Effects of Plasmodium vivax malaria in pregnancy

Effects of Plasmodium vivax malaria in pregnancy F Nosten, R McGready, J A Simpson, K L Thwai, S Balkan, Thein Cho, L Hkirijaroen, S Looareesuwan, N J White

Summary Background Plasmodium vivax is more common than P falciparum as a cause of malaria in many parts of the tropics outside Africa. P falciparum infection has harmful effects in pregnancy, but the effects of P vivax have not been characterised. We investigated the effects of P vivax infection during pregnancy. Methods Since 1986, pregnant Karen women living in camps for displaced people on the western border of Thailand have been encouraged to attend antenatal clinics. Karen women were screened for malaria and anaemia at each week of pregnancy until delivery, and pregnancy outcome recorded. We compared the effects of P vivax infection on anaemia and pregnancy outcome with those of P falciparum and no malaria infection in the first pregnancy recorded at the antenatal clinics. Findings There were 634 first episodes of pure P vivax malaria in 9956 women. P vivax malaria was more common in primigravidae than in multigravidae and was associated with mild anaemia and an increased risk of low birthweight (odds ratio 1·64 [95% CI 1·29–2·08], p<0·001). The birthweight was a mean of 107 g (95% CI 61–154) lower in women with P vivax infection than in uninfected women. By contrast with P falciparum malaria, the decrease in birthweight was greater in multigravidae. P vivax malaria was not associated with miscarriage, stillbirth, or with a shortened duration of pregnancy. Interpretation P vivax malaria during pregnancy is associated with maternal anaemia and low birthweight. The effects of P vivax infection are less striking than those of P falciparum infection, but antimalarial prophylaxis against P vivax in pregnancy may be justified. Lancet 1999; 354: 546–49

Introduction Of the four species of malaria parasite that infect human beings (Plasmodium falciparum, P vivax, P malariae, and P ovale), P falciparum is the most studied. P vivax is prevalent in Asia and South America, and may be more common in these areas than P falciparum. In pregnancy, only the harmful effects of infection with P falciparum have been recognised. Maternal mortality associated with P falciparum malaria is highest in areas of low and unstable transmission or in epidemics, whereas in areas of higher transmission the main complications of P falciparum malaria during pregnancy are maternal anaemia and low birthweight.1 These effects are most severe in primigravidae, but in areas of low transmission,women of higher gravidity are also affected.2 By contrast, the effects of the other malaria species in pregnancy have not been characterised. We studied the effects of P vivax in pregnancy.

Patients and methods Study population Karen women, living in open camps to the north and south of Mae Than, a town on the northwestern border of Thailand were studied. The total population in these camps increased from 9000 in 1986 to more than 60 000 in 1997. Multidrug-resistant P falciparum and chloroquine-sensitive P vivax are endemic in the area. In this population, malaria affects all age-groups and most parasitaemic episodes are symptomatic. The two groups most at risk of malaria complications are children and pregnant women.3 Before 1986, the malaria-related maternal mortality in this community was estimated to be more than 1% of all pregnant women per year. Since then, an intensive programme of weekly antenatal screening and treatment of all parasitaemic episodes lowered the mortality.2 Information from this 11-year period (1986–97) on each malaria episode and the outcome of pregnancy has been collected prospectively in the antenatal clinics. In each of the camps, pregnant women were invited to attend weekly antenatal clinics for physical examination and blood screening by microscopy for malaria parasites. Blood slides were stained with giemsa and read for 200 fields before being classified as negative. Haematocrit was measured every 2 weeks in some camps. P vivax was treated with chloroquine phosphate (25 mg base/kg) for 3 days. Anaemia (defined clinically or by haematocrit <30% at any stage of pregnancy) was treated with ferrous sulphate and folic acid until delivery. At delivery, information on birthweight, sex, and Dubowitz score4 of gestational age was collected within 24 h and stored on a database.

Malaria and outcomes Shoklo Malaria Research Unit (SMRU), Mae Sot, Thailand (F Nosten MD , R McGready MD, K L Thwai, T Cho, L Hkirijaroen MB ); Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand (F Nosten, R McGready, J A Simpson MSc, S Looareesuwan MD, Prof N J White FRCP); Centre for Tropical Medicine, Nuffield Department of Medicine, John Radcliffe Hospital, Headington, Oxford, UK (F Nosten, R McGready, Prof N J White); and Médecins sans Frontières, Paris, France (S Balkan) Correspondence to: Dr F Nosten, Shoklo Malaria Research Unit, PO Box 46, Mae Sot 63110, Thailand (e-mail: [email protected])

We defined P vivax malaria cases as the presence of asexual forms of P vivax in peripheral blood. Symptomatic P vivax episodes were defined by the presence of fever or history of fever, with one or more of headache, joint pain, or abdominal pain. The rate of reinfection or relapse was defined as the proportion of patients with a second P vivax episode within 28–63 days of the first episode. Stillbirth was defined as delivery after more than 28 weeks’ gestation of a fetus that did not breathe, and miscarriage as the expulsion of a fetus before 28 weeks’ gestation. The Dubowitz score allowed retrospective calculation of the gestational age at the time of infection. In this population, the relation between gestational age and fundal height was linear:

Northern camps Southern camps Primigravida Gravida 2–3 Gravida >3 Gravida unknown Mean (SD) age (years)* Mean (SD) gestational age (weeks)† Median number of attendances (range)

Control (n=7447)

P vivax (n=634)

p

3016 (40·5%) 4431 (59·5%) 1692 (22·7%) 2535 (34·0%) 2942 (39·5%) 278 (3·7%) 26·0 (6·3) 19·1 (7·5) 15 (1–51)

333 (52·5%) 301 (47·5%) 192 (30·3%) 180 (28·4%) 246 (38·8%) 16 (2·5%) 25·5 (6·7) 15·9 (7·6) 18 (1–36)

<0·001 <0·001

<0·063 <0·001 <0·001

*Control n=7377, P vivax n=632. †At first antenatal-clinic consultation, control n=4583, P vivax n=433.

Table 1: Baseline characteristics and antenatal-clinic attendance of women in the P vivax and control groups gestational age (weeks)=fundal height (cm)⫻0·887+4·968; (95% of normal range: predicted gestational age±6·26 weeks). We used this equation to calculate the gestational age at any time in the pregnancy from the fundal height when the Dubowitz score had not been done. Birthweight was measured in the first 24 h to within 50 g. Low birthweight was defined as a bodyweight less than 2500 g.

Statistical analysis We entered all data on a database (Dbase IV). Analyses were done with SPSS for Windows (version 8.1). We excluded twins from the analysis of outcome. The records of the first 1358 women seen between 1986 and 1989 and reported earlier2 were included in this analysis. We compared pregnant women who had first pure P vivax infections with women who had no evidence of malaria during pregnancy. To compare normally distributed continuous variables we used Student’s t test, and for non-normally distributed variables we used the Mann Whitney U test. Categorical variables were compared by χ2 test, with Yates’ correction for dichotomous variables. Univariate analysis identified variables associated with maternal anaemia, low birthweight, stillbirth, and miscarriage. We used multiple logistic regression to adjust for any identified confounding variables. The variable antenatal-clinic attendance was square-root transformed before being included in the multiple-regression analysis.

Results Between July, 1986, and April, 1997, 11 004 pregnancies were monitored in the antenatal clinics. We excluded 1048 women who had previous pregnancies recorded in antenatal-clinic records. 9956 women with first recorded pregnancies were, therefore, eligible for analysis. The median gravidity was three (range 1–19). There were 2448 (25%) primigravidae and 7130 (72%) multigravidae. Gravidity was unknown in 378 (4%) women. 44·4% of the records were from the northern camps and 55·6% from the southern camps. Mean age was 26 years (SD 7). The mean gestational age at presentation to the antenatal clinic was 19 weeks (8) and average attendance was once every 8 days, which gave a median of 15 consultations per woman (1–51). Anaemic

P vivax No malaria Primigravida Gravida 2–3 Gravida >3 Northern camps Southern camps Mean (SD) age (years) Median antenatal-clinic attendance (range)

Yes

No

249 (77·3%) 1242 (61·2%) 320 (54·2%) 537 (65·2%) 622 (67·7%) 861 (65·5%) 630 (60·7%) 26·1 (6·63) 20 (1–39)

73 (22·7%) 789 (38·8% ) 270 (45·8%) 286 (34·8%) 297 (32·3%) 454 (34·5%) 408 (39·3%) 24·8 (6·33) 11 (1–36)

Mean haematocrit during pregnancy for women infected with P vivax only, P falciparum only, or no malaria

7447 women had no documented malaria episodes during pregnancy (controls), 1402 had only P falciparum malaria, 473 had mixed infections, and 634 had only P vivax infections. Pregnant women whose first episode was caused by P vivax had a lower risk than women whose first episode was caused by P falciparum of presenting subsequently with P falciparum parasitaemia (relative risk 0·48 [95% CI 0·40–0·58], p<0·001). 485 (77%) of 634 women in the P vivax group had only one parasitaemic episode, whereas 149 (23%) had two or more episodes. Only 194 (23%) of 847 P vivax episodes were symptomatic. Women infected with P vivax were slightly younger, more likely to be primigravid, to present earlier in their pregnancy, and had a better compliance to attendance at the antenatal clinic than the women with no malaria (table 1). The proportion of infected women was higher in the northern camps than in the southern camps (p<0·001). After adjustment for the above confounders, primigravidae were 1·63 (1·37–2·06) times more likely to be infected with P vivax than multigravidae (p<0·001). Of 322 women infected with P vivax only, for whom serial haematocrit data were available, 249 (77·3%) were anaemic compared with 1242 (61·2%) of 2031 controls (p<0·001). After multiple-regression analysis, P vivax remained independently associated with an increased risk of anaemia (odds ratio 1·91 [1·42–2·56], p<0·001; table 2). This association persisted throughout pregnancy, although the decrease in haematocrit was less than that with low haematocrit in P falciparum malaria (figure). Low birthweight was associated with primigravidity (p<0·001), less antenatal-clinic attendance (p<0·001), and P falciparum and P vivax infections (tables 3 and 4). The proportions of low-birthweight babies in P vivax infected women with and without symptoms were similar (31 [23%] of 136 vs 65 [17%] of 381, p=0·18). Overall, there was no correlation between gestational age at time of Odds ratio (95% CI)

Adjusted odds ratio (95% CI)*

p†

2·17 (1·65–2·86) 1·00 1·00 1·58 (1·28–1·97) 1·77 (1·43–2·19) 1·23 (1·04–1·45) 1·00 1·03 (1·02–1·05) 1·69 (1·58–1·81)‡

1·91 (1·42–2·56) 1·00 1·00 1·40 (1·10–1·78) 1·30 (0·94–1·79) 1·29 (1·07–1·55) 1·00 1·02 (1·00–1·04) 1·68 (1·57–1·81)‡

<0·001

0·007 0·114 0·007 0·037 <0·001

*In each case, odds ratio was adjusted for other variables listed. †p value derived from multiple logistic regression. ‡Odds ratio associated with increase in 1 square-root unit of number of attendances.

Table 2: Factors associated with anaemia in pregnancy

Low birthweight

P vivax No malaria Primigravida Gravida 2–3 Gravida >3 Northern camps Southern camps Mean (SD) age (years) Median antenatal-clinic attendance (range)

Yes

No

98 (18·7%) 741 (13·2%) 320 (21·6%) 247 (11·7%) 242 (10·3%) 348 (13·6%) 491 (13·7%) 24·6 (6·5) 14 (1–35)

427 (81·3%) 4878 (86·8%) 1159 (78·4%) 1863 (88·3%) 2116 (89·7%) 2206 (86·4%) 3099 (86·3%) 26·1 (6·2) 17 (1–51)

Odds ratio (95% CI)

Adjusted odds ratio (95% CI)*

p†

1·51 (1·20–1·91) 1·00 1·00 0·48 (0·40–0·58) 0·41 (0·35–0·50) 1·00 (0·87–1·16) 1·00 0·96 (0·95–0·97) 0·81 (0·77–0·86)‡

1·64 (1·29–2·08) 1·00 1·00 0·52 (0·43–0·63) 0·46 (0·36–0·60) 1·06 (0·91–1·24) 1·00 1·00 (0·98–1·01) 0·80 (0·76–0·86)‡

<0·001

<0·001 <0·001 0·446 0·628 <0·001

*In each case, odds ratio adjusted for other variables listed. †p value derived from multiple logistic regression. ‡Odds ratio associated with increase in 1 square-root unit of number of attendances.

Table 3: Factors associated with low birthweight

infection and birthweight in women with only one documented P vivax episode (r=–0·026, p=0·67). P falciparum infections were associated with an increased risk of low birthweight in primigravidae and in multigravidae, and P vivax also had harmful effects on birthweight in all gravidity groups (table 4). After controlling for age, site, gravidity, and attendance at antenatal clinic, the mean birthweight of babies born to women infected with P vivax was lower than that of babies born to uninfected women (2847 [516] vs 2954 g [513], p<0·001; mean difference –107 g [95% CI –154 to –61]). The difference was significant only in multigravidae (table 4) and remained when women with only a single P vivax episode were compared with the control group (2888 [512] vs 2954 g [513], p=0·015 after stratification; mean difference –65·7 g [–119 to –13]). Mean gestational age at delivery and the proportion of pregnancies ending in stillbirth did not differ significantly between the P vivax and control groups (p=0·96 and p=0·30, respectively), but the rate of miscarriage was higher in the control group (44 [6·3%] of 7098 vs 25 [4·1%] of 604, p=0·04). No data were available on the rate of congenital malformations.

Discussion P vivax was associated with mild maternal anaemia, and significantly decreased birthweight compared with babies born to women who had no documented malaria during pregnancy. These harmful effects could not be accounted for by other confounding factors. They were, however, balanced by an apparent protective effect by P vivax infection against subsequent episodes of P falciparum malaria. A similar protective effect was seen in children of the same population recruited in a malaria vaccine trial (C Luxemburger, unpublished data). An immunological interaction between P vivax has been suggested.5 As for P falciparum malaria, primigravidae were more at risk of

being infected by P vivax than multigravidae. Whether this effect reflected a higher risk of relapse from the liver stage of an infection acquired before pregnancy, or an increased susceptibility to new infection because of pregnancy cannot be distinguished with certainty. The latter is more likely since the relapse rate in pregnant women was lower than that in non-pregnant women in this region (unpublished observations). The anaemia associated with P vivax malaria was more striking in primigravidae but was also evident in multigravidae. Therefore, the increased susceptibility of primigravidae to P vivax infections and the resulting anaemia, and the fact that multigravidae were also affected, are similar to the harmful effects of P falciparum described previously. 2 The anaemia induced by P vivax was less severe than that caused by P falciparum. Infection with P vivax during pregnancy was not associated with shorter gestation or with an increased rate of miscarriage or stillbirth, although the antenatal-clinic screening programme was intensive and curtailed infections at an early stage. The apparently high rate of miscarriages in the control group can be explained by the higher risk of acquiring malaria in pregnancies that were not interrupted prematurely. P vivax was associated with a mean decrease of birthweight of more than 100 g and, therefore, an increased risk of low birthweight. In this area, P falciparum was associated with a mean decrease in birthweight of 135 g, although this decrease was more pronounced in primigravidae, whereas in women with P vivax malaria, the difference compared with controls was greater in multigravidae. Low birthweight is associated with increased infant mortality.6 The mechanisms underlying the adverse effects of malaria in pregnancy are not well understood. P falciparum sequesters in the placenta and may damage the fetus through interference with fetomaternal nutrient exchange7 and local immunological changes that lead to

P falciparum Mean (SD) gestational age (weeks)

38·4 (3·2)

P vivax 38·6 (2·6)

Control 38·6 (3·2)

Primigravidae Mean (SD) birthweight (g) Proportion with low birthweight (%) Odds ratio for risk of low birthweight (95% CI, p)†

2568 (590)* 102/322 (31·7%) 1·83 (1·39–2·40, <0·001)

2748 (578) 40/155 (25·8%) 1·50 (1·01–2·22, 0·042)

2784 (517) 280/1324 (21·1%) 1·00

Gravida 2–3 Mean (SD) birthweight (g) Proportion with low birthweight (%) Odds ratio for risk of low birthweight (95% CI, p)†

2864 (515)* 66/342 (19·3%) 1·87 (1·38–2·55, <0·001)

2859 (534)* 28/155 (18·1%) 1·83 (1·18–2·84, 0·007)

2983 (507) 219/1955 (11·2%) 1·00

Gravida >3 Mean (SD) birthweight (g) Proportion with low birthweigh (%) Odds ratio for risk of low birthweight (95% CI, p)†

2937 (531)‡ 58/395 (14·7%) 1·58 (1·15–2·17, 0·005)

2957 (569)* 29/203 (14·3%) 1·68 (1·10–2·57, 0·003)

3029 (512) 213/2155 (9·9%) 1·00

*p<0·01; compared with control group. †Compared with controls. ‡p<0·05; compared with control group.

Table 4: Effect of malaria on pregnancy outcome by gravida

secretion of potentially harmful type 1 cytokines (interleukin 2, interferon gamma, and tumour necrosis factor).8,9 Sequestration results from cytoadherence, which may be compounded by rosetting and agglutination. Of these, only rosetting is known to occur in P vivax.10 Women who were parasitaemic with P vivax only once in their pregnancy had low-birthweight babies, which suggests either that P vivax is able to sequester for a long time in the placenta or to relapse at undetectable concentrations, or that parasite sequestration in the placenta is not the major cause of the adverse impact of uncomplicated P falciparum or P vivax malaria on the fetus. In maternal-fetal immunological interactions, type 2 cytokines (interleukins 4, 6, and 10) are thought to be protective to the fetus.11 P vivax may trigger a shift towards a type-1 response, with similar consequences to those described in P falciparum infections, in which raised concentrations of type 1 proinflammatory cytokines in the placenta lead to lower birthweights.9 A switch to a Thelper-1 type response could have adverse consequences on the placenta, fetus, or both that persist long after parasitisation. Our findings should encourage more studies of P vivax, P malariae, and P ovale malaria in pregnancy. Chemoprophylaxis against P vivax malaria in pregnancy may be justified.

programme overall and contributed to the analysis and the preparation of the manuscript.

Acknowledgments We thank the Karen staff of the Shoklo Malaria Research Unit who did most of the clinical and laboratory work. This study was part of the Wellcome Mahidol University Oxford Tropical Medicine Research Programme, funded by the Wellcome Trust, UK.

References 1 2

3

4 5

6

7

8

9

Contributors F Nosten set up the antenatal clinics in the Karen camps and was the principal investigator. R McGready ran the antenatal clinic for the Shoklo Malaria Research Unit and S Balkan for Médecins sans Frontières. J A Simpson analysed the data. K L Thwai was responsible for data entry and management. Thein Cho and L Hkirijaroen were in charge of the local staff in the antenatal clinics and the laboratory, respectively. S Looareesuwan and N J White were responsible for the research

10

11

Brabin BJ. An analysis of malaria in pregnancy in Africa. Bull World Health Organ 1983; 61: 1005–16. Nosten F, ter Kuile FO, Maelankirri L, Decludt B, White NJ. Malaria during pregnancy in an area of unstable endemicity. Trans R Soc Trop Med Hyg 1991; 85: 424–29. Luxemburger C, Thwai KL, White NJ, et al. The epidemiology of malaria in a Karen population on the western border of Thailand. Trans R Soc Trop Med Hyg 1996; 90: 105–11. Dubowitz LMS, Dubowitz V. Gestational age of the newborn. London: Addison-Wesley Publishing Company, 1977. Maitland K, Williams TN, Newbold CI. Plasmodium vivax and P falciparum: biological interactions and the possibility of cross-species Immunity. Parasitology 1997; 13: 227–31. Bloland P, Slutsker L, Steketee RW, Wirima JJ, Heymann DL, Breman JG. Rates and risk factors for mortality during the first two years of life in rural Malawi. Am J Trop Med Hyg 1996; 55: 82–86. Watkinson M, Rushton DI. Plasmodial pigmentation of placenta and outcome of pregnancy in West African mothers. BMJ 1983; 287: 251–54. Fried M, Duffy PE. Adherence of Plasmodium falciparum to chondroitin sulfate A in the human placenta. Science 1996; 272: 1502–04. Fried M, Muga RO, Misore AO, Duffy PE. Malaria elicits type 1 cytokines in the human placenta: IFN-gamma and TNF-alpha associated with pregnancy outcomes J Immunol 1998; 160: 2523–30. Udomsangpetch R, Thanikkul K, Pukrittayakamee S, White NJ. Rosette formation by Plasmodium vivax. Trans R Soc Trop Med Hyg 1995; 89: 635–37. Wegmann TG, Lin H, Guilbert LJ, Mosmann TR. Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a Th2 phenomenon? Immunonol Today 1993; 14: 353–56.