The effect of antenatal monthly sulphadoxine–pyrimethamine, alone or with azithromycin, on foetal and neonatal growth faltering in Malawi: a randomised controlled trial

Authors


Corresponding Author Mari Luntamo, Kellonsoittajankatu 3-7 B 29, 20500 Turku, Finland. Tel.: +358 50 3427778; E-mail :mari.luntamo@gmail.com

Abstract

Objective

To examine the potential to reduce foetal and neonatal growth faltering through intermittent preventive treatment in pregnancy (IPTp) of malaria and reproductive tract infections with monthly sulphadoxine–pyrimethamine (SP), alone or with two doses of azithromycin.

Methods

We enrolled 1320 women with uncomplicated second trimester pregnancies into a randomised, partially placebo controlled, outcome assessor-blinded clinical trial in Malawi. The participants received either two doses of SP (control), SP monthly (monthly SP) or SP monthly and azithromycin (1 g) twice (AZI-SP). Newborn size was measured within two days of birth and infant growth at four weeks of age.

Results

Babies in the AZI-SP group were on average (95% CI) 140 g (70–200) heavier at birth and 0.6 cm (0.2–0.9) longer at four weeks of age than control group babies. Corresponding differences between the monthly SP and control groups were 80 g (20–140) and 0.3 cm (−0.0 to 0.6). Compared with controls, babies in the AZI-SP group had a relative risk of 0.61 (0.40–0.93) for low birthweight, 0.60 (0.44–0.81) for stunting and 0.48 (0.29–0.79) for underweight at four weeks of age. Corresponding differences were similar but smaller between the monthly SP and control groups.

Conclusions

An IPTp regimen with monthly SP given to all pregnant women is likely to increase mean birthweight and length at four weeks of age in malaria holoendemic areas. Adding azithromycin to the regimen appears to offer further benefits in reducing foetal and neonatal growth faltering.

Abstract

Objectif

Examiner la possibilité de réduire la croissance fœtale et néonatale chancelante grâce à un traitement préventif intermittent (TPI) pendant la grossesse du paludisme et des infections du tractus reproducteur avec de la sulfadoxine-pyriméthamine (SP) mensuellement, seule ou avec deux doses d'azithromycine.

Méthodes

Nous avons recruté 1320 femmes avec des grossesses non compliquées au second trimestre, dans un essai clinique randomisé au Malawi, partiellement contrôlé par placebo, en aveugle pour l’évaluateur des résultats. Les participantes ont reçu soit deux doses de SP (témoin), de la SP mensuellement (SP mensuelle) ou de la SP mensuellement avec deux doses d'azithromycine (1g) (AZI-SP). La taille du nouveau-né a été mesurée dans les deux jours suivant la naissance et la croissance des nourrissons à quatre semaines d’âge.

Résultats

Les bébés du groupe AZI-SP étaient en moyenne (IC 95%) 140 g (70 à 200) plus lourds à la naissance et 0.6 cm (0.2 à 0.9) plus grands à quatre semaines d’âge que les bébés du groupe témoin. Les différences correspondantes entre les groupes ‘SP mensuelle’ et témoin étaient de 80 g (20 à 140) et 0.3 cm (−0.0 à 0.6). Comparés aux témoins, les bébés du groupe AZI-SP avaient un risque relatif de 0.61 (0.40 à 0.93) pour un poids faible à la naissance, de 0.60 (0.44 à 0.81) pour un retard de croissance et de 0.48 (0.29 à 0.79) pour un poids insuffisant à quatre semaines d’âge. Les différences correspondantes étaient similaires, mais plus petites entre les groupes ‘SP mensuelle’ et témoin.

Conclusions

Un régime de TPI avec de la SP mensuelle donnée à toutes les femmes enceintes est de nature à augmenter le poids moyen à la naissance et la taille à quatre semaines d’âge dans les zones holoendémiques pour le paludisme. L'ajout d'azithromycine au régime semble offrir des avantages supplémentaires en réduisant la croissance fœtale et néonatale chancelante.

Abstract

Objetivo

Evaluar el potencial de reducir el retraso en el crecimiento fetal y neonatal mediante el tratamiento preventivo intermitente de la malaria durante el embarazo (IPTp) y las infecciones del tracto reproductivo con dosis mensuales de sulfadoxina-pirimetamina, sola o con dos dosis de azitromicina.

Métodos

Incluimos 1320 mujeres que estaban en el segundo trimestre de un embarazo sin complicaciones, en un ensayo clínico aleatorizado, parcialmente controlado con placebo, enmascarado para el evaluador y realizado en Malawi. Las participantes recibieron o bien dos dosis de SP (control), SP mensual, o SP mensual más dos dosis de azitromicina (1g) (AZI-SP). El tamaño del recién nacido se midió dentro de los dos días siguientes al día de nacimiento, y el crecimiento del neonato a las cuatro semanas de edad.

Resultados

Los bebés dentro el grupo AZI-SP tenían como media (IC 95%) 140 g (70 a 200) más de peso al nacer y eran 0.6 cm (0.2–0.9) más largos a las cuatro semanas de edad que los bebés del grupo control. Las diferencias correspondientes entre el grupo con SP mensual y el grupo control eran de 80 g (20 a 140) y 0.3 cm (−0.0–0.6). Comparados con los controles, los bebés en el grupo AZI-SP tenían un riesgo relativo de 0.61 (0.40–0.93) de un bajo peso al nacer, 0.60 (0.44 a 0.81) de retraso en el crecimiento y 0.48 (0.29 a 0.79) de bajo peso a las cuatro semanas de edad. Las diferencias correspondientes eran similares pero menores entre el grupo recibiendo SP mensual y el grupo control.

Conclusiones

Un régimen de IPTp con SP mensual ofrecido a todas las mujeres embarazadas podría aumentar el peso medio al nacer y la longitud del neonato a las cuatro semanas de vida en áreas holoendémicas para malaria. El añadir azitromicina a dicho régimen parece ofrecer más beneficios reduciendo el retraso en el crecimiento fetal y neonatal.

Introduction

In sub-Saharan Africa (SSA), the incidence of pre-term delivery is high, about 14% of babies are born with low birthweight (LBW) and more than 40% of children under five years old are stunted (Osman et al. 2001; van den Broek et al. 2005; Beck et al. 2010; Luntamo et al. 2010; UNICEF 2011). These conditions are interrelated in that babies born pre-term often have LBW, which in turn predisposes the infants in resource-poor settings to growth faltering and undernutrition (ACC/SCN 2000; Espo et al. 2002; Lawn et al. 2010). Pre-maturity, LBW and undernutrition all have serious long-term consequences and contribute to high neonatal and infant mortality (ACC/SCN 2000; Desai et al. 2007; Black et al. 2008; Moster et al. 2008; Saigal & Doyle 2008; Victora et al. 2008; Lawn et al. 2010; UNSCN 2010).

Maternal malaria and reproductive tract infections (RTI) are common in SSA and believed to contribute to increased risk of intrauterine growth retardation (IUGR) and pre-term delivery (Guyatt & Snow 2004; Mullick et al. 2005; Desai et al. 2007; Swadpanich et al. 2008; Chico et al. 2012). Many African countries have therefore adopted a policy of providing intermittent preventive treatment in pregnancy (IPTp) against malaria to all pregnant women (WHO 2010). However, the effectiveness of the standard IPTp regimen with two doses of sulphadoxine–pyrimethamine (SP) has been questioned and more frequent dosing or other treatment regimens have been suggested (White 2005; Gill et al. 2007; ter Kuile & Steketee 2007; ter Kuile et al. 2007; Menéndez et al. 2007; Sridaran et al. 2010; Chico & Chandramohan 2011a; van Eijk et al. 2011). Pooled results of three previous trials showed that human immunodeficiency virus (HIV)-infected paucigravidae treated monthly with SP, rather than twice, had lower malaria prevalence at delivery, and their babies had a higher mean birthweight (ter Kuile et al. 2007). There are, however, little published data on the impact of more frequent SP dosing on birth outcomes among pregnant women who have not been screened for known risk factors.

We have previously reported an approximately 35% lower incidence of pre-term delivery and LBW among newborns of rural Malawian women treated with monthly SP and two doses of azithromycin, rather than the standard two-dose SP regimen (Luntamo et al. 2010). Azithromycin is effective against many RTI and has some antimalarial activity (Chico & Chandramohan 2011b; Pfizer Inc. 2012). Also newborns of women treated with monthly SP but no azithromycin had a lower incidence of pre-term delivery and LBW, but the differences to control group were smaller. There were no major differences between the study groups in serious adverse events, perinatal or neonatal mortality (Luntamo et al. 2010). In this article, we report the neonatal outcomes from this trial, including anthropometrics at birth and at one month of age.

Methods

Study design

We undertook a single-centre, randomised, partially placebo controlled, outcome assessor-blinded, three-arm clinical trial in rural Malawi. The study hypothesis was that IPTp with monthly SP, alone or in combination with two doses of azithromycin, improves foetal growth and decreases the incidence of pre-term delivery, thus leading to an increased infant size at birth and at one month of age. The primary efficacy and safety outcome measures were the incidence of pre-term delivery and serious adverse events, respectively; these results and other details of the study have been reported previously in Luntamo et al. 2010. In the current article, we report the results on pre-defined secondary outcomes regarding newborn and neonate size.

Participants, study site and ethics

The study enrolled women with uncomplicated second trimester pregnancies (gestational age 14–26 weeks by ultrasound assessment) who started antenatal care between December 2003 and October 2006 at Lungwena Health Centre, southern Malawi. Malaria is holoendemic, and earlier evidence suggested a high prevalence of RTI among pregnant women at the study site (Kulmala et al. 2001; University of Malawi, Save the Children Federation USA, Malawi Ministry of Health & Population & MEASURE Evaluation 2004). Appropriate ethical approvals were obtained (for details see Luntamo et al. 2010).

Study interventions

Participants in the control group received standard Malawian antenatal care, which included IPTp with SP (three tablets orally, each containing 500 mg of sulphadoxine and 25 mg of pyrimethamine) twice: at enrolment and between 28–34 weeks of gestation. At the same visits, they also received a placebo to azithromycin. Participants in the monthly SP intervention group received SP monthly from enrolment until 37 gestational weeks and a placebo to azithromycin as the control group. Participants in the AZI-SP intervention group received monthly SP and two doses of active azithromycin (two tablets orally, each containing 500 mg of azithromycin): at enrolment and between 28–34 weeks of gestation. All participants received ferrous sulphate (200 mg/day) and folic acid (0.25 mg/day) throughout their pregnancy. SP tablets were purchased from Malawi Central Medical Stores, and active azithromycin and its placebo were manufactured and donated by Pfizer Inc.

Enrolment

At enrolment, research personnel interviewed interested individuals about their socio-economic status and obstetric history, gave pre-test HIV counselling and performed an antenatal examination. The duration of pregnancy was determined with an ultrasound imager. All participants were tested for syphilis (VDRL carbon antigen, Biotec Laboratories, Ipswich, UK or Determine Syphilis TP, Abbott Laboratories; positive results were confirmed with TPHA kit, Lorne Laboratories). Those with confirmed syphilis were treated at the study site with intramuscular benzathine penicillin (2.4 mU) and their babies received the same drug (50 kU/kg) at the age of 4 weeks. Standard methods were used to assess blood haemoglobin concentration and peripheral blood malaria parasitaemia from all participants and to perform HIV tests for those who opted for it (for details, see Luntamo et al. 2010).

Eligible individuals signed or thumb-printed informed consent and picked one randomisation envelope that contained an identification number (for details on randomisation, see Luntamo et al. 2010). A research assistant not involved in outcome assessment gave the corresponding pre-packed study drugs to the participant under direct observation and monitored her for possible adverse reactions.

Follow-up

At follow-up visits (at four-week intervals until 36 completed gestational weeks and weekly thereafter), research personnel conducted an antenatal examination. The participants were offered post-test HIV counselling and, if needed, prevention of mother to child transmission with nevirapine. At each visit, the participant took the appropriate pre-packed study drugs under direct observation.

After delivery, a research assistant measured the newborn's head and chest circumference with a non-stretchable plastic tape (Lasso-o tape, Harlow Printing Limited, reading increment 1 mm). If the measuring happened at home, she weighed the infant with a spring scale (Super Samson, Salter Brecknell) and recorded the result to the nearest 50 g. At a health facility, birthweight was measured with an electronic infant weighing scale (in Lungwena health centre: SECA 834, Chasmors Ltd, reading increment 10 g). The newborn was given nevirapine or placebo based on maternal HIV status.

At one-month post-natal visit, research personnel measured the infant's weight with an electronic infant weighing scale (SECA 834, Chasmors Ltd), length with a length board (Kiddimetre, Raven Equipment Ltd, reading increment 1 mm) and chest and head circumference as well as MUAC with a non-stretchable plastic tape (same as at delivery).

Statistical analyses

Based on earlier results from Malawi and Mozambique (Schultz et al. 1994; Kulmala et al. 2000; Osman et al. 2001; van den Broek et al. 2005), we estimated that a target sample size of 1320 participants would provide 80% power at a 5% level of significance to detect a 7.5% reduction in the incidence of LBW from 20% to 12.5% and a 95 g difference in mean birthweight (for details, see Luntamo et al. 2010).

Statistical analyses were conducted with Stata 9.2 (StataCorp, College Station, USA) on the principle of intention to treat. Birthweights measured more than two days after delivery, other birth measurements taken more than one week after delivery, and post-natal variables with a measuring date that differed more than two weeks from the target, were excluded from the analyses. The anthropometric analyses at one month after delivery were adjusted for age at measurement by regression methods. Age- and gender-standardised anthropometric indices (length-for-age, weight-for-age and weight-for-length Z-scores) were calculated using the WHO Child Growth Standards (WHO 2006), and values below −2.0 were considered indicative of stunting, underweight and wasting, respectively.

We estimated risk ratio (RR) and risk difference for comparison of binary end-points at a single time-point. To prevent inflated type I errors by multiple comparisons, we began hypothesis testing with the global null hypothesis of all three groups being identical before doing the two-group comparisons. We tested the hypotheses either with Fisher's exact test (binary end-points) or analysis of variance (quantitative end-points).

Pre-term delivery was defined as birth before 37 completed gestational weeks and LBW as birthweight < 2500 g. We included two sets of sensitivity analyses due to digit preference (8.1% of newborns were recorded to weigh exactly 2500 g): one defined LBW as 2500 g, another considered a birthweight of 2500 g a missing value and used a multiple imputation method to replace the value (van Buuren et al. 1999; Royston 2004). The imputation utilised gestational duration, maternal weight gain, gender and study groups as covariates. We also used a chest circumference < 30 cm at birth as a proxy for LBW (WHO 1993; Walraven et al. 1994).

We performed tests for interaction between the interventions and the number of previous pregnancies (categorised as none, one and two or more), HIV status and bed net use at enrolment using the likelihood ratio test and did analyses stratified by the number of previous pregnancies for LBW. The proportion of women with none or one previous pregnancy and with malaria parasitaemia at enrolment was higher in the control than in the intervention groups. As sensitivity analyses for birthweight and LBW, we adjusted for these two covariates and child sex as categorical variables by generalised linear models. To determine whether the differences in birthweight between the control and intervention groups were due to differences in the duration of pregnancy, we performed an analysis further adjusting for gestational age at birth.

Results

Of the 3358 pregnant women whom we invited to participate in the study, 1320 were randomised into the three study groups: control, monthly SP and AZI-SP (Figure 1). At enrolment, the three groups were comparable, except for small differences in the number of previous pregnancies and the prevalence of malaria parasitaemia (Table 1). The enrolled individuals and those not enrolled had approximately the same mean age (25 vs. 26 years) and number of previous pregnancies (2.3 vs. 2.5).

Table 1. Baseline characteristics of participants at enrolment
CharacteristicControl (SP twice)Monthly SPAZI-SP
  1. SD, standard deviation; BMI, body mass index; Hb, haemoglobin.

  2. a

    Includes those with a positive VDRL/Determine result, but negative TPHA result (9, 6 and 7 participants in the control, monthly SP and AZI-SP groups respectively).

Number of enrolled women436441443
Mean (SD) age in years25 (7)25 (7)25 (6)
Mean (SD) height in cm155.0 (5.5)154.8 (5.4)155.3 (5.6)
Mean (SD) BMI as kg/m221.7 (2.2)21.8 (2.1)21.9 (2.1)
Mean (SD) gestational age at enrolment in weeks20.3 (3.0)20.0 (3.2)20.0 (3.0)
Number of previous pregnancies, proportion (%)
None110/436 (25.2)107/441 (24.3)89/443 (20.1)
One86/436 (19.7)78/441 (17.7)82/443 (18.5)
Two or more240/436 (55.1)256/441 (58.1)272/443 (61.4)
Maternal HIV status, proportion (%)
Positive48/436 (11.0)64/441 (14.5)49/443 (11.1)
Negative348/436 (79.8)336/441 (76.2)349/443 (78.8)
Unknown40/436 (9.2)41/441 (9.3)45/443 (10.2)
Maternal syphilis status, proportion (%)
Positive18/436 (4.1)27/441 (6.1)21/443 (4.7)
Negativea415/436 (95.2)408/441 (92.5)419/443 (94.6)
Unknown3/436 (0.7)6/441 (1.4)3/443 (0.7)
Mean (SD) blood Hb concentration as g/l110 (19)111 (17)110 (20)
Maternal anaemia status, proportion (%)
Mild or moderate (Hb < 110)215/436 (49.3)205/441 (46.5)215/443 (48.5)
Moderate (Hb <80 g/l)22/436 (5.1)11/441 (2.5)30/443 (6.8)
Proportion (%) with microscopic peripheral blood malaria parasitaemia49/435 (11.3)41/441 (9.3)27/443 (6.1)
Proportion (%) of literate participants116/436 (26.6)129/441 (29.3)139/443 (31.4)
Mean (SD) years of schooling completed2.1 (2.7)2.1 (2.6)2.4 (2.8)
Proportion (%) of those owning a bed-net320/436 (73.4)318/441 (72.1)330/443 (74.5)
Proportion (%) who used bed-net during previous night268/436 (61.5)262/441 (59.4)267/443 (60.3)
Proportion (%) with previous delivery complication118/436 (27.1)116/441 (26.3)118/443 (26.6)
Figure 1.

Participant flow in CONSORT recommended format.

The mean (standard deviation, SD) number of scheduled SP treatments received was 2.0 (0.2, range 1–2) in the control, 4.0 (1.0, range 1–6) in the monthly SP and 4.0 (0.9, range 1–6) in the AZI-SP group. Women in the AZI-SP group received a mean (SD) of 2.0 (0.2) azithromycin doses. Against the trial protocol, few SP doses were given also at non-scheduled visits (mean 0.008 doses/participant, = 0.31).

Follow-up data were obtained from 91% for birthweights and from 85% for post-natal anthropometrics (Figure 1). The success rate of follow-up was similar between study groups, except slightly lower in the monthly SP group for neonatal anthropometrics (P = 0.03). There were no differences between the groups in the number of birthweight measurements excluded from the analysis due to delay of measurement beyond two days (P = 0.15).

Mean (SD) birthweight was 2890 (470) g among the controls, 2970 (480) g in the monthly SP and 3020 (450) g in the AZI-SP group (Table 2). Compared with the control group, newborns in the monthly SP group had on average (95% CI) 80 g (20–140, P = 0.02) and those in the AZI-SP group 140 g (70–200, < 0.001) higher birthweight. The adjustment of birthweight for baseline malaria, number of previous pregnancies and child sex reduced the mean difference between each intervention group and the control group by about 13%. When gestational age was added to the adjustments, the further reduction was 29–42%. Newborns in the AZI-SP group had on average (95% CI) 0.5 cm (0.2–0.8, P = 0.002) larger chest circumference and 0.6 cm (0.3–0.8, P < 0.001) larger head circumference than the control babies.

Table 2. Birthweight, unadjusted and adjusted analyses, chest and head circumference at birth
OutcomeResult by study groupComparison between monthly SP and control groupComparison between AZI-SP and control group
ControlMonthly SPAZI-SPP-valueMean difference (95% CI)P-valueMean difference (95% CI)P-value
  1. SD, standard deviation.

Mean (SD) birthweight (g), unadjusted2890 (470)2970 (480)3020 (450)<0.00180 (20 to 140)0.02140 (70 to 200)<0.001
Mean (SD) birthweight (g), adjusted for malaria at enrolment, the number of previous pregnancies and child sex    70 (10 to 130)0.03120 (50 to 180)<0.001
Mean (SD) birthweight (g), adjusted for malaria at enrolment, the number of previous pregnancies, child sex and gestational age at birth    50 (−10 to 100)0.1070 (20 to 130)0.009
Mean (SD) chest circumference at birth (cm)32.6 (2.2)32.7 (2.2)33.0 (2.1)0.0060.1 (−0.2 to 0.4)0.500.5 (0.2 to 0.8)0.002
Mean (SD) head circumference at birth (cm)34.2 (1.8)34.4 (1.7)34.8 (1.5)<0.0010.2 (−0.0 to 0.5)0.050.6 (0.3 to 0.8)<0.001

The overall proportion of LBW was 10.0%. Compared with the controls, babies in the AZI-SP group had a RR of 0.61 (0.40–0.93, P = 0.02) and absolute risk reduction of 5.1% (0.9–9.3%) for LBW, whereas babies in the monthly SP group did not differ significantly from the controls (Table 3). The sensitivity analysis adjusting for baseline malaria, number of previous pregnancies and child sex, as well as analyses that used the two alternative definitions of LBW or chest circumference <30 cm as a LBW proxy (see 'Methods') as outcome variables, yielded comparable RRs as did the main analysis for LBW.

Table 3. Incidence of low birthweight (LBW) with different definitions and incidence of low chest circumference at birth
OutcomeNumber with outcome/infants with outcome data (%)Comparison between monthly SP and control groupComparison between AZI-SP and control group
ControlMonthly SPAZI-SPP-valueRisk ratio (95% CI)P-valueRisk ratio (95% CI)P-value
  1. CI, confidence interval; LBW, low birthweight.

  2. a

    Based on 10 imputed data sets. Denominator is the number of subjects analysed, numerator is the average (over 10 data sets) number of LBW subjects rounded to the nearest integer.

  3. b

    Adjusted for malaria at enrolment, the number of previous pregnancies (categorised as none, one, and two or more) and child sex.

Birthweight  < 2500 g, unadjusted52/402 (12.9)36/394 (9.1)32/406 (7.9)0.0480.71 (0.47 to 1.06)0.090.61 (0.40 to 0.93)0.02
Birthweight ≤ 2500 g, unadjusted92/402 (22.9)73/394 (18.5)52/406 (12.8)0.0010.81 (0.62 to 1.06)0.140.56 (0.41 to 0.76)<0.001
Birthweight < 2500 g, partly imputeda, unadjusted63/402 (15.7)47/394 (12.0)38/406 (9.3)0.040.77 (0.52 to 1.12)0.170.60 (0.40 to 0.89)0.01
Birthweight < 2500 g, adjusted analysisb    0.74 (0.50 to 1.10)0.140.64 (0.42 to 0.96)0.03
Birthweight < 2500 g stratified based on number of previous pregnancies
None29/101 (28.7)16/92 (17.4)6/78 (7.7)0.0010.61 (0.35 to 1.04)0.090.27 (0.12 to 0.61)<0.001
One8/79 (10.1)12/73 (16.4)9/73 (12.3)0.511.62 (0.70 to 3.75)0.341.22 (0.50 to 2.99)0.80
Two or more15/222 (6.8)8/229 (3.5)17/255 (6.7)0.210.52 (0.22 to 1.20)0.140.99 (0.50 to 1.93)>0.99
Low chest circumference (<30 cm) at birth as LBW proxy36/403 (8.9)28/392 (7.1)24/411 (5.8)0.240.80 (0.50 to 1.28)0.370.65 (0.40 to 1.08)0.11

There was a statistically significant interaction on the incidence of LBW between study group and the number of previous pregnancies (P = 0.01), but not with HIV status (P = 0.63), or bed net use at enrolment (P = 0.96). This interaction was mainly related to variation in the LBW incidence in the control group, ranging from 28.7% among women with no prior pregnancies to 6.8% among women with two or more earlier pregnancies (Table 3). In the AZI-SP group, primigravidae had only slightly higher LBW incidence (7.7%) than multigravidae (6.7%).

At one month after birth, the mean differences between groups in weight, head and chest circumference were similar to the corresponding differences at birth (Table 4). The overall prevalence of infant stunting was 21.2%, underweight 9.1% and wasting 1.7%. Groups differed in terms of stunting (< 0.001) and underweight (= 0.007), but not in wasting (P = 0.92). Compared to the control group, infants in the AZI-SP group had a RR of 0.60 (= 0.001) for stunting and 0.48 (= 0.003) for underweight. Figure 2 shows the size distribution for infants in each study group at birth and at one month of age.

Table 4. Growth outcomes for infants, adjusted for age at assessment and prevalence of infant undernutrition at one month of age
OutcomeResult by study groupComparison between monthly SP and control groupComparison between AZI-SP and control group
ControlMonthly SPAZI-SPP-valueMean difference (95% CI)P-valueMean difference (95% CI)P-value
Mean (SD) weight at 4 weeksa (g)3840 (570)3920 (580)4010 (510)<0.00170 (−10 to 150)0.07160 (80 to 240)<0.001
Mean (SD) length at 4 weeksa (cm)51.3 (2.3)51.6 (2.3)51.9 (2.1)0.0020.3 (−0.0 to 0.6)0.060.6 (0.2 to 0.9)<0.001
Mean (SD) head circumference at 4 weeksa (cm)36.9 (1.5)37.0 (1.4)37.2 (1.3)0.0010.1 (−0.1 to 0.3)0.310.4 (0.2 to 0.6)<0.001
Mean (SD) chest circumference at 4 weeksa (cm)35.3 (2.0)35.4 (2.1)35.6 (1.9)0.060.1 (−0.2 to 0.4)0.370.3 (0.1 to 0.6)0.02
Mean (SD) MUAC at 4 weeksa (cm)11.1 (0.9)11.2 (0.9)11.4 (0.9)<0.0010.1 (−0.0 to 0.2)0.080.2 (0.1 to 0.4)<0.001
 Number with outcome/infants with outcome data(%)Risk ratio (95% CI)P-valueRisk ratio (95% CI)P-value
Control Monthly SP AZI-SP P-value
  1. CI, confidence interval; SD, standard deviation; MUAC, mid upper arm circumference; LAZ, length-for-age Z-score; WAZ, weight-for-age Z-score; WLZ, weight-for-length Z-score.

  2. a

    Adjusted for age at assessment in days.

Stunted (LAZ<-2)92/377 (24.4)90/361 (24.9)57/391 (14.6)<0.0011.02 (0.79 to 1.31)0.930.60 (0.44 to 0.81)0.001
Underweight (WAZ<-2)44/376 (11.7)37/363 (10.2)22/391 (5.6)0.0070.87 (0.58 to 1.32)0.560.48 (0.29 to 0.79)0.003
Wasted (WLZ<-2)6/373 (1.6)7/361 (1.9)6/388 (1.6)0.921.21 (0.41 to 3.55)0.790.96 (0.31 to 2.95)>0.99
Figure 2.

Distribution of birthweight and head circumference at birth and weight and length at one month after delivery (adjusted for age at assessment in days) in the three study groups.

Discussion

This trial was carried out in a malaria holoendemic area in rural Malawi to compare the impact on birth outcomes and infant growth of the standard two-dose SP IPTp regimen, with a regimen consisting of monthly SP, alone or in combination with azithromycin. Compared with the controls, babies born to participants treated with monthly SP and two doses of azithromycin had a 140 g higher birthweight, about 40% lower incidence of LBW and a 40–50% lower prevalence of stunting and underweight at one month of age. Analyses of other anthropometric outcome measures yielded similar results, and the intergroup difference was observed across the entire outcome distribution. Adjusted analyses suggested that about one-third of the lower birthweight in the control group was due to shorter gestation, and two-thirds were due to IUGR. The babies born to the monthly SP group had a higher birthweight than those born to the control group. Most of the other outcome differences between these groups were to the same direction, but smaller than between the AZI-SP and control groups.

The probabilities of bias or random error were minimised by broad inclusion criteria, random group allocation, comprehensive follow-up, ultrasound-based assessment of the duration of pregnancy, partial use of placebo control, blinding of the outcome assessors and a large sample size. There were small intergroup differences at baseline in the number of previous pregnancies and malaria parasitaemia, but adjusted analyses suggested that these did not bias the results. Confidence interval calculations and hypothesis testing indicated that the probability of type I error was small. We therefore believe that the sample findings are reliable and sufficiently representative of the target population in suggesting that pregnancy outcomes and infant growth in the study area can be improved by treating pregnant women with monthly SP and two doses of azithromycin. Monthly SP alone is likely to increase mean birthweight and may have a positive effect on other outcomes. However, the higher likelihood of random error makes population inference less conclusive for this regimen.

Three earlier studies have tested the effect of an IPTp regimen containing monthly SP, but all of them enrolled only women belonging to defined risk groups (Parise et al. 1998; Filler et al. 2006; Hamer et al. 2007). The pooled results of these studies suggested that compared with the standard regimen, monthly SP results in higher birthweight among HIV-positive paucigravidae (ter Kuile et al. 2007). Our results corroborate this finding and suggest that monthly SP increases mean birthweight also when given to unscreened pregnant women, who typically comprise the actual target group for IPTp. A similar conclusion was drawn in a recent meta-analysis that used data from seven controlled trials (including ours) conducted in sub-Saharan Africa (Kayentao et al. 2013). Based on these results and other emerging evidence, WHO now recommends IPTp with SP for all pregnant women at each scheduled antenatal care visit in areas of moderate-to-high malaria transmission (WHO 2012).

The effect on birthweight of antibiotic treatment against RTI when given for unscreened pregnant women has previously been tested in two studies in Africa. A trial in Uganda found a reduced incidence of LBW proxies among participants treated with several antibiotics (Gray et al. 2001). Another study in Malawi (APPLe) recorded no significant improvement in mean birthweight or the incidence of pre-term delivery among women who received both SP and azithromycin (1 g) twice during pregnancy (van den Broek et al. 2009). The seemingly conflicting results between that trial and our study could be explained by the differences in the frequency of SP dosing (two doses in APPLe, monthly in our trial), malaria prevalence at enrolment (25% microscopy positive in APPLe, 9% in our trial) and participant characteristics (35% primigravid in APPLe, 23% in our trial) (van den Broek et al. 2009; Luntamo et al. 2010). The combination of a higher burden of malaria, less frequent dosing of SP and a high prevalence of SP resistance may have compromised the antimalarial activity of the intervention in the APPLe study (White 2005; Sridaran et al. 2010; Taylor et al. 2012). The slightly higher burden of syphilis (7% at enrolment in APPLe, 5% in our trial) together with suboptimal syphilis treatment (1 g benzathine penicillin) in the APPLe might have further diluted the effect of azithromycin in that trial (van den Broek et al. 2009; Chico & Chandramohan 2011b). Non-study antibiotics used to treat other conditions during pregnancy might have modified the effect of the study medications in both studies. We also cannot rule out an undisclosed bias in either study.

We have previously reported that both at 32 gestational weeks and at delivery, the monthly SP and AZI-SP interventions were associated with similar reductions in the prevalence of peripheral blood malaria parasitaemia when compared with the control group (Luntamo et al. 2010; Luntamo et al. 2012). This similarity in malaria prevalence between the two intervention groups could be explained by a limited antimalarial effect of azithromycin at the dose used in our trial (Chico & Chandramohan 2011b). Thus, the greater improvement of birth outcomes in the group that received azithromycin was probably due to some other activity of the antibiotic, such as its effects against RTI or other infections. However, due to limited information on specific infections among the study participants, we cannot make further conclusions on the mechanism of action of the interventions.

In conclusion, our results suggest that an IPTp regimen with monthly SP given to all pregnant women would increase mean birthweight in malaria holoendemic areas in SSA. Given the epidemiology and consequences of LBW and the feasibility of repeated administration, safety and low price of SP, an IPTp policy based on monthly SP seems warranted in such conditions. Adding azithromycin to the IPTp regimen might offer further benefits to the foetus and the neonate. However, further studies on this approach with rigorous interim outcome assessment are necessary due to the higher cost of azithromycin, the paucity of data on its mechanism of action on birth outcomes and the earlier study result that provided no evidence of positive effects on birth outcomes.

Acknowledgements

We thank the study participants and the people of Lungwena; the staff at the Lungwena Training Health Centre, Malindi and Mangochi Hospitals, and our research nurses and assistants for their positive attitude, support and help in all stages of the study; B Mbewe for supervising part of the data collection; E Molyneux, SA White and G Kafulafula for monitoring the study as the data safety and monitoring board members; J Kumwenda for making the site monitoring visit; L Csonka for designing the data entry program. This study was supported by grants from the Academy of Finland, the Foundation for Paediatric Research in Finland and the Medical Research Fund of Tampere University Hospital. Azithromycin and its placebo were provided free of charge by Pfizer Inc (New York), which also provided funding for the PCR testing of the sexually transmitted infections. ML is supported by a grant from the Finnish Cultural Foundation and YBC is supported by the Singapore Ministry of Health's National Medical Research Council under its Clinician Scientist Award.

Ancillary