Background. Fosmidomycin is a new antimalarial drug with a novel mechanism of action. Studies in Africa that have evaluated fosmidomycin as monotherapeutic agent demonstrated its excellent tolerance, but 3-timesdaily treatment regimens of ⩾4 days were required to achieve radical cure, prompting further research to identify and validate a suitable combination partner to enhance its efficacy.
Methods. We conducted a randomized, controlled, open-label study to evaluate the efficacy and safety of fosmidomycin combined with clindamycin (n = 12; 30 and 5mg/kg body weight every 12 h for 5 days, respectively), compared with fosmidomycin alone (n = 12; 30 mg/kg body weight every 12 h for 5 days) and clindamycin alone (n = 12; 5 mg/kg body weight every 12 h for 5 days) for the clearance of asymptomatic Plasmodium falciparum infections in schoolchildren in Gabon aged 7–14 years.
Results. Asexual parasites were rapidly cleared in children treated with fosmidomycin-clindamycin (median time, 18 h) and fosmidomycin alone (25 h) but slowly in children treated with clindamycin alone (71 h; P = .004). However, only treatment with fosmidomycin-clindamycin or clindamycin alone led to the radical elimination of asexual parasites as measured by day 14 and 28 cure rates of 100%. Asexual parasites reappeared by day 28 in 7 children who received fosmidomycin (day 14 cure rate, 92% [11/12; day 28 cure rate, 42% [5/12]). All regimens were well tolerated, and no serious adverse events occurred.
Conclusion. The combination of fosmidomycin and clindamycin is well tolerated and superior to either agent on its own with respect to the rapid and radical clearance of P. falciparum infections in African children.
The increasing resistance of Plasmodium falciparum to commonly used antimalarial drugs makes the development of new treatment options a prerequisite for continuing efforts at efforts controlling malaria . Fosmidomycin inhibits the DOXP pathway localized to the apicoplast of P. falciparum . Fosmidomycin as monotherapy has been shown to be very effective and well tolerated in previous studies, but late recrudescences with treatment regimens of 3 times daily for <4 days preclude its use as monotherapy [3, 4].
Clindamycin has been identified as a potential combination partner through systematic in vitro and animal studies, which conclusively showed the synergistic action of fosmidomycin and clindamycin . Additional strong lines of reasoning in favor of clindamycin as a combination partner include matching pharmacokinetic properties, no preexisting parasite resistance to clindamycin, and extensive experience with shortcourse combinations of clindamycin with quinine [6–8]. As an overall objective, we hypothesized that the addition of clindamycin to fosmidomycin would lead to a highly effective twice-daily regimen of short duration for the treatment of pediatric patients with uncomplicated falciparum malaria in Africa.
As a first step toward the outlined development objective, we conducted a phase 2a study that was designed to establish the efficacy and safety of a 5-day regimen of the combination of fosmidomycin and clindamycin, compared with fosmidomycin or clindamycin alone. The efficacy of 5-day regimens of fosmidomycin or clindamycin alone for the treatment of uncomplicated P. falciparum malaria has been previously established [3, 9], but clindamycin alone is not recommended for the treatment of P. falciparum malaria because of its slow onset of action . Therefore, we chose to study the efficacy of the combination of fosmidomycin and clindamycin in comparison to the single agents in asymptomatically infected schoolchildren aged 7–14 years.
We know from previous studies in the study area [10, 11] that the majority of asymptomatic infections will eventually develop into malarial attacks. The interventions in the present study were therefore judged to confer a benefit to these children. The study cohort of children aged 7–14 years is at the upper range of the most important target group for new antimalarial drugs: young children in areas of high transmission in Africa.
Materials and Methods
Study area. The study took place in the town of Lambaréné, Gabon, an area that is characterized by high perennial transmission of P. falciparum and an entomological inoculation rate of ∼50 infective bites/person/year [12, 13]. The study protocol was approved by the ethics committee of the International Foundation of the Albert Schweitzer Hospital, Lambaréné.
Study design and end points. The study was designed as a randomized, controlled, open-label phase 2a study to evaluate the efficacy and safety of fosmidomycin and clindamycin, compared with fosmidomycin or clindamycin alone, for the clearance of asymptomatic P. falciparum infections in schoolchildren in Gabon aged 7–14 years.
Efficacy end points were the parasite clearance times and the parasitological cure rates by days 7, 14, and 28, as defined by the absence of asexual parasites on the peripheral thick blood smear (polymerase chain reaction [PCR]-corrected results for day 28). The safety end point was the incidence of adverse events during the study period.
Enrollment and treatment of patients. The study took place in the Primary School of Lalala, Lambaréné, Gabon. This is a place where numerous clinical studies have been carried out in the past [14–16], and a relationship of mutual trust has been successfully maintained over the years. We explained the study to the schoolchildren in their classrooms and distributed the informed consent form to all children, who were asked to return the form when it had been signed by a parent. Parents were repeatedly encouraged by schoolteachers and by us to approach the study team in case of questions. Schoolchildren were then screened for asymptomatic P. falciparum infection by means of a thick blood smear from peripheral blood obtained by a finger prick. We stopped the screening once 36 eligible schoolchildren were identified. The schoolchildren were subsequently given consecutive study numbers and randomly assigned by the investigators (S.B., P.-B.M., A.S., S.I., and A.A.A.) to oral regimens of either the combination of fosmidomycin and clindamycin (n = 12; 30 and 5 mg/kg body weight every 12 h for 5 days, respectively), fosmidomycin alone (n = 12; 30 mg/ kg body weight every 12 h for 5 days), or clindamycin alone (n = 12; 5 mg/kg body weight every 12 h for 5 days) if they fulfilled all of the following inclusion and exclusion criteria: female and male schoolchildren aged 7–14 years; inclusive, asymptomatic asexual P. falciparum parasitemia of 1200 parasites/mL; written, informed consent by the legal representative of the subject (preferably the parents) and oral agreement of the child; no intake of drugs with known activity against P. falciparum during the preceding week; no antibiotic treatment for a current infection; hemoglobin concentration ⩾7 g/dL, hematocrit ⩾25%, and leukocyte count ⩽12,000 cells/µL; no symptoms of malaria; no other severe underlying disease (cardiac, renal, hepatic diseases, malnutrition, or known HIV infection); and no other febrile illness. Treatment was initiated in <12 h after the detection of asymptomatic parasitemia on the same day.
Study drugs and administration. Study drugs were provided by Jomaa Pharmaka. Fosmidomycin was formulated as capsules that contained 150 mg of active substance. Clindamycin was formulated as granular powder in bottles. Each bottle contained 3.2 g of active substance. Before dosing, the powder was dissolved in 50 mL of drinking water. Dosing was performed using scoops that contained 37.5 mg of clindamycin. A dedicated study physician supervised the administration of study medication either at school or at home, depending on the time of day, and children who vomited or spit out the study drug within 30 min received a second full dose. According to the study protocol, vomiting of the second dose would have led to withdrawal and the administration of a suitable rescue treatment.
Study flow and procedures. The counting of study days started with the first dose on day 0. Children were seen by a study physician every 12 h during the treatment period (days 0–4) and again on days 7, 14, 21, and 28, unless otherwise indicated. On each visit during the treatment and follow-up phases, the medical history was obtained, vital signs (blood pressure and pulse) were checked, tympanic temperature was measured, a thick blood smear was prepared from a finger prick (during the treatment period until 2 consecutive negative smear results were obtained), and starts and ends of adverse events— defined as any new or worsening condition not present at baseline— were documented and graded according to predefined scales of seriousness, severity, causality, and course of action. Venipunctures were performed on study days 0, 2, 7, and 28, to monitor hematologic parameters (hemoglobin, hematocrit, and differential white blood cell count), as well as parameters for liver (alanine aminotransferase [ALT] and total bilirubin) and renal (creatinine) function. In addition, standard urine dipstick tests were done at the same time intervals. To distinguish recrudescent infections from reinfections, EDTA-preserved blood samples from day 0 and the day of reappearance of asexual parasites were kept at −80°C until transport for PCRbased genotyping analysis.
Laboratory testing. The dried thick blood smears were stained with 20% Giemsa solution for 20 min (pH 7.2; Sigma). Parasite species were identified using standard morphological characteristics, and the parasitemia level was determined using the volumetric Lambaréné method and expressed as parasites per microliter [17, 18]. Slides from admission and the last follow-up visit were double checked by a second microscopist. Hematologic parameters were assessed with an automatic analyzer (Celldyn 3000; Abbott), and clinical chemistry was assessed by a semiautomatic dry-slide method (Vitros; Ortho- Clinical Diagnostics). Genotyping of the polymorphic P. falciparum merozoite surface antigens 1 and 2 was performed to distinguish recrudescences from new infections and compared matched pairs of parasite isolates obtained on admission and on the day of reappearing parasitemia [10, 19]. We classified reappearing parasitemia as reinfection if all electrophoretically separated PCR product bands detected on the day of reappearing parasitemia were distinct from those detected on the day of admission (the samples were tested in parallel).
Randomization, sample-size calculation, and statistical analysis. The randomization code was generated by computer in blocks of 9 by one of us (S.B.), and the resulting randomization list was used to allocate the schoolchildren to the 3 treatment regimens according to study number. The sample-size calculation for the pilot study was based on the following assumptions: parasite clearance time with clindamycin, fosmidomycin, and the combination of both of 120 , 49 , and 46 (our estimation) h, respectively. To see a significant difference in the parasite clearance time of 60 h (SD, 30 h) between the clindamycin group on the one hand and the fosmidomycin or combination group on the other hand, 8 patients in each treatment group were needed at a 1% significance level with a power of 90%. To allow for children being lost to follow-up, 12 patients were recruited into each group.
Data were double entered into 2 databases (FileMaker Pro 5.5 for Windows; FileMaker), and discrepancies were resolved by a third person during a subsequent blind verification. Data were analyzed using Stata Statistical Software (release 8 for Mac OS X; StataCorp).
Descriptive and categorical baseline data were summarized in a table. Differences in the times to parasite clearance between treatment groups were evaluated with the Kruskal-Wallis equality of populations rank test. Cure rates were calculated from the number of patients who had clearance of asexual parasitemia within 7 days of beginning treatment and who did not have subsequent recrudescence until day 28, divided by the total number of patients with a known efficacy outcome (per protocol population). Fisher's exact test was used to assess differences in rates between treatment groups. Exact binomial confidence intervals (CIs) were also calculated. For the calculation of proportions of gametocyte carrier status, we considered thick blood smears from day 0 (hours 0 and 12) and those of subsequent study days for the determination of the treatment- associated gametocyte carrier rate (emerging gametocytemia after day 0 in children without gametocytes at initiation of treatment). The safety analysis used abnormal laboratory data and adverse events from all subjects who received at least 1 dose of the study drug (the intent-to-treat population was the safety population). Only clinical adverse events occurring during the first 7 days after the start of treatment (2 days after the last dose) were analyzed in detail, to allow an appreciation of the temporal association of adverse events with the administration of study drugs. Incidence rates were compared with an exact test . Differences of changes in the laboratory parameters over time across the treatment groups were assessed using repeated-measures analysis of variance, with Box's conservative ε correction for the repeated measures variable (study day). All P values are reported as 2-sided values. Differences between groups were considered to be significant at P < .05.
Profile of study cohorts. Thirty-six children were enrolled into the study from October 2001 until January 2002; 12 each were treated with the combination of fosmidomycin and clindamycin, fosmidomycin alone, or clindamycin alone. Figure 1 shows the profile of the study cohorts. Two children were wrongly included: 1 child in the fosmidomycin-clindamycin group had a false-positive thick blood smear result, and another child in the clindamycin-alone group had only a single infection with P. malariae and was therefore promptly treated with chloroquine. We did not withdraw a subject because of failure to administer a study drug dose. In total, 26 children completed follow-up up to day 28 (11, 6, and 9 children in the fosmidomycin- clindamycin, fosmidomycin alone, and clindamycin alone groups, respectively). Six participants were prematurely withdrawn from the study because of treatment failure before day 28 (all 6 in the fosmidomycin-alone group). Two study participants in the clindamycin-alone group were excluded because of protocol violations (1 child was treated with an unknown dose of chloroquine at home on day 1, and another child received an unknown quantity of halofantrine on day 18). Thus, in total, 32 children had an evaluable efficacy outcome by day 28 (11, 12, and 9, respectively, in the fosmidomycin-clindamycin, fosmidomycin alone, and clindamycin alone groups).
Baseline characteristics and treatments. All treatment groups had similar baseline characteristics except for a nonsignificant difference in sex distribution (P = .084, Fisher's exact test). Details of baseline parameters are summarized in table 1. All children were free of malarial symptoms. There were 2 children with mixed infections of P. falciparum and P. malariae on admission, but both children were retained in the study and in the efficacy population (1 child each in the fosmidomycin-alone and clindamycin-alone groups).
Mean doses of fosmidomycin and clindamycin in the fosmidomycin- clindamycin combination group were 30.1 (SD, 1.5; range, 28.1–32.6) and 5.0 (0.3; 4.7–5.4) mg/kg of body weight, respectively. Mean fosmidomycin and clindamycin doses in the fosmidomycin-alone and clindamycin-alone groups were similar, with a mean value of 30.5 mg/kg of body weight for fosmidomycin (1.5; 27.8–33.2) and of 5.0 mg/kg of body weight for clindamycin (0.2; 4.7–5.2), respectively. Median dosing intervals were 12.1 (range, 10.0–14.0 h), 12.0 (range, 9.1–14.1 h), and 12.1 (range, 9.8–14.6 h) h for fosmidomycin-clindamycin, fosmidomycin alone, and clindamycin alone, respectively. One child in the clindamycin group was underdosed (4 scoops of clindamycin instead of the correct 5 scoops, resulting in a dose of 4.3 mg/kg of body weight). She was withdrawn from the study for another reason on day 1 (intake of chloroquine). Vomiting of study medication occurred once in 1 patient who received clindamycin alone (first dose of a total of 10), but the repeated administration 45 min later was well tolerated.
Efficacy.Table 2 summarizes the parasitological responses to the study drugs. Fosmidomycin-clindamycin and fosmidomycin alone led to the rapid elimination of asexual parasites from the peripheral blood, with median times to parasite clearance of 18 (range, 12–47 h) and 25 (range 13–46 h) h, respectively, but not clindamycin alone (median, 71 h; range, 13–85 h) (P = .004). In fact, clindamycin did not lead to a steady decline of parasitemia, and levels of asexual parasites resurged on day 2 (median parasite density, 550 parasites/µL), up from an initial trough by day 1 (median parasite density, 66 parasites/µL). By day 7, all patients were free of parasites. Parasites reappeared in 1 child in the fosmidomycin-clindamycin group (at day 28) and in 7 children in the fosmidomycin-alone group (1 by day 14, 5 by day 21, and 1 by day 28), leading to PCRuncorrected cure rates by day 14 of 100% (11/11; 95% CI, 72–100), 92% (11/12; 95% CI, 62–100), and 100% (10/10; 95% CI, 69–100) (P = 1.0) and by day 28 of 91% (10/11; 95% CI, 59–100), 42% (5/12; 95% CI, 15–72), and 100% (9/9; 95% CI, 66–100) in the fosmidomycin-clindamycin, fosmidomycin, and clindamycin groups, respectively (P = .003). Parasite genotype analysis revealed that 1 child treated with fosmidomycin-clindamycin carried different parasite strains at days 0 and 28. Thus, the PCR-corrected day 28 cure rates were 11 of 11 (100%; 95% CI, 72–100), 5 of 12 (42%; 95% CI, 15–72), and 9 of 9 (100%; 95% CI, 66–100) with fosmidomycin-clindamycin, fosmidomycin, and clindamycin, respectively (P < .001). We did not detect a significant relationship between initial parasite density and subsequent risk of failure by day 28 in a logistic regression model (P = .71; treatment group as an additional variable was dropped because of collinearity). Anecdotally, the 2 children who had mixed infection at baseline remained free of P. malariae parasites throughout the follow-up period, until day 28. Subjects with reappearing parasitemia received a single dose of sulfadoxine-pyrimethamine (25 and 1.25 mg/kg of body weight, respectively).
An analysis of gametocytemia detected on at least 1 time point during the follow-up period showed a higher treatmentassociated gametocyte carrier rate in the fosmidomycin-alone group (6/12 [50%]), compared with the fosmidomycin-clindamycin (3/11 [27%]) and clindamycin-alone groups (3/10 [30%]), but the difference was not significant (P = .46). Treatment- associated gametocyte carrier status seemed to be associated with a failure of treatments to radically eliminate the parasites (57% of children with subsequent failures developed gametocytemia after the start of treatment vs. only 28% of children with radical cure), but this association equally lacked significance (P = .2). If children with gametocytes at baseline were considered, we could not detect a gametocytocidal effect because of the small sample size. In total, there were only 3 children with baseline gametocytemia: 1 child in the fosmidomycin group (not cleared) and 2 in the clindamycin group (1 was cleared and the other was not). Gametocyte densities in carriers after day 0 were 6–60, 12–170, and 6–30 gametocytes/µL in the fosmidomycin-clindamycin, fosmidomycin-alone, and clindamycin-alone groups, respectively.
Safety and tolerability. Median follow-up times (times at risk of the safety population) were similar for all treatment groups, with values of 28 (range, 25–28), 23 (range, 14–29), and 27 (range, 21–28) days, resulting in total follow-up times of 304, 284, and 267 person-days, respectively, for fosmidomycin-clindamycin, fosmidomycin alone, and clindamycin alone (P = .14). All treatment regimens were well tolerated, and no serious adverse event was reported. In addition, there was no treatment-limiting adverse event. There were a total of 46 clinical adverse events (38 mild and only 8 moderately adverse events), but the majority of these occurred until day 7 (35 adverse events; 31 graded as mild and 4 as moderate). Until day 7, 11 mild and 2 moderate (1 case of vomiting and 1 case of dizziness) adverse events occurred in the fosmidomycinclindamycin group, 5mild and 2 moderate (1 case of abdominal pain and 1 case of cough) adverse events in the fosmidomycinalone group, and 15 mild adverse events in the clindamycinalone group (table 3). Gastrointestinal adverse events (8 cases of abdominal pain, 7 cases of diarrhea, 4 cases of vomiting, 2 cases of loose stool, and 2 cases of nausea) were the most frequently reported adverse events until day 7 (table 3). All gastrointestinal adverse events were judged to be possibly or probably related to the treatment. Nonetheless, all gastrointestinal adverse events were self-limiting, and no intervention was required. There was no difference in the incidence of gastrointestinal adverse events between the fosmidomycin-clindamycin and clindamycin-alone groups (9 cases in each of these treatment groups until day 7), whereas treatment with fosmidomycinclindamycin and clindamycin alone led to a nonsignificantly higher incidence rate of gastrointestinal adverse events until day 7 than treatment with fosmidomycin alone (0.12; 9 cases/78 person-days and 0.12; 9 cases/72 person-days, respectively, vs. 0.06; 5 cases/84 person-days; P = .24 and P = .19, respectively, for the comparisons).
There were no significant changes in hemoglobin concentrations (P = .80 for the comparison between groups and P = .14 for the comparison within groups), leukocyte counts (P = .67 and P = .51, respectively), or ALT levels (P = .84 and P = .69, respectively) over the study period (days 0, 2, 7, and 28). There were 2 laboratory adverse events: 2 children with abnormally elevated ALT values after start of treatment (1 child in the fosmidomycin alone group who had a value of 127 U/ L on day 2, which decreased to a normal 7 U/L by day 28, and 1 child in the fosmidomycin-clindamycin group who had an unexplained increase to 195 U/L on day 28, up from 30 U/L on day 7). The latter subject could not been followed up beyond day 28 because of travelling away from the study area (during the Christmas holidays). The urine dipstick tests did not reveal any abnormality during the study period.
The present study was primarily designed to test the working hypothesis that the combination of fosmidomycin with clindamycin can benefit from the distinct pharmacodynamic properties of both drugs while at the same time preserving the advantageous safety profiles of the single agents. Our results demonstrate that a 5-day twice-daily regimen of fosmidomycinclindamycin is well tolerated and effective for the clearance of asymptomatic P. falciparum infections in African children aged 7–14 years. Its efficacy exceeded the singly administered components of the combination by eliminating both rapidly and radically the asexual parasites from the body, as demonstrated by a median time to asexual parasite clearance of 18 h and a 100% cure rate by day 28. Fosmidomycin alone also led to a rapid median parasite clearance of 25 h, but it did not radically eliminate the parasites, which became apparent by a low day- 28 cure rate of only 42%. Clindamycin administered alone led to a protracted median asexual parasite clearance time. On the other hand, clindamycin did eventually clear all parasites, as determined by a 100% cure rate by day 28.
The efficacy of fosmidomycin alone, as determined by day 14, was comparable to a previously reported 5-day regimen for the treatment of uncomplicated P. falciparum malaria in adult patients in Gabon (89% [8/9] vs. 92% [11/12]) . However, in the present study, most failures occurred after day 14, and the cure rate decreased to 42% (5/12) by day 28, which is well below the rate of the earlier study. The main share of this difference can be attributed to the increased sensitivity of the day- 28 cure rate as an assessment of the radical elimination of asexual parasites. On the other hand, the higher degree of parasitespecific immunity in the patients in the former study might have prevented a similar decrease in the cure rate by day 28, but this remains speculation. It is impossible to conclude fromthe present data how the reduction of the frequency of administration influenced the present results or whether such a reduction had any effect at all (treatment with fosmidomycin alone was every 8 h in the previous study but only every 12 h in the present study). Also, care must be taken not to overinterpret the data, owing to the very small sample sizes and correspondingly extremely wide confidence intervals, as well as with regard to the inherent difficulties in comparing efficacy data from asymptomatic P. falciparum infections with data from malaria patients.
All treatment regimens were associated with gametocyte carrier status in children who had no gametocytes at baseline, but this effect was seen primarily in the group of children treated with fosmidomycin alone (50% vs. 27% and 30% with fosmidomycinclindamycin and clindamycin alone, respectively), which could partly be explained by the lack of radical elimination of asexual parasites in some of these children, as has been shown elsewhere for other antimalarials . Similar findings have also been reported for fosmidomycin alone [3, 4].
All treatment regimens were safe and well tolerated. Slight differences in the incidence rates of gastrointestinal adverse events judged to be related to the study medications between fosmidomycin-clindamycin and fosmidomycin did not reach statistical significance, and all gastrointestinal adverse events were self-limiting and did not require an intervention. The occurrence of mild cases of diarrhea in previously healthy children infected with P. falciparum is worrying and needs to be interpreted in light of the potential antibacterial effects of clindamycin  and, possibly, fosmidomycin on the intestinal flora. In anticipation of such adverse events, we specifically questioned the children and their parents at each study visit to ascertain gastrointestinal adverse events, which could have introduced a bias toward an increased reporting of gastrointestinal adverse events. Conversely, the antimalarial action of the study drugs accelerates the release of parasite antigens, resulting in an intensified immune response. The ensuing symptoms (in our case reported as adverse events) can mimic an acute malarial attack. Our adverse event assessment reflects this uncertainty in terms of the assignment of relationships to the study drugs. Undoubtedly, rigorous monitoring in ongoing and future studies is required to see whether the addition of clindamycin to fosmidomycin does indeed reduce its tolerability.
In analogy with a previous study with fosmidomycin alone in adult patients in Gabon who had uncomplicated falciparum malaria , we detected 1 patient who had a slightly elevated ALT value after treatment with fosmidomycin-clindamycin. In addition, another subject with a slightly elevated ALT value occurred in the clindamycin-alone group, which raises the question of potential liver toxicity caused by the combination of fosmidomycin with clindamycin. However, no sign of kidney toxicity was seen in the present study, as opposed to the previous study, in which 1 instance of increased serum creatinine levels were noted .
In summary, the combination of fosmidomycin and clindamycin bears great potential in terms of high efficacy against malaria and good tolerability. Future studies are now focusing on the effects of shorter regimens for the treatment of pediatric patients with P. falciparum malaria.
We are particularly indebted to the participating children and their parents. We thank Bertrand Lell and the anonymous reviewers of the manuscript for their helpful comments.