Because the indication for prophylaxis is to prevent disease in nonimmune travelers to all endemic regions, trials should be performed in persons in several regions of the world who are completely inexperienced with (that is, are not immune to) malaria. The clinical end point should be malaria (parasitemia plus symptoms) rather than parasitemia alone. We performed a randomized, double-blind, placebo-controlled clinical trial of primaquine prophylaxis against clinical malaria in nonimmune Colombian soldiers on patrol in a region in which P. falciparum and P. vivax malaria is endemic.

Methods

Top Methods Results Discussion Author & Article Info References

The study was performed between 18 April and 19 September 1997 in the province of Uraba, Colombia. Participants were 176 soldiers in the Colombian army. On specific questioning, all participants denied previous exposure to a malaria-endemic area.

Eligible participants were male; were 18 to 42 years of age; did not have important medical problems as determined by history, physical examination, and routine laboratory tests (complete blood count and measurement of blood urea nitrogen, glutamic oxalacetic aminotransferase, and methemoglobin); and were not G6PD deficient (G6PDH Diagnostic Test, Sigma, St. Louis, Missouri).

The study was conducted according to the regulations of the Colombian Army and was approved by the institutional review board of the Universidad Militar Nueva Granada. Written informed consent was required from all participants.

Drug Administration

Patients were randomly allocated to receive prophylaxis with primaquine diphosphate (two tablets of 15-mg base each, for a total of 30 mg/d) or matching placebo (two tablets) in a 2:1 allocation. Primaquine (Parmamed, Ltd., Malta) and placebo (Sanofi-Winthrop, New York, New York) were identical-looking brown-orange tablets obtained by the Colombian Ministry of Health. Investigators and participants were unaware of the randomization code and remained blinded to treatment group assignment throughout the study.

On 16 May 1997, participants began 15 weeks of patrol in Mutapa, a malaria-endemic region of Uraba province, after which they returned to barracks duty in Carepa, a nonmalarial area. Prophylaxis (two tablets) was taken daily from 15 May, the day before patrol, during the 15 weeks of patrol, and for 7 days after the return to Carepa. Colombian soldiers on patrol have access only to items issued by their commanders and therefore could not acquire or take unauthorized antimalarial agents.

Each day, immediately before a breakfast of coffee, bread, and, sometimes, an egg, the senior sergeant personally gave two tablets to each soldier from a bag labeled with the soldier's patient number and watched the soldier swallow the tablets. At the end of each week, the number of pills remaining in the bags was counted and recorded.

Determination of Infection

During the 15 weeks of exposure and for 4 weeks after the soldiers returned to the barracks, thick and thin smears were obtained weekly or sooner if participants had symptoms compatible with malaria infection. Giemsa-stained smears were read by a microscopist blinded to the symptom status of the participant. A smear was considered negative if no parasites were found in 200 fields of x1000 magnification. All positive smears were combined with two negative smears and were reread by a second microscopist blinded to participant status. Discordant readings were resolved by a third blinded reader in Bogota. There were no discordant readings for smears that resulted in study end points.

Diagnosis of Malaria

A participant was considered to have malaria if he had parasitemia of at least 1500 Plasmodia organisms/µL or lesser parasitemia with at least two symptoms compatible with malaria (fever, headache, myalgia, nausea, vomiting, diarrhea, or icterus).

Determination of Drug Toxicity

Each day, participants were asked, "Do you feel at all sick? Do you have any complaints?" If the answer was "no," questioning was stopped. If the answer was "yes," the complaints were recorded. Toxicity was graded as severe if the side effects of the drug caused the participant to withdraw from the trial, moderate if the complaints resulted in some impairment of normal duty, and mild if duty was not impaired. Laboratory tests other than G6PD were repeated during 2 to 4 weeks of follow-up.

End Points

Study end points were prophylaxis failure, indicated by P. falciparum or P. vivax malaria, and withdrawal from the study because of severe drug toxicity, other medical reasons, or noncompliance (defined as missing more than 2 days of drug administration in any 2-week period). All end points and toxicity grades were assigned before the code was broken.

Statistical Analysis

Protective efficacy was based on the rate of disease. This rate was calculated by dividing the total number of cases by the total number of weeks during which the drug was administered (primaquine group, 2137 weeks; placebo group, 709 weeks). For each participant, the number of weeks of drug administration was the week of the study in which an end point was reached or the week just before an end point was reached. Protective efficacy was defined as 1 –[(rate in primaquine group)/(rate in placebo group)]. Values are expressed as the estimate of efficacy with 95% CIs (calculated by using the method of Ederer and Mantel [3]).

On the basis of assumptions of 50% incidence of malaria in the placebo group, 90% prophylactic efficacy of primaquine, a desired lower limit of confidence of the prophylactic efficacy of 70%, and a 2:1 allocation ratio, the sample size was approximately 120 participants in the primaquine group and 60 participants in the placebo group.

Results

Top Methods Results Discussion Author & Article Info References

Patient Characteristics

One hundred seventy-nine volunteers were screened for this study. Because 3 volunteers declined to participate, 176 persons (mean age, 23.6 years) were enrolled. One hundred twenty-two (69%) participants were randomly allocated to the primaquine group and 54 (31%) were randomly allocated to the placebo group.

Compliance and Withdrawal

Of the 16 821 doses intended to be administered, 16 789 doses (99.8%) were observed by the senior sergeant to have been swallowed. A total of 32 pills were not taken by a total of 6 participants. Ten participants withdrew (Table 1).

Efficacy

The protective efficacy, based on the rate of all malaria, was 89% (95% CI, 75% to 96%). All patients had fever (temperature 38.3 °C) and two to four of the other six symptoms of malaria (chills, headache, myalgia, nausea, vomiting, and icterus).

The protective efficacy for P. falciparum infection was 94% (CI, 78% to 99%). Two participants in the primaquine group became parasitemic in week 6. Eleven participants who received placebo became parasitemic (2 in week 5, 1 in week 7, 2 in week 8, 1 in week 10, 3 in week 12, 1 in week 13, and 1 in week 14.

The protective efficacy of primaquine against P. vivax infection was 85% (CI, 57% to 95%). Six participants in the primaquine group became parasitemic (2 in week 5, 2 in week 7, 1 in week 8, and 1 in week 17). Among participants who received placebo, 13 became parasitemic (3 in week 4, 2 in week 5, 1 in week 6, 1 in week 7, 1 in week 8, 2 in week 9, 1 in week 10, 1 in week 13, and 1 in week 18).

Toxicity

Three participants in the primaquine group (2.5% [CI, –0.2% to 5.2%]) had epigastric pain, abdominal pain, or vomiting of sufficient severity that they had to withdraw from the study. Six other participants in the primaquine group (5.0% [CI, 1.0% to 9.0%]) had mild or moderate gastrointestinal symptoms. In contrast, only one placebo recipient (2.0% [CI, –3.0% to 7.0%]) had mild gastrointestinal symptoms.

Discussion

Top Methods Results Discussion Author & Article Info References

Prophylaxis with primaquine, 30 mg/d, or placebo was administered to nonimmune Colombian soldiers on patrol in a malaria-endemic area. The treatments were begun 1 day before the start of patrol, were administered during the 15 weeks of patrol, and were continued for 1 week after return to base camp. Participants were followed for parasitemia and symptoms for the 16 weeks of drug administration and for 3 more weeks at base camp. The protective efficacy of primaquine was 89% against all types of malaria, 94% against P. falciparum malaria, and 85% against P. vivax malaria.

Members of the military are tightly supervised. Troops that patrol malaria-endemic regions often come from malaria-free regions. Using participants in the military made it possible to fulfill our aims of conducting the trial according to exacting protocol conditions, adopting the clinically relevant criterion of disease as the end point, and performing the study in nonimmune persons who reflect the population requiring prophylaxis.

Although primaquine was administered immediately before a modest breakfast, moderate gastrointestinal toxicity occurred. Our results contrast with the lack of increased gastrointestinal distress due to primaquine seen in other studies [2]. Other side effects of this regimen that are listed in the literature but were not evaluated in this study are an increase in methemoglobin value to a mean of 0.06% (range, 0.01% to 0.13%) without clinical manifestations [2] and, rarely, cases of leukopenia and arrhythmias [4].

Acute intravascular hemolysis in G6Pd-deficient persons is a potentially severe side effect of primaquine. Variant A deficiency results in a modest decrease in enzyme function and is widespread among persons of African descent. Variant B (Mediterranean) and Asian variants are generally less prevalent but result in greater decreases in enzyme function. G6PD deficiency must be ruled out by testing before primaquine is administered. In addition, pregnant women should not take primaquine because of the possibility of hemolytic anemia in the fetus.

Demonstration of clinical utility usually requires comparable clinical data from two separate trials. The protective efficacies of 94% and 85% found in this trial confirm the 94% and 90% values for P. falciparum and P. vivax infection, respectively, found in the Indonesian trial [2]. These trials indicate that primaquine, 30 mg/d, has a prophylactic efficacy similar to that reported for the standard agents mefloquine (100%) and doxycycline (98%) against P. falciparum infection and somewhat less than the value of 100% against P. vivax infection reported in Indonesian soldiers [5]. For all three agents, lack of compliance in routine clinical use may cause effectiveness to be less than the efficacy demonstrated in study situations. The choice among these agents depends on drug toxicity, efficacy against all Plasmodium species in the endemic region, and feasibility of administration. Persons who are concerned about neuropsychiatric effects should not take mefloquine, those who are susceptible to dermatologic or gastrointestinal effects should avoid doxycycline, and those who may have gastrointestinal effects or are G6PD deficient should not be given primaquine. Pregnancy is a contraindication for doxycycline and primaquine, and the risk of mefloquine prophylaxis (which is teratogenic in animals) has not been clinically evaluated. Primaquine would be a lesser choice for a region in which P. vivax is predominately endemic. Suggestions from volunteer studies [6, 7] that primaquine administration may be stopped 1 week after departure from a malarial region rather than the 4 weeks needed for mefloquine or doxycycline is an important potential advantage for travelers, who often tour an endemic area for less than 1 week.

Mr. Sanchez and Dr. Herrera: Direccion de Sanidad Ejercito, XVII Brigada, Chigorodo, Colombia.

Dr. Padilla: Ministerio de Salud, Cra 7 con Calle 34, Santafe de Bogota, DC, Colombia.

Dr. Berman: Walter Reed Army Institute of Research, Washington, DC 20307.