Objective To assess the safety of chloroquine (CQ) as prophylaxis against Plasmodium vivax infection during pregnancy.
Method One thousand pregnant Karen women were enrolled in a randomized, double-blind, placebo-controlled trial of chemoprophylaxis with chloroquine (500 mg phosphate (or 300 mg base) weekly). Women received a median (range) chloroquine phosphate total dose of 9500 (1500–17 500) mg. The mothers were actively followed from inclusion to delivery and their infants until 12 months of age.
Results Chloroquine prophylaxis completely prevented P. vivax episodes; 10.1% (95%CI: 7.3–14.5) of women in the placebo group experienced at least one episode of vivax malaria but no episode occurred in women in the CQ group. By contrast, the numbers of P. falciparum episodes were similar in each group: 7.4% (95%CI: 3.7–11.1) and 5.6% (95%CI: 3.3–7.9) in the placebo and CQ groups respectively (P = 0.56). Chloroquine prophylaxis was well tolerated and there was no difference in the proportions of reported side effects between CQ treated and placebo groups except for the duration of palpitations and sleeping disorders which were more frequent in those who had received CQ. Chloroquine prophylaxis had no impact on maternal anaemia, birth weight, gestational age, development of newborns or on growth, neurological development or visual acuity in infants at 1 year of age.
Conclusion Chloroquine is safe and effective as prophylaxis against P. vivax during pregnancy in this population.
Objectif Evaluer la sûreté de la chloroquine comme prophylaxie contre l'infection de P. vivax pendant la grossesse.
Méthode Mille femmes enceintes Karen ont été enrôlées dans une étude randomisée placebo/contrôle à double aveugle sur la chimioprophylaxie avec la chloroquine (500 mg de chloroquine sulfate ou 300 mg de chloroquine base hebdomadaire). Les femmes ont reçu une dose médiane de phosphate de chloroquine de 9500 [1500–17 500] mg. Les mères ont été activement suivies de inclusion jusqu’à l'accouchement et à leur enfants jusqu ’à l’âge de 12 mois.
Résultats La prophylaxie à chloroquine a complètement évité les épisodes de P. vivax. 10,1% (IC 95%: 7,3–14,5) des femmes dans le groupe placebo ont eu au moins un épisode de vivax mais aucun épisode n'a été observé chez les femmes dans le groupe chloroquine. En revanche, le nombre d’épisodes de malaria àP. falciparuméta similaire dans les deux groupes: 7,4% (IC95%: 3,7–11,1) et 5,6% (IC95%: 3,3–7,9) dans les groupes placebo et chloroquine respectivement (P = 0.56). La prophylaxie à la chloroquine a été bien tolérée et il n'y avait aucune différence dans les taux d'effets secondaires rapportés entre les groupes chloroquine et placebo à l'exception de la durée des palpitations et des troubles de sommeil qui étaient plus fréquents dans le groupe chloroquine. La prophylaxie à la chloroquine n'a eu aucun impacte sur l'anémie maternelle, le poids de naissance, la durée de la gestation, le développement des nouveaux-nés ou la croissance, le développement neurologique ou l'acuité visuelle des enfants à l’âge de un an.
Conclusion La prophylaxie à la chloroquine est sûre et efficace contre P. vivax pendant la grossesse.
Objetivo Determinar la seguridad de la cloroquina (CQ) como profilaxis contra la infección por P. vivax durante el embarazo.
Método Se incluyeron mil mujeres Karen embarazadas en un ensayo aleatorizado, doble ciego y controlado con placebo de quimioprofilaxis con CQ (500 mg fosfato(o 300 mg base) semanales). Las mujeres recibían una dosis media [rango] de fosfato de CQ de 9500 [1500–17 500] mg. Las madres fueron seguidas activamente desde su inclusión hasta el parto, y sus bebés hasta los 12 meses de edad.
Resultados La profilaxis con CQ previno por completa los episodios de P.vivax; un 10.1% (95%IC: 7.3–14.5) de las mujeres dentro del grupo placebo tuvo al menos un episodio de malaria por vivax, mientras que no hubo ningún episodio entre las mujeres del grupo CQ. En cambio, el número de episodios de P. falciparum fue similar en ambos grupos: 7.4% (95%IC: 3.7–11.1) y 5.6% (95%IC: 3.3–7.9) en el grupo placebo y el de CQ respectivamente (P = 0.56). La profilaxis con CQ fue bien tolerada y no hubo diferencias en la proporción de efectos secundarios reportados por los dos grupos, excepto por la duración de palpitaciones y desórdenes del sueño que fueron más frecuentes entre las mujeres que habían recibido CQ. La profilaxis con CQ no tuvo impacto en la anemia materna, peso al nacer, edad gestacional, desarrollo de los recién nacidos o en el crecimiento, desarrollo neuronal o agudeza visual de los lactantes al cumplir el año.
Conclusión La profilaxis con cloroquina es segura y eficaz contra la infección por P.vivax durante el embarazo.
Plasmodium vivax is the most common malaria parasite in Asia and Latin America where it is responsible for 50–70% of all malaria cases (Mendis et al. 2001). P. vivax infection in pregnancy is harmful to both the mother and her foetus, causing maternal anaemia and low birth weight, although these are not as marked as in P. falciparum infections (Nosten et al. 1999a; Singh et al. 1999). Chloroquine (CQ) is the drug of choice for treatment of vivax malaria in areas where the parasite remains sensitive to this drug but its usefulness in chemoprophylaxis against P. vivax in pregnancy is unknown. Therefore, we have undertaken a randomized, placebo controlled trial of chloroquine chemoprophylaxis in a large group of pregnant women in Thailand.
Although tonnes of CQ have been consumed by pregnant women in tropical countries as prophylaxis against P. falciparum malaria, few studies have examined the effects of CQ on the developing human brain (Brabin 1991; Ostensen & Ramsey-Goldman 1998). In animals, CQ crosses the placenta and accumulates in fetal tissues where it has an affinity for melanin-containing tissues such as the retina, iris and choroid of the eye. This deposition is irreversible and leads to loss of vision. The effect is dose related (Parke 1988). Chloroquine and hydroxychloroquine (HCQ) have been used in very high dosages (average HCQ sulphate 200 mg per day) for the treatment of connective tissue diseases and a recent review identified more than 250 pregnancies resulting in live births during which there had been exposure to a high concentration of HCQ (Costedoat-Chalumeau et al. 2005). There was no reported toxicity including ototoxicity. There are only a few isolated reports of longer-term follow-up of these offspring (Levy et al. 2001; Petri 2003). Thus, during this large randomized double-blind, placebo-controlled trial of chloroquine prophylaxis against vivax infections in pregnancy undertaken in Thailand, where P. vivax remains sensitive to this drug, we have paid particular attention to the development of infants of mothers who received the drug.
Study sites and populations
This study took place in the antenatal clinics (ANCs) of the Shoklo Malaria Research Unit (SMRU). Pregnant Karen women who lived in Maela Refugee Camp (60 km north of Mae Sot) or in the vicinity of Maw Ker Tai village (60 km south of Mae Sot) on the Moei River were invited to participate. The epidemiology of malaria in this area and the effects of P. falciparum and/or P. vivax malaria during pregnancy have been described in detail previously (Nosten et al. 1991, 1999a; ). Briefly, in this hilly forested region, malaria transmission is low and seasonal. P. falciparum and P. vivax are the main plasmodial species; P. malariae and P. ovale are detected only occasionally. The epidemiology of malaria changed in the 10 years preceding this study with an inversion of the P. falciparum/P. vivax ratio associated with early diagnosis and treatment with artemisinin-based combination therapy (ACT) (Nosten et al. 2000). Residents of all age groups experience clinical malaria but pregnant women and children are at higher risk of severe malaria. The main health providers in the camp are Médecins Sans Frontières (MSF) and SMRU. Outside the camp, SMRU provides ANCs and offers the non-pregnant population access to malaria diagnosis and free treatment. Pregnant women receive a supplementary ration of eggs, rice and beans which are provided by a consortium of charities.
In 1986, a system of regular antenatal clinic visits was set up in the camps to prevent maternal death from malaria. Since then, more than 90% of pregnant women in the camps have attended on a weekly basis (Nosten et al. 1991). All women are invited to come to an ANC as soon as they are aware of their pregnancy. All women who attend ANCs are screened weekly for malaria by finger prick blood sample and every 2nd week for anaemia by haematocrit. All anaemic women routinely receive supplementation with ferrous sulphate 250 mg three times a day and folic acid 5 mg once a day until delivery. All women are invited to deliver at SMRU although Karen women traditionally deliver at home.
Sample size calculations
Assuming an incidence of vivax parasitaemic episodes of 10% among pregnant women (Nosten et al. 2000), 474 women were required in each group in order to detect a 50% reduction in the incidence of malaria (10% to 5%) between the groups (CQ and placebo) with 95% confidence and 80% power. To allow for drop-out rate a final sample size of 1000 was chosen.
Criteria for inclusion in the study were a confirmed pregnancy (pregnancy test or positive fundal palpation) of any gestational age, a negative malaria smear and an ability to comply with the study protocol. Exclusion criteria were allergy to CQ, an inability to tolerate oral drugs, clinical or laboratory findings of severe renal or hepatic impairment, tuberculosis treatment, a history of epilepsy and/or diabetes mellitus or signs of labour. The purpose of the trial, treatment and randomisation procedures, the benefits and possible side effects of chemoprophylaxis with CQ were explained in Karen or Burmese languages to each pregnant woman. Written informed consent was obtained from all participants. The study was approved by the Ethics Committee of the Faculty of Tropical Medicine of Mahidol University and the Ethics Committee of the London School of Hygiene & Tropical Medicine.
Pregnant women were randomized to CQ or placebo, identical in appearance and taste. Participants were assigned unique identification numbers sequentially. All identification numbers were allocated randomly by computer to a number between one and ten, in blocks of ten (five randomly allocated to CQ and five to placebo in each block). Each unique identification number was linked to a brown paper envelope which contained the study drugs in weekly allotments, sealed into zippered plastic bags, each labelled with week number of the study. The preparation of the study drugs was done in Mae Sot by the SMRU pharmacist who was not involved with any other aspect of the study. The study codes and randomisation list was retained by a clinician at SMRU who was not involved in the study in any other way. The investigators and staff participating in the trial were unaware of the study codes until data collection was completed.
At enrolment, a full physical examination was performed and questionnaires on obstetric and medical history were completed. Data on weight, fundal height (FH), foetal presentation, blood pressure (BP), axillary temperature, foetal heart sound (FHS), anaemia, gestational age and presence of oedema were recorded.
Chloroquine treatment and prophylaxis
Four tablets of chloroquine [250 mg chloroquine phosphate (153 mg base), The Government Pharmaceutical Organization, Bangkok] or placebo (placebo of chloroquine, Mahidol University Pharmaceutical Pilot Plant, Bangkok), identical in taste and appearance, were given on enrolment. Two tablets of the same type, CQ or placebo, were then given on a weekly basis until delivery. Trial drugs were administered under direct supervision. If fever was present, trial medication was given only after malaria had been excluded as the cause of the fever. At each visit, clinical evaluation was carried out as per the routine ANC procedure (see above). Women were asked weekly about any adverse effects that might be related to medication, using a standardized recording form, until delivery. The number of days that the adverse effects were present was calculated from the date of onset to the end-date for each event. The severity of the adverse event and its relationship to drug administration were recorded. Subjects who developed malaria were excluded from the analysis of efficacy and adverse effects from that date forwards. All malaria cases diagnosed during the trial were treated according to the SMRU guidelines. For P. falciparum or mixed infections, a first episode was treated with quinine sulphate 10 mg/kg three times a day for 7 days, a second episode was treated with artesunate 2 mg/kg once per day for 7 days and subsequent episodes were treated with artesunate (as above) plus clindamycin 5 mg/kg three times a day for 7 days. P. vivax infections were treated with 25 mg base/kg of CQ given over 3 days (10, 10, 5 mg base/kg/day).
All women had a haematocrit and blood smear on admission. Thick and thin blood films were prepared by standard Giemsa staining. Parasitaemia was assessed on a thick blood film as the number of parasites per 500 white blood cells (WBCs) or on thin blood films as the number of parasites per 1000 red blood cells (RBCs). At least 200 high-power fields of a thick blood film were read before a malaria smear was considered negative.
Labour and delivery
All study women were encouraged to deliver at the SMRU delivery unit with trained midwives. Data on the outcome of deliveries (routinely conducted with WHO partograph), including abortion, stillbirth, birthweight, sex and/or congenital abnormalities were recorded on the ANC card. Newborns were weighed on a Salter scale (Salter, Birmingham, UK), accurate to 50 g, and head, arm circumference and length were measured. Estimated gestational age was assessed by the Dubowitz method (Dubowitz 1977) and, when possible, infants had a newborn neurological examination (McGready et al. 2000; Haataja et al. 2002). When a Dubowitz evaluation could not be done, estimated gestational age was calculated by a formula developed from a cohort of Karen pregnant women with known gestation age [fundal height on admission × (0.887 + 4.968)] weeks (Nosten et al. 1999a).
Growth, developmental and visual assessment
Only mothers resident in Maela Refugee Camp were invited to bring their infant back to SMRU for assessment of growth, (weight, height, arm circumference, head circumference), neurodevelopment and visual acuity when they were 1 year of age. One tester with experience in the methods performed all examinations. These examinations adapted from the Griffiths scales, Movement ABC and Hammersmith Infant neurological evaluation, have been standardized for Karen infants (McGready et al. 2000; Haataja et al. 2002) and were done in a quiet room set aside for testing.
Outcome and definitions
The primary study outcome was the first episode of P. vivax parasitaemia (regardless of the presence of symptoms) observed up to the time of delivery. There was a 1-week interval from the time of consent to participate to the trial until administration of the loading dose to allow time for laboratory results to be obtained. Some women delivered during this period. Only women with a negative baseline blood smear who had received the loading dose and at least 1 week of prophylaxis were included in the final analysis. Absences of 1 or 2 weeks were permitted but any woman who was absent for three consecutive weeks was censored before the 3rd absent week from efficacy analysis i.e. from the per-protocol analysis. Subsequent analysis of pregnancy outcome and infant follow-up included the same women. For analysis of efficacy, women were censored at the time of first malaria parasitaemia of any plasmodial species.
Anaemia was defined as a haematocrit <30%. Low birth weight was defined as a birth weight (measured within 24 h) of less than 2500 g and a premature birth was defined as delivery before 37.0 estimated weeks of pregnancy. Abortion was defined as delivery before 28 weeks of gestation and stillbirth as delivery of a dead baby after 28 weeks of gestation. Women who used CQ for ≥25 weeks of the pregnancy were defined as heavy users.
Data was entered using Dbase IV and SPSS v.11 software (SPSS, Benelux Inc., Gorinchem, The Netherlands) was used for statistical per protocol analysis. The statistical significance of categorical variables was assessed using the Chi-square or Fisher's exact test; the Student's t-test or Wilcoxon rank sum test was used for continuous variables. Survival times to the first episode of P. vivax or P. falciparum were estimated by the Kaplan–Meier method.
One thousand women were enroled into the study between November 1998 and January 2000 whilst infant follow-up was completed in December 2001. A total of 21 women in the placebo group and 28 in the CQ group were excluded from the final analysis of efficacy against P. vivax for the reasons indicated in Figure 1. Characteristics at presentation in the remaining 951 women were comparable between the CQ (n = 479) and the placebo groups (n = 472) (Table 1). A further 44 women, 22 in each group, were included in the efficacy analysis up to the time that they were censored. The reasons for censoring include: one woman in the placebo group committed suicide; one woman in the CQ group was started on tuberculosis treatment; 17 and 16 women in the placebo and CQ groups respectively were absent for three or more consecutive weeks; three women in each group withdrew consent and three women (one placebo and two CHQ) dropped out due to side effects.
Table 1. Patient characteristics on enrolment
Placebo (n = 479)
Chloroquine (n = 472)
Age, years, mean ± SD (range)
25.4 ± 6.3 (15–42)
26.1 ± 6.4 (15–47)
Gravidity, median (range)
Parity, median (range)
Weight, kg, mean ± SD (range)
48.6 ± 6.2 (33–75)
48.2 ± 6.7 (33–74)
Haematocrit, %, mean ± SD (range)
32.0 ± 3.7 (19.0–44.0)
32.1 ± 3.7 (23.0–40.0)
Residence in refugee camp, n (%)
Proportion of smokers, n (%)
Enrolment by trimester, n (%)
Trimester 1 (<13.0 weeks)
Trimester 2 (13.0 to <26.0 weeks)
Trimester 3 (26.0 weeks or more)
Malaria before enrolment
Proportion women infected % (n)
Median number episodes, median (range)
Geometric mean parasitaemia/μl (range)
1388 (32–489 779)
1439 (16–128 825)
Proportion of episodes with fever %
Proportion women infected % (n)
Median number episodes, median (range)
Geometric mean parasitaemia/μl (range)
437 (16–42 658)
Proportion of episodes with fever %
There was a small but statistically significant difference in the median number of weeks that the study drug was administered: 15 (1–32) and 17 (1–33) weeks in the placebo and CQ groups respectively (P = 0.01). This was probably related to the higher rate of censorship in the placebo than in the CQ group as there was no significant difference in the median number of weeks of drug administration: 16 (1–32) and 17 (1–33) weeks respectively among women who did not contract a P. vivax infection (P = 0.28). The proportions of women who did not have 100% of doses supervised were very similar in each group: 7.3% (35/479) and 7.4% (35/472) in the placebo and CQ groups respectively. The median (range) total dose of chloroquine phosphate received by women in the CQ group was 9500 (1500–17 500) mg or 195 (21–460) mg/kg (equivalent to 119 (13–281) mg base/kg).
Efficacy of prophylaxis on parasitaemia
There were nine women in the placebo group who had P. falciparum infection and treatment before their first P. vivax episode and who were censored. By survival analysis, the proportion of women who had at least one P. vivax infection was 8.3% (95%CI: 5.2–11.4) in the placebo group compared with 0% in the CQ group (P < 0.001) (Figure 2). The 31 P. vivax infected women in the control group had 49 distinct episodes with a median (range) of 1 (1–5) episodes and geometric mean (range) parasitaemia of 347 (16–46 774)/μl.
A history of vivax parasitaemia prior to study enrolment but during the current pregnancy, was strongly associated with an increased risk of a vivax episode during the surveillance period: RR = 8.1 (95% CI: 4.6–14.2), (P < 0.001). None of the 24 women in the CQ group, who had a history of vivax malaria prior to inclusion in the study, had a parasitaemic episode of P.vivax detected by weekly microscopy.
Similar numbers of women in each group were infected with P. falciparum malaria: 5.4% (26/479) and 4.7% (22/472) in the placebo and CQ groups respectively, P = 0.66. There was no significant difference in the severity of P. falciparum episodes before or during the period of observation between women in either group as measured by the number of days of fever, the parasite density, haematocrit, or the presence of schizonts and/or malaria pigment on the blood smear (data not shown). Three women had a history of P. malariae (one placebo and two chloroquine) before entry to the study and one woman in the placebo group was infected with this malaria parasite during the surveillance period.
Effect of prophylaxis on anaemia
The proportions of women who were anaemic on admission did not differ significantly between the groups: 23.4% (112/479) and 22.5% (106/472) in the placebo and CQ groups respectively (P = 0.76) nor did the proportions of women who were observed to be anaemic during the period of surveillance: (71.4% (342/479) and 71.2% (336/472) in the placebo and CQ groups, respectively)(P = 0.76).
Three women stopped taking the study drug due to side-effects. One woman in the placebo group complained of visual problems and refused further medication after 5 weeks of medication: her eye examination was normal. One woman in the CQ group withdrew after 2 weeks because of constipation, and another withdrew because of intolerable nausea, after 6 weeks of prophylaxis.
Adverse events questionnaires, which enquired at each visit about 25 potentially drug related adverse events, were obtained on 3397 occasions from women in the CQ group and on 2754 occasions from women who had received placebo. This large difference is explained by the higher number of women in the placebo group that were excluded from this analysis after their first P. vivax episode. The five most commonly reported adverse events by women in each group were headache, anorexia, sleep disorder, dizziness and weakness, although the rank order varied slightly between groups. There were no significant differences between the two groups in the proportions of affected women or in the length of time the event continued.
Nine hundred and fifty-one women were included in the analysis of outcome of pregnancy (Table 2). Forty-one women (4.3%), 27 in the placebo and 14 in the CQ group were lost to follow-up before their pregnancy outcome was established. There were 865 (90.9%) singleton births, 10 (1.1%) twin births, 34 (3.6%) abortions and 1 (0.1%) maternal death. The twins, were all live born and without congenital abnormalities and were not included in further analyses. There were five stillbirths (all unrelated to malaria) and nine abnormal infants (see Table 3), with no significant differences between the groups. Ninety-nine per cent (855/865) of infants were weighed but only 85.7% (733/855) of live born infants were weighed in the first 72 h and included in the birth weight analysis. There was no significant difference in mean birth weight, gestational age or in the proportion of low birth weight, premature or IUGR infants between the groups (Table 2). The total neurological optimality score for 664 tested term newborns was not significantly different between the groups.
Table 2. Birth outcomes for women included in analysis of efficacy
Placebo (n = 479)
Chloroquine (n = 472)
Percentage (n) of:
Lost to follow-up
Percentage (n) singleton pregnancy outcome
Birthweight, mean ± SD (range) grams
2777 ± 435 (1650–4200), n = 77
2741 ± 481 (1100–3700), n = 64
2956 ± 471 (1200–4000), n = 296
2954 ± 423 (1400–4000), n = 296
2919 ± 469 (1200–4200), n = 373
2916 ± 441 (1100–4000), n = 360
Percentage (n) low birth weight
Gestational age, mean ± SD (range), weeks
39.6 ± 1.4 (33.8–42.0)
39.4 ± 2.0 (29.7–41.5)
39.9 ± 1.6 (28.3–43.7)
39.9 ± 1.2 (34.0–43.0)
39.9 ± 1.6 (28.3–43.7)
39.8 ± 1.3 (29.7–43.0)
Percentage (n) premature
Percentage (n) IUGR
Neurological optimality score
17 (9–20.5) n = 334
17 (10–20.5) n = 330
Table 3. Details of congenital abnormalities in the placebo and chloroquine groups
EGA at enrolment (weeks)
Weeks study drug
EGA, estimated gestational age; G, gravidity; P, parity; Rx, treatment; PF, P. falciparum; UQ7, unsupervised quinine 10 mg/kg t.i.d. ×7 days; AS7, artesunate 2 mg/kg once daily ×7 days.
Infant follow-up was conducted only in Maela Refugee Camp; 71.8% (497/692) of eligible infants were examined at 1 year of age. The demographic details of their mothers (age, gravidity, proportion of primipara) and the characteristics of the infants (birth weight, gestational age) did not differ significantly between the groups (data not shown). Although there were no significant differences between the groups for infant growth, development and visual acuity (Table 4) there were some very low scores in the CQ group. The seven lowest scores for coordination were all in the CQ group and below the lowest value of the range for the placebo group: one infant had Down's syndrome and four others performed poorly on the coordination test but displayed normal development on all other aspects of the test. There were two further infants whose development was grossly delayed and who also scored poorly on social and speech development. Both of them had uncomplicated births and were born at term and of normal birth weight but each of their mothers had had a complicated pregnancy: cholera, a urinary infection, anaemia and severe vitamin B 1 deficiency in one case and an undiagnosed illness and anaemia in the other.
Table 4. Summary of infant growth, development and visual acuity results at 1 year of life, Thai-Burmese Border, Jan 2000-Dec 2001
Placebo (n = 246)
Chloroquine (n = 251)
Statistical tests were: * Student's t-test and †Mann–Whitney U-test.
Growth parameters, mean ± SD (range)*
Head circumference, cm
44.2 ± 1.4 (40.5–49.6)
44.2 ± 1.5 (40–49)
Arm circumference, cm
13.8 ± 1.2 (9.5–17.5)
13.8 ± 1.1 (10.5–17.5)
8230 ± 962 (6100–11 000)
8272 ± 1003 (5600–11 000)
71.4 ± 2.7 (62–79)
71.2 ± 2.7 (62.5–78.5)
Visual acuity and developmental scores, median (range)†
Median acuity cycles/degree
Social and speech
There was no difference in the mean visual acuity in infants whose mothers took chloroquine prophylaxis for ≥25 weeks (n = 54), compared with those who took it for less than 25 weeks (n = 166): 7.3 (3.9–10.8) vs. 7.3 (3.9–10.8) cycles/degree respectively (P = 0.41) (Figure 3).
This double-blind, randomized controlled trial has shown that chloroquine prophylaxis is highly effective in preventing vivax malaria episodes in pregnant Karen women. Prophylaxis was highly effective even in women who already had documented P. vivax infections prior to entry to the study and who were, therefore, likely to present with relapses (Nosten et al. 1999a). As P. vivax infection in pregnancy has previously been shown to be associated with maternal anaemia and low birthweight it might have been anticipated that this high level of protection would have been associated with reductions in the levels of maternal anaemia and low birthweight among women in our study. This was not observed, probably because only a relatively small proportion of women in the control group (8.3%) were infected. For example if prophylaxis had resulted in a 150 g increase in mean birthweight in the 44 women in the CQ group who may have been protected from malaria, this would have resulted in only a 15 g increase in the group overall and our study was not powered to detect small difference of this kind. Thus, the absence of a significant effect on birthweight seen in our study does not preclude the possibility of an important population effect if CQ prophylaxis was introduced into a large population where P. vivax malaria is highly prevalent and remains sensitive to CQ.
Chloroquine prophylaxis was recommended for many years as the primary means of preventing P. falciparum malaria in pregnancy in Africa. However, the policy was never widely implemented, in part because of the reluctance of pregnant women to take the drug. In this study, weekly CQ was well tolerated and only two women withdrew because of side effects. However, this high level of compliance may have been due in part to the high level of attendance at weekly antenatal services provided to the population. The Karen midwives who supported the trial have a close relationship with the study participants and the majority of study women have had some personal experience of the seriousness of malaria in pregnancy. Compliance might not be so high if CQ was administered through routine health services.
There are very few studies that have investigated the possible harmful effects to a mother and her infant of exposure to CQ during pregnancy, against any plasmodial species. Our results are reassuring as no evidence was found for any impairment in growth, audio or visual development of infants whose mothers had received CQ in pregnancy. There were two infants in the CQ group with severe developmental delay but these were probably unrelated to the drug as both the infants and their mothers had intercurrent health problems.
In Africa, intermittent preventive treatment has replaced chemoprophylaxis as the preferred method for the prevention of P. falciparum malaria in pregnancy, in part because of problems of compliance and also because of the spread of CQ resistance in P. falciparum. We are unaware of any studies that have investigated the role of intermittent preventive treatment in the prevention of P. vivax malaria in pregnancy. Sulphadoxine-pyrimethamine (SP), the recommended drug for intermittent preventive treatment in Africa may not be very effective because of the generally poor response of P. vivax malaria to treatment with this drug. Intermittent preventive treatment with CQ might be worth investigation in areas where P. vivax remains sensitive to this drug. In a recent study in Mali (Kayentao et al. 2005), where P. falciparum is the dominant parasite, intermittent preventive treatment with CQ was almost as effective as weekly chemoprophylaxis with this drug although neither of these regimens was as effective as intermittent preventive treatment with SP. Mefloquine is a possible alternative but its use in pregnancy was associated with an increased incidence of stillbirth in one study undertaken at the same site as that employed for the current trial (Nosten et al. 1999b). Dihydroartemisinin/piperaquine is a possible alternative that would provide protection against both P. vivax and P. falciparum infections. However, the safety of this drug in pregnancy has yet to be ascertained.
Chloroquine proved very effective at preventing P. vivax malaria in pregnancy and it would be anticipated that if this control measure was adopted as a public health intervention in areas where P. vivax is prevalent and chloroquine-sensitive, it would lead to improvements in maternal health. Acceptability to mothers and/or cost effectiveness remains to be determined.
This study was part of the Wellcome Trust – Mahidol University – Oxford Tropical Medicine Research Programme. We especially thank the mothers for their participation and patience in this study. Thanks also to the staff of the Shoklo Malaria Research Unit, the Wellcome Trust Unit, Mahidol University, Bangkok and Elizabetta Leonardi and Michele van Vugt for their monitoring of the trial. Rose McGready and François Nosten are supported by the Wellcome Trust, UK. Leopoldo Villegas was supported by the Ministerio de Salud de Venezuela (Proyecto Control de Enfermedades) and the UNDP/World Bank/WHO Special Programme for Research training in Tropical Diseases (Research Training Grant).