Therapeutic efficacy test in malaria falciparum in Antioquia, Colombia

After the use of any antimicrobial, treatment failure or resistance occurs in a percentage of cases, due to the variants or genetic mutants. In Plasmodium, mutations appear with a frequency of 9.5-71.2 × 10^-10 mutations/site/year in coding regions [1] and with a frequency of 1.5910 × 10^-4 mutations/site/year in microsatellites [2].

The in vitro evaluation of resistance of Plasmodium to an antimalarial drug consists of assessing growth inhibition when exposed to different concentrations of the drug, and several methods have been described [3‐6]. The in vivo evaluation, or therapeutic efficacy test (TET), measures the clinical and parasitological response of malaria patients to a specific antimalarial treatment. The clinical response is indicated by the clearance of symptoms and signs, and the parasitological response by the clearance of blood parasites, after oral administration of recommended doses of one or several drugs, under supervision of the sanitary staff, and with a follow-up during a determined period (usually, 7 to 28 days). The TET evaluation protocol, approved by the World Heath Organization (WHO) in 1998 [7], has been used in many South-American studies. In this protocol, TET has been classified into three categories: early treatment failure, late treatment failure and adequate response. This treatment failure refers exclusively to inability to be cured of malaria, not to the inability to prevent the disease when the drug has been used in a prophylactic fashion. Treatment failure has been described in two out of the four species of Plasmodium, which affect humans: Plasmodium vivax and, especially, Plasmodium falciparum; the latter has shown resistance to many antimalarial drugs [8].

In Colombia, between 1961 and 2003 (Table 1), the non-weighed mean of treatment failure in falciparum malaria was 66% for CQ (13 reports), 22% for AQ (9 reports) and 15% for SP (14 reports). In recent publications, the failure of the malaria control programme in Colombia has been described and the causes analysed [9‐11]. The geographic, environmental, socio-demographic, economic characteristics and general malaria morbi-mortality, in Antioquia, Urabá and Bajo Cauca regions as well as in El Bagre and Turbo municipalities have been detailed by others articles [10, 12‐15]. Demographic and epidemiological changes ocuring in Colombia during the XXth century have been reviewed and detailed elsewhere [16]. These reports constitute an integral reference framework to interpret the antimalarial treatment failure and resistance and to outline policies and programmemes to reduce their effects.

On this background, this report: a) collects and summarizes the available information about treatment failure in Colombia; b) presents the results of the evaluation of TET for various treatment schedules, between 2000 and 2004, in the municipalities of Turbo (Urabá region), El Bagre and Zaragoza (Bajo Cauca region), in the north-west and north-east of the state of Antioquia, Colombia; c) collects and summarizes information about use of the combination AQ-SP for the treatment of non-complicated falciparum malaria; and d) discusses the creation of an epidemiological surveillance system of antimalarial treatment failure as part of the malaria programmeme operating in the country.

Of the 520 patients included in the study, 47% were recruited in Turbo (Urabá zone) and 53% in The Bajo Cauca zone (El Bagre 37% and Zaragoza 16%). Data were analysed independently from the residence place of the patient since important differences were not found when this variable was considered.

The patients exhibited the following characteristics:

TET could be measured in 93.7% of the patients (487 out of 520 patients). The reason for not evaluating TET in 33 patients was the impossibility of finding them (90%: 30/33) or voluntary withdrawal from the investigation (10%: 3/33). Failure to attend increased further into the follow-up time. Such defaulting was statistically different among treatments, and the four monotherapies had less defaulting overall (4%) than the four treatment combinations (8%). This percentage of defaulting per treatment varied from major to minor: a) in monotherapies: it was 9% with MQ, 3% with AQ, none with CQ and SP; b) in combinations: it was 12% with MQ-SP, 10% with AQ-SP and AS-SP, none with CQ-SP. Defaulting was higher (9.3%) in El Bagre-Zaragoza, which largely depends for communication on the Nechí River, than in Turbo (5.7%), in which basic transport is by land. The group of 33 non-attending patients did not differ from the non-defaulters (n = 487) in relation to the residence zone (urban, rural), sex, ethnicity, age, days of evolution of the current disease and initial parasitaemia.

The mean of asexual parasitaemia in 487 patients before treatment administration was 7,700 parasites/μL, the median was 4,480, the 25 percentile was 2,000; the 75 percentile was 10,130 and the standard deviation was 8,472 parasites/μL. A 26% of patients had as many as 2,000 parasites/μL, 45% as many as 4,000, 50% up to 5,000, 75% up to 10,000 and only 8% over 20,000 parasites/μL.

The mean asexual parasitaemia before treatment administration was not statistically equal among the treatments; it varied between 8,954 (MQ) and 5,203 parasites/μL (CQ-SP) (p < 0.05); between MQ and SP (5,663 parasites/μL) there was a statistically significant difference. Gametocytes were found in 17% of patients before treatment, in 23% on day 1, 33% on day 2, 43% on day 3, 65% on day 7, 64% on day 14, 51% on day 21, and 15% on day 28. Before the treatment, the mean number of gametocytes ranged from 22 (AS-SP) to 140 gametocytes/μL (AQ-SP), with a general mean of 78 and a standard deviation of 421 gametocytes/μL (minimum 0 and maximum 7,600).

Monotherapy with CQ showed 82% failure, with AQ it showed 30%, with MQ 4%, and with SP 24%, while the combinations of SP with other drugs showed the following failure rates: CQ-SP 17%, AQ-SP 2%, MQ-SP 0% and AS-SP 3% (Table 4). There were 62 failures (12,7%) in 487 patients in whom the final effect could be measured. Early failure contributed with 42% of the total failure (26/62) and the late failure accounted for 58% (36/62) of it. Considering defaulting patients as failures the percentage of failure reached 18% (62 observed failures + 30 supposed failures/520). The frequency of early and late failures changes from one treatment to another: a) with monotherapies, early and late failures were equal to the those with CQ, early failures were predominant with AQ (ratio 1:4), and the late ones were the majority with SP (2:1); b) with the CQ-SP combination, the late failures widely prevailed (9:1). With the rescue treatment, which varied according to the primary treatment, all treatment responses were successful; which meant no failure of two treatments was observed.

Treatment failure was not associated with initial asexual parasitaemia (pre-treatment), although there were important differences as follows: those who successfully responded and those who had early failure had 7,760 and 7,951 parasites/μL, respectively, while those who had late failure showed 6,731 parasites/μL; despite that, the differences were not statistically significant. Neither was there any relation of TET with initial gametocytaemia. Thus, although initial asexual parasitaemia was not equal in all eight treatments, this fact did not have any effect in the TET.

It should be pointed out that there was no significant association of the variable TET with any one of these explicative variables: municipality, zone, sex, age group, ethnic group, malaria antecedents, and hospitalization or transfusion antecedents.

Antimalarial schemes administered (dose, scheme, application time) corresponded to those used for many years in Colombia and recommended by other authors [8, 26]. Direct surveillance of patients by the researchers during drug intake assured that the drugs were taken, and ruled out losses due to vomiting or diarrhea.

The protocol used for measuring TET was identical in most studies, with exception of the studies of 2000–2001, when CQ, SP and CQ-SP were evaluated over 21 days; in those undertaken between 2001 and 2004, the follow-up lasted 28 days.

The number of patients in whom TET could not be measured (lost patients) was low (6% in the eight studies) and was not affected by the minimum sampling size. The strategy used to avoid defaulting consisted of monetary support (of up to 5,000 Colombian pesos, i.e. less than US$2) to cover transportation costs, as well as the Immediate Active Search (IAS) for any patient who missed a follow-up examination. In fact, defaulting rate has been higher than 40%, but was reduced to 6% through the application of that strategy. IAS was performed by the physician or clinical laboratory technician, accompanied by the health assistant, using motorcycles supplied by the local hospital. Even in the worst case scenario, i.e. if all patients in whom TET was could not be measured are considered as failures, the percentage of treatment failure for each treatment schedule was under 14%, except in the case of AQ, which had net failure (excluding defaulters) of 30%, and gross failure of 33% (when counting defaulters as failures).

From the evaluated schizonticide monotherapies (CQ, AQ, MQ, SP), only CQ has been used in this way in Colombia, although associated to primaquine as gametocide (45 mg in a single dose). CQ had become useless in 1998 (according to López et al. [27] and Blair et al [28‐30]), due to recorded failures between 71% and 67% now reaching 82%. The national average of CQ failure from 1961–2003 is 66% (Table 1), which indicates that it should no longer be used anywhere in the country. The CQ-SP combination (failure 17%) seems to remains useful, but efficacy should be evaluated in other areas (e.g. Orinoquia, Amazonia, Catatumbo).

The accelerated growth of failure presented by the monotherapy AQ is serious because it increased from 7% in 1998 in El Bagre [27], Zaragoza [28, 30] and Turbo [28, 29] to 30% in 2004, which excludes the drug for monotherapy, according to the 25% threshold proposed by WHO [7]. The national average of AQ failure is 22%, but there are recent data (2002–2003) coming from Tadó (Chocó) indicating a failure rate of 27% (Table 1) and from Tumaco (Nariño) which reports 50%, although there was no failure in El Charco, in the same state [31].

Failure of the monotherapy with SP also increased strongly, going from 12% in 1998 [27‐30] to 24% current. The MQ-SP combination showed a zero failure rate, but AS-SP showed a failure of 3%, although never officially used in Colombia.

The AQ-SP combination is highly efficient (98%). The relatively scantly use of the AQ was a result of the adverse reactions which had been observed when AQ was used for prophylaxis [32], a situation in which it is taken for weeks or months, but reports on serious adverse effects of its use as an antimalarial are practically absent [33‐35]. In a different study carried out by us in Turbo, 57 patients treated with AQ monotherapy were clinically and laboratory (blood and liver tests) assessed before treatment and 10 days after it. We concluded that, at the standard dose (25 mg/kg peso) and period (3 days), as recommended for uncomplicated falciparum malaria, AQ had no adverse effects or toxicity (paper in preparation).

A systematic review of 2001 compared AQ-SP and CQ-SP combinations for the treatment of non-complicated malaria; on the basis of seven studies in 1,277 patients, it concluded that there was no evidence of serious collateral effects with those combinations and that they, more than monotherapy with SP, "could make people feel better more rapidly", and caused a better steady elimination of blood parasites [36]. In 2005, a Cochrane review aimed at assessing " the combination of CQ or AQ plus SP compared with SP alone for first-line treatment of uncomplicated falciparum malaria", concluded that "the evidence base is not strong enough to support firm conclusions. The available evidence suggests that AQ plus SP can achieve less treatment failure than SP, but this might depend on existing levels of parasite resistance to the individual drugs" [37].

According to the results obtained in this study and to the additional information presented here, treatment with AQ-SP should remain as the first line treatment of non-complicated falciparum malaria in Colombia, as it is at present [24], leaving options based on MQ and on AS for the future, once AQ-SP has failed.

Colombia should establish an "Antimalarial Failure Epidemiological Surveillance System" (AFESS). The AFESS should have, as primary objective, the early detection of failure by means of a screening test. For this, a selected screen test should be low cost, applicable in field conditions and have a high sensitivity. Our recommendation is to install an AFESS coupled to the antimalarial programmeme, to measure TET during a following up period of 7 – 14 days and to install sentinel posts in strategic places of malaria regions. In addition the AFESS, should: a) carry out specific research projects aimed at confirmation of therapeutic failure after a following up period of 28 days or more, b) evaluation of P. falciparum in vitro resistance; c) P. falciparum y P. vivax genetic analysis and d) assessment of antimalarial blood concentrations in patients studied.