ARTICLES

Atovaquone-proguanil versus chloroquine-proguanil for malaria prophylaxis in non-immune travellers: a randomised, double-blind study

Birthe Hgh, Paul D Clarke, Daniel Camus, Hans Dieter Nothdurft, David Overbosch, Matthias Gnther, Izak Joubert, Kevin C Kain, Dea Shaw, Neil S Roskell, Jeffrey D Chulay, and the Malarone International Study Team

Summary
Background Chloroquine plus proguanil is widely used for malaria chemoprophylaxis despite low effectiveness in areas where multidrug-resistant malaria occurs. Studies have shown that atovaquone and proguanil hydrochloride is safe and effective for prevention of falciparum malaria in lifelong residents of malaria-endemic countries, but little is known about non-immune travellers. Methods In a double-blind equivalence trial, 1083 participants travelling to a malaria-endemic area were randomly assigned to two treatment groups: atovaquoneproguanil plus placebos for chloroquine and proguanil, or chloroquine, proguanil, and placebo for atovaquoneproguanil. Follow-up was by telephone 7 and 60 days after travel and at a clinic at 28 days. Serum samples were tested for antibodies to a malaria circumsporozoite protein. Blood and serum samples of participants with a potential malaria diagnosis were tested in a reference laboratory. Findings 7 days after travel, at least one adverse event was reported by 311 (61%) of 511 participants who received atovaquone-proguanil and 329 (64%) of 511 who received chloroquine-proguanil. People receiving atovaquone-proguanil had a lower frequency of treatmentrelated gastrointestinal adverse events (59 [12%] vs 100 [20%], p=0001), and of treatment-related adverse events of moderate or severe intensity (37 [7%] vs 56 [11%], p=005). There were fewer treatment-related adverse events that caused prophylaxis to be discontinued in the atovaquone-proguanil group than in the chloroquineproguanil group (one [02%] vs ten [2%], p=0015). Interpretation Overall the two preparations were similarly tolerated. However, significantly fewer adverse gastroInternational Travel Vaccination Centre, Copenhagen, Denmark (B Hgh MD); Medical Advisory Services for Travellers Abroad (MASTA), London, UK (P D Clarke FRCP); Institut Pasteur, Lille, France (D Camus MD); Department for Infectious Diseases and Tropical Medicine, Munich, Germany (H D Nothdurft MD); Harbour Hospital and Institute of Tropical Medicine, Rotterdam, Netherlands (D Overbosch MD); Institute for Tropical Medicine, Berlin, Germany (M Gnther MD); Travelsafe Clinic, Cape Town, Republic of South Africa (I Joubert MD); Toronto General Hospital, Toronto, Canada (K C Kain MD); and Glaxo Wellcome, Research Triangle Park, NC, USA and Greenford, UK (D Shaw RN, N S Roskell MSc, J D Chulay MD) Correspondence to: Malarone Publication Coordinator, Room 50-3505B, Glaxo Wellcome Inc, 5 Moore Drive, Research Triangle Park, NC 27709, USA (e-mail: Malarone@glaxowellcome.com)

intestinal events were observed in the atovaquoneproguanil group in than in the chloroquine-proguanil group. Lancet 2000; 356: 188894

Introduction
Malaria is one of the greatest threats to health for international travellers. The number of reported cases of malaria in Europe increased from 6840 in 1985, to 8438 in 1995, with a case-fatality rate as high as 36%.1 Imported malaria is also a problem in North America, with more than 1000 cases reported annually in the USA and Canada.2,3 Imported malaria mainly occurs when travellers fail to use appropriate chemoprophylaxis or do not use it correctly.3 Widespread parasite resistance renders chloroquine poorly effective in most malaria-endemic countries.4 Mefloquine is highly effective in preventing malaria but has been associated with neuropsychiatric side-effects that may be severe and can restrict its use.5 The combination of chloroquine plus proguanil has fewer serious side-effects than mefloquine6 and is more effective than chloroquine alone, but still less effective than mefloquine.7 An additional drawback with drugs widely used for malaria prophylaxis is that they must be continued for 4 weeks after leaving a malaria-endemic area.8 There is a clear need for better drugs to prevent malaria. In controlled clinical trials with more than 600 participants, a fixed-dose combination of atovaquone and proguanil hydrochloride was 98% effective in prophylaxis of malaria caused by Plasmodium falciparum and had a safety profile similar to that of placebo.912 However, these studies were done on lifelong residents of malaria-endemic countries, who may have had some immunity to malaria. Additionally, the frequency and nature of adverse events might differ between travellers and residents of endemic areas. We have undertaken a randomised, double-blind trial to compare the safety and efficacy of atovaquone-proguanil versus chloroquine-proguanil in non-immune travellers. The hypothesis was that the frequency of adverse events in participants receiving atovaquone-proguanil was not higher than in those receiving chloroquine-proguanil. The frequency of treatment-limiting adverse events and efficacy of prophylaxis were secondary endpoints.

Methods
Study participants Participants were enrolled into study MAL30011 at 21 travel clinics in Denmark, UK, France, Germany, Netherlands, South Africa, and Canada. They were of both sexes, at least age 14 years, weighed more than

1888

THE LANCET Vol 356 December 2, 2000

Copyright 2000 All Rights Reserved

ARTICLES

50 kg, were in good general health, and planned to travel for up to 28 days to an area where P falciparum was endemic. Reasons for exclusion were: known hypersensitivity to atovaquone, proguanil, or chloroquine; a history of seizures, psychiatric disorders, generalised psoriasis, or alcoholism; renal, hepatic, or cardiac dysfunction; severe blood or neurological disorders; pregnancy or lactation; malaria within the previous 12 months; or travel to a malaria-endemic area within the previous 60 days. Written informed consent was obtained from all individuals or their parents. The study was approved by the ethics committee at each study site. At enrolment, information was obtained about demographic characteristics, malaria history, travel destination, current medical conditions and medications; a physical examination was done; and a pregnancy test was carried out for women of childbearing potential. Design and procedures Participants were randomly assigned to one of two treatment groups, atovaquone-proguanil or chloroquine-proguanil. Those in the atovaquoneproguanil group received active atovaquone-proguanil and placebos for chloroquine and proguanil, and those in the chloroquine-proguanil group received chloroquine, active proguanil and a placebo for atovaquone-proguanil (figure 1). For each active drug, capsules or film-coated tablets were identical in appearance to the matching placebo. A computer-generated code was used to randomly assign a treatment number to the three bottles of study drug for every individual. At all sites consecutively enrolled individuals who satisfied all entry criteria received the next treatment number. Treatment codes were provided to investigators in opaque sealed envelopes, to be opened only if knowledge of study drug assignment was required for management of a medical emergency. Atovaquone-proguanil hydrochloride (Malarone tablets, containing 250 mg atovaquone and 100 mg proguanil hydrochloride, GlaxoWellcome, Mississauga, Canada) or matching placebo was given according to standard recommendations of the manufacturer (figure 1).12 Chloroquine phosphate (Avloclor, 250 mg tablets equivalent to 155 mg chloroquine base) and proguanil hydrochloride (Paludrine, 100 mg tablets) were manufactured by AstraZeneca and encapsulated by GlaxoWellcome to achieve blinding. Chloroquine and proguanil or matching placebos were given according to standard recommendations of WHO (figure 1).8 At enrolment, participants were given a diary card to record details of study drug administration, and a malaria diagnosis kit. They were instructed to use this kit if they were diagnosed with malaria. The kit contained a card on which a health care provider could record details about the diagnosis; slides and a slide holder (to take blood films); filter strips (Isocode Stix, Schleicher and Schull, Keene, NH, USA) on which blood could be spotted for parasite DNA analysis; tubes for whole blood (for parasite DNA analysis) and plasma (to measure plasma drug concentrations); and instructions on how to take samples and arrange for prepaid courier service to return samples to a central laboratory. Individuals were told that if they developed malaria they should ask the health care provider making the diagnosis to obtain the samples and send the whole-

Before

Travel period During

After

12 days

Group

ATQ-PGN active one tablet daily Atovaquoneproguanil Chloroquine placebo two capsules per week

Proguanil placebo two capsules daily

ATQ-PGN placebo one tablet daily Chloroquineproguanil

Chloroquine active two capsules per week

Proguanil active two capsules daily Figure 1: Drug dosing regimens

ATQ=atovaquone; PGN=proguanil.

blood and plasma samples to the central laboratory. The participant was instructed to bring the slides and filter strips to the investigator at the follow-up visit. Follow-up surveys People were interviewed by telephone 7 and 60 days after leaving the malaria-endemic area and at a clinic visit at 28 days after departure. At the 7-day contact, information was obtained about the exact dates and locations of travel in malaria-endemic areas. At enrolment and each follow-up assessment, participants were asked if they had been diagnosed with malaria and queried about symptoms (fever, nausea, vomiting, abdominal pain, diarrhoea, mouth ulcers, itching, headache, insomnia, strange or vivid dreams, dizziness, anxiety, depression, visual difficulties, and fits or seizures). They were also asked an open-ended question about any other medical conditions or adverse events they had had since the last visit. Each adverse event was assessed for date of onset, duration, and intensity (mild=fairly trivial, moderate=interfered with daily activities, severe=sought medical advice). Investigators assessed whether there was a reasonable possibility that the adverse event was caused by the study drug. An adverse event was recorded as treatment-emergent if it started while the participant was taking the study drug, treatment-limiting if it resulted in permanent discontinuation of study drug, and serious if it were fatal, life-threatening, disabling, resulted in admission, or otherwise seriously jeopardised the participant. Compliance with study drug use was assessed at the 28 day follow-up visit by interview, reviewing the diary card, and counting unused pills. Measurement of malarial indices At the baseline and 28 day follow-up visits, serum was collected from participants at all sites, and blood was obtained for routine haematological tests (haemoglobin,

THE LANCET Vol 356 December 2, 2000

1889

Copyright 2000 All Rights Reserved

28 days

7 days

7 days

ARTICLES

white blood cell count, platelet count) and chemistry (creatinine, alanine aminotransferase) from all people at one site. The serum was used to measure antibodies to P falciparum circumsporozoite protein by ELISA13 for all participants, and to measure antibodies to blood-stages of the four human species of plasmodium by immunofluorescence14 for those diagnosed with potential malaria. The circumsporozoite antibody test was judged positive when the result for a post-travel serum sample was more than 2 SD above the mean of results for the negative control sera and also more than 2 SD above the baseline serum sample result. Antibodies to blood-stage parasites were recorded as positive when they were detected at a titre of 1/64 or greater in a post-diagnosis serum sample and the baseline sample was negative, or when there was a 16-fold or greater increase in antibody titre in a postdiagnosis serum sample compared with baseline.14 Blood smears obtained at the time of malaria diagnosis were examined at the malaria reference laboratory of the Toronto General Hospital. DNA was extracted from whole blood samples and filter strips; parasite ribosomal RNA genes were amplified by PCR to identify the Plasmodium spp causing malaria.15 As defined in the trial protocol, a potential diagnosis of malaria was thought definite if parasite DNA was detected by PCR or parasites were seen on a blood smear slide; possible if a diagnosis of malaria was recorded by a health care provider, but blood smears and parasite DNA analyses were negative or missing and antibodies to blood-stage parasites were missing; and negative if a diagnosis of malaria was recorded by a health care provider, but blood smears and parasite DNA analyses were negative or missing and antibodies to blood-stage parasites were negative. P falciparum isolates obtained during the course of this study underwent DNA analysis for putative molecular markers of chloroquine and proguanil resistance as described1620 and by sequencing. Statistical analyses The primary study endpoint was the overall frequency of adverse events assessed at 7 days after leaving the malaria-endemic area, analysed in the intent-to-treat population of all participants who received at least one dose of study drug. The overall frequency of adverse events at 28 days after leaving the malaria-endemic area was also determined. The sample size was based on Jones and colleagues formula.21 An equivalence trial with 400 suitable participants per group has power of 82% to detect non-inferiority of atovaquone-proguanil compared with chloroquine-proguanil, if the overall proportion of adverse events is 40% and a 10% difference in adverse event proportions is regarded as clinically important. The power is also more than 80% if the overall proportion of adverse events is 6070%. Non-inferiority was assessed by comparison of the two-sided 95% CI for the adverse event proportion difference against the non-inferiority range (100%, +10%). The target enrolment was increased to 1000 (500 per group) to allow for unsuitable participants. Secondary study endpoints were the frequency of treatment-limiting adverse events and the efficacy of prophylaxis. Percentage efficacy was calculated from: 100(1[number of participants with confirmed malaria/number of participants at risk]). To estimate the minimum efficacy of each drug, the number of participants regarded as at high risk was the number

1083 eligible participants R 540 randomised to receive atovaquone-proguanil 543 randomised to receive chloroquine-proguanil

11 6 4 8

did not travel lost to follow-up consent withdrawn other/unknown*

14 3 3 12

did not travel lost to follow-up consent withdrawn other/unknown

511 received study drug 462 completed dosing 49 stopped dosing early

511 received study drug 468 completed dosing 43 stopped dosing early

2 2 1 1 4

did not travel lost to follow-up adverse event consent withdrawn other

2 did not travel 1 lost to follow-up 1 adverse event

501 completed the trial Figure 2: Trial profile

507 completed the trial

*Unknown (3), work-related (2), wanted more protection from malaria (1), permanently left country (1), and protocol violation (1). Unknown (9), concern about potential side-effects (1), protocol violation (1), and adverse event (1). Could not swallow capsules (1), drug lost (1), protocol violation (1), and treatment with another antimalarial drug (1).

who had confirmed malaria or who developed anticircumsporozoite antibodies and for whom 60-day efficacy data were available. To estimate the maximum efficacy of each drug for prevention of malaria, the number of those thought at risk was the number for whom 60-day efficacy data were available. For both minimum and maximum estimates of efficacy, the 95% CI around the proportion was calculated from the binomial distribution. Proportions of people with adverse events were compared with Yates corrected 2 test. Because the p-values are unadjusted and multiple comparisons will inflate the type I error, care should be taken when interpreting these comparisons.

Results
From April to October, 1999, 1083 participants were randomly assigned to receive atovaquone-proguanil (n=540) or chloroquine-proguanil (543). 61 people did not receive their first dose of study drug because they did not travel to a malaria-endemic area (25), were lost to follow-up (nine), withdrew consent (seven), or other reasons (20). Of the 1022 who received at least one dose of study drug, 1008 (99%) completed the trial (figure 2). The two groups were well balanced with respect to baseline demographics, history of malaria, travel destination, and duration of travel (table 1). 24 participants were more than 65 years old. 16 participants (16%) reported a previous episode of malaria a median of 12 (range 260) years before enrolment. The average duration of travel was about 25 weeks and 642 (63%) people travelled to Africa.

1890

THE LANCET Vol 356 December 2, 2000

Copyright 2000 All Rights Reserved

ARTICLES

Characteristic Mean (SD) age (years) Range Male/Female Race* White Black Asian Other Mean (SD) height (cm) Mean (SD) weight (kg) History of malaria Median (range) time since last episode (years) Travel destination East Africa West Africa Southern Africa Central Africa South America Other Mean (SD) travel duration (days)

Atovaquone-proguanil (n=511) 360 (133) 1372 251/260 (49/51%) 496 (97%) 6 (1%) 3 (1%) 6 (1%) 1731 (96) 714 (134) 9 (18%) 13 (260)

Chloroquine-proguanil (n=511) 350 (133) 1374 276/235 (54/46%) 486 (95%) 12 (2%) 7 (1%) 6 (1%) 1737 (97) 723 (145) 7 (14%) 11 (230)

Event

Number of participants with adverse events Atovaquone-proguanil Chloroquine-proguanil (n=511) (n=511) p-value 0024 0001 0113 0001 0033 0350 0002 0754 0865 0479 0498 0999 0370 0999 0872 0999

Any adverse event* Any gastrointestinal event Diarrhoea Nausea Abdominal pain Mouth ulcers Vomiting Any neuropsychiatric event Dizziness Strange or vivid dreams Insomnia Visual difficulties Anxiety Depression Headache Itching

110 (22%) 59 (12%) 24 (5%) 9 (2%) 15 (3%) 18 (4%) 0 49 (10%) 17 (3%) 19 (4%) 8 (2%) 10 (2%) 1 (<1%) 3 (<1%) 21 (4%) 6 (1%)

142 (28%) 100 (20%) 37 (7%) 34 (7%) 30 (6%) 25 (5%) 11 (2%) 53 (10%) 19 (4%) 14 (3%) 12 (2%) 10 (2%) 4 (<1%) 2 (<1%) 19 (4%) 5 (1%)

132 (26%) 90 (18%) 77 (15%) 15 (3%) 50 (10%) 151 (30%) 169 (75)

152 (30%) 102 (20%) 68 (14%) 12 (2%) 39 (8%) 140 (27%) 176 (69)

*Among participants who reported a drug-related adverse event, the mean (SD) number of adverse events per person was 18 (11) while receiving atovaquone-proguanil and 22 (18) while receiving chloroquine-proguanil.

Table 2: Treatment-emergent adverse events attributed to study drug

*Percentages may not add up to 100 because of rounding. Some participants visited more than one part of Africa

Table 1: Baseline characteristics

Participants and study personnel remained unaware of treatment assignment apart from ten individuals (seven in the atovaquone-proguanil group and three in the chloroquine-proguanil group). Reasons for breaking the treatment code were: an adverse event for which knowledge of study treatment was deemed essential (n=five); participants who opened the chloroquine and proguanil capsules themselves (four) and loss of study drug before completion of postexposure prophylaxis (one). The mean (SD) duration of treatment was 26 (8) days for atovaquone-proguanil, 48 (11) days for chloroquine, and 45 (10) days for proguanil. The proportion of participants who took at least 80% of prescribed doses in the pretravel, travel and post-travel periods was: 484 (95%), 489 (96%), and 473 (93%), respectively, for atovaquone-proguanil; 478 (94%), 466 (91%), and 408 (80%) for chloroquine; and 461 (90%), 480 (94%), and 443 (87%) for proguanil. In the post-travel period the difference from atovaquone-proguanil was significant for chloroquine (difference=65 [13%], p=0001) and proguanil (difference=30 [6%], p=0003). At 7 days after return from a malaria-endemic area, one or more adverse events were reported by 311 of 511 (61%) participants in the atovaquone-proguanil group and 329 of 511 (64%) in the chloroquine-proguanil group. The difference in frequency of adverse events in the intent-to-treat population was 35% (95% CI 95%, +24%, not significant). Because of the different pretravel dosing regimens, participants randomly assigned to atovaquone-proguanil started chloroquine placebo about 5 days before starting atovaquone-proguanil. Excluding events that occurred while people were receiving only placebo, adverse events were reported by 296 of 511 (58%) of those receiving atovaquone-proguanil and 329 of 511 (64%) receiving chloroquine-proguanil. The difference in number of participants who reported adverse events while receiving active treatment was 33 (65%) (95% CI 124%, 05%). Most events were thought by the investigator to be unrelated to study drug. Treatment-emergent adverse

events attributed to study drug occurred in more participants who received chloroquine-proguanil than those who received atovaquone-proguanil (142 [28%] vs 110 [22%], p=0024). The difference was significant for gastrointestinal events but not for neuropsychiatric events (table 2). Most adverse events were mild in intensity. Moderate or severe events attributable to study drug occurred in 37 (7%) participants (54 events) receiving atovaquone-proguanil and 56 (11%) (97 events) on chloroquine-proguanil (difference= 19 [4%], p=005). These events were severe in five (1%) people (six events) receiving atovaquone-proguanil and 11 (2%) (20 events) on chloroquine-proguanil. One participant receiving chloroquine-proguanil had an epileptic seizure. 49 people in the atovaquone-proguanil group and 43 on chloroquine-proguanil discontinued study drug prematurely because of adverse events (11 vs 16), protocol violation (nine vs six), non-compliance (five vs six), or other reasons (24 vs 15). Other reasons included: did not travel, drugs lost or stolen, or consent withdrawn.
Event Number of participants with an event Atovaquone-proguanil (n=511) Any treatment-limiting event Any gastrointestinal event Nausea Abdominal pain Diarrhoea Vomiting Other Any neuropsychiatric event Dizziness Anxiety Other 1* 0 0 0 0 0 0 0 0 0 1 Chloroquine-proguanil (n=511) 10 10 6 5 4 2 4 2 1 1 1

*Two additional participants developed a treatment-limiting adverse event while receiving chloroquine placebo before starting active atovaquone-proguanil (rash or abdominal pain and diarrhoea) and one additional person developed a treatment-limiting adverse event (dizziness) while receiving chloroquine and proguanil placebos after completing prophylaxis with active atovaquone-proguanil. p=0015. Other events were burning in pharynx, constipation, mouth ulcers, dysphagia, and oesophageal reflux, each of which occurred in one person receiving chloroquineproguanil. Other events were allergic reaction in one person receiving atovaquone-proguanil and headache in one participant on chloroquine-proguanil.

Table 3: Treatment-limiting adverse events attributed to study drug

THE LANCET Vol 356 December 2, 2000

1891

Copyright 2000 All Rights Reserved

ARTICLES

Atovaquone-propguanil Number of participants 60-day efficacy data available CS* antibodies present Confirmed P falciparum malaria Minimum efficacy (95% CI) Maximum efficacy (95% CI) 501 7 0 100% (59100%) 100% (99100%)

Chloroquine-proguanil 507 8 3 70% (3593%) 99% (9899%)

*CS=circumsporozoite. One participant with confirmed malaria also developed CS* antibodies.

Table 4: Estimates of minimum and maximum efficacy for malaria prophylaxis

Among those who discontinued study drug because of an adverse event, the event was attributed to treatment in 14 of 27. Five treatment-limiting adverse events attributed to study drug arose in four participants, randomly assigned to receive atovaquone-proguanil and 26 such events happened in ten randomly assigned to chloroquine-proguanil (table 3). A serious adverse event occurred in six participants who received atovaquoneproguanil (infectious illnesses in four, Wolff-ParkinsonWhite syndrome in one, and pituitary tumour in one) and in six people who were on chloroquine-proguanil (malaria in three, other infectious illnesses in two, and depression in one). None of these serious events was thought to be caused by study drug. For the 180 people who had laboratory safety samples collected, there were no significant differences in either treatment group between baseline and follow-up values for haematological and clinical chemistry tests. No clinically important laboratory abnormalities were identified. A potential diagnosis of malaria was made in four participants. In three, P falciparum malaria was diagnosed while they were receiving chloroquoneproguanil; 3, 6, or 11 days after leaving Mali, Nigeria, or Uganda. The diagnoses were confirmed by review of slides (two), parasite DNA analysis (three), and diagnostic increase in blood-stage antibodies (three). In one person, symptoms of malaria began 28 days after completing prophylaxis with atovaquone-proguanil and a subsequent blood smear was reported as positive for P ovale. No blood smears or parasite DNA samples were returned to the reference laboratory, but serological testing for blood-stage antibodies showed a diagnostic titre increase against P ovale. DNA from the three P falciparum isolates carried the K76T mutation in the pfcrt gene associated with chloroquine resistance.19 One of three isolates also contained the N86Y allele in pfmdr1.1618 All three isolates contained the S108N mutation in dhfr, which has been associated with resistance to pyrimethamine and a 4-fold to 22-fold increase in the 50% inhibitory concentration to cycloguanil.20 Additionally, all three isolates had N51I and C59R mutations associated with high resistance to pyrimethamine.20 1008 participants completed the 60-day follow-up and had efficacy information recorded. 987 had paired samples available for serological testing, of these, circumsporozoite antibodies developed in 15 (15%), including one of three people with confirmed P falciparum malaria. The minimum efficacy for prevention of P falciparum malaria was estimated to be 100% (95% CI 59100%) in the atovaquone-proguanil group and 70% (95% CI 3593%) in the chloroquineproguanil group (table 4).

Discussion
Chloroquine and proguanil are generally regarded as the safest drugs available for malaria prophylaxis, and are

very rarely associated with severe adverse reactions in the recommended doses.22 The favourable safety profile of these drugs accounts for their continued wide use, despite evidence of decreasing effectiveness as the prevalence of multidrug-resistant P falciparum increases. In this study, the proportion of individuals who had at least one adverse event was similar in the two treatment groups, but evaluation of several indices indicated that atovaquone-proguanil was significantly better tolerated with respect to gastrointestinal events. The findings that treatment-emergent gastrointestinal adverse events attributed to study drug occurred more often in the chloroquine-proguanil group and were the most common reason for premature discontinuation are in accord with those of previous studies showing that gastrointestinal symptoms are the most common adverse events in people receiving chloroquine-proguanil for malaria prophylaxis.6,23,24 Failure to complete the full course of antimalarial prophylaxis will increase the risk of developing malaria. For this reason, a low frequency of treatment-limiting adverse events is important for the effectiveness of prophylaxis. In our study, chemoprophylaxis was discontinued prematurely in only one person receiving atovaquone-proguanil, compared with ten on chloroquine-proguanil. We assessed the efficacy of chemoprophylaxis as a secondary endpoint. Because many other illnesses can be mistaken for malaria, we obtained blood specimens at the time of suspected malaria and serum samples before travel and several weeks after treatment. Testing of these specimens in reference laboratories allowed us to confirm a diagnosis of falciparum malaria in three people receiving chloroquine-proguanil and in none on atovaquone-proguanil. In-vitro drug susceptibility testing of parasite isolates was not feasible but, as a surrogate, isolates were analysed for mutations in genes associated with resistance to chloroquine and proguanilcycloguanil. All three isolates carried at least one mutation previously linked to chloroquine resistance, and all three had mutations in the dhfr gene linked to antifolate drug resistance. These data provide a potential molecular basis for failure of prophylaxis with chloroquine-proguanil. Prospective collection of paired serum samples to measure antibodies to the circumsporozoite protein of P falciparum allowed us to estimate the minimum efficacy of chemoprophylaxis. In a study in which serum was obtained every 10 days for 2 months from soldiers exposed to malaria, detection of anti-circumsporozoite antibodies by ELISA was reported to have high sensitivity and specificity for identifying individuals who were bitten by a P falciparum-infected mosquito and developed clinical malaria.25 People taking effective antimalarial prophylaxis can develop anticircumsporozoite antibodies after being bitten by infected mosquitoes without developing malaria.26 Because we obtained only one postexposure serum sample, the proportion of participants in our study who developed a circumsporozoite-antibody response is a minimum estimate of the proportion actually bitten by a malaria-infected mosquito. This assumption is lent support by the observation that circumsporozoite antibodies were detected in only one of three people with confirmed P falciparum malaria. Based on the circumsporozoite antibody results, the minimum efficacy of chemoprophylaxis was higher for atovaquone-proguanil than for chloroquineproguanil.

1892

THE LANCET Vol 356 December 2, 2000

Copyright 2000 All Rights Reserved

ARTICLES

Because of the small number of endpoints, the difference in efficacy rates between treatment groups was not analysed statistically. The risk of developing malaria for travellers to East Africa who take no chemoprophylaxis is about 12% per month and chloroquine-proguanil has prophylactic effectiveness of 72%.7 If the average duration of travel is 25 weeks, the expected rate of malaria in travellers taking prophylaxis would be 02% per month with chloroquine-proguanil and 0035% per month with a drug that has 95% effectiveness. With these expected rates, and assuming is 5% and power is 80%, a study in travellers designed to show that a new antimalaria drug with 95% efficacy is better than chloroquine-proguanil would require more than 16 000 participants.27 Although our study was clearly underpowered to show higher efficacy than the chloroquine-proguanil combination, atovaquoneproguanil is more effective than these drugs for treatment of falciparum malaria28 and therefore could be more effective than chloroquine-proguanil for prophylaxis in areas where multidrug resistant P falciparum infections occur. One participant developed malaria caused by P ovale, 28 days after completing the standard course of prophylaxis with atovaquone-proguanil. Both atovaquone and proguanil have causal prophylactic activity directed against developing hepatic forms of P falciparum,29,30 but neither drug is active against hypnozoites.29,31 Thus atovaquone-proguanil, like chloroquine and mefloquine, acts only as a suppressive prophylactic agent against P vivax and P ovale. After dosing is discontinued and drug elimination occurs, a delayed onset of malaria caused by a relapsing type of malaria parasite can therefore occur, and is consistent with the course of events we observed. Travellers with intense exposure to P vivax or P ovale, and those who develop malaria caused by either of these parasites, will require additional treatment with primaquine.31
Contributors
B Hgh, P D Clarke, D Camus, H D Nothdurft, D Overbosch, M Gnther, and I Joubert did the study and contributed to data review and interpretation of results. K C Kain did the parasite DNA analyses and contributed to interpretation of results and writing the paper. N S Roskell and J D Chulay contributed to study design and data analysis. D Shaw supervised the study. B Hgh and J D Chulay had primary responsibility for writing the paper. Other members of the Malarone International Study Team include: R H Behrens (London, UK), J Beytout (Clermont Ferrand, France), U Bienzle (Berlin, Germany), O Bouchaud (Paris, France), J Delmont (Marseilles, France), E Dutoit (Lille, France); K Fleischer (Wrzburg, Germany), B Gachot (Paris, France), T J L M Goud (Rotterdam, Netherlands), P Inglebert (Lille, France); J Knobloch, Tbingen, Germany; B Marchoux (Toulouse, France), V Masson (Lille, France); M Peters (Hamburg, Germany), E Petersen (Copenhagen, Denmark), H Schilthius (Amsterdam, Netherlands); R Tan (Vancouver, Canada), S Toovey (Sunninghill, South Africa), S Waner (Johannesburg, South Africa), and B Zieger (Dresden, Germany). GlaxoWellcome members of the Study Team include: Amod (South Africa), R Balschmidt (Denmark), C Brennan (USA), A Goetschel (France), C Hedgley (UK), T Hendrickx (Netherlands), P de Jonge (Netherlands), M Lebaddi (France), R Marina (Canada), G B Miller (USA), M OHare (UK), J Olsen (Denmark), H Richter (Germany), S van Delft (Netherlands), and T Scott (USA).

References
1 Muentener P, Schlagenhauf P, Steffen R. Imported malaria (198595): trends and perspectives. Bull World Health Organ 1999; 77: 56066. Williams HA, Roberts J, Kachur SP, et al. Malaria surveillance United States, 1995. MMWR CDC Surveill Summ 1999; 48: 123. Kain KC, Harrington MA, Tennyson S, Keystone JS. Imported malaria: prospective analysis of problems in diagnosis and management. Clin Infect Dis 1998; 27: 14249. Lobel HO, Kozarsky PE. Update on prevention of malaria for travellers. JAMA 1997; 278: 176771. Nosten F, van Vugt M. Neuropsychiatric adverse effects of mefloquine. What do we know and what should we do? CNS Drugs 1999; 11: 18. Barrett PJ, Emmins PD, Clarke PD, Bradley DJ. Comparison of adverse events associated with use of mefloquine and combination of chloroquine and proguanil as antimalarial prophylaxis: postal and telephone survey of travellers. BMJ 1996; 313: 52528. Steffen R, Fuchs E, Schildknecht J, et al. Mefloquine compared with other malaria chemoprophylactic regimens in tourists visiting East Africa. Lancet 1993; 341: 1299303. WHO. International travel and health. Geneva: WHO, 1997: 6778. Shanks GD, Gordon DM, Klotz FW, et al. Efficacy and safety of atovaquone/proguanil for suppressive prophylaxis against Plasmodium falciparum malaria. Clin Infect Dis 1998; 27: 49499. Lell B, Luckner D, Ndjav M, Scott T, Kremsner PG. Randomised placebo-controlled study of atovaquone plus proguanil for malaria prophylaxis in children. Lancet 1998; 351: 70913. Sukwa TY, Mulenga M, Chisaka N, Roskell NS, Scott TR. A randomized, double-blind, placebo-controlled field trial to determine the efficacy and safety of Malarone (atovaquone/proguanil) for the prophylaxis of malaria in Zambia. Am J Trop Med Hyg 1999; 60: 52125. Shanks GD, Kremsner PG, Sukwa TY, et al. Atovaquone and proguanil hydrochloride for prophylaxis of malaria. J Travel Med 1999; 6 (suppl 1): S21S27. Ballou WR, Hoffman SL, Sherwood JA, et al. Safety and efficacy of a recombinant DNA Plasmodium falciparum sporozoite vaccine. Lancet 1987; 1: 127781. Sulzer AJ, Wilson M, Hall EC. Indirect fluorescent-antibody tests for parasitic diseases. V. An evaluation of a thick-smear antigen in the IFA test for malaria antibodies. Am J Trop Med Hyg 1969; 18: 199205. Snounou G, Viriyakosol S, Zhu XP, et al. High sensitivity of detection of human malaria parasites by the use of nested polymerase chain reaction. Mol Biochem Parasitol 1993; 61: 31520. Foote SJ, Kyle DE, Martin RK, et al. Several allelles of the multidrug-resistance gene are closely linked to chloroquine resistance in Plasmodium falciparum. Nature 1990; 345: 25558. Reed MB, Saliba KJ, Caruana SR, Kirk K, Cowman AF. Pgh1 modulates sensitivity and resistance to multiple antimalarials in Plasmodium falciparum. Nature 2000; 403: 90609. Duraisingh MT, Drakeley CJ, Muller O, et al. Evidence for selection for the tyrosine-86 allele of the pfmdr 1 gene of Plasmodium falciparum by chloroquine and amodiaquine. Parasitology 1997; 114: 20511. Fidock DA, Nomura T, Talley AK, et al. Mutations in the Plasmodium falciparum digestive vacuole transmembrane protein PfCRT and evidence for their role in chloroquine resistance. Mol Cell 2000; 6: 86171. Foote SJ, Galatis D, Cowman AF. Amino acids in the dihydrofolate reductase-thymidylate synthase gene of Plasmodium falciparum involved in cycloguanil resistance differ from those involved in pyrimethamine resistance. Proc Nat Acad Sci USA 1990; 87: 301417. Jones B, Jarvis P, Lewis JA, Ebbutt AF. Trials to assess equivalence: the importance of rigorous methods. BMJ 1996; 313: 3639. Luzzi GA, Peto TE. Adverse effects of antimalarials: an update. Drug Safety 1993; 8: 295311. Carme B, Peguet C, Nevez G. Compliance with and tolerance of mefloquine and chloroquine + proguanil malaria chemoprophylaxis in French short-term travellers to sub-Saharan Africa. Trop Med Int Health 1997; 2: 95356. Durrheim DN, Gammon S, Waner S, Braack LE. Antimalarial

4 5

8 9

10

11

12

13

14

15

16

17

18

19

20

21

Acknowledgments
We thank Kent Kester and Carolyn Holland of the Walter Reed Army Institute of Research, Washington, DC, USA, for the circumsporozoite antibody assays, and Marianna Wilson of the Centers for Disease Control and Prevention, Atlanta, GA, USA, for the blood-stage antibody assays. K C Kain was supported in part by a career award from the Ontario Ministry of Health and a grant from PSI. This study was funded by GlaxoWellcome. 22 23

24

THE LANCET Vol 356 December 2, 2000

1893

Copyright 2000 All Rights Reserved

ARTICLES

25

26

27 28

prophylaxisuse and adverse events in visitors to the Kruger National Park. S Afr Med J 1999; 89: 17075. Webster HK, Boudreau EF, Pang LW, Permpanich B, Sookto P, Wirtz RA. Development of immunity in natural Plasmodium falciparum malaria: antibodies to the falciparum sporozoite vaccine 1 antigen (R32tet32). J Clin Microbiol 1987; 25: 100208. Bwire R, Slootman EJ, Verhave JP, Bruins J, Docters van Leeuwen WM. Malaria anticircumsporozoite antibodies in Dutch soldiers returning from sub-Saharan Africa. Trop Med Int Health 1998; 3: 6669. Fleiss JL. Statistical methods for rates and proportions. 2nd edn. New York: John Wiley, 1981: 3842. Looareesuwan S, Chulay JD, Canfield CJ, Hutchinson DB.

Malarone (atovaquone and proguanil hydrochloride): a review of its clinical development for treatment of malaria. Am J Trop Med Hyg 1999; 60: 53341. 29 Fairley NH. Researches on paludrine (M.4888) in malaria. An experimental investigation undertaken by the L.H.Q. Medical Research Unit (A.I.F), Cairns, Australia. Trans R Soc Trop Med 1946; 40: 10551. 30 Shapiro TA, Ranasinha CD, Kumar N, Barditch-Crovo P. Prophylactic activity of atovaquone against Plasmodium falciparum in humans. Am J Trop Med Hyg 1999; 60: 83136. 31 Looareesuwan S, Wilairatana P, Glanarongran R, et al. Atovaquone and proguanil hydrochloride followed by primaquine for treatment of Plasmodium vivax malaria in Thailand. Trans R Soc Trop Med 1999; 93: 63740.

Clinical picture

Short stature and rickets

A S Kashyap, Vivek Kumar

A 14-year-old male was referred for growth arrest, muscle pain, and genu-valgum. On examination he was found to have short stature, moderate pelvic girdle muscle weakness, genu-valgum, and a painful waddling gait. Physical examination was otherwise normal. His pelvic radiograph showed homogeneous ground glass appearance and deformity. Knee radiographs showed widened femoral, tibial, and fibular metaphyses with transverse growth arrest lines, abnormal trabecular patterns, and a malunited femoral shaft fracture. Laboratory investigations were consistent with Fanconi syndrome with proximal renal tubular acidosis. He was being managed as a case of renal rickets. However, closer examination revealed Kayser-Fleischer rings. With the combination of Kayser-Fleischer rings, clinical and radiological features of rickets, and laboratory investigations consistent with renal tubular acidosis, a diagnosis of Wilson's disease presenting with Fanconis syndrome leading to rickets was made. The clinical diagnosis of Wilson's disease was confirmed by serum caeruloplasmin levels of 150 mg/L (270370 mg/L).
Department of Medicine, Armed Forces Medical College, Pune 411 040, India ( V Kumar MD, A S Kashyap MD)

1894

THE LANCET Vol 356 December 2, 2000

Copyright 2000 All Rights Reserved