Screening for HTLV-I requires a venous blood sampling, which can have certain drawbacks during large-scale epidemiologic surveys. Our aim was to assess the ability to detect antibodies to HTLV-I from blood collected on filter paper, a more convenient method that is already in use and has been validated in screening for the human immunodeficieney virus (HIV) . A previous study was conducted by our team in Burundi and showed that the agreement between the two sampling methods to detect the seropositivity for HTLV-I was poor . In order to assess this method on a higher number of patients, the present study was conducted in Benin in March 1995 where a cohort of HTLV-I positive subjects is surveyed since 1991 .
The study group was comprised of 109 subjects. Paired specimens were obtained: 10 ml of blood by venipuncture, and capillary blood obtained by lancet finger puncture which provided three blood spots, each I cm indiameter on chromatography paper (Whatman 311, Polylabo). The filter paper was let to dry in ambient air and then kept in a refrigerator at 40 Celsius in an air-tight container with desiccant. The biological analysis were performed within three weeks in Limoges (France) in the same laboratory without patient identification. Serologic testing for HTLV-I was performed using the same diagnostic kits for the serum and filter paper (screening with ELISA technique-HTLV-1/11, Abbott Laboratories, Chicago, USA-and confirmation of the samples considered positive using a Western Blot assay - Bioptim 2.3, Diagnostic Biotechnology, Singapore-). A 5mm diammeter punch from a filter paper spot of capillary blood was obtained and eluted in 200 microliters of the diluent supplied with the kit over a 12 hours period preceding the serologic testing. The tests were performed on the eluate thus obtained. Criteria for HTLV-I seropositivity were lhe recognition of two different proteins: gag (pl9 and/or p24) and env (gp46 and/or rgp21) .
The sensitivity of the method was 81% (Table 1). The specificity was 100%.
HTLV-I: Human T cell Leukemia Virus type 1; WB: Western Blot; ELISA: Enzyme Linked ImmunoSorbentAssay.
Concordance between serologic testing for HTLV-1 by venipuncture as compared with collection on filter paper was poor. The latter technique has the disadvantage of global decrease in sensitivity. These results do not confirm those by Jeannel et al.  who found complete concordance between the two serologic testing methods for HTLV during a study conducted in Zaire in 122 subjects. However, details on the sampling method used by these authors and the number of seropositive subjects in their study were not given.
Compared to our previous study in Burundi  carried out with a smaller sample of positive subjects, the sensitivity of filter paper has increased from 66.7% to 81.0%. Difference in the methods of conversation of the samples could explain this improvement (shorter delay to perform biological analysis; filter papers kept in a refrigerator). However, this sensitivity is still too low for screening purposes. To evaluate if structural or functional differences between diagnostic kits may explain this lack of sensitivity, other kits must be tested. It may be also be useful to assess antibodies stability under various conditions, particularly climatic.
Table 1: Serologic testing for HTLV-I using filter paper and serum after confirmation by WB in subjects that tested positive by ELISA in 109 subjects, Benin. March 1995.