Results
A total of 10,364 articles were identified initially. After removing 10324 articles, of which 4284 were duplicates, 4191 articles were screened out through title review and 1849 through abstract review, 30 articles were still for further consideration. After excluding 25 studies, the full-text articles of five studies were included in final synthesis (the reasons for their exclusion were given in Figure 1).
Five included RCTs were conducted in 2019, 2018, 2016, 2016, and 2015 respectively and in areas of high malaria-transmission intensity. A total of 4660 HIV-seronegative pregnant were included in this meta-analysis, 2628 in DP groups, 2032 in SP groups (Table 1 presented the characteristics of the included articles). All participants received long-lasting insecticide-treated nets (LLINs) at enrolment. Each dose of DP consists of 40 mg of dihydroartemisinin and 320 mg of piperaquine which was given with a standard 3-d course. SP was composed by 500 mg of sulfadoxine and 25 mg of pyrimethamine per dose. In all five studies, Mwayiwawo Madanitsa22 compared IPTp-SP with ISTp-DP, where study participants made scheduled antenatal visits every 4–6 weeks (IPT-SP-4-6w). Abel Kakuru, M.D.14 and Prasanna Jagannathan23 compared IPTp-SP with two IPTp-DP regimens: one was monthly IPTp-DP regimen (IPT-DP-4w) from 16 or 20 GA, another IPTp-DP group received DP at 20, 28, and 36 GA (IPT-DP-8w). Richard Kajubi24 compared IPTp-SP with IPTp-DP, every 4 weeks starting at 16 or 20 GA(IPT-DP-4w and IPT-SP-4w). Meghna Desai13 compared IPTp-SP with IPTp-SP and ISTp-DP. The study participants were enrolled at 16–32 GA and IPTp groups received the interventions at intervals of 4–6 weeks (IPT-DP-4-6w and IPT-SP-4-6w). In subgroup analysis, we treated 4w as 4-6w when comparing IPT-DP with IPT-SP.
Risk of bias based on the Cochrane Collaboration tool were presented in figure 2 and figure 3. All the five studies were randomized controlled trials and had similar group characteristics at baseline13, 14, 22-24. Of the five included studies, four had random sequence generation, allocation concealment and selective reporting 13, 22-24. Two mentioned the blinding of outcome assessment13, 22 and the blinding of participants and personnel23, 24. Only one study had the unclear risk of selective reporting22. No studies had incomplete outcome data and other bias.