Authors Walter Mulombo Kazadi, Burstein Nganga Makina and Jean Coran Mantshumba, Programme National de Lutte contre le Paludisme (PNLP), Kinshasa, DRC. Sirenda Vong, Epicentre, Kinshasa, DRC. Willy Kabuya, BASICS, Kinshasa, DRC. Benoit Ilunga Kebela, Direction de la Lutte contre la Maladie (4ème Direction du Ministère de la Santé), Kinshasa, DRC. Nkuku Pela Ngimbi, Institut National de Recherche Bio-médicale (INRB), DRC.
We evaluated the in vivo responses to chloroquine (CQ), the first line antimalarial, and to sulfadoxine–pyrimethamine (SP), the most readily available and affordable alternative treatment, in children under 5 with acute uncomplicated Plasmodium falciparum malaria in seven sites of Democratic Republic of Congo (DRC) between May 2000 and November 2001, using the standard 14-day WHO protocol. In the CQ group, the overall treatment failure rate was 45.4% (95% CI: 40.1–50.8) of 350 infections successfully tested; in the SP group it was 7.5% (95% CI: 5.0–11.0) of 333 infections. Of 191 patients who had an adequate clinical response (ACR) in the CQ group, 127 (66.5%; range: 62.5–71.4) still had parasitaemia on day 14. In the SP group, only 21 (6.8%; range: 2.2–12.8) of 308 patients with an ACR were still parasitaemic on day 14. Using pooled data from three rural sites, haematological recovery was better in the SP group (mean of haematocrit difference between days 14 and 0 among anaemic children: 4.7 vs. 3.2; P < 0.01, Wilcoxon test). These findings suggest that CQ is no longer effective in DRC and that SP may be a good alternative for its replacement as first line antimalarial treatment. The Ministry of Health (MOH) therefore now recommends SP as the first line antimalarial drug in DRC, as an interim step, 18 months after launching the first study. Additional studies are needed to select alternative therapies that might replace SP or improve its efficacy, should it prove ineffective in the future.
Sub-Saharan Africa, which bears the greatest burden of malaria morbidity and mortality, faces an impending catastrophe because of rising rates of chloroquine (CQ) resistance and associated increases in childhood mortality (Campbell 1991; Trape et al. 1998). Resistance of Plasmodium falciparum to CQ has been documented in Eastern Democratic Republic of Congo (DRC) since 1983 (Delacollette et al. 1983). Many other studies have confirmed that it has spread to the west of the country as well, and that sulfadoxine–pyrimethamine (SP) had a relatively lower resistance. However, these studies were done using various protocols, making it difficult to compare trends over time and between different places. Table 1 shows the results of these studies.
Table 1. In vivo sensitivity of Plasmodium falciparum strains to chloroquine (CQ) and sulfadoxine–pyrimethamine (SP), 7-day test, 1983–1989, Democratic Republic of Congo
* Sensitive = absence of asexual parasites by day 7.
Source: Programme National de Lutte contre le Paludisme.
The CQ resistance is a serious impediment to proper treatment of malaria infections and disease, a situation that can result in potentially severe or fatal consequences, such as cerebral malaria and anaemia. In order to update its drug use policy, the National Malaria Control Program with the support of international partners, decided to reassess the efficacy of CQ, the first line antimalarial treatment, and SP, the most readily available and affordable alternative treatment. These drugs were assessed in selected urban and rural sites, using the WHO standard protocol (WHO 1996), slightly modified for local conditions. We report here findings in these sites.
We undertook the studies between May 2000 and November 2001, successively in Kinshasa (6 000 000 inhabitants; urban), Mikalayi (20 000 inhabitants; Central DRC, rural), Kapolowe (20 000 inhabitants; South-East DRC, rural), Vanga (10 000 inhabitants; West DRC, rural), Kimpese (30 000 inhabitants; West DRC, rural), Kisangani (480 000 inhabitants; East DRC, urban) and Bukavu (369 610 inhabitants; East DRC, urban). The studies took place either in health centers (Kinshasa, Mikalayi, Kapolowe, Kimpese), or at the outpatient department of the District Referral Hospitals (Vanga, Kinshasa second site, Bukavu, Figure 1).
Febrile patients aged 6–59 months presenting at the health facilities were enrolled in the study if we had informed consent from the parent or guardian and if they had an axillary temperature ≥37.5 and <39.5 °C; mono-infection with P. falciparum, with parasitaemia in the range of 2000–100 000 asexual parasites per microlitre blood; no other cause for fever than malaria; no general danger signs (unable to sit or stand up, unable to drink or breastfeed, lethargy or unconsciousness, recent history of convulsions, persistent vomiting) or signs of severe and complicated falciparum malaria according to the definition given by WHO (1996). To be included, participants should be able to come for the stipulated follow-up and needed access to the health facility. Parents/guardians were then interviewed by a study physician about symptoms, duration of the illness, previous antimalarial therapy and other medications.
Children were examined for pallor and jaundice or any other danger sign, and their axillary temperature and weight were measured. A 10% treatment failure rate was expected per drug. An arbitrary sample of 50 patients per regimen and per site was required, with a maximum acceptable drop out rate of 10%.
Enrolled children were either treated with a standard dosed CQ (25 mg/kg body weight CQ base over 3 days) or SP (1.25 mg/kg body weight of pyrimethamine plus 25 mg/kg body weight of sulfadoxine in one single day). Drugs were obtained from either Mumbay, India or IDA Denmark for CQ, and from either CREAT S.A. International, PO B17-28501 Vernouillet, France or IDA Denmark for SP. Drug content was confirmed at CDC by high performance liquid chromatography, and/or at LAPHAKI, the governmental control office for drugs in DRC. Medications were dosed according to the modified weight-based guidelines from WHO for administration of fractions of tablets. In five sites (Kapolowe, Kimpese, Kinshasa, Mikalayi and Vanga), assignment of patients to drug regimens was made alternatively, either by completing the sample for one drug before the other (four sites), or by alternating the drug on every other day (Kimpese). In the two other sites (Bukavu and Kisangani), patients were randomly assigned CQ or SP, using computer-generated random numbers tables. Personnel were not blinded to treatment. Drugs were administered by a medical doctor and repeated in case of vomiting within 30 min.
In the case of a CQ treatment failure, SP was given, while failures to SP and severe cases were treated with quinine for 7 days. In addition, paracetamol was administered on days 0 and 1 in case of fever. In case of emergence of a concomitant infection during the follow-up, an antibiotic without antimalarial activity was administered.
Follow-up appointments were scheduled for days 1, 2, 3, 7 and 14, and consisted of physical examination and completion of a standardized form. Parents and guardians were encouraged to return to the clinic at any time if their children felt unwell. Blood was collected by finger prick on days 3, 7 and 14, for repeat thick and thin blood smears, and to repeat haematocrit or haemoglobin on day 14. This was also performed on any other day as required, based on clinical examination.
Patients who did not turn up on a scheduled day were visited at home. Patients were excluded from the study: (1) if they reported self-administration of antimalarial drug during the follow-up; (2) they withdrew consent; (3) a concomitant disease occurred that would interfere with the clear classification of treatment outcome; (4) or the patient moved out of the follow-up area. In any other condition, patients who did not turn up on a critical day or were not seen at home were considered lost to follow-up.
Patients recruited for the study had a blood sample obtained by finger prick on day 0 for thick and thin blood film, haematocrit or haemoglobin test. The smears were stained with 3% Giemsa for 30 min. Parasite density was calculated by counting the number of asexual parasites against 200 White Blood Cells (WBC), and then multiplied by 40 on the assumption of a normal WBC count of 8000/μl. Thin blood smears were also examined for other Plasmodium sp. Haematocrit was determined by the micro-haematocrit method (Haematokrit 24, Hettich). In two sites, the haemoglobin level was measured by a spectrophotometric method (Bukavu: Haemocue; Vanga: Haematokrit 24).
The clinical outcomes were classified into three categories of therapeutic responses. Early treatment failure (ETF) referred to development of danger signs or severe malaria on days 1, 2 or 3 with parasitaemia; axillary temperature ≥37.5 °C on day 2, with parasitaemia higher than day 0 count; parasitaemia on day 3 with axillary temperature ≥37.5 °C; parasitaemia on day 3 ≥25% of count on day 0. Late treatment failure (LTF) was defined as development of danger signs or severe malaria after day 3 with parasitaemia, without previously meeting any of the criteria of ETF;or parasitaemia and axillary temperature ≥37.5 °C on any day from days 4 to 14, without previously meeting any of the criteria of ETF. Adequate clinical response (ACR) was constituted by absence of parasitaemia on day 14 irrespective of axillary temperature, without previously meeting any of the criteria of ETF or LTF; or axillary temperature <37.5 °C on day 14 irrespective of parasite count, without previously meeting any of the criteria of ETF or LTF.
We also assessed parasitological failure (PF) by breaking down the results of the ACR group into two groups: ACR without parasitaemia at day 14 [(adequate clinical and parasitological response (ACPR)] and ACR with the presence of parasites at day 14 but without fever [adequate clinical response and parasitological failure (ACR/PF)].
From May 2000 through November 2001, starting in Kinshasa, and then in Mikalayi, Kapolowe, Vanga, Kimpese, Kisangani and Bukavu, 743 children were enrolled, of whom 40 (5.4%) were excluded, 20 (2.7%) lost to follow-up, yielding 683 (91.9%) who completed the study. Of these, 350 and 333 received CQ and SP respectively.
In Kinshasa, 22 children who did not respond to CQ treatment were subsequently enrolled in the SP group. Among them, nine had an axillary temperature <37.5 °C. In addition, other children were enrolled with either parasitaemia out of range (<2000 or more than 100 000 parasites/μl blood), or haematocrit of <15%, or axillary temperature >39.5 °C. However these cases accounted for <5% of the overall enrolled patients. As illustrated in Table 2, characteristics of patients enrolled were similar between CQ and SP groups in each site.
Table 2. Baseline characteristics at enrolment of children under 5 with acute uncomplicated falciparum malaria treated with chloroquine (CQ) or sulfadoxine–pyrimethamine (SP) in seven study sites of Democratic Republic of Congo, May 2000–November 2001
Kimpese (Bas Congo)
Kisangani (Province Orientale)
Mikalayi (Kasai occ.)
* Mean haemoglobin level (±SD) in g/dl.
Number of patients
Mean age (±SD) in months
Mean weight (±SD) in kg
Mean temperature (±SD) in °C
Mean haematocrit (±SD) in %
Mean parasitaemia density (range) per μl
30 555 (2020– 97 164)
24 345 (2050– 90 150)
36 989 (3423– 98 500)
38 856 (2444– 99 321)
32 349 (2029– 95 167)
29 562 (2010– 98 258)
29 768 (1120– 366 000)
35 525 (1440– 308 000)
40 440 (3350– 96 333)
28 451 (2057– 98 783)
32 385 (2030– 107 122)
39 830 (2511– 103 025)
27 443 (2120– 97 600)
27 051 (2547– 92 545)
Therapeutic response to CQ and SP
Overall, the treatment failure rate was 45.4% (95% CI: 40.1–50.8) in the CQ group and 7.5% (95% CI: 5.0–11.0) in the SP group. Treatment failure rates varied between sites from 29.4% to 80% in the CQ group and from 0 to 19.2% in the SP group. The highest treatment failure rates were observed in Eastern DRC sites, both in the CQ group (Bukavu: 80%) and the SP group (Kisangani: 19.2%).
Among patients with ACR, the overall proportion of patients still parasitaemic on day 14 was higher in the CQ group (66.5%) compared with those in the SP group (6.8%). This proportion varied from 62.5% to 71.4% in the CQ group and from 2.2% to 12.8% in the SP group (Tables 3 & 4).
Table 3. Therapeutic responses of CQ and SP in children under 5 with acute uncomplicated falciparum malaria, May 2000– November 2001, Democratic Republic of Congo
Table 4. Adequate clinical responses in children ≤5 years old with acute uncomplicated falciparum malaria* treated with CQ and SP in seven study sites, May 2000–November 2001, Democratic Republic of Congo
Parasitaemia at day 14
* Temperature ≥37.5 °C.
CQ, chloroquine; SP, sulfadoxine–pyrimethamine; ACR/PF, adequate clinical response and parasitological failure; ACPR, adequate clinical and parasitological response; TOTACR, total adequate clinical response.
Final parasitaemia and haematological response among patients with ACR
After pooling data from three rural sites (Kapolowe, Kimpese, Mikalayi) which had used the same method to assess haematological response (micro-haematocrit method), a highly significant difference was observed on PF rate between children with ACR in the CQ group vs. SP group [67% (67 of 100) for CQ vs. 3.6% (five of 139) for SP group; chi-squares (Mantel–Haenszel): 58.18; P < 0.000001]. Furthermore, when comparing children with ACR in the CQ group to those in the SP group, results showed a more significant improvement in their anaemic status after treatment with SP (Wilcoxon test, P < 0.01; Table 5).
Table 5. In adequate clinical response group, children under 5 with acute uncomplicated falciparum malaria treated with chloroquine (CQ) or sulfadoxine–pyrimethamine (SP). Comparison between CQ and SP groups from pooled data sets of three study sites of Democratic Republic of Congo (Mikalayi, Kapolowe, Kimpese). November 2000–April 2001
P value (Wilcoxon test)
* Definition of anaemia (haematocrit ≤ 30%).
Figures in parentheses are the range.
Mean of haematocrit difference between days 14 and 0 among anaemic children*(100 vs. 139)
According to these results, approximately half (45%) of children with uncomplicated P. falciparum malaria did not respond to CQ treatment whereas only 8% of those who were treated with SP did not recover appropriately. We also observed that eastern DRC (Bukavu) bears the highest level of CQ treatment failures (80%) while lowest rates are in the central and western DRC. Such findings are not unexpected and do support the reported trends in the spread of antimalarial drug resistance from East Africa to the West. Similar high levels of resistance were consistently documented in neighbouring countries such as Rwanda and Burundi.
Our results also indicate that two-thirds of patients who were considered ACR under CQ medication still had parasitaemia on day 14. In contrast, only about 7% of those who recovered under SP regimen did not clear their parasites on day 14. This suggests that SP might have retained its efficacy over time, while that of CQ had seriously decreased. However, our studies have some limitations to be mentioned.
First, our results may not be representative of the whole country. For practical reasons, we have selected study sites where a minimum of health care system was functional and we could expect high frequentation rates. Most of the sites were in rural settings. All the rural sites used had been supported by non-governmental partners. Population of these areas may have had easier access to drugs leading to a higher drug pressure, a situation that could lead to higher levels of resistance.
Secondly, patients were not systematically recruited from health centers. In two sites (Bukavu and Vanga), patients were recruited at the out patient department (OPD) of the district referral hospital and in one site (Kinshasa), one enrollment post was set up at the OPD of the referral hospital. Recruitment of patients from hospitals may lead to a selection of cases with special problems, including drug resistance, a situation that might in turn induce some bias in the overall results. However, in the context of DRC, with a destroyed health infrastructure, although members of the community tend to seek care in hospitals for severe cases, there is no evidence to date that uncomplicated cases are preferentially addressed to health centers. More over, the three OPDs are practically functioning rather like health centers than referral consultation services, thus receiving both uncomplicated and severe malaria cases from the community, as are currently doing health centers. Also, the purpose of these studies was to inform policy, not to measure the ‘true’ state of resistance. Therefore, even if the people attending a clinic were in some way a selected group, they are still the ones that need to be treated effectively. These studies, while maybe not telling you the true status of resistance at the community level, do tell you how to treat uncomplicated malaria in health facilities – which is what public health policy intends to do.
Thirdly, we did not measure CQ blood levels to demonstrate drug and metabolite concentrations. This may have led either to overestimation or to underestimation of PF rates. However, studies of drugs for which resistance is well known (such as CQ), now rarely include confirmation of drug absorption and metabolism; demonstration of persistence of parasites in a patient receiving directly observed therapy is usually considered sufficient, except in areas where CQ resistance has never been demonstrated (Bloland & Ettling 1999). In our studies, the treatment was directly observed. History of prior intake of CQ was not retained for the analysis because we did not rely on the mothers’ answers and there is no evidence on any different risk of clinical failure or poor parasitological response between patients with or without recent drug intake (Bloland et al. 1993). The Protein Chain Reaction (PCR) tests were not used because the re-infection risk is assumed to be low during the 14 first days of follow up.
Fourthly, in one site, 22 children were recruited from the CQ group, following a CQ treatment failure. However, no significant differences in responses to SP were noted between these children and the other children in the SP group.
Fifthly, the sampling method, with alternative assignment of drug regimen in some sites and a random procedure in others might affect the representativeness of the results, which may not be appropriate to calculate a point estimate. However, this protocol has a public health surveillance goal, and does represent a tool to monitor and guide decision makers in updating the antimalarial drug use policy in different settings. In addition, no difference was observed in baseline characteristic among children enrolled in the various study sites.
Despite these limitations, our results confirm trends already described in sub-Saharan Africa. Our studies showed that SP was much more effective than CQ in treating children under 5 with uncomplicated malaria. They are consistent with recent reports in our region (Ringwald et al. 2000a, b; 39.7% of therapeutic failures for CQ and 12.1% for SP in Yaounde, Cameroon; 38.9% (89 of 229) of therapeutic failure of CQ in Zambia (Barat et al. 1998).
The objectives of any national antimalarial drug policy are first to make effective, safe, low-cost essential drugs available and affordable for the entire population, and secondly to ensure that drugs are of good quality and are used rationally. CQ has been the longest-lived first-line antimalarial drug because it satisfies the above objectives well (Kitua 2000).
The level of therapeutic efficacy to a particular drug is an important determining factor in the process of developing or revising an antimalarial drug policy. The decision to change a policy is extremely difficult for policy and decision-makers of many malaria-endemic areas of Africa, because there are not many other antimalarials with similar qualities to CQ on the market. In order to simplify the process leading to decision making, a systematic approach has been suggested (WHO antimalarial-drug policy guidelines; Kitua 2000). This dynamic process can be divided into four major periods, based on the level of the clinical failure: the period of ‘grace’ (clinical failure: 0–5%) during which the drug will be functioning at the highest level of its effectiveness; the period of ‘alert’ (clinical failure: 6–15%) which calls for clear and systematic preparation of eventual actions as the situation deteriorates; the ‘action’ period (clinical failure: 16–24%) during which specific actions must be taken, including countrywide ascertainment of treatment failure, assessment of other options available, and their cost and distribution, and reaching final consensus on the need to change; and the period of change (clinical failure: ≥25%). Increasing P. falciparum resistance to CQ in sub-Saharan Africa necessitates use of alternative and affordable antimalarial agents.
Uganda decided to change its antimalarial drug use policy after high levels of treatment failure rates to CQ had been documented in the country, and is currently using CQ + SP as the first line antimalarial treatment. This choice was based on evidence of good efficacy of this regimen, as suggested by preliminary results of drug studies in Rwanda. Unfortunately, in-country monitoring of the first line treatment in 2002 has shown that resistance to this regimen is worsening. Therapeutic failures rates >25% have been reported [Uganda NMCP oral communication, Multilateral Initiative on Malaria (MIM)]. Rwanda itself shifted directly to SP plus amodiaquine, as the first line antimalarial treatment after documenting high level of CQ resistance in the country. Burundi has recently adopted amodiaquine plus artesunate, though still using SP alone as an interim step, after shifting from CQ.
Based on our results, there were numbers of points to be considered prior to recommending a change in first line therapy for malaria in DRC. The level of treatment failure currently recommended for change of the first line treatment is 25%, based on a 14-day follow-up of treated mild malaria cases (WHO antimalarial-drug guidelines; Shapira et al. 1993). When the period of ‘change’ is reached (treatment failure ≥25%), change must be made and it must be quick because otherwise the cost of delay, in terms of unnecessary human lives lost may be high (Kitua 2000). In four of seven sites, the CQ therapeutic failure is thought to be significantly above 25% (Bukavu, Kimpese, Kisangani and Vanga).
In a context of high PF rates and a 45% cost differential between CQ and SP, SP is more cost effective than CQ (Sudre et al. 1992). Approximations from parasitological outcomes suggested that in DRC, where PF rates are high in the CQ group and where SP is as inexpensive as CQ, using SP should be cost effective.
Haematological recovery is also argued to be an indicator of a change in first-line treatment (Bloland & Ettling 1999). Using pooled data, we showed that anaemia improved to a significantly greater degree after treatment with SP compared with CQ among children with ACR. We also observed a higher proportion of parasitaemia asymptomatic children among those treated with CQ [range from 62.5% to 71.4% with an overall of 66.5% (127 of 191)] than ones treated with SP (range from 2.2% to 12.8%) at the end of the follow-up period. This situation implies a higher risk of morbidity and mortality because of anaemia, potential for increased risk of human immunodeficiency virus transmission because of avoidable blood transfusions, and an increase in the reservoir of infection in the community among incompletely treated children (Greenwood 1987, Greenwood et al. 1987; White et al. 1999).
Our findings suggested that there was an urgent need for the country to revise its antimalarial drug use policy and that SP appeared to be the only feasible alternate for this shift at this time. Thus, we recommended halting use of CQ as first line treatment. In reaching a final consensus on malaria treatment policy, the following issues were considered: (1) data suggesting a potential for a rapid increase of resistance to SP when used as monotherapy (Watkins et al. 1997); (2) evaluation of alternatives to SP, including artemisinin-containing combinations; (3) relative costs of alternative treatment strategies and availability of resources to pay for alternatives; (4) monitoring changing patterns of resistance through a network of sentinel sites; (5) slowing down the resistance occurrence by improving rational medical prescription, patient or caregiver compliance and the use of selective and sustainable prevention measures such as the use of insecticide treated bed nets; and (6) improving the effectiveness of policy by developing a greater understanding of the patient or caregiver perception of the new drugs, treatment seeking behavior and provision of care from community and private sectors.
Based on the findings of these studies and on careful consideration of the other issues outlined above, the MOH decided to update the antimalarial drug use policy in DRC and recommended the replacement of CQ with SP for the treatment of uncomplicated malaria cases, 18 months after starting the first study. This policy is considered an interim policy to be used for 18–24 months. This interim strategy will allow for substantially improved malaria treatment in DRC while providing additional time to assess the efficacy and use of alternatives to SP (including combination therapies). As importantly, this interim strategy will also provide time to identify sources of the levels of funding that would be required to implement most of these alternatives.
This study was conducted with the financial support of BASICS/United States Agency for International Development (USAID), the European Union (Programme d'Appui Transitoire à la Santé) and Médecins Sans Frontières France (MSF/F). Materials and drugs were tested and provided by the Centers for Disease Control and Prevention (CDC), Atlanta, USA and MSF/F. CDC staff also provided with a strong technical support. We thank mothers and children from the seven sites who accepted to participate in the studies and the field teams for their assistance.