Dr D. S. Tarimo Department of International Health, Institute of Public Health, University of Copenhagen, Panum Institute, Blegdamsvej 3, 2200 Copenhagen N, Denmark. Tel.: +45 35 32 7495; Fax: +45 35 32 7736; E-mail: firstname.lastname@example.org
Following widespread chloroquine (CQ) resistance, sulfadoxine plus pyrimethamine (SP) is now the first line antimalarial drug in a number of African countries including Tanzania. Unlike CQ, SP has no antipyretic effects, a feature that might delay fever clearance, and by acting on late stage parasites, SP could theoretical be slow in parasite clearance. We therefore assessed the antipyretic effects of CQ in therapeutic combination with SP, and the speed of parasite clearance by SP in an open-labelled, randomized trial of CQ alone (n=39), SP alone (n=39), SP plus CQ (n=37) and SP plus paracetamol (PCM) (n=38) in children with uncomplicated malaria. Over 72 h, there were eight (20.5%) treatment failures in the CQ group but none in the other groups. Although not significant (P > 0.1), irrespective of resistance CQ alone had a shorter median survival time to fever clearance than SP alone (54 vs. 60 h). SP plus CQ had a highly significantly shorter median survival time to fever clearance than SP alone (48 vs. 60 h) (P < 0.001). Although borderline (P=0.038), the median survival time to parasite clearance was significantly longer in SP plus PCM (72 h) than SP alone (48 h). Irrespective of resistance, CQ alone had a median survival time to parasite clearance equal to that of SP alone (48 h). Parasite clearance by SP was rapid and at the end of 72 h, most (77.3%, 95% confidence interval: 69.6, 85.0) of the children on SP (as a group) had become aparasitaemic. The findings rule out concerns on possible delayed parasitological and clinical responses to SP that could result from its action on late stage parasites. Despite its diminishing antimalarial activity, CQ has beneficial in vivo antipyretic effects in therapeutic combination with SP.
The increase in childhood mortality attributable to the escalating Plasmodium falciparum resistance to chloroquine (CQ) (Trape et al. 1998) has compelled a number of African countries such as Kenya, Malawi, Botswana and South Africa to change to the fixed combination of sulfadoxine and pyrimethamine (SP) (White et al. 1999) as the first line antimalarial. In Tanzania, the prevalence of P. falciparum resistance to CQ has surpassed the upper threshold (≥ 25%) for replacement, but for SP resistance is still ≤ 10%, and CQ is being replaced with SP (WHO 1996; Ministry of Health Tanzania 2000).
Chloroquine has been very popular in the treatment of fevers over the last 50 years because of its rapid schizontocidal action as well as its anti-inflammatory and antipyretic effects (Rollo 1971; Warrell 1993). Although CQ was developed primarily as an antimalarial drug, its anti-inflammatory property is well known and its usefulness in the treatment of rheumatoid arthritis is established (Rollo 1971). In vitro evidence shows that the antipyretic effect of CQ is mediated through the downregulation of inflammatory cytokines following schizonts' rupture (Picot et al. 1993; van de Borne et al. 1997).
The adoption of SP as the first line drug raises theoretical concerns that, by acting on the late trophozoite and schizont stages of the parasite (White & Krishna 1989; White 1992), SP will give slow clinical responses (White 1999), and the absence of anti-inflammatory and antipyretic effects (Warrell 1993) could potentially lead to slow fever clearance which might negatively influence satisfaction and acceptability of SP as the first line drug by individuals, families and communities (Williams et al. 1999). An evaluation based on the absence of a health seeking event after treatment showed that the combination of SP with CQ gives significantly better symptomatic relief than SP alone (Bojang et al. 1998; McIntosh & Greenwood 1998).
This study assessed the antipyretic effects of CQ in therapeutic combination with SP among children with uncomplicated malaria and examined the theoretical concerns on delayed clinical and parasitological responses to SP as a result of its action on the late stage parasites (White 1999). The objectives were: (1) to systematically and prospectively assess the in vivo antipyretic effects of CQ in therapeutic combination with SP; (2) to prospectively evaluate the hypothesis that, by acting on the late stages of the parasites, SP will be slow in parasite clearance thus delaying the desired clinical effects; and (3) to prospectively evaluate whether paracetamol (PCM) antipyresis is superior to CQ antipyresis.
Patients and methods
The study was conducted from May to August 1998 at Kibaha Hospital in the holoendemic Kibaha district, coastal Tanzania. Children 12–59 months old attending the mother and child health (MCH) clinic with fever were enrolled using the WHO criteria (WHO 1996). Briefly, haemoglobin was estimated by the Haemocue, and parasitaemia assessed by Giemsa stained thick and thin films. Parasites were counted against 200 WBC, multiplied by 40 and converted into count per microlitre assuming a mean count of 8000 WBC/μl of blood. Strict inclusion criteria were adopted. Children were selected if they had axillary temperature ≥ 37.5 °C, haemoglobin level > 5 g/dl, mono P. falciparum infection 2000–250 000 parasites/μl, if they were not severely sick, not malnourished, and without other coexisting diseases. Selected children were enrolled if parents or guardians gave informed verbal consent to participate in the study.
Sample size estimation
The sample size was estimated for a qualitative outcome measure as described by Pocock (1983). The main purpose was to see if the antipyretic effect of CQ is clinically beneficial in therapeutic combination with SP vs. SP monotherapy, synergy in parasite clearance was also relevant but not the prime purpose. The main outcome measure was fever clearance over the crucial period of 72 h after initiation of therapy, and treatment outcome was classified as fever cleared (survival) or not cleared (failure). A power calculation based on survival analysis and the logrank test was performed on the basis that all patients had fever (temperature ≥37.5 °C) at time 0 (before treatment). Assuming at time t, after treatment, the proportion of children who would have cleared fever in SP alone was 20%, and in SP plus CQ was 50%, we calculated a two-sided test so as to detect a 30% difference in survival proportion between two treatments at a significance level of 0.05 and power=0.8. This resulted in a minimum required sample size of 36 patients in each treatment group.
Study design and treatment assignment
A randomized, open-labelled comparison was performed on four treatment regimens containing CQ alone (150 mg tablets of CQ phosphate, Helm Pharmaceuticals GMbH, Hamburg, Germany), SP alone (Fansidar®, F. Hoffman – LaRoche, Basel, Switzerland: 25 mg pyrimethamine and 500 mg sulfadoxine), SP plus CQ and SP plus PCM. Treatment groups were labelled alphabetically as A (CQ), B (SP), C (SP and CQ) and D (SP and PCM). Patients were consecutively enrolled and admitted to the study and hospitalized for 72 h during the acute stage of illness. On each day of enrolment, the first patient was allocated to a treatment group by the spin of a coin, and then consecutive patients were sequentially allocated to receive either of the treatments in alphabetical order. Doses were calculated according to the WHO standard doses for infants and children (WHO 1996). As inclusion criteria permitted to enrol only children with uncomplicated malaria, selection bias on the basis of severity was unlikely. Treatments were administered to each child under supervision, and if vomiting occurred within 30 min, the same dose was repeated. The children were closely observed for the timely detection of early treatment failures (defined as positive smear with parasite density > day 0 count on day 2 with fever, or positive smear of any density on day 3 with fever, or positive smear on day 3 ≥ 25% of day 0 count) in which case alternative treatment was given accordingly (WHO 1996). Axillary temperature (digital thermometer) and other vital signs were recorded every 6 h. Parasitaemia was normally assessed every 24 h, and at any other time as the clinical condition of the child dictated. A patient was declared aparasitaemic 24 h after the last positive film (White & Krishna 1989). The children were discharged after 72 h to continue with routine outpatient follow-up.
The study was approved by the Commission for Science and Technology in Tanzania and ethically cleared by the Tanzanian Ministry of Health. Verbal informed consent for participation was obtained from parents and guardians after thorough information on the study was provided in the local language. Temperature was closely monitored, giving paracetamol every 8 h to the children on SP plus PCM, and giving mechanical antipyresis to children in the other arms to keep their temperature at 38.5 °C (WHO 1996). Although resistance to CQ was already very high, CQ was still the first line drug, but in the presence of very strict monitoring for early treatment failures.
Two experienced technicians blindly (unaware of patients' status) did microscopy independently, each time comparing their results. At enrolment and thereafter, a blood film was pronounced negative when the examination of 100 fields was negative. The principal investigator (D.S.T.) closely supervised the study team to ensure consistency and accuracy of the data.
The data were cleaned, entered into SPSS version 10.0, checked for consistency and analysed. Survival analyses of (i) time to fever clearance (from initiation of therapy to the first four consecutive six hourly normal temperature readings, corresponding to an afebrile period of 24 h) and (ii) time to parasite clearance (from initiation of therapy to the first negative smear 24 h after the last positive smear) were performed to compare the proportion of children cleared of fever and the speed of parasite clearance. To assess the 72-h parasite clearance rates, parasitaemia status (+/–) and geometric mean parasites densities were compared among treatments. Proportions were compared by chi-square test and means by the one-way ANOVA. Survival proportions were compared by the Kaplan–Meier curves (Altman 1994) and the difference in survival times assessed by the Mantel–Cox logrank test in a pairwise comparison of treatments (Clayton & Hills 1998). Significance was set at 0.05 level.
A total of 153 children were enrolled and evaluated clinically and parasitologically over the crucial period of 72 h of inpatient observation. There were no deaths or severe morbidity or complications. The baseline demographic, clinical and laboratory data were compared for the different treatment regimens (Table 1). There were no significant differences between CQ alone, Fansidar (SP) alone, Fansidar plus CQ (SP plus CQ) and Fansidar plus paracetamol (SP plus PCM) regimens with respect to the distribution of baseline attributes.
Table 1. Baseline characteristics of the children enrolled with uncomplicated malaria (n=154)
The clinical responses (Table 2) showed that over the crucial period of 72 h, there were eight (20.5%) early treatment failures in children treated with CQ alone. There were no treatment failures in the SP alone, SP plus CQ or SP plus PCM regimens.
Table 2. Early treatment failure at 72 h by treatment regimen among children with uncomplicated malaria
Figure 1 shows the survival (Kaplan–Meier) curves of the time to fever clearance by the four treatment regimens. The median survival times to fever clearance were: CQ alone, 54 h; SP alone, 60 h; SP plus CQ, 48 h; and SP plus PCM, 48 h. A pairwise comparison of survival times showed that SP plus CQ, and SP plus PCM had equal median survival times to fever clearance. Although non-significant (P > 0.1), irrespective of resistance CQ alone had a shorter median survival time to fever clearance than SP alone. SP plus CQ and PCM had a highly significantly shorter median survival time to fever clearance than SP alone (P < 0.001). Because of the high level of resistance to CQ, SP plus CQ had a highly significantly shorter median survival time to fever clearance than CQ alone (P=0.002).
Survival analysis of the time to parasite clearance by the four treatment regimens over the crucial period of 72 h showed that the median survival times to parasite clearance were: on CQ alone, 48 h; SP alone, 48 h; SP plus CQ, 72 h; and SP plus PCM, 72 h. A pairwise comparison showed that although of borderline significance (P=0.038), the median survival time to parasite clearance was significantly longer in SP plus PCM than SP alone. Irrespective of resistance, CQ alone had a median survival time to parasite clearance equal to that of SP alone.
The geometric mean parasite densities (GMPDs) among treatment regimens were dropping at the same rate (Figure 2) and there was no significant difference in the GMPDs at the different points of observation (P > 0.05). The treatments showed the same rate of reversion to negative parasitaemia status (Figure 3), and there was no significant difference in the percentage uncleared at the different observation points (P > 0.05). On average, 77.3% (95% CI: 69.6, 85.0) of the children in the SP group (combined) had reverted to aparasitaemia after 72 h.
Chloroquine has been very popular in the treatment of malarial fevers because of its anti-inflammatory and antipyretic effects (Rollo 1971; Warrell 1993). The findings show that, in vivo, the combination therapy of SP plus CQ produces a faster resolution of fever than SP monotherapy. The absence of a significant difference in the median survival times to parasite clearance between SP plus CQ combination therapy and SP monotherapy indicates that the significantly faster fever clearance in SP + CQ is because of the antipyretic effects of CQ. The findings provide an evidence of in vivo antipyretic effects of CQ in therapeutic combination with SP confirming an earlier evaluation based on the absence of a health seeking event after treatment with SP alone or SP plus CQ, which showed that the combination gave significantly better symptomatic relief than SP alone (Bojang et al. 1998; McIntosh & Greenwood 1998).
These observations have important implications on the acceptance of SP as a first line drug. In the implementation of policy change from CQ to SP, attention should be given to the practical aspects of community acceptance of SP as a replacement of CQ. Fever control is important as it brings comfort, reduces water loss by perspiration and makes the child less anorexic with increased fluid and food intake. Fever, level of activity and fluid/food intake are the proximate measures that parents use to assess wellness and efficacy of therapy. The therapeutic combination of SP plus CQ produces rapid and dramatic symptomatic relief which might affect perceptions of wellness and efficacy in a positive direction. In the absence of this, parents may perceive that SP is not working despite the documented high biological efficacy. Parents may therefore be reluctant to use SP as a first line drug and continue to use CQ even in the face of a high decline in biological efficacy, or seek alternative/additional therapy that might delay attainment of appropriate and effective therapy.
Thus, in the implementation of policy change to SP, the Ministry of Health should provide training to health workers on the pharmacological differences between SP and CQ focusing on the length of time SP takes to show clinical improvement, approaches to fever control and comfort measures that can be used at home such as regular use of antipyretics and tepid baths. Health workers should be trained to offer specific counselling to parents on what they should expect regarding the course of the illness after SP therapy. Parents rely on fever as a primary sign for recognizing malarial fevers in their children and for evaluating whether the medication is working (Tarimo et al. 2000), so in the absence of adequate counselling, persistence of symptoms might be perceived as a sign of medication failure.
Our data do not support concerns on possible delayed parasitological and therefore clinical responses to SP (Warrell 1993; White 1999) as a result of its action on the late stage parasites (White & Krishna 1989; White 1992) that could conceivably make SP a slow-acting drug. Thus, at the end of 72 h, a significantly high percentage (77.3%) of the children in the SP group (combined) had cleared parasitaemia. Moreover, the GMPDs were dropping rapidly both for the SP and CQ groups, and there was no significant difference in the GMPDs at the different observation points.
The outcome of antimalarial treatment depends on parasite clearance (White & Krishna 1989). Our data show that, although of borderline significance (P=0.038), the median survival time to parasite clearance was significantly longer in SP plus PCM than SP alone confirming the earlier finding that although PCM has beneficial antipyretic effects, it has the disadvantage of prolonging parasite clearance time for almost 1 day (Brendit et al. 1997). Whether this also prolongs morbidity could not be ruled out in this study because of the small sample size. Moreover, PCM antipyresis was not superior to mechanical antipyresis (Brendit et al. 1997) and did not influence clients' satisfaction with SP treatment (Williams et al. 1999) suggesting that CQ could replace PCM in combination therapy with SP for the purpose of symptoms relief.
Even when CQ is only partially effective and SP fully effective, the two drugs had equal median survival time to parasite clearance suggesting a possibility of synergy between them (White 1999). This is plausible because the early parasite stage specificity of action by CQ is synergistic to the late stage action of SP (White & Krishna 1989). Such a synergy could delay the emergence of parasites resistant to SP (White & Olliaro 1996). The therapeutic combination of SP with CQ could therefore be one of the strategies to delay the development and spread of resistance to SP (White et al. 1999).
The findings have ruled out concerns on possible delayed parasitological and clinical responses to SP that could result from its action on late stage parasites, and that, despite its diminishing antimalarial activity, CQ has beneficial in vivo antipyretic effects in therapeutic combination with SP, and a theoretical possibility of synergy based on their parasite stage specificity of action that might conceivably delay the spread of resistance to SP.
This work was supported by DANIDA/ENRECA under the Tanzanian–Danish Collaborative Research and Training Programme, Phase III 1998–2001, Project III and partly by SAREC under the Muhimbili–Karolinska Collaboration. We thank the Permanent Secretary, Ministry of Health, Tanzania, for approving the study. Special thanks go to Mr Frederick Kalokola, for tirelessly handling the laboratory work in Tanzania. We are grateful to the mothers and staff of Kibaha Hospital for their co-operation in this study. This work is dedicated to the late Jeremiah Masunga (C.O.) for working with us tirelessly and making this work a success.