Multicentre evaluation of an antigen-detection ELISA for the diagnosis of Trypanosoma brucei rhodesiense sleeping sickness.

Multicentre evaluation of an antigen-detection ELISA for the diagnosis of Trypanosoma brucei rhodesiense sleeping sickness E. Komba,' M. Odiit,2 D.B. Mbulamberi,3 E.C. Chimfwembe,4 & V.M. Nantulya5 The performance of an enzyme-linked immunosorbent assay (antigen ELISA) for the detection, in serum or cerebrospinal fluid, of an invariant trypanosome antigen to diagnose Trypanosoma brucei rhodesien- se sleeping sickness was evaluated in four clinical treatment centres. The test, which was carried out in polystyrene test-tubes, was positive in 88 (88.9%) of 99 parasitologically confirmed cases that were tested at the National Institute for Medical Research, Tabora, United Republic of Tanzania; 99 (94.3°%) of 105 cases tested at the National Sleeping Sickness Control Programme, Jinja, Uganda; 86 (87.8%) of 98 cases tested at the Uganda Trypanosomiasis Research Organisation, Tororo, Uganda; and 59 (96.7%) of 61 cases tested at the Tropical Diseases Research Centre, Ndola, Zambia. The over- all detection rate was 91.5%. There was no cross-reactivity with the agents of the common bacterial, viral, or parasitic diseases prevalent in the areas where the studies were conducted. The only false- positive result involved a blood donor from a trypanosomiasis endemic focus. The test was simple to perform, was read visually, and is therefore a potential tool for diagnosing human African trypanosomiasis. Introduction Human African trypanosomiasis is one of the major diseases that afflicts the continent. There are two forms of the disease: the chronic variety encountered in West and Central Africa, which is caused by Trypanosoma brucei gambiense, and the more acute form, which occurs in East Africa and is caused by T. b. rhodesiense. Conservative estimates put the number of new cases diagnosed at 25 000 per annum, with 50 million people at risk in the affected countries in tropical Africa. Detection of trypanosomes in the blood of an infected individual would be the ideal method for the laboratory diagnosis of African trypanosomiasis. This, however, is not easy, mainly because some of the techniques used have very low sensitivity, while I Scientist, National Institute for Medical Research, Tabora, United Republic of Tanzania. 2 Medical Officer, Uganda Trypanosomiasis Research Organisation, Tororo, Uganda. 3 Manager, National Sleeping Sickness Control Programme, Jinja, Uganda. 4 Research Officer, Tropical Diseases Research Centre, Ndola, Zambia. 5 Senior Scientist, International Laboratory for Research on Animal Diseases, P.O. Box 30709 Nairobi, Kenya. Requests for reprints should be sent to this address. Reprint No. 5247 those that are sensitive are not readily applicable in the field (1,2). Several serological techniques have therefore been developed and evaluated extensively in the field as altematives to the detection of trypa- nosomes (3,4). Serological tests, however, enable only a presumptive diagnosis to be made, since they cannot determine whether or not a patient has an active infection. Consequently, there is still a need to develop new techniques, particularly those that can distinguish between current and cured infections.a,b Recently an approach has been described that involves the detection of trypanosome-invariant anti- gens in tissue fluids of infected individuals, using an enzyme-linked immunosorbent assay (antigen ELISA) as a means of diagnosis (1). The detection of trypanosome antigens in host tissue fluids would provide evidence of an active infection. The assay, which is based on a monoclonal antibody raised against an invariant antigen of T. b. rhodesiense pro- cyclic trypomastigotes (5), was evaluated in the a Report of the Second Scientific Working Group on African Trypanosomiasis, Arusha, Tanzania, 26-30 October 1981. WHO unpublished document TDR/TRY-SWG (2)/81.30. b Report of the Meeting of the Joint Steering Committees of the Scientific Working Group on African Trypanosomiasis, Nairobi, Kenya, 13-16 March 1988. WHO unpublished document TDR/TRY/SC-SWG (1)/88.3. Bulletin of the World Health Organization, 70 (1): 57-61 (1992) © World Health Organization 1992 57 E. Komba et al. laboratory using sera from patients with T. b. rhode- siense infection and was highly sensitive and specific (1). It has also shown promise as a method for the diagnosis of T. b. gambiense sleeping sickness (6). The studies described here were conducted from July 1989 to July 1990 at four treatment centres (one in the United Republic of Tanzania, two in Uganda, and one in Zambia) to evaluate this antigen ELISA for the field diagnosis of T. b. rhodesiense sleeping sickness. Materials and methods Patient selection Patients who presented to the study treatment centres with histories indicative of sleeping sickness were investigated to confirm the diagnosis, using the fol- lowing techniques: thick blood film examination (7); haematocrit centrifugation (8); miniature anion- exchange chromatography (9); single and double centrifugation for trypanosomes in cerebrospinal fluid (CSF) (10); and CSF cell count and protein determination (11). Serum and CSF samples from parasitologically confirmed cases were stored frozen at -20 °C and tested later for trypanosome antigens. To determine the specificity of the antigen ELISA, random serum samples from patients with common viral, bacterial, and parasitic infections, as well as from blood donors, were screened using it. Monoclonal antibody Polystyrene tubes were coated with the monoclonal antibody TR7/47.34.16, as described by Nantulya et al. (12). 500 11l) of a 1:500 dilution of peroxidase-labelled monoclonal antibody TR7/47.34.16 were added to each tube without first washing the tubes. The tubes were re-incubated for 15 minutes, rinsed three times, refilled with the same buffer, and left to stand for 15 minutes. The washing buffer was then decanted and the tubes were rinsed three times before 10 drops per tube (250jl) of substrate (hydrogen peroxide) and chromogen (2,2'-azino-bis(3-ethylbenzthiazoline-6- sulfonic acid) (ABTS) were added to each tube. The tubes were incubated at room temperature with inter- mittent mixing and the colour changes observed after 30 minutes. For each assay, negative and positive serum controls were included. Negative reactions pro- duced no colour change, while positive reactions resulted in a clear green solution, which made it easy to distinguish visually between positive and negative results. The assay for antigens in CSF was performed as described for serum except that the CSF (10-20 drops) was tested without prior dilution. Results Specificity of the antigen ELISA To assess the specificity of the antigen ELISA, we analysed sera obtained from patients with various other infections that were endemic in the study areas. Table 1 shows the pooled results from the four study centres. No cross-reactivity was observed with 36 samples of sera from malaria patients, 33 samples from cases of tuberculosis, or 17 samples from cases of human immunodeficiency virus (HIV) infection. One sample of serum from a blood donor from an endemic area was weakly positive for antigen; it was, however, not possible to make exhaustive in- Antigen detection To detect circulating trypanosome antigens, we per- formed a simplified sandwich ELISA using mono- clonal-antibody-coated polystyrene tubes (12). Briefly, l-ml polystyrene tubes were filled at 4 °C with a solution containing 2.5,g/ml of purified IgM monoclonal antibody TR7/47.34. 16; the tubes were capped and then sent from the Intemational Laboratory for Research on Animal Diseases (ILRAD) to the collaborating centres, where they were stored at 4 'C. Immediately before carrying out the assay, the tube contents were tipped out and 25 drops (ca.700 gil) of washing buffer added to each tube using a Pasteur pipette, followed by one drop of test serum. The tubes were capped and incubated at room temperature for 15 minutes, with intermittent mixing. The contents were then decanted and 15 drops (ca. Table 1: Results of the antigen-ELISA on samples of sera obtained from various categories of patients to assess the specificity of the assaya Category of patient Malaria Tuberculosis HIV infection b Blood donors Syphilis Schistosomiasis Filariasis Total No. tested No. positive 0 0 0 36 33 17 20 3 111 0 0 0 1 (0.9%) a Pooled results for the four collaborating centres. b HIV = human immunodeficiency virus. WHO Bulletin OMS. Vol 70 199258 Antigen-detection ELISA for diagnosing human African Table 2: Pattern for the detection of trypanosome anti- gens in samples of serum and cerebrospinal fluid (CSF) of Trypanosoma brucei rhodesiense patients No. of cases antigen-ELISA- positive in: Study Both serum centrea Serum CSF and CSF Total UTRO 34 7 40 81 b NSSCP 55 10 34 99 Total 89 (49.4)c 17 (9.4) 74 (41.1) 180 (99.9) a UTRO = Uganda Trypanosomiasis Research Organisation, To- roro; NSSCP = National Sleeping Sickness Control Programme, Jinja, Uganda. b Five other patients in the series were excluded from the ana- lyses because they did not have paired CSF and serum results. c Figures in parentheses are percentages. vestigations on this patient to rule out a trypanosome infection. The assay thus demonstrated a high degree of specificity. Sensitivity of the antigen ELISA For some patients, trypanosome antigens were detec- ted in serum only; for others, they were detected in CSF only; and for yet others the antigens were detec- ted in both serum and CSF. The pattem for two centres where paired CSF and serum analyses were performed is shown in Table 2. Nearly half the patients (49.4%) tested positive for antigen in serum only; 9.4% in CSF only; and the rest in both serum and CSF. For calculation of the detection rates, we therefore considered patients to have tested positive if their serum and/or CSF was ELISA-positive (Table 3). A total of 363 parasitologically proven cases were tested using the antigen ELISA in the four study centres. The results are summarized in Table 3. Of these patients, 332 gave positive results, corresponding to an overall sensitivity of 91.5%. At the National Institute for Medical Research (NIMR), United Republic of Tanzania, 99 cases were tested, of which 88 (88.9%) were antigen-positive in the ELISA. At the National Sleeping Sickness Control Programme (NSSCP), Uganda, 99 (94.3%) of 105 cases tested gave positive results, while at the Uganda Trypanosomiasis Research Organisation (UTRO) 86 (87.8%) of the 98 cases tested were posi- tive. At the Tropical Diseases Research Centre (TDRC) Zambia, 59 (96.7%) of 61 patients tested were positive. A total of 30 patients (8.3%) in the whole series tested negative for antigens (Table 3). Analysis of CSF results A comparison between the results of the antigen ELISA for CSF and other changes in CSF, such as the presence of trypanosomes, increased protein levels, and elevated white cell counts, is given in Table 4 and Table 5. Trypanosomes were detected in the CSF of 66.7% of the patients whose CSF gave a positive antigen ELISA result; the protein levels were elevated in 48.7%; and the white cell counts were elevated in 87.2% (Table 4). Conversely, 70.4% of patients whose CSF contained trypanosomes (Table 5) had detectable levels of trypanosome anti- gens in their CSF; 65.5% had elevated protein levels; and 96.3% had elevated white cell counts. Table 3: Ability of the antigen ELISA to detect trypanosome antigens in samples of sera from patients with parasitologically proven Trypanosomia brucei rhodesiense sleeping sickness No. of cases No. of cases Total number of cases No. of cases with positive with positive CSF with positive CSF Study centrea tested serum but negative serum and/or serum NIMR 99 88 (88.9)b N.A.C 88 (88.9) NSSCP 105 89 (84.8) 10 (9.5) 99 (94.3) UTRO 98 79 (80.6) 7 (7.1) 86 (87.8) TDRC 61 59 (96.7) N.A.c 59 (96.7) Total 363 314 (86.5) 18 (5.0) 332 (91.5) a NIMR = National Institute for Medical Research, Tabora, United Republic of Tanzania; NSSCP = National Sleeping Sickness Control Programme, Jinja, Uganda; UTRO = Uganda Trypanosomiasis Research Organisation, Tororo; TDRC = Tropical Diseases Research Centre, Ndola, Zambia. b Figures in parentheses are percentages. c N.A. = not available. WHO Bulletin OMS. Vol 70 1992 59 E. Komba et al. Table 4: White cell counts, protein levels, and presence of trypanosomes in cerebrospinal fluid (CSF) of patients with Trypanosoma bruce! rhodeslense sleep- ing sickness whose CSF was positive In the antigen ELISAa CSF parameter No. tested No. positive Presence of trypanosomes 57 38 (66.7)b Elevated cell count 86 75 (87.2) Elevated protein level 39 19 (48.7) a Pooled results for the Uganda Trypanosomiasis Research Organisation (UTRO) and the National Sleeping Sickness Control Programme (NSSCP). b Figures in parentheses are percentages. Table 5: Antigen ELISA results, white cell counts, and protein levels in the cerebrospinal fluid (CSF) of patients with Trypanosoma bruce! rhodesiense whose CSF contained trypanosomes" No. tested No. positive Antigen ELISA 54 38 (70.4)b Elevated cell count 54 52 (96.3) Elevated protein 29 19 (65.5) level a Pooled results from the Uganda Trypanosomiasis Research Organisation (UTRO) and National Sleeping Sickness Control Programme (NSSCP). b Figures in parentheses are percentages. Discussion In the study we have described, an antigen-trapping ELISA (13) was evaluated in four treatment centres for the diagnosis of T. b. rhodesiense infection. The detection rate varied from 87.8% to 96.7%, and the overall detection rate was 91.5%, which is in agree- ment with that previously reported for an assessment of the same monoclonal antibody in a microplate ELISA (1). This level of sensitivity also compared favourably with that obtained using anti-trypanoso- me antibody detection techniques, such as the indi- rect immunofluorescent antibody test (13), ELISA (14), card agglutination (4), and procyclic agglutina- tion (15) tests. Whereas antibody assays can provide only a presumptive diagnosis, demonstration of spe- cific trypanosome antigens in the tissue fluids of a patient is evidence of a current infection. Antigen positivity is thus synonymous with parasitological diagnosis, and hence provides a more direct basis for initiation of chemotherapy. Cross-reactivity was not observed with other diseases that are endemic in the study areas, such as malaria, schistosomiasis, tuberculosis, syphilis, fila- riasis, and acquired immunodeficiency syndrome (AIDS). Only one blood donor from an endemic area gave a false positive reaction. The assay thus appears to be specific and this represents an improvement over the other immunodiagnostic tests, which on average give false positivity rates of 5% (14,16). Despite its high sensitivity, the test nevertheless failed to detect a total of 30 (8.3%) out of 363 para- sitologically diagnosed cases from the four collabo- rating centres. These false negative cases probably represent a group at an early stage (less than 2 weeks) of infection, whose circulatory antigens were below the assay's limit of detection (1,7,17). However, these antigen-ELISA-negative patients were readily detected by parasitological techniques, which indicates that antigen detection and parasito- logical diagnosis are complementary. Because in experimental infections in animals, antigen levels fluctuate throughout the course of infection (17,18), the assay should be repeated at least once on patients with a highly indicative clini- cal history but who do not have demonstrable parasi- taemia. With some patients, antigens were detected in CSF but not in serum. Apart from the diagnostic implication of this, i.e., that a proper diagnostic investigation of a patient should include a search for trypanosome antigens in both CSF and serum, this finding emphasizes the independent nature of the blood and CSF compartments, as has also been observed for the distribution of anti-trypanosome antibodies in the infected host (19). There was a strong correlation between the pres- ence of antigens in CSF and other pathological changes associated with involvement of the central nervous system, i.e., an increase in protein levels and white cell counts, and the presence of trypanosomes in the CSF compartment. A similar correlation was observed in T. b. gambiense infections, where the antigen ELISA has major diagnostic potential (7). These results suggest that, apart from its use as a dia- gnostic tool, the assay can be interpreted alongside existing criteria for clinically staging the disease. Another potential application of the assay is for eva- luating the success of treatment (7, 17), and this is currently being investigated. Acknowledgements This investigation received financial support from the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR) (Grant No. ID 890041). We wish to thank all the individuals who assisted in collecting the samples and managing patients, Mr J.T. Njuguna for technical assistance, and Mrs Risper Okonji for secretarial assistance. WHO Bulletin OMS. Vol 70 199260 Antigen-detection ELISA for diagnosing human African trypanosomiasis Resume Evaluation multicentrique d'un ELISA pour la recherche de l'antigene servant au diagnostic de la trypanosomiase a Trypanosoma brucei rhodesiense On a evalue dans quatre centres de traitement clinique les r6sultats d'une r6action immunoenzy- matique (ELISA-antigene) visant a mettre en 6vi- dence dans le serum ou le liquide c6phalorachi- dien un antigene trypanosomique invariant, qui permettrait de diagnostiquer une trypanosomiase a Trypanosoma brucei rhodesiense. Ce test, qui a 6t6 effectue dans des tubes a essai en polysty- rene, a ete positif pour 88 (88,9%) des 99 cas parasitologiquement confirm6s test6s au National Institute for Medical Research de Tabora, R6publique-Unie de Tanzanie, pour 99 (94,3%) des 105 cas testes dans le cadre du National Sleeping Sickness Control Program de Jinja, Ouganda, pour 86 (87,8%) des 98 cas test6s par l'Uganda Trypanosomiasis Research Organisation de Tororo, Ouganda, et pour 59 (96,7) des 61 cas test6s au Tropical Diseases Research Centre de Ndola, Zambie. Le taux global de mise en 6vi- dence a et6 de 91,5%. II n'y a pas eu de r6activi- t6 crois6e avec les germes des maladies bact6- riennes, virales ou parasitaires r6pandues dans ces r6gions. Le seul faux positif obtenu 6tait dO a un donneur de sang provenant d'un foyer d'end6- mie de la trypanosomiase. Ce test est d'ex6cu- tion simple, la lecture du resultat est visuelle et il constitue par cons6quent un outil potentiellement utile pour le diagnostic de la trypanosomiase afri- caine chez I'homme. References 1. Nantulya, V.M. An antigen detection enzyme immu- noassay for the diagnosis of T. rhodesiense sleep- ing sickness. Parasite immunology, 11: 69-75 (1989). 2. Nantulya, V.M. Trypanosomiasis in domestic ani- mals: the problems of diagnosis. Revues scienti- fiques et techniques des Office internationale des Epizootiques, 9: 357-367 (1990). 3. Aiyedun, B.A. et al. Parallel evaluation of serologi- cal tests applied in African trypanosomiasis: a WHO collaborative study. Bulletin of the World Health Organization, 54: 141-280 (1976). 4. Magnus, E. et al. A card agglutination test with stained trypanosomes (CATT) for serodiagnosis. Annales de la Societe belge de M6dicine tropicale, 58: 169-176 (1978). 5. Nantulya, V.M. et al. Monoclonal antibodies that distinguish Trypanosoma congolense, T. vivax and T. brucei. Parasite immunology, 9: 421-431 (1987). 6. Nantulya, V.M. et al. Antigen detection ELISA for diagnosis of Trypanosoma brucei gambiense sleeping sickness. Transactions of the Royal Society of Tropical Medicine and Hygiene, (in press). 7. Baker, J.R. Techniques for the detection of trypano- some infections. In: Mulligan, H.W., ed. The African trypanosomiases. London, Allen Unwin, 1970, pp. 67-88. 8. Woo, P.T.K. Evaluation of the haematocrit centri- fuge and other techniques for the field diagnosis of human trypanosomiasis and filariasis. Acta tropica, 28: 298-303 (1971). 9. Lumsden, W.H.R. et al. Trypanosoma brucei: miniature anion-exchange centrifugation technique for the detection of low parasitaemias: adaptation to field use. Transactions of the Royal Society of Tropical Medicine and Hygiene, 73: 312-317 (1979). 10. Cattand, P. et al. Human African trypanosomiasis: use of double centrifugation of cerebrospinal fluid to detect trypanosomes. Bulletin of the World Health Organization, 66: 83-86 (1988). 11. Trypanosomiasis control manual. Geneva, World Health Organization, 1983. 12. Nantulya, V.M. et al. Two simple antigen-detection enzyme immunoassays for the diagnosis of Trypanosoma evansi infections in the dromedary camel (Camelus dromedarius). Tropical medicine and parasitology, 40: 415-418 (1989). 13. Wery, M. et al. The diagnosis of human African try- panosomiasis (T. gambiense) by the use of fluores- cent antibody test. 1. Standardization of an easy technique to be used in mass surveys. Annales de la Societ6 belge de M6dicine tropicale, 50: 613-634 (1970). 14. Voller, A. Serology of African trypanosomiasis. Annales de la Soci6t6 belge de M6dicine tropicale, 57: 273-279 (1977). 15. Liu, M.K. et al. Immunodiagnosis of sleeping sick- ness due to Trypanosoma brucei gambiense by detection of anti-procyclic antibodies and trypano- some antigens in patients' sera. Acta tropica, 46: 257-266 (1989). 16. van Meirvenne, N. & le Ray, D. Diagnosis of African and American trypanosomiasis. British medi- cal bulletin, 41: 156-161 (1985). 17. Nantulya, V.M. & Lindqvist, K.J. Antigen-detection enzyme immunoassays for the diagnosis of Trypanosoma vivax, T. congolense and T. brucei infections in cattle. Tropical medicine and parasitol- ogy, 40: 267-272 (1989). 18. Masake, A.R. & Nantulya, V.M. The sensitivity of an antigen-detection enzyme immunoassay for the diagnosis of Trypanosoma congolense infections in goats and cattle. Journal of parasitology, 77: 231-236 (1991). 19. Masake, R. et al. Cerebral trypanosomiasis in cattle with mixed Trypanosoma congolense and T. brucei infections. Acta tropica, 41: 237-246 (1984). WHO Bulletin OMS. Vol 70 1992 61