- Email: [email protected]
Comparative evaluation of three new tools for diagnosis of bancroftian filariasis based on detection of specific circulating antigens
TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1999)93,278-282
Comparative evaluation of three new tools for diagnosis based on detection of specific circulating antigens
of bancroftian
filariasis
Paul E. Simonsenl and Samuel K. Dunyo* ‘Danish Bilharziask Laboratory, Jaegersborg Alle 1 D, 2920 Charlottenlund, Denmark; ‘Noguchi Memorial Institute for Medical Research, University of Ghana, I? 0. Box 25, Legon, Ghana Abstract Three new and commercially available tools for diagnosis of Wuchereria bancrofii infections based on detection of specific circulating antigens were evaluated and compared in the same group of individuals from a highly endemic village in southern Ghana. The tests were: (1) the ICT card test for serum specimens; (2) the TropBio ELISA test for serum specimens; and (3) the TropBio ELISA test for filter-paper specimens. A high degree of positive/negative response similarity was observed for the 3 tests, and the sensitivity for detecting microfilaraemic cases was 100% for all tests. The antigen levels measured in the TropBio serum test and the TropBio filter-paper test were statistically significantly correlated. Among antigen-positive endemic individuals the antigen levels in these 2 tests furthermore showed a positive association with the microfilarial intensity, but a statistical significant correlation was seen only for the filter-paper version of the test. The results are promising for the use of the 3 tests as diagnostic tools in bancroftian filariasis. Keywords:
Wuchereria bancrofi, filariasis, diagnosis, circulating antigens, Ghana
Introduction Diagnosis of bancroftian filariasis, resulting from infection with the mosquito-borne filarial nematode Wuchereria bancrofti, relied until recently almost exclusively on detection and identification of microfilariae (mf) in blood specimens (MCMAHON & SIMONSEN, 1996; NICOLAS, 1997). In most geographical areas, W. bancrofti mf have a nocturnal periodicity, with highest intensity in the peripheral blood at night and few or none during the davd ~SIMONSEN et al.. 1997a). Blood sneci: mens for parasrtological diagnosis m&t therefo;e be collected at night. Many attempts have been made to develop simple and reliable immunological tests, which can also be used during daytime. While tests for antibody detection have generally suffered from lack of specificity and inability to distinguish between current and past infection, other tests based on detection of specific circulating W! bancrofiifilarial antigens (CFA) have been highly successful, and a number of kits have recently become commercially available. The first of these kits to appear on the market was made by Tropical Biotechnology Pty Ltd (TropBio) in Australia. It utilizes a monoclonal antibody, Og4C3 (MORE & COPEMAN, 1990), in a sandwich enzymelinked immunosorbent assay (ELISA) to detect CFA in serum specimens. A high specificity and sensitivity of the test has been confirmed in several studies (e.g., MORE & COPEMAN, 1990; CHANTEAU et al., 199aa, 1994b; ROCHA et al.. 1996: NICOLAS. 1997; NICOLAS et al.. 1997; WEIL et al., (997). Another version of this tes< adapted for analysis of finger-prick blood specimens collected on filter-paper disks, has been developed by the same company. More recently, ICT Diagnostics in Australia produced a simple immunochromatographic card test for detection of CFA in serum specimens (WEIL et al., 1997). In this test, antigen capture is accomplished by another monoclonal antibody, AD. 12 @VEIL & L~TIS, 1987), with high sensitivity and specificity to W. bancroftiantieen N[IEIL et al.. 1987). Serum and wash buffer is ippliedyo specifically indicated pads on the test card, and within a few minutes a coloured reaction develops which indicates whether the serum specimen is positive or negative for the antigen. This test requires no advanced equipment, which makes it suitable for many endemic areas in the developing world. The present study evaluated and compared 3 tests for diagnosis of bancroftian filariasis based on detection of Address for correspondence: Dr Paul E. Simonsen, Danish Bilhaniasis Laboratory, Jaegersborg Alle 1 D, 2920 Charlottenlund, Denmark; phone +45 77327732, fax +45 77327733, e-mail [email protected]
CFA, namely (1) the ICT card test for serum specimens, (2) the TropBio ELISA test for serum specimens, and (3) the TropBio ELISA test for filter-paper specimens. The same group of individuals from an endemic village in southern Ghana was examined with all 3 tests, and results were compared to the mf status and mf intensity for the same individuals. Materials and Methods Study individuals and study design Following a baseline survey for bancroftian filariasis in Gomoa Okyereko village in the Central Region of Ghana (DZODZOMENYO et al., 1999), a number of villagers aged 12 years or more were requested to participate in the study. Fifty-one individuals responded positively and these constitute the endemic study population. After registration, blood specimens were sampled between 21:00 and 24:O0. Finger-prick blood was collected in capillary tubes for quantification of W. bancrofi mf and on filter-paper disks for determination of specific W. bancrofti CFA. Immediately thereafter, 5 mL of venous blood was collected from the same individual (using the Vacutainer system) for serum preparation. The exact time of sampling was noted, and all specimens from the same individual were collected within a period of 5 min. Ten individuals from Denmark, where no transmission of W. bancrofti occurs, served as non-endemic controls. The procedure for sampling of blood specimens from the control individuals was similar to that described for the endemic individuals, with the exception that sampling took place at daytime and that no specimens were collected and examined for mf. Examination of blood for mf et al., The counting-chamber technique (MCMAHON 1979) was used to quantify the blood microfilaraemia. In brief, 100 ~.LL of finger-prick blood was collected in heparinized capillary tubes, and immediately transferred to specimen tubes containing 900 ~.ILof 3% acetic acid. The number of mf in the specimens was later counted in a counting chamber, using a microscope. To compensate for the effect of sampling time on mf intensities, timeadjusted mf intensities were calculated by multiplying the observed intensities with a time-specific adjustment factor (see SIMONSEN et al., 1997a). Only time-adjusted mf intensities were used in the data analysis. Serum preparation Venous blood was left in a fridge at 5°C overnight for clotting, Thereafter, the clot was removed and the remaining specimen was centrifuged. The supernatant of serum was frozen at -20°C until use.
CIRCULATING
ANTIGENS FOR DIAGNOSIS OF BANCROFTIAN
The ICT card test The serum version of the ICT card test (ICT Diagnostics, Balgowlah, Australia; catalogue no. ICT FLOl) was used, according to the manufacturer’s instructions. In brief, 50 uL of serum was applied to the pink specimen pad, and 2 drops of wash reagent were applied at the indicated white pad. The card was closed, and the result read after 5-10 min. The card was re-examined 1 and 24 h after application of serum, but the results were similar to those obtained at the first reading. The TropBio serum test The TropBio ELISA kit for detecting and quantifying W bancrofti antigen in serum specimens (JCIJ Tropical Biotechnology Pty Ltd, Townsville, Australia; catalogue no. 03-010-01) was used according to the manufacturer’s instructions. After boiling pre-treatment, each serum specimen was tested in duplicate. The 7 standards with known antigen content enclosed in the kit were similarly tested in duplicate, and the results were used to prepare a standard curve relating optical density (OD) and antigen content. The mean OD of each serum specimen was used to determine the response in antigen units (from the standard curve) as well as the antigen titre group (in relation to the 7 standards) to which the specimen belonged. Serum specimens with 2 32 antigen units (2 titre group 3) were considered positive for CFA, and specimens with a response 3 the OD of standard 7 were assigned a fixed value of 32 000 antigen units (titre group 8). The TropBio filter-paper test Finger-prick blood was collected on TropBio filterpaper disks (catalogue no. 05-001-07). All 6 protrusions on 1 disk were saturated with blood from 1 individual, whereafter the disks were dried in air at room temperature overnight. The dried disks were packed individually in small punched plastic bags. These bags were placed, together with a silica-gel sachet, in a larger plastic bag, which was then sealed and kept frozen at -80°C until use. The TropBio ELISA kit for detecting and quantifying W. bancrofti antigen in filter-paper specimens (catalogue no. 03-010-05) was used accordine; to the manufacturer’s instructions. Three protrusions from each filterpaper disk were cut into small pieces with scissors and placed in an elution tube with 200 p.L sample diluent. After boiling treatment and centrifugation, the supernatant was tested in dunlicate. The mean OD of the 2 tests was used to convert the response to antigen units and to allocate specimens to titre groups as described above for the TropBio serum test. Results Among the 5 1 endemic study individuals, 24 were mf positive (range 25-10 948 mf/mL) and 27 were mf negative. All sera were examined for CFA with the 3 tests, and results expressed as positive and negative responses are shown in the Table. Sera from all mfpositive endemic individuals were CFA positive, and sera Table. Characteristics of the study individuals, response, in the three examined tests for specific Category of individuals Endemic, mf positive Endemic, mf negative Non-endemic controls
Number examined (males/females)
Mean age in years (range)
279
FILARIASIS
from all non-endemic control individuals were CFA negative, in all 3 tests. For the mf-negative endemic individuals, sera from 10 (37%) were CFA positive in the TropBio serum test. Sera from 9 of these individuals were also positive in the TropBio filter-paper test, and 8 of these were also uositive in the ICT serum card test. A high degree of resionse similarity was therefore seen for the 3 tests, when the outcome was expressed as a positive/negative response. The sensitivity for detecting microfilaraemic cases was 100% for all tests. The TrooBio tests give not onlv a nositive/neeative result, but also a graded response expressed’in either antigen units or titre group. The graded responses of the filter-paper version and the serum version of the TropBio test for the endemic individuals are compared in Figure 1. Responses obtained with the 2 tests were statistically significantly correlated both when expressed in antigen units (Fig. IA; Pearson correlation coefficient r = 0.679; P < 0.01) and when expressed in titre groups (Fig. 1B; Pearson correlation coefficient r = 0.925; P < 0.01). To analyse the relationship between mf intensity and the response intensity in the 2 TropBio test versions, the antigen-positive individuals were divided into 3 categories: those who were mf negative (category 1; n = 7), those who had low-level microfilaraemia of lo999 mf/mL (category 2; n = 13), and those who had high-level microfilaraemia of 2 1000 mf/mL (category 3; n = 11). The single individual who was CFA positive in the serum test and CFA negative in the filter-paper test was excluded from this analvsis. The mean CFA responses for the 3 categories are shown in Figure 2. Also shown in this Figure are the mean CFA responses for endemic individuals who were both mf and CFA negative (category 4; n = 19) and for the non-endemic control individuals (category 5; n = 10). Both when expressed in antigen units (Fig. 2A) and in titre groups (Fig. 2B), and both for the filter-paper version and for the serum version of the test, the mean CFA response increased from category 1 to category 2 and from category 2 to category 3, i.e., with increasing level of microfilaraemia. For the filter-paper version of the test, this increase was statistically significant (t-test; P < 0.05 from category 1 to 2 and P < 0.01 from category 2 to 3; both for responses expressed in antigen units and responses expressed in titre group), whereas it was not statistically significant for the serum version (t-test; P > 0.05 for all tests). Further analysis of individual data from the mf-positive individuals indicated a statistically significant level of correlation between mf intensity and antigen response for the filter-paper test (Pearson correlation coefficient r = 0.532 and r = 0.559 for resnonses exnressed in antigen units and titre group, respectively; P’< 0.0 1 for both tests) but not for the serum version of the test (Pearsoncorrelation coefficient r = 0.359 and r = 0.311 for responses expressed in antigen units and titre group, respectively; P > 0.05 for both tests). Discussion Research carried out during the past decade to develop reliable tests for detection of CFA in W. bancrofti-
and the number of individuals W. bancrofti circulating antigen
with
a positive
serum
Number positive in the TropBio serum test
Number positive in the TropBio filter-paper test
Number positive in the ICT serum card test
24 (18/6)
37.4 (12-70)
24
24
24
27 (16/11)
34.4 (12-67)
10”
9
8
10 (6/4)
34.4 (26-45)
0
0
0
“Nine of these individuals were positive in the TropBio filter-paper test, and 8 of these were positive in the ICT serum card test.
PAUL E. SIMONSEN AND SAMUEL K. DUNYO
280
5
50
500
5000
50000
TropBio serum test (antigen units)
TropBio serum test (titre group) Fig. 1. The correlation between circulating lV’. bancrojci antigen levels measured with the TropBio serum test and the TropBio filterpaper test, with results expressed in antigen units (A) or titre group (B) for the endemic study individuals. The regression line is shown. Stippled lines indicate cut-off levels between positive and negative tests. Petals on the dots indicate number of individuals, if several individuals have similar results.
infected individuals has been highly successful. Based on gained information and products, diagnostic kits with high sensitivity and specificity have been developed and are now commercially available. The use of CFA as diagnostic markers instead of mf provides several advantages. Thus, blood specimens can be collected at any time of the day (WEIL et al., 1987; hh4MIE et al., 1994), and large numbers of specimens can be screened in a short period oftime (WEILetal., 1997; NICOLAS, 1997). Moreover, the presently available kits detect antigens originating from adult worms (WEIL & LIlTIS, 1987; CHANTEAU et al., 1994a; WEIL et al., 1996) and therefore diagnose adult worm infections, as opposed to the parasitological techniques, which diagnose mf. These kits therefore also constitute novel research tools for investigation of infection dynamics and disease development in bancroftian filariasis. The present study evaluated and compared 3 commercially available kits for diagnosis of bancroftian fllariasis based on detection of CFA. Sera from mf-positive individuals responded positively in all 3 antigen tests, thus giving a sensitivity of 100% for
detection of microfilaraemic cases. In addition, approximately one-third of the mf-negative endemic individuals responded positively in these tests. A higher prevalence of CFA positivity than mf positivity has also been reported from other endemic areas &AMMIE et al., 1994; ROCHA et al., 1996; SIMONSEN et al., 1996; NICOLAS et al., 1997; WEIL et al., 1997). Since the specificity of the antigen-capturing monoclonal antibodies used in these tests is extremely high (MORE & COPEMAN, 1990; WEIL et al., 1987; 1997) it is likely that most CFA-positive but mf-negative individuals are not falsely positive for CFA, but rather have undetected ultralow mf intensities or cryptic adult worm infections without mf. Use of CFA detection for diagnosis therefore provides more accurate information on infection status than mf detection in infected individuals with few or no mf. Discrepancies in positive/negative response outcomes between the tests were limited and were seen only in the group of mf-negative endemic individuals. Among these, 1 individual was positive in the TropBio serum test only, and another was positive in both versions of the TropBio
CIRCULATING
ANTIGENS
FOR DIAGNOSIS
OF BANCROFTIAN
100000,
A
50000
10000 u) x $s
1000
5
500
5 s
100 50
.z
10 50005 i 3
Category
4
n. 5
of individuals
Category
of individuals
Fig. 2. Relationship between circulating W. bancrojii antigen (CFA) level and infection status category for the TropBio serum test (black bars) and the TropBio filter-paper test (white bars), with results expressed in antigen units (A) or titre group (B). Category 1 = CFA positive, mf negative; Category 2 = CFA Category 3 = CFA positive, positive, lo-999 mf/mL; 2 1000 mf/mL, Category 4 = CFA negative, mf negative; Category 5 = non-endemic controls.
test but not in the ICT card test. These individuals both had low antigen titres in the TropBio tests (titre group 3 and 4 in the serum test, and titre group 2 and 3 in the filter-paper test, respectively). A slightlylower sensitivity of the ICT card test than of the TronBio serum test among endemic mf-negative individuals was also reported by WEIL et al. ( 1997). The graded responses of the serum and filter-paper versions of the TronBio test showed a clear and statistically significant coielation, when results were expressed both in antigen units and in titre groups. Overall, however, responses in the filter-paper version were slightly lower than those of the serum version, probably because the amount of serum absorbed in filter paper for a test in the filter-paper version (- 3.5 &) is 3-4 times less than the amount of serum used for a test in the serum version (12.5 pL). Owing to the 4-fold difference in antigen concentration between each of the 7 standards used for preparation of the standard curve, and owing to the relatively high antigen content in antigen-positive serum specimens, this difference in response appears not to have a marked influence on the results obtained. However, further harmonization in design of the 2 test
FILARIASIS
281
versions would be desirable in order to make comparison of test results an easier task. The TropBio filter-paper version has obvious practical advantages for large-scale field work, when compared to the serum version. However, a poor sensitivity of the test was noted by several laboratories in the early period of marketing (1996), and resulted in serious concern about et al., 1997b; GYAPONG et its performance (SIMONSEN al.. 1998). The oroblems were at least nartlv caused bv a defective batch of antiserum, which has later been replaced (TropBio, personal communication). It is our personal experience that tests carried out with kits obtained from earlv 1997 onwards nerform satisfactorilv, as was also observed in the present study in which the filter-paper test version produced results comparable to those of the serum version. LUITHA et al. (1998) and ITOH et al. (1998) also reported a high level of correlation between the concentration of CFA in filter-paper and serum specimens collected from the same individuals, by use of the TropBio ELISA kit. These studies, however, did not use the standardized TropBio filter-paper collection kit, and the type of filter paper and the amount of blood used differed from those recommended by the manufacturer. Analysis of the relationship between mf intensity and the response intensity in the 2 TropBio test versions indicated a positive association, which was statistically significant for the filter-paper version but not for the serum version. Since both versions detect antigens from adult worms, these associations were probably indirect and resulting from a positive association between the number of adult worms and the number of mf in infected individuals. A positive association between circulating antigenaemia and microfilaraemia has also, at various degrees, been reported from other studies carried out in endemic areas, by use of the serum version of the 1990; LAMMIE et al., TropBio kit (MORE & COPEMAN, 1994; NICOLAS et aZ., 1997). The results obtained with the 3 tests thus showed a good level of agreement. The tests have various qualities for application in different settings and for different purposes. The ICT card test can be performed by technical staff after a minimum of training and with limited laboratory facilities. It seems ideal for diagnosis in clinical laboratories and for surveillance in large control programmes, when only a positive or negative result is required. The TropBio tests provide a graded response, probably reflecting the load of adult worm infection, but they also require more advanced laboratory equipment (including an ELISA reader) and well-trained laboratory staff, and therefore they are more suitable for research projects. However, all 3 tests appear to be promising future tools for diagnosis of IV. bancrofti infection. Acknowledgements We are nrateful to Collins Ahorlu and TohnFentene Noauchi
Memorial-Institute for Medical Research) for exce&t
assis-
tance in the field, and to Mette Lund and i‘Jernille Lund Strom (Danish Bilharziasis Laboratory) for excellent assistance in the immunology laboratory. The study was supported by funds from the Danish Bilharziasis Laboratory.
References Chanteau, S., Glaziou, I’., Luquiaud, I’., Plichart, C., MouliaPelat, J. I’. &Cartel, J. L. (1994a). Og4C3 circulating antigen, anti-&&a malayi IgG and IgG4 titres in Wucheretiu bancrofi infected patients, according to their parasitological status. Tropical Medicine and Parasitology, 45,255-257. Chanteau, S., Moulia-Pelat, J. I’., Glaziou, I’., Nguyen, N. L., Luquiaud, I’., Plichart, C., Martin, P. M. V. & Cartel, J. L. (1994b). Og4C3 circulating antigen: a marker of infection and adult worm burden in Wuchereria bancrofti filariasis. Journal oflnfectious Diseases, 170, 247-250. Dzodzomenyo, M., Dunyo, S. K., Ahorlu, C. K., Coker, W. Z., Appawu, M. A., Pedersen, E. M. & Simonsen, P. E. (1999). Bancroftian filariasis in an irrigation project community in southern Ghana. TropicalMedicine andlntemational Health, 4, 13-18.
7.82
PAUL E. SIMONSEN AND SAMUEL K. DUNYO
Gyapong, J. O., Omane-Badu, K. & Webber, R. H. (1998). Evaluation of the filter paper blood collection method for detecting Og4C3 circulating antigen in bancroftian filariasis. Transactions of the Royal Society of Tropical Medicine and Hygiene, 92,407-410.
Itoh, M., Gunawardena, N. K., Qiu, X.-G., Weerasooriya, M. V. & Kimura. E. ( 1998). The use of whole blood absorbed on filter paper to detect Ekchereria bancrofti circulating antigen. Transactions of the Royal Society of Tropical Medicine Hygiene, 92,513-515.
and
Lalitha, P., Ravichandran, M., Suba, S., Kaliraj, I’., Narayanan, R. B. & Jayaraman, K. (1998). Quantitative assessment of circulating antigens in human lymphatic filariasis: a field evaluation of monoclonal antibody-based ELISA using blood collected on filter strips. Tropical Medicine and Inmttational Health, 3,41-45.
Lammie, P. J., Hightower, A. W. & Eberhard, M. L. (1994). Age-specific prevalence of antigenemia in a kuchereria bancroft-exposed population. American Journal of Tropical Medicine and Hygiene, 51?348-355. McMahon, J. E. & Stmonsen, P. E. (1996). Filariases. In: Manson’s TropicalDiseases, Cook, G. C. (editor), 20th edition. London: WB Saunders, pp. 1322-1368. McMahon, J. E., Marshall, T. F. de C., Vaughan, J. P. AtAbaru, D. E. (1979). Bancrofttan tilariasis: a comparison of microfilariae counting techniques using counting chamber, standard slide and membrane (nuclepore) filtration. Annals of Tropical Medicine and Parasitology, 13,457-464. More, S. J. & Copeman, D. B. (1990). A highly specific and sensitive monoclonal antibody-based ELISA for the detection of circulating antigen in bancroftian filariasis. Tropical Medicine and Parasitology, 41, 403-406. Nicolas, L. (1997). New tools for diagnosis and monitoring of bancroftian filariasis parasitism: the Polynesian experience.
Rocha, A., Addiss, D., Ribeiro, M. E., Noroes, J., Baliza, M., Medeiros, Z. & Dreyer, G. (1996). Evaluation of the Og4C3 ELISA in Wuchereria bancrofti infection: infected personswith undetectable or ultra-low microfilarial densities. Tropical Medicine and International Health, 1,859-864. Simonsen, P. E., Lemnge, M. M., Msangeni, H. A., Jakobsen, P. H. & Bygbjerg, I. C. (1996). Bancroftian filariasis: the patterns of filarial-specific immunoglobulin Gl (IgGl), IgG4, and circulating antigens in an endemic community of northeastern Tanzania. Am&an Journal of Tropical Medicine and Hygiene, 55,69-75.
Simonsen, P. E., Niemamr, L. & Meyrowitsch, D. W. (1997a). Wuchereria bancrofti in Tanzania: microfilarial periodicity and effect of blood sampling time on microfilarial intensities. Tropical Medicine and International Health, 2, 153-158. Simonsen, P. E., Pedersen, E. M., Lemnge, M. M. & Michael, E. (1997b). Lymphatic tilariasis research and control in Africa. Parasitology Today, 13,413-415. Weil, G. J. & Liftis, F. (1987). Identification and partial characterization of a parasite antigen in sera from humans infected with Wuchereriabancrofki. Journal of Immunology, 138, 3035-3041. Weil, G. J., Jain, D. C., Santhanam, S., Malhotra, A., Kumar, H., Sethumadhavan, K. V. P., Liftis, F. & Ghosh, T. K. (1987). A monoclonal antibody-based enzyme immunoassay for detecting parasite antigenemia in bancroftian filariasis.
Journalof InfectiousDiseases, 156,350-355.
Weil, G. J., Ramzy, R. M. R., Chandrashekar, R., Gad, A. M., Lowrie, R. C., Jr & Faris, R. (1996). Parasite antigenemia without microfilaraemia in bancroftian filariasis. American Journal of Tropical Medicine and Hygiene, 55,333-337.
Weil, G. J., Lammie, P. J. &Weiss, N. (1997). The ICT filariasis test: a rapid-format antigen test for diagnosis of bancroftian filariasis. Parasitology Today, 13,401-404.
Parasitology Today, 13,370-375.
Nicolas, L., Plichart, C., Nguyen, L. N. & Moulia-Pelat, J. P. (1997). Reduction of Wuchereria bancrofti adult worm circulating antigen after annual treatments of diethylcarbamazine combined with ivermectin in French Polynesia. Journal of Infectious Diseases, 17.5,489-492.
1 Announcement
Received 8 December 1998; revised 2 March for publication 2 March 1999
1999; accepted
1 2000 Award in honour of Fred L. Soper (1893-1976) for publications in the field of Inter-American Health
This is an announcement and call for submission of nominations for the 2000 award in honour of Fred L. Soper, former Director of the Pan American Health Organization (the World Health Organization Regional Office for the Americas) from 1947 to 1958. The Award is presented annually to the author or authors of an original scientific contribution comprising new information on, or new insights into, the broad field of public health, with special relevance to Latin America or the Caribbean or both. This may consist of a report, an analysis ofnew data, experimental or observational, or a new approach to analysing available data. Preference is given to studies involving more than one discipline and to papers related to infectious diseases,a life-long concern of Dr Soper. Only papers already published in scientific journals listed in the Index Medicus or in the official journals of the Pan American Health Organization are eligible for consideration. Furthermore, the award is limited to contributions by authors whose principal affiliation is with teaching, research or service institutions located in the countries of Latin America and the Caribbean (including the Centers of the Pan American Health Organization). The Award consists of a certificate and a prize of US$ 1000. Papers submitted by or on behalf of their authors may be considered for the Fred L. Soper Award. For the 2000 Award, only papers published during the calendar year 1999 will be considered; all submissions must be received by 31 March 2000 at the following address: Fred L. Soper Award Program, PAHEF, 525 23rd Street NW, Washington, DC 20037, USA.
Comparative evaluation of three new tools for diagnosis based on detection of specific circulating antigens
of bancroftian
filariasis
Paul E. Simonsenl and Samuel K. Dunyo* ‘Danish Bilharziask Laboratory, Jaegersborg Alle 1 D, 2920 Charlottenlund, Denmark; ‘Noguchi Memorial Institute for Medical Research, University of Ghana, I? 0. Box 25, Legon, Ghana Abstract Three new and commercially available tools for diagnosis of Wuchereria bancrofii infections based on detection of specific circulating antigens were evaluated and compared in the same group of individuals from a highly endemic village in southern Ghana. The tests were: (1) the ICT card test for serum specimens; (2) the TropBio ELISA test for serum specimens; and (3) the TropBio ELISA test for filter-paper specimens. A high degree of positive/negative response similarity was observed for the 3 tests, and the sensitivity for detecting microfilaraemic cases was 100% for all tests. The antigen levels measured in the TropBio serum test and the TropBio filter-paper test were statistically significantly correlated. Among antigen-positive endemic individuals the antigen levels in these 2 tests furthermore showed a positive association with the microfilarial intensity, but a statistical significant correlation was seen only for the filter-paper version of the test. The results are promising for the use of the 3 tests as diagnostic tools in bancroftian filariasis. Keywords:
Wuchereria bancrofi, filariasis, diagnosis, circulating antigens, Ghana
Introduction Diagnosis of bancroftian filariasis, resulting from infection with the mosquito-borne filarial nematode Wuchereria bancrofti, relied until recently almost exclusively on detection and identification of microfilariae (mf) in blood specimens (MCMAHON & SIMONSEN, 1996; NICOLAS, 1997). In most geographical areas, W. bancrofti mf have a nocturnal periodicity, with highest intensity in the peripheral blood at night and few or none during the davd ~SIMONSEN et al.. 1997a). Blood sneci: mens for parasrtological diagnosis m&t therefo;e be collected at night. Many attempts have been made to develop simple and reliable immunological tests, which can also be used during daytime. While tests for antibody detection have generally suffered from lack of specificity and inability to distinguish between current and past infection, other tests based on detection of specific circulating W! bancrofiifilarial antigens (CFA) have been highly successful, and a number of kits have recently become commercially available. The first of these kits to appear on the market was made by Tropical Biotechnology Pty Ltd (TropBio) in Australia. It utilizes a monoclonal antibody, Og4C3 (MORE & COPEMAN, 1990), in a sandwich enzymelinked immunosorbent assay (ELISA) to detect CFA in serum specimens. A high specificity and sensitivity of the test has been confirmed in several studies (e.g., MORE & COPEMAN, 1990; CHANTEAU et al., 199aa, 1994b; ROCHA et al.. 1996: NICOLAS. 1997; NICOLAS et al.. 1997; WEIL et al., (997). Another version of this tes< adapted for analysis of finger-prick blood specimens collected on filter-paper disks, has been developed by the same company. More recently, ICT Diagnostics in Australia produced a simple immunochromatographic card test for detection of CFA in serum specimens (WEIL et al., 1997). In this test, antigen capture is accomplished by another monoclonal antibody, AD. 12 @VEIL & L~TIS, 1987), with high sensitivity and specificity to W. bancroftiantieen N[IEIL et al.. 1987). Serum and wash buffer is ippliedyo specifically indicated pads on the test card, and within a few minutes a coloured reaction develops which indicates whether the serum specimen is positive or negative for the antigen. This test requires no advanced equipment, which makes it suitable for many endemic areas in the developing world. The present study evaluated and compared 3 tests for diagnosis of bancroftian filariasis based on detection of Address for correspondence: Dr Paul E. Simonsen, Danish Bilhaniasis Laboratory, Jaegersborg Alle 1 D, 2920 Charlottenlund, Denmark; phone +45 77327732, fax +45 77327733, e-mail [email protected]
CFA, namely (1) the ICT card test for serum specimens, (2) the TropBio ELISA test for serum specimens, and (3) the TropBio ELISA test for filter-paper specimens. The same group of individuals from an endemic village in southern Ghana was examined with all 3 tests, and results were compared to the mf status and mf intensity for the same individuals. Materials and Methods Study individuals and study design Following a baseline survey for bancroftian filariasis in Gomoa Okyereko village in the Central Region of Ghana (DZODZOMENYO et al., 1999), a number of villagers aged 12 years or more were requested to participate in the study. Fifty-one individuals responded positively and these constitute the endemic study population. After registration, blood specimens were sampled between 21:00 and 24:O0. Finger-prick blood was collected in capillary tubes for quantification of W. bancrofi mf and on filter-paper disks for determination of specific W. bancrofti CFA. Immediately thereafter, 5 mL of venous blood was collected from the same individual (using the Vacutainer system) for serum preparation. The exact time of sampling was noted, and all specimens from the same individual were collected within a period of 5 min. Ten individuals from Denmark, where no transmission of W. bancrofti occurs, served as non-endemic controls. The procedure for sampling of blood specimens from the control individuals was similar to that described for the endemic individuals, with the exception that sampling took place at daytime and that no specimens were collected and examined for mf. Examination of blood for mf et al., The counting-chamber technique (MCMAHON 1979) was used to quantify the blood microfilaraemia. In brief, 100 ~.LL of finger-prick blood was collected in heparinized capillary tubes, and immediately transferred to specimen tubes containing 900 ~.ILof 3% acetic acid. The number of mf in the specimens was later counted in a counting chamber, using a microscope. To compensate for the effect of sampling time on mf intensities, timeadjusted mf intensities were calculated by multiplying the observed intensities with a time-specific adjustment factor (see SIMONSEN et al., 1997a). Only time-adjusted mf intensities were used in the data analysis. Serum preparation Venous blood was left in a fridge at 5°C overnight for clotting, Thereafter, the clot was removed and the remaining specimen was centrifuged. The supernatant of serum was frozen at -20°C until use.
CIRCULATING
ANTIGENS FOR DIAGNOSIS OF BANCROFTIAN
The ICT card test The serum version of the ICT card test (ICT Diagnostics, Balgowlah, Australia; catalogue no. ICT FLOl) was used, according to the manufacturer’s instructions. In brief, 50 uL of serum was applied to the pink specimen pad, and 2 drops of wash reagent were applied at the indicated white pad. The card was closed, and the result read after 5-10 min. The card was re-examined 1 and 24 h after application of serum, but the results were similar to those obtained at the first reading. The TropBio serum test The TropBio ELISA kit for detecting and quantifying W bancrofti antigen in serum specimens (JCIJ Tropical Biotechnology Pty Ltd, Townsville, Australia; catalogue no. 03-010-01) was used according to the manufacturer’s instructions. After boiling pre-treatment, each serum specimen was tested in duplicate. The 7 standards with known antigen content enclosed in the kit were similarly tested in duplicate, and the results were used to prepare a standard curve relating optical density (OD) and antigen content. The mean OD of each serum specimen was used to determine the response in antigen units (from the standard curve) as well as the antigen titre group (in relation to the 7 standards) to which the specimen belonged. Serum specimens with 2 32 antigen units (2 titre group 3) were considered positive for CFA, and specimens with a response 3 the OD of standard 7 were assigned a fixed value of 32 000 antigen units (titre group 8). The TropBio filter-paper test Finger-prick blood was collected on TropBio filterpaper disks (catalogue no. 05-001-07). All 6 protrusions on 1 disk were saturated with blood from 1 individual, whereafter the disks were dried in air at room temperature overnight. The dried disks were packed individually in small punched plastic bags. These bags were placed, together with a silica-gel sachet, in a larger plastic bag, which was then sealed and kept frozen at -80°C until use. The TropBio ELISA kit for detecting and quantifying W. bancrofti antigen in filter-paper specimens (catalogue no. 03-010-05) was used accordine; to the manufacturer’s instructions. Three protrusions from each filterpaper disk were cut into small pieces with scissors and placed in an elution tube with 200 p.L sample diluent. After boiling treatment and centrifugation, the supernatant was tested in dunlicate. The mean OD of the 2 tests was used to convert the response to antigen units and to allocate specimens to titre groups as described above for the TropBio serum test. Results Among the 5 1 endemic study individuals, 24 were mf positive (range 25-10 948 mf/mL) and 27 were mf negative. All sera were examined for CFA with the 3 tests, and results expressed as positive and negative responses are shown in the Table. Sera from all mfpositive endemic individuals were CFA positive, and sera Table. Characteristics of the study individuals, response, in the three examined tests for specific Category of individuals Endemic, mf positive Endemic, mf negative Non-endemic controls
Number examined (males/females)
Mean age in years (range)
279
FILARIASIS
from all non-endemic control individuals were CFA negative, in all 3 tests. For the mf-negative endemic individuals, sera from 10 (37%) were CFA positive in the TropBio serum test. Sera from 9 of these individuals were also positive in the TropBio filter-paper test, and 8 of these were also uositive in the ICT serum card test. A high degree of resionse similarity was therefore seen for the 3 tests, when the outcome was expressed as a positive/negative response. The sensitivity for detecting microfilaraemic cases was 100% for all tests. The TrooBio tests give not onlv a nositive/neeative result, but also a graded response expressed’in either antigen units or titre group. The graded responses of the filter-paper version and the serum version of the TropBio test for the endemic individuals are compared in Figure 1. Responses obtained with the 2 tests were statistically significantly correlated both when expressed in antigen units (Fig. IA; Pearson correlation coefficient r = 0.679; P < 0.01) and when expressed in titre groups (Fig. 1B; Pearson correlation coefficient r = 0.925; P < 0.01). To analyse the relationship between mf intensity and the response intensity in the 2 TropBio test versions, the antigen-positive individuals were divided into 3 categories: those who were mf negative (category 1; n = 7), those who had low-level microfilaraemia of lo999 mf/mL (category 2; n = 13), and those who had high-level microfilaraemia of 2 1000 mf/mL (category 3; n = 11). The single individual who was CFA positive in the serum test and CFA negative in the filter-paper test was excluded from this analvsis. The mean CFA responses for the 3 categories are shown in Figure 2. Also shown in this Figure are the mean CFA responses for endemic individuals who were both mf and CFA negative (category 4; n = 19) and for the non-endemic control individuals (category 5; n = 10). Both when expressed in antigen units (Fig. 2A) and in titre groups (Fig. 2B), and both for the filter-paper version and for the serum version of the test, the mean CFA response increased from category 1 to category 2 and from category 2 to category 3, i.e., with increasing level of microfilaraemia. For the filter-paper version of the test, this increase was statistically significant (t-test; P < 0.05 from category 1 to 2 and P < 0.01 from category 2 to 3; both for responses expressed in antigen units and responses expressed in titre group), whereas it was not statistically significant for the serum version (t-test; P > 0.05 for all tests). Further analysis of individual data from the mf-positive individuals indicated a statistically significant level of correlation between mf intensity and antigen response for the filter-paper test (Pearson correlation coefficient r = 0.532 and r = 0.559 for resnonses exnressed in antigen units and titre group, respectively; P’< 0.0 1 for both tests) but not for the serum version of the test (Pearsoncorrelation coefficient r = 0.359 and r = 0.311 for responses expressed in antigen units and titre group, respectively; P > 0.05 for both tests). Discussion Research carried out during the past decade to develop reliable tests for detection of CFA in W. bancrofti-
and the number of individuals W. bancrofti circulating antigen
with
a positive
serum
Number positive in the TropBio serum test
Number positive in the TropBio filter-paper test
Number positive in the ICT serum card test
24 (18/6)
37.4 (12-70)
24
24
24
27 (16/11)
34.4 (12-67)
10”
9
8
10 (6/4)
34.4 (26-45)
0
0
0
“Nine of these individuals were positive in the TropBio filter-paper test, and 8 of these were positive in the ICT serum card test.
PAUL E. SIMONSEN AND SAMUEL K. DUNYO
280
5
50
500
5000
50000
TropBio serum test (antigen units)
TropBio serum test (titre group) Fig. 1. The correlation between circulating lV’. bancrojci antigen levels measured with the TropBio serum test and the TropBio filterpaper test, with results expressed in antigen units (A) or titre group (B) for the endemic study individuals. The regression line is shown. Stippled lines indicate cut-off levels between positive and negative tests. Petals on the dots indicate number of individuals, if several individuals have similar results.
infected individuals has been highly successful. Based on gained information and products, diagnostic kits with high sensitivity and specificity have been developed and are now commercially available. The use of CFA as diagnostic markers instead of mf provides several advantages. Thus, blood specimens can be collected at any time of the day (WEIL et al., 1987; hh4MIE et al., 1994), and large numbers of specimens can be screened in a short period oftime (WEILetal., 1997; NICOLAS, 1997). Moreover, the presently available kits detect antigens originating from adult worms (WEIL & LIlTIS, 1987; CHANTEAU et al., 1994a; WEIL et al., 1996) and therefore diagnose adult worm infections, as opposed to the parasitological techniques, which diagnose mf. These kits therefore also constitute novel research tools for investigation of infection dynamics and disease development in bancroftian filariasis. The present study evaluated and compared 3 commercially available kits for diagnosis of bancroftian fllariasis based on detection of CFA. Sera from mf-positive individuals responded positively in all 3 antigen tests, thus giving a sensitivity of 100% for
detection of microfilaraemic cases. In addition, approximately one-third of the mf-negative endemic individuals responded positively in these tests. A higher prevalence of CFA positivity than mf positivity has also been reported from other endemic areas &AMMIE et al., 1994; ROCHA et al., 1996; SIMONSEN et al., 1996; NICOLAS et al., 1997; WEIL et al., 1997). Since the specificity of the antigen-capturing monoclonal antibodies used in these tests is extremely high (MORE & COPEMAN, 1990; WEIL et al., 1987; 1997) it is likely that most CFA-positive but mf-negative individuals are not falsely positive for CFA, but rather have undetected ultralow mf intensities or cryptic adult worm infections without mf. Use of CFA detection for diagnosis therefore provides more accurate information on infection status than mf detection in infected individuals with few or no mf. Discrepancies in positive/negative response outcomes between the tests were limited and were seen only in the group of mf-negative endemic individuals. Among these, 1 individual was positive in the TropBio serum test only, and another was positive in both versions of the TropBio
CIRCULATING
ANTIGENS
FOR DIAGNOSIS
OF BANCROFTIAN
100000,
A
50000
10000 u) x $s
1000
5
500
5 s
100 50
.z
10 50005 i 3
Category
4
n. 5
of individuals
Category
of individuals
Fig. 2. Relationship between circulating W. bancrojii antigen (CFA) level and infection status category for the TropBio serum test (black bars) and the TropBio filter-paper test (white bars), with results expressed in antigen units (A) or titre group (B). Category 1 = CFA positive, mf negative; Category 2 = CFA Category 3 = CFA positive, positive, lo-999 mf/mL; 2 1000 mf/mL, Category 4 = CFA negative, mf negative; Category 5 = non-endemic controls.
test but not in the ICT card test. These individuals both had low antigen titres in the TropBio tests (titre group 3 and 4 in the serum test, and titre group 2 and 3 in the filter-paper test, respectively). A slightlylower sensitivity of the ICT card test than of the TronBio serum test among endemic mf-negative individuals was also reported by WEIL et al. ( 1997). The graded responses of the serum and filter-paper versions of the TronBio test showed a clear and statistically significant coielation, when results were expressed both in antigen units and in titre groups. Overall, however, responses in the filter-paper version were slightly lower than those of the serum version, probably because the amount of serum absorbed in filter paper for a test in the filter-paper version (- 3.5 &) is 3-4 times less than the amount of serum used for a test in the serum version (12.5 pL). Owing to the 4-fold difference in antigen concentration between each of the 7 standards used for preparation of the standard curve, and owing to the relatively high antigen content in antigen-positive serum specimens, this difference in response appears not to have a marked influence on the results obtained. However, further harmonization in design of the 2 test
FILARIASIS
281
versions would be desirable in order to make comparison of test results an easier task. The TropBio filter-paper version has obvious practical advantages for large-scale field work, when compared to the serum version. However, a poor sensitivity of the test was noted by several laboratories in the early period of marketing (1996), and resulted in serious concern about et al., 1997b; GYAPONG et its performance (SIMONSEN al.. 1998). The oroblems were at least nartlv caused bv a defective batch of antiserum, which has later been replaced (TropBio, personal communication). It is our personal experience that tests carried out with kits obtained from earlv 1997 onwards nerform satisfactorilv, as was also observed in the present study in which the filter-paper test version produced results comparable to those of the serum version. LUITHA et al. (1998) and ITOH et al. (1998) also reported a high level of correlation between the concentration of CFA in filter-paper and serum specimens collected from the same individuals, by use of the TropBio ELISA kit. These studies, however, did not use the standardized TropBio filter-paper collection kit, and the type of filter paper and the amount of blood used differed from those recommended by the manufacturer. Analysis of the relationship between mf intensity and the response intensity in the 2 TropBio test versions indicated a positive association, which was statistically significant for the filter-paper version but not for the serum version. Since both versions detect antigens from adult worms, these associations were probably indirect and resulting from a positive association between the number of adult worms and the number of mf in infected individuals. A positive association between circulating antigenaemia and microfilaraemia has also, at various degrees, been reported from other studies carried out in endemic areas, by use of the serum version of the 1990; LAMMIE et al., TropBio kit (MORE & COPEMAN, 1994; NICOLAS et aZ., 1997). The results obtained with the 3 tests thus showed a good level of agreement. The tests have various qualities for application in different settings and for different purposes. The ICT card test can be performed by technical staff after a minimum of training and with limited laboratory facilities. It seems ideal for diagnosis in clinical laboratories and for surveillance in large control programmes, when only a positive or negative result is required. The TropBio tests provide a graded response, probably reflecting the load of adult worm infection, but they also require more advanced laboratory equipment (including an ELISA reader) and well-trained laboratory staff, and therefore they are more suitable for research projects. However, all 3 tests appear to be promising future tools for diagnosis of IV. bancrofti infection. Acknowledgements We are nrateful to Collins Ahorlu and TohnFentene Noauchi
Memorial-Institute for Medical Research) for exce&t
assis-
tance in the field, and to Mette Lund and i‘Jernille Lund Strom (Danish Bilharziasis Laboratory) for excellent assistance in the immunology laboratory. The study was supported by funds from the Danish Bilharziasis Laboratory.
References Chanteau, S., Glaziou, I’., Luquiaud, I’., Plichart, C., MouliaPelat, J. I’. &Cartel, J. L. (1994a). Og4C3 circulating antigen, anti-&&a malayi IgG and IgG4 titres in Wucheretiu bancrofi infected patients, according to their parasitological status. Tropical Medicine and Parasitology, 45,255-257. Chanteau, S., Moulia-Pelat, J. I’., Glaziou, I’., Nguyen, N. L., Luquiaud, I’., Plichart, C., Martin, P. M. V. & Cartel, J. L. (1994b). Og4C3 circulating antigen: a marker of infection and adult worm burden in Wuchereria bancrofti filariasis. Journal oflnfectious Diseases, 170, 247-250. Dzodzomenyo, M., Dunyo, S. K., Ahorlu, C. K., Coker, W. Z., Appawu, M. A., Pedersen, E. M. & Simonsen, P. E. (1999). Bancroftian filariasis in an irrigation project community in southern Ghana. TropicalMedicine andlntemational Health, 4, 13-18.
7.82
PAUL E. SIMONSEN AND SAMUEL K. DUNYO
Gyapong, J. O., Omane-Badu, K. & Webber, R. H. (1998). Evaluation of the filter paper blood collection method for detecting Og4C3 circulating antigen in bancroftian filariasis. Transactions of the Royal Society of Tropical Medicine and Hygiene, 92,407-410.
Itoh, M., Gunawardena, N. K., Qiu, X.-G., Weerasooriya, M. V. & Kimura. E. ( 1998). The use of whole blood absorbed on filter paper to detect Ekchereria bancrofti circulating antigen. Transactions of the Royal Society of Tropical Medicine Hygiene, 92,513-515.
and
Lalitha, P., Ravichandran, M., Suba, S., Kaliraj, I’., Narayanan, R. B. & Jayaraman, K. (1998). Quantitative assessment of circulating antigens in human lymphatic filariasis: a field evaluation of monoclonal antibody-based ELISA using blood collected on filter strips. Tropical Medicine and Inmttational Health, 3,41-45.
Lammie, P. J., Hightower, A. W. & Eberhard, M. L. (1994). Age-specific prevalence of antigenemia in a kuchereria bancroft-exposed population. American Journal of Tropical Medicine and Hygiene, 51?348-355. McMahon, J. E. & Stmonsen, P. E. (1996). Filariases. In: Manson’s TropicalDiseases, Cook, G. C. (editor), 20th edition. London: WB Saunders, pp. 1322-1368. McMahon, J. E., Marshall, T. F. de C., Vaughan, J. P. AtAbaru, D. E. (1979). Bancrofttan tilariasis: a comparison of microfilariae counting techniques using counting chamber, standard slide and membrane (nuclepore) filtration. Annals of Tropical Medicine and Parasitology, 13,457-464. More, S. J. & Copeman, D. B. (1990). A highly specific and sensitive monoclonal antibody-based ELISA for the detection of circulating antigen in bancroftian filariasis. Tropical Medicine and Parasitology, 41, 403-406. Nicolas, L. (1997). New tools for diagnosis and monitoring of bancroftian filariasis parasitism: the Polynesian experience.
Rocha, A., Addiss, D., Ribeiro, M. E., Noroes, J., Baliza, M., Medeiros, Z. & Dreyer, G. (1996). Evaluation of the Og4C3 ELISA in Wuchereria bancrofti infection: infected personswith undetectable or ultra-low microfilarial densities. Tropical Medicine and International Health, 1,859-864. Simonsen, P. E., Lemnge, M. M., Msangeni, H. A., Jakobsen, P. H. & Bygbjerg, I. C. (1996). Bancroftian filariasis: the patterns of filarial-specific immunoglobulin Gl (IgGl), IgG4, and circulating antigens in an endemic community of northeastern Tanzania. Am&an Journal of Tropical Medicine and Hygiene, 55,69-75.
Simonsen, P. E., Niemamr, L. & Meyrowitsch, D. W. (1997a). Wuchereria bancrofti in Tanzania: microfilarial periodicity and effect of blood sampling time on microfilarial intensities. Tropical Medicine and International Health, 2, 153-158. Simonsen, P. E., Pedersen, E. M., Lemnge, M. M. & Michael, E. (1997b). Lymphatic tilariasis research and control in Africa. Parasitology Today, 13,413-415. Weil, G. J. & Liftis, F. (1987). Identification and partial characterization of a parasite antigen in sera from humans infected with Wuchereriabancrofki. Journal of Immunology, 138, 3035-3041. Weil, G. J., Jain, D. C., Santhanam, S., Malhotra, A., Kumar, H., Sethumadhavan, K. V. P., Liftis, F. & Ghosh, T. K. (1987). A monoclonal antibody-based enzyme immunoassay for detecting parasite antigenemia in bancroftian filariasis.
Journalof InfectiousDiseases, 156,350-355.
Weil, G. J., Ramzy, R. M. R., Chandrashekar, R., Gad, A. M., Lowrie, R. C., Jr & Faris, R. (1996). Parasite antigenemia without microfilaraemia in bancroftian filariasis. American Journal of Tropical Medicine and Hygiene, 55,333-337.
Weil, G. J., Lammie, P. J. &Weiss, N. (1997). The ICT filariasis test: a rapid-format antigen test for diagnosis of bancroftian filariasis. Parasitology Today, 13,401-404.
Parasitology Today, 13,370-375.
Nicolas, L., Plichart, C., Nguyen, L. N. & Moulia-Pelat, J. P. (1997). Reduction of Wuchereria bancrofti adult worm circulating antigen after annual treatments of diethylcarbamazine combined with ivermectin in French Polynesia. Journal of Infectious Diseases, 17.5,489-492.
1 Announcement
Received 8 December 1998; revised 2 March for publication 2 March 1999
1999; accepted
1 2000 Award in honour of Fred L. Soper (1893-1976) for publications in the field of Inter-American Health
This is an announcement and call for submission of nominations for the 2000 award in honour of Fred L. Soper, former Director of the Pan American Health Organization (the World Health Organization Regional Office for the Americas) from 1947 to 1958. The Award is presented annually to the author or authors of an original scientific contribution comprising new information on, or new insights into, the broad field of public health, with special relevance to Latin America or the Caribbean or both. This may consist of a report, an analysis ofnew data, experimental or observational, or a new approach to analysing available data. Preference is given to studies involving more than one discipline and to papers related to infectious diseases,a life-long concern of Dr Soper. Only papers already published in scientific journals listed in the Index Medicus or in the official journals of the Pan American Health Organization are eligible for consideration. Furthermore, the award is limited to contributions by authors whose principal affiliation is with teaching, research or service institutions located in the countries of Latin America and the Caribbean (including the Centers of the Pan American Health Organization). The Award consists of a certificate and a prize of US$ 1000. Papers submitted by or on behalf of their authors may be considered for the Fred L. Soper Award. For the 2000 Award, only papers published during the calendar year 1999 will be considered; all submissions must be received by 31 March 2000 at the following address: Fred L. Soper Award Program, PAHEF, 525 23rd Street NW, Washington, DC 20037, USA.