world health organisation--supervised interlaboratory comparison

Epidemiol. Infect. (1996), 117, 69-77 Copyright © 1996 Cambridge University Press

World Health Organisation - supervised interlaboratorycomparison of ELISAs for the serological detection ofSalmonella enterica serotype Enteritidis in chickens

P. A. BARROW', M. DESMIDT2, R. DUCATELLE2, M. GUITTET3,H. M. J. F. VAN DER HEIJDEN4, P. S. HOLT5, J. H. J. HUIS IN'T VELT6,P. McDONOUGH,7 K. V. NAGARAJA8, R. E. PORTER9, K. PROUX3, F. SISAK10,C. STAAK", G. STEINBACH12, C. J. THORNS13, C. WRAY13AND F. VAN ZIJDERVELD'4'Institute for Animal Health, Compton Laboratory, Compton, Nr. Newbury, Berkshire, RG20 7NN, UK

2Faculty of Veterinary Medicine, University of Gent, Casinoplein 24, B-9000, Gent Belgium3Laboratoire Central de Recherches Avicole et Porcine, Centre Nationale d'Etudes Veterinaires etAlimentaires (CNEVA), Beaucemaine, BP53, 22440 Ploufragan, France4Stichting Gezondheidsdienst voor Dieren in Zuid Nederland, Postbus 4, 5280 AA Boxtel The Netherlands5 USDA Southeast Poultry Research Laboratory, PO Box 5657, Athens, Georgia 30604, USA

6Department of the Science of Food of Animal Origin, Faculty of Veterinary Medicine, Utrecht University,The NetherlandsI Diagnostic Laboratory, New York State College of Veterinary Medicine, Cornell University, PO Box 5786,Ithaca, New York 14853, USA8Department of Veterinary Pathobiology, College of Veterinary Medicine, University of Minnesota, 1971Commonwealth Avenue, St Paul, Minnesota 55108, USA9Animal Disease Diagnostic Laboratory, Purdue University, 1175 VADL, West Lafayette, Indiana 47907-1175, USA10 Veterinary Research Institute, Hudcova 70, 621-32 Brno, Czech Republic11 Bunderinstitut far gesundheitlichen Verbraucherschutz und Veterindrmedizin, Postfach 33 00 13, D-14191,Berlin 33, Germany12 Bunderinstitut fur gesundheitlichen Verbraucherschutz und Veterindrmedizin, Bereich Jena, D-07722 Jena,Germany13 Central Veterinary Laboratory, New Haw, Weybridge, Surrey, KT15 3NB, UK14 Central Veterinary Institute (CD-DLO), PO Box 65, NL-8200 AB, Lelystad, The Netherlands

(Accepted 27 January 1996)

SUMMARY

A collaborative exercise, supervised by the World Health Organisation, was set up to compareELISAs used for the serological detection of Salmonella enteritica serotype Enteritidis inchickens. The aim was to ascertain how far agreement could be reached on the interpretationof optical density readings for high titre, intermediate titre and low titre sera. Two sets of sera

were sent to 14 participants. The first set compared high, medium and low titre sera raised in

specified-pathogen-free and commercial broiler breeder chickens. The second set comprised 20

sera of different antibody titres raised in commercial birds reared under laboratory conditions

and sent blind. Both indirect and double-antibody sandwich blocking ELISAs were used with a

number of different detecting antigens. With a few exceptions good agreement was reached on

the interpretation of results obtained from high and low titre sera from the optical densityobtained with a single serum dilution. Differences were observed in the interpretation of

medium titre sera. The results suggested that most ELISAs produce reasonably comparableresults and that practical problems may arise from interpretation of the results mainly as a

result of the choice of the criteria used for differentiating sera obtained from infected and

uninfected chickens. These problems are discussed.

70 P. A. Barrow and others

INTRODUCTION

In many countries Salmonella enterica serotype Enteri-tidis has replaced S. Typhimurium as the predominantserotype in both human disease and in poultry [1].Bacteriological methods are generally required bynational and international legislation for monitoringprocedures to provide epidemiological evidence fromwhich control measures might be devised [2, 3]. Thedesire for detailed and rapidly obtainable informationon the infection status of flocks has led to theindependent development in several countries ofELISAs for the detection of circulating specific IgGwhose production is induced by invasive serotypessuch as S. Enteritidis, S. Typhimurium, S. Galli-narum, S. Pullorum and S. arizonae (for review see

[4]). The advantages of the ELISA over otherserological assays in coping with large numbers ofsamples of either serum or egg yolk together withminiaturization, mechanization and the availability ofhigh quality reagents has led to the extensive use ofthis system.European Union legislation [3] allows for the use of

serological methods for screening purposes providedthat the method chosen is able to provide similarguarantees of success produced by hatchery investi-gations. Bacteriological confirmation of infection isrequired nevertheless. ELISAs are already in use in a

number of countries for this purpose, particularly inthe Netherlands, where they are an integral part of theS. Enteritidis control programme, and in the UK. Insuch countries ELISAs have generated a great deal ofuseful information on infection frequencies withinand between flocks.As a result of successful standardisation of ELISA

procedures for brucellosis [5, 6], the World HealthOrganisation (WHO) recognized the value of suchassays in monitoring salmonella infections in bothpoultry and cattle. Since different ELISAs had alreadybeen developed, largely independently, it was sugges-ted at recent meetings of the WHO working group on

salmonella immunization in animals that a standard-ization process should be instigated, at the very leastto enable different groups working in this area to

agree on the interpretation of results obtained byassaying standard sera. An interlaboratory trial was

therefore set up to assess the performance of differentELISAs using a bank of sera of known antibodystatus raised in chickens against S. Enteritidis. Theaim was not to compare such assays with otherprocedures such as immunoblotting or micro-

antiglobulin [4] but rather to assess whether standard-ization procedures might be necessary and if so howthis might be done.

MATERIALS AND METHODS

Trial laboratories

A number of laboratories were contacted which haddeveloped assays and were using them either forresearch purposes or for screening poultry flocks forinvasive Salmonella serotypes.

ELISAs

Amongst the different laboratories taking part twomain types of assay were used, details of which havebeen published elsewhere (see below). Althoughdetails will not be presented in detail here, a summaryof the stages of the different ELISAs is shown in Table1. An indirect ELISA, comprising antigen-coatedplates, for the detection of IgG in serum or yolk wasused by a number of groups. A variety of detectingantigens have been used including lipopolysaccharideLPS [7-9], whole flagella [10], recombinant flagellinprotein containing the serotype specific flagellinfragment [11], SEF14 fimbrial antigen [12], outermembrane proteins [13] and disrupted whole bacterialcell proteins [7]. In the present study several partici-pants' assays used LPS preparations. Flagella antigenwas used by participants C. S., J. H. V., P. M. D.,P.H./R.P. and F.V.Z., outer membrane protein byF. S. and K. N. and SEF14 fimbrial antigen by C.T. Inaddition, G. S. used whole-bacterial-cell protein anti-gen, repeating some of the tests with LPS. The otherbasic method used, the double antibody sandwich(DAS), ELISA, uses plates coated with monoclonalantibody, followed by a pure or crude antigenpreparation. Test samples are then applied eitherfollowed by a conjugate (CT) or by a conjugatedmonoclonal antibody which will not bind if thesample contains specific antibodies [14]. Variouspreparations, such as IM mineral acid, 0 5 M NaOHor 1/20-1/50 diluted dishwashing detergent, wereused to stop the reaction. In some cases nothing wasused and the plate was read immediately.

Sera

Two sets of sera were used.The first set was prepared by oral inoculation of a

group of five 3-week-old specified-pathogen-free

Standardisation of salmonella ELISAs 71

u0

11.0 'Itt

(:i,=.-4= 00 rA. m

CL4u0

r-.o cn

6'-,== 00 ..-04,

0

o -

zz

A4E O

to to_/ _):,

c00

ceq bO

X,q

C>all

e0 0

Cd z

Cd

s- -

CT

C>

Cd ICs._,

_- CAeM0

't-CtX

Cis_

Cis=s- ca

v

0 v

Nt

_ v v ~~ oUC>

O

( o

0W

Od C

C) u_:> cvY OO

~E .-m b4 0

2 to cqt

1-104 E

A4 _ 0

WI)C)Itt

bo^

O0 b

< E= E t_ E E~~~C

O E=-E = tW-~ ~

Z

z _ Y o~~C

tn. 4I.j(Z

tL.(Z)-Q-11.*.Azt46)

Itzi;1-11-Qcn

'Ilt.

cnN-N.-444;l.

1.21(Z14.)

tn14.);6.

14.).j(Z16..Qll-.

.111

Z-3E.E.C.-44)

4

ct

Ld.d

d

Cl)

._*c_C.

1-1ulqt 01-4 $-4

U) >, ;3"4 +-jw U. z

C. U) P64

0r.

s.. 0V) >,.< co +--.U) 0.z -.14 r-Lo

,0oo I'*

;>°-r -- c

48 0 E; MO 4


(SPF) Light Sussex chickens, housed at the Institutefor Animal Health, Compton, UK, under conditionsdescribed previously [15] with 0 3 ml of a 10 mlnutrient broth culture of S. Enteritidis P125589 (phagetype 4, [9]) which had been incubated for 18 h at 37 °Cwith shaking and which contained c. 109cfu/ml.These chickens were then placed in the same pen as 25uninfected SPF chickens. All the chickens were bledevery week for 7 weeks and then every second weekuntil sufficient sera had been obtained.

This set also included sera obtained from acommercial company which had screened sera fromcustomers, some of whose birds had shown bac-teriological evidence of S. Enteritidis infection.The second set was obtained by hatching com-

mercial broiler breeder eggs from a flock free of S.Enteritidis, S. Typhimurium, S. Gallinarum and S.Pullorum. They were hatched and reared under SPFconditions using commercial food. They were infectedand bled in the same way as the first group of LightSussex chickens.

Sera were titrated using an indirect ELISA with S.Enteritidis LPS as antigen [7] and were stored frozen.They were sent to the participants containing either0-02 % w/v sodium azide (first set) or 0-02 % w/vsodium merthiolate (second set) as preservative.

Protocol of trial

The trial comprised two stages. In the first, 3 strongpositive, 3 weak positive and 3 negative sera from SPFchickens were distributed to participants. In addition,2 strong positive and 3 negative sera obtained fromcommercial chickens were sent, in view of reports thatsera from uninfected commercial chickens give highertitres than SPF birds [10, 16]. The sera were tested inthe participants' laboratories using their own ELISA,run in triplicate on 3 different days (9 times in total).The results were reported as 9 individual opticaldensity values (OD) with the blank already subtracted,together with the mean and standard deviation andthe OD values of the positive and negative controlsera in routine use.

In the second stage 20 chicken sera were sent to alllaboratories and run blind in triplicate. The resultsand their interpretation (positive or negative) of theresults were then returned.

Bacteriological data were not collected since thiswas not a field trial to assess the sensitivity of assaysin comparison with culture.

RESULTS

Stage one

The results presented in Table 2 are the arithmeticmean of the nine results from each laboratory. Inaddition, the optical density values obtained for thestandard positive and negative sera are showntogether with the standard deviation. It must be notedthat competitive blocking ELISAs (column FVZ) givelow OD values for high titre sera and high values forlow titre sera.

In general, most of the assays behaved in a similarmanner such that sera identified as high titre gave highOD values and medium and low titre sera gaveintermediate and low values. Three participants(H.V.H., R.D. and F.V.Z.) found that the mediumtitre sera gave similar values to the high titre sera. Thiswas also the case with K. N. who also recorded highvalues with the commercial low titre sera, higher thanthe values obtained by P.M. D. and C. S. for thehigher titre sera. Whether this relates to the use byC. S. and P. M. D. of flagellar antigen is unclear. G. S.found that one of the commercial low titre seraproduced a high OD with a sonicated salmonellaantigen. Other results supplied by this participantshowed that this was eliminated by using LPS asdetecting antigen.The positive control sera used by the participants

produced high OD values with two exceptions (M. G.,P. M. D.) and the negative control sera gave very lowreadings with one exception (K. N.). There wasconsiderable variation in the standard deviationsobtained by the different participants with greatervariation seen with the assays of P. B., C.W. and F. S.and negligible variation seen with the assay of P. M. D.

Stage two

Although the sera were sent blind they have beenarranged in Table 3 in order of decreasing titre. Withthe exception of the KN and CT assays the extent ofcorrelation between the results produced by thedifferent participants was again high. There was asmaller gradation between high and low OD values inone set of results (KN) and one participant (HVH)produced much higher values generally although a

strong gradation from high to low titres was demon-strated. This participant was using assay conditionsfor optimum sensitivity.The optical densities from Table 3 were interpreted

by the participants and these are shown in Table 4.


_

- N R C k N

-00~~~ C~ CV~O- ClsC

0 0 0 .0 .0 .

'Io n r-o

.00000ml

_ t_R_)_ ts

ON C) w w w

Lo 00^ N oC Co Co

mO0 w cs w

oo . -4 -_4 o

64 C~ OC~4 Cl46 l 0006

N o 00 Co 3o C - o -

b-000000__

O00000Clm00-^mo

- Cl Clo

Cl Cl 0

- Cl r^ d / 00 00 Oh 0000 £ ONFoNooo

N N 00000oo

l Cl l Cl ClC0000o

- Cl - Cl 0 o o Cl Cl

Cl l Cl N N NO O O---000000.

It Qo. -0

-C Cl0. ..

0

C.I,)

O o) 80o o o

en 0 - Cl Cl Cl4_q

tf) O' r- £% O) 0 00O00£Nb00 0000

e~nCI CQ --o

- 2. rl1 tn rnO O £Cl- 66 66

0000 99O>e£oo N'dor^ 0

N - o

66 666ZZ

_ oob00Nt00 NCl- _ NNe a o- - 666 6o

66 000N 0 0

N lt o0 o0 (oN̂ o

R - C C C o C

66qC~ 66 6AZ:

00 c oot 'tNoot- cr

0 Q- Oo-0-O- o o - - o

C o'IC0 o0

C1 -q o o -o 4

AA 666Ao0O-00 000 0t 9-lC- oo 0o-

66o oo o-oo

b00 0_ o co-o oo'oCo olC

Cl

9

6090

6

Cl

0

9

7o

09

9Cl

oo

0

60

69

69

6

Cl

6

(5IR

.)

00

.)

(5

*-

to

- C _ 4 ^

W=

c o

~~~~~~~~~ . -~~~~~~~

ot Uo

Q~~~~~~~~~Zu

90

0

en

0

6I-)

9

Cl

6Cl

0

9

0

00

060

906

0

9

00

9

6Cl

0

6

90

6Cl4

9

6

0

06

U

I-

i

C ) C

C-)

.

U

0Co

Ct C

._)

ri.

(-3

z

(

0

0

QL

0)

0

0

Cl,

0

.l

~0

Cq

U.

F-

0

0

C-

U

.-

u

,0

0

0

t--4

0

riIno0

uz0m

0

0

(U

u:

0

u

C l,

C.i

cld

"0 ce

o

._U

to

+._

Cl,

u,CO)

0

0-4

+-j

U)

rU i0

_Nq" _ M_ N4- IL .4IL .14

Qo (2en Cl4

0 Cl

00 I- l W) M QNoO-C - O'sV- r N 0 r) "o, ,eCM - r~t Cl"Oeo C -0 eW Nr-"t6 66r Cic :.6 :A LA66

0 00 N -9 9 -_0

. . 6o

0 I - Wf 00 O's W O 00- 000 o - qt -66666666o

I) 00 N- N- - - N- N 0 t qt 00 'R 000 ~c 'f) "D mON ON aN 000 O 0 N- Nl em em 1.O 9 q* 1O

90 0 ~~ - '/~ 00 oCl -C CNe~ O'00O ,. r- t--

N7- Cl0 Cl/ 0 NO- Cl Nt0 r CV~Vn00Nm ~C Cl e

,~omN OOWWM-N D-.O0-

00~ Cl l - -.,N.I0 0O--W ~ON N 4N

( mCClC - ONItt (CO C C - 09: CLCl--0000000000

C)C Cl

o o o>o o o--_ _- t

q~Tq " 'C 14 10 OC 00 00 0000 qt

,o ~o o o o o o o o o o o


F-(-5

z

-I

*(o-

(A

cn

rfi

04

(A

to

._.

cn

¢dCA

su

io

0

v2

C-n

r-0

C-0

0

u-

'0

C,,

bo

0

0

1-

--4

11.4.ct4--

o-ce10

(D0

c°a4)

0

A

II11

0.

- o

oo

) CD

o ;,-tE'0Q

0cD¢v) *-

.< m u C) w v. 0 = . " .4 -. :. Z 0 O. a04 cn .-


Table 4. Participants' own interpretation of the data in Table 3

Interpretation of absorbances by the following participants

C.W.H.v.H. M.G. P.B. R.D. G.S. C.S. J.H.V. P.H./R.P. F.S. K.N. F.V.Z. C.T.

DAS ELISATitre Indirect ELISA using, as antigen using

Serum againstno. LPS LPS LPS LPS LPS LPS WCP Fla Fla Fla OMP Fla SEF14

A 51200 + + + + + + + + + + + + + +++ + + + + +B 51200 + ++++ + + + + +++ ++ + + + + +C 25600 + ++++ + + + + ++++ +++ + + + + +D 6400 + + ++ + + + + + + + + + + + + + + + + +E 6400 + +++++ + + + + ++ + + + + + + + + +F 6400 + ++++ + + + + ++++ +++ + + + + +G 6400 + ++++ + + + + ++++ + + + + + -H 3200 + + + + + + + + ++ + ++ + + + + +I 1600 + + + + + + + + + + + + + + + +J 1600 - ++++ + + + + + ± + +K 1600 + + + + + + + + + + + + + + + +L 1600 - + + + + + - + + - - + +M 1600 - + + - - + - + + _ - + +N 800 - + + + + + - - + - + + + + + +0 800 - ++ + - ± + ++ + + + + + +P 800 - + + - - + ++ + + - + + +Q 800 - + - - + + +R 400- + - - - + + - - - - + +S 400- - - - - - - - - - + + +T 400- + --+ - + + +

See Table 2 for LPS, Fla, OMP, SEF14.For key to participants see Table 1.

There was again no problem in interpreting the ODvalues of the high titre sera whereas a mixture ofpositive, weak positive and negative interpretationswere obtained with the intermediate titre group. Mostof the negatives were found in the low titre group. Theinterpretations of four participants (P. B., R. D., F. S.and C. W.) were very similar. Some participants(M. G., K. N., G. S., C. T., F. V. Z.) interpreted at leastone low titre serum with a value as high as some of thehighest titre sera.

DISCUSSION

The results from the first part of this comparativeevaluation showed that, by and large, a goodcorrelation was obtainable between the OD valuesand IgG titre of the sera when tested in differentlaboratories with different assays. A similar degree ofcorrelation has been found in a joint exercise to

compare ELISAs for the serological detection of S.Dublin infection in cattle, also supervised by WHO[17]. The main problems arising from this exercise lay

in the high OD values obtained by some participantswith the intermediate titre sera. Since these had beenobtained from chickens which had been infected withS. Enteritidis and the IgG titres in these sera were somuch higher than those of the low titre sera thisproblem may not be intractable.

Because of this a greater number of intermediateand low titre sera were included in the second part ofthe exercise. Although in a small number of cases thisproblem re-occurred, for low titre sera most assaysproduced OD values considerably lower than thosefor the intermediate titre sera. Despite this, a numberof participants interpreted these results as positive.The extensive use of these assays in the field should bequestioned until such problems have been resolved. Inmost cases these discrepancies of interpretation arethe results of different criteria used in determining cut-off OD values above which a reading is consideredpositive. In practice most of these assays wouldtherefore be expected to produce similar results andthus the major problem is one of the choice of a cut-off value. This is of considerable significance to the


practical use of ELISAs for serological screening ofpoultry flocks for S. Enteritidis and other invasiveserotypes. Absolute unanimity in interpretation oflow titre sera is more important than the same for hightitre sera because when used as a flock test the decisionto regard a flock as serologically positive may rest ona small number of sera which produce OD values justabove the cut-off value. A possible way of overcomingthis problem could be to use, as a criterion, thepresence of individual sera in a flock sample whichcontain very high titre antibodies and which thereforeshould produce high optical densities, well above thecut-off usually obtained from sera from uninfectedbirds. This has been discussed at length elsewhere[18, 19]. This does not negate the value of positivecontrol sera which should anyway be included. Oneinterpretation of the discrepancies in the actual valuesis that the temporal antibody response to antigensother than LPS and flagellin, such as SEF14 andOMPs, may be different. The generation of infor-mation on these points from experimental infectionswould assist in the interpretation of individual assays.One weakness of this comparative exercise is that

sera were produced and titrated by one of theparticipants using one assay. Sera produced byanother laboratory may have produced differentresults through, amongst other things, choice of testor antigen. This does not, however, appear to be amajor problem since a number of quite differentassays compared produced consistent results.Another problem raised by a recent workshop on

ELISAs for salmonella in poultry [19] was one ofspecificity in using LPS-based ELISAs for group Band D Salmonella serotypes. Mild periodate treatmentof group D LPS has been shown to remove the cross-reacting 12 epitope, allowing increased specificity forS. Enteritidis and other serotypes such as S. Gallina-rum and S. Pullorum [20]. For screening for S.Enteritidis only, SEF14 antigen [12] may be used,since this is found only in this and one or two otherrare serotypes such as S. Moscow and S. Blegdam.Research into the use of other antigens such as flagellawhich are already incorporated into several of theELISAs in this study is continuing [11]. Flagella wereused because of the potentially greater antigenicdiversity although cross reactions may occur. Cloningthe serotype-specific fragment away from the con-served N- and C-terminal parts may allow greaterspecificity for a number of serotypes and enablerelatively large amounts of antigen to be produced[21]. The use of combinations of detecting antigens to

enhance specificity has yet to be studied in sufficientdetail.

In a similar comparative exercise the WHO wereable to recommend a standardized procedure for theuse of ELISA for the serological detection ofbrucellosis in cattle and a number of other hostsincluding man [5, 6]. With the present assays forsalmonella already in extensive use it is uncertainwhether standardization can be taken to similarlengths. Rather than reach a point where an exactprotocol is recommended it might be more appro-priate for official testing to be carried out by assayswhich have been approved by standardisation againsta set of sera similar to that used here. This work couldbe organized by an international organization such asthe WHO or European Union.

REFERENCES

1. Rodrigue DC, Tauxe RV, Rowe B. Internationalincrease in Salmonella enteritidis: a new pandemic?Epidemiol Infect 1990; 105: 21-7.

2. Statutory Instruments. The poultry breeding flocks andhatcheries (Registration and Testing) Order no. 1963.London: HMSO, 1989.

3. Council Directive 92/117/EEC of 17 December 1992concerning measures against specified zoonoses andspecified zoonotic agents in animals and products ofanimal origin in order to prevent outbreaks of food-borne infections and intoxications. Of J Europ Com-munities 1993; L62: 38-48.

4. Barrow PA, ELISAs and the serological analysis ofsalmonella infections in poultry: a review. EpidemiolInfect 1992; 109: 361-9.

5. Report. Unpublished report of the World HealthOrganisation informal meeting on enzyme immuno-assay for brucellosis management. Nepean, Ontario,Canada, 29-30 May 1991. VPH/89.84. Geneva: WorldHealth Organisation, 1989.

6. Report. Unpublished report of the World HealthOrganisational informal consultation on enzyme im-munoassay for brucellosis diagnosis and research.Geneva, 5-6 April 1990. VPH/90.90. Geneva: WorldHealth Organisation, 1990.

7. Hassan JO, Barrow PA, Mockett APA, McLeod S.Antibody response to experimental Salmonella typhi-murium infection in chickens measured by ELISA. VetRec 1990; 126: 519-22.

8. Nicholas RAJ, Cullen GA. Development of an ELISAfor detecting antibodies to Salmonella enteritidis inchicken flocks. Vet Rec 1991; 128: 74-6.

9. Barrow PA, Lovell MA. Experimental infection of egg-laying hens with Salmonella enteritidis. Avian Pathol1991; 20: 339-52.

10. Timoney JF, Sikora N, Shivaprasad HL, Opitz M.Detection of antibody to Salmonella enteritidis by a gmflagellin-based ELISA. Vet Rec 1990; 127: 168-9.


11. Baay MFD, Huis in't Veld JHJ. Alternative antigensreduce cross-reactions in an ELISA for the detection ofSalmonella enteritidis in poultry. J Appl Bact 1993; 74:243-7.

12. Thorns CJ, Sojka MG, Chasey D. Detection of a novelfimbrial structure on the surface of Salmonella enteri-tidis by using a monoclonal antibody. J Clin Microbiol1990; 28: 2409-14.

13. Kim CJ, Nagaraja KV, Pomeroy BS. Enzyme-linkedimmunosorbent assay for the detection of Salmonellaenteritis infections in chickens. Amer J Vet Res 1991;52: 1069-74.

14. Van Zijderveld FG, Van Zijderveld-van Bemmel AM,Anakotta J. Comparison of four different enzyme-linked immunosorbent assays for serological diagnosisof Salmonella enteritidis infections in experimentallyinfected chickens. J Clin Microbiol 1992; 30: 2560-6.

15. Smith HW, Tucker JF. The effect of antibiotic therapyon the faecal excretion of Salmonella typhimurium byexperimentally infected chickens. J Hyg 1975; 75:275-92.

16. McLeod S, Barrow PA. Lipopolysaccharide-specificIgG in egg yolk from two chicken flocks infected withSalmonella enteritidis. Lett Appl Microbiol 1991; 13:294-7.

17. Dilling GW, Smith BP. International standardisation ofELISA for serologic detection of Salmonella dublin

carrier cows. Proceedings of the International Sym-posium on Salmonella and Salmonellosis. Ploufragan,France 1992; 87-8.

18. Barrow PA. Use of ELISAs for monitoring Salmonellain poultry. Vet Rec 1994; 134: 99.

19. Barrow PA, Wray C, eds. Proceedings of the EuropeanCommission Workshop on ELISAs for serologicaldiagnosis of Salmonella infection in poultry. Brussels,7-9 June 1993. Directorate General for Agriculture,Commission of the European Communities, Brussels,1994.

20. Van der Heijden HMJF. Effect of periodate oxidationof LPS on the specificity of Salmonella-ELISAs inpoultry. Proceedings of the European CommissionWorkshop on ELISAs for serological diagnosis ofSalmonella infection in poultry, Brussels, 7-9 June1993. Directorate General for Agriculture, Commissionof the European Communities, Brussels, 1994.

21. Huis in't Veld JHJ, van Asten AJAM, Baay MFD,Kusters JG, van der Zeijst BAM. Selection of alterna-tive antigens for the development of a specific ELISAfor the detection of Salmonella enteritidis in poultry.Proceedings of the European Commission Workshopon ELISAs for serological diagnosis of Salmonellainfection in poultry, Brussels, 7-9 June 1993. Direc-torate General for Agriculture, Commission of theEuropean Communities, Brussels, 1994.

world health organisation--supervised interlaboratory comparison

Documents