High uptake of Intermittent Preventive Treatment of malaria in pregnancy is associated with improved birth weight among pregnant women in Ghana

Despite the clinically proven advantages of intermittent preventive treatment of malaria in pregnancy (IPTp) with sulfadoxine-pyrimethamine (SP), utilisation has been low in many African countries. To increase uptake and achieve the desired effect, the World Health Organization revised the policy to a monthly administration. Assessing the coverage and impact of the revised policy on pregnancy and neonatal outcomes is, therefore, a necessity. A 2-parallel cross-sectional hospital-based study was carried out among pregnant women attending first antenatal care (ANC) and delivery. Maternal and cord blood samples were assayed for malaria parasites by quantitative PCR targeting both the 18S rDNA and the acidic terminal segment of Plasmodium falciparum var genes, and plasma SP levels were measured by liquid chromatography coupled to tandem mass spectrometry. Parasite prevalence was similar between the two study sites but decreased significantly between the first ANC (9% or 43%) and delivery (4% or 11%) based on the qPCR target. At delivery, 64.5% of women received ≥3 IPTp-SP dose, 15.5% received 2 doses and 6% had 1 dose. Taking ≥3 IPTp-SP doses was associated with an average birth weight increase of more than 0.165 kg. IPTp-SP uptake was associated with plasma SP level at delivery (OR = 32.3, p ≤ 0.005, 95% CI (13.3;78.4) for those that reported ≥3 IPTp-SP doses) while the same trend of improved birth weight was observed with high plasma SP levels. The new IPTp policy is well implemented and well utilised by women in the sites considered in this study and translates to the improved birth weight observed. This study confirms the interest and the clinical benefit expected from this policy change.

Coverage and impact of IPTp in the delivery cohorts. At delivery, information on IPTp-SP uptake was taken from the participants ANC booklets as a reliable source because of the practice of directly observed treatment. Among the 935 women recruited at delivery, 603 (64.5%) had taken ≥3 IPTp-SP doses, 143 (15.3%) had received 2 IPTp-SP doses while 53 (5.7%) had taken 1 IPTp-SP dose. One hundred and thirty-six (14.5%) did not receive any IPTp-SP. When the data were stratified by site, of the 635 participants at Maamobi, 365 (57.5%) had taken ≥3 doses, 106 (16.7%) had taken 2 IPTp-SP doses and 39 (6.1%) had taken 1 IPTp-SP dose. One hundred and twenty-five (125) participants did not receive any IPTp-SP, and among them 82 (65.6%) were disqualified at the hospital due to glucose-6-phosphate dehydrogenase (G6PD) deficiency detected by a nitroblue tetrazolium test ( Fig. 1).
Women who took ≥3 IPTp-SP doses had a lower risk of P. falciparum infection detected by qPCR, compared to those who received only 1 or 2 doses, (OR = 0.72, p = 0.36 at Maamobi; OR = 0.69, p = 0.59 at Kpone on-Sea,) ( Table 4). In contrast, the overall linear regression analysis showed that uptake of ≥3 IPTp-SP doses increase the average birth weight by more than 0.165 Kg. This increase was only statistically significant at Maamobi (P < 0.005) although the same trend was visible at Kpone on-Sea (Table 5). This effect was lower and not significant in women taking only 1 or 2 IPTp-SP doses although an increase in birth weight of about 0.72 Kg was observed as compared with women who did not receive any IPTp-SP treatment.
Sulfadoxine-pyrimethamine concentration measurement. To assess and confirm uptake of SP by pregnant women, we measured SP residual levels among a sub-group of participants from both the ANC and delivery cohorts. For ANC, this was to assess whether pregnant women attending their first ANC at the study sites had taken any SP. In the first ANC group, we measured SP residual levels among 353 and 408 plasma samples from Maamobi and Kpone on-Sea, respectively. Of the samples analysed at Maamobi, sulfadoxine was detected in 99 (28%) of them, with median concentration of 3777 mg/ml and interquartile range (IQR: 263-38335), while we detected pyrimethamine in 54 women (15.3%) and median concentration of 224 ng/ml (IQR: 18.75-410.3). At Kpone on-Sea, sulfadoxine and pyrimethamine were detected in 10 (2.5%) plasma samples at median concentration of 2160 mg/ml (IQR: 256.8-4378) and in 1 (0.2%) plasma sample at a concentration of 216 ng/ml, respectively. Women in whom we detected pyrimethamine were also sulfadoxine-positive, however sulfadoxine was detected in more women. So, we considered the presence of sulfadoxine as a proxy for SP in the analyses.
In the delivery group, we measured SP residual levels in peripheral and also in cord blood to explore whether the amount of SP observed in the mother's peripheral blood could be detected in the cord compartment.  www.nature.com/scientificreports www.nature.com/scientificreports/ pyrimethamine residual levels among 77 (59.2%) peripheral plasma at median concentration of 27 ng/ml (IQR: 13.3-80.5) and 55 (42.3%) cord samples at median concentration of 50 ng/ml (IQR: 20-120).
SP level in both maternal peripheral and cord plasma was significantly associated with IPTp-SP uptake ( Table 6). When the plasma SP levels in the maternal peripheral blood at delivery were categorized into quartiles,   Table 3. The concordance and discordance between infections detected targeting the 18 SqPCR and the ATS qPCR.   www.nature.com/scientificreports www.nature.com/scientificreports/ the same pattern of improvement in birth weight was observed for women with high concentrations of SP in their blood at delivery. When the data were stratified based on site, this significant association was only observed at Maamobi (Data not shown).

Discussion
Despite the recommendation and implementation of IPTp-SP in some countries in tropical Africa some decades ago, pregnant women are still affected by malaria disease and its effect on their neonates. This is partly due to the low utilisation of the IPTp policy irrespective of high coverage of ANC visits 5,19 . To address this, WHO revised the policy to a monthly administration of SP to reduce inadvertent low or non-uptake of SP 18 . Thus, determining the coverage and effectiveness of this new policy is of utmost importance to its sustainability or otherwise across the different SP-resistant areas in tropical Africa.
In this study, we enrolled two different sets of pregnant women; those who were visiting the health centres for their first antenatal consultation and those coming to deliver at the health centres. The design of this study makes it possible to better capture the characteristics of women appearing at their first ANC and also to determine the use of the services set up for malaria prevention in the delivery cohort. The study design also makes it possible to limit the introduction of potential biases on the envisaged observations. The first observation is the early arrival of women at first ANC, at 18 weeks of pregnancy on average at both study sites, the earliest of which is at 11 weeks. This observation is important and very reassuring for the policy change. Our data which shows over 60% of pregnant women taking ≥3 IPTp-SP doses, together with other recent studies in Ghana that report between 87 and 99% IPTp-SP uptake of ≥3 IPTp-SP doses 20,21 indicate a high IPTp coverage across Ghana. The observed satisfactory coverage indeed results from the effective campaigns by the Ghana Health Service (GHS) for the IPTp-SP policy. Measuring SP molecules in the blood of women under IPTp corroborated the documented information on treatment received by the participants.
A high proportion of women (more than 75%) with detectable SP in both peripheral and cord plasma at delivery is also indicative of a high IPTp-SP coverage in Ghana as expected by the new policy. Further, this clearly shows that SP can cross the placenta barrier and raises the question whether these high doses of SP that correlated with that of the mother's compartment are appropriate for the condition of the fetus. However, future studies on the impact of prolonged use of SP in IPTp on the health of the fetus are needed. This is very interesting as this biological measurement correlated with the number of SP doses recorded when the measurement was adjusted for ANC attendance. However, we detected SP in an unexpected and relatively high proportion of women (28%) visiting the urban site for their first ANC, which falls out of line with their declaration. As we did not find this in the peri-urban site; we cannot exclude the hypothesis of possible unreliable disclosure due to self-medication or a first undeclared visit in a private clinic. Our possibilities of verifying what these women in this group disclosed before their first ANC attendance at the study health centres are minimal. Thus, the use of this analytical tool could lead to definition of a follow-up marker for SP uptake.  Table 6. Relationship between IPTp-SP uptake and residual SP a level measured in the blood at delivery. a sulfadoxine as a proxy for SP categorised into binary variables (either high or low SP level) base on the median; b Odds-Ratio, *Participants that reported to have received 1 or 2 SP doses, #Participants that reported to have received ≥3 SP doses. The reference classes are the multigravidae, the women who reported no dose of SP taken, and Maamobi site.

Coefficients (Kg) 95% confidence interval p-value
The significant finding of this study strongly confirms increased birth weight associated with increased SP uptake as recommended by the revised policy. In this case uptake of ≥3 IPTp-SP doses improves birth weight by more than 0.165 Kg. P. falciparum infection (either microscopically or sub-microscopically) during pregnancy is one of the major factors underlying low birth weight, and several studies have shown the beneficial effect of IPTp-SP on the birth weight with some indication of a dose-dependent relationship with birth weight 13,16,22 . The low prevalence rates of infections observed at delivery compared with those amongst the first ANC group is an indication of IPTp-SPs' curative and prophylactic effect as a whole. However, the fact that taking ≥3 doses of IPTp-SP did not show a particular impact on the infection rate at delivery probably suggests that this variable does not provide information on all infections that would have been prevented since initiation of IPTp. This variable can only measure infections observed at delivery which may include more recent infections, while the indirect effect of IPTp that was visible on the improvement of birth weight would better reflect a broader window of coverage.
Sulfadoxine like other sulfonamides is a broad-spectrum antibiotic that is known to likely exert an inhibitory effect against potential non-malarial infections such as sexually transmitted and reproductive tract infections that can cause low birth weight and preterm birth. Thus, SP uptake can sufficiently curb these infections and indirectly improve birth weight as observed elsewhere 22 . This could partially explain why in areas where SP efficacy has reduced, SP continued to improve birth weight outcomes. The effect size is similar in both sites but probably due to lack of power, statistical significance was only achieved at the Maamobi site. A possible explanation may lie in the power of analysis facilitated by a larger sample size of participants that reported of not receiving any IPTp-SP at Maamobi due to G6PD status compared to Kpone on-Sea where virtually every participant received SP, or other possible factors such as maternal nutrition, malaria history or other infections that may be specific to the site but which we are unaware of.
In our study populations the prevalence of microscopically detectable Plasmodium infections was low (less than 4%), however, the use of qPCR revealed the huge difference in the proportion of women who were parasite-positive with microscopy and PCR. The high sensitivity of the 18 S qPCR (detection limits of 0.5-3 parasites/µl of blood) and ATS-qPCR (0.06-0.15 parasites/µl of blood) 23 has been an asset for optimal detection of these infections that are missed by conventional microscopy with a detection limit of only 50 parasites/µl of blood 24 . The reasons for low placental infections compared to peripheral infections are not known, it is possible that all the peripheral infections detected are not necessarily of placental type. Although densities of infection in the placenta are generally higher compared to the peripheral blood, the asymptomatic carriage of certain non-placental type peripheral infections cannot be ruled out. The use of a more sensitive detection PCR in this study, such as ATSqPCR which not only numerically recognizes more targets on the parasite genome than do the 18SqPCR, but also the fact that it was expended over 45 cycles of amplification, allowed to detect these populations probably more prevalent at lower densities.The study demonstrates that low-density parasitemia is very common in high malaria-endemic areas.
To conclude, the new IPTp policy is well implemented and utilised by pregnant women in our study sites confirmed by plasma SP level measurements. Our findings also showed that ≥3 SP administrations of IPTp were associated with increased birth weight as compared to the previous standard 2-dose regimen. Also, as in any observational field study, we did not control for all possible confounding factors. At the same time, this study also reveals that despite the reduction of adverse effects on birth weight by administering of more doses of IPTp-SP, malaria infections remain a threat outside the window of application of IPTp, mainly before women present to antenatal care. Therefore, the need to complement this strategy with an approach that can protect women from infections in the first trimester of pregnancy, such as the use of a vaccine, remains a necessity.

Methods
Study population and design. The study was a cross-sectional hospital-based survey conducted in two distinct communities in southern Ghana; Kpone on-Sea, a semi-urban community and Maamobi, an urban community about 50 Km apart. Malaria transmission in both communities is perennial, with two peaks in April-July and September-November. However, the entomological inoculation rate (EIR) at Kpone on-Sea is 62.1 25 while that of Maamobi is 44.7 26 . At Kpone on-Sea, primary health care is mainly provided by a health centre that also serves surrounding communities. In Maamobi, primary health care service is provided by the Maamobi General Hospital located in the heart of the Maamobi township and also serves several communities in its vicinity. In each site, two independent, HIV-naïve, groups of pregnant women were enrolled in cross-sectional surveys; one survey targeted pregnant women attending their first ANC to get an overview of the appearance of women at first ANC and their characteristics, while the other targeted pregnant women attending the maternity wards for delivery, and was used to study the IPTp coverage and its impact.
Ethics statement. All experimental protocols were approved by the Institutional Review Board (IRB) of the Noguchi Memorial Institute for Medical Research (NMIMR) and the Ethics Review Committee of the Ghana Health Service (GHS). Written informed consent was obtained from each participant before they were admitted into the study. All methods were carried out in accordance with the relevant guidelines and regulations.
Data and biological sample collection. Standardised questionnaires and checklists were used to collect demographic, clinical and ultrasound data for the cohort at the first ANC. As standard practice, pregnant women attending ANC in these two centres are given IPTp-SP under direct observation, and the health staff make entries into their booklets after this directly observed therapy (DOT). Three (3) ml of peripheral blood were collected into EDTA vacutainer tubes before administration of IPTp-SP. Malaria screening by light microscopy and haemoglobin (Hb) level determinations by Hemocue ® were performed.
For the cohort at delivery, standardised questionnaires and checklists were also used to collect demographic and clinical data. Information on IPTp-SP uptake was obtained from participants' ANC booklets. Three (3) ml of peripheral blood and 3 ml of cord blood were collected into separate EDTA vacutainer tubes, and peripheral Antimalarial drug level measurements. Plasma sulfadoxine and pyrimethamine levels were measured by liquid chromatography coupled to tandem mass spectrometry (TSQ Quantum Ultra, Thermo Fisher, France) with electrospray ionisation in positive mode using an Atlantis T3 (100 mm × 2.1 mm, 3 µm) column (Waters, France). Separation was assessed using a 0.30 ml/min flow rate and a gradient mode using two solvent phases, A (water/methanol/formic acid, 95/4.9/0.1% v/v) and B (methanol/formic acid 99.9/0.1% v/v). Specific multiple reaction monitoring (MRM) transitions were selected for detecting sulfadoxine (311.10/245.10) and pyrimethamine (249.10/233.10). Quinide-d3 (328.09/163.04) was used as an internal standard. Separation was achieved in less than 10 min. Using 96-well micro plates (OASE Waters, France), 100 µl of plasma were mixed with 300 µl of acetonitrile containing quinide-d3 (50 ng/ml) as internal standard proteins and phospholipids were eliminated by positive pressure (20 psi during 1 min.). Eluents were evaporated at room temperature, and dry residues were dissolved in 20 mM ammonium formate buffer with formic acid (0.5% v/v) before 10 µl was injected into the system. Method was linear between 10 and 10000 ng/ml for sulfadoxine and between 10 and 1000 ng/ml for pyrimethamine. Lower limit of quantification was 10 ng/ml for both sulfadoxine and pyrimethamine. Upper limits of quantification were 25000 and 5000 ng/ml for sulfadoxine and pyrimethamine, respectively.
For homemade and external controls (WWARN controls), coefficients of variation were below 10%, and bias values were ±10%. Appropriate dilutions were performed in blank serum for all samples and concentrations higher than 10,000 ng/ml for sulfadoxine or higher than 1,000 ng/ml for pyrimethamine and re-analysed. Statistical analysis. At delivery, the primary exposure variable of interest was IPTp-SP usage, and it was grouped into 3 categories; women who did not receive any SP, women who received 1 or 2 doses, and women who received ≥3 doses of SP. We analyzed the beneficial effect of SP on birth weight based on the number of SP intakes using a multivariate linear regression model. The risk of P. falciparum infection during pregnancy both at first ANC and at delivery was analyzed using a multivariate logistic regression model. To evaluate the association between SP uptake and residual SP level, we performed a logistic regression after categorizing SP into a binary variable based on the median level. In these analyses, the adjustment factors were gravidity (primigravidae or multigravidae), infection status (positive or negative by PCR), gestational age and SP intake information from the ANC booklets. The Cohen Kappa test was used to test for concordance between the 18SqPCR and the ATSqPCR methods used in P. falciparum detection. All statistical analyses were done with Stata(version 13.0). All tests were conducted using 0.05 level of significance.