Comparison of passive haemagglutination and enzyme-linked immunosorbent assay for serodiagnosis of plague*

Bulletin of the World Health Organization, 60 (5): 777-781 (1982) Comparison of passive haemagglutination and enzyme-linked immunosorbent assay for serodiagnosis of plague* JAMES E. WILLIAMS,' LORRAINE ARNTZEN,2 DAVID M. ROBINSON,3 DAN C. CAVANAUGH,4 & MARGARETHA ISAACSON5 Sera of42 suspect plague casesfrom Ovamboland, Namibia, were examined. ELISA proved more sensitive than passive haemagglutination for the detection ofFI antibody and increased positive serodiagnoses ofplague by 7-42%. The enzyme-linked immunosorbent assay (ELISA) has been investigated as a method for detecting plague antibodies in man, carnivores, and rodents (1, 2). Results indicate that ELISA might be an alterna- tive to the standard passive haemagglutination test (PHA) for the diagnosis of human plague and for plague surveillance and control activities. Unlike PHA, ELISA can be used to measure specific im- munoglobulins (IgG, IgM, etc.) in patients' sera or other specimens. In this paper, additional advantages of ELISA over PHA for serodiagnosis of plague are described, and the merits and limitations of these techniques are compared. MATERIALS AND METHODS Human sera from the Ovamboland plague focus in Namibia (3, 4) were examined for specific plague antibodies. Paired sera collected during 1976-78 from 30 patients were tested by PHA, within two weeks of sampling, and then stored at - 20 °C until November-December 1980, when the sera were examined by ELISA. In addition, single sera col- lected in October-November 1980, from 12 cases presumptively diagnosed as plague on the basis of clinical symptoms, were tested. * From the Department of Hazardous Microorganisms, Walter Reed Army Institute of Research, Washington, DC 20012, USA, and the Department of Tropical Pathology and Epidemiology, South African Institute for Medical Research, PO Box 1038, Johannesburg 2000, South Africa. ' Chief, Department of Hazardous Microorganisms. 2 Senior Research Technologist, Department of Tropical Pathology and Epidemiology. 3Chief, Department of Veterinary Medicine, US Naval Medical Research Unit No. 3, Cairo, Egypt. 4 Formerly Chief, Department of Hazardous Microorganisms (Retired). 5 Head, Department of Tropical Pathology and Epidemiology. The PHA test used (3, 5) during 1976-78 and in 1980 was essentially the procedure currently recom- mended by WHO for serodiagnosis of plague (6, 7). A series of twofold serum dilutions, from 1:4, were prepared from 25 1l of serum, and tested. The titre was taken to be the highest dilution exhibiting com- plete haemagglutination. For ELISA, the reagents were prepared as de- scribed by Voller et al (8). Each well of the ELISA substrate platesa was sensitized with 1 yg of the frac- tion 1 (Fl) antigen of Yersinia pseudotuberculosis subsp. pestis (previously named Yersinia pestis) in 100 1d of pH 9.6 coating buffer. Plates were incubated at 37 °C until dry. Excess antigen was removed by washing four times with phosphate- buffered saline containing 0.5 ml of Tween 20 per litre (PBST). Using 25 ul of serum, twofold dilutions (50 il/well), from 1:4, were prepared with PBST directly in the substrate plate using a multichannel pipettor, and incubated for 1 hour at 37 'C. After four washes with PBST, 50 gl of a 1:150 dilution of horseradish peroxidase-conjugated immunoglobulin were added to each well and incubated for 1 hour at 37 'C. All sera were tested with anti-human IgG + IgM + IgA.b Sera collected in 1980 were also tested with anti-human IgG (y-chain specific)c and with anti-human IgM (j4-chain specific).d After incubating with conjugate and another four washes with PBST, 150 u1 of freshly prepared indicator solution (9 parts of 0.8 g/litre 5-aminosalicylic acid, pH 6.0, and 1 part of 0.05 ml/litre H202) were added to each well a M129A flat bottom Microelisa plates, Dynatech Laboratories, Inc., Alexandria, VA, USA. b Code No. 61-231-1 Ly, Miles Laboratories, Inc., Elkhart, IN 46515, USA. ' Code No. 61-130-1, Miles Laboratories, Inc., Elkhart, IN 46515, USA. d Code No. 61-132-1, Miles Laboratories, Inc., Elkhart, IN 46515, USA. 4228 -777 J. E. WILLIAMS ET AL. and the plates were incubated for 1 hour at 37 'C. In recording the results, the titre was taken to be the highest serum dilution producing an obvious colour reaction. Positive sera were retested to confirm the titre. The specificity of ELISA was determined by performing an ELISA-inhibition test (EI) at the same time as a positive serum was retested by ELISA. The El differed from ELISA only in that the PBST used for preparing serum dilutions contained 100 ug of F1 antigen/ml. An ELISA titre was considered to be a measure of specific Fl antibody if the El endpoint was at least eight times lower than the ELISA value. A fourfold or greater increase in titre to 1: 32 or more was accepted as primafacie evidence for a diag- nosis of plague. However, owing to transport diffi- culties in Ovamboland, patients sometimes arrived at the hospital many days after the onset of symptoms and early rises in titre may have been missed. Consequently, a titre of 1: 32 or more concurrent with typical symptoms and no history of plague vacci- nation were considered sufficient evidence for a posi- tive diagnosis. RESULTS ELISA gave results that were very similar to those obtained with PHA during 1976-78 (Table 1). The titres given by the two techniques were similar for most sera, and retests of positive sera demonstrated a normal test error for both PHA and ELISA of ± 1 twofold dilution. However, ELISA results for sera with titres of 1: 4-1:16 were somewhat more variable Table 1. Serological tests of human sera collected in Ovamboland during 1976-78 Change in titre° Serodiagnosis of plague Age Date of Days toand sex first serum second serum PHA ELISA (IgG + IgM + IgA) 9 patients (30%) - negative by PHA 8F 12 Oct. 77 5 both < 1:4 both < 1:4 and ELISA 4M 20 Oct. 77 4 both < 1:4 both < 1:4 4F 20 Oct. 77 5 both < 1:4 both < 1:4 11F 24 Oct. 77 10 both < 1:4 both < 1:4 14M 31 Oct. 77 4 < 1:4 - 1:16 < 1:4- 1:8 7M 14 Nov. 77 3 non-specific both < 1:4 8M 9 Dec. 77 3 both < 1:4 both < 1:4 13F 2 Oct. 78 3 both < 1:4 both < 1:4 Fb 3 Nov. 78 7 < 1:4- 1:16 both < 1:4 19patients(63%)-positivebyPHA 6F 9 Oct. 76 3 1:4 -1:16 1:16 - 1:8192 and ELISA 8F 3Nov. 76 5 1:8 -1:128 <1:4 - 1:32 12F 12 Sep. 77 3 <1:4 -1:256 <1:4 - 1:64 12M 20Sep. 77 34 <1:4 -1:512 <1:4 - 1:512 16Fc 13Oct. 77 11 <1:4 -1:64 <1:4 -.1:256 9F 20Oct. 77 4 1:16 -.1:128 <1:4 -.1:16 10M 20Oct. 77 5 <1:4 -1:128 <1:4 -.1:64 35F 24 Oct. 77 7 1:512 -1:256 both 1:128 37M 24 Oct. 77 7 1:128- 1:64 both 1: 64 11M 25 Oct. 77 9 <1:4 - 1:64 <1:4 -'1:32 13F 4 Nov. 77 6 <1:4 -'1:32 <1:4 -'1:32 16F 14Dec. 77 9 <1:4 -'1:128 <1:4 -'1:128 13F 2 Oct. 78 7 1:8 - 1:128 1:4 - 1:32 16Fc 3Oct. 78 6 1:8 -.1:64 1:4 -.1:32 21F 3Oct. 78 9 <1:4 -.1:64 <1:4 -.1:256 14F 4Oct. 78 5 1:8 -.1:256 <1:4 -.1:128 5F 31 Oct. 78 16 <1:4 - 1:256 <1:4 - 1:512 8F 16 Nov. 78 4 both 1:256 1:256 - 1:2048 44F 16 Nov. 78 4 both 1:128 both 1:512 2 patients (7%)-positive by ELISA; 22M 2 Oct. 78 7 1:8 - < 1:4 1:32 - 1:256 negativebyPHA 14M 5 Oct. 78 14 1:4 - < 1:4 <1:4 - 1:1024 e Antibody to the specific F1 antigen of Y. pseudotuberculosis subsp. pestis. PHA was done within 1- 2 weeks of drawing serum. ELISA was performed on same serum in November-December 1980. b Adult, age unknown. c Plague confirmed by isolation of Y. pseudotuberculosis subsp. pestis. 778 SERODIAGNOSIS OF PLAGUE Table 2. Serological tests of human sera collected in Ovamboland during October-November 1980 ELISA titrea Patient titre a g+g+g g gPatient PHrA ~~~~~~~~IgG+ IgM + IgA IgG IgM 7-year-old girl, in hospital < 1:4 1:256 1:64 1:32 3-year-old boy, in hospital < 1:4 1:32 1:32 1:4 45-year-old woman, in hospital with large femoral bubo < 1:4 1:512 1:256 1:64 8-year-old boy, in hospital with axillary bubo < 1:4 1:256 1:64 < 1:4 14-year-old girl, ambulatory out-patient < 1:4 1:32 < 1:4 1:32 Seven other patients < 1:4 < 1:4 < 1:4 < 1:4 a Antibody to the specific Fl antigen of Y. pseudotuberculosis subsp. pestis. (± 2 dilutions), possibly owing to a light "back- ground" colouration occasionally present at these low serum dilutions. Non-specific ELISA reactions were not observed. Two sera that gave non-specific PHA reactions were negative by ELISA. Serodiag- nosis for plague was identical by PHA and ELISA for 28 (93Gb) of 30 patients (Table 1). In two patients (70/o) who were negative by PHA, ELISA detected significantly higher titres of Fl antibody in the con- valescent sera, suggesting that positive diagnoses of plague were appropriate. In contrast with earlier tests, the results of PHA done in 1980 differed from those of ELISA in 42%o of patients. Five of 12 patients bled during October- November 1980 had Fl antibody titres of 1:32 or more by ELISA but all patients were negative by PHA (Table 2). DISCUSSION Human plague in Ovamboland, Namibia, during the period 1976-80 was confirmed by isolations of Yersinia pseudotuberculosis subsp. pestis from patients. There were 2 bacteriologically confirmed cases in 1976, 3 in 1977, 18 in 1978, 3 in 1979, and 1 in 1980. Additional cases of plague were diagnosed on the basis of serological reactions, including those listed in this paper. The results of PHA done during 1976-78 cor- related better with the ELISA results than did those of PHA done in 1980. Our studies illustrate that differences in PHA reagents can occur even after every effort has been made to achieve standard- ization. Standardization may be easier to achieve with ELISA, especially if photometrically defined endpoints are employed that precisely measure differ- ences between ELISA and corresponding EI tests. ELISA was more sensitive than PHA for detecting plague antibody and increased the frequency of sero- diagnosis of plague by 7-42% in this study. For screening sera, ELISA with a conjugate of pooled immunoglobulins (IgG+IgM+IgA) was used and can be recommended. Subsequent tests for IgG and IgM antibodies were informative. An ambulatory patient, who exhibited a clinical course characteristic of the "mild" plague frequently seen in Ovamboland (3, 4), had an IgM titre of 1: 32 but no IgG antibody. Patients with IgM antibody to Fl but without detect- able IgG antibody have been detected in other studies employing radioimmunoassay (9). Additional case reports with appropriate clinical and laboratory data will be required to determine what interpretations are valid for this and other serological responses. Immunoglobulin patterns may indicate whether there are anamnestic responses in patients with pyrexial illnesses other than plague, or whether certain serological responses are representative of infections with atypical Y. pseudotuberculosis subsp. pestis (10). ELISA can be recommended because of its sensi- tivity and versatility for measuring various types of immunoglobulins. However, the procedure for ELISA is somewhat more complicated than PHA and requires about 4 hours to complete when previously sensitized plates are used. Results are obtained in about 2 hours with PHA. The specificity of both tests can be ensured using inhibition controls; a reaction is non-specific if endpoints are not lowered when puri- fied Fl antigen is added to the serum diluent. 779 780 J. E. WILLIAMS ET AL. ACKNOWLEDGEMENTS The authors thank Dr 0. van Niekerk, Secretary for Health and Welfare, Government Service of Owambo, Dr S. S. Grove, Regional Director, South African Institute for Medical Research (SAIMR), Namibia, and Professor J. Metz, Director, South African Institute for Medical Research, Johannesburg, for their interest in and support of these studies. The collection and transport of clinical specimens by SAIMR personnel was accomplished with the kind assistance of the South African Medical Service under direction of Surgeon General N. J. Nieuwoudt. RtSUME COMPARAISON ENTRE L'HEMAGGLUTINATION PASSIVE ET LE TITRAGE AVEC IMMUNOADSORBANT LIE A UNE ENZYME POUR LE SERODIAGNOSTIC DE LA PESTE Une etude a e effectuee pour comparer les avantages et les inconvenients du titrage avec immunoadsorbant lie a une enzyme (ELISA) et de l'epreuve d'hemagglutination passive classique (HAP) pour le serodiagnostic de la peste humaine. Des s6rums humains provenant du foyer pesteux d'Ovambo- land en Namibie ont et examines et on y a recherche les anti- corps a l'6gard de l'antigene de la fraction 1 (Fl) de Yersinia pestis. Des paires de serums recueillies en 1976-1978 a partir de 30 malades ont e eprouv6es par HAP dans les semaines qui ont suivi le prelevement, puis conservees dans des conge- lateurs jusqu'en novembre-d6cembre 1980 et examinees alors a l'aide de la technique ELISA. Douze autres serums non apparies, recueillis en octobre-novembre 1980 de cas pour lesquels un diagnostic presomptif de peste avait e pose sur la base des sympt8mes cliniques, ont e eprouves par les deux methodes serologiques. Une augmentation de quatre fois ou plus du titre jusqu'a 1: 32 ou plus, ou bien un titre de 1: 32 ou plus accompagne d'une symptomatologie typique et sans antecedents de vaccination antipesteuse ont e consi- deres comme des criteres suffisants pour un diagnostic positif de peste. La technique HAP a ete effectu&e selon la methode recom- mand&e par l'OMS et on a retenu comme titre la dilution de serum la plus pouss6e donnant une hemagglutination com- plete. Dans le cas de la technique ELISA (qui est decrite), le titre a e la dilution de serum la plus pouss6e produisant une reaction coloree nette. La specificite de la technique ELISA a e determinee par l'execution simultan&e d'epreuves d'inhi- bition d'ELISA (IE), qui ne differaient d'ELISA que par le fait que le diluant utilise pour preparer les dilutions de serum contenait 100 ,g de Fl/ml. Le titre ELISA etait consid&e comme une mesure de l'anticorps specifique de F1, si IE avait un point final au moins huit fois inferieur a celui d'ELISA. Les resultats obtenus par la technique HAP en 1976-78 etaient les memes que ceux qui ont e obtenus avec ELISA. Les serodiagnostics de peste etaient identiques pour 28 des 30 malades, soit 93%. Deux malades (70o) etaient negatifs en HAP, alors que ELISA a revele des titres nettement plus eleves d'anticorps anti-Fl dans les serums de convalescents, suggerant qu'un diagnostic positif de peste etait justifie. Les titres obtenus par les deux techniques etaient similaires et pour l'une et I'autre l'erreur etait normalement de ± 1 dilution d'ordre 2; cependant, les resultats fournis par ELISA pour les serums dont les titres etaient compris entre 1: 4 et 1:16 etaient un peu plus variables (± 2 dilutions). La technique HAP appliquee en 1980 a fourni des resultats differents de ceux d'ELISA pour 42% des malades. Sur les 12 malades chez qui l'on a preleve du sang en octobre-novembre 1980, 5 avaient des titres d'anticorps anti-Fl de 1:32 ou plus en ELISA, alors que tous ces malades etaient negatifs en HAP. Nos etudes montrent qu'il se produit des differences dans les reactifs HAP, meme lorsqu'on s'est efforce le plus possible de les standardiser. La standardisation peut etre plus facile a obtenir avec la technique ELISA, surtout si l'on utilise des points finals definis par photometrie, qui permet- tent de mesurer precisement les differences entre les epreuves ELISA et les 6preuves IE correspondantes. La technique ELISA peut egalement etre recommandee en raison de sa sensibilite et de sa capacite de mesurer differents types d'immunoglobulines (IgG, IgM, etc.). Le mode operatoire d'ELISA est cependant un peu plus complique que celui de HAP et exige environ 4 heures lorsqu'on utilise des plaques presensibilisees, alors qu'il est possible d'obtenir des resultats en 2 heures environ avec HAP. REFERENCES 1. CAVANAUGH, D. C. ET AL. Application of the ELISA technique to problems in the serologic diagnosis of plague. Bulletin of the Pan American Health Organiz- ation, 13: 399-402 (1979). 2. WILLEBERG, P. W. ET AL. Epidemiologic survey of syl- vatic plague by serotesting coyote sentinels with enzyme immunoassay. American journal ofepidemiology, 110: 328-334 (1979). SERODIAGNOSIS OF PLAGUE 781 3. ISAACSON, M. Human plague in Southern Africa during the period 1965-1969. MD thesis. University of the Witwatersrand, Johannesburg, 1972, 159 pp. 4. IsAAcsoN, M. & JAMISON, J. R. Epidemiological and clinical aspects of plague in Southern Africa. In: Gear, J. H. S., ed., Proceedings of the International Sym- posium on Medicine in a Tropical Environment, South Africa, 1976, Rotterdam, A. A. Balkema, 1977, pp. 479-486. 5. ISAACSON, M. & HALLETT, A. F. Serological studies on human plague in Southern Africa. Part I. Plague anti- body levels in a population during a quiescent and a sub- sequent active period in an endemic region. South African medicaljournal, 49: 1165-1168 (1975). 6. WHO Technical Report Series, No. 447, 1970 (WHO Expert Committee on Plague: fourth report). 7. BAHMANYAR, M. & CAVANAUGH, D. C. Plague manual. Geneva, World Health Organization, 1976, 76 pp. 8. VOLLER, A. ET AL. Enzyme immunoassays in diagnostic medicine. Theory and practice. Bulletin of the World Health Organization, 53: 55-65 (1976). 9. HUDSON, B. W. ET AL. The use of solid-phase radio- immunoassay techniques for serodiagnosis of human plague infection. Bulletin of the Pan American Health Organization, 14: 244-250 (1980). 10. ISAACSON, M. ET AL. Unusual cases of human plague in Southern Africa. South African medical journal, 47: 2109-2113 (1973).