Parasitological efficacy of antimalarials in the treatment and prevention of falciparum malaria in pregnancy 1998 to 2009: a systematic review

Authors


Dr R McGready, Shoklo Malaria Research Unit (SMRU), PO Box 46 Mae Sot, Tak 63110, Thailand. Email rose@shoklo-unit.com

Abstract

Please cite this paper as: McGready R, White N, Nosten F. Parasitological efficacy of antimalarials in the treatment and prevention of falciparum malaria in pregnancy 1998 to 2009: a systematic review. BJOG 2011;118:123–135.

Background  Pregnant women are at increased risk from malaria. Resistance to all classes of antimalarials has affected the treatment and prevention of malaria in pregnancy.

Objectives  To review the therapeutic efficacy of antimalarials used for treatment and intermittent preventive treatment (IPT) in pregnancy.

Search strategy  We searched MEDLINE and the Cochrane Library between January 1998 and December 2009 for publications using the medical subject headings: efficacy, antimalarials, malaria, pregnancy, pharmacokinetics, treatment, IPT and placenta positive. In May 2010 we searched the register of clinical trials (http://clinicaltrials.gov/) and of WHO (http://apps.who.int/trialsearch/) using ‘malaria’, and ‘pregnancy’ and ‘treatment’.

Selection criteria  We identified 233 abstracts, reviewed 83 full text articles and included 60 studies.

Data collection and analysis  Two authors entered extracted data to an excel spreadsheet.

Main results  Parasitological failure rates, placenta positivity rates (assessed by microscopy) or both were reported in 44% (21/48), 46% (22/48) and 10% (5/48) of articles, respectively. Most pharmacokinetic studies (9/12) suggested dose optimisation. In 23 treatment studies 17 different antimalarial drugs were delivered in 53 study arms; 43.4% (23/53) reported a failure rate of < 5%; 83.3% of sulphadoxine-pyrimethamine (SP) arms and 9% of artemisinin combination therapy (ACT) arms had failure rates ≥ 10%. Placenta-positive rates (mostly reported in the context of IPT in pregnancy) were > 10% in 68% (23/34) of SP trial arms and > 15% in all seven chloroquine arms. The ACT provided lower parasitological failure and gametocyte carriage rates.

Author’s conclusions  Drugs used in pregnancy should aim for 95% efficacy but many currently deployed regimens are associated with much lower cure rates.

Introduction

The African Roll Back Malaria Summit in April 2000 adopted the Abuja Declaration in which regional leaders committed to ensuring that 60% of pregnant women in communities where malaria is endemic should have access to effective prevention and treatment of malaria by 2005.1 It is estimated that about 25 million pregnant women are at risk of malaria each year and one in four women in SubSaharan Africa stable transmission areas have placental malaria at the time of giving birth.2 Approximately 10 000 pregnant women die from malaria-related anaemia per year.3 The three components of the control strategies are: effective antimalarial drugs, insecticide-impregnated bednets and indoor residual spraying. The drugs are used either for the treatment of malaria or for malaria prevention.

The World Health Organization (WHO) guidelines recommend an artemisinin-based combination therapy (ACT) for the treatment of uncomplicated Plasmodium falciparum malaria and they are based on evidence.4 They recommend a change in the ‘national malaria treatment policy if the total treatment failure proportion is ≥ 10% as assessed through in vivo monitoring of therapeutic efficacy’. New or alternative antimalarials should have a parasitological cure rate of > 95% determined from active follow up (usually weekly) in treatment trials for a fixed interval (usually 42 [28–63] days).4 Even without symptoms, malaria can cause adverse outcomes for both the mother and fetus.5 An effective treatment cures the woman of infection by eliminating parasites from peripheral and placental blood. Intermittent preventive treatment (IPT) is the delivery of a curative treatment dose of an antimalarial at predefined intervals, regardless of the parasitological status of the women. To be truly preventive, IPT in pregnancy (IPTp) must eliminate all parasites in the woman’s blood and placenta and prevent any new parasitaemia. The current recommendation1 for IPTp is sulfadoxine-pyrimethamine (SP): ‘at least two doses after quickening (18–20 weeks) and not more frequently than monthly’. It is uncertain whether the efficacy of IPTp should also fulfil the same WHO recommended criteria of a parasitological cure rate of > 95% over 28 days.4 Pregnancy also affects the pharmacokinetic properties of many drugs,6 including antimalarials,7 and this has been demonstrated to affect efficacy.8

Although miscarriage, stillbirth and preterm birth are frequently associated with malaria in pregnancy9–12 this may not result directly from parasitaemia per se but be mediated by the disease effects of malaria, with symptoms such as fever and maternal anaemia.13 The cause of maternal death from malaria has been described as either direct: an acute systemic illness leading to severe disease and death, more frequent in low or hypo-endemic areas,14,15 or indirect: chronic parasitisation resulting in severe maternal anaemia and death from congestive cardiac failure or postpartum haemorrhage at delivery, a feature of high-transmission areas.16 Malaria during pregnancy causes a reduction in birthweight in all transmission areas, regardless of symptoms. While P. falciparum causes greater morbidity (maternal anaemia and low birthweight) and mortality than non-falciparum infections,2,15,17–19 there is mounting evidence that P. vivax is not as benign as previously thought.20–22

The primary endpoint of trials of antimalarials given as IPTp is usually birthweight or anaemia but what evidence is available on parasitological efficacy? We have reviewed publications from the past 12 years in pregnant women that report parasitological efficacy by comparable standards to nonpregnant people as advocated by WHO, and placenta-positive rates. We also review ongoing and planned trials in pregnancy.

Methods

In March 2010 we searched MEDLINE and the Cochrane Library to identify potentially relevant publications, in English, between January 1998 and December 2009, using the medical subject headings: efficacy, antimalarials, malaria, pregnancy, treatment, intermittent preventive treatment in pregnancy, chemoprophylaxis, pharmacokinetics and individual antimalarial drug names. The year 1998 was chosen as the starting point because SP for IPT was strongly advocated in that year.23,24 In September 2010 we searched the register of clinical trials (http://clinicaltrials.gov/) and the WHO international clinical trials registry platform (http://www.who.int/ictrp/en/) using ‘malaria’, and ‘pregnancy’ and ‘treatment’ and summarised the main features of currently recruiting or proposed trials.

We extracted three measures of parasitological efficacy results. The first was failure rates obtained when the woman was included with a confirmed parasitaemia with uncomplicated malaria and treated and then followed up at a minimum to day 14 with parasitological confirmation.25 In these manuscripts we aimed to extract the type of study, sample size including number of women screened, the type of treatment/IPT, microscopy-based recrudescence rates and polymerase chain reaction (PCR) genotyping confirmation rates, and the length of follow up. The second measure was transmission potential after treatment (gametocyte carriage) in the manuscripts that reported failure rates. The third measure was microscopy-based placenta malaria positivity rates. These were extracted from the manuscripts reporting failure rates (when available) and from a further set of manuscripts that reported IPTp-SP use. Those IPT trials that did not include SP in at least one arm of the trial were not included in this review. For parasitological confirmation, we excluded trials that followed the woman passively after treatment because WHO recommends active follow up to determine the day of reappearance of parasitaemia. As resistance worsens treatment failures manifest progressively earlier after initial treatment.25 For microscopy-based placenta positivity rates no exclusions were applied. All pharmacokinetic studies on antimalarials in pregnancy published in the same time frame were included regardless of efficacy reporting.

Titles and abstracts were reviewed for possible exclusion by two reviewers (RM and FN). If both reviewers excluded a citation, then that publication was excluded from further review. The remaining articles were reviewed in full. Most of the abstracts were excluded because they did not report parasitological efficacy according to the inclusion criteria, or because they were studies using the same data as the original trial, reported severe malaria, did not include pregnant women, or were commentaries, letters or abstracts. We cross-checked the reference list of all related systematic reviews for possible additional studies.

We included randomised controlled trials (RCT), cohort studies, surveys and pharmacokinetic studies. As the focus of this study was parasitological cure and placenta positivity with antimalarial use, all studies corresponding to the predefined criteria from 1998 onwards were included. The quality of trials was not assigned. Priority was given to PCR-adjusted failure rates when reported for purposes of comparison. The RCTs are indicated by asterisks in Table 180–86,88–90,92–102 and by shading in the Supporting information tables.

Table 1.   The proportion of placenta-positive women (%) according to antimalarial prescribed, sample size and parasitological failure rates
ReferenceSP (+other) IPTCQOther
  1. AQ, amodiaquine; AS, artesunate; AZ, azithromycin; COA, artemether-lumenfantrine; CQ, chloroquine; n, sample size; na, not available; PYR, pyrimethamine; SP, sulphadoxine-pyrimethamine; TS, trimethoprim-sulphamethoxazole.

  2. Parasitological failure rates (polymerase chain reaction confirmed [bold] or unconfirmed [non-bold]) are given in parentheses when available.

  3. *Randomised controlled trials. Tx indicates treatment trials the remainder being IPT-SP trials or surveys.

  4. **Placenta parasitaemia for all women, that is not specified in relationship to the small number recruited in the efficacy sub-study.

30*
Tx
  5.1 COA n = 78 (13)
1.3 AS n = 78 (8)
34*
Tx
16.1 SP n = na (35)
27.3 SP + AZ n = na (10)
11.4 SP + AS n = na (8)
  
35*
Tx
13 SP n = 174 (11)
15 SP + AQ n = 176 (0)
20.0 CQ TX n = 184 (14)12 AQ n = 177 (3)
80*10.6 SP IPTp2 dosesn = 14118.8 CQ weekly n = 137
15.9 CQ IPTp2 dosesn = 145
 
81*4.4 SP IPT2 dosesn = 656 1.7 MQ IPT2 dosesn = 663
82*12 SP IPT2 dosesn = 617  
83*,**28.7 SP IPTup to 3 dosesn = 202 (6.1)
27.6 SP + AQ IPTup to 3 dosesn = 308 (3.2)
 20.8 AQ IPTup to 3 dosesn = 138 (0)
84*7.0 ITN + SP IPT2 dosesn = 426 13.6 ITN + Placebo n = 419
85*8.3 SP IPT2 dosesn = 15120.6 CQ IPT2 dosesn = 150 
86*4.0 SP IPT3 dosesn = 232
2.0 SP IPTmonthlyn = 224
  
88*21.5 SP IPT2 doses HIV+ n = 93
6.3 SP IPT2 doses HIV−n = 128
7.8 SP IPTmonthly HIV+ n = 102
2.3 SP IPTmonthly HIV−n = 175
  
89*24.5 SP IPT2 dosesn = 36932.1 CQ IPT2 dosesn = 360
29.7 CQ weekly n = 381
 
90*13.3 SP IPT-SP2 dosesn = 120 2.4 Placebo n = 124
52*12 SP IPT2 doses (< 2) n = na
7 SP IPTmonthly (< 2) n = na
  
489.0 SP IPTp2 doses HIV−n = 336 6.0 TS daily HIV+ n = 150
92*3.0 SP IPT2 dosesn = 58518.8 CQ weekly n = 1090 
9315.3 SP IPT2 dosesn = 59 34.9 PYR weekly n = 839
9410.5 IPT SP2 dosesn = 598 16.8 PYR weekly n = 294
17 (no IPT/chemo) n = 171
95SP IPT2 dose + Haematinics
11.8 HIV+ n = 178
10.6 HIV−n = 618
 Haematinics 29.3 HIV+ n = 257
16.2 HIV−n = 883
Nil 24.8 n = 308
16.0 HIV−n = 864
9612.7 IPT-SP≥ 2 dosesn = 55 HIV+
7.9 IPT-SP≥ 2 dosesn = 164 HIV−
 19.4 No SP n = 72 HIV+
12.6 No SP n = 206 HIV−
9719.7 IPT SP2 dosesn = 196  
9813.7 IPT-SP3 dosesn = na 30.3 No IPT SP n = na
9919.0 SP(used in 2.1%)n = 2502  
10022.8 SP IPT≥ 2 dosesn = 303 33.2 No SP n = 232
10160.5 IPT SP2 doses HIV+ n = na
20.8 IPT SP2 doses HIV−n = na
 19.0 No IPT HIV+ n = na
12.8 No IPT HIV−n = na
10237.0 IPT SP0–2 dosesn = 173
(only primiparas and 86.4%≥ 1 dose)
  

Results

We identified 233 abstracts, retrieved 83 full text articles and included 60 that fufilled the inclusion criteria (Figure 1). The 48 studies included for efficacy are listed in Supplementary material Table S1 by (a) treatment (n = 23) and (b) IPT (n = 25).77–102 Nine studies (12 publications) on pharmacokinetics including 390 pregnant women have been reported separately because of overlap with some of the efficacy studies.

Figure 1.

 Flow diagram.

A total of 29 176 pregnant women were included in the efficacy review: 13.2% (3845) in 23 treatment studies (12 RCT and 11 cohort) and 86.8% (25 331) in 25 IPTp-SP prevention trials or surveys (12 RCT and 13 surveys) (Table S1). The studies were lengthy, taking an average of 22.5 (2–97) months to complete. The median number of women per arm in treatment studies was 85 (18–258). Parasitological failure rates, placenta-positive rates by microscopy or both were reported in 44% (21/48), 46% (22/48) and 10% (5/48), respectively.

Parasitological efficacy

There were 53 study arms with 17 different antimalarial drugs in the 23 treatment studies, not accounting for dosing differences (Table S2). The studies were for uncomplicated malaria but the inclusion criteria still varied. Some of the treatment studies in both Africa26–29 and Asia30–33 intentionally included women who had failed a prior treatment. The drug treatment they had failed was variable. Sowunmi et al. only included women (100%; 45/45) who had failed chloroquine or SP or both.26 In the latest published trial from the Thai-Burmese border more than half the women (56.1%; 142/253) had PCR-confirmed failure of quinine or artesunate ± clindamycin30 at entry to the trial. Higher failure rates would be anticipated in such trials compared with treating women with their primary infection in pregnancy. The reporting of parasitological efficacy varied: PCR confirmation (one study heteroduplex tracking34) was available for 75% (8/12) of randomised controlled treatment trials and 25% (3/12) of randomised controlled IPT efficacy trials (Table S2). More than half of the women included in the 12 randomised controlled treatment trials came from one of two sites: one site 34% (838/2496) in a single study35 and the other site 22% (557/2496) in four studies30,36–38 (Table S2).

Most antimalarial treatment arms were monotherapies, 70% (37/53); ACTs comprised 21% (11/53) and non-artemisinin-based combinations were 9% (5/53) (Table S2). The single most common antimalarial treatment arm was SP monotherapy 22.6% (12/53) (Table S2). Each treatment arm in Table S2 was categorised into failure rates of < 5%, 5–10%, and ≥ 10% failure rates (Figure 2). Overall 41.5% (22/53) of study arms achieved a WHO failure rate of < 5%. Of the SP arms 58.3% had failure rates ≥ 10% and most of these arms had short (14-day) follow-up periods. There were 9% of women treated with an ACT who had failure rates ≥ 10%. Poor cure rates were observed with the combination of SP and azithromycin in an RCT in 42 women from Malawi, where 10% of women had PCR-confirmed failures by day 2834 (Table S2).

Figure 2.

 The proportion of reported antimalarial efficacy results with failure rates at < 5% (□), 5 to < 10% (▮) and ≥ 10% (▮) in pregnant women from 1998 to 2009. A, artemisinin; ACT, artemisinin combination therapy; CT, combination therapy; mono, monotherapy.

On the Thai-Burmese border, an area with multidrug resistant P. falciparum, artemether-lumefantrine (3 days) and artesunate (7 days) both had PCR-confirmed parasitological failure rates of > 5% (Table S2) by the time of delivery.30 The PCR-confirmed cure rates at day 42 and delivery for artemether-lumefantrine and artesunate rose from 13% and 8%, to 18% and 11%, respectively. More than one-third of the PCR-confirmed reappearances occurred after day 42. Prolonged time to reappearance has been reported in pregnant women from Africa and Asia and was reported in three of the trials included in this review (133 days,34 98 days30 and 85 days37) and in two further studies that are not part of this review (187 days,39 121 days40).

There were six studies (five treatment and one IPTp81) using mefloquine. Three different dosing regimens were used. Three RCTs with mefloquine and artesunate combination therapy all achieved the WHO criteria of < 5% failure rates26,37,41 when mefloquine was given as a split dose: as 25 mg/kg (15 and 10 mg/kg) in the two studies on the Thai-Burmese border37,41 and in Nigeria as 15 mg/kg (7.5 and 7.5 mg/kg).26 Single-dose mefloquine (25 mg/kg) monotherapy was given in the Sudanese27 and Thai-Burmese border33 studies, with markedly different failure rates, 2.5% and 38% respectively. In the RCT in Benin, women received two doses of mefloquine (15 mg/kg single dose) or SP for IPTp81 (Table 1). More than 90% of women did not have malaria on admission, that is they were healthy pregnant women, and high rates of adverse events occurred with mefloquine, 78% versus 32% for SP-IPTp, P < 0.001. Although mefloquine efficacy was better than SP, this high single dose still resulted in clinical malaria (26 cases per 10 000 person-months) and the prevalence of infected placentas was 1.7%.

The highest reported quinine failure rate for treatment of primary infections was reported in 50 Gabonese women with a failure rate of 40% (Table S2).42 In the same year of publication no failures were reported in 25 pregnant women treated in Sudan.43 In Sudan, women were hospitalised and every dose of quinine was supervised, compared with Gabon where only the first dose was supervised.

Two studies included first-trimester treatments of P. falciparum, both with quinine.28,44 Failure rates with 7 days of quinine in the first trimester were high in pregnant women on the Thai-Burmese border: 29% by day 28 and significantly higher when quinine was used for re-treatment, 44% by day 2844 (Table S2). Re-treatment episodes had higher parasitological failure rates than primary episodes confirmed by PCR genotyping24 and in PCR-unconfirmed infections33,44,45 (Table S2).

Gametocyte carriage

Fewer than half, 48% (11/23) of the efficacy studies30–33,36–38,44–47 and only two from Africa46,47 reported gametocyte carriage rates and these were not standardised to a set reporting methodology or time post-treatment. We can still gain an impression of gametocyte rate post-treatment by categorising the results into < 10%, 10–20% and > 20% with gametocytes per antimalarial group (Figure 3). Artemisinin combination therapies resulted in the lowest gametocyte rates post-treatment (Figure 3). While SP + AQ had < 5% parasitological failure rate in Tanzania, the combination was also associated with the highest gametocyte carriage compared with the other regimens in that trial.47 Similarly, quinine and clindamycin resulted in low failure rates on the Thai-Burmese border but the gametocyte carriage rate post-treatment in women who did not have them on admission was 13-fold higher than with a 7-day course of artesunate monotherapy.36 Gametocyte rates also increased with re-treatment.33

Figure 3.

 The proportion of treatment arms with gametocytes after treatment categorised into rates of < 10% (□), 10 to < 20% (▮) and ≥ 20% (▮) in pregnant women from 1998 to 2009. A, artemisinin; ACT, artemisinin combination therapy; CT, combination therapy; mono, monotherapy.

Placental positivity rates

Placental parasite positivity rates were mostly reported in the context of IPT RCT or IPT surveys (Table 1) and only reported in 12.5% (3/24) of treatment studies.30,34,35 Overall, 68% (23/34) of SP trial arms reported placenta-positive rates ≥ 10% and all of the chloroquine trial arms reported placenta-positive rates ≥ 15% (Table 1). Monthly IPT-SP dosing always had placenta-positive rates < 10% and performed better than standard dosing regimens (at least two doses). Daily cotrimoxazole in women with HIV resulted in one of the lowest reported rates of placenta positivity in a cross-sectional survey in an area of high malaria transmission in Uganda in (February 2008 to February 2009).48

In the treatment trials, low parasitological failure rates during follow up did not always correspond to low placental positivity rates (Table 1). In the four-arm study in Ghana35 the best efficacy result was with SP and amodiaquine (0% failure) but this did not correspond to a lower placental positivity rate (Table 1). Similarly, in Malawi, SP and azithromycin produced the best efficacy result (10%) at day 28, but the highest placenta positivity rate (27.3%) in that study34 (Table 1).

Pharmacokinetic studies

The pharmacokinetics of proguanil (and cycloguanil), atovaquone-proguanil, dihydroartemisinin, artemether-lumefantrine, quinine, azithromycin, chloroquine, sulphadoxine and pyrimethamine were reported (Table 2).103–108 Again the inclusion criteria varied as did the use of controls. Lower peak drug concentration, increased clearance, and lower area under the blood or plasma concentration–time curve (AUC) were recurring themes in the results of these studies. Nine of the twelve publications recommended higher doses or dose optimisation studies.

Table 2.   Antimalarial pharmacokinetic studies in pregnancy from 1998 to 2009
No.ReferencesYearSiteDrugReported pharmacokinetic results and dose recommendationSample sizeType of studyEfficacy
(failure %)
  1. AM-LN, artemether-lumefantrine; AP, atovaquone-proguanil; AS, artesunate; AUC, area under the concentration time curve; AZ, azithromycin; CL, clearance; Cmax, maximum concentration; CQ, chloroquine; DHA, dihyroartemisinin; HV, healthy volunteer; IPT, intermittent preventive treatment; LN, lumefantrine; OCP, oral contraceptive pill; PE, protective efficacy; PG, proguanil; PW, pregnant women; Q, quinine; SP, sulphadoxine-pyrimethamine; t½, half life; TBB, Thai-Burmese border; Tx, treatment; UM, uncomplicated malaria; VC/F, volume distribution in central compartment.

11032003TBBPGLate pregnancy and OCP reducedformation of cycloguanil (active metabolite)
Dose should be increased 50%
24
24 OCP
Tx
(HV)
Not applicable
21042003TBBAPCL, VD increased, Cmax, AUC decreased both drugs
Dose increase needed but not the same for each drug
24Tx
(UM)
Day 28, 0%
1052005AS/DHACL, VD increased, Cmax, AUC decreased compared with nonPW
Dose optimisation needed
31062005Thailand
Zambia
APCmax and AUC decreased approximately two-fold
Dose increase required
8
18
Tx
(UM)
Day 28, 0%
All placenta negative
4652006TBB
dense PK
AM-LNReduced AUC both drugs
Dose modifications needed
13Tx
(UM)
Day 42, 0%
82009TBB
population PK
LN40% day 7 concentrations more than threshold associated with failure
Increase dose and duration for example twice daily 5 days
103Tx
(UM)
Day 42, 16.5%
51072007SudanQNo significant difference
No dose adjustment needed
8
8 nonPW
Tx
(UM)
Not reported
6592008KenyaSPt½ significantly shorter, AUC ∼40% lower in PW
Need dose optimisation (higher or more frequent dosing)
17 HIV−
16 HIV+
IPT
(UM)
Not reported
7622009Mali, Zambia, Mozambique, SudanSPConcentrations PYR were higher, SDOX lower on day 7 in pregnant compared with same women postpartum
Unable to recommend how to modify dose
98IPT
(HV)
Day 28, 1%(PE)
8502009PNGAZ
+SP/CQ
Increase AZ VC/F PW no significant change in the AUC
No dose adjustment likely when given with SP or CQ
31
29 nonPW
IPT
(UM)
Day 42, 100%
day 42, 23%(PE)
9492009PNGCQ+SPSDOX, N-acetylsulfadoxine (NASDOX), PYR lower AUC
Higher dose SP recommended
30IPT
(UM)
Day 28, 38%
 1082010 (include as same study as Reference 49)Reduced AUC CQ and DECQ
Higher IPTp CQ doses may be desirable
30 nonPW  

Pharmacokinetic studies are generally small and not designed as efficacy studies. Nevertheless, efficacy is frequently available because the woman is seen repeatedly and loss to follow up is usually low. In Papua New Guinea49,108 13 women who were positive for P. falciparum on admission and received SP and chloroquine had parasite reappearance before day 28, an uncorrected failure rate of 38%. In another pharmacokinetic study from the same site, women received azithromycin and chloroquine or azithromycin and SP. There were six women with P. falciparum on admission (one with P. vivax) and none of them had reappearance before day 42. However, the protective efficacy in the 22 remaining women who were negative on admission was poor (seven became parasitaemic, five with P. falciparum, two with P. vivax), it was 23% (5/22) for P. falciparum and 32% (7/22) for any malaria by day 42.50

Future studies

A search on 25 May 2010 of http://clinicaltrials.gov/andhttp://apps.who.int/trialsearch/using‘malaria’ and ‘pregnancy’ and ‘treatment’ resulted in 58 trials of which 18 involved drugs with antimalarial activity and were in the stage of recruiting or not yet recruiting. These trials have been summarised (Table S3). Assuming all proposed trials would go ahead, 43 522 women would be included and only four of these would be treatment trials (5489 women). There would be 13 prevention trials, mostly IPT based, and one pharmacokinetic trial. Some of the proposed trials include extended follow up for parasitological failure rates, have embedded pharmacokinetic studies, and account for long-lasting insecticide-treated net use. Most of the ongoing or proposed trials on P. falciparum include SP 62% (11/18) and 35% (15/43) of trial arms include SP as monotherapy. Most of the women in ongoing and not yet recruiting studies will receive SP-IPT (Figure 4). Eighty per cent (12/15) of the SP trials are already recruiting. Two of the SP arms specifically state that they are aiming to achieve more than two doses in pregnancy. There will only be 34% (15/43) of treatment arms with an ACT.

Figure 4.

 The number of women in ongoing and planned antimalarial treatment (▮) and prevention/IPT (▮) trials by antimalarial drug (available from http://clinicaltrials.gov/ct2/ and http://www.who.int/ictrp/en/; accessed May 2010). AM-LN, artemther-lumefantrine; AQ, amodiaquine; AS, artesunate; AZ, azithromycin; CQ, chloroquine; CTX, cotrimoxazole; DP, dihyroartemisinin-piperaquine; MQ-AS, mefloquine-artesunate; MQ, mefloquine; SP, sulphadoxine-pyrimethamine.

This review has focused on P. falciparum malaria in pregnancy. There is a current dearth of treatment and prevention efficacy studies for P. vivax in pregnancy. Chloroquine efficacy when used for treatment44 and prevention51 of P. vivax in pregnancy comes only from one site. Three of the proposed trials specifically include the treatment of P. vivax in pregnancy (Table S3).

The three trials that have elected to trial mefloquine as IPT have opted for the 15 mg/kg dose (either single or split). The four treatment trials will use the fixed mefloquine-artesunate combination, which provides a mefloquine dose of approximately 8 mg/kg/day. Four studies have included daily cotrimoxazole for its antimalarial properties (Table S3) and no treatment studies have proposed to include pregnant women in their first trimester.

Discussion

The current studies on malaria in pregnancy include a wide array of methodologies, drugs, inclusion criteria, follow-up durations, care (in treatment trials) of the woman at the end of the efficacy period and come from a limited number of study sites. Antimalarial efficacy meta-analysis relies on systematic methodology and reporting. While the 2010 WHO Guidelines for treatment of malaria4 in nonpregnant women are based on the GRADE system of guideline development this was not possible for the pregnancy data because of the paucity and inconsistencies and heterogeneity in the available data. The levels of premunition and parasite resistance are highly relevant to the assessment of efficacy but these are often not reported or are difficult to measure and grade, especially because they can shift with time. The problem of prolonged carriage in pregnancy needs to be explored in future studies.

The first randomised controlled trial on artemisinins in pregnancy was in 1998 by Sowunmi and colleagues in Nigeria26 (Table S2). Despite excellent efficacy of both artemether and the combination of mefloquine and artemether, it was nearly 10 years before the artemisinins were again included in a RCT for treatment in pregnancy in Africa.34 Chloroquine was included in a P. falciparum efficacy study in Ghana during a trial conducted as recently as 2003/04 and Burkina Faso in 2003 when this drug was already considered ineffective against P. falciparum. The respective failure rates of chloroquine at day 28 were 30%35 and 47%46 (Table S2). The SP monotherapy data (Table S2) demonstrate that although the first RCT with SP in pregnancy (data from 1994–96) reported exceptionally low failure rates,52 the subsequent RCTs have reported unacceptably high failure rates (58% ≥ 10%). WHO recommends the replacement of an antimalarial if the efficacy of a treatment dose under trial conditions fails, that is 10% or higher failure rates.4 The current consensus of opinion is that doing otherwise is harmful.

The drug that most pregnant women will continue to receive in ongoing and planned studies is SP. This is of concern because the published evidence suggests that SP now always fails to achieve the > 95% cure rate recommended by WHO.4 The limited duration of follow up in some of the SP trials (< 28 days) and the fact that late reappearance occurs in pregnancy suggests that these parasitological failure rates are indeed underestimates. Placenta positivity rates with IPTp-SP were also high. Resistance to SP (and chloroquine) has both spread and intensified.53 Use of IPTp-SP relies on the efficacy of SP,53 a policy based on studies conducted from 1992 to 1999.54 Few of the ongoing and planned studies have addressed the findings of the pharmacokinetic studies suggesting dose alteration. None of the studies attempt to examine efficacy (regular weekly follow up to day 42, at least) when SP is used for re-treatment. There seems to be a disconnect between the abundant evidence of resistance and current planning and projections.

Partially effective antimalarials may exacerbate malaria in pregnancy.53,55 With increasing drug resistance, the concentration at which parasite growth is inhibited increases and the window of time during which drug concentrations exceed the minimum inhibitory concentrations shortens. The prophylactic effect post-treatment shortens progressively.54 Parasites with triple DHFR mutations have an approximately 1000-fold reduction in susceptibility to pyrimethamine, which translates into a reduction of the duration of the post-treatment prophylaxis of 1 month, compromising the recommended two-dose regimen, which can have a 3-month interval between doses. As resistance worsens, preventive efficacy is further impaired. Recent work suggests that partially effective SP-IPTp may exacerbate malaria infections in pregnancy because the most highly resistant parasites outcompete less fit parasite populations and overgrow under drug pressure.55 Use of SP-IPTp was associated with an increased fraction of parasites carrying the resistance allele at DHPS codon 581, an increase in the level of parasitaemia and more intense placental inflammation.55 Recent SP-IPTp use was associated with the lowest mean level of parasite diversity and highest mean level of parasitaemia. This counters the suggestion that SP could still be useful in pregnancy even if the failure rate in children reaches 40%,56 probably because pregnancy requires immunity to specific parasite phenotypes.

How well are current studies addressing the problem of finding a suitable replacement for SP? In Asia the loss of SP for treatment resulted in the introduction of mefloquine and later mefloquine-artesunate, and now other ACTs. The excellent efficacy of artemether and artemether-mefloquine in pregnancy in Africa was met with a 10-year gap in the trials of ACT in pregnancy in Africa. Yet the small amount of evidence available here demonstrates that ACTs have lower failure rates (when they are partnered with a drug that is not already lost to resistance) than the alternatives. A much larger and convincing data set is available from studies in nonpregnant people.4,57 In addition, ACTs reduce gametocyte carriage, which can reduce transmission. Unfortunately, as demonstrated here, the gametocyte data for antimalarial drug use in pregnant women is limited.

Seven years after the endorsement of IPTp by WHO,58 the first publication on the pharmacokinetics of SP in pregnancy demonstrated significantly lowered drug concentrations of sulphadoxine but not pyrimethamine and dose adjustments were recommended (more frequent or higher dosing).59 The complexity of this was highlighted given that this fixed combination and the synergistic interaction between the two drugs depends also ‘on the drug-resistance genotype of the parasite and corresponding degree of resistance’.60,61 Two further publications including pregnant women also reported lowered drug concentrations of sulphadoxine49,62 and one also lowered pyrimethamine concentrations.49 The pharmacokinetic properties of azithromycin in pregnancy with a half-life of 78 hours (i.e. extremely short)50 make it a poor choice as an IPTp where longer protection is needed. The performance of azithromycin as an antimalarial is weak (Table S4). A large RCT on prevention of preterm birth in Malawi administered 1 g azithromycin at 16–24 weeks and 28–32 weeks of gestation and no protective effect on maternal peripheral parasitaemia at the time of the second dose was observed.75 The recommendation has been to combine azithromycin with a longer-acting antimalarial such as chloroquine or SP because of potential synergism.50,63 This may be possible in limited circumstances, for example in Malawi where chloroquine for P. falciparum treatment showed a return of efficacy (2005) following withdrawal of the drug in 1993.64 The pharmacokinetics of artemether-lumefantrine, the most widely recommended first-line ACT for uncomplicated malaria, also appear altered in pregnant women and this affected cure rates. A dense pharmacokinetic study of artemether-lumefantrine in pregnancy demonstrated significantly lower concentrations of both artemether and lumefantrine when compared with nonpregnant people.65 Population pharmacokinetics at day 7 demonstrated concentrations of lumefantrine that were approximately 60% lower in pregnant than nonpregnant adults and children from the same area.8 In pregnant women each 100-ng/ml decrease in the day 7 lumefantrine concentration was associated with a 64% higher odds of PCR-confirmed recrudescent infection. Although pregnant women with high levels of immunity may still be cured by the standard adult dose as demonstrated recently in Uganda,66 unlike women with lowered levels of immunity in Thailand,30 exposing large numbers of women to subtherapeutic drug concentrations may well be a driving force for loss of the drug by optimising conditions for resistance to occur.67

One of the major reasons that chemoprophylaxis was abandoned was the perceived difficulties with daily or weekly compliance and yet ongoing and planned studies include chemoprophylaxis with daily cotrimoxazole. Mefloquine chemoprophylaxis (5 mg/kg/week) trials in pregnant women in Asia68,69 and Africa70,71 (not reported here because they were published before 1998) were highly effective but appear not to have been considered as an alternative for the loss of SP-IPTp. The current studies are grappling with the issue of delivering mefloquine to predominantly asymptomatic women at treatment doses (usually 25 mg/kg4 but defined as 15 mg/kg in proposed IPT trials) to maintain the IPT approach. We already know that mefloquine as prophylaxis (5 mg/kg/week) results in higher rates of adverse effects than when it is given as treatment to people with malaria and that women have more neuropsychiatric reactions than men and adults more than children.72 It is likely that 8 mg/kg mefloquine given for three consecutive days each month is a low enough dose to be tolerated and to provide the protective efficacy required for IPTp.76 In pregnanct women on the Thai-Burmese border, active weekly screening with early detection and treatment (recently described as frequent intermittent screening73) was implemented before the loss of mefloquine as monotherapy and has been adopted for routine use.15

The data of Adegnika et al.42 is likely to represent the failure rate of quinine in practice in more areas than just Gabon. Quinine’s poor adverse effects profile makes it very difficult to comply with a 7-day treatment course without supervision.74 On the Thai-Burmese border, quinine failure rates are high despite supervised treatment because of resistance. There is a dearth of efficacy results on quinine for the treatment of uncomplicated malaria in pregnancy in Africa.

This review is limited because it did not attempt to include fetal birthweight and maternal anaemia into the picture of efficacy. This was based on published evidence that indicates parasitaemia, P. falciparum or P. vivax, even in asymptomatic women is harmful. The efficacy of IPT interventions by weekly follow up and survival analysis, as well as placental infection and birth outcomes at delivery, would provide much needed information on this intervention in pregnancy.

Conclusion

There is an obvious disparity between results of parasitological efficacy and pharmacokinetics and the ongoing and proposed studies for pregnant women living in countries where malaria is endemic. Chloroquine and SP should no longer be used for treatment or prevention of malaria in pregnancy because they fail to eliminate parasites and to prevent re-infections. Repeat treatment with ineffective drugs is harmful.33,44 Supporting fixed combinations of highly active, artemisinin-based combination antimalarials that are safe for use in pregnant women would be beneficial for this distinct group of patients. Supporting drugs that should be abandoned or that do not have the pharmacokinetic properties for the task being proposed for them, is potentially unsafe. Pregnant women living in the tropics must be fully protected against malaria parasites and if infected they have the right to be treated with an efficacious regimen. Much more data are needed to recommend effective antimalarials for pregnancy so that this group is no longer forgotten.6

Disclosure of interests

The authors do not have competing interests.

Contribution to authorship

RM and FN reviewed the abstracts and manuscripts and analysed the data. RM, FN and NJW participated in developing the initial concept of the study and in manuscript revision. All authors read and approved the final manuscript.

Funding

Mahidol-Oxford Tropical medicine Research Unit is funded by the Wellcome Trust of Great Britain.

Acknowledgements

The Mahidol-Oxford Tropical medicine Research Unit is funded by the Wellcome Trust of Great Britain. The authors acknowledge Jeanne Packer, Kanchana Pongsaswat, Georges Snounou and Marcus Rijken for help in accessing manuscripts.

Ancillary