Overall study design
The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1
and Protocol S1
. It was an individually randomised, open, controlled trial was undertaken to investigate whether screening for malaria with an RDT and treatment of women with a positive test with either SP or AS+AQ was not inferior to SP-IPTp in the prevention of anaemia in pregnancy or low birth weight.
The study was conducted in the Ejisu-Juaben and Afigya-Sekyere East districts of the Ashanti Region of Ghana from March 2007 to September 2008. Malaria transmission in this area is perennial but with a peak in the rainy season. The predominant parasite is P. falciparum. The entomological inoculation rate in the neighbouring area of Kintampo is about 250 infectious bites per year. HIV prevalence in Ashanti region is reported to be 3.0% and 2.2% in the general population and in pregnant women respectively (Ghana Health Service 2007 Annual Report).
Enrolment was undertaken at the antenatal clinic of three district hospitals and three health centres in the study area. The study population comprised pregnant women of all parities who presented at the antenatal clinics with a gestational age of 16 to 24 weeks at their first booking. Women who were temporary residents, had had a prior dose of SP-IPTp, had a haemoglobin concentration <5 g/dl, gave a history of sensitivity to SP, amodiaquine or an artemisinin, had an illness requiring hospital admission or declined to join the trial were excluded ().
CONSORT chart showing enrolment and follow-up status of women in each study arm.
The study protocol was approved by the ethics committee of the London School of Hygiene and Tropical Medicine and the Committee on Human Research and Publications Ethics of the School of Medical Sciences, Kumasi – Ghana. A Data, Safety and Monitoring Board (DSMB) approved the protocol, standard operating procedures and analysis plan.
Sample size was calculated on the assumption that the prevalence of moderately severe anaemia (Hb<8 g/dl) in the third trimester of pregnancy and of low birth weight in women in the SP-IPTp arm of the study would be at least 12% and 6% respectively, figures based on findings from a study undertaken in The Gambia 
. To establish that IST was not inferior to SP-IPTp we decided that it was necessary to show that the differences in the percentages of women with moderately severe third trimester anaemia or low birth weight between the IST groups and the SP-IPTp group should not be more than 5% or 4% respectively, differences that would be clinically important. To meet these criteria with 90% statistical power and allowing for 20% loss to follow-up it was calculated that 1,110 women were needed in each study arm to give a total sample size of 3,330.
After written informed consent had been obtained, eligible women of all parities were randomised to one of the three treatment groups described above. Women who declined to participate were given SP-IPTp according to the national guidelines. The following process was followed to randomise women into the study arms. A list of random numbers was computer generated as identification numbers, randomly allocated to treatment groups and grouped in blocks of 15 by an IT specialist who did not participate further in the study. The list of identification numbers with their corresponding treatment groups was printed and cut into slips. Fifteen slips each with an identification number and an allocated treatment group were sealed in an envelope. During enrolment an eligible pregnant woman was asked by the recruitment team to pick a slip from the sealed envelope. The allocated treatment group printed on each slip indicated which treatment arm the women belonged to and the corresponding identification number was used to identify her. This constituted entering the study, and treatment group allocation was binding on the recruiting team and the woman from this point. Another envelope was opened only when the contents of the previous one had been exhausted. The PI and all other project staff were blinded to the randomisation process and treatment allocation.
At enrolment, a finger prick blood sample was obtained for measurement of haemoglobin concentration, preparation of thin and thick blood films for malaria parasite counts and preparation of a filter paper blood spot. During the last 3 months of the trial, women in treatment group 1 who had a positive blood film and received SP-IPTp were seen again on day 14 and day 28 after treatment and a further blood film and blood spot obtained to check on the response to treatment with SP. HIV screening was offered to all pregnant women study women as part of the routine antenatal services recommended in Ghana with an option for treatment but the results of HIV screening were not available to the investigators. Women in the SP-IPTp arm received an initial dose of SP (1500 mg sulphadoxine/75 mg pyrimethamine) as a single dose. Pregnant women in treatment groups 2 and 3 were screened for malaria infection with an OptiMAL® dipstick, a lactate dehydrogenase (LDH) based RDT. The OptiMAL® dipsticks were purchased from DiaMed AG, Cressier, Switzerland who supplied and organised transportation to the study site in batches as and when needed,. The test kits were kept at room temperature always at the study site. The viability of the kits were tested monthly throughout the study period using positive test controls obtained from the manufacturer. The tests were performed and interpreted by the study team following the manufacturer's instructions. Women in treatment group 2 were treated with a single dose of SP (1500 mg sulphadoxine/75 mg pyrimethamine) if the RDT was positive. Women in treatment group 3 were treated with AQ+AS (AQ-300 mg + AS-100 mg twice daily for 3 days) if the RDT was positive. All treatments with SP and the first dose of AQ+AS treatment were administered by a member of the study team but doses 2 and 3 of AQ+AS were given to study women to take at home unsupervised. Women in groups 2 and 3 received no antimalarial treatment or IPT if their RDT results were negative. All women received a daily supplement of ferrous sulphate (200 mg) and folic acid (4 mg) tablets as part of the routine antenatal care services provided in Ghana throughout the study period. SP was purchased from Kinapharma Ltd, Ghana and AQ+AS was provided by the Ministry of Health, Ghana. All study women received a long-lasting insecticide treated bed net (LLIN) and instructions on how to use this at enrolment.
Study women were asked to attend for follow-up antenatal care and IPTp or screening at 24, 32 and 36 weeks of gestation. At the 24 and 32 week visits, women in treatment group 1 received SP-IPTp whilst women in treatment groups 2 and 3 were screened with the RDT and, if positive, treated with SP or AS+AQ according to the allocation arm. As part of safety assessment, all women were visited at home by a trained community health worker to record any complaints that the women might have seven days after each scheduled antenatal visit.
Blood samples were obtained from all study women between 36 and 40 weeks of gestation (before delivery) for determination of haemoglobin concentration and for preparation of thin and thick blood films. However, the smears were read retrospectively and so the results were not available for the point of care. Any study woman who presented with a history of fever or other features suggestive of malaria between scheduled antenatal visits was screened for malaria using an RDT and treated with quinine (30 mg/kg in divided doses daily for 5 days) regardless of treatment group.
Information on the outcome of pregnancy was obtained for 2706 of the 3333 study women (81.2%), 2144 of whom delivered at a health centre or hospital and 562 at home. If a woman delivered at a health facility, birth weight was recorded by a midwife who was unaware of the treatment group of the woman who she was attending. Women who delivered at home were traced through a network of trained community health workers within 72 hours of delivery and the infant's weight was measured at home. The occurrence of miscarriages, still births, neonatal deaths and the presence of congenital abnormalities was recorded by midwives. If any congenital abnormality was suspected, a full examination, including a neurological assessment, was undertaken by a qualified medical doctor. Field teams visited all women and babies at approximately 6 weeks post delivery to obtain reports of any neonatal adverse events and a blood film was obtained from mothers at this time. We could not determine the prevalence of placenta malaria and efficacy of SP in postpartum study women as proposed due respectively to inadequate funding and difficulty in finding postpartum women with parasitaemia.
An experienced microscopist, unaware of treatment group assignment, read all the blood films and quantified parasitaemia against 200 leucocytes in thick blood smears. A thick blood film was declared negative only after examination of 100 high power fields (HPF). An independent expert microscopist from the Noguchi Memorial Institute of Medical Research read ten percent of all blood slides obtained at enrolment and on follow up visits for quality assurance. Agreement between the study microscopist and the reference microscopist was 91.7% and a kappa of 0.83. Haemoglobin (Hb) concentration was measured using Hb 301 Hemocues (HemoCue AB, Angelholm, Sweden). Paired samples from women who failed treatment with SP were tested for molecular markers to differentiate reinfections from recrudescences using genetic markers as described previously 
. HIV testing was not part of the study but it is likely that some HIV positive may have been included in the sample.
Stata version 10 (StataCorp, College Station, Texas) was used for data analyses. The primary objective of the study was to demonstrate that the risk of third trimester moderately severe anaemia (Hb<8 g/dl) in the IST groups was no more than 5% greater than in the SP-IPTp arm. Secondary objectives were to demonstrate that the risks of low birth weight (BW<2500 g), spontaneous abortions, intrauterine deaths/stillbirths, neonatal and maternal mortality were not more than 4% higher in women in the IST groups than in women who received SP-IPTp. The principal analysis of primary and secondary outcomes was per protocol but an "intention-to-treat" analysis was also undertaken according to a statistical analysis plan approved by the Data and Safety Monitoring Board. In the per protocol analysis, only data from women who remained within their randomization group, received two courses of SP-IPTp (Group 1) or were screened twice using an RDT at scheduled visits (groups 2 and 3) and in whom the primary outcome had been recorded were considered for analysis. In the intent-to-treat analysis, women were included if they had received an initial treatment of IPTp or had had an initial screening test done and provided that an outcome had been recorded.
The proportion of the per protocol and intention-to-treat populations experiencing each primary and secondary outcome for the treatment groups, and the associated 2-sided 95% CI for the difference, was estimated using the generalized linear model. To declare non-inferiority with a significance level of 0.05%, the upper boundary of the 2-sided 95% CI for the estimated treatment effect (risk difference) had to be below the pre-defined non-inferiority margins (Δ) of 5% and 4% for third trimester severe anaemia and low birth weight respectively. We controlled for gestational age at enrolment, gravidity, baseline parasitaemia and anaemia using binomial regression.