Intermittent Preventive Treatment of Malaria in Pregnancy with Mefloquine in HIV-Infected Women Receiving Cotrimoxazole Prophylaxis: A Multicenter Randomized Placebo-Controlled Trial

Clara Menéndez and colleagues conducted a randomized controlled trial among HIV-positive pregnant women in Kenya, Mozambique, and Tanzania to investigate the safety and efficacy of mefloquine as intermittent preventative therapy for malaria in women receiving cotrimoxazole prophylaxis and long-lasting insecticide treated nets. Please see later in the article for the Editors' Summary


Introduction
This Analysis Plan (AP) provides detailed description of the descriptive (with the skeleton of the tables) and inferential statistical analysis for the efficacy and safety data collected in the Trial 2 of the MiPPAD study to compare the IPT p -MQ versus IPT p -Placebo in HIV infected women receiving Cotrimoxazol (CTX) and Long lasting insectide treated nets (LLITNs). A Separate AP is available for the analysis of exploratory endpoints and the analysis of children, Any planned modification of the analysis should be documented in this AP.

Primary
To determine the safety and efficacy of IPTp with mefloquine among HIV infected women receiving CTX prophylaxis for opportunistic infections

Secondary
• To evaluate the efficacy of CTX in the prevention of malaria infection in pregnant women.
• To compare immune status of HIV infected women receiving CTX + IPTp-MQ to those receiving CTX + IPTp-placebo.
• To assess the safety of study drugs in the development of infants 3 Endpoints

Primary
• Prevalence of peripheral parasitaemia (microscopic and submicroscopic) at delivery

Secondary
• Prevalence of P. falciparum parasitaemia in cord blood (microscopic and submicroscopic 1 ) • Prevalence of placental P. falciparum infection (histology or blood smear, submicroscopic 1 ) • Mean maternal haemoglobin (g/dL) at delivery • Prevalence of maternal anaemia at delivery (<11 g/dL) • Prevalence of severe maternal anaemia at delivery (<7 g/dL) • Mean of CD4 counts and viral load at delivery • Prevalence of neonatal anemia (Hb<12.5 g/dL in cas e of cord blood, or Hb<13g/dL in case of peripheral blood) • Mean birth weight (in grams) 1 Only in a subsample of participants • Prevalence of low birth weight babies (<2500 g) • Prevalence of prematurity

Study Design Overview
This is a randomized double-blind superiority clinical trial to compare the efficacy of MQ as IPTp (IPT p -MQ ) with placebo-IPTp (IPT p -Placebo ) in HIV-infected pregnant women receiving CTX prophylaxis.The table 1 shows the visits and procedures schedule to the mother and the table 2 shows the visits and procedures schedule to the child.

Population for analysis
The assignment of participants into analysis populations will be performed prior to database lock and any data analysis. The following populations are defined:

According to protocol (ATP)
This population includes all women who fulfill all the inclusion-exclusion criteria and took the three IPT doses, received the LLITN, received prophylaxis with CTX and from whom data is available for the analysis. Subjects will be excluded from the ATP population in case of: • No fulfill inclusion criteria • No data on outcome

• No three doses of IPTp
• Wrong study drug administered • Do not receive LLITN • Less than 4 weeks (28 days) between IPTp doses

Intention to treat (ITT)
This population includes all randomized women. Following the intention-to-treat principle, patients will be analyzed according to the preventive treatment they were assigned to at randomization. This population is the target population for the efficacy analysis

Safety
This population includes all patients who received at least one dose of IPT and had at least one postbaseline safety assessment. Patients will be analyzed according to preventive treatment assigned. This is the target population for the safety and tolerability analysis 6 Statistical methods The primary analysis of this trial is the comparison of the proportion of mothers with peripheral parasitaemia at delivery in the ATP cohort, adjusted by gravity, country, seasonality and other variables associated with the prevalence of peripheral parasitemia at delivery.
Proportions are compared between group using fisher exact test and presented as relative risk ratio (RR) or reduction of the RR (1 − RR * 100%) if RR lower than 1. Adjustment for co-variates and possible confounder are done using poisson regression with a log link and robust estimate of the covariance (Huber method), using the method proposed by Zou (A modified poisson regression approach to prospective studies with binary data, American Journal of Epidemiology, 2004:159(7): 702-706) Continuous variables are compared between groups using Wilcoxon rank sum test and the effect presented as Mean Difference. Adjustment for co variates and possible con founders are done using ordinary least square regression. Variables will be transformed to the logarithm scale if normality is improved and result presented as Proportional Difference. If after transformation, non-normality of the residuals is detected using diagnostic regression plots (q-q plots of the residuals, and plot of the residuals against the predicted values), robust intervals of confidence and Wald test will be resented instead.
Incidence of clinical malaria, overall admissions and outpatient attendances will be estimated as the number of episodes over the time at risk. Time at risk is estimated as the time from the start of follow up (dose 1 in mothers) until the end of follow-up (visit one month after deliver for mothers) or withdrawal due to censoring or death, whatever occurs first. In order to avoid to count twice the same clinical malaria episode, subjects will not contribute to the denominator nor the numerator during an arbitrary period of 28 days after episode of clinical malaria is defined. A maximum of one episode of admission or outpatient visit will be count per day. The total number of events will be compare between groups using Negative Binomial regression models which take into account a possible extra Poisson variation due to different frailty of the subjects. The comparison will be expressed as relative rate ratio (RRate).
To calculate the compliance with CTX medication, the number of tables taken (from number of tables given minus the number of tablets returned) will be related to the number of days under chemoprophylaxis (from date of deliver -date of first dose of CTX given). A Mother will be considered compliant if more than 80% of the expected medication is taken.
7 Skeleton of tables and graphs 7.1 Trial profile (ATP,ITT,Safety)