Background

In Sub-Saharan Africa malaria is a leading cause of morbidity and mortality, especially in children under five years [1]. Despite intensive campaigns, over 10 Million malaria cases occurred in Tanzania in 2003, about 30% more compared to the previous year [2]. Most of these cases were caused by Plasmodium falciparum. This may be partly due to better surveillance systems, but raising drug resistance is the most likely reason for this tremendous increase.

Chloroquine (CQ) was the antimalarial treatment of choice during the second half of the 20th century. But increasing rates of CQ resistance led Tanzania to change its first line treatment of uncomplicated malaria to sulfadoxine-pyrimethamine (SP) in 2001 [3]. This antifolate combination seemed to be an effective and reasonable alternative, but resistance to SP was rapidly gaining ground, facilitated by the slow elimination from the body. New data show a high level (45%) of SP treatment failures in Muheza, northeast Tanzania [4]. Other effective drugs, where resistance is as yet not a frequent problem, such as atovaquone-proguanil or mefloquine, are of limited value due to their high current costs [5]. A useful alternative is artemisinin-based combination therapy (ACT), e.g., artemether-lumefantrine which is introduced as the new first line drug in Tanzania in mid 2006. At the moment the availability of ACT is limited; while artemisinin monotherapy formulations and SP are commonly used, especially in the private sector. Some drug shops still sell the relegated CQ.

CQ resistance is associated with an amino acid change from lysine to threonine in codon 76 of the P. falciparum chloroquine resistance transporter gene (pfcrt) [6], and a mutation from asparagine to tyrosine in codon 86 of the multidrug resistant gene (pfmdr) [7–10]. Besides there are indices that pfmdr1 N86 is associated with resistance to lumefantrine that is widely used in combination with artemether [11], and also with decreased sensitivity to artemisinins [12].

SP resistance is associated with mutations in the dihydrofolate reductase (pfdhfr) and dihydropteroate synthase (pfdhps) genes. Pyrimethamine is a selective, competitive inhibitor of dihydrofolate reductase and earlier in the folate pathway sulfa drugs inhibit dihydropteroate synthase [13, 14]. Several point mutations are connected to antifolate drug resistance. The quintuple mutation (triple pfdhfr: I51, R59, N108 and double pfdhps: G437, E540) is discussed as a relevant molecular marker of SP treatment failure [15, 16].

Atovaquone-proguanil is a relatively new antimalarial drug that inhibits mitochondrial electron transport. Point mutations in the cytb codon 268 are associated with resistance to this combination [17–19].

Methods

Overall 86 blood samples were collected from patients with clinically diagnosed uncomplicated P. falciparum malaria (fever ≥ 38.0°C, parasitaemia ≥ 2 000/μl) from August to October in 2004 and from June to July in 2005 in Mbeya region of South-western Tanzania. Patients were enrolled at two locations: Matema Health Care Center located in a poor rural area at the shores of Lake Malawi with holoendemic malaria transmission and the Mbeya Referral Hospital, a tertiary hospital at an altitude of 1 700 meters that admits complicated cases from surrounding mesoendemic areas or people that have travelled within Tanzania. Written informed consent was obtained from each patient or the parental guide. The study was reviewed and approved by the local IRB at the Mbeya Referral Hospital and the Tanzanian National Ethics Board of the National Institute for Medical Research, Dar es Salaam.

From each patient a finger prick blood sample was taken for thick and thin blood film and another for filter paper blood sample. Giemsa-stained blood films were examined for malaria parasites (per 200 white blood cells) and densities were assessed based on a assumed mean WBC count of 8 000/μl. DNA extraction from filter paper bloodspots was done using Chelex® (Bio-Rad, Germany) as described elsewhere [20] Nested PCR assays were used to verify the parasite species [21]. The DNA was amplified by nested PCR and digested by the RFLP-method to detect the mutations of P. falciparum pfcrt 76, pfmdr1 86, dhfr 16, 51, 59, 108, 164, dhps 436, 437, 540, 581, 613 and cytb 268 [17, 22, 23]. Mixed alleles (wild type and mutant) were assessed as mutant. The median age of the 86 patients with P. falciparum infection (45 female, 41 male) was 21 years (range 8 month to 55 years). All 86 isolates showed P. falciparum mono-infection, there was no P. vivax, P. ovale, or P. malariae infection. The geometric mean parasite density was 29 992/μl (range 3 320/μl to 127 440/μl).

Results

Table 1 displays the prevalence of pfcrt, pfmdr1, pfdhfr, pfdhps and cytb alleles in Mbeya & Matema. Pfcrt T76 mutation is expressed by 58.8% in Mbeya and 47.8% in Matema. 70.6% of the isolates from Mbeya and 47.1% from Matema showed the pfmdr1 Y86 mutation. Almost 100% of both settings exhibited the pfdhfr N108 mutation. Likewise, nearly all samples showed the I51 mutation. All but one displayed the R59 mutation in Mbeya. Amino acid changes from alanine to glycine at codon 437 and from lysine to glutamine at codon 540 of the pfdhps gene were detected in 81% and 86.9%, respectively. Only one specimen in Matema displayed the G581 variant. No pfdhfr L164 mutation was seen. T108 and V16 variants which are linked with cycloguanil resistance were not present as well. In line with previous reports no evidence for cytb codon 268 mutations were found in south-western Tanzania [24].

Table 1 Prevalence of mutations conferring resistance to chloroquine, sulfadoxine-pyrimethamine and atovaquone-proguanil in Plasmodium falciparum isolates from Mbeya & Matema, southern Tanzania
Full size table

Discussion

More relevant for predicting treatment failure or emerging resistance are combinations of the point mutations described above (Table 2). In Matema, the rural setting, the Pfdhfr quintuple mutation was more common than in the Mbeya Referral Hospital (67.2% vs 52.9%), while the Pfdhfr triple mutation, suggested to be an early molecular marker for SP resistance in Tanzania [25], was more frequent in Mbeya (79.7% vs 94.1%). Although the predictive value of these markers for SP treatment failure has not been established in the study regions, these results are in line with the high level of treatment failure (42.3%) in a multi-site survey in Tanzania [26].

Table 2 Prevalence of pfcrt, pfmdr 1, dhfr and dhps genotype combinations conferring chloroquine and sulfadoxine-pyrimethamine resistance in Mbeya & Matema, together and each seperated. The risk ratio is the prevalence ratio of the combinations between Mbeya and Matema.
Full size table

Even so this study investigated a limited number of patients and therefore differences may not reach significance, differences in mutation rates might reflect differences of access to health care between the two locations. While individuals in the rural area receive their malaria treatment, currently SP, almost exclusively through the Matema Health Care Center, patients from Mbeya can choose upon a wide variety of health care facilities and antimalarial drugs. Ongoing usage of CQ may be assumed there.

Although considerably higher resistance rates before 2001 are probable, the still relatively high rate of pfcrt and pfmdr1 mutations is contrary to other reports that have demonstrated a complete regression of pfcrt mutations several years after leaving CQ as first – line antimalarial drug [27, 28].

In vivo selection of pfmdr1 86N allele by artemether-lumefantrine has been found in Tanzania [11] and pfmdr copy numbers seem to influence susceptibility to lumefantrine and artemisinin [29]. There is no prediction possible due to our results but continuing surveillance would be interesting concerning pfmdr1 polymorphisms.

Conclusion

This study confirms the high prevalence of point mutations in the pfdhfr and pfdhps genes in Tanzania which are associated with SP treatment failure [25, 26]. The rate of quintuple pfdhfr/pfdhps mutations in the Mbeya region, south-western Tanzania, is in the upper range of frequencies reported in East Africa. Data from Malawi, Kenya, Tanzania and Ethiopia range from 10 to 78% [25, 30–32]. The absence of cytb codon 268 supports atovaquone-proguanil as a possible second- or third-line drug for treatment of uncomplicated malaria.

The study data might be used as a basis for surveillance of resistance markers after introduction of ACT and might later indicate the possibility for reintroduction of one of the other drugs.

Conflict of interest

The authors do not have a commercial or other association that might pose a conflict of interest.