TlUNSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (2001) 95,33-35 The fibre-web blood sampling technique applied to serological diagnosis of schistosomiasis mansoni Lars-Ake Nilsson’, Govert J. van Damz, Andri M. Deeldeti, Bo Eriksson3, Reverianus M. Gabone4* and Susanne Schmeisser’ ’ Depanment of Medical Microbiology and Immunology, University ojG&eborg, Giiteborg, Sweden; 2Depament of Parasitology, Leiden University Medical Centre, Leiden, The Netherlands; 3The Nordic School of Public Health, Giiteborg, Sweden; 4National Institute for Medical Research, Mwanza, Tanzania Abstract The fibre-web technique for sampling, storing and transport ofvenous or capillary blood has been evaluated, in 84 schoolchildren from the Mwanza region of Tanzania, with regard to diagnostic efficacy for determination of the schistosome circulating anodic antigen (CAA) under conditions similar to those prevailing in the field. Although the average concentrations determined in fibre-web eluates were only about half of those determined in serum, the prevalences of CAA-positive individuals for the 2 sample materials were approximately the same. The average coefficient of variation calculated on determination of CAA in venous-blood fibre-web eluates amounted to 7%. The study shows that the fibre-web technique is well suited for use under field conditions. Keywords: schisrosomiasis, Schistosoma munsoni, serodiagnosis, blood sampling, methodology, ELLSA, circulating anodic antigen, Tanzania Introduction Demonstration of the circulating anodic antigen (CAA) in blood, released during active infections with the trematode Schistoroma mansoni (DEELDER et al., 1980), has during recent years proven to be a valuable tool complementary 10 the demonstration of eggs ex- creted in the stool, e.g., by the Kato-Katz technique (DE JONGE et al., 1988; H&ANGhD et al., 1996). In an attempt to use conventional filter paper to collect blood for demonstration of CAA, it became evident that CAA adsorbed strongly to the cellulose paper and therefore could not be satisfactorily eluted and subsequently detected by ELISA. In order to solve this problem we tested several synthetic-fibre materials and recently described a polypropylene fibre-web material which was desiened for collection of blood (TAMALY et al.. 1997). Incontrast to conventionally us&Y cellulose fiite; paper this fibre-web material allowed satisfactory elution of both CAA and CCA (circulating cathodic antigen) as well as specific schistosome antibodies. In the present study we have evaluated the fibre-web technique for qualitative and quantitative determination of CAA when used under field conditions. To this end. we compared the CAA concentrations in serum and in eluates of fibre-web material on to which capillary and venous blood samples were dried, respectively. Materials and Methods Blood and stool specimens were obtained from 84 schoolchildren, aged 7-18 years, living in the village Bwiru, Mwanza region, Tanzania, an area endemic for S. manroni with high prevalence. After sampling of venous blood, 100-JJL aliauots were immediatelv transferred to the Ulglass’shapeh receptacles of the fi”bre-web sheet. Triplicate venous-blood fibre-web samples were pre- pared from each of 44 children to determine the variation in the procedure. Capillary blood was also sampled from 60 of the children and similarly transferred to the fibre- web material. After drying of the blood, the fibre-web sheets were stored at -20°C until analysed. Serum was prepared from the remaining part of the venous-blood samples and stored at -20°C until transported to the laboratory in Giiteborg for analysis of CAA. *Deceased on 8 August 1998. Author for correspondence: Dr Lars-he Nilsson, Department of Medical Microbiology and Immunology, Box 435, S-405 30 Gheborg, Sweden; fax +46 3 1 82 67 9 1, e-mail lars-ake.nilsson@@crobio.gu.se Determination of faecal egg excretion Duplicate stool samples (42-mg samples) were exam- ined by the Kato-Katz technique as previously described (I-IAKANG~RD et aZ., 1996) and egg excretion was ex- pressed as the mean of the 2 counts. Determination ojCAA The circular receptacles of the fibre-web sheet con- taining dried blood were cut off the plastic support and immersed in 500 &L PBS-Tween (phosphate-buffered saline containing 0.05% Tween 20). CAA was quantita- tively determined in serum and in capillary- and venous- blood eluates by ELISA as previously described (I-lAKANGARD et aZ., 1996). The lower detection level of the ELISA, as carried out in this study, was estimated to be about 0.7 ng CAA/mL. Statistical procedures Because of the positively skewed distribution of data, logarithmic transformations were performed before cor- relation and regression analysis of the relationship be- tween CAA concentrations and egg counts. Individuals who were negative in both the CAA and the Kato-Katz tests were excluded in the regression analysis since they were regarded as non-infected, whereas individuals who were positive in either one or both of the tests were included, assuming that CA&positive but egg-negative individuals (and vice versa) had light infections. Egg counts were analysed as log,,(epg+l) of the means of duplicate samples, as is conventional. CAA concentra- tions were also log,,-transformed and a value of half of the detection limit was assigned before the log-transfor- mation to CAA-levels below the detection level. The latter procedure was applied since, as pointed out by VAN LIESHOLJT er al. (1995), exclusion of the latter type of patients could lead to bias because of the ditYerence in sensitivity between the 2 methods used. For calculation of the average error of the method the logarithmically transformed CAA values were used since it was not shown that the standard deviations for these values were dependent on the CAA concentration over a range of l-50 ng/mL. Calculations performed on triplicate de- terminations of CAA in venous-blood fibre-web eluates from 44 individuals showed that the methodological error within samples, expressed as the average coefficient of variation, was 7%. The relationship between CAA concentrations ob- tained from the various sampling materials was expressed in percent as the slopes of the regression lines. This procedure was found to be acceptable since the intercept with the y-axis was close to the origin. Furthermore, when the variables on the X- and y-axes were switched,