Blood spots collected onto filter paper are an attractive alternative to collection of plasma or serum for serological studies, particularly in remote field settings with limited laboratory facilities. Blood spots are an even more attractive option in the context of malaria studies where they are frequently already collected either for malaria diagnosis (as rapid diagnostic tests) or as a source of DNA for PCR-based detection and genotyping of parasites. However, whereas the ability to discern simply the presence or absence of parasites is frequently sufficient for DNA-based assays, antibody measurements tend to be of more value if information about the quantity of the antibodies present can be reliably obtained. This requires implementation of carefully validated protocols, for collection and storage of spots and elution of antibodies, designed to maximize the recovery of intact immunoglobulins.
The utility of the procedures presented here rests on the assumptions that the blood spot is evenly impregnated with blood, so that a disc of a standard diameter represents a defined volume of blood, and the adsorbed proteins can subsequently be eluted with high efficiency. A number of alternative substrates have been used for collecting blood spots and we have not been able to evaluate all of these here. Some of these substrates have been optimized for the recovery of DNA, but it is not clear that treatments to reduce nuclease activity and improve nucleic acid recovery are as beneficial for protein recovery. Of those that have been evaluated here, glass fibre mats and Whatman No 3 MM had high adsorption capacities while 3 MM and Whatman no 1 were robust for handling in a field environment. Glass fibre paper, however, was fragile, especially when punching discs, had inferior wetting characteristics and the efficiency of antibody elution was non-linear. Although antibody recoveries from No 1 paper were not systematically compared with those from 3 MM, limited experience suggests that recoveries are similar when the different adsorption capacities are allowed for. On balance, Whatmann 3 MM paper appears to offer the ideal combination of strength, wettability and adsorption. Moreover, comparison of amounts of antibodies recovered from freshly-prepared blood spots on 3 MM showed good correspondence with levels determined in the equivalent plasma.
The utility of this protocol also depends on establishing storage conditions under which immunoglobulins are stable in dried form, and can be quantitatively recovered from the paper at a later date. To estimate the stability of intrinsically stable preparations it is necessary to carry out accelerated stability testing at elevated temperatures, ideally under conditions where a substantial fraction of the activity has decayed during the period of measurement [30–32]. Stability testing is an integral part of pharmaceutical development [41, 42], although most regimes are intended to address a relatively modest shelf-life. The choice of a first-order kinetic model is an empiric one, but a choice that appears compatible with experience of accelerated stability testing of lyophilized biological reference materials . In the current case, it appears justified as a satisfactory approximation of the kinetic behaviour observed, though using higher temperatures than previously reported studies  in order to obtain adequate losses of activity. Encouragingly, this analysis demonstrates that, when the results are extrapolated below 23°C, it should be possible to obtain good quality antibody from blood spots kept for several years at -20°C and, more importantly, from blood spots stored for a few days at ambient temperature, and a few weeks at 4–10°C before either assay or transfer to -20°C.
The other major factor affecting the stability of antibodies in blood spots is the humidity at which they are stored. This was not examined in detail, partly because of the added experimental complexity of manipulating RH, but, more practically, because of the impossibility of monitoring and integrating changing levels of RH in field use. However, storage of spots for as little as two weeks at ambient humidity reduced the recovery of intact immunoglobulins by more than 50%, a far greater decrease than would have been expected if the spots had been scrupulously desiccated. Given the importance of keeping blood spots dry, the simple visual check provided by self-indicating silica gel appears to be the simplest if not the only practical method of monitoring storage humidity. Silica gel containing CoCl2 as indicator remains classified as safe for use in desiccation despite the recent classification of CoCl2 as being "possibly carcinogenic to humans" by the International Agency for Research on Cancer  and its classification as a carcinogen by the European Community. However, proprietary alternatives have recently been introduced (eg Sigma, Poole, UK #136767; Geejay, Sandy, UK; Silgel, Telford, UK) which appear to change colour at RH <20%, and may therefore be equally suitable.
Although the rates of loss of antibodies to MSP-119 and MSP-2 in the accelerated degradation study were similar, recoveries from field samples suggested that antibodies to MSP-119 (predominantly IgG1) appeared to be more stable than antibodies to MSP-2 (predominantly IgG3). IgG3 antibodies have a greatly extended and unstructured hinge region [44, 45] and it is possible that IgG3 antibodies are more susceptible to degradation by endogenous proteases when RH is not rigidly controlled. Given these observations, prudence suggests that the results of antibodies recovered from blood spots should not be combined with results from serum or plasma samples.
One unexpected advantage of this method of sample collection is that the quality of blood spots of questionable provenance (in terms of their storage conditions) can be judged visually by examining the colour of the filter papers discs after overnight incubation in elution buffer. Discs that have been properly stored, and from which there is a very high recovery rate of antibodies, turn from deep red to pale pink or white after incubation, indicating that the adsorbed proteins (including haemoglobin) have been eluted. By comparison, discs stored at high temperatures remain dark red, indicating that the indicator protein haemoglobin can no longer be readily eluted from the filter paper. When blood spots from different malaria field studies, which had been collected for diagnosis or DNA recovery and had been stored for extended periods at ambient temperature and humidity were examined, the eluates of spots which gave poor antibody recovery were uniformly pale in colour. Thus, whilst visual inspection may not be a fool-proof means of assessing blood spot quality it will allow seriously degraded samples to be identified and discarded prior to ELISA analysis.
Finally, the utility of blood spot collection of serum antibodies has been demonstrated by showing that it is possible to derive serological estimates of malaria endemicity from blood spot samples and that these serological estimates agree well with entomological data. Close agreement was observed between the serological and entomological data for Lower Moshi and for Rukungiri; agreement was less good for Kabale which may be due to the longer gap between the two estimates (the only estimate of EIR available was from 1996/7) or the inherent inexactitude of entomological assessments at low transmission [46, 47].