WHO/BS/09.2124 ENGLISH ONLY EXPERT …whqlibdoc.who.int/hq/2009/WHO_BS_09.2124_eng.pdf · WHO/BS/09.2124 ENGLISH ONLY EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION Geneva, 19 to
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WHO/BS/09.2124
ENGLISH ONLY
EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION
Geneva, 19 to 23 October 2009
An International Collaborative Study to Value Assign
the 6th International Standard for Unfractionated Heparin and
the US Pharmacopeial Heparin Reference Standard for Assay Lot F
Elaine Gray
1,3, John Hogwood
1, Peter Rigsby
1, Barbara Mulloy
1, Anita Y. Szajek
2,
Michael R. Ambrose2, Fabian Jameison
2, Walter Hauck
2 and Tina S. Morris
2
1National Institute for Biological Standards and Control,
Blanche Lane, South Mimms, Potters Bar, Herts EN6 3QG, UK
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: [email protected]). 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: [email protected]).
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 [or editors as appropriate] alone are responsible for the views expressed in this publication.
WHO/BS/09.2124
Page 2
Summary Thirty-three laboratories from 18 countries contributed data obtained from 12 different assay
methods to value assign the 6th International Standard for Unfractionated Heparin against the 5th
International Standard for Unfractionated heparin, 97/578. Six candidates were included in the
study. Candidate W, 07/328, gave the lowest range of intra-laboratory variation (%GCV: 2.2 – 6.4)
for the different assay methods and the lowest inter-laboratory variation by all methods
(%GCV = 3.6). Candidate W, 07/328, is therefore recommended to be the WHO 6th International
Standard for Unfractionated Heparin with an assigned value of 2145 IU/ampoule by all methods.
This collaborative study also served to harmonise the International Unit and the United States
Pharmacopoeia (USP) unit for unfractionated heparin. A new (USP) Heparin Sodium reference
standard (RS) Lot F for assay has been selected and value assigned by sub-group analysis of data
obtained using the proposed new USP monograph potency method (anti-factor IIa antithrombin
dependent assay; aIIa:AT). Nine participants carried out this assay. Candidate X, 07/330, gave low
intra- and inter-laboratory variation for the USP anti-IIa chromogenic assay and it also gave the best
agreement of potency estimates by the three present and future USP monograph methods. Candidate X, 07/330, was accepted by the USP as the USP heparin standard for assay, Lot F, with
an assigned value of 2144 USP unit/ampoule.
As both the 6th International Standard for Unfractionated Heparin and the USP heparin standard for
assay, Lot F, are traceable and value assigned against the 5th International Standard for
Unfractionated Heparin, the long-standing disparity between the USP unit and the International Unit
for Unfractionated Heparin is resolved.
Introduction The 5th International Standard for Unfractionated Heparin, 97/578 was established in 1998 (1) and
it was calibrated against the 4th International Standard. This standard has been widely used for
calibration of clinical products and controls for diagnostic kits for heparin. Due to the depletion of
stock, a replacement for 97/578 is now required. The programme for the replacement of the WHO
IS for Unfractionated Heparin was endorsed by the WHO Expert Committee on Biological
Standardisation in October 2006 and was introduced to the SSC/ISTH Subcommittee on Control
of Anticoagulation at the 53rd SSC meeting in July 2007 (Geneva).
For the last 30 years or more, there has been a 7–13% disparity between the United States
Pharmacopoeia (USP) unit and the International Unit (IU) for unfractionated heparin (1). It is
therefore desirable to harmonize these two units for the measurement of heparin. The need to
replace the 5th International Standard for Unfractionated Heparin and to establish the USP heparin
standard for assay, presents an ideal opportunity for harmonization. The USP is currently
undertaking a revision of the Heparin Sodium anticoagulant potency monograph assay, with the
intention to establish a new USP heparin standard for assay that will enable the harmonization of
the USP unit and the IU for unfractionated heparin. This harmonization will require the calibration
of the new USP heparin standard for assay, Lot F, against the current International Standard (IS)
for Unfractionated Heparin and as the new USP potency assay, the anti-factor IIa chromogenic
assay using purified antithrombin (aIIa:AT) will also be implemented at the same time, the
potency of Lot F would be value assigned using this new aIIa:AT potency assay.
Because of the recent contaminated heparin events, heparin manufacturers have been encouraged
by regulatory authorities to replace clotting assays that can be influenced by contaminants and
process related impurities to methods that are specific for heparin, for example, anti-IIa and anti-
WHO/BS/09.2124
Page 3
Xa assays using purified antithrombin, for measurement of heparin potency. This study also
allowed us to examine the robustness of these chromogenic methods.
This report describes two important findings of this collaborative study that involved 6
candidates:
1. The analysis of data and conclusions that lead to the value assignment of the proposed
6th International Standard for Unfractionated Heparin
2. Information on the calibration of the USP heparin standard for assay, Lot F and the
performance of anti-IIa and anti-Xa chromogenic assays
Candidates The coding of the candidates is as indicated in Table 1. Candidates T, W, X, Y and Z are all
typical clinical heparin Active Pharmaceutical Ingredients (APIs) from different European and
North American manufacturers. Candidate V is not a routine API and has been manufactured for
research studies only. Preliminary investigations including specific activity, molecular weight
profiles and NMR spectroscopy were carried out at NIBSC. Specific activities of the candidates
were measured against the 5th IS, using antithrombin dependent anti-IIa and anti-Xa assays, on
candidates weighed out at ambient temperature and humidity. With the exception of candidate V,
which has a specific activity close to 300 IU/mg by both anti-IIa and anti-Xa assays, all other
candidates have specific activities between 170 – 230 IU/mg. The higher specific activity
exhibited by candidate V could be due to the higher proportions of high molecular weight
material in this sample (Figure 1). With the exception of 08/254, candidate T, all the other
candidates were filled and freeze-dried at NIBSC in accordance with the conditions normally
used for international biological standards (2). Table 1 shows the fill characteristics of the
candidates. The excipient for all the candidates was sterile water.
The CV of the fills showed that all the candidates were filled with acceptable precision.
Homogeneity of the fill by functional activity was also investigated by assay of ampoules
sampled across the fill. For each candidate filled at NIBSC, 4 ampoules (2 independent assay
per ampoule) at every 5000 ampoules filled were assayed by the Activated Partial
Thromboplastin Time (APTT) against the 5th IS and results were compared by analysis of
variance. Table 2 shows that the variabilities by functional assay for candidates V, W, X, Y and Z
were all low with GCV below 3% and analysis of variance showed that there were no significant
differences between the activities of the ampoules across the fill.
Molecular weight determination of the candidates was carried out according to Mulloy et al. (3).
The profiles and measurements are shown in Figure 1 and in Table 3. Candidate V is noticeably
different from the others, having the highest values for peak molecular weight (Mp), number
average molecular weight (Mn) and weight average molecular weight (Mw). The other candidates
fall within a relatively narrow range, similar to that of the 5th International Standard for
Unfractionated Heparin.
1H NMR spectra at 500 MHz were recorded for each of the candidates in accordance with McEwen
et al. (4) and spectra of the candidates are shown in Figures 2 - 7. The NMR spectra confirmed that
the candidates are all free of oversulphated chondroitin sulphate, a contaminant which has been
found in some recent heparin batches. A small signal in the spectrum of sample W at 2.18 ppm can
be seen, but a spiking experiment showed that this does not originate from oversulphated
chondroitin sulphate (data not shown) .
WHO/BS/09.2124
Page 4
Participants Thirty-six laboratories from 18 countries (1 Australia, 4 Austria, 1 Bosnia, 1 Brazil, 1 Canada, 1
China, 2 Denmark, 4 France, 5 Germany, 1 India, 2 Italy, 1 Japan, 1 The Netherlands, 1
Portugal, 3 Spain, 3 Switzerland, 1 UK, 3 USA) agreed to take part in the study. Thirty-three
laboratories returned results, each of which is referred to in this report by an arbitrarily assigned
number, not necessarily representing the order of listing in Appendix 1. Lab 20 (Austria) withdrew
from the study, while results from Lab 4 and 14 were not received in time for analysis.
Assay methods All the methods used by the participants are listed in Appendix 2. In total, 12 different methods
were used in the study:
• Anti-IIa chromogenic assay, purified antithrombin (AT), n = 18
• Anti-Xa chromogenic assay, purified AT, n = 23
• Anti-IIa chromogenic assay, human plasma, n =1
• Anti-Xa chromogenic assay, human plasma, n = 3
• Anti-Xa clotting assay, n = 1
• US Pharmacopoeial assay (USP), clot-based sheep plasma assay n = 9
• European Pharmacopoeial assay (EP), clot-based sheep plasma assay n = 12
• Japanese Pharmacopoeial assay (JP), n = 1
• Chinese Pharmacopoeial assay (CP), n = 1
• Human plasma APTT, n = 12
• Thrombin Time (TT), n = 2
• Prothrombinase induced clotting time (PiCT), n = 1
Some laboratories requested assay protocols for anti-IIa and anti-Xa chromogenic assay using
purified antithrombin. NIBSC provided these laboratories with methods for both of these assays
(Appendix 3).
Study design The detail of the study protocol is shown in Appendix 4. Each laboratory was asked to perform
their chosen methods. Where a laboratory performed more than one assay method, each method is
treated as if performed by different laboratories.
Participants were requested to perform four independent assays for each method, an independent
assay being defined as one with a completely fresh set of ampoules and dilutions. To allow for day
to day variation, the participants were asked to carry out the four assays, for each method, on four
separate days. For those laboratories that performed two or more methods, an assay for each
method could be carried out on the same or different days, from one set of reconstituted solutions,
but making fresh dilutions for each method.
Participants were asked to assay concurrently a series of at least three, preferably four dilutions of
each of the five coded samples. The assay order of the materials (including replicates) was to be
varied to give a balanced order overall. Some examples of different assay order schemes were given
in the protocol for participants to choose from.
WHO/BS/09.2124
Page 5
Raw assay data were to be returned together with a summary of their estimates for the potency of
materials T, V, W, X, Y and Z using S, the 5th IS for Unfractionated Heparin, as the standard.
Statistical analysis An independent statistical analysis of raw data was performed at NIBSC. Potencies were
calculated relative to the 5th International Standard for Unfractionated Heparin, 97/578 (Study
code: S), for each candidate sample using a parallel-line or slope-ratio model (5). All data were
plotted and the assay validity was assessed both visually and by analysis of variance. When
doses did not appear to lie on the linear section of the dose-response curve they were excluded
from analysis. All assays that showed no significant deviations from the model at the 1% level
have been included in this report. Where non-parallelism was found to be significant at the 5%
level in parallel-line models (0.01<p<0.05) the results are included and highlighted. This is also
the case when the intercept difference was found to be significant at the 5% level in slope-ratio
models (0.01<p<0.05). In a small number of cases the results reported by the laboratory have
been used directly and these are indicated below. All mean potencies given in this report are
unweighted geometric mean potencies. Variability between assays and laboratories has been
expressed using geometric coefficients of variation (%GCV) (6). Grubbs’ Test (7) was applied to
the log transformed laboratory mean estimates in order to detect any significant outliers (p<0.05)
and analysis of variance (with Tukey’s test) was used to make comparisons between assay
methods.
Due to the late submission of results by Lab 4 and Lab 14, their data obtained by the European
Pharmacopoeia method and APTT method respectively could not be analysed and was therefore
excluded from the overall potency estimates. For comparison purpose, the potency estimates
from Lab 4 and 14 are shown in Appendix 5.
Anti-IIa chromogenic assays using purified AT A parallel-line model comparing untransformed assay response to log concentration was used in
all laboratories except laboratory 05 where a log transformation appeared to be more suitable.
Anti-Xa chromogenic assays using purified AT A slope-ratio model comparing log transformed assay response to concentration was used for
almost all laboratories. For laboratory 10 a parallel-line model comparing untransformed assay
response to log concentration appeared to be more suitable. For laboratory 36 the assay
responses were transformed to percentages relative to the estimated upper and lower limits of the
dose-response curve and a weighted regression of logit response on log dose was used to
calculate potency estimates.
USP assays No further analysis was carried out as the USP sheep plasma assay is based on a single point
estimate. The results reported by the laboratories are shown and used for overall potency
estimates.
EP assays A parallel-line model comparing log transformed assay response to log concentration was used
in all laboratories. For laboratory 17 deviations from parallelism were compared to deviations
from linearity when assessing the validity of the model due to the low level of residual error
observed.
WHO/BS/09.2124
Page 6
APTT assays A slope-ratio model comparing log transformed assay response to concentration was used for
almost all laboratories. For laboratory 10 a parallel-line model comparing untransformed assay
response to log concentration appeared to be more suitable. For laboratories 34a and 34b, neither
model appeared to provide a good fit and the reported estimates (calculated using a third order
polynomial calibration curve) are shown.
Anti-Xa chromogenic assay using plasma, Thrombin Time and other methods For laboratories 18 (Japanese Pharmacopoeial method), 30 (anti-Xa assay, using plasma as
source of antithrombin) and 34 (Thrombin Time), the reported estimates are shown as neither
model appeared to fit the data (labs 30 and 34) or only single dilutions of each test sample were
reported (lab 18). In laboratory 34 (anti-Xa Clotting) the untransformed responses were used.
Results
Assay data The 33 participants contributed data from a total of 690 assays; with 642 assays analysed by NIBSC.
For 48 assays, no further analysis was carried out and the potency estimates reported for each of
these assays were used to calculate the overall potency estimates and the intra- and inter-laboratory
variability. Table 4 shows the breakdown of the number of assays per assay method.
Assay validity The assumptions of linearity and parallelism of the log dose-transformed/untransformed response
lines of samples T, V, W, X, Y, Z and S held in 79 % of the assays analysed. In some of the assays,
certain candidates were found to be non-linear and/or non-parallel to S and the potency estimates for
these samples were not included in the overall potency estimation. Tables 6 - 11 show the validity
and potency estimates of all the assays. Some of the invalidities may have been due to inadequate
replication in the assay design, as well as from real deviations from the model assumptions. There
was no evidence of any overall non-parallelism when the 6 candidates were compared with the 5th
IS, however, candidate V did yield the highest number of assays that were non-parallel or had
significantly different intercepts and this was especially true for the EP sheep plasma assay (Table
5).
Potencies relative to the 5th International Standard for Unfractionated Heparin,
97/578 The individual assay potency estimates of candidates T, V, W, X, Y and Z taken relative to sample
S (5th IS for Unfractionated Heparin) by different methods are shown in Tables 6 - 11. Detailed
values of the individual laboratory mean potencies and overall mean potencies of samples T, V, W,
X, Y and Z relative to sample S, the 5th IS for Unfractionated Heparin for each method are listed in
Tables 12 – 18. As there was an insufficient number of data sets returned for the anti-IIa
chromogenic assay using human plasma, the anti-Xa clotting assay, Japanese Pharmacopoeial assay,
Chinese Pharmacopoeial assay and Prothrombinase induced Clotting Time, overall mean potency
estimates were not calculated for these methods and the laboratories’ mean potencies for each
candidate are classified as “other methods” and shown in Table 19. A summary of the potency
estimates by the different methods is also presented in Table 21. The data are also shown in
histogram forms in Figures 7 - 12. Only 9 outliers were identified by Grubb’s test. These were
excluded from the overall mean potency estimates as indicated in Tables 12 -19 and Figures 1 – 6.
Although the Japanese Pharmacopoeial assay produced 5 outliers, only one laboratory carried out
this assay method. It was not possible to conclude whether the problem is related to this method or
WHO/BS/09.2124
Page 7
with the laboratory’s performance. There was no correlation between assay methods and outliers
identified. The histograms (Fig 7 – 12) illustrate good agreement between both the laboratories
and the assay methods for candidates T, W, X, Y and Z. Analysis of variance on the log potencies
to compare methods shows no significant differences for W (p=0.832) and X (p=0.999). Sample
Y shows significant differences (p=0.036) due to the US Pharmacopoeial assay results being
significantly higher than those for anti-IIa and anti-Xa chromogenic assays using purified AT
(p<0.05 in Tukey’s test). Sample T shows significant differences (p=0.008) due to the anti-IIa
chromogenic assays using purified AT results being significantly lower than the USP results
(p<0.05 in Tukey’s test). Sample V shows significant differences (p<0.001) due to the anti-IIa
chromogenic assays using purified AT results being significantly higher than all other methods
and the anti-Xa chromogenic assays using purified AT results being significantly higher than the
APTT, the European Pharmacopoeial and the US Pharmacopoeial assays results (p<0.05 in
Tukey’s test). Sample Z shows significant differences (p<0.001) due to the anti-IIa chromogenic
assays using purified AT results being significantly higher than the APTT, European
Pharmacopoeial assay and US Pharmacopoeial assay results and the anti-Xa chromogenic assays
using purified AT results being significantly higher than the APTT results (p<0.05 in Tukey’s
test).
Table 21 shows the overall potency estimates by all methods relative to the 5th IS for
Unfractionated Heparin, together with the respective 95% confidence limits for all the candidates.
Figure 13 shows the potency estimates by each assay method expressed as the % of the overall
mean potency estimates for each candidate. It is clear that with the exception of the potency
estimates for candidate V which varied with different methods (PiCT = 82.6% to anti-IIa AT =
123.7%), potency estimates by the different types of assays for the other candidates were close to
the overall mean potency estimates for that candidate; the maximum difference being 6% for
sample X by the Chinese Pharmacopoeial assay and sample T by the Prothrombinase induced
Clotting Time. Potency estimates for candidate W, followed by sample X when assayed against the
5th IS, were the least influenced by the method type.
Intra-laboratory variability: all methods The variability within each laboratory, expressed as geometric coefficients of variation (% GCV's)
for each sample, is given in Tables 12 - 19. Table 20 summarises the %GCV showing the number
under 5%, 7% and 10%. In general, the laboratories were able to perform their chosen methods with
good precision and reproducibility with more than half of the GCVs less than 5% and only a small
percentage of laboratory potency estimates gave over 10%. The clot-based assays, especially the
pharmacopoeial assays, gave lower ranges of intra-laboratory %GCVs than the chromogenic assays.
This may be because the pharmacopoeial assays are well-established methods and the laboratories
are more experienced at carrying out these assays. Nonetheless, the chromogenic assays performed
well and only a few participants produced GCVs over 10%.
Inter-laboratory variability: all methods Variability between laboratories of the potency estimates of samples T, V, W, X, Y and Z relative to
sample S, the 5th IS for Unfractionated Heparin, is shown in Tables 12 – 18 for anti-IIa
chromogenic assay with AT, anti-Xa chromogenic assay with AT, European Pharmacopoeial assay,
US Pharmacopoeial assay, APTT, Anti-Xa chromogenic assay with plasma and Thrombin Time
and is also summarised as % GCV's in Table 19. Inter-laboratory variability is not given for
Thrombin Time as only two laboratories carried out this method and for anti-IIa chromogenic assay
with plasma, anti-Xa clotting assay, Chinese Pharmacopoeial assay, Japanese Pharmacopoeial assay
and Prothrombinase induced Clotting Time as these assays were carried out by one laboratory only.
WHO/BS/09.2124
Page 8
Variability between laboratories ranged from 0.1 to 18.0% and was consistently higher for
sample V and consistently lower for sample W. The clot based methods (European Pharmacopoeial
assay, US Pharmacopoeial assay and the APTT) tended to give lower inter-laboratory variability,
but the variability found with the chromogenic assays was acceptable, with the majority of the inter-
laboratory GCVs being less than 10%. Pooling the different assay methods (Table 22), the
variability between laboratories is the highest for sample V (GCV = 20.7%) and the lowest for
candidate W (GCV = 3.6%). The variability for samples T, X, Y and Z was also relatively low,
with respective GCVs of 6.6, 4.8, 5.7 and 5.5%.
Comparison of the candidates as putative standards Potencies of candidates were recalculated relative to each other, assuming each sample has an
assigned unitage equivalent to the overall mean potency by all methods, relative to sample S, shown
in Table 22. The overall mean potencies, 95% confidence limits and inter-laboratory variation
expressed as % geometric coefficient of variation (%GCV) for each candidate relative to each other
are shown in Tables 23 – 28 and the potency estimates are summarised in Table 29. These were
similar to those relative to the 5th IS for Unfractionated Heparin, which is not surprising since the
assumed assigned potencies of each sample were calculated relative to the 5th IS for Unfractionated
Heparin.
Variability between laboratories of the potency estimates of Candidates T, V, W, X, Y and Z
relative to each other, for all assay methods, is summarized as % GCVs in Table 30. High
variability was observed when Candidate V was used as the putative standard and as the test
sample against all the other putative standards. Amongst the other candidates, the differences in
inter-laboratory variation were marginal. Generally, variability between laboratories was lowest
when Candidates W or X were used as the putative standard and they could be considered as the
potential replacement international standard.
Sub-group analysis to select and value assign USP RS Lot F Sub-group analyses were carried out to assess the performance of the USP aIIa:AT, USP aXa:AT
and the USP sheep plasma clot based assays and to select and value assign the USP heparin
standard for assay.
Three different USP assay methods related to potency were carried out by the participants: a. the
current sheep plasma clot based assay, b. the current aXa:AT assay and c. the new (to be
implemented in October, 2009) aIIa:AT assay. Appendix 2b lists the laboratories by number and
the USP assay methods that they have performed. For anti-IIa and anti-Xa chromogenic assays,
laboratories 02, 19 and 25 declared that they have performed the USP anti-Xa and the proposed
anti-IIa chromogenic assays. The remainder of the laboratories used protocols provided by
NIBSC. The NIBSC protocols were simplified versions of the USP methods, with the reagents
and assay conditions as described in the USP assay methods (Appendix 3). In total,
9 laboratories carried out the USP anti-IIa, anti-Xa and sheep plasma clot-based assays. For
information, the laboratories that carried out antithrombin dependent anti-IIa and anti-Xa assays
based on commercial kits, in-house or European Pharmacopeoial Low Molecular Mass
Monograph methods are listed in Appendix 2c.
Due to the nature of the USP sheep plasma clot based assay, it was not possible to analyse the raw
data and the laboratory’s own potency estimates were used for comparison. Only statistically valid
USP aIIa:AT assays were used for estimation of potency for the USP RS Lot F.
WHO/BS/09.2124
Page 9
Intra-laboratory variation: USP sheep plasma clot based, aIIa:AT and aXa:AT
assays Tables 31, 32 and 34 show the intra-laboratory variation, expressed as %GCVs for the USP sheep
plasma clot based, aIIa:AT, aXa:AT assays respectively. Tables 36 and 37 summarise the %GCV
showing the number under 5%, 7% and 10% for the all the antithrombin dependent anti-IIa and
anti-Xa assays. The intra-laboratory variation was low for the USP sheep plasma clot based assay,
GCVs ranging from 0 – 11.5%. The higher %GCVs were found with results from 1 laboratory (Lab
19); the GCVs for the other participants using this method were below 5% indicating that the
laboratories were able to carry out this assay with good precision. The intra-laboratory variation for
the USP aIIa:AT method was also reasonably low, GCV ranging from 1.4 – 29.1%, with the higher
GCVs coming from 2 laboratories (Lab 03 and 19) and the majority of the GCVs being < 7%.
These results were in line with data from other antithrombin dependent anti-IIa chromogenic assays
(Tables 33, 36). The antithrombin dependent anti-Xa assays showed similar reproducibility to the
anti-IIa assay; for both the USP aXa:AT and other chromogenic assays, the majority of the GCVs
were below 7% (Tables 34, 35, 37). These results indicate that the majority of the laboratories
carrying out these chromogenic assays were able to perform these methods with good precision and
reproducibility. There was no obvious correlation between the degree of intra-laboratory variation
and the different candidates. The sheep plasma clot based assay did give lower ranges of intra-
laboratory variation and this may be because this pharmacopoeial assay has been in use for more
than 30 years so that some laboratories are more experienced at carrying out this assay than the new
chromogenic assays.
Inter-laboratory variation: USP sheep plasma clot based, aIIa:AT and aXa:AT
assays Variability between laboratories of the potency estimates of candidates T, V, W, X, Y and Z relative
to sample S, the 5th IS for Unfractionated Heparin, is shown in Table 38. For both the USP aIIa:AT
and aXa:AT methods, it is clear that the variability between laboratories is the highest for candidate
V, the lowest for candidate W, with candidate X having GCVs close to those obtained for candidate
W. The higher inter-laboratory variation exhibited by V could be explained by the molecular
weight differences between V and the 5th IS. It is also important to note that the inter-laboratory
variation for both the USP antithrombin dependent anti-IIa and anti-Xa assays are not too different
to those obtained by the sheep plasma clot based assay.
Potency estimates by USP sheep plasma clot based, aIIa:AT and aXa:AT assays
relative to the 5th IS for unfractionated heparin The individual laboratory potency estimates of candidates T, V, W, X, Y and Z taken relative to
sample S (5th IS for Unfractionated Heparin) by the USP sheep plasma clot based, aIIa:AT,
aXa:AT and all methods are shown in Tables 39 – 41. A summary of the potency estimates by the
different methods is also presented in Table 42. Analysis of variance of log potency estimates
indicated that there were no significant differences between the potencies obtained using the USP
anti-IIa and anti-Xa chromogenic methods and other chromogenic assays for any of the candidates
(p>0.05).
Stability study Preliminary accelerated degradation study, of all the candidates, monitored using both anti-IIa and
anti-Xa chromogenic assays, showed no sign of change after 9 months storage at temperatures of -
150, -70, -20, +4, +20, +37, +45 and +56°C. Continual real time degradation study of the –20°C
against ampoules stored at –70°C and further accelerated degradation study at elevated temperature
will be carried out to monitor the stability of the replacement standard.
WHO/BS/09.2124
Page 10
Discussion The main aims of this collaborative study were to select from the six candidate materials, firstly a
replacement for the 5th International Standard for Unfractionated heparin and secondly a new USP
heparin standard for assay. It also provided an opportunity to examine performance of the anti-IIa
and anti-Xa chromogenic assays.
Selection of the proposed 6th International Standard for Unfractionated
Heparin
Intra-laboratory (between assay) variability for all 6 candidate materials was good when assayed
against the 5th IS, with the majority of the %GCVs well under 10% and more than half of the
%GCVs being 5% or less (Tables 12 - 19). This indicates that the participants of the study were able
to perform their choice of assay with precision. The clot based assays (European Pharmacopoeial
assay, US Pharmacopoeial assay and APTT) generally gave lower between assay %GCVs and the
chromogenic methods gave slightly higher. The higher %GCVs for anti-IIa chromogenic assay
with AT came mostly from Lab 03 and Lab 19 (Table 12), while the higher GCV for anti-Xa
chromogenic assay with AT were from Lab 16 and Lab 36 (Table 13). Taking into account that the
higher variability was from 4 laboratories only, the %GCVs for other laboratories were excellent for
these two chromogenic assays indicating that the majority of the participants were able to perform
these chromogenic assays reproducibly. There were no other obvious correlations between
performance and parameters such as methods or reagents.
With the exception of candidate V, the inter-laboratory variations were excellent for all the other
candidates when assayed against the 5th International Standard for Unfractionated Heparin (Tables
12 – 19, 21, 22); the %GCVs were under 11% by any of the assay methods used, with at least half
under 4% (range 0.1 – 10.7%). Of all the candidates, W gave marginally lower variability than the
other samples (Tables 21, 22). These data suggest good reproducibility of results by the various
methods amongst the different laboratories. The higher variability (range 5.9 – 18.0%) obtained
with candidate V could be due to the differences between the molecular weight profile of the 5th
International Standard and sample V, with sample V having a higher proportion of high molecular
weight material than the 5th International Standard and the other candidates in the study (data not
shown). The inter-laboratory variation when all methods were taken into account indicated that W
gave the lowest and V gave the highest %GCV (W, GCV = 3.6%; V, GCV = 20.7%). The other
candidates only gave marginally higher %GCVs than sample W (6.6, 4.8, 5.7 and 5.5 for T, X, Y
and Z respectively).
In terms of assay methods, there were no significant differences amongst the potency estimates
obtained using the different methods for candidates W and X against the 5th International Standard
for Unfractionated Heparin. The best agreement was obtained for candidate W where the potency
estimates by various methods were between 96.7 to 101.5 % of the overall mean potency estimate
by all methods (Fig. 14). Potency estimates for candidate T by anti-IIa chromogenic assay using
purified AT were significantly lower than the US Pharmacopoeial assay results. For sample V,
potency estimates were significantly higher by anti-IIa chromogenic assay with purified AT than
results from all other methods and potencies by anti-Xa chromogenic assay with purified AT were
significantly higher than the data from the APTT and the European Pharmacopoeial assay. For
sample Y, the US Pharmacopoeial assay was significantly higher than both the anti-IIa and anti-
Xa chromogenic assay using purified AT. For sample Z, the anti-IIa chromogenic assays using
purified AT results were significantly higher than the APTT, European Pharmacopoeial assay
and US Pharmacopoeial assay results, with the anti-Xa chromogenic assays using purified AT
results being significantly higher than the APTT results. Although not all the assay methods
gave complete agreement of potency estimates for all the candidates, with the exception of
WHO/BS/09.2124
Page 11
sample V (range of potency estimates by different assay methods = 82.6 – 123.7 of the overall mean
potency), the overall differences between potency estimates obtained with the various methods were
not large and were probably found to be statistically significant because of the good precision of the
assays. This is confirmed by the close agreement in the potency estimates of the different
preparations (except V) by all assay methods, with all except four of the average percentages of
combined potency estimates for each method within 5% of the overall combined potency estimates
from all assay methods (Fig 14).
When each candidate was used as the putative standard, candidate V gave the highest inter-
laboratory variation (GCV range 17.8 – 24.8 %). Although high %GCVs were obtained when V
was assayed against W and X, these two candidates gave low GCVs for all the other samples (GCV
range 4.0 – 6.0%, Table 30). Since candidate W also gave the lowest range of intra- and inter-
laboratory variation when assayed against the 5th IS, W would be the best candidate to go forward
as the 6th IS for Unfractionated Heparin.
Selection of the USP heparin standard for assay As 9 laboratories carried out the USP aIIa:AT, there were sufficient data to carry out sub-group
analysis and obtain estimation of potency by this proposed USP monograph potency method for
heparin sodium. Comparison of log potency estimates obtained using the USP sheep plasma, USP
aIIa:AT and USP aXa:AT assays by one-way analysis of variance (ANOVA) and Tukey’s test
showed that there were no significant differences amongst the potency estimates for candidates W
(p = 0.277), X (p = 0.979) and Z (p = 0.06). Candidate T shows significant differences due to the
USP aIIa:AT assay results being significantly lower than the USP sheep plasma assay results
(p<0.05 in Tukey’s test). Candidate V shows significant differences (p<0.001) due to the USP
aIIa:AT and aXa:AT results being significantly higher than the USP sheep plasma assay data.
Candidate Y shows significant differences due to the USP aIIa;AT assay results being
significantly lower than the USP sheep plasma assay results (p<0.05).
Potency estimates for candidates W and X when assayed against the 5th IS, were the least
influenced by the method type (Figures 14, 15). Table 43 also shows that candidates W and X gave
an anti-Xa to anti-IIa ratio of 1.01 and that both candidates are within the USP specification for anti-
Xa to anti-IIa ratio of 0.9 – 1.1 for Heparin Sodium. Figure 15 illustrates the potency estimates by
the different USP methods expressed as the % of the overall mean potency estimates obtained by all
methods employed in the study. It is clear that for candidate X, the three USP methods gave
excellent agreement in potency estimates. So in considering continuity and traceability of the
international unit and that Lot F will be the RS that unites the international and USP units, candidate
X would be a good choice.
Conclusions and Recommendation • Although the clot based assays gave lower intra- and inter- laboratory variation, the
performance of the anti-IIa and anti-Xa chromogenic assay was good. Some laboratories
performed better than others with intra-laboratory GCV of less than 1%. In general, the
intra-laboratory GCVs were around or less than 5%. Good laboratory agreement was
also obtained by the anti-IIa and anti-Xa chromogenic assays for all candidates except V.
However sample V is very different in physico-chemical characteristics to the current
International Standard and to the other candidates in the study and poor laboratory
agreement was also obtained for V when other assay methods were used. This illustrates
the importance of assaying “like against like” in obtaining good agreement between
laboratories.
WHO/BS/09.2124
Page 12
• Currently there are insufficient data to assess and compare the stability of these
candidates. However, all the physical parameters such as moisture and oxygen content
are within the specifications set out in the WHO guidelines for the preparation of
international standards and reference materials. Stability data on preparations with
similar characteristics e.g. the 5th International Standard for Unfractionated Heparin and
the 2nd International Standard for Low Molecular Weight Heparin have shown no loss of
measurable activity after more than 10 years storage at -20°C. So these preparations are
expected to be very stable over the lifetime of reference material.
• All candidates except V performed well in the study, with the majority of the intra- and
inter- laboratory GCVs below 10%. Overall, candidate W gave the lowest range of intra-
and inter-laboratory variation whether as a test sample or as the putative standard. It is
therefore recommended that Candidate W, 07/328, to be the 6th International Standard
for Unfractionated Heparin with an assigned value of 2145 IU/ampoule obtained from all
assay methods against the 5th International Standard for Unfractionated Heparin, 97/578.
• The sub-group analysis presented in this report shows that both candidates W and X are
also suitable materials for the next USP heparin standard for assay, Lot F. Both
candidates gave comparably low intra- and inter-laboratory variation for the USP anti-IIa
chromogenic assay, but candidate X is marginally better than candidate W in that it gave
the best agreement of potency estimates by the three present and future USP monograph
methods (figure 15). The molecular weight distribution of candidate X is similar to that
of other typical clinical APIs in the study. The 1H NMR spectrum also indicates that
candidate X is a clean heparin with no detectable over-sulphated chondroitin sulphate and
extremely low levels of other process related impurities such as dermatan sulphate,
acetate or ethanol.
• It was therefore recommended that candidate X, 07/330, be the USP heparin
standard for assay, Lot F, with a value of 2144 USP unit/ampoule as assigned by
USP proposed aIIa:AT assays. The USP considered the recommendation and agreed
with this proposal, and candidate X has been established as the USP heparin standard
for assay Lot F.
• As the calibration of both the proposed 6th IS and USP heparin standard for assay,
Lot F was against the 5th IS, the USP unit and the International Unit for
Unfractionated Heparin are now harmonized.
• A consequence of this harmonization is a discontinuous change in the USP unit. The
effects of this change on users are being monitored by the USP. The US FDA is aware of
this change as well.
Responses from the participants and Scientific and Standardization
Committee (SSC) of the international Society on Thrombosis and
Haemostasis All submitted responses from participants agreed with the recommendations for the proposed 6th
International Standard for Unfractionated heparin. There was a comment from a participant with
regards to the classifications of the method used by their laboratory and the error has been corrected
in this current report. This error has no impact on the recommended replacement IS and its value
assignment. One of the SSC experts required clarification of the rationale for the use of slope ratio
and parallel line analysis and abbreviations used in the report. These comments have also been
considered and amended as required. At the SSC Annual Business Meeting (the 55th SSC
Meeting in Boston, 15 July 2009), the SSC endorsed the proposal that Candidate W, 07/328, be the
6th International Standard for Unfractionated Heparin.
WHO/BS/09.2124
Page 13
Proposal and recommendation to the ECBS • Candidate W, 07/328, be the 6
th International Standard for
Unfractionated heparin : 2145 IU/ampoule
The proposed Instruction for Use for the proposed 6th International Standard for Unfractionated
Heparin, 07/328 is illustrated in Appendix 6.
References (1) Gray E, Walker AD, Mulloy B, Barrowcliffe TW. A collaborative study to establish the 5th
International Standard for Unfractionated Heparin. Thromb Haemost. 2000 Dec;84(6):1017-
22.
(2) Campbell PJ. Procedures used for the production of biological standards and reference