194, by taking the ratio of the mean titre for a sample to the mean titre for
09/194 obtained by a laboratory, and multiplying by an “assigned” value for 09/194. For the purposes of this report, the assigned value for 09/194 was the overall mean titre for HI and VN assays. However the impact on the between laboratory agreement, and the %GCV, is
independent of the units chosen.
The distribution of geometric mean titres between laboratories for each sample is also displayed in histogram form, where each box represents the mean obtained by an individual laboratory, and is labelled with the laboratory code number. Where a laboratory obtained a negative result, the results are plotted as “-ve” (ie negative) in a position equivalent to a titre of 5. This was to distinguish between a negative <20 and a positive 10 from different laboratories. Results using different viruses are indicated where appropriate.
The relationship between HI and VN test results within laboratories was assessed by plotting the log HI titre against the log VN titre across study sample for each individual laboratory, and calculating the correlation coefficient.
Participants
Although fifteen participants were invited, only ten participants from seven countries took part in the study. Participants were selected for their previous experience of influenza serology, for their public health importance and for geographical distribution. The participants are listed in
Appendix 1 and were randomly assigned a laboratory code number, not necessarily in the order listed.
Assay methods
As there are no established standard protocols for HI or VN assays, participants were requested to use their in-house methods. A checklist was supplied to each participant to indicate the key parameters used in their methods. One of the benefits of an IS would be to reduce errors associated with protocol variation.
VN protocols
Six of 9 (66%) laboratories returning VN data also supplied checklists (table 1). VN protocols could be generally grouped into two methodologies.
(i) Cell suspension method and short incubation time of assay to endpoint (<26 hr): used by labs 4, 5, 6, 9 and 10
(ii) Preformed cell monolayer and long incubation time of assay to endpoint (≥3 days): used by lab 8
Another variable was the method for calculating the starting dilution, which affects the VN titre. Labs 5, 6, 9 and 10 (virus #2) used the volume of serum in diluent as the starting dilution whereas lab 8 and 10 (virus #1) used the volume of serum in diluent plus the volume of virus as the starting dilution.
HI protocols
Six of 10 laboratories (60%) returning HI data also supplied checklists (table 2). HI protocols were generally similar. The variable parameters considered to be critical are:
(i) Red cell type: turkey rbc (used by labs 4, 5, 9 and 10) or chicken rbc (used by labs 3 and 8)
(ii) Rbc concentration ranged from 0.4% (lab 8) to 0.7% (lab5)
(iii) Virus added was either 4 Agglutination Doses (Ads) (labs 3, 4 and 9) or 8Ads (labs 5, 8 and 10)
(iv) Adsorption of sera with packed rbcs to remove nonspecific agglutinins (lab 4)
Results and data analysis
Data Received
Study samples were sent to 15 laboratories. Data were received from 10 laboratories. One laboratory only returned data from HI assays, while the other nine performed HI and VN assays.
All laboratories that returned data used the NYMC-X179A virus (X-179A). In addition six of the laboratories also provided data for HI assays using A/California/7/2009 (A/California). Five of these six provided data for VN assays using A/California. Laboratory 10 provided two sets of VN assay results, described as using viruses from different sources (virus #1or #2, both A/California and X-179A). The definition of titre appeared to differ, depending on whether addition of virus was considered to be part of the dilution. The following explanatory notes were sent by participants:
Laboratory 10
1) VN virus # 1 - initial serum dilution is 1:10 and the final titre does not take into account the addition of virus (final titres 1:10 to 1:5120). The fifty percent neutralizing titre (NT50) is calculated using the Karber formula from duplicate tests performed on the same plate (two rows per sample).
2) VN virus # 2 - initial serum dilution is 1:5 and final titre does take into account the addition of virus (final titres 1:10 to 1:5120). The titre (geometric mean titre) is calculated from titres from two independent tests of each sample (singlets on two different plates within the same assay).
3) HAI - initial serum dilution is 1:5 and final titre does take into account the addition of virus (final titres 1:10 to 1:5120). The titre (geometric mean titre) is calculated from the titres from three replicates of each sample on the same plate.
The difference in definition of what constitutes a “dilution” (ie whether it is just serum plus diluent or whether it is serum plus diluent and virus) will add to the variability between
laboratories. There is no accepted convention so this is an added justification for using a standard Laboratory 5 reported problems with their third VN assay. They commented:
“We are getting highly variable virus control and background read outs and assay 3 was invalid due to high background and virus control plate responses.”
The third VN assay from laboratory 5 was not included in subsequent analysis.
Results
The VN results from laboratory 10 using the two different supplies of X179A and protocols for definition of titre (see above) showed differences of around 1.5-fold to 2-fold. The laboratory geometric mean titres (GMTs) are shown in table 14. For subsequent analysis, both set of results were combined, giving single VN GMTs for each sample for this laboratory (tables 4 and 5).
Assay reproducibility – replicate assays:
The within laboratory reproducibility was assessed by comparing the replicate titres within each laboratory for each sample across replicate tests. For laboratory 10, this was done before
combining results from the two protocols. The percentages with a range greater than 2-fold, and greater than 4-fold, are shown in table 3. The reproducibility of the HI assays is good for the majority of the laboratories, with only laboratory 5 having any replicate tests differing by more than 4-fold (8%), and few laboratories with replicate tests differing by more than 2-fold, except laboratory 5 at 50%. The reproducibility of the VN assays appears to be slightly weaker than for HI in some laboratories, but in most laboratories the reproducibility is good. In four laboratories, there were no replicates that differed by more than 2-fold.
Assay reproducibility – duplicate samples:
Samples B and E were duplicates of the same medium titre serum. When the laboratory GMTs for samples B and E were compared, there were only two cases where the GMTs differed by more than 2-fold. These were for laboratory 12 (HI assay: 5-fold difference) and laboratory 8 (VN assay: 2.2-fold difference). Comparing titres of samples B and E within individual assays, none of the HI assays, with the exception of laboratory 12, had titres that differed by more than 2-fold. For the VN assays, laboratories 5, 6, 8, 11 and 12 each had a single assay where titres differed by more than 2-fold. There were no assays where titres differed by more than 4-fold.
Laboratory 12 had a consistent difference in HI results between sample B & E (overall GMTs of 508 and 2560 respectively; table 4). The reason for this is not clear. Apart from this, there was good reproducibility within assays for the duplicate samples.
Absolute Titres:
All laboratories returned negative results for sample D, an H1N1 negative sample.
The laboratory GMTs for the other samples are shown in table 4 for the HI assays, and table 5 for the VN assays. The overall GMT (Geometric mean of the laboratory GMTs) and the between laboratory percentage geometric coefficient of variation (%GCV) are also shown. Results obtained with X179A and A/California viruses are shown separately. All laboratories obtained positive results for all samples (except D), with the exception of the VN assay from laboratory 13, with A/California, which had a negative result for the low-titre sample A.
The laboratory GMTs for the different samples are also shown in histogram form in figure 1.
Each box represents the GMT from an individual laboratory, and is labelled with the laboratory code number. Results obtained with X179A and A/California viruses are shown separately, with the latter shaded.
The histograms illustrate that there was considerable variation between the GMTs from different laboratories. This is also reflected in the high %GCVs (tables 4 and 5, and summarised in tables 6 and 7), and in the fold range (max/min) between labs for the different samples, summarised in table 8. The range (max/min) for the HI GMTs is between 8-fold (samples A & C) and 80-fold (sample E). For the VN assays, the range is between 31-fold (sample C) and 109-fold (sample E). These figures are for results using X179A.
The overall GMTs for the candidate standard 09/194, based on assays using X179A, were 1:183 for HI assays and 1:516 for VN assays.
Comparison of Absolute Titres: X179A and A/California:
From the tables and histograms, there is a suggestion that A/California gives generally lower titres than X179A. It should be noted that the laboratory with the highest titres (laboratory 9) only used X179A however.
A paired test, comparing the results with the two viruses in those laboratories that provided results with both viruses, based on geometric means of the GMTs for all positive samples, found a significant difference (p=0.03) for the HI assay, with an overall 50% increase in titre for X179A over A/California. For laboratories 4 and 5 however, the differences were very small or not apparent. For the VN assays, there was no significant difference (p=0.37). Although there was a general trend for an increase in titre for X179A over A/California (33%), there was variability between laboratories, and in particular, laboratory 6 obtained higher titres with A/California than with X179A (table 5 and figure 1).
Use of 09/194 as a standard:
To assess the value of expressing results relative to a standard serum, the titres for samples A, B, C, E and F were expressed relative to candidate standard 09/194, using a hypothetical assigned titre of 160 for HI and 640 for VN. These were the closest rounded dilution steps to the overall GMTs for 09/194 of 183 for HI assays and 516 for VN assays using X179A.
To convert an observed absolute titre to a relative titre, the value is scaled by the ratio of the assigned titre for the standard to the observed titre for the standard. For example, if the standard has an assigned titre of 160, and an assay gives titres of 640 for sample X and 320 for the standard, then the titre of X relative to the standard is 320 ( = 640 x 160/320).
The resulting relative GMTs are shown in tables 9 & 10, and in histogram form in figure 1. The summaries of the between laboratory %GCVs are shown in tables 6 & 7, and the fold-range between laboratories is shown in table 8.
It is clear from the tables and figures that in the majority of cases there is a reduction in the between laboratory variation when titres are expressed relative to a standard. One instance where there is not a reduction in %GCV is in the HI assays for sample F, with X179A. It can be seen from figure 1 that the main source of between laboratory variability is a single laboratory (lab 12) with a particularly high result for sample F, which is not fully accounted for by expressing the titre relative to 09/194.
The most marked reduction in variability is for sample C. This sample was the liquid material that was used to produce the freeze-dried candidate standard 09/194 (ie close homology between test and standard). Calculating the titres for sample C relative to 09/194 would therefore be expected to produce the “best case” agreement.
Relationship between HI and VN assays:
To investigate the relationship between the HI assay and the VN assay, “conversion factors”
were calculated for each laboratory and sample, as the ratio of the VN titre to the HI titre, using results with X179A. A conversion factor of greater than 1.0 indicates that the VN assay gave a higher titre than the HI assay. Table 11 shows the conversion factors calculated for each laboratory and sample. Although in general it appears that the VN assay gave higher titres than
the HI assay, there is considerable variation between samples, and particularly between laboratories. Based on the geometric mean conversion factors for laboratories, laboratory 10 obtain VN titres more than 10-fold higher than HI titres, but laboratories 8 and 12 have lower VN titres than HI titres.
An equivalent analysis was carried out using the titres relative to 09/194. The calculated conversion factors are shown in table 12. There are still differences between laboratories and samples. The differences between laboratories have been reduced by basing the conversion factors on relative titres, but conversion factors still range from around 3 to around 12.
The relationship between the HI assay and the VN assay is not sufficiently consistent across different laboratories to determine a “universal” conversion factor, which would be appropriate for all laboratories.
The relationship between HI and VN assays within individual laboratories is shown in figure 2, with the log HI titre plotted against the log VN titre, for all study samples. The associated correlation coefficients are shown in table 13, along with the results of tests for significance of the correlation.
There was a significant correlation between HI and VN titres of the five test sera and 09/194 for all laboratories except 5, 6 and 13. For these three laboratories, and to a lesser extent for some other laboratories, the poor correlation is due to the results for the duplicate samples B and E (post-vaccination sera), which had much higher VN titres than HI titres. The results for the other study samples showed good correlations. It may be relevant that laboratories 5 and 6 were among those with weakest within-laboratory reproducibility.
Stability studies on 09/194
Stability studies on ampoules stored at different temperatures are underway at NIBSC, UK using X-179A virus in order to predict the estimated loss of activity when stored at the recommended storage temperature of -20oC. Ampoules have been stored at -70oC, -20oC, +4oC, +20oC, +37oC and +45oC and after 10 months they will be selected at random for assay.
Discussion and conclusions
There was considerable variation observed between participating laboratories in the level of titres obtained for the same serum samples. Differences of up to 80-fold between laboratories were observed for HI and up to 109-fold for VN tests. This level of variability is consistent with that seen in previous collaborative studies [1,2]. For most of the laboratories, there were higher titres for tests using the reassortant X-179A virus than with the wild type A/California virus.
Expressing titres relative to the candidate IS 09/194 gives a reduction in the variation between laboratories for tests with both A/California and X-179A viruses in nearly all cases. Exceptions were HI tests of the high titre serum F where laboratory 12 had a particularly high result and VN tests of the low titre serum A where laboratory 13 was very low or negative. These extreme results were not brought into better agreement with other laboratories by use of 09/194.
The relationship between the HI titres and the VN titres appears dependent on the laboratory, and to a lesser degree on the serum being tested. Taking means across all laboratories, the conversion factor varied from1.5 for low titre serum sample A and 9.1 (VN 9.1-fold higher than HI) for
medium titre serum sample B. Nevertheless the correlation between HI and VN was generally good within each laboratory. Where there was a poor correlation, it was due to the results for duplicate samples B and E, a post-vaccination serum. For the HI test, a ‘protective’ titre of 1:40 is widely used to assess results but, based on results of this collaborative study, it does not appear appropriate to assign an equivalent VN titre that would be applicable to all laboratories.
The use of a common human serum standard for pandemic H1N1 antibody assays will improve agreement between laboratories, but the designation of unitage to 09/194 needs some
consideration. It is usual to assign an arbitrary potency, in International Units, to an IS but should this be assigned in isolation or should it be linked to the previous IS for antibody to H5N1 clade 1 antibody (1000 IU)? The units of either H5N1 or H1N1pdm IS will be used to harmonise results from different laboratories but they may also be used to predict a measure of protective immunity. This is conventionally done for the HI assay where an HI titre of 1:40 is accepted to indicate protective immunity in 50% of subjects. If the IU values of both H5N1 and H1N1pdm IS are linked by HI or VN titres there would be some continuity in serology assays, which would help in regulatory acceptability of data. On the other hand there is a case for not linking the two antibody standards as it is not known whether antibody to H5N1 is more or less protective than the equivalent level of antibody to H1N1pdm virus. Due to the pandemic situation, 09/194 has been prepared, evaluated and distributed before evaluation by ECBS could take place and it has been assigned a consensus HI titre of 1:183 and VN titre of 1:516 (obtained by using X-179A virus) so that users could adjust their results accordingly. The titres using X-179A and not the wild type A/California virus were used as it is envisaged that most vaccine clinical trials and possibly some serum surveys would be done with the reassortant virus.
Stability studies are underway and results will be reported to ECBS when available.
Proposal
09/194 should be established as the 1st IS for antibody to influenza A/California/7/2009 (H1N1)pdm virus with an assigned potency of 1300 IU per ampoule (1300 IU/ml on
reconstitution). This potency has been linked to the potency of the 1st IS for antibody to H5N1 antibody clade 1 07/150 using the HI titre of 07/150 (1:140) and the HI titre of 09/194 (1:183), to give equivalent conversion factors from IU to consensus HI titre for both standards It is also proposed that the consensus HI titre of 1:183 and VN titre of 1:516 be declared in the
“Instructions for Use” that accompany 09/194.
The material is suitable for use in HI and VN assays of human sera for antibody to A/California/7/2009 (H1N1)pdm virus.
Comments from participants
Laboratory 4
• This report reads really well and is very clear. I only had a few minor edits. – text corrected
• One point of clarification on lab 4 protocol is that we do use 4 HAU of turkey RBC - – text corrected
Laboratory 9
• Some minor editorial changes – text corrected
• Sentence in Introduction. “The aims of the collaborative study were: To evaluate the suitability of a freeze-dried serum pool to serve as an IS.” Comment: There’s no discussion about the freeze-dried (Sample B, E) vs liquid (Sample C) form of the IS. – International Standards must be suitable not only in terms of reactivity but also in terms of stability. Therefore it is not relevant to evaluate the merits of the liquid serum
preparation as an IS.
• Sentence on P8 “It is clear from the tables and figures that in the majority of cases there is a reduction in the between laboratory variation when titres are expressed relative to a standard.” Comment: the method needs to have acceptable reproducibility in order to consistently benefit from expressing titers relative to 09/194 – we agree, but don’t see how it affects the sentence. We are just saying that in most cases there was a benefit. This sentence should be left unchanged.
• Sentence on P8 “Calculating the titres for sample C relative to 09/194 would therefore be expected to produce the “best case” agreement.” Comment: Example of a benefit
expressing titers relative to 09/194 if the sample is similar to the 09/194 sample, not sure if this benefit would apply to a diverse sample set – This comment agrees with our statement that homologous sample and reference will result in better agreement. This sentence should be left unchanged.
• Sentence on P9 “An equivalent analysis was carried out using the titres relative to 09/194.” Comment: Not sure if this is biologically meaningful. One can use IS to normalize the HI and VN titers but not sure about using this normalized titers to see individual lab’s two assay relationship between HI and VN – introducing other labs variation factors into each lab’s assay performance. - We don’t agree. We are looking to see if there is a pattern of results between HI and VN, consistent across samples, after removing different levels of assay sensitivity between labs, i.e. would using results normalised to the IS allow a translation of the HI 1:40 to a VN titre?
• It is valuable to use the 09/194 candidate International Standard (IS) to evaluate the variability of the HI and VN within and among laboratories. However, the 09/194 IS should be used with caution. For example, both the HI and VN (with <26 hr incubation time) measure antibodies against influenza HA and should be correlated. The > 3 day VN methodology may be measuring other antibody effects. As such the use of a IS may influence correlations between the two assays differently. We are of the opinion that 09/194 should not be linked to the previous IS for antibody to H5N1 clade 1.
Additionally, the 09/194 candidate IS would be most valuable if the assigned potency was determined with data from HI and VN assays with certain inclusion criteria; for example, assigned IS potency with data generated from assays measuring same type of immune response(s) and with no extreme outliers (outside the 2SD range of the GMT) contributing to the assignments. The current assigned value of 09/194 was established from labs with variable assay reproducibility or measuring different types of immune responses, which may misrepresent the true value of the 09/194 candidate IS. – We accept that the various serology assays may be measuring different antibody populations, but in this study the candidate IS 09/194 reduced inter-laboratory variability for all the assays used. The issue of the consensus titre being derived from a variety of labs each with different assay sensitivities and validity criteria is difficult to address. It would be difficult to select which labs’ results to use, or be able to justify it, as the IS needs to be accepted for use by all labs. However if the potency of 09/194 is described in IU, then the actual consensus GMT is not important.
Laboratory 10
• We did provide checklists for both HAI and MN (sections 3.3.1 and 3.3.2) – updated sections 3.3.1 and 3.3.2
• Could you please not indicate source of virus for our two VN assays. Could you just indicate they are the same virus (A/California/7/2009 NYMC X-179A) from two different sources with 2 different methods for calculating titers? Just indicating VN # 1 and VN # 2 is fine. – text revised
• The section on HAI in Section 4 for Laboratory ten (listed 3)) is not needed since we only provided one method for HAI similar to all the others. Also 1) says VN # 1, while 2) says MN # 1, both should be VN. – text corrected but this section remains since laboratory 10 used the virus volume as part of the final serum dilution in contrast to other laboratories.
• I don’t think the titers for the two VN assays should be combined as they are not the same methods for titer calculation. The Karber method (VN#1) usually has a higher titer than the GMT method (VN#2). These should really be presented separately, although Table 14 already does this. - As they are separate in table 14, it is not essential that any of the other tables are changed. The overall means would be unaffected.
• Table 14 should indicate VN # 1 and VN # 2 not HAI and VN. – text corrected
References
1. Stephenson I, Das R, Wood J, Katz J. Comparison of neutralising antibody assays for detection of antibody to influenza A/H3N2 viruses: an international collaborative study.
Vaccine 2007; 25: 4056-65
2. Stephenson I, Heath A, Major D, Newman R, Hoschler K, Wang-Junzi, Katz J, Weir J, Zambon MC, Wood JM. Reproducibility of serologic addays for influenza virus A (H5N1).
Emerging Infectious Diseases 2009;15: 1250-1259.
3. WHO. Availability of a candidate international standard for antibody to A/California/7/2009 (H1N1)v-like viruses.
http://www.who.int/csr/resources/publications/swineflu/cp139_international_standard/en/
Table 1. Comparison of variable parameters in VN assay protocols
Parameter or variable Most frequent variables used Range A. Stock virus preparation
Cell substrate for virus growth
10-11 day old embryonated eggs
10-12 day old embryonated eggs Conditions of virus growth 2 days at 35oC 2-3 days, 32-35 oC
Stock virus infectivity and method of determination
~106 TCID50/ml 105.5-6.0 TCID50/ml titrated by ELISA/cytopathic effect B. Serum preparation
Storage of sera following receipt from NIBSC
-60oC storage; 1 freeze thawing cycles
-20oC to -70oC; 1 thawing cycle Pre-assay treatment of sera Heat treatment, 56oC for
30mins
RDE then heat treatment, 56oC for 30- 40mins
Initial serum dilution 1:10 (10µl in 90µl diluent) 1:5 to 1:10
Serum diluent Various used MEM no trypsin, Ultra MEM+trypsin, DMEM+BSA+HEPES
Serial dilution steps 1:2 dilution steps 1:2
Range of serum dilutions Various used 1:10 to 1:1280; 1:20 to 1:20,480 C. virus preparation
Virus conc added per well 50-100 TCID50 50-100 TCID50;
Dilution of stock virus to achieve assay virus conc
1:50- 1:3000 1:50 to 1:3500
Vol of virus solution added 50µl 50µl
Virus diluent Various used See serum diluent
Virus/serum mix incubation
1-2 hour, room temp or 37 oC 1-2 hour, room temp or 37 oC Calculated starting serum
dilution
1:10 1:10 (either include virus volume or
not) D. Cell preparation
Preparation of cells and number of cells added
Cell suspension or preformed monolayer method
Cell suspension or preformed monolayer method
Cell type used MDCK MDCK
Assay diluent Various used See serum diluent
E. Assay set up Total assay volume per well
200µl 200µl
Incubation time of assay to endpoint reading
<26 hours <26 hours; ≥3 days Incubation conditions 36-37 OC in 5% CO2 35-37 OC in 5% CO2 F. Endpoint estimation and
calculation
Endpoint determination
Viral antigen detection by ELISA using anti-nucleoprotein antibody
Detection of viral antigen by ELISA/
Cytopathic (light microscopy) Endpoint titer calculation
method
50% neutralization 50% neutralization or highest serum dilution showing no CPE
Table 2. Comparison of variable parameters in HI assay protocols
Parameter or variable Most frequent variables used
Range A. Stock virus preparation
Cell substrate for virus growth
10-11 day old embryonated eggs
10-12 day old embryonated eggs,
Conditions of virus growth 3 days at 34-35oC 2-3 days at 34-35oC Stock virus
haemagglutination units titre
1:480-1:1024 1:240 to 1:1024
B. Serum preparation Storage of sera following receipt from NIBSC
-20 to -70oC storage 1 freeze thawing cycles
-20oC to -70oC storage 0-2 freeze thawing cycles Treatment of sera 4 RDE to 1 sera, Overnight
37°C, Heat for 30-60 min 56OC
3-4 parts RDE to 1 part sera, Heat for 30-60 min 56OC +/- adsorption with turkey rbcs
Serum diluent PBS PBS
Initial serum dilution 1:10 1:5 to 1:10
Serial dilution steps/volume
1:2 dilutions in 25µl vol 1:2 dilutions in 25-50µl vol Range of sera dilutions 1:10 to 1:1024 1:10 to 1:1024 or 20480 C. Cell preparation
Cell type Turkey/chicken Turkey/chicken
Preparation of rbcs Within 72 hrs to 8 days of bleed
Within 12 hrs to 8 days of bleed
Red cell diluent PBS PBS or NaCl + 0.5%BSA
Red cell suspension concentration
0.4-0.5% v/v 0.4-0.7% v/v D. Virus preparation
Virus HA titration 4 HA units 4-8 HA units
Volume of virus added 25µl 25-50µl
Virus/serum mix incubation conditions
30-60 min at room temp or 37°C
20-60 min room temp or 37°C E. HI Assay set up
Total volume per well 75-200 µl 75-200µl
Incubation conditions to HI endpoint
45-60 min room temp 30-130 min room temp F. Endpoint estimation and
calculation
Endpoint determination Reciprocal of last well giving complete inhibition shown by streaming of RBC button
Reciprocal of last well giving complete inhibition as shown by streaming of RBC button
Table 3: Within Laboratory Reproducibility of VN and HI Titres
% > 2-fold % > 4-fold Lab HI VN HI VN
3 0 - 0 -
4 0 0 0 0
5 50 75 8 25
6 8 33 0 0
8 0 0 0 0
9 0 0 0 0
10 0 0 0 0
11 17 17 0 0
12 0 33 0 0
13 8 8 0 0
Table 4: HI - Laboratory GMTs by Sample and Virus
Sample A Sample B Sample C Sample E Sample F 09/194
Lab X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal
3 51 640 320 359 1810 285
4 20 25 101 101 160 127 80 80 1280 806 160 160
5 58 40 80 80 202 231 80 105 2560 2032 183 291
6 40 13 32 13 160 101 40 20 1280 806 160 101
8 20 10 160 63 101 80 160 101 1280 1016 160 80
9 160 1280 640 1280 3225 806
10 40 173 202 160 2370 160
11 20 10 20 13 101 40 32 16 1016 320 101 50
12 32 508 320 2560 10240 160
13 20 10 40 34 80 67 40 20 905 640 95 80
Overall 36 15 137 37 188 93 158 42 1925 806 183 107
%GCV 95 81 294 151 89 81 345 146 105 83 83 85
Table 5: VN - Laboratory GMTs by Sample and Virus
Sample A Sample B Sample C Sample E Sample F 09/194
Lab X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal 3
4 57 71 1810 1280 1016 640 1810 1280 16255 20480 508 453
5 81 49 1157 1126 617 604 2249 1064 6746 6347 336 373
6 40 101 1613 2560 1016 2560 2560 2560 2560 2560 640 1613
8 9 201 142 90 453 142
9 293 9612 4433 9799 39326 4305
10 142 4833 1918 4561 17222 1437
11 80 25 160 50 508 127 127 40 5120 1613 640 160
12 48 751 381 1331 3840 210
13 10 <10 403 254 160 127 508 320 2560 2560 202 127
Overall 52 34 1048 543 660 437 1153 537 5463 4242 516 354
%GCV 211 228 294 382 204 255 383 412 274 175 192 173
Table 6: Between Laboratory %GCV
%GCV
HI VN
A B C E F A B C E F
X179A 95 294 89 345 105 211 294 204 383 274 Absolute Titres
A/Cal 81 151 81 146 83 228 382 255 412 175 X179A 34 161 36 231 109 77 133 44 214 111 Titres relative to 09/194
A/Cal 20 96 12 92 36 259 152 37 182 250
Table 7: Summary of Between Laboratory %GCV
%GCV median (min – max)
HI VN
X179A 105 (89-345) 274 (204-383) Absolute Titres
A/Cal 83 (81-151) 255 (175-412) X179A 109 (34-231) 111 (44-214) Titres relative to 09/194
A/Cal 36 (12-96) 182 (37-259)
Table 8: Fold Range (max/min) Between Laboratories
Fold Range (max/min) between laboratories
HI VN
A B C E F A B C E F
X179A 8 64 8 80 11 33 60 31 109 87 Absolute Titres
A/Cal 4 8 6 7 6 20 51 20 64 13 X179A 1.6 16 1.3 64 16 5 14 2.5 34 10 Titres relative to 09/194
A/Cal 2.5 6 3.1 6 2.5 20 10 2 11 28
Table 9: HI - Laboratory GMTs by Sample and Virus Titres Relative to 09/194
Sample A Sample B Sample C Sample E Sample F
Lab X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal
3 28 359 180 202 1016
4 20 25 101 101 160 127 80 80 1280 806
5 50 22 70 44 176 127 70 58 2236 1118
6 40 20 32 20 160 160 40 32 1280 1280
8 20 20 160 127 101 160 160 202 1280 2032
9 32 254 127 254 640
10 40 173 202 160 2370
11 32 32 32 40 160 127 50 50 1613 1016
12 32 508 320 2560 10240
13 34 20 67 67 135 135 67 40 1522 1280
Overall 32 23 120 56 164 138 138 62 1682 1204
%GCV 34 20 161 96 36 12 231 92 109 36
Table 10: VN - Laboratory GMTs by Sample and Virus Titres Relative to 09/194
Sample A Sample B Sample C Sample E Sample F
Lab X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal X179A A/Cal 3
4 71 101 2281 1810 1280 905 2281 1810 20480 28953 5 155 85 2200 1931 1174 1036 4278 1825 12831 10887
6 40 40 1613 1016 1016 1016 2560 1016 2560 1016
8 40 905 640 403 2035
9 44 1429 659 1457 5847
10 63 2153 854 2032 7672
11 80 101 160 202 508 508 127 160 5120 6451
12 145 2293 1163 4064 11722
13 32 <10 1280 1280 508 640 1613 1613 8127 12902
Overall 64 44 1300 983 819 791 1429 972 6772 7676
%GCV 77 259 133 152 44 37 214 182 111 250
Table 11: Relationship between HI and VN titres using X-179A
Conversion Factor: HI to VN
Lab Sample
A B C E F 09/194 Mean
3
4 2.8 18.0 6.3 22.6 12.7 3.2 8.2 5 1.4 14.5 3.1 28.1 2.6 1.8 4.5 6 1.0 50.8 6.3 64.0 2.0 4.0 7.4 8 0.4 1.3 1.4 0.6 0.4 0.9 0.7 9 1.8 7.5 6.9 7.7 12.2 5.3 6.0 10 3.6 27.9 9.5 28.5 7.3 9.0 11.0 11 4.0 8.0 5.0 4.0 5.0 6.3 5.2 12 1.5 1.5 1.2 0.5 0.4 1.3 0.9 13 0.5 10.1 2.0 12.7 2.8 2.1 3.0 Mean 1.5 9.1 3.7 8.0 2.8 3.0
Table 12: Relationship between HI and VN titres using X-179A based on titres relative to 09/194
Conversion Factor: HI to VN
Lab Sample
A B C E F Mean
3
4 3.6 22.6 8.0 28.5 16.0 12.4 5 3.1 31.5 6.7 61.2 5.7 11.8 6 1.0 50.8 6.3 64.0 2.0 8.4 8 2.0 5.7 6.3 2.5 1.6 3.1 9 1.4 5.6 5.2 5.7 9.1 4.6 10 1.6 12.4 4.2 12.7 3.2 5.1 11 2.5 5.0 3.2 2.5 3.2 3.2 12 4.6 4.5 3.6 1.6 1.1 2.7 13 0.9 19.0 3.8 24.0 5.3 6.1
Mean 2.0 12.2 5.0 10.8 3.8
Table 13: Relationship between HI and VN titres using X-179A Correlation coefficients
Lab R R2 p
4 0.893 0.798 0.017*
5 0.662 0.483 0.152 6 0.350 0.122 0.497 8 0.910 0.828 0.012*
9 0.994 0.989 <0.001*
10 0.866 0.751 0.026*
11 0.984 0.968 <0.001*
12 0.986 0.971 <0.001*
13 0.762 0.580 0.079 R – Correlation coefficient
p – Level of significance. * indicates correlation significantly different from 0
Table 14: Comparison of HI and VN results from Lab 10 GMT based on three tests for each virus
Sample
A B C E F 09/194
VN #1 113 3225 1437 2873 14482 1140 VN # 2 180 7241 2560 7241 20480 1810
Figure legends
Figure 1 Histograms showing the distribution of geometric mean titres across laboratories. Each box represents the GMT for one laboratory, and is labelled with the laboratory code number.
Results for the wild type A/California/7/2009 virus are shown in grey and for NYMC X-179A in white. The results for the HI and VN tests are shown side by side. The respective titres relative to 09/194 are shown below the absolute titres.
Figure 2 Correlation between HI and VN titres against X-179A virus for each laboratory. The correlation coefficients are shown in table 13
Figure 1: Sample A
Sample A - Absolute HI titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 08 11 13
04 04 08 11 13
05 03 06 10 12
05 09
NYMC-X179A A/California
Sample A - Absolute VN titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
13 08 13
11 05 06 12
04 06 04 05 11
10 09
NYMC-X179A A/California
Sample A - VN relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
13 06
06 08 09 13
04 05 11 04 10 11
05 12
NYMC-X179A A/California
Sample A - HI relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
04 05 06 08 13 04 08
11 03 05 06 09 10 11 12 13
NYMC-X179A A/California
Figure 1: Sample B
Sample B - Absolute HI titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 11
11 13 06 13
04 05 08 04 05
08 10
03 12
09
NYMC-X179A A/California
Sample B - Absolute VN titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 08
11 13 13
12 04 05 04 05 06
06 10 09
NYMC-X179A A/California
Sample B - VN relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 11
08 04 06 13 06 09 13
05 04 05 10 12
NYMC-X179A A/California
Sample B - HI relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 05 11 06 11
04 13 04 05 13
08 08 10
03 09
12
NYMC-X179A A/California
Figure 1: Sample C
Sample C - Absolute HI titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 06 08 13 08 11 13
04 04 05 06 10
05 03 12
09
NYMC-X179A A/California
Sample C - Absolute VN titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 13 08 13
12 04 05 05 11
04 06
06 10
09
NYMC-X179A A/California
Sample C - VN relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
04 11 13 08 09 10 11 13
05 06 04 05 06 12
NYMC-X179A A/California
Sample C - HI relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
08 04 05 06 08 11 13 03 04 05 06 09 10 11 13
12
NYMC-X179A A/California
Figure 1: Sample E
Sample E - Absolute HI titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 11 13
06 11 13
04 05 08 04 05
08 10
03 09 12
NYMC-X179A A/California
Sample E - Absolute VN titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 08 11 13 13 04
05 04 12
06 05 06
10 09
NYMC-X179A A/California
Sample E - HI relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 11 13 06 11
04 05 04 05 13
08 03 08 10
09 12
NYMC-X179A A/California
Sample E - VN relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 11
08 04
06 13 09 13
05 04 06 10
05 12
NYMC-X179A A/California
Figure 1: Sample F
Sample F - Absolute HI titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
11 04 06 13
08 03 04 06 08 11 13
05 05 09 10
12
NYMC-X179A A/California
Sample F - Absolute VN titres
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
08 11 06 13 06 13
05 05 11 12
04 04 10
09
NYMC-X179A A/California
Sample F - VN relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
VN Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
06 06 08
11 09 11
05 13 05 10 12 13
04 04
NYMC-X179A A/California
Sample F - HI relative to 09/194
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
HI Titre relative to 09/194
-ve 10 20 40 80 160 320 640 1280 2560 5120 10240 20480 40960 81920
04 09
05 06 11 13 03 04 06 08 11 13
08 05 10
12
NYMC-X179A A/California
