•
WORLD HEALTH ORGANIZATION ORGANISATION MONDIALE DE LA SANTE
WHO/BS/68.939 ENGLISH ONLY
EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION Geneva, 30 September to 5 October 1968
REPORT ON THE INTERNATIONAL COLLABORATIVE ASSAY OF ANTI-TOXOPLASMA SERUM Additional information by P. Krag, M.D. Department of Biological Standardization and M. Weis Bentzon Actuary, Bio-Statistical Department Statens Seruminstitut, Copenhagen, Denmark
The International Standard for Anti-Toxoplasma Serum was established in 1967 (WId Hlth Org. techno Rep. Ser., 1968, 384, 18). In June 1967 and April 1968 further reports were received from two laboratories (see Table 1) and the results have been included in the material reported on as WHO/BS/890 under the designation laboratory 7 and 8; the evaluation has been undertaken following the same principles as those mentioned in this main report. The sera tes.ted are listed in Table 2; laboratory 8 did not receive serum BK'
The dye test results from laboratory 7 (see Table 3) are fitting well into the general picture, although the log titres are lying lower than for laboratories 1 to 4. Only results from the third day are at the same level as the results from laboratory 1. The results from laboratory 7 include two unexpected deviations (high values found on day 1 B20 and day 3 BK). All the dose/response curves are covering the 50 per cent. values. They are of the usual form except in five cases when the curve contained a plateau (about 67 per cent. changed) or unexpected low values (15-20 per cent. below an expectation of about 90 per cent. - see notes in Annex I, Table 1.7 abc). One of the two high values for sLopes corresponds to one of the above irregular curves. Laboratory 8, in addition to the dye test which was not among their routine methods, performed three other tests: Indirect Haemagglutination Test (IRA) Indirect Fluorescent Antibody Test (IFA) Complement-fixation Test (CF) The log titres for these three tests were estimated locally as end-point titres, because noquantitation of the readings per single test-tube was available. For the dye test the laboratory 8 reported the readings of the two dilutions giving results on both sides of the 50 per cent. value. For the other dilutions only a code showing ~ 20 and ~30 per cent. changed organisms, respectively, was indicated. The slopes estimated for the central part of the dose/response •.C\n;:"~~ were of the same magnitude as in the ~ain ::;~< -~""" ",,:<~:""t> assay. "", ,\ \ .~ ...... -J '; The 188ue of this document does not constitute ,: ' , <'0' Ce document ne con8titue pas une publication.
liEn]);} .
formal publication. It 8hould not be reviewed, ab8tracted or quoted without the agreement ~f the World Health Organization. Authors alon~,; are responsible for views expressed In signed
14 S{i:q ..... ~u.' resume nl d'aucune Citation sans I autorl$atlon de ,'," _ .. -. 1_ ......... n ••• AII.e
~:;: :, II ne dolt falre l'obJet d'aucun cO,mpte .ren~u ou
11'Organlsation Mondlale de la s~nte. Le~ opinions ." > exprimees dans les articles signe. n engagent
WHO/13S/G8.939 page 2 The results from the dye-test (see Table 3) were lying lower than the results for the above laboratori~s 1:4; the serum B results were - especially on day 1 - greatly below the expected value (H = B). The two other tests, IlIA and IFA (see Table 6) gave results of similar magnitude and showed also a tendency to lower results for serum B. The CF results (see also Table 6) were lying at a much lower level, and serum B was always non-reactive. ~on-reactive
sera
The distributed non-reactive serum was found non-reactive except in laboratory 8:
Day 1 Dye test IFA test
Day 2
Day 3
Average
O.Oll No. react.
0.ll6 0.903
0.938 0.602
0.355 (0.753)
These findings are in contrast to the results in laboratories 1-7. Statistical analysis An analysis of variances (Table 4), performed for laboratory 7, on the results for all sera except 2, has shown a residual variance of 0.033341 (f = 12). The variance components corresponding to sera are not exceeding the limits of significance, but a significant day-today variation was found as for laboratory 4. Laboratories 2 and 7 have standard deviations of the same magnitude sR = 0.190 which exceeds the value found for laboratories 1, 3 and 4 (0.0 6 4). It is noted that the within laboratory variation is clearly different in laboratories 2 and 7 from that found in laboratories ~ 3 and 4, but in both cases the standard error estimated will cover the differences fairly well. The analysis of variances for laboratory 8 (see also Table 4) was performed not only on the total material (3 days against 6 sera) but also after the elimination of all results for sera "B". 'rhis elimination was made for all the four tests in laboratory 8, although the above above variation for serum B results were most pronounced for the dye test and IFA test. The dye test and the IFA test in laboratory 8 showed much larger residual variances, 0.176376 (f = 10) and 0.300852 (f = 10), when results for all sera were tested. After the elimination of the serum B results, the residual variances were reduced to 0.108897 (f = 6) and 0.042784 (f = 6). The IHA test showed only a small reduction in residual variance, when comparing the total serum material and the restricted one. The latter gave the value 0.027684 (f = 6). The limited results from the CF test showed a residual variance 0.075276 (f = 4). It was decided to use the values from the limited material as basis for the standard error for single log titres. TIle values were: the dye test: 0.330, the lHA and IFA tests a common value: 0.188, the CF test: 0.274. From those standard errors the significance of the differences between log titres given in Table 5 and the following tables was estimated. The average results per serum from laboratories 1 to 4 and 7 have been examined in an analysis of variances which has given a residual variance of 0.008495 (f = 24) (Table 4). This result does not deviate from the value 0.008256 found for laboratories 1 to 4, and it is of no importance whether the whole material is examined on the basis of the sR = 0.091 or = 0.092. 1-4 The results from laboratory 8 have not been included in such an overall analysis due to the excess variation.
WHO/BS/68.939 page 3 All the relative potencies for the dye test in laboratory 7 and for the sera H20 and B20 in laboratory 8 (Table 5a) are within the ranges of the relative potencies for laboratories 1 to 4 (-0.20 to 0.25). High relative potencies were noted for sera H <0.50) and HAT (0.43) in laboratory 8, while serum B (days 2 and 3) gave a value (-0.49) which is below the above group of relative potencies. The results of IHA' IFA and CF tests gave relative potencies (Table 7a) within the above ranges in 7 out of 8 well-described cases. All results for sera Band B in these three tests were very low (outside the above ranges). It is noted that the es~~mate of relative potencies has compensated for the low testing level in laboratory 7 and 8. The differences B - H are for the dye test results noted in Table 5b and for the HA, IFA test results in Table 7b. The differences found for laboratory 7 are corresponding to those seen for laboratories 1 - 4, while it is remarkable that many of the B - H differences for laboratory 8 are passing even the 0.1 per cent. limit of significance. In this connexion it has been considered whether the reconstitution in laboratory 8 of the freeze-dried material in some of the ampoules Band B2 has been less effective or has given a less stabile solution causing values below the expe2tation. It is not likely that such a plain explanation covers the deviations found, because the four tests in laboratory 8 were performed, per testing day, on material from one single ampoule of B and another of B • A plain loss of potency of the material from say the B ampoule on day 1 should have resu~~ed in a general reduction of all B titres on that particular day, but it is only for the dye test and for the IFA test that such major reductions in titre were noted. Laboratory 7 performed also on the first day a comparison between the effect on the local accessory factor and that of the one distributed from Copenhagen. It is clear from Table 5c that the titres are higher when using the local accessory factor; the average difference (0.47) is passing the 0.1 per cent. limit of significance. Table 8 gives a compilation of all average log relative potencies from the whole study; all variations are random except for serum B in the CF test. Conclusion The results from laboratory 7 and laboratory 8 are supplementing the picture seen for laboratories 1 to 4 regarding the random variations. Three laboratories showed in the dye test a low standard error, 0.084, while two had a considerably higher value, 0.190, and one laboratory, which did not-use the dye test as a routine method, had a still higher standard error, 0.330. The standard error for the IHA and IFA tests are lying at 0.188, while the CF test has a value at 0.274. The titre variations from laboratory to laboratory are greatly reduced when the results are expressed as potencies relative to the log value for serum H , but some of the results from K laboratory 8 gave relative potencies outside the expected range. The average relative potencies for the dye test in laboratory 7 for serum H and serum B were of the same strength, close to zero (-0.05 and +0.03). The corresponding average values for laboratory 8 were more deviating from zero (0.34 and -0.12) but not passing the limit of significance. Serum HAT showed in laboratory 3 a relative potency Significantly (5 per cent. level) larger than zero and all other dye test results were larger than or equal to the serum H results. The average for serum HAT including all six laboratories was 0.19 which is passing the 5 per cent. limit of Significance. The additional tests from laboratory 8 gave for serum B lower average relative potencies than the dye test.
i
WHO/BS/68.939 page 4 Stability The stability of the International Standard for Anti-Toxoplasma Serum was examined October 1967: two ampoules were stored for 1 month at each of the temperatures _20°, +37' and +56·C. The testing of the reconstituted material (under code numbers) showed the following titres: Log titres Storage temperature 1st testing day _20° C +37° C +56'C 3.43* 3.43 3.34 2nd testing day 3.47 * 3.44 3.47 Average loss in: Log potency per cent.
/ 0.015 0;045 3.5 11.0
*
Average of results from two ampoules.
The average losS in potency was regarded as of no importance and the material was found acceptable regarding thermostabi1ity. The International Unit The content of dry material in the ampoules of the International Standard was examined for 10 ampoules: (1) (2) ( 3) (4) (5) l76.45mg 181.93 .. 181.61 .. 176.71 .. 190.73 .. Average: ( 6) (7) ( 8)
176.26 mg 183.50·" 185.26 .. 177.77 189.12
(9) (10)
.. tI
181.9340 mg
The variation between ampoules was described by the standard deviation 5.27 mg (2.9 per cent.). The participants have in the autumn 1967 accepted a unit corresponding to 2000 I.U. per ampoule. Thus one I.U. corresponds to 0.090967 mg.
When the total content of one ampoule is reconstituted by 2.00 ml distilled water, the solution will contain 1000 I.U. per ml.
WHO/BS/68.939 page 5
LIST OF TABLES
Table I Table 2 Table 3 Table 4 Table 5a Table 5b Table 5c Table 6 Table 7a
Participating laboratories Anti-toxoplasma sera tested in the assay Dye test; Dye test; Dye test; Dye test; Dye test; IRA, log titres per testing day and serum (labs 7 and 8) analysis of variances (labs 7 and 8) IRA, IFA and CF tests (lab. 8)
average values per serum and log relative potencies (labs 7 and 8) differences between log titres B-H (labs 7 and 8) comparison between Accessory factors (lab. 7) per serum and day (lab. 8) average titres per serum and log relative
IFA and CF results;
IRA, IFA and CF results; potencies (lab. 8) IRA and IFA tests;
Table 7b Table 8
differences between log titres (lab. 8)
Comparison of results from all 8 labs
Annex I Table I Table 2
(available on request): The readings for the dye test for labs 7 and 8 Technical details in the dye test for labs 7 and 8
WHO/BS/68.939 page 6
TABLE 1. Atlanta
PARTICIPATING LABORATORIES IN FINAL PART OF THE ASSAY Georgia, United States of America Parasitic Serology Unit Parasitology Section National Communicable Disease Center
Syracuse
New Tork, United States of America Department of Preventive Medicine Upstate Medical Center State University of New York
TABLE 2.
THE ANTI-TOXOPLASMA SERA TESTED IN THE ASSAY
Distributed under code No. H HK H 2O B BK* B 20 HAT ,
Copenhagen pool of human serum
H 6, 1, 9, 12, 14 16
"
"
"
"
"
H diluted 1 + 19
Copenhagen bovine serum
B 2, 3, 11,
"
"
"
15
B diluted 1 + 19
7,
18
Copenhagen bovine serum (different from B) non-reactive pool of human serum Serum BK not ditributed to lab. 8.
5,
10,
13
Neg
4,
8,
17
*
The sera H, B and HAT had titres about 1200 (log titre 3.08), while sera H and B were 20 20 expected to give titres about 60 (log titre 1.78).
WHo/BS/68.939
page 7 TABLE 3. DYE-TEST RESULTS FROM LAB. 7 & 8 AS LOG TITRE PER DAY AND PER SERUM H K 1.301 + H 20
H Laboratory 7 Micro a Day 1 Cop.a Day 2 localDay 3 local Laboratory 8 Day 1 Day 2 Day 3
B
BK
1.301 + B 20
RAT
L :
Average per day
3.028 2.957 3.301 2.472 3.096 3.142
2.922 3.244 3.265 2.150 1.940 3.130
2.966 3.273 3.037 2.526 2.337 2.906
2.906 2.881 3.494 0.969 1.870 2.224
2.806 3.341 3.637
3.313 3.204 3.672 1. 326 2.327
3.207 3.451 3.569 3.035 2.428 3.056 9.467
, 2.919 3.215 3.232 3.153 3.370 3.398
3.021 3.193 3.425 2.080 2.333 2.947
SE for titre differences lab. 7:
3.225 0.269, lab. 8:
~ Origin of Accessory factor. TABLE 4. Laboratory 2
MAIN RESULTS OF ANALYStS OF VARIANCES FOR THE DYE TEST, THE IRA, IFA AND CF TESTS (labs 7 & 8) Residual variance s R 2
Between days 2 8.6 3.9 5.7 13.9 1.9
'2 ~
Between sera s S 0.060949 0.158149 0.089598 0.007726 0.203402 6 3 3 3 2 2 2
sR 12 6 6 6 4 0.183 0.330 0.166 0.207 0.274
s D 0.287536 0.428001 0.158552 0.596456 0.144005 2 2 2 2 2
s Dis R xX
s Sis R 1.7 1.5 3.2 0.2 2.7
2~
7 8 8 8 8
b DYE testc DYE testc IRA testc IFA testCF tes~ The index for s 2
0.033341 0.108897 0.027684 0.042784 0.075276
values indicates the degrees of freedom (f). 2 2 Weighted average of s R for laboratories 2 & 7: s R :::; 0.036271 3 2 0.035234 Weighted average of s R for IRA & IFA methods : s R (7 sera by 5 laboratories) Residual variance 0.008495 (f :::; 24)
s
R.
= 0.190 0.188
sR
Laboratories 1,2,3,4 & 7
I
I
sR 1 _ 4 & 7 = 0.092 a
In the original document are the variance ratios denoted by v
2
b 7 sera and 3 days
~ 4 sera and 3 days
WHO/BS/68.939 page 8 TABLE 5a. THE DYE TEST - LABS 7 &< 8 AVERAGE VALUES PER SERUM AND LOG RELATIVE POTENCIES BASED ON HK Deviations from average value H H K 1. 301 + H 2O B 1.301 + B 20
fLaboratory
BK
HAT
pveral
~ver~ge
:
days 1,2,3 a log re1.pot.8 : days 1,2,3 b log rel.pot.8 : days 2 &< 3 c log rel.pot-
3.095 -0.049
3.144
3.092 -0.052
3.094 -0.050
3.261 -0.117
3.396 0.252
3.409 x
3.213
0.265 d 0.081
-
2.903 0.496
2.407
2.590 0.183
2.840 0.433
2.685 0.371
3.119 0.584
2.535
2.622 0.087
2.047 -0.488 0.130 B
2.776 0.241
2.742 0.207 1.301 + B20 1.326 )-0. 824xx)
2.640 0.126
~ SE for differences between average results b
= \,,2 x 0.008495
Day 1 gave B values significantly below expectation:
log ti tres log reI. pot.
0.969 xxx -1.181
(SE for differences between single results 0.266) Average for B values days 1, 2 &< 3 were not estim ated, due to the unexpected low values on day l. SE for differences between average results (days 1, 2 & 3) 0.270
£ d
SE for differences between average results (days 2 &<.3
0.330
x,
Significant deviation from zero passing 5, 1 &< 0.1 per cent. limits were marked with xx &< xxx respectively.
TABLE 5b.
a Laboratory 7-
Laboratory
#
DIFFERENCES BETWEEN LOG TITRES IN THE DYE TEST (LABS 7 &< 8) B H (B ovi Human Serum H) ne S erum agalnst Average per day Single values per day B 20 - H2O B - H XXX day 1 0.347 -0.119 -0.069 0.011 " 2 xxx xXX 0.635 0.283 " 3 XX x _1.200 day 1 -1.342 -0.010 -0.648 " 2 0.319 -0.912 " 3
-
- -
I
SE b - B replaced by B;
~
0.092/';; = 0.065 SE = 0.330 x B - H
0.092 0.467
Vi":5
= 0.563
Average per laboratory Laboratory 7 Laboratory 8 diff. SE diff. SE 0.117 0.130
B 0.304 0.130 XXX x
-
H
-
-0.001 0.130 _1.215 0.270
0.058 0.130/12
0.092
-0.297 0.270
WHO/BS/68.939 page 9 TABLE 5c. COMPARISON BETWEEN TWO ACCESSORY FACTORS (LAB. 7) (Local titre estimates)
*
Log titre for sera from day 1 ccessory factor H
1.301 + H
B
20 3.107 3.107 0.000 2.806 2.806 0.000
1.301 + B 20 3.204 4.107 0.903
HAT
Average
,
Flom Copenhagen Local Difference
2.806 3.709 0.903
3.408 3.408 0.000
2.903 4.107 1.204
3.408
3.092 3.565 0.473
?3. 709 ~0.301
The variation between sera may correspond to the principle of the estimates (last reactive dilution) . TABLE 6. H~
*
RESULTS FOR IHA, IFA &; CF TESTS (LAB. 8) LOG TITRES PER DAY AND SERUM 1.301 + H 20 2.505 2.505 2.505 3.408 2.505 2.505 B 1.301 + B 20 2.204 1.903 2.204 Sl.452 1.602 2.204 HAT L Average per day 6 sera 2.340 2.491 2.641 2.441 2.190 2.491 0.70# 0.753 1.104
H K
IHA Test: Day 1 Day 2 Day 3 IFA Test: Day 1 Day 2 Day 3 CF
2.408 3.010 3.010 3.311 2.408 2.709
2.408 2.709 3.010 3.010 2.709 3.010
1.806 1.806 2.107 ~0.151
2.709 3.010 3.010 3.311 2.408 2.709 0.903 0.903 1.204
2.408 2.408 2.709 2.709 2.709 3.010 1.806 1.806 1.806
1.505 1.806
Test: Day 1 0.602 Sera H ' Band B 0.602 2O 20 Day 2 ,!:0.151 1.204 non-reactive 0.903 Day 3 1.204 in CF test a - Average for CF-test based on 3 sera per day.
TABLE 7a. IHA, I FA AND CF TESTS IN LAB. 8 AVERAGE RESULTS AND LOG REL. PCYI'ENCIES BASED ON HK PER SERUM H H K 2.709 1.301 + H 2O 2.505 -0.204 2.910 2.806 -0.104 1.003 B 1.301 + B 20 2.104 (2.054) _0.605 (-0.806) (1.903) (-0.957 Average H,H &; HAT K 2.809
HAT 2.910 0.201 2.809 -0.101 1.003 0.000
IHA~ log reI. pot. a I FAlog reI. pot. b CFlog re1. pot.
2.809 0.100 2.809 -0.101 0.552 -0.451 c
1.906 (1.957) _0.803 (-0.903) (1.656) (-1.204) no reaction
xXX
xX
2.843
0.853
x
~ Values in ( ) are based on results from days 2 and 3 and are all passing the 0.1 per cent. limit of significance. SE for differences between average of 2 results of 0.188; of 3 results 0.152. b - SE for differences of average between 3 results 0.223.
.£ Significant deviation from zero passing 5, 1 and 0.1 per cent. limits were marked with x, xx and xxx respectively. -... ...... ' ,
WHO/BS/68.939 page 10 TABLE 7b. DIFFERENCES BETWEEN LOG TITRES IN IHA & IFA TESTS (LAB. 8) B - H (BOVINE SERUM AGAINST HUMAN SERUM IHA B - H
SE x xXX xXX 0.266
IFA
SE 0.266
day 1
-0.602 _1.204 _0.903 -0.301 -0.602 -0.301 -o.903
"
2
_0.903 xXX _0.903 xXX
" "
3
day 1
2 3
x
_0.903 -0.301
xXX
" Average B - H
xXX
0.153
_0.903 _0.602
xxx xx a
a
0.188
" a
-0.401 xx - Average for days 2 and 3.
~ As above test results are based on end-point readings of twofold dilution series, the difference of 0.301 is the smallest to be observed and therefore the one-x significance is of dubious value. TABLE 8. COMPARISON OF RELATIVE POTENCIES LABS 1-4(-6) AND 7 & 8 H & H 2O DYE test labs 1 - 4 CF test labs 2, 4 DYE test lab. 7 Dye test lab. 8 IHA test IFA test CF test -0.16 to 0.08 B(B ) & B ~ 20 K 0.13 to 0.18 -0.65? to 1.10? 0.03 HAT 0.00 to 0.27 0.20 to 0.29 0,27
-6
-0.06 to 0.20? -0.05
0.34
(~0.12)
0.43
-0.05 -0.10 ( -0.45)
-0.70 ( -1.08) no reaction
0.20 -0.10 0.00 value in ( ) corresponds to days
I
- Average for B in lab. 8 based on Band 1.301 + B ; 20 2 and 3 only.
a