BACKGROUND: The efficacy of anti-malarial drugs is assessed over a period of 28-63 days (depending on the drugs' residence time) following initiation of treatment in order to capture late failures. However, prolonged follow-up increases the likelihood of new infections depending on transmission intensity. Therefore, molecular genotyping of highly polymorphic regions of Plasmodium falciparum msp1, msp2 and glurp loci is usually carried out to distinguish recrudescence (true failures) from new infections. This tool has now been adopted as an integral part of anti-malarial efficacy studies and clinical trials. However, there are concerns over its utility and reliability because conclusions drawn from molecular typing depend on the genetic profile of the respective parasite populations, but this profile is not systematically documented in most endemic areas. This study presents the genetic diversity of P. falciparum msp1, msp2 and glurp markers in selected sub-Saharan Africa countries with varying levels of endemicity namely Malawi, Tanzania, Uganda, Burkina Faso and São Tomé. METHODS: A total 780 baseline (Day 0) blood samples from children less than seven years, recruited in a randomized controlled clinical trials done between 1996 and 2000 were genotyped. DNA was extracted; allelic frequency and diversity were investigated by PCR followed by capillary electrophoresis for msp2 and fragment sizing by a digitalized gel imager for msp1 and glurp. RESULTS AND CONCLUSION: Plasmodium falciparum msp1, msp2 and glurp markers were highly polymorphic with low allele frequencies. A total of 17 msp1 genotypes [eight MAD20-, one RO33- and eight K1-types]; 116 msp2 genotypes [83 3D7 and 33 FC27- types] and 14 glurp genotypes were recorded. All five sites recorded very high expected heterozygosity (HE) values (0.68 - 0.99). HE was highest in msp2 locus (HE=0.99), and lowest for msp1 (HE=0.68) (P<0.0001). The genetic diversity and allelic frequency recorded were independent of transmission intensity (P=0.84, P=0.25 respectively. A few genotypes had particularly high frequencies; however the most abundant showed only a 4% probability that a new infection would share the same genotype as the baseline infection. This is unlikely to confound the distinction of recrudescence from new infection, particularly if more than one marker is used for genotyping. Hence, this study supports the use of msp1, msp2 and glurp in malaria clinical trials in sub-Saharan Africa to discriminate new from recrudescent infections.