Chloroquine failure rates were higher than anticipated, and since administration was directly observed, failure was due to resistance rather than poor adherence. Analysis showed that with 51% failure in CQ40 and 84% failure in CQ25, chloroquine is no longer suitable for falciparum malaria treatment among Afghan refugees, either as first or second-line with short or extended regimens, and usage should stop. Second-line CQ40 achieved a higher failure rate than did first-line, demonstrating lack of suitability for this purpose.
While the authors were unable to make use of PCR genotyping to distinguish recrudescent from new infections, other in vivo trials conducted in the area that did include genotyping indicated fewer than 5% would be new infections . The most compelling evidence for subsequent infections being therapeutic failures is the very high failure rates by Day 28 and apparent absence of any new parasitaemias over 60 days of follow up.
The finding of the pfcrt 76T mutation in 100% of isolates analysed is consistent with a low degree of heterogeneity in the parasite population, as also shown by Khatoon et al in isolates from nearby Bannu district NWFP . The pfcrt 76T allele is strongly associated with CQ and amodiaquine (AQ) resistance in falciparum isolates from Asia, Papua New Guinea, Africa, and South America [9, 15]. Pfmdr1 86Y and 184Y alleles, which are also associated with CQ and AQ resistance, were present in only a minority of our isolates from Adizai camp or from Bannu district . In a clinical trial of AQ in nearby Afghanistan, which also resulted in high rates of recrudescence, pfmdr1 alleles were not strongly selected among treatment failures . These findings indicate that the pfcrt codon 72-76 haplotype SVMNT present in Pakistan is sufficient by itself (i.e. without pfmdr1 86Y and 184Y) to cause high-level CQ and AQ resistance [9, 22]. By contrast, in Africa where the CQ-resistant variant pfcrt codon 72-76 CVIET appears to be the predominant haplotype, AQ remains relatively effective . In the one clinical trial in East Africa where AQ did demonstrate high levels of in vivo resistance, the CQ-resistant variant CVIET haplotype was present with pfmdr1 86Y and 184Y alleles which presumably added to the resistance there [9, 23–26].
The rate of parasitological failure was higher after second-line than after first-line CQ40 treatment. Recrudescent infections presumably started with higher proportions of resistant parasites than did initial infections. However, this cannot explain why the initial CQ40 course seemed to eliminate around 39% of resistant infections, as indicated by the improved cure rates over 60 days following initial five-day (51% recrudescence) as compared to three-day treatment (84% recrudescence). Among this 39%, any resistant parasites must have been removed by the additional two days of treatment and did not reappear over the subsequent 60 days. Ursing et al have, in parallel, undertaken clinical studies with high-dose CQ in Guinea-Bissau [27, 28]. They found that high-dose CQ (75 mg/kg as split-dose over five days) was well-tolerated (as was the 40 mg/kg administered in the present trial) and 78% of infections carrying pfcrt 76T were successfully treated compared to only 34% with 25 mg/kg . This was a higher treatment success rate than in Pakistan.
While pfcrt 76T was highly prevalent in the Pakistan samples, pfcrt 76T prevalence in the Guinea-Bissau population, discussed above, remained stable at a much lower 25% between the years 1990 and 2005 [27, 29]. These contrasts in pfcrt 76T between continents are likely due to differences in the fitness of resistance alleles, as the pfcrt 72-76 SVMNT resistance haplotype dominant in India, Iran, Pakistan and Afghanistan is not associated with re-emergence of CQ sensitivity or fluctuations in seasonal prevalence shown by the CVIET haplotype in some parts of Africa [27, 28, 30–32]. Drug pressure may also affect stability. If most infections were treated with a quinoline, pfcrt 76T frequency would remain high. In the African settings, sensitive parasites may have found a niche in the many untreated infections, where their greater fitness would allow them to compete better than any co-infecting resistant parasites. It has been more difficult for CVIET-carrying parasites to gain the same high prevalence in Africa as SVMNT-carrying parasites have achieved in parts of Asia.
The present trial did not measure adherence, as it was designed to assess efficacy rather than effectiveness. Consequently, it cannot challenge the initial assumption that refugees fail to adhere to either three-day or five-day courses. However, recent research demonstrates that with appropriate instructions - as in a trial of unsupervised 14-day primaquine treatment - Afghan refugees will adhere to much longer treatment regimens than the five-day course described here [22, 33]. What is clear from directly-observed treatment is that neither CQ25 nor CQ40 regimens can be justified, as neither can provide acceptable cure rates given the high prevalence of CQ-resistant falciparum malaria now in Pakistan [4, 5].
In Afghanistan, where many refugees have returned, most malaria cases are still treated with CQ, without parasitological diagnosis by either microscopy or rapid diagnostic test . This makes the present study, though conducted more than ten years ago, still highly relevant. About 70-90% of cases in Afghanistan are due to vivax and will respond to CQ. However, most of the falciparum cases treated with CQ - whether for three days or longer - are likely to fail . While the total number of falciparum cases will be small, without effective treatment these risk developing into severe malaria. As up to one-fifth of suspected malaria cases arrive at government clinics with detectable chloroquine present in their urine, irregular or intermittent treatment with chloroquine might be common . The need for routine parasitological diagnosis by microscopy or rapid diagnostic test to allow differential and, most importantly, effective treatment for both falciparum and vivax malaria is paramount in Pakistan and Afghanistan.
Although combination therapy using artesunate-SP has been adopted as policy for treatment of confirmed falciparum malaria in Pakistan and Afghanistan, implementation is patchy and uptake slow . While follow-up and numbers of SP patients were too few in this trial to determine significance, the 12% SP failure rate raises questions about SP's long-term efficacy. While SP has a role as combination partner in Pakistan and Afghanistan, if administered without artesunate, resistance to SP may select rapidly [36, 37].