0264-410X(94)00084-0

Vaccine, Vol. 13, No. 14, pp. 1346-1352, 1995 Comriaht 0 1995 Elsevier Science Ltd Printed in &eat Britain. All rights reserved 0264-410X/95 $lO+O.OO

Assessment of foot and mouth disease vaccine potency by liquid-phase blocking ELISA: a proposal for an alternative to the challenge procedure in Argentina
B. Robiolot, P.R. Grigerat, O.H. Periolot, E. Maradeiz and J.L. La Torre*t C. Sekit, T. Bianchil,
The lowest expectedprotection (LEP) at a 95% confidence of 245 foot and mouth disease (FA4D) commercial vaccines was calculated from the titres of liquid-phase blocking sandwich ELISA (1pELISA) of cattle sera obtained from 3920 animals at 60 days post-vaccination (d.p.v.) and challenged with live virus at 90 d.p.v. It was found that LEP evaluation is highly spectfic (i.e. it is able to predict the failure in 100% of the cases) although its ability to predict the challenge (PG test) approval (i.e. sensitivity) comprised only 65% of the vaccines that passed the trial. It was possible, nevertheless, to improve the sensitivity of the evaluation by using an alternative coeficient (Ro), exclusively dependent on the number of animals exhibiting the highest and lowest 1pELISA titres in a particular vaccine trial. This coe#icient was capable of predicting the PG approval of 90% of the vaccines, yet maintaining acceptable levels of safety (87% of specificity). Based on these results and as a first step towards the replacement of the challenge protocol in Argentina, we propose a swift approval for commercialization of FA4D vaccines which are able to reach the highly restricting LEP passmark of 82%, and the rejection of those not reaching the 50% LEP limit. More extensive experience with this new protocol will allow a finer adjustment of the LEP and Ro values and to set more precisely the cut-oflpoints for direct approval or disapproval of vaccines by lpELISA, eliminating the use of live FMDV in the field.
Keywords: Foot and mouth disease;

Ip ELISA

predicting

vaccine potency;

protective

immunity;

live virus challenge;

FMDV

vaccine quality

In Argentina current tests for evaluating the potency of FMD vaccines are very costly and cumbersome. They require long-distance transportation of thousands of cattle to allow challenge with infectious prototype strains of FMDV. Although reliable in showing the immunogenic properties of the vaccines, the challenge procedure gives little information about mucosal immunity against natural airborne infections2 and can also be criticized with regard to the likelihood of releasing infectious viruses into the environment. In addition, ethical questions and increasing pressure against this test are raised by animal welfare organizations3. We have already reported experimental data on 1634 vaccinated animals which provided strong support for
*Corresponding author. tCentro de Virologia Animal (CEVAN), Serrano 669-1414-Capital Federal, Argentina. SServicio National de Sanidad Animal (SENASA), Av. Fleming 1653-1640-Martinez Pcia, Buenos Aires, Argentina. (Received 20 April 1994; revised 7 November 1994; accepted 18 November 1994)

the use of a lpELISAh6; for the evaluation of FMD vaccine potency in Argentina, suggesting that this could be a first step towards the elimination of the challenge procedure. In this work, we extend our previous studies by presenting two independent numerical coefficients that offer a high degree of reliability and safety in predicting the potency of commercial vaccines in the absence of live virus challenge. The results are based on 1pELISA titres of 3920 vaccinated/challenged cattle. MATERIALS
Viruses

AND

METHODS

Production of the specific antigens for the 1pELISA test have been described previously. The four strains used in this study were FMDV 01 Caseros (OlCas), A Argentina 79 (A79), A Argentina 87 (A87) and C3 Argentina 85 (C85), the initial virus seeds were provided by the Argentine Animal Health Service (GELABSENASA).

1346

Vaccine

1995 Volume

13 Number

14

Assessment of FMD vaccine potency: B. Robiolo et al.

Virus challenge

The challenge protocol for measuring vaccine potency (PG test)T8 in Argentina has been previously described . Briefly, groups of 16 animals were vaccinated with a commercial vaccine containing serotypes 01, A79, A87 and C8.5, in a FMD free Southern province of the country, transported approximately 600 km to a FMDV enzootic area to be vaccinated, and then other 1000 km to the SENASA facilities in Buenos Aires where they are challenged with infectious virus at 90 d.p.v. The strict protocol for approval of FMD vaccines used by SENASA requires that 13 of 16 challenged animals (82%) show no signs of disease 7 days after challenge. If at least 11 animals were protected in the first trial, the manufacturer could apply for a second trial (retrial). Failure to protect 13 animals in the retrial leads to rejection of the vaccine.
Analysis of serum cattle by IpELISA

The EP induced by a given vaccine in a particular individual with a given 1pELISA titre is operatively defined as the percentage of protected animals with regard to the total number of vaccinated animals that share the same serum 1pELISA titre (Figure 1). The mean expectancy of protection (MEP) was then calculated as follows: MEP=X EPIn where n=16.
LEP coefficient

The lower limit of MEP to a 95% level of confidence, herein defined as LEP, was determined by the equation: LEP=MEP - (t,O.O5(n- 1) x S.E.) where t,O.O5(n- 1) is the value obtained by Students t-test for (n - 1) degrees of freedom, (n) is the number of animals challenged and S.E. is the standard error of the mean (S.D.I&). For details see Table 1.
Ro coefficient

The liquid-phase blocking sandwich ELISA used here has already been described in detail. Briefly, dilutions of test cattle sera were independently incubated overnight at 4C with a pretitrated dose of virus in a saline buffer liquid phase. The mixtures were then incubated for 1 h at 37C on wells coated with a specific polyvalent antibody and followed by a standard ELISA titration with a second specific antibody. Colour-developing reagents for detecting the virus that did not react with the cattle serum in the initial step and was trapped on the wells were added. The original technique was modified by replacing the specific rabbit polyvalent antiserum for a guinea pig antiserum as trapping agent and using a mixture of four different monoclonal antibodies (mAbs) specific for each of the standard FMDV strains tested, as secondary antibodies. Virus stocks from the different strains ( z 1 mg ml ~ ) were used either in a 1:200 (01 and C85 strains) or 1: 100 (A strains) dilutions. In all cases each serum was assayed using the four viral vaccine strains (see above).
Reproducibility of the test

Coefficient Ro was empirically derived by fitting the actual values of protection against generalization (i.e. number of animals protected/total number of animals challenged) with respect to the corresponding 1pELISA titres for each individual serum. Ro was expressed as the ratio between the number of animals showing a lpELISA serum titre equal to or greater than 2.10 and those showing a titre equal to or less than 1.80. For the specific case of 01 Caseros these values were found to be 2.00 and 1.70, respectively. It is worth noting that unlike the LEP coefficient, the calculation of the Ro coefficient is not dependent on the protection percentages showed in Figure I.

RESULTS
Correlation between protection and IpELISA titres from 3232 control and vaccinated cattle

Titration of each standard virus is repeated in every test to ensure conditions of maximum sensitivity. The test was highly reproducible as is demonstrated for the low dispersion shown by the positive control sera reacting against the four FMDV strains. The mean values and the standard dispersion obtained for the same batch of positive control sera for each of the four FMDV strains in 30 independent assays (n=30) are: A79: 1.90f0.11; A87: 1.99*0.12; C85: 1.72+0.08 and 01: 1.73 & 0.11. Twenty different batches of positive control sera have been tested with similar results. In order to mantain a high degree of standardization, four control sera of known titres were assayed simultaneously, as internal standards, in each ELISA plate. In addition it is worth mentioning that the introduction as a second antibody, in the 1pELISA procedure, of a pool of monoclonal antibodies improved the technique, contributing to the obtention of low dispersion values. Expectancy of protection (EP), LEP and Ro coefficients: the EPs and Probit values were originaly calculated from the results of 3232 sera plotted in Figure 19.

Figure 1 shows the proportion of protected to non-protected individuals in groups with a number of animals sharing the same 1pELISA titre at 60 d.p.v. The results show that as the 1pELISA titres increase there is an increase in the ability of animals to survive virus challenge. The large size of the sample together with the modification of the test (see Materials and Methods), allowed for the use of 0.1 titre unit intervals in the graph which gave a more precise estimate of the particular expectancy of protection (EP) for each individual vaccine. This study showed that the relationship between lpELISA serum titres and percentages of protected animals is almost identical for all four different FMDV prototype strains (Figure 1), including the A79 strain, which was shown to be different from a previous study. These results prompted us to carry out a more rigorous statistical analysis of the raw data presented in Figure 1, to find a practical and reliable numerical coefficient capable of predicting the potency of FMD commercial vaccines in the absence of challenge.
Experimental challenge correlation between Ro, LEP and

As reported in Materials and Methods, each vaccine was assayed in a group of 16 bovines. Animals were bled

Vaccine 1995 Volume 13 Number 14 1347

Assessment

of FMD vaccine potency: B. Robiolo et al. A79 n = 689

A79 n = 589

20 10 0 <O.Q ~1.51.5 1.6 1.7 1.8 1.9 2.0 2.1>2.1

n =

e1.5

1.5

1.6

1.7

1.8

1.9

2.0

2.1 ~2.1

(Contra) A&J
1 I,

766

A87 n = 766

20 10 0 co.9 <1.51.5 1.6 1.7 1.8 1.9 2.0 2.122.1

o,

<1.5

1.5

1.6

1.7

1.8

1.9

2.0

2.1

>2.'

(Control)

cas

= 1014

01 Cas

n = 1014

20 10 0 co.9 <1.51.5 1.6 1.7 1.8 1.9 2.0 2.122.1

n =

!J---__
<1.5 1.5 1.6 1.7 1.8

1.9

2.0

2.1 ~2. 1

(Cotio) (-85
110 1M) 90 60 3 E60 850 $40 II 30 20 10 0 co.9 (Control) 70

863

C85 ._
1M) /

n = 863

~1.51.5 1.6 1.7 1.8 1.9 2.0 2.1>2.1

0.

<1.5

1.5

1.6

1.7

1.8

1.9

2.0

2.1 ~2.1

IpELlSA titre (log 10)

ELISA titre (log 1O)-

Figure 1 (A) Correlation between protection and IpELlSA titres in vaccinated/challenged animals. Cattle (n=3232) were vaccinated with different lots of tetravalent oiled vaccines, bled and challenged with live virus (for details see Materials and Methods). Each bar represents (a) A79 (n=589; protected, n=509, groups of animals sharing the indicated titre (reciprocal log 10). q , protected; n , non-protected. non-protected, ~80); (b) A87 (1~766; protected, 1~560, non-protected, 1~206); (c) 01 Cas (n=1014; protected, n=834, non-protected, n=180); (d) C85 (-863; protected, n=706, non-protected, n=157). (B) Correlation between protection and IpELlSA titres calculated by Probit (x-x) experimental values and (m-W ) theoretical values

at 60 d.p.v. and the 1pELISA titres against each of the four vaccine viruses were determined. From those values Ro and LEP coefficients were calculated; in addition,

each particular titre was correlated with protection (P) or non-protection (NP) after challenge with live virus at 90 d.p.v.

1348

Vaccine 1995 Volume 13 Number 14

Assessment
Table 1 Correlation of LEP and Ro coefficients, IpELlSA titres A79 1.65 >2.10 >2.10 22.10 >2.10 >2.10 >2.10 r2.10 >2.10 >2.10 >2.10 >2.10 >2.10 2.10 cl .50 2.00 6.5 A87 cl .50 1.95 1.70 1.70 1.80 1.90 1.70 1.70 >2.10 1.70 <1.50 2.00 2.10 <1.50 cl .50 2.10 01 1.70 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 >2.10 1.65 2.00 7.0 G85 1.55 >2.10 >2.10 >2.10 22.10 >2.10 >2.10 >2.10 >2.10 >2.10 s2.10 >2.10 22.10 2.00 1.50 2.10 calculated on individual

of FMD vaccine potency: B. Robiolo et al.

IpELlSA titres, with protection Expectancy A79 75.0 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.9 97.7 30.0 84.8 83.7 of protection A87 21.3 83.8 67.3 67.3 66.7 83.8 67.3 67.3 97.7 67.3 21.3 89.8 93.9 21.3 21.3 93.9 52.2 (EP) 01 80.6 98.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 96.9 72.7 96.7 91.3 AP ND, not done C85 44.8 98.4 98.4 98.4 98.4 98.4 98.4 98.4 98.4 98.4 98.4 98.4 98.4 95.7 44.8 93.1 83.2 AP

Animal number 1 2 3 4 : 7 8 9 10 11 12 13 14 15 18 LEP (a)

RO

0.3

6.5 ND NAP

PG test (b) *P: Protected; NP: non-protected (for details see Results). Lowest EP (%). AP: approved;

NAP: non-approved.

Table 2 A

Double-entry

analysis of LEP vs PG test for 245 commercial

vaccines B

PG test iPPR b RETRIAL DIS APP 4 8 3 15 b: Approved 1 5 3 9

C

PG test DISAPP iPPR b RETRIAL APP DIS 5 7 3 15 2 4 3 9

C

DISAPP

LEP

32% 62-82 <62%

H M L

124 60 7 191

0 15 15 30

129 88 28 n245

LEP

z-75 % 62-75% <62 %

H M L

157 27 7 191

4 11 15 30

168 :: n 245

a: Approved

(first instance).

and disapproved

after retrial. c: Disapproved

(first instance)

Table 1 shows 1pELISA titres and their respective EPs, Ro, LEP and challenge results obtained with 16 cattle immunized with a commercial vaccine and challenged with virus A87. As can be seen EPs, Ro and LEP values for viruses A79, 01 and C85 were very high, whereas they show rather low values for virus A87. The results of the challenge for virus A87 showed that only 10 out of the 16 challenged cattle (10116) were protected. In this particular case two additional groups of 16 cattle were vaccinated with the same vaccine batch and challenged with viruses 01 and C85, obtaining in the PG test values of 15/l 6 and 16/l 6, respectively (data not shown). Due to the failure with one virus (A87) this vaccine batch was rejected. Prediction of protection for 245 commercial vaccines (3920 cattle) LEP coeficient. The lower limit of the mean expectancy of protection (LEP) was calculated for 245 vaccines tested (representing 3920 animals) and their values were plotted as a function of either approval or rejection of each individual vaccine. In Figure 2, the distribution of individual LEPs has been defined as three regions, termed high (H), medium
Figure 2 Correlation between LEP and PG performance for 245 commercial vaccines (3920 animals). LEP values of 245 individual vaccines (see Materials and Methods) plotted as a function of their 0 approval or n rejection in PG trial

(M) and low (L) according to their respective proportions of approved and non-approved vaccines. The H region (n=129 vaccines), with LEP values above the 80% level, contained almost exclusively approved vaccines (128 of 129) while the M region (n=88) (LEP between 60 and 80%) showed a significant ratio of

Vaccine 1995 Volume 13 Number 14 1349

Assessment
Table 3 A

of FMD vaccine potency: B. Robiolo et al.

analysis of Ro vs

Double-entry

PG test for 245 commercial vaccines

B

PG test iPPR b RETRIAL APP DIS 4 8 3 15 b: Approved

C

PG test DISAPP iPPR b RETRIAL APP DIS 7 z 15 4 2 3 9

C

DISAPP

Ro

A.5 4.5-0.5 <0.5

H M L

96 94 1 191

& 3 9

0 15 ;z

101 122 22 n 245

Ro

>l .o 0.5-l .O% <0.5

: L

172 18 1 191

4 :: 30

187 36 22 n 245

a: Approved

(first instance).

and disapproved

after retrial. c: Disapproved

(first instance)

non-approved to approved vaccines (20 and 68). The L region (n=28), below 60%, contained the majority of rejected vaccines (10 approved, 18 disapproved). A more quantitative analysis of the data appearing in Figure 2 was obtained by using a double-entry diagram (Table 2A). The lower limit of the H region was arbitrarily set at 82% in order to match the approval percentage (i.e. passmark) required for the highly restrictive official trial (see Materials and Methods). Table 2A also shows vaccines that were initially sent to retrial by the Animal Health Authorities (either 11 or 12 animals protected out of the total 16 challenged), which were either approved or rejected in the second instance. The grouped LEP values showed that 124 out of the 191 (65%) vaccines initially approved by SENASA had values above the 82% passmark (H), whereas 60 were in the range between 62 and 82% (M) and I below the 62% level (L). This distribution also showed that none of the LEP values of the 30 vaccines rejected by the trial in the first instance were in the H area, suggesting that under the restrictions already described, the LEP coefficient could be considered a reliable and safe (i.e. specific) predictor of vaccine failure. Under these conditions, however, approximately one-third of the approved vaccines (67 out of 191) showed LEP values below the 82% LEP limit. Table 2B shows the variation of these same parameters when a less severe LEP limit of 75% is used. Under these conditions, 157 vaccines of the 191 approved vaccines (82%) and 4 of the 30 disapproved vaccines (13%) attained the H area defined by this new lower limit. Approval of vaccines based on a hypothetical 75% LEP passmark would therefore increase the sensitivity of the procedure (i.e. the ability of the indirect immunological method to predict the approval by the PG procedure) from 65 to 82%; this, however, would be associated with a decrease in the specificity (i.e. ability to predict rejection) from 100 to 87%. Analysis of the LEP values of vaccines sent to retrial confirmed the reliability of the IpELISA-based coethcient to reveal the protective ability of the vaccine. Five of the 129 vaccines that exhibited LEP values above 82% were sent for retrial by the authorities and four were found to pass the new challenge, thus confirming the LEP prediction. It should be noticed that the remaining vaccine did induce protection in 12 animals, being rejected by the SENASA PG standards.

50

150 100 Vaccine number

200

245

1991

1992

1993

Figure 3 Correlation between Ro and PG performance for 245 commercial vaccines (3920 animals). Ro values of 245 individual vaccines (see Materials and Methods) plotted as a function of their q approval or n rejection in PG trial

It is also interesting to note that 15 of the 30 disapproved vaccines were located in the M area of the graph in Figure 2, suggesting that the 1pELISA was able to discriminate the relative efficiency of vaccines among those that failed to pass the trial. More experience is needed to set a more rigorous limit in this region to minimize the rejection of vaccines with good protective properties, thus increasing the sensitivity with no harm to the overall safety of the evaluation.
Ro coeficient. Figure 3 shows the distribution obtained from the analysis of the same 245 vaccines grouped according to the Ro coefficient. This coefficient, like the LEP, is based exclusively on 1pELISA serum data. The distribution of Ro values from approved and non-approved vaccines defined three areas in the graph: an area with Ro values above 4 with the majority of approved vaccines; an intermediate region between 4 and 0.5, in which both approved and non-approved vaccines occur, and an area with Ro values below 0.5 containing almost exclusively rejected vaccines. The double entry analysis in Table 3A shows that none of the 30 rejected vaccines showed Ro coefficicients greater than 4.5 and suggested that this value could match the 82% LEP passmark described above. This Ro limit is, however, less suitable for the evaluation of vaccines since only 96 of the 191 that passed the trial (50%) in the first instance showed Ro values equal to or greater than 4.5. This parameter was increased by setting a passmark limit of R=l, as shown in Table 3B. Under

1350

Vaccine 1995 Volume 13 Number 14

Assessment

of FMD vaccine potency: B. Robiolo et al.

these conditions, 172 of the 19 1 approved vaccines (90%) exhibit Ro values higher than the unit. On the other hand, the amount of non-approved vaccines having Ro values above 1 increases to 4, or 13% of a total of 30 vaccine and rejected by the officials. The results suggested that Ro evaluation could be more sensitive than the LEP under conditions in which it is possible to reduce moderately the stringent margins of specificity reached by the 82% LEP passmark.

DISCUSSION Replacement of the current PG test evaluation of FMD vaccines by a simple serological method capable of predicting the vaccine performance in the absence of challenge is highly desirable2*7.0~1 . Besides avoiding the use of live virus, the alternative protocols should guarantee that vaccines released for the market are at least as potent as those approved by the PG test and that they have good protective activities which correlate to the PG test in terms of sensitivity. This would minimize the rejection of vaccines that would have otherwise passed the current standard trial, thereby saving public and private resources. The new protocol should also be capable of assessing the potency using the four different prototype strains used in the formulation of commercial FMD vaccines in Argentina. This would have a clear advantage over the PG protocol in which only one of the four strains is tested. Previous data obtained from the analysis of approximately 1600 vaccinated and challenged animals suggested that the 1pELISA method developed by Hamblin et al.M could meet these requirements. The availability of an enlarged sample of 3920 vaccinated/challenged cattle allowed a better statistical analysis of the significance of the correlation between IpELISA titres and protection. In this regard LEP and Ro coefficients were found to be highly reliable in predicting vaccine potency. None of the vaccines that failed to pass the challenge had LEP or Ro values above 82% or 4.5, respectively, which strongly suggests that setting these limits as the passmark values for approval would be safe enough to prevent vaccines with poor protective characteristics to reach market. Interestingly, starting in 1992 and as a consequence of a collaboration between our laboratory and the SENASA authorities, the challenge FMDV strain for vaccinated animals was selected from the four different prototypes the one exhibiting the lowest Ro coefficient (i.e. oriented selection) at 60 d.p.v. The change in the ratio of approved to non-approved vaccines during that period suggests that rationalization of the selection introduced by the use of the Ro coefficient instead of a random selection, resulted in a more rigorous trial and in the rejection of vaccines that would otherwise have passed the challenge control using one of the other three strains for challenge. The decrease in the relative number of rejected vaccines during the last part of 1993 suggests an improvement of the overall quality of the vaccines sent to trial by the veterinary companies as a consequence of a more restrictive protocol. It is important to realize that the possibility of carrying out a selective choice of challenge strain based on use

of the LEP or Ro coefficients is equivalent to testing the protection against the four different prototype strains using the same serum sample. This is of critical importance in the production of vaccines targeted to those animals present in enzootic areas where the relative proportion of the different FMDV field strains changes periodically. Based on our data and as a first step in the progressive replacement of the current FMD vaccine challenge trial, we suggest that commercial vaccines which are able to induce LEP values above the 82% passmark (or Ro 2 than 4) at 60 d.p.v. should be directly approved for commercialization, and that vaccines exhibiting LEP values below 50% (or Ro < 0.5) should be rejected. The remaining vaccines with intermediate LEP or Ro values should be reassessed using other selected challenge strains exhibiting the lowest Ro coefficient. Results have suggested that implementation of this protocol would result in a 60% decrease in the number of animals challenged, thus reducing the load on the Animal Health Sanitary facilities and allowing more time to make a more thorough analysis on a reduced number of FMD vaccines sent to trial. At a later stage and with the overall improvement of vaccine quality, a less severe Ro passmark of 1 could be used to minimize the chances of rejecting vaccines with acceptable potency, thus avoiding an eventual shortage of the supply of FMD vaccines at times where demand may increase. Finally, we would like to emphasize that the reliability of 1pELISA to evaluate the potency of FMD vaccines as well as it easy adaptability for many different virus serotypes (0. Periolo, personal communication) could lead, in the near future, to the complete replacement of the challenge procedures not only in Argentina but in other regions of the world, allowing the total elimination of live virus specially for vaccine testing, eliminating therefore the dangers of its massive use even in regular biosafety facilities.

ADDENDUM

IN PROOF

While this paper was sent for publication we have extended our studies to 435 vaccines series involving a total of 7390 vaccinated/challenged cattle, confirming the results presented in this work. Based on those figures the National Sanitary Authorities (SENASA) took the decision of gradual replacement of the PG test by the 1pELISA for the licensing of commercial vaccines in Argentina.

ACKNOWLEDGEMENTS The authors are grateful to C. De Vicenzo, M. Iglesias and M. Rodriguez for technical assistance and MS A. Llorensi for typing the manuscript. This kork was supported by the Consejo National de Investigaciones Cientificas y Tecnicas (CONICET) of Argentina.

REFERENCES
1 Periolo, O.H., Seki, C., Grigera, P.R. et al. Large scale use of liquid-phase blocking sandwich ELISA for the evaluation of

Vaccine 1995 Volume 13 Number 14 1351

Assessment

of FMD vaccine potency: 5. Robiolo et al.

bodies after infection and vaccination. Epidemiol. Infect. 1987, 99,733 Willems, R. Projet dune technique pour la standardisation du controle des vaccins antiaphteuse. Bull. Off. Int. fpiz. 1953, 40, 487 Vianna Filho, Y.L., Astudillo, V., Gomes, I. eta/. Potency control of Foot and Mouth Disease Vaccine in cattle. Comparison of the 50% protection against generalization. Vaccine 1993, 11, 1424 Sutmoller, P., Gomes, S. and Astudillo, V.M. Potency estimation of foot and mouth disease vaccines according to antibody assay results. Bol. Centr. Panam. Fiebre Aftosa 1984, 53, 49-50, 31 Pay, T.W.F. and Parker, N.J. Some statistical and experimental design problems in the assessment of FMDV vaccine potency. Develop. Biol. Stand. 1977, 35, 369 Pay, T.W.F. Quelques problems statistiques et biologiques inherents aux tests du pouvoir immunisant du vaccin antiaphteux. In: Joumees Argent. de Microbial, Buenos Aires, 7-11 June 1976, Proceedings of the AAM, BSAS, Argentina

protective immunity against aphthovirus in cattle vaccinated with oil-adjuvanted vaccines in Argentina. Vaccine 1993, 11, 754 Amadori, M., Archetti, I.L., Tollis, M., Buonavoglia, C. and Panina, G.F. Potency assessment of foot and mouth disease vaccines in cattle by means of antibody assays. Biologicals 1991,19, 191 Barteling, S.J. and Vreeswijk, J. Developments in FMDV vaccines. Vaccine 1991, 9, 75 Hamblin, C., Barnett, I.T.R. and Crowther, J.R. A new enzyme linked immunosorbent assay (ELISA) for the detection of antibodies against FMDV. I. Development of a method for ELISA. J. lmmunol. Methods 1986, 93, 123 Hamblin, C., Barnett, I.T.R. and Crowther, JR A new enzyme linked immunosorbent assay (ELISA) for the detection of FMDV. II. Application. J. Immunol. Methods 1986, 93, 123 Hamblin, C., Kitching, R.P., Donaldson, A.I., Crowther, J.R. and Barnett, I.T.R. Enzyme-linked immunosorbent assay for the detection of antibodies against FMDV. Ill. Evaluation of anti-

10

11

1352

Vaccine 1995 Volume 13 Number 14