Mefloquine for preventing malaria in pregnant women.

BACKGROUND: The World Health Organization recommends intermittent preventive treatment in pregnancy (IPTp) with sulfadoxine-pyrimethamine for malaria for all women who live in moderate to high malaria transmission areas in Africa. However, parasite resistance to sulfadoxine-pyrimethamine has been increasing steadily in some areas of the region. Moreover, HIV-infected women on cotrimoxazole prophylaxis cannot receive sulfadoxine-pyrimethamine because of potential drug interactions. Thus, there is an urgent need to identify alternative drugs for prevention of malaria in pregnancy. One such candidate is mefloquine.
OBJECTIVES: To assess the effects of mefloquine for preventing malaria in pregnant women, specifically, to evaluate:• the efficacy, safety, and tolerability of mefloquine for preventing malaria in pregnant women; and• the impact of HIV status, gravidity, and use of insecticide-treated nets on the effects of mefloquine. SEARCH
METHODS: We searched the Cochrane Infectious Diseases Group Specialized Register, the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library, MEDLINE, Embase, Latin American Caribbean Health Sciences Literature (LILACS), the Malaria in Pregnancy Library, and two trial registers up to 31 January 2018. In addition, we checked references and contacted study authors to identify additional studies, unpublished data, confidential reports, and raw data from published trials. SELECTION CRITERIA: Randomized and quasi-randomized controlled trials comparing mefloquine IPT or mefloquine prophylaxis against placebo, no treatment, or an alternative drug regimen. DATA COLLECTION AND ANALYSIS: Two review authors independently screened all records identified by the search strategy, applied inclusion criteria, assessed risk of bias, and extracted data. We contacted trial authors to ask for additional information when required. Dichotomous outcomes were compared using risk ratios (RRs), count outcomes as incidence rate ratios (IRRs), and continuous outcomes using mean differences (MDs). We have presented all measures of effect with 95% confidence intervals (CIs). We assessed the certainty of evidence using the GRADE approach for the following main outcomes of analysis: maternal peripheral parasitaemia at delivery, clinical malaria episodes during pregnancy, placental malaria, maternal anaemia at delivery, low birth weight, spontaneous abortions and stillbirths, dizziness, and vomiting. MAIN
RESULTS: Six trials conducted between 1987 and 2013 from Thailand (1), Benin (3), Gabon (1), Tanzania (1), Mozambique (2), and Kenya (1) that included 8192 pregnant women met our inclusion criteria.Two trials (with 6350 HIV-uninfected pregnant women) compared two IPTp doses of mefloquine with two IPTp doses of sulfadoxine-pyrimethamine. Two other trials involving 1363 HIV-infected women compared three IPTp doses of mefloquine plus cotrimoxazole with cotrimoxazole. One trial in 140 HIV-infected women compared three doses of IPTp-mefloquine with cotrimoxazole. Finally, one trial enrolling 339 of unknown HIV status compared mefloquine prophylaxis with placebo.Study participants included women of all gravidities and of all ages (four trials) or > 18 years (two trials). Gestational age at recruitment was > 20 weeks (one trial), between 16 and 28 weeks (three trials), or ≤ 28 weeks (two trials). Two of the six trials blinded participants and personnel, and only one had low risk of detection bias for safety outcomes.When compared with sulfadoxine-pyrimethamine, IPTp-mefloquine results in a 35% reduction in maternal peripheral parasitaemia at delivery (RR 0.65, 95% CI 0.48 to 0.86; 5455 participants, 2 studies; high-certainty evidence) but may have little or no effect on placental malaria infections (RR 1.04, 95% CI 0.58 to 1.86; 4668 participants, 2 studies; low-certainty evidence). Mefloquine results in little or no difference in the incidence of clinical malaria episodes during pregnancy (incidence rate ratio (IRR) 0.83, 95% CI 0.65 to 1.05, 2 studies; high-certainty evidence). Mefloquine decreased maternal anaemia at delivery (RR 0.84, 95% CI 0.76 to 0.94; 5469 participants, 2 studies; moderate-certainty evidence). Data show little or no difference in the proportions of low birth weight infants (RR 0.95, 95% CI 0.78 to 1.17; 5641 participants, 2 studies; high-certainty evidence) and in stillbirth and spontaneous abortion rates (RR 1.20, 95% CI 0.91 to 1.58; 6219 participants, 2 studies; I2 statistic = 0%; moderate-certainty evidence). IPTp-mefloquine increased drug-related vomiting (RR 4.76, 95% CI 4.13 to 5.49; 6272 participants, 2 studies; high-certainty evidence) and dizziness (RR 4.21, 95% CI 3.36 to 5.27; participants = 6272, 2 studies; moderate-certainty evidence).When compared with cotrimoxazole, IPTp-mefloquine plus cotrimoxazole probably results in a 48% reduction in maternal peripheral parasitaemia at delivery (RR 0.52, 95% CI 0.30 to 0.93; 989 participants, 2 studies; moderate-certainty evidence) and a 72% reduction in placental malaria (RR 0.28, 95% CI 0.14 to 0.57; 977 participants, 2 studies; moderate-certainty evidence) but has little or no effect on the incidence of clinical malaria episodes during pregnancy (IRR 0.76, 95% CI 0.33 to 1.76, 1 study; high-certainty evidence) and probably no effect on maternal anaemia at delivery (RR 0.94, 95% CI 0.73 to 1.20; 1197 participants, 2 studies; moderate-certainty evidence), low birth weight rates (RR 1.20, 95% CI 0.89 to 1.60; 1220 participants, 2 studies; moderate-certainty evidence), and rates of spontaneous abortion and stillbirth (RR 1.12, 95% CI 0.42 to 2.98; 1347 participants, 2 studies; very low-certainty evidence). Mefloquine was associated with higher risks of drug-related vomiting (RR 7.95, 95% CI 4.79 to 13.18; 1055 participants, one study; high-certainty evidence) and dizziness (RR 3.94, 95% CI 2.85 to 5.46; 1055 participants, 1 study; high-certainty evidence). AUTHORS'
CONCLUSIONS: Mefloquine was more efficacious than sulfadoxine-pyrimethamine in HIV-uninfected women or daily cotrimoxazole prophylaxis in HIV-infected pregnant women for prevention of malaria infection and was associated with lower risk of maternal anaemia, no adverse effects on pregnancy outcomes (such as stillbirths and abortions), and no effects on low birth weight and prematurity. However, the high proportion of mefloquine-related adverse events constitutes an important barrier to its effectiveness for malaria preventive treatment in pregnant women.

Background

Description of the condition

Malaria is the most important parasitic disease worldwide and is endemic in parts of Africa, Asia, and South America. Pregnant women are at higher risk of malaria infection than non‐pregnant women in the same age group, and are at higher risk of severe illness (Brabin 1983; Desai 2007). Malaria infection during pregnancy, particularly the first or second pregnancy, is also associated with adverse outcomes for both mother (severe anaemia) and infant (low birth weight, neonatal mortality; Ataíde 2014; Guyatt 2004; Menendez 2010; Radeva‐Petrova 2014; Schwarz 2008; Steketee 2001). Symptoms most commonly reported by semi‐immune pregnant women with clinical malaria include headache, arthromyalgias, and fever (Bardaji 2008). In areas of low transmission, pregnant women with malaria parasitaemia frequently present with symptoms and signs such as fever, malaise, headache, and vomiting. The infection may develop into severe complications such as cerebral malaria and pulmonary oedema if untreated, and may be a cause of maternal mortality (Bardaji 2008).

To reduce the burden and consequences of malaria in pregnancy, the World Health Organization (WHO) currently recommends that pregnant women who live in moderate to high malaria transmission areas in Africa sleep under an insecticide‐treated net (ITN), as described in Gamble 2006, and receive intermittent preventive treatment (IPT) with sulfadoxine‐pyrimethamine at each scheduled antenatal care visit (provided that doses are at least one month apart) (WHO 2013). IPT is a form of malaria chemoprevention that was tested and adopted as policy in response to both malaria parasites developing resistance to weekly prophylaxis with chloroquine and low compliance with the weekly regimen (WHO 2004). The long elimination half‐life of sulfadoxine‐pyrimethamine allows intermittent dosing while still providing prophylactic cover for the intervening weeks (White 2005). IPT is therefore defined as "administration of a curative treatment dose of an effective antimalarial drug at predefined intervals during pregnancy" regardless of the presence or absence of current infection (White 2005).

Sulfadoxine‐pyrimethamine remains the drug used for IPT in pregnancy, even though resistance has spread in many parts of southern and eastern Africa (ter Kuile 2007; WHO 2012a), which is spurring researchers and policy makers to seek safe and effective alternatives to sulfadoxine‐pyrimethamine (Desai 2018).

Description of the intervention

Mefloquine is a 4‐methanolquinoline that is related to quinine. It was originally developed by the US military for preventing malaria in soldiers and has been widely used for preventing malaria in travellers (Schlagenhauf 2010). Like sulfadoxine‐pyrimethamine, mefloquine has a long elimination half‐life of two to four weeks; in travellers, weekly dosing consists of 250 mg (FDA 2004), and in pregnant women monthly dosing at treatment doses is feasible (Briand 2009).

Mefloquine was first investigated in the 1990s as prophylactic treatment for pregnant women. An observational study raised concerns that mefloquine may be associated with increased risk of stillbirth (Nosten 1999); however other trials did not confirm this finding (Pekyi 2016; Steketee 1996). A systematic review considered the safety of mefloquine in pregnancy and concluded that no evidence indicates that mefloquine use in pregnancy carries increased risk for the foetus (Gonzalez 2014). The drug is known to be associated with a range of mild dose‐related transient side effects, such as vomiting, nausea, and dizziness (Bardaji 2012; Lee 2017; Sevene 2010; ter Kuile 1995). Researchers have described severe neuropsychiatric side effects that occur in about one in 10,000 travellers taking mefloquine as chemoprophylaxis (Phillips‐Howard 1995; Steffen 1993). Studies conducted in Beninese pregnant women found that dizziness and vomiting are the most frequent adverse effects related to use of mefloquine as IPT in pregnancy (Briand 2009; Denoeud‐Ndam 2012).

Data show resistance to mefloquine in multi‐drug resistance areas of Thailand (Carrara 2009; Nosten 2000), but it remains rare in Africa (Aubouy 2007; MacArthur 2001; Oduola 1987).

How the intervention might work

Malaria chemoprevention is thought to work through clearance or suppression of asymptomatic malaria infection in the peripheral blood of the mother and the placenta (White 2005). This reduction in malaria parasitaemia may, however, be insufficient to justify recommendations for widespread prophylactic prescriptions that do not provide subsequent benefit for clinically important outcomes for mother and baby. These outcomes may include a reduction in episodes of maternal malaria, reduced risk of anaemia, and improved birth weight, as well as more substantive outcomes such as a reduction in severe maternal illness or lower rates of spontaneous pregnancy loss and maternal, neonatal, and infant mortality (see Figure 1).

1

Indicators and impact of malaria infection in mothers and infants.

Effects of malaria chemoprevention may depend on the local malaria epidemiology and thus the level of acquired immunity against malaria in pregnant women. In stable transmission areas, women of reproductive age may be partially immune to malaria, presenting parasitaemia without clinical disease; however, asymptomatic infections may have detrimental effects, such as anaemia and low birth weight. In contrast, in unstable malaria transmission areas, naturally acquired malaria immunity is usually low among adults and malaria infection may be associated with clinical episodes and severe illness.

Primigravidae women are at higher risk of adverse effects of malaria infection than multigravidae women. This is thought to result from women developing antibodies specific to placental‐type parasites when exposed to Plasmodium falciparum during their first pregnancy. These antibodies are then present in subsequent pregnancies (Ataíde 2014). This is seen in multigravidae women as a more specific and efficient immune response and clearing the infection at an earlier stage than in primigravidae women (Walker 2013).

Another potential effect modifier of the susceptibility to malaria infection is HIV status (Menéndez 2011). In many malaria‐endemic areas, data show that the prevalence of HIV infection, which has been observed to increase the risk of malaria infection, is high among pregnant women (Gonzalez 2012; van Eijk 2003). Compared with HIV‐uninfected women, HIV‐infected women are more likely to carry malaria parasites in their blood, to have higher parasite densities, and to develop placental parasitaemia, anaemia, and malaria symptoms (Ayisi 2003; van Eijk 2002; van Eijk 2003). This increased risk of malaria is the same in multigravidae (women in their third pregnancy or higher) and in women in their first or second pregnancy (ter Kuile 2004; van Eijk 2003). Placental malaria infection may also increase the risk of perinatal mother‐to‐child transmission of HIV (Ayisi 2003).

Use of ITNs during pregnancy has been shown to have a beneficial impact on pregnancy outcomes (reduced prevalence of low birth weight, miscarriage, and placental parasitaemia) in malaria‐endemic Africa (Gamble 2007), and this approach could modify the effect of IPT (Menéndez 2008).

Why it is important to do this review

The WHO recommends IPT with sulfadoxine‐pyrimethamine for all pregnant women who live in moderate to high malaria transmission areas in Africa (WHO 2004; WHO 2013). However, studies have shown that resistance to sulfadoxine‐pyrimethamine in some regions of Eastern Africa has been increasing steadily during the past two decades (Iriemenam 2012; Mockenhaupt 2008). Thus, there is an urgent need for more effective antimalarials to prevent malaria during pregnancy.

This review aims to evaluate the efficacy and safety of mefloquine for preventing malaria in pregnant women. These findings could serve as the basis for future guidelines on preventive agents for malaria in pregnant women.

Objectives

To assess the effects of mefloquine for preventing malaria in pregnant women ‐ specifically, to evaluate:

the efficacy, safety, and tolerability of mefloquine for preventing malaria in pregnant women; and

the impact of HIV status, gravidity, and use of insect‐treated nets (ITNs) on the effects of mefloquine.

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials (RCTs) and quasi‐RCTs.

Types of participants

Pregnant women of any gravidity regardless of HIV status, living in malaria‐endemic areas (CDC 2017).

Types of interventions

Interventions

Mefloquine given to pregnant women as intermittent preventive treatment or as chemoprophylaxis.

Controls

Placebo, no intervention, or an alternative drug regimen.

Types of outcome measures

Maternal

Maternal peripheral parasitaemia during pregnancy

Maternal peripheral parasitaemia at delivery

Placental malaria¹

Mean haemoglobin and maternal anaemia (moderate and severe) at delivery

Clinical malaria episodes during pregnancy

Foetal/infant

Cord blood parasitaemia

Cord blood haemoglobin and anaemia (as defined in the original studies)

Mean birth weight

Low birth weight prevalence (< 2500 g)

Prematurity prevalence (< 37 weeks of gestation)

Morbidity in first year of life

Adverse events

Serious adverse events (SAEs)²

Illnesses that were life threatening or required hospitalization during pregnancy (SAEs in pregnancy)

Adverse pregnancy outcomes: spontaneous abortion, stillbirth, congenital malformation

Maternal mortality

Perinatal, neonatal, infant mortality

Mother‐to‐child transmission of HIV frequency (at six weeks of age)

Non‐serious adverse events

Frequency and severity of reported all‐cause and drug‐related adverse events

¹Placental malaria diagnosed by histology, microscopy, or any method used in the included study. Figure 2 shows the relations between different outcomes.

2

Conceptual framework of malaria chemoprevention. Reproduced under the terms of a Creative Commons Licence from Radeva‐Petrova 2014.

²Review authors acknowledge the limitation of analyzing rare serious adverse events because randomized controlled trials usually are not powered enough to detect them.

Search methods for identification of studies

We attempted to identify all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).

Electronic searches

We searched the following databases using the search terms and strategy described in Appendix 1: the Cochrane Infectious Diseases Group Specialized Register (up to 31 January 2018); the Cochrane Central Register of Controlled Trials (CENTRAL), published in the Cochrane Library (January 2018); MEDLINE (PubMed; from 1966 to 31 January 2018); Embase (OVID; 1974 to 31 January 2018); and Latin American Caribbean Health Sciences Literature (LILACS) (BIREME; 1982 to 31 January 2018). We also searched the Malaria in Pregnancy (MiP) Library (www.mip‐consortium.org/resources/index.htm), the WHO International Clinical Trial Registry Platform (ICTRP; www.who.int/ictrp/search/en), ClinicalTrials.gov, and the International Standard Randomized Controlled Trial Number (ISRCTN) registry (www.isrctn.com/), using ‘mefloquine', ‘malaria', and ‘pregnan*' as search terms.

Searching other resources

We contacted researchers working in the field to ask for unpublished data, confidential reports, and raw data from published trials. We also checked the citations of all trials identified by the methods described.

Data collection and analysis

Selection of studies

Two review authors independently screened all trials identified by the search strategy by title or abstract, or both (Appendix 1). We coded studies as ‘retrieve' or ‘do not retrieve'. We retrieved the full‐text copies of trials deemed potentially relevant. Two review authors then independently assessed study eligibility using a form based on the review inclusion criteria. We resolved disagreements through discussion or by consultation with a third review author. Any review author who participated in trials that potentially met the review inclusion criteria did not participate in the procedure to select studies for inclusion. We listed all studies excluded after full‐text assessment and reasons for their exclusion in a ‘Characteristics of excluded studies' table. We illustrated the study selection process in a PRISMA diagram.

Data extraction and management

Three review authors (RG, CPD, and MP) used a data extraction form to independently extract data on trial characteristics, including trial site, year, local malaria transmission estimates, antimalarial resistance pattern of mefloquine and the comparator drug (when possible), trial methods, participants, interventions, doses, and outcomes.

We extracted the number of participants randomized and the number of participants analyzed in experimental and control groups for each outcome. For dichotomous outcomes, we extracted the number of participants experiencing the event and the number assessed in each treatment group. For continuous outcomes, we extracted the arithmetic means, standard deviations for each treatment group (when provided), and the number of participants assessed in each group. We also extracted medians and ranges when provided. For outcomes reported as incidences, we extracted the number of participants experiencing the event (cases) and the person‐years at risk.

Any review author who participated in any of the trials included in the review did not participate in data extraction nor ‘Risk of bias' assessment of their own articles.

Assessment of risk of bias in included studies

Two review authors independently assessed the risk of bias for each included trial using the Cochrane ‘Risk of bias' assessment tool (Higgins 2011). This approach assesses the risk of bias across seven domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other potential sources of bias (Higgins 2011). For each domain, we assigned a judgment of low, high, or unclear risk of bias. We judged the risk of bias for blinding on the presence of blinding and whether lack of blinding could potentially influence the results.

Measures of treatment effect

We presented dichotomous outcomes using risk ratios (RRs), count outcomes as incidence rate ratios (IRRs), and continuous outcomes as mean differences (MDs). We presented all measures of effect with 95% confidence intervals (CIs).

Unit of analysis issues

When conducting a meta‐analysis, we ensured that participants and cases in the placebo group were not counted more than once.

Dealing with missing data

We aimed to conduct the analysis according to the intention‐to‐treat principle. However, when there was loss to follow‐up, we used a complete‐case analysis such that participants for whom no outcome was reported were excluded from the analysis. This analysis assumes that participants for whom an outcome is available are representative of the original randomized patients. We aimed to conduct a sensitivity analysis to evaluate the robustness of this method, but this was not possible, as described below. If data from trial reports were insufficient, unclear, or missing, we contacted the study authors for additional information.

Assessment of heterogeneity

We calculated the I2 statistic using values of 30% to 59%, 60% to 89%, and 90% to 100% to denote moderate, substantial, and considerable levels of heterogeneity, respectively.

Assessment of reporting biases

We aimed to assess the risk of publication bias by constructing funnel plots and looking for asymmetry, but the small number of trials included in each comparison of the meta‐analysis made this assessment impossible.

Data synthesis

We performed data analysis using Review Manager 5 (RevMan 5) (RevMan 2014). We intended to perform subgroup analysis by gravidity and HIV status when possible. HIV status subgroup analysis was not possible in any case owing to different study designs for different HIV status populations. In the absence of heterogeneity, we used a fixed‐effect model for the meta‐analysis; when we detected moderate or considerable heterogeneity, we used a random‐effects model. Additionally, we assessed the certainty of evidence using the GRADE approach (GRADEpro GDT 2015) for the following main outcomes of analysis: maternal peripheral parasitaemia at delivery, clinical malaria episodes during pregnancy, placental malaria, maternal anaemia at delivery, low birth weight, spontaneous abortion and stillbirth, dizziness, and vomiting.

Subgroup analysis and investigation of heterogeneity

We aimed to investigate heterogeneity by conducting prespecified subgroup analysis to evaluate the contributions of differences in trial characteristics such as risk of bias, geographical region, malaria transmission pattern, antimalarial resistance, drug regimen, use of ITNs, gravidity (primigravidae versus multigravidae), HIV status (uninfected, infected, unknown), and trial methods. Only the gravidity subgroup analysis was possible for one outcome of the main comparison. The other subgroup analyses were not possible because of the small number of trials included in each comparison.

Sensitivity analysis

We planned to conduct a sensitivity analysis to restore the integrity of the randomization process and to test the robustness of our results; however, the small number of trials included in each comparison – two at most – made this impossible. Additionally, missing outcome data were balanced in numbers across intervention groups, and reasons for missing data were similar across groups.

Results

Description of studies

Results of the search

The literature search, conducted up to 31 January 2018, identified 254 references, of which two were duplicate trial reports. Of the 252 remaining articles, we excluded 231 articles and one ongoing trial after title/abstract screening. We assessed 20 full‐text articles for eligibility, of which we excluded 14 articles. Six trials (in six publications) met the inclusion criteria of the review (Figure 3).

3

Study flow diagram.

Included studies

Six chemoprevention trials that included 8192 pregnant women met our inclusion criteria (see the Characteristics of included studies section). These trials were conducted between 1987 and 2013 in Thailand (one trial), Benin (three trials), Gabon (one trial), Kenya (one trial), Mozambique (two trials), and Tanzania (two trials).

The included trials recruited women of all gravidities of all ages (four trials) or over 18 years of age (two trials). Gestational age at recruitment was greater than 20 weeks (one trial), between 16 and 28 weeks (three trials), or ≤ 28 weeks (two trials).

Two trials evaluated mefloquine against sulfadoxine‐pyrimethamine as IPTp in HIV‐uninfected pregnant women. Three trials evaluated mefloquine IPTp alone (or in combination with daily cotrimoxazole) against cotrimoxazole in HIV‐infected pregnant women. Finally, one trial in Thailand compared weekly mefloquine prophylaxis against placebo in women of unknown HIV status. All included trials reported that drug administration was supervised.

All included trials recruited women in all gravidity groups; five reported aggregate results and one disaggregated by gravidity for the primary outcome. In five trials, all women in both intervention and control groups received a long‐lasting ITN at recruitment and iron, and investigators routinely administered folic acid.

Excluded studies

We excluded one trial for the reasons given in the ‘Characteristics of excluded studies' table.

Risk of bias in included studies

See Figure 4 and Figure 5 for a summary of the ‘Risk of bias' assessments. We have presented further details in the ‘Characteristics of included studies' table.

4

‘Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

5

‘Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

Random sequence generation (selection bias)

Two trials adequately described methods of sequence generation (Gonzalez 2014a BEN GAB MOZ TAN; Gonzalez 2014b KEN MOZ TAN), three described a non‐random component in the sequence generation process (Briand 2009 BEN; Denoeud‐Ndam 2014a BEN; Denoeud‐Ndam 2014b BEN), and in the remaining trial, the risk was unclear (Nosten 1994 THA).

Allocation concealment (selection bias)

Four trials described adequate methods of allocation concealment (Denoeud‐Ndam 2014a BEN; Denoeud‐Ndam 2014b BEN; Gonzalez 2014a BEN GAB MOZ TAN; Gonzalez 2014b KEN MOZ TAN), one trial reported no concealment of allocation (Briand 2009 BEN), and in the remaining trial, the risk was unclear (Nosten 1994 THA).

Blinding

Blinding of participants and personnel (performance bias)

Four trials were open (Briand 2009 BEN; Denoeud‐Ndam 2014a BEN; Denoeud‐Ndam 2014b BEN; Gonzalez 2014a BEN GAB MOZ TAN), and we assessed these as having high risk of performance risk. Two trials were double‐blind and placebo‐controlled (Gonzalez 2014b KEN MOZ TAN; Nosten 1994 THA), and we assessed these as having low risk of performance bias.

Blinding of efficacy outcome assessment (detection bias)

For five trials, we judged the efficacy outcome as not influenced by blinding or lack of blinding. In the remaining trial, the risk of detection bias for efficacy outcomes was unclear (Nosten 1994 THA).

Blinding of safety outcome assessment (detection bias)

For the four open trials, we judged the risk of detection bias as high for assessment of safety outcomes (Briand 2009 BEN; Denoeud‐Ndam 2014a BEN; Denoeud‐Ndam 2014b BEN; Gonzalez 2014a BEN GAB MOZ TAN). In one trial, the risk of detection bias was unclear (Nosten 1994 THA). For the remaining trial, which was double‐blinded, we judged the risk of detection bias as low (Gonzalez 2014b KEN MOZ TAN).

Incomplete outcome data

In all included trials, missing outcome data were balanced in numbers across groups, and we judged the risk of attrition bias to be low.

Selective reporting

We considered the risk of reporting bias as low in five trials and unclear in one (Nosten 1994 THA).

Other potential sources of bias

All included trials appeared to be free of other sources of bias, and we judged this risk as low.

Effects of interventions

See: Table 1; Table 2

Summary of findings for the main comparison

Mefloquine compared with sulfadoxine‐pyrimethamine for preventing malaria in pregnant women

Mefloquine compared with sulfadoxine‐pyrimethamine for preventing malaria in pregnant women
Patient or population: HIV‐uninfected pregnant women Setting: Benin, Gabon, Mozambique, and Tanzania Intervention: mefloquine Comparison: sulfadoxine‐pyrimethamine
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (trials)	Certainty of the evidence (GRADE)	Comments (compared with sulfadoxine‐pyrimethamine)
	Risk with sulfadoxine‐pyrimethamine	Risk with mefloquine
Clinical malaria episodes during pregnancy	‐	‐	IRR 0.83 (0.65 to 1.05)	‐ (2 RCTs)	⊕⊕⊕⊕ HIGHa	Mefloquine results in little or no difference in the incidence of clinical malaria episodes during pregnancy
Maternal peripheral parasitaemia at delivery	43 per 1000	28 per 1000(20 to 37)	RR 0.65(0.48 to 0.86)	5455(2 RCTs)	⊕⊕⊕⊕ HIGHa	Mefloquine results in lower maternal peripheral parasitaemia at delivery
Placental malaria	52 per 1000	54 per 1000 (30 to 97)	RR 1.04 (0.58 to 1.86)	4668 (2 RCTs)	⊕⊕⊝⊝ LOWa,b,cDue to imprecision and heterogeneity	Mefloquine may result in little or no difference in placental parasitaemia
Maternal anaemia at delivery	219 per 1000	184 per 1000 (166 to 206)	RR 0.84 (0.76 to 0.94)	5469 (2 RCTs)	⊕⊕⊕⊝ MODERATEa,dDue to imprecision	Mefloquine probably results in fewer women anaemic at delivery
Low birth weight	117 per 1000	111 per 1000 (91 to 137)	RR 0.95 (0.78 to 1.17)	5641 (2 RCTs)	⊕⊕⊕⊕ HIGHa,e	Mefloquine results in little or no difference in low birth weight
Stillbirths and abortions	31 per 1000	37 per 1000 (28 to 49)	RR 1.20 (0.91 to 1.58)	6219 (2 RCTs)	⊕⊕⊕⊝ MODERATEa,bDue to imprecision	Mefloquine probably results in little or no difference in stillbirths or abortions
AEs: vomiting	82 per 1000	390 per 1000 (338 to 449)	RR 4.76 (4.13 to 5.49)	6272 (2 RCTs)	⊕⊕⊕⊕ HIGHa	Mefloquine results in a four‐fold increase in vomiting
AEs: dizziness	94 per 1000	396 per 1000 (316 to 496)	RR 4.21 (3.36 to 5.27)	6272 (2 RCTs)	⊕⊕⊕⊝ MODERATEa,fDue to risk of bias	Mefloquine probably results in a four‐fold increase in dizziness
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). Abbreviations: CI: confidence interval; IRR: incidence rate ratio; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence. High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.

aNot downgraded for risk of bias: although one trial has serious risk of bias, the other is of good quality and exclusion of the smaller trial has little effect on the estimate of effect.
 bDowngraded by 1 for imprecision: confidence intervals range from considerable benefit to considerable harm.
 cDowngraded by 1 for heterogeneity: substantive qualitative heterogeneity is evident in the meta‐analysis.
 dDowngraded by 1 for imprecision: CIs include little or no important difference to a 24% reduction in anaemic women.
 eNot downgraded for imprecision: we consider that a 22% reduction or 17% increase in birth weight is not a clinically significant change.
 fDowngraded by 1 for performance bias: both trials are unblinded.

Summary of findings 2

Mefloquine plus cotrimoxazole compared with cotrimoxazole for preventing malaria in pregnant women

Mefloquine plus cotrimoxazole compared with cotrimoxazole for preventing malaria in pregnant women
Patient or population: HIV‐infected pregnant women Setting: Benin, Kenya, Mozambique, and Tanzania Intervention: mefloquine plus cotrimoxazole Comparison: cotrimoxazole
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (trials)	Certainty of the evidence (GRADE)	Comments (compared with cotrimoxazole)
	Risk with cotrimoxazole	Risk with mefloquine plus cotrimoxazole
Clinical malaria episodes during pregnancy	‐	‐	IRR 0.76 (0.33 to 1.76)	‐(1 RCT)	⊕⊕⊕⊕ HIGH	Mefloquine results in little or no difference in the incidence of clinical malaria episodes during pregnancy
Maternal peripheral parasitaemia at delivery (PCR)	66 per 1000	34 per 1000(20 to 62)	RR 0.52(0.30 to 0.93)	989(2 RCTs)	⊕⊕⊕⊝ MODERATEaDue to risk of bias	Mefloquine probably results in lower maternal peripheral parasitaemia at delivery
Placental malaria (PCR)	68 per 1000	19 per 1000 (10 to 39)	RR 0.28 (0.14 to 0.57)	977 (2 RCTs)	⊕⊕⊕⊕ MODERATEa	Mefloquine plus cotrimoxazole results in fewer women with placental malaria at delivery
Maternal anaemia at delivery	178 per 1000	168 per 1000 (130 to 214)	RR 0.94 (0.73 to 1.20)	1197 (2 RCTs)	⊕⊕⊕⊝ MODERATEaDue to risk of bias	Mefloquine plus cotrimoxazole probably results in little or no difference in maternal anaemia cases at delivery
Low birth weight	118 per 1000	141 per 1000 (105 to 188)	RR 1.20 (0.89 to 1.60)	1220 (2 RCTs)	⊕⊕⊕⊝ MODERATEaDue to risk of bias	Mefloquine plus cotrimoxazole probably results in little or no difference in the proportion of births that are low birth weight
Spontaneous abortions and stillbirths	50 per 1000	56 per 1000 (21 to 149)	RR 1.12 (0.42 to 2.98)	1347 (2 RCTs)	⊕⊝⊝⊝ VERY LOWa,b,c	We do not know if mefloquine plus cotrimoxazole results in a difference in spontaneous abortions and stillbirths
AEs: vomiting	30 per 1000	239 per 1000(144 to 396)	RR 7.95(4.79 to 13.18)	1055(1 RCT)d	⊕⊕⊕⊕ HIGH	Mefloquine plus cotrimoxazole results in an eight‐fold increase in vomiting
AEs: dizziness	75 per 1000	296 per 1000(214 to 411)	RR 3.94(2.85 to 5.46)	1055(1 RCT)e	⊕⊕⊕⊕ HIGH	Mefloquine plus cotrimoxazole results in a four‐fold increase in dizziness
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). Abbreviations: CI: confidence interval; IRR: incidence rate ratio; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence. High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.

aDowngraded by 1 for risk of bias: one of the included trials is at serious risk of bias.
 bDowngraded by 1 for inconsistency: trials showed substantial heterogeneity.
 cDowngraded by 1 for imprecision: confidence intervals range from considerable benefit to considerable harm.
 dA second RCT, Denoeud‐Ndam 2014a BEN, reported 50 events in the mefloquine+cotrimoxazole group and 0 in the control group (cotrimoxazole), with RR 101 (95% CI 6.29 to 1621.68). This trial was open and participants knew to which group they were allocated. Meta‐analysis causes a paradoxically very wide CI. Because of this distortion, we have used the results from Gonzalez 2014b KEN MOZ TAN in the grade table.
 eA second RCT, Denoeud‐Ndam 2014a BEN, reported 52 events in the mefloquine+cotrimoxazole group and 0 in the control group (cotrimoxazole), with RR 105 (95% CI 6.54 to 1685.03). This trial was open and participants knew to which group they were allocated. Meta‐analysis causes a paradoxically very wide CI with the lower 95% CI. Because of this distortion, we have used the results from Gonzalez 2014b KEN MOZ TAN in this ‘Summary of findings' table.

Comparison 1: Mefloquine versus sulfadoxine‐pyrimethamine (HIV‐uninfected pregnant women)

See Table 1.

Maternal outcomes

We included in this comparison two trials that evaluated two doses of IPTp (Briand 2009 BEN; Gonzalez 2014a BEN GAB MOZ TAN). Data show a decrease in the number of clinical malaria episodes during pregnancy among mefloquine recipients, but this does not clearly constitute an effect of mefloquine because the 95% CIs do not exclude the possibility of no different effects (IRR 0.83, 95% CI 0.65 to 1.05; 2 studies; high‐certainty evidence; Analysis 1.1). Overall, IPTp‐mefloquine was associated with a 35% reduction in the risk of maternal peripheral parasitaemia at delivery (RR 0.65, 95% CI 0.48 to 0.86; 5455 participants, 2 studies; I2 statistic = 16%; high‐certainty evidence; Analysis 1.2), but the absolute difference between treatments was small. We found no significant evidence of an effect of mefloquine or sulfadoxine‐pyrimethamine on placental malaria infections (RR 1.04, 95% CI 0.58 to 1.86; 4668 participants, 2 studies; I2 statistic = 63%; low‐certainty evidence; Analysis 1.3). The mefloquine group showed a slight increase in the mean haemoglobin level at delivery (MD 0.10, 95% CI 0.01 to 0.19; 5588 participants, 2 studies; I2 statistic = 0%; Analysis 1.4) and a decrease in maternal anaemia cases at delivery (RR 0.84, 95% CI 0.76 to 0.94; 5469 participants, 2 studies; I2 statistic = 0%; moderate‐certainty evidence; Analysis 1.5), but the data show no significant differences in severe maternal anaemia at delivery between groups (RR 0.93, 95% CI 0.58 to 1.48; 5469 participants, 2 studies; I2 statistic = 41%; Analysis 1.6). The original definitions of maternal moderate anaemia and severe maternal anaemia were different in the two trials included in the analysis (Gonzalez 2014a BEN GAB MOZ TAN defined anaemia as haemoglobin < 11 g/dL and severe anaemia as haemoglobin < 7 g/dL), but we homogenized data for the analysis as < 9.5 g/dL and < 8 g/dL (as defined in Briand 2009 BEN), respectively.

1.1

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 1 Clinical malaria episodes during pregnancy.

1.2

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 2 Maternal peripheral parasitaemia at delivery.

1.3

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 3 Placental malaria.

1.4

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 4 Mean haemoglobin at delivery.

1.5

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 5 Maternal anaemia at delivery.

1.6

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 6 Severe maternal anaemia at delivery.

Foetal/infant outcomes

No effect was evident for the outcomes of cord blood parasitaemia (RR 0.44, 95% CI 0.13 to 1.46; 5309 participants, 2 studies; I2 statistic = 33%; Analysis 1.7) and cord blood anaemia (RR 1.04, 95% CI 0.87 to 1.23; 4006 participants, 1 study; Analysis 1.8).

1.7

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 7 Cord blood parasitaemia.

1.8

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 8 Cord blood anaemia.

Regarding newborn outcomes, mean birth weight did not show significant differences between groups (MD 2.52, 95% CI ‐25.66 to 30.69; 5241 participants, 2 studies; I2 statistic = 0%; Analysis 1.9). Low birth weight (RR 0.95, 95% CI 0.78 to 1.17; 5641 participants, 2 studies; I2 statistic = 33%; high‐certainty evidence; Analysis 1.10) and prematurity prevalence (RR 1.03, 95% CI 0.76 to 1.40; 4640 participants, 2 studies; I2 statistic = 0%; Analysis 1.12) also showed no differences between groups. Subgroup analysis of low birth weight by gravidity yielded results that did not vary (primigravidae: RR 1.02, 95% CI 0.80 to 1.30; 1576 participants, 2 studies; I2 statistic = 3%; Analysis 1.11; multigravidae: RR 0.94, 95% CI 0.78 to 1.14; 4065 participants, 2 studies; I2 statistic = 0%; Analysis 1.11).

1.9

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 9 Mean birth weight.

1.10

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 10 Low birth weight.

1.12

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 12 Prematurity.

1.11

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 11 Low birth weight by gravidity.

Only one trial reported data on infant morbidity, and results followed the same trend; the IRR was near 1, and the CIs did not discard the possibility of no difference between mefloquine and sulfadoxine‐pyrimethamine. Chosen proxies for infant morbidity were malaria in the first year of life (IRR 0.97, 95% CI 0.82 to 1.15; 1 study; Analysis 1.13) and hospital admissions in the first year of life (IRR 0.93, 95% CI 0.75 to 1.17; 1 study; Analysis 1.14).

1.13

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 13 Malaria in first year of life.

1.14

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 14 Hospital admissions in first year of life.

Safety outcomes

No difference was evident between mefloquine and sulfadoxine‐pyrimethamine in overall serious adverse events reporting (RR 0.98, 95% CI 0.81 to 1.20; 4674 participants, 1 study; Analysis 1.15). Definitions of stillbirth and abortion were different for the two trials included in this comparison; therefore we aggregated both outcomes into a single outcome (RR 1.20, 95% CI 0.91 to 1.58; 6219 participants, 2 studies; I2 statistic = 0%; moderate‐certainty evidence; Analysis 1.16). Congenital malformation cases were also similar in both intervention groups (RR 1.10, 95% CI 0.51 to 2.37; 5931 participants, 2 studies; I2 statistic = 33%; Analysis 1.17).

1.15

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 15 SAEs during pregnancy.

1.16

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 16 Stillbirths and abortions.

1.17

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 17 Congenital malformations.

Regarding maternal mortality, one of the trials reported maternal deaths only in the mefloquine group, and the other trial showed a similar proportion of maternal deaths in both IPTp groups; the CI of the meta‐analysis was wide, and heterogeneity was moderate (RR 2.41, 95% CI 0.27 to 21.23; 6219 participants, 2 studies; I2 statistic = 54%; Analysis 1.18). Only one of the trials reported neonatal and infant mortality (Gonzalez 2014a BEN GAB MOZ TAN), but we obtained neonatal mortality rates for the other trial by contacting the study authors (Briand 2009 BEN). Neither of the two outcomes showed a significant effect of mefloquine or sulfadoxine‐pyrimethamine (neonatal deaths: RR 0.98, 95% CI 0.67 to 1.43; 6134 participants, 2 studies; I2 statistic = 0%; Analysis 1.19; incidence of infant deaths: IRR 1.00, 95% CI 0.66 to 1.52; 1 study; Analysis 1.20).

1.18

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 18 Maternal mortality.

1.19

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 19 Neonatal mortality.

1.20

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 20 Infant mortality.

Overall, IPTp‐mefloquine increased the risk of adverse events; results of individual trials and of meta‐analyses were significant for vomiting (RR 4.76, 95% CI 4.13 to 5.49; 6272 participants, 2 studies; I2 statistic = 0%; high‐certainty evidence; Analysis 1.21), fatigue/weakness (RR 4.62, 95% CI 1.80 to 11.85; 6272 participants, 2 studies; I2 statistic = 91%; high‐certainty evidence; Analysis 1.22), and dizziness (RR 4.21, 95% CI 3.36 to 5.27; 6272 participants, 2 studies; I2 statistic = 66%; moderate‐certainty evidence; Analysis 1.23), with the exception of headache (RR 0.70, 95% CI 0.25 to 1.94; 6272 participants, 2 studies; I2 statistic = 85%; Analysis 1.24).

1.21

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 21 AEs: vomiting.

1.22

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 22 AEs: fatigue/weakness.

1.23

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 23 AEs: dizziness.

1.24

Analysis

Comparison 1 Mefloquine versus sulfadoxine‐pyrimethamine, Outcome 24 AEs: headache.

Comparison 2: Mefloquine plus cotrimoxazole versus cotrimoxazole (HIV‐infected pregnant women)

See Table 2.

This comparison included two trials evaluating three IPTp doses of mefloquine (Denoeud‐Ndam 2014a BEN; Gonzalez 2014b KEN MOZ TAN). Only one of the trials reported clinical malaria episodes during pregnancy, noting no significant differences in malaria episodes between groups (IRR 0.76, 95% CI 0.33 to 1.76; 1 study; high‐certainty evidence; Analysis 2.1). IPTp‐mefloquine plus cotrimoxazole prophylaxis was associated with a 48% reduction in the risk of maternal peripheral parasitaemia at delivery measured by polymerase chain reaction (PCR) (RR 0.52, 95% CI 0.30 to 0.93; 989 participants, 2 studies; I2 statistic = 0%; moderate‐certainty evidence; Analysis 2.2), a 49% reduction in the risk of placental malaria measured by blood smear (RR 0.51, 95% CI 0.29 to 0.89; 1144 participants, 2 studies; I2 statistic = 0%; Analysis 2.3), and a 72% reduction in the risk of placental malaria measured by PCR (RR 0.28, 95% CI 0.14 to 0.57; 977 participants, 2 studies; I2 statistic = 0%; moderate‐certainty evidence; Analysis 2.4). The other maternal‐related outcomes at delivery included in this comparison did not show evidence that they were effects of mefloquine owing to the wideness of the CIs (mean haemoglobin: MD 0.07, 95% CI ‐0.32 to 0.46; 1167 participants, 2 studies; I2 statistic = 62%; Analysis 2.5; maternal anaemia: RR 0.94, 95% CI 0.73 to 1.20; 1197 participants, 2 studies; I2 statistic = 12%; moderate‐certainty evidence; Analysis 2.6; severe maternal anaemia: RR 0.93, 95% CI 0.41 to 2.08; 1167 participants, 2 studies; I2 statistic = 0%; Analysis 2.7). The original definitions of maternal anaemia were different in the two trials included in the analysis (Gonzalez 2014b KEN MOZ TAN defined anaemia as haemoglobin < 11 g/dL), but we homogenized definitions for the analysis as < 9.5 g/dL (as defined in Denoeud‐Ndam 2014a BEN). The two trials defined severe maternal anaemia as haemoglobin < 7 g/dL.

2.1

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 1 Clinical malaria episodes during pregnancy.

2.2

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 2 Maternal peripheral parasitaemia at delivery (PCR).

2.3

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 3 Placental malaria (blood smear).

2.4

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 4 Placental malaria (PCR).

2.5

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 5 Mean haemoglobin at delivery.

2.6

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 6 Maternal anaemia at delivery (< 9.5 g/dL).

2.7

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 7 Maternal severe anaemia at delivery.

Meta‐analyses of foetal and neonatal outcomes were underpowered to detect significant effects of mefloquine on cord blood parasitaemia (RR 0.33, 95% CI 0.03 to 3.13; 1166 participants, 2 studies; I2 statistic = 0%; Analysis 2.8), mean birth weight (MD ‐25.75, 95% CI ‐86.99 to 35.49; 1220 participants, 2 studies; I2 statistic = 0%; Analysis 2.9), low birth weight rates (RR 1.20, 95% CI 0.89 to 1.60; 1220 participants, 2 studies; I2 statistic = 0%; moderate‐certainty evidence; Analysis 2.10), and prematurity rates (RR 1.07, 95% CI 0.58 to 1.72; 824 participants, 2 studies; I2 statistic = 32%; Analysis 2.11). These CIs did not exclude the possibility of no different effects between groups.

2.8

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 8 Cord blood parasitaemia.

2.9

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 9 Mean birth weight.

2.10

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 10 Low birth weight.

2.11

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 11 Prematurity.

Overall, serious adverse events during pregnancy were significantly less frequent in the group of IPTp‐mefloquine plus cotrimoxazole prophylaxis than in the cotrimoxazole alone group (RR 0.69, 95% CI 0.50 to 0.95; 1347 participants, 2 studies; I2 statistic = 0%; Analysis 2.12). However, analysis of individual adverse events did not show differences between groups, for example, spontaneous abortions and stillbirths (RR 1.12, 95% CI 0.42 to 2.98; 1347 participants, 2 studies; I2 statistic = 69%; very low‐certainty evidence; Analysis 2.13) and congenital malformations (RR 0.61, 95% CI 0.22 to 1.67; 1312 participants, 2 studies; I2 statistic = 0%; Analysis 2.14). Definitions of spontaneous abortion and stillbirth were different in the two included trials (that is, difference in the gestational age cutoff for classifying miscarriage or stillbirth); therefore, we combined both indicators and analyzed them as one. Only one trial included information on maternal deaths (Gonzalez 2014b KEN MOZ TAN), and we obtained this information by contacting the authors in the other trial (Denoeud‐Ndam 2014a BEN). Analyses of maternal deaths revealed no significant differences between groups (RR 0.51, 95% CI 0.13 to 2.01; 1347 participants, 2 studies; I2 statistic = 0%; Analysis 2.15). Also, we found that neonatal mortality rates were not significantly different among groups, as revealed by the CI (RR 1.32, 95% CI 0.65 to 2.69; 1239 participants, 2 studies; I2 statistic = 0%; Analysis 2.16). It is important to note that mefloquine plus cotrimoxazole recipients were at 1.92 times greater risk of mother‐to‐child transmission of HIV than the group that took only cotrimoxazole (RR 1.92, 95% CI 1.13 to 3.25; 1019 participants, 2 studies; I2 statistic = 0%; Analysis 2.17).

2.12

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 12 SAEs during pregnancy.

2.13

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 13 Spontaneous abortions and stillbirths.

2.14

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 14 Congenital malformations.

2.15

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 15 Maternal mortality.

2.16

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 16 Neonatal mortality.

2.17

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 17 Mother‐to‐child transmission HIV.

Vomiting, fatigue/weakness, and dizziness displayed substantial and considerable levels of heterogeneity in the meta‐analysis. Individual trials showed significant increases in three drug‐related adverse events in the groups given IPTp‐mefloquine plus cotrimoxazole prophylaxis, but random‐effects analyses show a significant effect of IPTp‐mefloquine only in the case of vomiting (RR 20.88, 95% CI 1.40 to 311.66; 1347 participants, 2 studies; I2 statistic = 74%; Analysis 2.18), while fatigue (RR 2.95, 95% CI 0.26 to 32.93; 1347 participants, 2 studies; I2 statistic = 91%; Analysis 2.19) and dizziness (RR 16.34, 95% CI 0.39 to 684.99; 1347 participants, 2 studies; I2 statistic = 86%; Analysis 2.20) show no significant evidence. In the three cases, CIs are considerably wide. Headache cases were not significantly different across groups (RR 0.76, 95% CI 0.28 to 2.10; 1347 participants, 2 studies; I2 statistic = 30%; Analysis 2.21).

2.18

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 18 AEs: vomiting.

2.19

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 19 AEs: fatigue/weakness.

2.20

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 20 AEs: dizziness.

2.21

Analysis

Comparison 2 Mefloquine plus cotrimoxazole versus cotrimoxazole, Outcome 21 AEs: headache.

Comparison 3: Mefloquine versus cotrimoxazole (HIV‐infected pregnant women)

Only one trial conducted in Benin provided data on this comparison of three IPTp‐mefloquine doses versus cotrimoxazole prophylaxis (Denoeud‐Ndam 2014b BEN). The few observations reported in the trial made the analyses, in general, underpowered to detect differences between groups. Efficacy outcomes directly related to malaria yielded RR indicating beneficial effects of IPTp‐mefloquine in reducing infection, but CIs did not exclude the possibility of no difference between groups (maternal peripheral parasitaemia during pregnancy measured by PCR: RR 0.21, 95% CI 0.03 to 1.72; 98 participants, 1 study; Analysis 3.1; placental malaria measured by PCR: RR 0.73, 95% CI 0.13 to 4.15; 94 participants, 1 study; Analysis 3.2; placental malaria measured by blood smear: RR 0.35, 95% CI 0.01 to 8.30; 108 participants, 1 study; Analysis 3.3). Data show no differences across groups for mean haemoglobin (MD ‐0.10, 95% CI ‐0.67 to 0.47; 100 participants, 1 study; Analysis 3.4) or maternal anaemia at delivery (RR 0.90, 95% CI 0.26 to 3.16; 100 participants, 1 study; Analysis 3.5).

3.1

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 1 Maternal peripheral parasitaemia at delivery (PCR).

3.2

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 2 Placental malaria (PCR).

3.3

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 3 Placental malaria (blood smear).

3.4

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 4 Mean haemoglobin at delivery.

3.5

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 5 Maternal anaemia at delivery (< 9.5 g/dL).

All newborn outcomes included in the trial displayed wide CIs, providing no evidence of differences between groups (mean birth weight: MD ‐102.00, 95% CI ‐255.52 to 51.52; 120 participants, 1 study; Analysis 3.6; low birth weight rate: RR 1.52, 95% CI 0.56 to 4.13; 120 participants, 1 study; Analysis 3.7; prematurity rate: RR 1.08, 95% CI 0.33 to 3.56; 125 participants, 1 study; Analysis 3.8).

3.6

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 6 Mean birth weight.

3.7

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 7 Low birth weight.

3.8

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 8 Prematurity.

Serious adverse events reported in the trial were balanced across groups and were infrequent. The CIs reveal the possibility of no different effects between interventions in overall serious adverse events (RR 1.06, 95% CI 0.28 to 4.07; 140 participants, 1 study; Analysis 3.9), stillbirths (RR 4.30, 95% CI 0.49 to 37.49; 139 participants, 1 study; Analysis 3.10), spontaneous abortions (RR 1.07, 95% CI 0.07 to 16.84; 139 participants, 1 study; Analysis 3.11), and congenital malformations (RR 1.07, 95% CI 0.16 to 7.41; 139 participants, 1 study; Analysis 3.12). No maternal deaths occurred during the trial (139 participants, 1 study; Analysis 3.13), and only one neonate in each intervention group died (RR 1.05, 95% CI 0.07 to 16.39; 129 participants, 1 study; Analysis 3.14). The trial did not record infant mortality and regarded infant deaths after seven days of birth until six weeks of age as a proxy; small numbers of observations and infant deaths made demonstration of differences between groups impossible (RR 2.10, 95% CI 0.19 to 22.54; 129 participants, 1 study; Analysis 3.15).

3.9

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 9 SAEs during pregnancy.

3.10

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 10 Stillbirths.

3.11

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 11 Spontaneous abortions.

3.12

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 12 Congenital malformations.

3.13

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 13 Maternal mortality.

3.14

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 14 Neonatal mortality.

3.15

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 15 Infant deaths after 7 days.

Drug‐related adverse events were significantly more frequent in the mefloquine group. Despite wide CIs, results show an effect of mefloquine in increasing the frequency of vomiting (RR 13.43, 95% CI 3.31 to 54.54; 139 participants, 1 study; Analysis 3.16), fatigue/weakness (RR 6.99, 95% CI 1.64 to 29.81; 139 participants, 1 study; Analysis 3.17), and dizziness (RR 52.60, 95% CI 3.26 to 848.24; 139 participants, 1 study; Analysis 3.18). Data show no differences between groups in drug‐related headache (RR 0.21, 95% CI 0.01 to 4.39; 139 participants, 1 study; Analysis 3.19).

3.16

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 16 AEs: vomiting.

3.17

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 17 AEs: fatigue/weakness.

3.18

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 18 AEs: dizziness.

3.19

Analysis

Comparison 3 Mefloquine versus cotrimoxazole, Outcome 19 AEs: headache.

Comparison 4: Mefloquine versus placebo (pregnant women of unknown HIV status)

Maternal and foetal/infant outcomes

Only one trial provided data on this comparison, which comprised two phases of mefloquine prophylaxis with different doses of the drug (Nosten 1994 THA); the results belong to the pooled samples of both trial phases. This trial did not report clinical malaria episodes during pregnancy, maternal anaemia at delivery, cord blood parasitaemia and anaemia, serious adverse events, neonatal mortality, and adverse events, or data reporting was incomplete.

The only observed significant effect that could be attributed to mefloquine was the decrease in maternal peripheral parasitaemia at delivery (RR 0.13, 95% CI 0.05 to 0.33; 339 participants, 1 study; Analysis 4.1). The other efficacy outcomes evaluated in this trial ‐ both maternal and newborn‐related outcomes ‐ showed wide CIs and did not demonstrate different effects between placebo and mefloquine prophylaxis (placental malaria: RR 0.14, 95% CI 0.01 to 2.68; 220 participants, 1 study; Analysis 4.2; mean birth weight: MD ‐80.00, 95% CI ‐184.65 to 24.65; 290 participants, 1 study; Analysis 4.3; low birth weight: RR 1.39, 95% CI 0.78 to 2.48; 290 participants, 1 study; Analysis 4.4; prematurity: RR 0.48, 95% CI 0.15 to 1.53; 199 participants, 1 study; Analysis 4.5).

4.1

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 1 Maternal peripheral parasitaemia during pregnancy.

4.2

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 2 Placental malaria.

4.3

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 3 Mean birth weight.

4.4

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 4 Low birth weight.

4.5

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 5 Prematurity.

This trial reported only serious adverse events, and adverse events data were not complete in the published article. Stillbirths were more prevalent in the group given mefloquine prophylaxis, but the small number of observed events made the analysis unpowered to detect differences between groups (RR 2.63, 95% CI 0.86 to 8.08; 311 participants, 1 study; Analysis 4.6). Investigators reported only three spontaneous abortions and five congenital malformations, thus the CIs of analyses were very wide to detect differences in effects (spontaneous abortion: RR 0.48, 95% CI 0.04 to 5.22; 311 participants, 1 study; Analysis 4.7; congenital malformation: RR 3.82, 95% CI 0.43 to 33.83; 311 participants, 1 study; Analysis 4.8). During the trial, only one maternal death occurred in the mefloquine group, but the power of the analysis was too low to attribute the effects to an intervention (RR 2.95, 95% CI 0.12 to 71.85; 339 participants, 1 study; Analysis 4.9). Infant deaths were equally frequent in both trial groups (RR 1.04, 95% CI 0.63 to 1.74; 288 participants, 1 study; Analysis 4.10).

4.6

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 6 Stillbirths.

4.7

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 7 Spontaneous abortions.

4.8

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 8 Congenital malformations.

4.9

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 9 Maternal mortality.

4.10

Analysis

Comparison 4 Mefloquine versus placebo, Outcome 10 Infant mortality.

Discussion

Summary of main results

We included in this Cochrane Review six trials, enrolling 8192 pregnant women.

For HIV‐uninfected women, two doses of intermittent preventive mefloquine treatment in pregnancy (IPTp‐mefloquine) reduced the risk of maternal peripheral parasitaemia at delivery by 35% (high‐certainty evidence) and the risk of anaemia by 16% (moderate‐certainty evidence) compared with two doses of intermittent preventive sulfadoxine‐pyrimethamine treatment in pregnancy (IPTp‐sulfadoxine‐pyrimethamine). Investigators have reported no significant evidence of an effect of mefloquine on placental malaria, cord blood parasitaemia and anaemia, mean birth weight, prevalence of low birth weight, prematurity, stillbirths and abortions, and congenital malformations. Overall, IPTp‐mefloquine increases by approximately four‐fold the risk of drug‐related adverse events including vomiting, fatigue/weakness, and dizziness (high‐ or moderate‐certainty evidence), when compared with sulfadoxine‐pyrimethamine.

For HIV‐infected women, three doses of IPTp‐mefloquine plus cotrimoxazole prophylaxis compared with cotrimoxazole alone reduced the risk of maternal peripheral parasitaemia at delivery (measured by polymerase chain reaction (PCR)) by 48% (moderate‐certainty evidence) and the risk of placental malaria (measured by PCR) by 72% (high‐certainty evidence). Meta‐analyses were underpowered to detect differences between effects of mefloquine plus cotrimoxazole and cotrimoxazole on other maternal, foetal, and neonatal outcomes. Regarding drug‐related adverse events, random‐effects analyses showed a significant effect of IPTp‐mefloquine plus cotrimoxazole prophylaxis compared with cotrimoxazole alone only in the case of vomiting (RR 7.95, 95% CI 4.79 to 13.18; 1055 participants; high‐certainty evidence). It is important to note that mefloquine plus cotrimoxazole recipients were at 1.92 times greater risk of mother‐to‐child transmission of HIV than the group that received cotrimoxazole alone (RR 1.92, 95% CI 1.13 to 3.25; 1019 participants). A secondary analysis of one of the included trials revealed this finding (Gonzalez 2014b KEN MOZ TAN).

One trial among HIV‐infected women comparing three doses of IPTp‐mefloquine and cotrimoxazole was underpowered to detect an effect of mefloquine on maternal, foetal, infant, and safety outcomes, except for drug‐related adverse events, which were more frequent in the mefloquine group.

Finally, the single trial conducted in Thailand (where Plasmodium vivax coexists) found a significant effect attributable to mefloquine weekly prophylaxis (compared with placebo) only in reducing the risk of maternal peripheral parasitaemia at delivery (RR 0.13, 95% CI 0.05 to 0.33; 339 participants).

Overall completeness and applicability of evidence

Trials were carried out in sub‐Saharan Africa, except for one conducted in Thailand, and were published between 1994 and 2014. Findings evidenced that mefloquine chemoprevention reduces the risk of maternal parasitaemia at delivery in both HIV‐uninfected and HIV‐infected women compared with other antimalarials or placebo. Additionally, in HIV‐infected women, Mefloquine was found to reduce the risk of placental malaria. Results from these trials show fairly consistent clinically important benefits for women and their infants. However, the risk of drug‐related adverse events was increased among mefloquine recipients, and it is notable that mefloquine increased the risk of mother‐to‐child transmission in one trial.

Included trials evaluated two or three IPTp doses of sulfadoxine‐pyrimethamine as per World Health Organization (WHO) recommendations, whereas current evidence suggests that monthly doses may provide a better prophylactic effect (Kayentao 2013). Additionally, the WHO currently recommends IPTp administration at each scheduled antenatal contact (WHO 2012b).

The findings of this review, derived from a variety of sub‐Saharan African settings and comparing mefloquine chemoprevention in pregnancy with varied antimalarial drugs and placebo, may be applied worldwide. Mefloquine is currently recommended as malaria chemoprevention for pregnant women of all gestational ages travelling to malaria‐endemic areas (CDC 2016). This drug is also recommended for treatment of uncomplicated malaria episodes in combination with artesunate (WHO 2015), and a fixed‐dose formulation is available in some malaria‐endemic countries. In 2013, the WHO Evidence Review Group (ERG) on IPTp met to assess evidence obtained from IPTp‐mefloquine trials, and the WHO Malaria Policy Advisory Committee (MPAC) reviewed ERG recommendations and agreed that mefloquine at the 15‐mg/kg dose regimen should not be recommended for IPTp, given its adverse events and poor tolerability (WHO MPAC 2013).

Certainty of the evidence

We assessed the certainty of evidence in this review by using the GRADE approach and presented the evidence in two ‘Summary of findings' tables for efficacy and safety outcomes (Table 1; Table 2).

For HIV‐uninfected pregnant women, evidence that IPTp‐mefloquine was superior to IPTp‐sulfadoxine‐pyrimethamine in reducing the risk of maternal peripheral parasitaemia and anaemia at delivery was of moderate certainty, and evidence that IPTp‐mefloquine increased drug‐related adverse effects (namely, vomiting and dizziness) compared with IPTp‐sulfadoxine‐pyrimethamine was of high and moderate certainty (respectively). We considered the effects of IPTp‐mefloquine in decreasing placental malaria risk compared with IPTp‐sulfadoxine‐pyrimethamine to be of low certainty because of substantial heterogeneity among trials. Finally, we considered evidence of no effects of mefloquine compared with sulfadoxine‐pyrimethamine on low birth weight and stillbirths and abortions to be of moderate certainty.

For HIV‐infected women, evidence that cotrimoxazole plus IPTp‐mefloquine was superior to cotrimoxazole in reducing the risk of maternal peripheral parasitaemia and anaemia at delivery was of moderate certainty, whereas evidence regarding lack of effect on risk of placental malaria was of high certainty. Evidence of no effects of cotrimoxazole plus IPTp‐mefloquine compared with cotrimoxazole on low birth weight and stillbirths and abortions was of moderate and very low certainty, respectively, because of serious risk of bias of one of the included trials and substantial heterogeneity. Finally, we considered evidence of mefloquine increasing risks of vomiting and dizziness to be of low certainty because heterogeneity among trials was substantial and the 95% CI was wide.

Potential biases in the review process

It seems unlikely that we have missed any trials examining mefloquine for prevention of malaria in pregnant women.

Agreements and disagreements with other studies or reviews

A previous Cochrane Review on drugs for preventing malaria in pregnant women in endemic areas analyzed the effects of mefloquine for prevention of malaria (Radeva‐Petrova 2014). Our results are consistent with those previously reported but include more trials and thus may be more robust.

The findings of this Cochrane Review are also consistent with those of a previous systematic review assessing the safety and tolerability of mefloquine in pregnancy (González 2013).

Authors' conclusions

In past decades, many clinical trials have tested mefloquine chemoprevention to prevent malaria and its consequences in pregnant women.

For HIV‐uninfected pregnant women, IPTp‐mefloquine better reduces malaria effects compared with IPTp‐sulfadoxine‐pyrimethamine, but the drug is worse tolerated than sulfadoxine‐pyrimethamine. For HIV‐infected pregnant women, IPTp‐mefloquine added to cotrimoxazole prophylaxis reduces the risk of important malaria consequences better than cotrimoxazole alone, but drug tolerability constitutes a health issue.

The data show that mefloquine is an efficacious and safe antimalarial drug in terms of pregnancy outcomes for prevention of malaria in pregnancy. However, the high proportion of mefloquine‐related adverse events constitutes an important barrier to its effectiveness for malaria preventive treatment in pregnant women.

Mefloquine efficacy to prevent malaria effects in pregnancy is well established. Future research should concentrate on finding a dose that would provide the same antimalarial beneficial effects while reducing its drug‐related adverse events, especially as weekly prophylaxis (for example, at a dose of 5 mg/kg) for HIV‐uninfected women living in areas of high sulfadoxine‐pyrimethamine resistance. Researchers also should further examine findings on the two‐fold increased risk of mother‐to‐child transmission of HIV among mefloquine recipients.

What's new

Akinyotu 2015 NIG

Trial name or title	A comparative study of mefloquine and SP as prophylaxis against malaria in pregnant HIV‐infected patients
Methods	Allocation: randomizedIntervention model: parallel assignmentMasking: single‐blind (outcomes assessor)Primary purpose: prevention
Participants	Inclusion criteria: Pregnant HIV‐infected patients Gestational age ≥ 16 weeks No history of use of mefloquine or sulphadoxine Pyrimethamine 4 weeks before recruitment Exclusion criteria: Anaemia packed cell volume < 30% Pre‐existing medical conditions ‐ diabetes mellitus, hypertension Allergy to sulphadoxine‐pyrimethamine or mefloquine Non‐consenting patients Multiple gestation Known psychiatric illness Known seizure disorder History of severe renal or hepatic disease
Interventions	Mefloquine: 250 mg 3 doses 4 weeks apart Sulfadoxine‐pyrimethamine: 500 mg sulphadoxine and 25 mg pyrimethamine, 3 tablets 4 weeks apart for 3 doses
Outcomes	No information available
Starting date	September 2015
Contact information	Oriyomi O Akinyotu, MBBS; Ibadan: +2348035044590; oriyomiddoc@yahoo.com
Notes	We contacted the study authors, but they could not provide the data to us because the study was part of a thesis not yet defended.

Comparison 1

Mefloquine versus sulfadoxine‐pyrimethamine

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Clinical malaria episodes during pregnancy	2		Rate Ratio (Fixed, 95% CI)	0.83 [0.65, 1.05]
2 Maternal peripheral parasitaemia at delivery	2	5455	Risk Ratio (M‐H, Fixed, 95% CI)	0.65 [0.48, 0.86]
3 Placental malaria	2	4668	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.58, 1.86]
4 Mean haemoglobin at delivery	2	5588	Mean Difference (IV, Fixed, 95% CI)	0.10 [0.01, 0.19]
5 Maternal anaemia at delivery	2	5469	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.76, 0.94]
6 Severe maternal anaemia at delivery	2	5469	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.58, 1.48]
7 Cord blood parasitaemia	2	5309	Risk Ratio (M‐H, Random, 95% CI)	0.44 [0.13, 1.46]
8 Cord blood anaemia	1	4006	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.87, 1.23]
9 Mean birth weight	2	5241	Mean Difference (IV, Fixed, 95% CI)	2.52 [‐25.66, 30.69]
10 Low birth weight	2	5641	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.78, 1.17]
11 Low birth weight by gravidity	2	5641	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.84, 1.13]
11.1 Primigravidae	2	1576	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.80, 1.30]
11.2 Multigravidae	2	4065	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.78, 1.14]
12 Prematurity	2	4640	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.76, 1.40]
13 Malaria in first year of life	1		Rate Ratio (Fixed, 95% CI)	0.97 [0.82, 1.15]
14 Hospital admissions in first year of life	1		Rate Ratio (Fixed, 95% CI)	0.93 [0.75, 1.17]
15 SAEs during pregnancy	1	4674	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.81, 1.20]
16 Stillbirths and abortions	2	6219	Risk Ratio (M‐H, Fixed, 95% CI)	1.20 [0.91, 1.58]
17 Congenital malformations	2	5931	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.51, 2.37]
18 Maternal mortality	2	6219	Risk Ratio (M‐H, Random, 95% CI)	2.41 [0.27, 21.23]
19 Neonatal mortality	2	6134	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.67, 1.43]
20 Infant mortality	1		Rate Ratio (Fixed, 95% CI)	1.00 [0.66, 1.52]
21 AEs: vomiting	2	6272	Risk Ratio (M‐H, Fixed, 95% CI)	4.76 [4.13, 5.49]
22 AEs: fatigue/weakness	2	6272	Risk Ratio (M‐H, Random, 95% CI)	4.62 [1.80, 11.85]
23 AEs: dizziness	2	6272	Risk Ratio (M‐H, Random, 95% CI)	4.21 [3.36, 5.27]
24 AEs: headache	2	6272	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.25, 1.94]

Comparison 2

Mefloquine plus cotrimoxazole versus cotrimoxazole

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Clinical malaria episodes during pregnancy	1		Rate Ratio (Fixed, 95% CI)	0.76 [0.33, 1.76]
2 Maternal peripheral parasitaemia at delivery (PCR)	2	989	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.30, 0.93]
3 Placental malaria (blood smear)	2	1144	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.29, 0.89]
4 Placental malaria (PCR)	2	977	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.14, 0.57]
5 Mean haemoglobin at delivery	2	1167	Mean Difference (IV, Random, 95% CI)	0.07 [‐0.32, 0.46]
6 Maternal anaemia at delivery (< 9.5 g/dL)	2	1197	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.73, 1.20]
7 Maternal severe anaemia at delivery	2	1167	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.41, 2.08]
8 Cord blood parasitaemia	2	1166	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.03, 3.13]
9 Mean birth weight	2	1220	Mean Difference (IV, Random, 95% CI)	‐25.75 [‐86.99, 35.49]
10 Low birth weight	2	1220	Risk Ratio (M‐H, Fixed, 95% CI)	1.20 [0.89, 1.60]
11 Prematurity	2	824	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.58, 1.96]
12 SAEs during pregnancy	2	1347	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.50, 0.95]
13 Spontaneous abortions and stillbirths	2	1347	Risk Ratio (M‐H, Random, 95% CI)	1.12 [0.42, 2.98]
14 Congenital malformations	2	1312	Risk Ratio (M‐H, Fixed, 95% CI)	0.61 [0.22, 1.67]
15 Maternal mortality	2	1347	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.13, 2.01]
16 Neonatal mortality	2	1239	Risk Ratio (M‐H, Fixed, 95% CI)	1.32 [0.65, 2.69]
17 Mother‐to‐child transmission HIV	2	1019	Risk Ratio (M‐H, Fixed, 95% CI)	1.92 [1.13, 3.25]
18 AEs: vomiting	2	1347	Risk Ratio (M‐H, Random, 95% CI)	20.88 [1.40, 311.66]
19 AEs: fatigue/weakness	2	1347	Risk Ratio (M‐H, Random, 95% CI)	2.95 [0.26, 32.93]
20 AEs: dizziness	2	1347	Risk Ratio (M‐H, Random, 95% CI)	16.34 [0.39, 684.99]
21 AEs: headache	2	1347	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.28, 2.10]

Comparison 3

Mefloquine versus cotrimoxazole

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Maternal peripheral parasitaemia at delivery (PCR)	1	98	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.03, 1.72]
2 Placental malaria (PCR)	1	94	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.13, 4.15]
3 Placental malaria (blood smear)	1	108	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.30]
4 Mean haemoglobin at delivery	1	100	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.67, 0.47]
5 Maternal anaemia at delivery (< 9.5 g/dL)	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.26, 3.16]
6 Mean birth weight	1	120	Mean Difference (IV, Fixed, 95% CI)	‐102.0 [‐255.52, 51.52]
7 Low birth weight	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	1.52 [0.56, 4.13]
8 Prematurity	1	125	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.33, 3.56]
9 SAEs during pregnancy	1	140	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.28, 4.07]
10 Stillbirths	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	4.30 [0.49, 37.49]
11 Spontaneous abortions	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.07, 16.84]
12 Congenital malformations	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.16, 7.41]
13 Maternal mortality	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14 Neonatal mortality	1	129	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.07, 16.39]
15 Infant deaths after 7 days	1	129	Risk Ratio (M‐H, Fixed, 95% CI)	2.10 [0.19, 22.54]
16 AEs: vomiting	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	13.43 [3.31, 54.54]
17 AEs: fatigue/weakness	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	6.99 [1.64, 29.81]
18 AEs: dizziness	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	52.60 [3.26, 848.24]
19 AEs: headache	1	139	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.01, 4.39]

Comparison 4

Mefloquine versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Maternal peripheral parasitaemia during pregnancy	1	339	Risk Ratio (M‐H, Fixed, 95% CI)	0.13 [0.05, 0.33]
2 Placental malaria	1	220	Risk Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 2.68]
3 Mean birth weight	1	290	Mean Difference (IV, Fixed, 95% CI)	‐80.0 [‐184.65, 24.65]
4 Low birth weight	1	290	Risk Ratio (M‐H, Fixed, 95% CI)	1.39 [0.78, 2.48]
5 Prematurity	1	199	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.15, 1.53]
6 Stillbirths	1	311	Risk Ratio (M‐H, Fixed, 95% CI)	2.63 [0.86, 8.08]
7 Spontaneous abortions	1	311	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.04, 5.22]
8 Congenital malformations	1	311	Risk Ratio (M‐H, Fixed, 95% CI)	3.82 [0.43, 33.83]
9 Maternal mortality	1	339	Risk Ratio (M‐H, Fixed, 95% CI)	2.95 [0.12, 71.85]
10 Infant mortality	1	288	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.63, 1.74]