Am. J. Trop. Med. Hyg., 67(4), 2002, pp. 383–387 Copyright © 2002 by The American Society of Tropical Medicine and Hygiene
MOLECULAR EPIDEMIOLOGY OF MALARIA IN CAMEROON. XII. IN VITRO DRUG ASSAYS AND MOLECULAR SURVEILLANCE OF CHLOROQUINE AND PROGUANIL RESISTANCE LEONARDO K. BASCO Unité de Recherche “Paludologie Afro-Tropicale,” Institut de Recherche pour le Développement (IRD) and Laboratoire de Recherche sur le Paludisme, Organisation de Coordination pour la lutte contre les Endémies en Afrique Centrale (OCEAC), B. P. 288, Yaoundé, Cameroon
Abstract. Chloroquine-proguanil combination is one of the options for chemoprophylaxis. The rapid evolution of drug resistance status requires a constant upgrade of epidemiologic data. Due to various difficulties in conducting prospective clinical studies on the prophylactic efficacy of the drug combination, especially in highly chloroquineresistant zones, in vitro drug sensitivity assays and specific molecular markers for chloroquine (Plasmodium falciparum chloroquine-resistance transporter, pfcrt) and cycloguanil (a biologically active metabolite of proguanil; dihydrofolate reductase, dhfr) resistance were evaluated as an alternative approach in this study. Of 116 isolates, 62 (53.4%) were doubly resistant in vitro to chloroquine (IC50 ⱖ 100 nM) and cycloguanil (IC50 ⱖ 15 nM). Likewise, 62 of 118 isolates (52.5%) carried both the mutant Thr-76 pfcrt allele and at least one dhfr mutant allele (1 with a single Asn-108 allele, 8 with double Arg-59 and Asn-108 mutations, and 53 with triple Ile-51, Arg-59, and Asn-108 mutations). The in vitro drug response corresponded with the presence or absence of key mutation(s) in the pfcrt and dhfr genes. These results suggest the high proportion of P. falciparum isolates in southern Cameroon that may not respond to chloroquineproguanil combination. resistance genes. Previous studies have established the high correlation between the in vitro response to chloroquine and cycloguanil (biologically active metabolite of proguanil) and P. falciparum chloroquine-resistance transporter (pfcrt) and dihydrofolate reductase (dhfr) gene mutations, respectively, both in reference clones and field isolates.8–13 Likewise, the key pfcrt mutation at codon 76 has been associated with clinical resistance to chloroquine.14 The correlation between dhfr mutations and clinical resistance to proguanil has not been established. In the present study, the efficacy of chloroquine-proguanil combination was indirectly evaluated by an alternative approach based on in vitro drug sensitivity assays and molecular markers with the aim to describe the current trend of drug efficacy in southern Cameroon.
INTRODUCTION One of the strategies for malaria control involves the use of chemoprophylaxis for individuals exposed to high risk, including nonimmune travelers from nonendemic regions and pregnant women residing in endemic regions. Drugs recommended for antimalarial prophylaxis depend on several factors, such as the degree of risk, level of chloroquine resistance, tolerance, physiologic conditions, and age. In subSaharan Africa, the World Health Organization currently recommends mefloquine as the first drug of choice, followed by doxycycline (second choice), and chloroquine-proguanil combination (third choice) for nonpregnant, nonimmune temporary visitors.1 The use of chloroquine alone for chemoprophylaxis of nonimmune subjects traveling to sub-Saharan Africa has been seriously compromised in the late 1980s and in the 1990s due to the rapid spread of chloroquine-resistant Plasmodium falciparum. At about the same period, the use of other alternative drugs for prophylaxis, such as sulfadoxine-pyrimethamine and amodiaquine, had been discontinued because of rare but potentially fatal adverse effects.2 In this context, chloroquine-proguanil combination appeared to be a safe, well tolerated, and effective option and was officially recommended for prophylaxis of temporary visitors, as well as for pregnant women residing in endemic areas, in the 1990s. Most studies on the efficacy of the chloroquine-proguanil combination have been based on retrospective surveys or case reports of malaria infections in nonimmune subjects under prophylaxis.3–5 Few prospective studies have been undertaken, which have generally shown a relatively good protection.6,7 Because prospective clinical studies on the current prophylactic efficacy of chloroquine-proguanil are time consuming, costly, and possibly unjustified in some highly chloroquine-resistant areas, alternative methods may be of value to monitor and assess the utility of this drug combination in sub-Saharan Africa. These methods include in vitro drug sensitivity assays and molecular analysis of drug-
MATERIALS AND METHODS Patients. After informed consent, venous blood samples were obtained in 2000–2001 from symptomatic children aged > 12 years old and adults consulting at the Nlongkak Catholic missionary dispensary if the following criteria were met: the presence of P. falciparum, without other Plasmodium species, at a parasitemia ⱖ 0.1%, signs and symptoms of acute uncomplicated malaria, and denial of recent self-medication with an antimalarial drug, confirmed by negative SakerSolomon’s urine test.15 Pregnant women and patients with signs and symptoms of severe and complicated malaria were excluded from the study. All enrolled patients were residents of Yaoundé. The patients were treated with oral amodiaquine and followed by the dispensary staff. This study was approved by the Cameroonian National Ethics Committee and Cameroonian Ministry of Public Health. In vitro drug sensitivity assay. Chloroquine sulfate and cycloguanil base were provided by Aventis (Antony, France) and AstraZeneca (Rueil-Malmaison, France), respectively. Stock solutions and working solutions were prepared in sterile distilled water. The final concentrations ranged from 25–
383
384
BASCO
1,600 nM for chloroquine (two-fold dilutions in triplicate) and 0.0488–51,200 nM (four-fold dilutions in duplicate) for cycloguanil. Blood samples were washed with para-aminobenzoic acid (PABA) and folic acid-free RPMI 1640 culture medium (designated NPNF RPMI medium hereafter) three times by centrifugation within 2 to 3 hr after blood extraction. Infected erythrocytes were resuspended in NPNF RPMI medium and standard RPMI medium containing 1 mg/L PABA and 1 mg/L folic acid (both media contained 25 mM HEPES and 25 mM NaHCO3 buffers) to determine the in vitro activity of cycloguanil and chloroquine, respectively. Culture media were supplemented with 10% human serum. Technical procedures of the isotopic microtest and determination of 50% inhibitory concentrations (IC50), defined as the concentration at which 50% of the parasite growth is inhibited as compared with drug-free control wells, were described in detail previously.11 Sequence analysis. Genomic DNA was extracted from 2 mL of red cell pellet, as described previously.11 The region spanning the key pfcrt codon was amplified by polymerase chain reaction, followed by restriction fragment length polymorphism using Apo I endonuclease.13,14 The entire DHFR domain was amplified by polymerase chain reaction, followed by purification of the amplified product, fluorescent labeling, and automatic sequencing (ABI Systems, Perkin Elmer, Les Ulis, France). These procedures were detailed in our previous study.11,16
Data interpretation. The threshold IC50 value for chloroquine resistance was fixed at ⱖ 100 nM.17,18 This cut-off value is based on the determination of in vitro chloroquine response of African field isolates obtained from malaria-infected, nonimmune individuals under chloroquine prophylaxis and semiimmune patients failing to respond to chloroquine treatment. The threshold IC50 for cycloguanil resistance is still undetermined. Although previous studies have arbitrarily fixed the threshold at ⱖ 50 nM, more recent studies using the isotopic microtest have suggested that it may lie between 10 and 50 nM.11 Since a reliable molecular marker for cycloguanil resistance exists, an inverse approach was applied to determine the approximate threshold value based on the distribution of cycloguanil IC50 values in respect to the presence of dhfr mutations. Thus, an IC50 value that distinguishes between parasites with wild-type and mutant dhfr was used as the threshold value for in vitro cycloguanil resistance. The in vitro response to chloroquine and cycloguanil was expressed as the geometric mean IC50 and 95% confidence interval. Wild-type pfcrt and dhfr were defined as DNA sequences that characterize reference clones of P. falciparum.8,9,12 Mutant pfcrt denotes the Lys-to-Thr amino acid substitution at position 76. Mutant dhfr denotes the presence of at least one amino acid substitution at position 108 (Ser-to-Asn), with possibly two additional substitutions, Asn51Ile and Cys59Arg. Other dhfr mutations at positions 16, 50, 108 (Serto-Thr), and 164 have not been observed in Cameroonian isolates.
FIGURE 1. Distribution of chloroquine IC50 values in relation to the key pfcrt codon (K, Lys; T, Thr) at position 76. K76 and T76 are wild-type and mutant alleles, respectively.12 The hypothetical threshold value (100 nM) for in vitro chloroquine resistance is indicated by the dotted line.
385
CHLOROQUINE-PROGUANIL RESISTANCE
RESULTS A total of 118 samples were analyzed for in vitro response to chloroquine and cycloguanil, and the corresponding DNA sequences of pfcrt and dhfr were determined. Two drug assays for chloroquine were uninterpretable due to bacterial contamination of the RPMI 1640 medium. The in vitro response of these 2 isolates to cycloguanil was available because a different culture medium is used to determine the IC50 for cycloguanil. Based on the threshold IC50 of ⱖ 100 nM, 49 isolates (42%) were chloroquine-sensitive (geometric mean, 34.5 nM; 95% confidence interval, 30.5–38.9 nM; range, 18.6– 90.0 nM), and 67 (58%) were chloroquine-resistant (geometric mean, 209 nM; 95% confidence interval, 189–231 nM; range, 102–714 nM). Of 118 isolates, 36 (31%) carried the wild-type Lys-76 pfcrt allele, and 68 (58%) carried the mutant Thr-76 pfcrt allele. Fourteen isolates (12%) had mixed Lys + Thr-76 alleles. The threshold value of 100 nM distinguished between isolates with wild-type and those with mutant pfcrt in a large majority of isolates, as illustrated in Figure 1. The distribution of cycloguanil IC50 values in relation to the dhfr profiles is presented in Figure 2. Ten isolates had wildtype dhfr with no mutation. Of the remaining isolates, 94 were dhfr mutants with at least Ser108Asn substitution (two possible additional amino acid substitutions at positions 51 and/ or 59), and 14 had mixed dhfr alleles. No new dhfr mutation was identified among the isolates. The highest cycloguanil IC50 among wild-type isolates and the lowest IC50 value
among mutants, excluding isolates with mixed dhfr alleles, were 9.14 nM and 15.3 nM, respectively. This observation led us to suppose that 15 nM may be the threshold value for in vitro cycloguanil resistance in our laboratory. Using this criterion, 18 isolates (10 wild-type and 8 with mixed dhfr alleles) were cycloguanil-sensitive (geometric mean, 1.30 nM; 95% CI, 0.752–2.26 nM; range, 0.249–9.14 nM), and 100 (94 with at least one mutant dhfr allele + 6 with mixed dhfr alleles) were cycloguanil resistant (geometric mean, 77.1 nM; 95% CI, 65.0–91.6 nM; range, 15.3–901 nM). Overall, 54 isolates (47%) were sensitive in vitro to at least one of the compounds (Table 1). The analysis of isolates with different combinations of pfcrt and dhfr alleles showed that 32 isolates (27%) were of wild-type in either, or both, of the two resistance genes and 62 isolates (53%) were double mutants (Table 2).
DISCUSSION The present study indicates that various technical and operational difficulties involved in conducting prospective clinical studies on the prophylactic efficacy of chloroquineproguanil combination may be partially overcome by resorting to an alternative method to assess the current trend of chloroquine and cycloguanil resistance in an endemic region. Although we must bear in mind that no in vitro or molecular methods can entirely substitute for in vivo studies at present, and that these results cannot be directly extrapolated to pre-
FIGURE 2. Distribution of cycloguanil IC50 values in relation to the presence of mutant dhfr alleles (N, Asn; R, Arg; I, Ile). The hypothetical threshold value (15 nM) for in vitro cycloguanil resistance is indicated by the dotted line.
386
BASCO
dict in vivo response, especially on an individual basis, pragmatic considerations lead us to explore indirect methods that may yield relevant information on the epidemiology of drugresistant malaria. Some of the host factors that are not taken into account by in vitro methods and therefore preclude an accurate prediction of in vivo protective efficacy of the drug combination include wide interindividual variation in pharmacokinetics, pharmacogenetic regulation of hepatic biotransformation of proguanil into its biologically active metabolite, cycloguanil, resulting in a significantly low plasma cycloguanil level in poor metabolizers (proguanil itself has a low blood schizontocidal activity), and tissue schizontocidal action of cycloguanil.19–21 In spite of these inherent limitations of in vitro approaches, the data presented in this study provide useful information on the drug-resistant status of malaria parasites. The isotopic microtest assay can accurately distinguish between sensitive and resistant isolates. The in vitro response to chloroquine can predict the presence of mutant Thr-76 pfcrt allele on the basis of the threshold value of 100 nM, and vice versa, in a large majority of isolates. A major confounding factor in testing field isolates by in vitro drug assay seemed to be the presence of multiple parasite populations within an isolate. The in vitro chloroquine response of mixed populations with both wild-type and mutant pfcrt cannot be predicted accurately, possibly due to the varying extent of predominance of each population in a given isolate and biologic fitness of each population to multiply under in vitro conditions. In the present study, 12% of the samples had detectable mixed populations with Lys- and Thr-76 pfcrt alleles. Using a more polymorphic genetic marker that allows a detailed analysis of the parasite population structure, we have observed a much higher level of mixed P. falciparum infections (57%) among symptomatic patients in Yaoundé.22 Although the presence of mixed parasite populations may lower the correlation between in vitro drug response and pfcrt, the latter may serve as a useful molecular marker within the context of an epidemiologic study, in which minor deviations are corrected due to a large sample size. Comparative studies between clinical response to sulfadoxine-pyrimethamine and dhfr mutations have suggested that a single Ser108Asn mutation does not lead to therapeutic failure and that three mutations, Ile-51, Arg-59, and Asn-108, may be required for clinical resistance to sulfadoxinepyrimethamine.23–25 Whether this observation is due to the synergistic action of the drug combination, the secondary role played by mutations in the molecular target of sulfadoxine, dihydropteroate synthase, or modification in the transport process of sulfadoxine into the parasites is still being debated. It is not known at present how many dhfr mutations are re-
TABLE 2 Response to chloroquine and cycloguanil determined by pfcrt and dhfr sequences dhfr allele* Pfcrt allele
Wild-type Ser-108
Mutant allele Asn-108
Mixed dhfr alleles (Ser- and Asn-108)
Wild-type Lys-76 Mutant allele Thr-76 Mixed Lys/Thr-76
8 (6.8%) 2 (1.7%) 0
22 (18.6%) 62 (52.5%) 10 (8.5%)
6 (5.1%) 4 (3.4%) 4 (3.4%)
* Results are expressed as the number of Plasmodium falciparum isolates (%). Of 22 Thr-76 pfcrt and mutant dhfr, there were 1 single mutant (Asn-108), 2 double mutants (Arg-59 + Asn-108), and 19 triple mutants (Ile-51 + Arg-59 + Asn-108). Of 62 Thr-76, there were 1 single mutant, 8 double mutants, and 53 triple mutants.
quired to lead to proguanil prophylactic failure in nonimmune subjects. However, if we assume that a single Lys76Thr pfcrt mutation and at least one dhfr mutation (Ser108Asn) and/or chloroquine IC50 ⱖ 100 nM and cycloguanil IC50 ⱖ 15 nM are associated with chloroquine-proguanil prophylactic failure, our results suggest an insufficient protection (about 53%) against the erythrocytic stage of P. falciparum isolates in Yaoundé. Even if we adopt a more stringent theoretical requirement for prophylactic failure by assuming a single Lys76Thr pfcrt mutation and triple dhfr mutations, 53 of 118 isolates (45%) would still be expected to display double resistance to the drug combination. These considerations have been partially confirmed by our clinical observation of several cases of prophylactic failure in expatriates residing in Yaoundé, despite the allegation of strict compliance with the daily dose of chloroquine (100 mg) and proguanil (200 mg). Our results suggest that a high proportion (> 50%) of P. falciparum isolates in Yaoundé is resistant to both chloroquine and cycloguanil, as indicated by in vitro drug sensitivity assays and specific molecular markers. The high level of chloroquine resistance is supported by clinical studies in Yaoundé that have shown a parasitologic failure rate of 49% on day 14 and concluded that chloroquine can no longer be recommended for local populations with acute uncomplicated falciparum malaria.26 This conclusion may be extended to the entire southern and central Cameroon where our recent clinical and molecular data have shown a high prevalence of in vivo chloroquine resistance and triple dhfr mutations.27 Taken together, these data suggest the probable inadequate efficacy of chloroquine-proguanil combination in approximately one half of individuals exposed to infective mosquito bites in southern Cameroon. Acknowledgments: The author thanks the personnel of Nlongkak Catholic missionary dispensary in Yaoundé for their precious aid in recruiting patients and Fleurette Meche for technical assistance. Financial support: This study was supported by AstraZeneca (RueilMalmaison, France).
TABLE 1 Response to chloroquine and cycloguanil determined by in vitro drug sensitivity assay
Author’s address: Leonardo Basco OCEAC B. P. 288 Yaoundé Cameroon, Central Africa, Telephone: +237 2 232 232, Fax: +237 2 230 061, E-mail:
[email protected]
Cycloguanil response* Chloroquine response
Sensitive (IC50 < 15 nM)
Resistant (IC50 ⱖ15 nM)
Sensitive (IC50 < 100 nM) Resistant (IC50 ⱖ 100 nM)
13 (11.2%) 5 (4.3%)
36 (31.0%) 62 (53.4%)
* Results are expressed as the number of Plasmodium falciparum isolates (%). Two isolates with uninterpretable results for chloroquine were both cycloguanil resistant.
REFERENCES 1. World Health Organization, 2001. International Travel and Health. World Health Organization, Geneva. 2. World Health Organization, 1990. Practical chemotherapy of ma-
CHLOROQUINE-PROGUANIL RESISTANCE
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
laria. Technical Report Series no. 805, World Health Organization, Geneva. Lobel HO, Bernard KW, Williams SL, Hightower AW, Patchen LC, Campbell CC, 1991. Effectiveness and tolerance of longterm malaria prophylaxis with mefloquine. Need for a better dosing regimen. J Am Med Assoc 265: 361–364. Lobel HO, Miani M, Eng T, Bernard KW, Hightower AW, Campbell CC, 1993. Long-term malaria prophylaxis with weekly mefloquine. Lancet 341: 848–851. Steffen R, Fuchs E, Schildknecht J, Naef U, Funk M, Schlagenhauf P, Phillips-Howard P, Nevill C, Sturchler D, 1993. Mefloquine compared with other malaria chemoprophylactic regimens in tourists visiting East Africa. Lancet 341: 1299–1303. Gozal D, Hengy C, Fadat G, 1991. Prolonged malaria prophylaxis with chloroquine and proguanil (chloroguanide) in a nonimmune resident population of an endemic area with a high prevalence of chloroquine resistance. Antimicrob Agents Chemother 35: 373–376. Garin D, Lamarque D, Ringwald P, Dupuy O, Chaulet JF, Chapalain JC, Flechaire A, 1993. Efficacy of chloroquineproguanil chemoprophylaxis against malaria in the Central African Republic. Trans R Soc Trop Med Hyg 87: 304–305. Foote SJ, Galatas D, Cowman AF, 1990. Amino acids in the dihydrofolate reductase-thymidylate synthase gene of Plasmodium falciparum involved in cycloguanil resistance differ from those involved in pyrimethamine resistance. Proc Natl Acad Sci USA 87: 3014–3017. Peterson DS, Milhous WK, Wellems TE, 1990. Molecular basis of differential resistance to cycloguanil and pyrimethamine in Plasmodium falciparum malaria. Proc Natl Acad Sci USA 87: 3018–3022. Parzy D, Doerig C, Pradines B, Rico A, Fusai T, Doury JC, 1997. Proguanil resistance in Plasmodium falciparum African isolates: assessment by mutation-specific polymerase chain reaction and in vitro susceptibility testing. Am J Trop Med Hyg 57: 646–650. Basco LK, Ringwald P, 2000. Molecular epidemiology of malaria in Yaounde, Cameroon. VI. Sequence variations in the Plasmodium falciparum dihydrofolate reductase-thymidylate synthase gene and in vitro resistance to pyrimethamine and cycloguanil. Am J Trop Med Hyg 62: 271–276. Fidock DA, Nomura T, Talley AK, Cooper RA, Dzekunov SM, Ferdig MT, Ursos LMB, Sidhu ABS, Naude B, Deitsch KW, Su XZ, Wootton JC, Roepe PD, Wellems TE, 2000. Mutations in the P. falciparum digestive vacuole transmembrane protein PfCRT and evidence for their role in chloroquine resistance. Mol Cell 6: 861–871. Basco LK, Ringwald P, 2001. Analysis of the key pfcrt point mutation and in vitro and in vivo response to chloroquine in Yaounde, Cameroon. J Infect Dis 183: 1828–1831. Djimde A, Doumbo OK, Cortese JF, Kayentao K, Doumbo S, Diourte Y, Dicko A, Su XZ, Nomura T, Fidock DA, Wellems TE, Plowe CV, Coulibaly D, 2001. A molecular marker for chloroquine-resistant falciparum malaria. N Engl J Med 344: 257–263. Mount DL, Nahlen BL, Patchen LC, Churchill FC, 1989. Adaptations of the Saker-Solomons test: simple, reliable colorimet-
16.
17. 18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
387
ric field assays for chloroquine and its metabolites in urine. Bull World Health Organ 67: 295–300. Basco LK, Ringwald P, 2002. Molecular epidemiology of malaria in Cameroon. X. Evaluation of pfmdr1 mutations as genetic marker for resistance to amino alcohols and artemisinin derivatives. Am J Trop Med Hyg 66: 667–671. Ringwald P, Bickii J, Basco LK, 1996. In vitro activity of antimalarials against clinical isolates of Plasmodium falciparum in Yaounde, Cameroon. Am J Trop Med Hyg 55: 254–258. Basco LK, Ndounga M, Keundjian A, Ringwald P, 2002. Molecular epidemiology of malaria in Cameroon. IX. Characteristics of recrudescent and persistent Plasmodium falciparum infections after chloroquine or amodiaquine treatment in children. Am J Trop Med Hyg 66: 117–123. Howells RE, Gilles HM, Bent NS, 1985. A comparison of the pyrimethamine and cycloguanil sensitivities of the preerythrocytic and erythrocytic stages of drug-sensitive and -resistant strains of Plasmodium yoelii. Ann Trop Med Parasitol 79: 247–251. Watkins WM, Mberu EK, Nevill CG, Ward SA, Breckenridge AM, Koech DK, 1990. Variability in the metabolism of proguanil to the active metabolite cycloguanil in healthy Kenyan adults. Trans R Soc Trop Med Hyg 84: 492–495. Ndounga M, Basco LK, Ringwald P, 1999. Variabilité de l’activité in vitro du proguanil et du cycloguanil sur les stades érythrocytaires de Plasmodium falciparum en fonction des conditions de culture. Bull Soc Pathol Exot 92: 313–316. Basco LK, Ringwald P, 2001. Molecular epidemiology of malaria in Yaounde, Cameroon. VIII. Multiple Plasmodium falciparum infections in symptomatic patients. Am J Trop Med Hyg 65: 798–803. Rallon N, Osorio L, Giraldo L, 1999. Lack of an association between the Asn-108 mutation in the dihydrofolate reductase gene and in vivo resistance to sulfadoxine/pyrimethamine in Plasmodium falciparum. Am J Trop Med Hyg 61: 245–248. Basco LK, Tahar R, Keundjian A, Ringwald P, 2000. Sequence variations in the genes encoding dihydropteroate synthase and dihydrofolate reductase and clinical response to sulfadoxinepyrimethamine in patients with acute uncomplicated falciparum malaria. J Infect Dis 182: 624–628. Nzila AM, Mberu EK, Sulo J, Dayo H, Winstanley PA, Sibley CH, Watkins WM, 2000. Towards an understanding of the mechanism of pyrimethamine-sulfadoxine resistance in Plasmodium falciparum: Genotyping of dihydrofolate reductase and dihydropteroate synthase of Kenyan parasites. Antimicrob Agents Chemother 44: 991–996. Ringwald P, Same-Ekobo A, Keundjian A, Kedy Mangamba D, Basco LK, 2000. Chimiorésistance de P. falciparum en milieu urbain à Yaoundé, Cameroun. Part I. Surveillance in vitro et in vivo de la résistance de Plasmodium falciparum à la chloroquine entre 1994 et 1999 à Yaoundé, Cameroun. Trop Med Int Health 5: 612–619. Basco LK, Ndounga M, Tejiokem M, Foumane Ngane V, Youmba JC, Ringwald P, Soula G, 2002. Molecular epidemiology of malaria in Cameroon. XI. Geographic distribution of Plasmodium falciparum isolates with dihydrofolate reductase gene mutations in southern and central Cameroon. Am J Trop Med Hyg 67: 378–382.
