Adverse pregnancy outcomes between the anti-malarial drugs: Is there a difference between the drugs recommended by World Health Organization? Results of a mixed treatment comparison analysis of randomized clinical trials and cohort studies

Abstract

BACKGROUND:

Data regarding the relative safety profile of anti-malarial drugs in pregnancy is sparse mainly limited by the absence of head-to-head clinical trials. The present study is a network meta-analysis of safety of anti-malarial drugs used to treat malaria in pregnant women.

METHODS:

A thorough literature search using the search strategy “Malaria [tiab] AND (Pregnant [tiab] OR Pregnancy [tiab])” was carried out for either randomized controlled trials or prospective cohort studies in pregnant malarial women prescribed any of the recommended anti-malarial drugs by World Health Organization (WHO) and that have reported adverse pregnancy outcomes such as miscarriage, still birth, and neonatal deaths. Odds ratio with 95% confidence interval was used as the effect estimate. Random-effects model and Markov Chain Monte Carlo simulation method was used to generate pooled estimates. Sensitivity analysis was performed excluding data from first trimester and GRADE approach was used to categorize the quality of evidence.

RESULTS:

A total of 1242 papers were obtained with the search strategy, of which seven evaluating 10 treatment arms in a total of 5510 participants were included in the present meta-analysis. The pooled estimates revealed significantly lower risks of abortion with quinine and artemisinin-lumefantrine compared to dihydroartemisinin-piperaquine, artesunate with mefloquine and artesunate with amodiaquine. But when a cohort study that was conducted in the first trimester of pregnancy was excluded, no significant differences were observed in the risk of abortion between the anti-malarial drugs. No significant differences in the risk of either stillbirths or neonatal deaths were observed with any of the drugs. The quality of evidence was found to be very low due to serious limitations in both the precision and indirectness.

CONCLUSION:

WHO recommended anti-malarials in pregnancy have similar risk profiles with regard to abortion, stillbirth and neonatal deaths.

Keywords

Pregnant malaria artemisinin ACT chloroquine quinine

1 Introduction

Malaria in pregnancy is an important public health concern associated with major morbidity and mortality [1]. Prevalence of malaria in pregnancy within malarious areas has been estimated to be as high as 47% [2]. Pregnant women with malaria are associated with additional health complications, such as anemia, which pose a higher degree of healthcare needs within areas that often have more limited healthcare options [3]. A review by Newman et al. reports that two-thirds of the fetuses exposed to malaria during their mother’s pregnancy had intrauterine growth retardation associated with 8% of infant mortality [3].

Several anti-malarial drugs are recommended by World Health Organization (WHO) to be used in pregnancy for treating malaria. These include quinine-clindamycin and artesunate (in severe malaria or when oral quinine is unavailable), used in first trimester and artemisinin combination treatments in the second and third trimesters [4, 5]. Additionally, WHO recommends use of sulphadoxine-pyrimethamine intermittently as a prophylactic measure in Africa with stable transmission of malaria [6]. Head-to-head comparison studies assessing the safety of anti-malarial drugs in pregnant women are limited due to ethical concerns. Over the past decade, the field of meta-analysis and comparative effectiveness has grown substantially. Specifically, the development of network meta-analysis (NMA) provides a method to simultaneously analyse all treatment options within a disease area [7, 8]. Moreover, these methods allow for comparisons to be made despite the absence of head-to-head evidence. As these methods combine direct and indirect evidence, they may provide more precise estimates, which lends itself well to a safety analysis, given that randomized controlled trials tend to be powered for efficacy and underpowered for safety. Hence, we sought the present systematic literature review and NMA to compare the adverse outcomes of anti-malarial drugs when used to treat malaria in pregnant women.

2 Methods

A thorough literature search was completed on 25 November 2016 with the following search strategy: Malaria [tiab] AND (Pregnant [tiab] OR Pregnancy [tiab]). Databases searched for the potential articles were: Medline (thorough PubMed), Cochrane CENTRAL and Google Scholar. No limits were placed with respect to either language or year of publication.

We included studies that have recruited pregnant women diagnosed as having malaria and was administered one of the anti-malarial drugs for treatment. Studies reporting at least one of the following adverse pregnancy outcomes were considered in the present review: spontaneous abortion, still birth and neonatal deaths. Studies that have administered anti-malarial drugs as prophylactic drugs or the administered anti-malarial drugs were other than WHO recommended were excluded. Only randomized clinical trials and cohort studies were included in the present review. Eligible interventions are those recommended by WHO that are as follows: Artemether-lumefantrine, artesunate with amodiaquine, aretesunate with mefloquine, dihydroartemisinin-piperaquine, artesunate with sulphadoxine-pyrimethamine, choroquine and quinine. Treatments were differentiated according to the specific drugs, doses and frequencies of administration. Two reviewers (KS and GS) independently scanned all abstracts and potentially eligible articles. Where discrepancies in judgments between reviewers occurred, agreement was reached through discussion. Using a standard data extraction sheet, information was extracted on trial characteristics, patient characteristics and outcomes. Risk of bias of randomized controlled trials was assessed using Cochrane tool of risk of bias [9] while New Castle Ottawa scale [10] was used for cohort studies. The protocol for this review was registered with PROSPERO (CRD42016051554).

2.1 Statistical analyses

We reported the elements as per the preferred reporting items for systematic reviews and meta-analysis extension for network meta-analysis (PRISMA-NMA) guidelines [11]. We conducted the network meta-analysis using Bayesian hierarchical models. Proportions of adverse pregnancy outcomes reported with the anti-malarial drugs were obtained from each study. We used a logistic regression model with the logit link function and a binomial likelihood. We chose to present results as odds ratios (OR) for these models so as to avoid the ceiling effect that limits relative risks for outcomes with proportions around 0.8 to 0.95. Both random-effects and fixed-effects model were used and the Markov Chain Monte Carlo method was used to run 1,000 iterations. Model selection was conducted using the deviance information criterion (DIC) according to NICE conventions. Inconsistency of the results between direct and indirect comparisons was evaluated by plotting the posterior mean deviance of individual data points for consistency and inconsistency model [12]. Step plot was used to compare various treatment arms and rankograms were generated based on surface under the cumulative ranking curve (SUCRA) [13]. Under Bayesian approach, SUCRA estimates the probability of a treatment being the best amongst many possibilities. Sensitivity analysis was carried out by excluding data from first trimester as well as studies that enrolled both complicated and uncomplicated malarial women individually for each of the outcome measures. NetMetaXL [14] and WinBUGS statistical analysis program version 1.4.3 (MRC Biostatistics Unit, Institute of Public Health, Cambridge, UK) were used for generating results of network meta-analysis. Grades of Recommendation, Assessment, Development and Evaluation (GRADE) working group approach was used to assess the quality of evidence [9].

3 Results

3.1 Overall search results

A total of 1242 papers were obtained using the search strategy, of which seven studies [15–21] evaluating 10 treatment arms were included for the final review (Fig. 1). Data from 5510 participants with 199 adverse pregnancy outcomes were included. Network diagram for individual outcome measures is depicted in (Fig. 2) and the key details of the included studies are provided in (Table 1). Of the total number of studies, five were randomized and two were cohort studies. Two studies [20, 21] enrolled both complicated and uncomplicated cases of malaria as well as women in their first trimester. Risk of bias of the randomized controlled trials is depicted in (Fig. 3). Both Cohort studies scored six in the New Castle Ottawa scale. No evidence of inconsistencies was observed between the direct and indirect comparisons (Fig. 4).

Fig.1

Study flow chart. A total of 1242 studies were obtained of which finally seven studies were found to be eligible for inclusion in this network meta-analysis.

Fig.2

Network diagram of the included studies. A – Abortion, B – Stillbirth and C – Neonatal death; DHP – Dihydroartemisinin-Piperaquine; S-P – Sulphadoxine-pyrimethamine.

Table 1

Summary of key characteristics of the included studies in this meta-analysis

Study ID; Year and country of publication; Study design	Population	Interventions	Outcomes
PREGACT study group [15]; 2016 and Sub-Saharan Africa; RCT	Pregnant women in 2nd or 3rd trimester diagnosed with either Plasmodium falciparum mono-infection of any density, regardless of symptoms with hemoglobin >7 g/dl without any other serious illness were recruited.	DHA-PQ – 40/320 mg once a day for 3 days; MQ-AS – 12 mg/kg AS and 24 mg/kg MQ twice a day for 3 days; AL-20 mg artemether and 120 mg lumefantrine once a day for 3 days; AS-AQ – 100 mg artesunate and 270 mg amodiaquine twice a day for 3 days.	PCR-adjusted cure rates; Adverse pregnancy outcomes; Number of treatment failures; Adverse and serious adverse events
Kalinani et al. [16]; 2007 and Malawi; RCT	15 to 49 year old pregnant with fetal gestation age of 14 to 26 weeks diagnosed to have peripheral parasitemia were included in the study. Those with multiple gestations, history of chronic diseases like tuberculosis, diabetes mellitus, mental disorders, known allergies to sulphonamides, macrolides or pyrimethamine, complications in pregnancy and consumption of any anti-malarial drug within 28 days before enrollment were excluded.	SP- 3 tablets of 500 mg sulphadoxine and 25 mg pyrimethamine each; SP + Azi – 3 tablets of 500 mg sulphadoxine and 25 mg pyrimethamine each with azithromycin 1 g/day for 2 days; SP + Artesunate – 3 tablets of 500 mg sulphadoxine and 25 mg pyrimethamine each with artesunate 200 mg/day for 3 days.	Parasite clearance time; Fever clearance time; Treatment failure; Maternal and birth outcomes; Adverse events
Piola et al. [17]; 2010 and Uganda; RCT	Women with viable pregnancy of estimated gestation of >13 weeks with microscopically detected Plasmodium falciparum either mono-infection or mixed infection were included in the study. Those with density of parasitemia >250,000 per μl, severe anemia (Hemoglobin <7 g/dl), signs and symptoms of severe/complicated malaria needing parenteral treatment or known allergy to study drugs were excluded.	Quinine hydrochloride – 300 mg base quinine per tablet was administered at a dose of 10 mg base/kg every 8 hourly for 7 days; AL- 20 mg artemether and 120 mg lumefantrine 4 tablets each at 0, 8, 24, 36, 48 and 60 hours.	PCR corrected cure rate; Fever clearance time; Parasite clearance time; Hemoglobin levels; Adverse pregnancy outcomes; Adverse events.
Tagbor et al. [18]; 2006 and Ghana; RCT	Pregnant women of any parity with gestational age >16 weeks with microscopically detected Plasmodium falciparum parasite	CQ-Chloroquine 600 mg on day 1 and day 2 and 300 mg on day 3; Amodiaquine 600 mg first 2 days with 300 mg on day 3; SP-1500 mg sulphadoxine and 300 mg pyrimethamine single dose; A + SP–600 mg amodiaquine with 1500 mg sulphadoxine and 300 mg pyrimethamine on day 1 followed by 600 mg amodiaquine on day 2 and 300 mg on day 3.	PCR-adjusted cure rate; Treatment failure; Adverse pregnancy outcomes; Re-infection rate; Adverse events
Ukah et al. [19]; 2015 and Nigeria; RCT	Consecutive pre-natal clinic attendees in 2nd or 3rd trimester with symptoms of malaria and microscopically detected Plasmodium falciparum without any evidence of urinary tract infection were included in the study. Those with complicated malaria, in their first trimester of pregnancy, consumed any anti-malarial drug in the previous week were excluded.	AS-A – 100 mg artesunate with 270 mg amodiaquine twice a day for 3 days; AL-80 mg artemether and 480 mg lumefantrine twice a day for 3 days.	Fever clearance time; Parasite clearance time; Adverse pregnancy outcomes; Treatment failure; Adverse events
Poespoprodjo et al. [20]; 2014 and Indonesia; Cohort study	All pregnant women admitted at a tertiary care hospital with peripheral asexual parasitemia were included.	DHP-PQ – 40 mg dihydroartemisinin and 320 mg piperaquine with a target dose of 2–4 mg/kg for dihydroartemisinin and 16–32 mg/kg for piperaquine for 3 days; Quinine hydrochloride – 300 mg base quinine per tablet was administered at a dose of 10 mg base/kg every 8 hourly for 7 days with clindamycin 20 mg base/kg/day divided into three doses administered for 7 days.	Fever clearance time; Adverse pregnancy outcomes; Adverse events.
Mosha et al. [21]; 2014 and Tanzania; Cohort study	Pregnant women with gestational age of 20 weeks and below with the use of anti-malarial drug in first trimester were recruited.	AL-20 mg artemether and 120 mg lumefantrine once a day for 3 days; Quinine hydrochloride – 300 mg base quinine per tablet was administered at a dose of 10 mg base/kg every 8 hourly for 7 days; 1500 mg sulphadoxine and 300 mg pyrimethamine single dose; Amodiaquine 600 mg day 1 followed by 300 mg each on day 2 and 3.	Adverse pregnancy outcomes

Fig.3

Summary of risk of bias of randomized controlled trials included in this meta-analysis. A total of 5 studies included in this meta-analysis were randomized controlled trials. Green circle with plus symbol indicates the reporting of that specific element in the study while red circle with minus symbol indicates the absence of such reporting.

Fig.4

Deviance plots for determining inconsistency. No significant deviations were observed between the consistency and inconsistency models indicating absence of any inconsistency between direct and indirect estimates.

3.2 Network meta-analysis results

3.2.1 Abortions

Six studies in the present review reported abortions with the use of anti-malarial drugs. However, in one study that compared chloroquine with sulphadoxine-pyrimethamine, this combination remained disconnected with rest of the treatment arms (Fig. 2) and so was not included in the pooled analysis. In a network meta-analysis, the pooled estimates will be derived from both head-to-head comparisons as well as indirectly through a common comparator. When a common comparator is not available, the comparison remains disconnected from the other drugs and cannot be pooled for analysis in the meta-analysis. Hence, a total of five studies in 4001 participants were included for analysing this outcome and the pooled Forest plot for all the possible comparisons is depicted in (Fig. 5). It can be observed that both quinine and artemisinin-lumefantrine had a significantly lower risk of abortion when compared to dihydroartemisinin-piperaquine, artesunate with mefloquine and artesunate with amodiaquine. However, no significant differences were observed between quinine and artemether with lumefantrine. When the results of a cohort study that had compared quinine with dihydroartemisinin-piperaquine in first trimester was excluded from the analysis, we did not observe any significant differences in the risk of abortions between the anti-malarial drugs (Fig. 6).

Fig.5

Forest plot of mixed treatment comparisons for abortions. Quinine and artemisinin plus lumefantrine were found to be statistically significantly better than dihydroartemisinin-piperaquine, artesunate with mefloquine, artesunate with amodiaquine but no significant differences were observed between quine and artemisinin lumefantrine.

Fig.6

Forest plot of mixed treatment comparisons of abortion for anti-malarials administered in second and third trimester. No significant differences were observed in the risk of abortions with any of the anti-malarial agents when administered in 2nd or 3rd trimester of pregnancy.

3.2.2 Stillbirth

All studies included in this review reported the incidence of stillbirth and the pooled estimates for various drug comparisons is depicted in (Fig. 7). No significant differences were observed in the risk of stillbirth with any of the comparisons.

Fig.7

Forest plot of mixed treatment comparisons for stillbirth. No significant differences were observed for stillbirth with any of the comparisons.

3.2.3 Neonatal deaths

Four studies reported neonatal deaths following the use of anti-malarial drugs in the present review. However, as can be observed in (Fig. 2), three arms that were evaluated in the two studies were disconnected and so were not included in the pooled analysis. Forest plot for neonatal deaths from two studies in 802 patients did not reveal any significant risk in any of the treatment comparisons (Fig. 8).

Fig.8

Forest plot of mixed treatment comparisons for neonatal deaths. No significant differences were observed in the risk of neonatal deaths in any of the treatment comparisons.

3.3 Rankogram

SUCRA was found to vary for each of the included interventions depending on the outcome measured. (Figs. 9–11) depicts the rankogram obtained in all three analyses and it can be observed that quinine, sulphadoxine-pyrimethamine and artemether-lumefantrine had the highest probability of being the best interventions for the incidence of abortions, stillbirths and neonatal deaths respectively. However, considering the small number of events included for each of the outcomes in the present meta-analysis, no clinically significant interpretations could be drawn from the plots.

Fig.9

Rankogram plot for abortion. Quinine had the highest probability of being the best interventions with respect to abortions.

Fig.10

Rankogram plot for stillbirth. Sulphadoxine-pyrimethamine had the highest probability of being the best interventions with respect to stillbirths.

Fig.11

Rankogram plot for neonatal death. Artemether-lumefantrine had the highest probability of being the best interventions with respect to neonatal deaths.

3.4 Sensitivity analyses

Data from studies that enrolled both complicated and uncomplicated cases of malaria were excluded to assess their influence on each of the evaluated outcomes. No significant differences were observed in any of the pooled estimates when compared to the overall analyses as well as rankograms. When a study that had included pregnant in their first trimester was excluded, no significant differences were obtained between the anti-malarial drugs regarding their risk of abortion.

3.5 Grading the quality of evidence

Grading of the evidence was carried out and (Table 2) lists the quality of evidence obtained for each of the outcomes assessed. Due to the serious limitations on the directness and precision of the estimates, the quality of evidence in the present review was observed to be low for the outcome measures.

Table 2
Grading the quality of evidence for key comparisons

Outcomes Illustrative comparative risks (95% confidence intervals) Odds ratio [95% credible intervals] No of Participants (studies) Quality of the evidence (GRADE)

Assumed risk¹ Corresponding risk

Abortion with quinine versus DHA-P 315 per 1000 27 per 1000 0.06 1071 (5) ⊕⊖⊖ ⊖

(5 to 104) [0.01, 0.25] very low ^2,3,4

Abortion with quinine in comparison to artemether-mefloquine 5 per 1000 1 per 1000 0.08 1066 (4) ⊕⊖⊖ ⊖

(0 to 2) [0.01, 0.42] very low ^2,3,4

Abortion with artemether-lumefantrine in comparison to DHA-P 315 per 1000 35 per 1000 0.08 2055 (6) ⊕⊖⊖ ⊖

(5 to 117) [0.01, 0.29] very low ^2,3,4

Abortion with quinine in comparison to artesunate-amodiaquine 10 per 1000 1 per 1000 0.10 1125 (5) ⊕⊖⊖ ⊖

(0 to 6) [0.01, 0.59] very low ^2,3,4

Abortion with artemether-lumefantrine in comparison to artesunate-mefloquine 5 per 1000 1 per 1000 0.11 2050 (5) ⊕⊕ ⊖⊖

(0 to 2) [0.01, 0.43] low ^2,3

Abortion with artemether-lumefatnrine in comparison to artesuante-amodiaquine 10 per 1000 1 per 1000 0.14 1701 (3) ⊕⊕ ⊖⊖

(0 to 6) [0.01, 0.56] low ^2,3

Outcomes	Illustrative comparative risks (95% confidence intervals)	Odds ratio [95% credible intervals]	No of Participants (studies)	Quality of the evidence (GRADE)
Abortion with quinine versus DHA-P	315 per 1000	27 per 1000	0.06	1071 (5)	⊕⊖⊖ ⊖
		(5 to 104)	[0.01, 0.25]		very low ^2,3,4
Abortion with quinine in comparison to artemether-mefloquine	5 per 1000	1 per 1000	0.08	1066 (4)	⊕⊖⊖ ⊖
		(0 to 2)	[0.01, 0.42]		very low ^2,3,4
Abortion with artemether-lumefantrine in comparison to DHA-P	315 per 1000	35 per 1000	0.08	2055 (6)	⊕⊖⊖ ⊖
		(5 to 117)	[0.01, 0.29]		very low ^2,3,4
Abortion with quinine in comparison to artesunate-amodiaquine	10 per 1000	1 per 1000	0.10	1125 (5)	⊕⊖⊖ ⊖
		(0 to 6)	[0.01, 0.59]		very low ^2,3,4
Abortion with artemether-lumefantrine in comparison to artesunate-mefloquine	5 per 1000	1 per 1000	0.11	2050 (5)	⊕⊕ ⊖⊖
		(0 to 2)	[0.01, 0.43]		low ^2,3
Abortion with artemether-lumefatnrine in comparison to artesuante-amodiaquine	10 per 1000	1 per 1000	0.14	1701 (3)	⊕⊕ ⊖⊖
		(0 to 6)	[0.01, 0.56]		low ^2,3

¹Assumed risk was the median control group risk across the studies. ²Downgraded one level for serious limitations in the precision as evident by a wide confidence interval. ³Downgraded one level for low number of events observed in the individual studies. ⁴Downgraded one level as one of the included studies for that outcome had also recruited pregnant during their first trimester in addition to women in second and third trimesters. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

4 Discussion

We carried out the present network meta-analysis to compare the safety of anti-malarial drugs in terms of adverse pregnancy outcomes from randomized controlled trials and cohort studies. Seven studies were included, of which five were randomized and two were cohort studies. We observed that all the anti-malarial drugs evaluated in the present review have similar risk profile in terms of abortion, stillbirth and neonatal deaths when administered in either second or third trimester of pregnancy. When the data on studies that were conducted in the first trimester were also included in the pooled analysis, quinine and artemether plus lumefantrine were observed to carry the lowest risk of abortion. Compared to the conventional meta-analysis, Bayesian network meta-analysis permits comparison of various treatment options available for a disease state even in the absence of any head-to-head comparison trials. We found that the treatment options evaluated in the present NMA were in agreement with the WHO guidelines. To the best of our knowledge, this is the first network meta-analysis conducted for anti-malarial drugs in pregnant women.

Malaria in pregnancy is associated with high morbidity and mortality in first trimester and guidelines recommend preventive measures such as medicated mosquito nets and intermittent preventive therapy for pregnant women living in high endemic areas [22]. A recent Cochrane review comparing various treatment strategies for uncomplicated malaria in pregnant women concluded that scant evidence precluding any conclusion exists as far as adverse pregnancy outcomes are concerned [23]. Kovacs et al. [6] observed that artemisinin combination therapies were safe in comparison with other antimalarial drugs, when used in the second or third trimester. However, the present network meta-analysis has the following advantages: Pooled estimates were obtained by both direct and indirect comparisons; Analysis of pooled estimates individually for each of abortions, stillbirths and neonatal deaths; ranking of the treatment arms and grading the quality of evidence. Additionally, we carried out sensitivity analyses by excluding trials conducted in women in their first trimester as well as complicated cases of malaria. McGready et al. [24] conducted a population based study on 48426 pregnant women with malaria in their first trimester and observed that the risk of miscarriage was similar for women treated with chloroquine (92, 26% of 354), quinine (95, 27% of 355), or artesunate (20, 31% of 64; p = 0.71). WHO recommends seven days of therapy with quinine and clindamycin for the treatment of malaria in first trimester and artemisinin combination treatment as a second line of choice, in the absence of considerable evidence. A large prospective cohort study is underway in three Sub-Saharan African countries to evaluate the safety of artemisinin combination treatments when administered during first trimester of pregnancy [25]. Although in the present review, the participants of the included studies were diagnosed to have either Plasmodium falciparum mono-infection or mixed with Plasmodium vivax species, the anti-malarial drugs recommended to be used in patients with Plasmodium vivax malaria include quinine, artemisinin combination treatments and chloroquine weekly prophylaxis. We observed that all the anti-malarial drugs used for Plasmodium vivax were comparably safe to be used in pregnant women. Network meta-analyses for cure rates and recrudescence/relapse following the use of anti-malarial drugs in pregnant women with malaria are the need of the hour. We also observed that quinine, sulphadoxine-pyrimethamine and artemether-lumefantrine had the highest probability of being ranked the best treatment in terms of risks of abortion, stillbirth and neonatal deaths. However, due to small number of events for each of the outcomes, no clinical significance should be attached for the same. Further, widespread resistance to sulphadoxine-pyrimethamine in many parts of Africa and Asia precludes its consideration in pregnant women for the treatment of Plasmodium falciparum malaria [26, 27].

The study is limited by the fact that small number of events was observed with each of the outcome measures and treatment options. In conclusion, from the present review we conclude that all the anti-malarial drugs recommended by WHO have similar risk profiles when administered in second or third trimester of pregnancy. However, quinine and artemether-lumefantrine carry the lowest risk of leading to abortions when administered in first trimester of pregnancy.

Author contributions

Kannan Sridharan had full access to all of the data in the study. Kannan Sridharan takes responsibility for the integrity of the data, the accuracy of the data analysis, and the final decision to submit for publication. Study concept and design: Sridharan, Sivaramakrishnan, Kanters; Acquisition, analysis, or interpretation of data: Sridharan, Sivaramakrishnan, Kanters; Drafting of the manuscript: Sridharan, Sivaramakrishnan, Kanters; Critical revision of the manuscript for important intellectual content: Sridharan, Sivaramakrishnan, Kanters; Statistical analysis: Sridharan, Sivaramakrishnan, Kanters.

Conflict of interest

The authors do not have any conflict of interest to report.

Funding

No funding was obtained for this review.

Footnotes

Acknowledgments

We thank PROSPERO for registering the review protocol and Canadian Agency for Drugs and Technologies in Health for utilizing NetMetaXL in generating pooled results. We are also grateful to MRC Biostatistics Unit, Cambridge for using WINBUGS software and Cochrane group for utilizing RevMan software and GRADE tool.

References

Malaria in pregnant women. World Health Organization. Available at: http://www.who.int/malaria/areas/high_risk_groups/pregnancy/en/ (Accessed on 14 Dec 2016).

van Eijk

Hill

Noor

Snow

, et al. Prevalence of malaria infection in pregnant women compared with children for tracking malaria transmission in sub-Saharan Africa: A Systematic review and meta-analysis. The Lancet Global Health. 2015;3:e617–e628.

Newman

Robalo

Quakyi

. Malaria during pregnancy: Epidemiology, current prevention strategies, and future directions. Emerg Infect Dis. [serial on the Internet]. 2004 November [14 Dec 2016]. https://dx-doi-org.web.bisu.edu.cn/10.3201/eid1011.040624_09

Moore

Salman

Davis

TME

. Treatment regimens for pregnant women with falciparum malaria. Expert Review of Anti-Infective Therapy. 2016;8:691–704.

Kovacs

Maria van Eijk

Sevene

, et al. The safety of artemisinin derivatives for the treatment of malaria in the 2nd or 3rd trimester of pregnancy: A Systematic review and meta-analysis. PLoS One. 2016;11:e0164963.

Intermittent preventive treatment in pregnancy. World Health Organization. Available at: http://www.who.int/malaria/areas/preventive_therapies/pregnancy/en/ (Accessed on 14 Dec 2016).

Mills

Ioannidis

Thorlund

, et al. How to use an article reporting a multiple treatment comparison meta-analysis. JAMA. 2012;308:1246–53.

Ades

. Assessing Evidence Inconsistency in Mixed Treatment Comparisons. Journal of the American Statistical Association. 2006;101:447–59.

Higgins

JPT

Green

(editors). Cochrane Handbook for Systematic Reviews of Interventions. 5.1.0 edition. Available from www.cochrane-handbook.org (Accessed on 14 Dec 2016).

10.

Wells

Shea

O’Connell

, et al. The Newcastle-Ottawa scale for assessing the quality of nonrandomized studies in meta-analysis. Available at: http://www.medicine.mcgill.ca/rtamblyn/Readings%5CThe%20Newcastle%20-%20Scale%20for%20assessing%20the%20quality%20of%20nonrandomised%20studies%20in%20meta-analyses.pdf (Accessed on 14 Dec 2016).

11.

Hutton

Salanti

Caldwell

, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: Checklist and explanations. Ann Intern Med. 2015;162:777–84.

12.

Valkenhoef

Brock

, et al. Automating network meta-analysis. Research Synth Methods. 2012;3:285–99.

13.

Rucker

Schwarzer

. Ranking treatments in frequentist network meta-analysis works without resampling methods. BMC Medical Research Methodology. 2015;15:8.

14.

Brown

Hutton

Clifford

, et al. A Microsoft-Excel based tool for running and critically appraising network meta-analyses-an overview and application of NetMetaXL. Systematic Reviews. 2014;3:110.

15.

The PREGACT study group. Four Artemisinin-based treatments in African pregnant women with malaria. New Engl. J. Med. 2016;374:913–27.

16.

Kalinani

Mofolo

Chaponda

, et al. A randomized controlled pilot trial of Azithromycin or Artesunate added to Sulfadoxine-Pyrimethamine as treatment for malaria in pregnant women. PLoS One 2(11):e1166.

17.

Piola

Nabasumba

Turyakira

, et al. Efficacy and safety of artemether–lumefantrine compared with quinine in pregnant women with uncomplicated Plasmodium falciparum malaria: An Open-label, randomised, non-inferiority trial. Lancet Infect Dis. 2010;10:762–9.

18.

Tagbor

Bruce

Browne

, et al. Efficacy, safety, and tolerability of amodiaquine plus sulphadoxine-pyrimethamine used alone or in combination for malaria treatment in pregnancy: A Randomised trial. Lancet. 2006;368:1349–56.

19.

Ukah

Badejoko

Ogunniyi

, et al. A randomized trial of artesunate-amodiaquine versus artemether-lumefantrine for the treatment of acute uncomplicated malaria in pregnancy. International Journal of Gynecology and Obstetrics. 2015;131:41–44.

20.

Poespoprodjo

Fobia

Kenangalem

, et al. Dihydroartemisinin-piperaquine treatment of multidrug resistant Falciparum and Vivax malaria in pregnancy. PLoS ONE. 2014;9(1):e84976.

21.

Mosha

Mazuguni

Mrema

, et al. Safety of artemether-lumefantrine exposure in first trimester of pregnancy: An Observational cohort. Malaria Journal. 2014;13:197.

22.

Brutus

Santalla

Schneider

, et al. Plasmodium vivax malaria during pregnancy, Bolivia. Emerging Infectious Diseases. 2013;19:1605–11.

23.

Orton

Omari

AAA

. Drugs for treating uncomplicated malaria in pregnant women. Cochrane Database of Systematic Reviews. 2008, Issue 4. Art. No.: CD004912.

24.

McGready

Lee

Wiladphaingern

, et al. Adverse effects of falciparum and vivax malaria and the safety of antimalarial treatment in early pregnancy: A Population-based study. Lan Infect Dis. 2012;12:388–96.

25.

Tinto

Sevene

Dellicour

, et al. Assessment of the safety of antimalarial drug use during early pregnancy (ASAP): Protocol for a multicenter prospective cohort study in Burkina Faso, Kenya and Mozambique. Reproductive Health. 2015;12:112.

26.

Taylor

Antonia

Parobek

, et al. Plasmodium falciparum sulphadoxine resistance is geographically and genetically clustered within the DR Congo. Scientific Reports. 2013;3:1165.

27.

Sibley

Hyde

Sims

, et al. Pyrimethamine-sulfadoxine resistance in Plasmodium falciparum: What next? Trends Parasitol. 2001;17:582–8.