The results showed that P. vivax parasites from Kolkata demonstrated an extremely high prevalence of VK210 type in the pvcs gene (with only one VK247 type), high polymorphism in both merozoite surface proteins (pvmsp1 and pvmsp3 alpha), and low rates (10.6%) of multiple genotype infection.
The predominance in Kolkata isolates of VK210 pvcs gene type (99.3%) has not been seen elsewhere to the same extent. Recent studies in Thailand have found rates of 70.5% , 78%  and 90% , though these results were in sharp contrast to those of earlier studies in which VK247 was found to be the predominant type; 83% in Thai and 90% in Papua New Guinea samples by PCR/Oligo Probe . Another study of Indian strains has shown the VK210 type to be predominant, but also demonstrated (using a different technique) significant numbers of VK247 type . This phenomenon may be attributed to selection by host immune pressure on a particular genotype, and/or the preferential production of sporozoites carrying a specific variant , such as VK210 for pvcs in a mosquito species. These differences may also be due to sampling biases or regional temporal fluctuations of individual genotypes frequencies. The frequency of absence of the pre-repeat insertion of VK210 in the pvcs gene was 4.7% (7/150) and that of the post-repeat region of the gene 56% (84/150) in Kolkata. In the absence of a multicentre study involving various regions of India it is difficult to define the extent to which the observed difference was in any way related to the functional aspects of the Kolkata strain.
In this study, two major types of the pvmsp1 marker containing 35 alleles were found. A similar degree of diversity (36 allelic types) was also found in the study from Thailand using the same laboratory protocol , and with a different protocol in Papua New Guinea and Indian strains [16,42]. For pvmsp3-alpha, the present study reveals 37 alleles with three size variants cut by two restriction enzymes. The frequencies of the three pvmsp3-alpha types were consistent with those found in Papua New Guinea and Thailand [18,19]. A recent study using a different protocol revealed 16 size and sequence polymorphic allele in Indian strains . This high degree of polymorphism in the merozoite surface protein (pvmsp1 and pvmsp3-alpha) is similar to that found in other studies from other regions.
All three markers show marked genetic diversity in Kolkata strains, with 11 alleles for pvcs, 35 alleles for the second fragment of pvmsp1, and 37 alleles for pvmsp3-alpha. This compares with another study of Indian vivax isolates which revealed 2 sequence variants of pvcs, 9 size variants in pvgam1, and 16 size variants with sequence polymorphism in pvmsp3-alpha gene .
The frequency of multiple genotype infections of P. vivax malaria has been estimated in many regions but it is difficult to compare these values due to differences in sampling and genotyping methods. The proportion of mixed gene infections estimated in Papua New Guinea, India, and Thailand ranges from 30% to 65% [16,20,24,25]. The present study with a relatively large sampling size, shows an overall 10.6% for multiple genotypes (1.3% for pvcs, 0.7% for pvmsp1 and 8.6% for pvmsp3 alpha). This rate may reflect a limitation in the sensitivity of PCR for the detection of multiple genotype infections, despite the high degree of polymorphism seen.
According to the Calcutta School of Tropical Medicine Report 2005, the malaria clinic treats around 6,000 malaria positive cases annually, of which an average of 65% are caused by P. vivax. The rate of mixed infection with both P. vivax and P. falciparum during the last few years has declined (1.0% in 1997, 0.7% in 1998 and 0.1% in 2001) . The samples from Kolkata were collected from an urban area where vivax malaria remains chloroquine-sensitive, but where recurrences are frequent. Only a few cases were contracted outside of central Kolkata. The high degree of polymorphism and low level of multiple genotype infection probably reflects the nature of this endemic setting. More study is needed to assess whether these discrepancies reflect true differences between disease populations, or are due to differences in sample sizes or the laboratory methodology [20,38].