Ethics statement

This study was approved by the Ethics Committee from the University of São Paulo (Plataforma Brasil, CAAE: 03930812.8.0000.5467 and 90474318.4.0000.5467), according to resolution n° 466/12 of the Brazilian National Health Committee and in compliance with the Declaration of Helsinki. All subjects or their legal guardians (if minors) gave written informed consent. This study followed a specified analysis plan (S1 Text) and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement guidelines (S1 STROBE Checklist).

Setting and study design

We have conducted an observational cohort study in five municipalities (Cruzeiro do Sul, Mâncio Lima, Rodrigues Alves, Marechal Thaumaturgo, and Porto Walter) in the Juruá Valley, located in the Brazilian Amazon southwest region (Acre State, Brazil). This region is considered highly endemic for malaria, with an annual parasite incidence (API) above 100 cases per 1,000 inhabitants. Juruá Valley is also characterized by a significant prevalence of P. vivax infections, responsible for 70–80% of total malaria cases [27,28]. A complete region characterization with detailed procedures and methods used for data and sample collection has been described elsewhere [11,13,21]. Briefly, 600 pregnant women were recruited during their first antenatal care (ANC) visit. These women were followed until delivery, between January 2013 and April 2015. Data was collected on socioeconomic, clinical, and obstetric variables at the time of recruitment, together with peripheral blood samples. In parallel, thick and thin peripheral blood smears were performed to screen for malaria parasites by microscopy, which were later confirmed by PCR. Each pregnant woman was followed by a trained nurse, which involved at least two domiciliary visits, at the second and third trimester, to monitor their clinical state and collect a peripheral blood sample, in addition to the usual prenatal care visit to the health care services. An additional blood sample was collected in each malaria episode during pregnancy. At the time of delivery, clinical data were collected from mother and newborn and performed a placental biopsy and blood sampling. Only women with P. vivax monoinfections and their non-infected counterparts were considered for this study (Fig 1A). We evaluated mother-child pairs at delivery in the Women and Children’s Hospital of Juruá Valley located in Cruzeiro do Sul.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Description of study participants and definitions.

(A) Flow chart detailing exclusion criteria applied prior data analysis; (B) Definitions of pregnancy partitions used in the report. Mixed infection–P. vivax and P. falciparum infections occurring at the same or different times during pregnancy; w—weeks; d—days.

https://doi.org/10.1371/journal.pntd.0009390.g001

Malaria screening and treatment

Malaria during pregnancy was first diagnosed from thin and thick blood smears by microscopy analysis, which was performed by the endemic surveillance team of Juruá Valley. Upon diagnosis, all women infected with P. vivax during pregnancy were treated with chloroquine (25mg/kg of body weight) in divided doses over a 3-day period. Successful treatment was confirmed subsequently. No hypnozontocidal treatment with primaquine was performed, following the Brazilian Ministry of Health (MoH) guidelines regarding medical prescription in pregnant women [29]. Of note, the Intermittent Preventive Treatment in pregnancy (IPTp) is not employed in Brazil. All samples collected throughout pregnancy (recruitment, malaria episodes, and domiciliary visits) and at delivery were further screened by molecular diagnosis using the real-time PCR technique (PET-PCR) as described elsewhere [13,30].

Exclusion criteria

Pregnant women were excluded from further analysis if they self-informed any alcohol consumption, smoking, and/or if presenting any other diagnosed infectious diseases (e.g. Toxoplasmosis, Rubella, Cytomegalovirus and Herpes simplex (TORCH), HIV, Hepatitis B and C virus, Syphilis), and/or other comorbidities (e.g. hypertension, pre-eclampsia/eclampsia, diabetes mellitus, and newborn with congenital malformation). Likewise, pregnant women infected with other Plasmodium species (P. falciparum or mixed infections) were excluded after confirmation by PET-PCR (Fig 1A).

Laboratory procedures

Peripheral and placental blood (from the maternal side) was collected aseptically in heparinized tubes and separated into plasma and whole blood cells after centrifugation. All samples were stored at -80°C. Total DNA was obtained from whole blood cells using a commercially available extraction kit (QIAmp DNA Mini Kit, Qiagen), following the manufacturer’s instructions. Fragments of placental tissue were collected and stored in 10% neutral buffered formalin at 4°C, until processing. Placental tissue was sectioned and embedded in paraffin using standard procedures; afterward, it was stained with Hematoxylin-Eosin (H&E) or Giemsa for histological examination. All histological evaluation methods were optimized by our group and are described elsewhere [11,13]. Placental images were captured using a Zeiss Axio Imager M2 light microscope equipped with a Zeiss Axio Cam HRc camera. Placental parameters were evaluated and analyzed using Image J software (http://imagej.nih.gov/ij). All measurements were performed by two investigators blinded to group and outcome. Cases that proved to be contradictory between observers were re-evaluated until consensus was reached.

Placental weight and newborn’s anthropometric measurements

Placental weight was measured (grams (g)) immediately after expulsion/extraction using a digital semi-analytical scale (Shimadzu), with a capacity of 3.2 kilograms (kg) and a precision of 0.01 g. Umbilical cord, membranes, and fetal blood contained within the large chorionic vessels and intervillous space, were included in weight measurement. Newborns’ anthropometric measures were obtained after birth (maximum within a 24-hour period) by trained obstetrician nurses. Weight was measured (g) using digital pediatric scales (Toledo), with a maximum capacity of 15 kg and a precision of 5 g. Length, head and chest circumferences were measured using a non-stretching flexible measuring tape and expressed as centimeters (cm). Rohrer’s ponderal index was calculated as the newborn’s weight (g) divided by the cube of the length (cm).

Gestational age estimation and other definitions

Gestational age (GA) was assessed during the first trimester of pregnancy using crown-rump length measured by trans-abdominal ultrasound. In the absence of this examination within this period, the GA was estimated by the woman’s last menstrual period (LMP) together with the earliest ultrasound performed. The LMP method is recommended by the Brazilian MoH for GA calculation. Low birth weight (LBW) was defined as birth weight below 2500 g. Preterm birth was defined as birth before the 37th week of gestation, abortion as birth occurring before the 22nd week of gestation, and stillbirth as fetal death occurring between the 22nd week of gestation and time of delivery (Fig 1B). Term LBW was defined as term birth (after 37th week of gestation) with weight below 2500 g. Small for gestational age (SGA), reduced head circumference, and reduced body length were defined as: birth weight, head circumference and length below the 10th percentile sex-specific for GA according to the INTERGROWTH-21st standards [31]. Maternal anemia was defined as hemoglobin levels lower than 11 g/dL. Placental malaria was diagnosed according to a previously established classification [12].

Angiogenic factors and leptin measurement

Angiogenic factors (angiopoietins 1 and 2 (ANG-1 and ANG-2), soluble TEK receptor tyrosine kinase (sTIE-2), vascular endothelial growth factor A (VEGF-A), receptors fms-like tyrosine kinase-1 (sFlt-1) and soluble vascular endothelial growth factor receptor 2 (sVEGFR-2)), and the leptin hormone were measured in placental plasma (1:20 dilution for all factors) using the DuoSet ELISA development kits (R&D), according to manufacturer’s guidelines.

Cytokines/anaphylatoxins measurement by cytometric bead array (CBA)

Interleukin (IL-) 1β, IL-6, IL-8, IL10, IL-12p70, and TNF-α were detected and quantified in placental plasma by CBA Human Inflammatory kit (BD Biosciences), according to the manufacturer’s protocol. Complement (C3a, C4a, and C5a) activation was evaluated in placental plasma through CBA Human Anaphylatoxin kit (BD Biosciences). Samples were analyzed in a two-laser BD FACSCalibur flow cytometer with CellQuest version 5.2 software (BD Biosciences) and concentrations computed using FCAP array software version 3.0.1 (BD Biosciences).

Measurement of total IgG antibodies against Plasmodium vivax merozoite surface protein 1 (PvMSP1₁₉)

For the study of total IgG antibodies against PvMSP1₁₉, samples collected at the first time-point (recruitment) were used; yet, we excluded infected pregnant women who were recruited after having already one or more malaria episodes during that pregnancy. Enzyme-linked immunosorbent assay (ELISA) was performed using high-binding 96-well microplates (Costar, Cambridge, MA) coated with Glutathione S-transferase (GST)-PvMSP1₁₉ (31.25 ng/well), produced as described elsewhere [32,33]. In parallel, GST was used as a control for non-specific binding. The optimal concentrations were determined by plasma serial titration of positive and negative controls. Plates were incubated at 4°C overnight and washed four times with phosphate buffered saline with 0.05% Tween 20 (PBS-T). PBS-T containing skimmed milk was used for blocking (4%) and diluting plasma and antibodies (1%). Each plasma sample (50 μL/well) was diluted 1:200. After a 1-hour incubation at room temperature, antibody-antigen complexes were detected with HRP-conjugated mouse anti-human IgG (IG266) (Novus, Cat# NBP2-34648H) at a 1:3,000 dilution. Enzyme-substrate reaction was revealed with tetramethylbenzidine (TMB)/H₂O₂ at room temperature (dark conditions). After 10 minutes, the reaction was interrupted with 2N H₂SO₄. Absorbance values were measured at 450/595 nm using a CLARIOstar Plus plate reader (BMG Labtech). Corrected absorbance values were obtained by subtracting absorbance readings with GST ran on the same microplate. Absorbance data used in statistical analyses correspond to corrected values. Reactivity indices (RIs) were calculated as the ratio between each test sample’s corrected absorbance values and a cut-off value for antigen, corresponding to the average corrected absorbance for samples from 10 malaria-naïve blood donors plus 3 standard deviations. Positive samples were considered as having RIs greater than 1.

Statistical analysis

Data were analyzed using Stata/SE 14.2 (Stata, College Station, TX, USA) and GraphPad Prism 6.0 software (GraphPad Software Inc., San Diego, CA, USA). Continuous variables were summarized using mean and standard deviation (SD) values as well as median values and interquartile ranges (IQRs). Categorical variables were summarized using frequencies and percentages. Differences between groups were evaluated using the nonparametric Kruskal-Wallis test followed by the Dunn’s post hoc multiple comparison test and the Mann-Whitney test as appropriate. Differences between categorical data and proportions were analyzed using Pearson′s chi-squared test or Fisher’s exact test depending on the type of variables. All p values were 2-sided and interpreted at a significance level of 0.05. To compare the association between malaria and other variables (clinical characteristics, newborns’ anthropometric measurements, placental parameters, angiogenic and inflammatory factors, and anti-PvMSP1₁₉ IgG), we used linear regression, adjusting for maternal age, number of gestations, residence, education level, and occupation. To assess associations between malaria and adverse pregnancy outcomes, the presence of anti-PvMSP1₁₉ IgG and outcomes in the context of MiP, adjusted odds ratios (aOR) with 95% confidence intervals (CI) were estimated using multiple logistic regression, which included infection by malaria (no / yes), maternal age (≥ 18 years old / ≤ 17 years old), number of gestations (two or more / one), residence (urban / rural), education level (≥ secondary / ≤ primary), occupation (others / housewife) and presence of anti-PvMSP1₁₉ IgG (no / yes) as explanatory variables and adverse pregnancy outcomes (yes / no) as response variables. The first category for each explanatory variable was considered as the reference.

Study population and baseline characteristics

From 600 pregnant women enrolled in the major study, 329 (54.8%) were eligible for further analysis after applying the exclusion criteria and malaria parasite species diagnosis by PET-PCR (Fig 1A). Of these, 170 (51.7%) were classified as non-infected and 159 (48.3%) as P. vivax (Pv)-infected during gestation (Fig 1A). The overall socio-demographic profile of the stratified participants is depicted in Table 1. Group stratification was performed by the trimester in which the first infection occurred, considering that time of infection is known to be associated with MiP poor pregnancy outcomes [34]. Fig 1B illustrates a standard gestational window with the hypothetical poor outcomes in unsuccessful pregnancies. Table 1 shows some heterogeneity between the groups since pregnant women in the infected groups tend to be younger, with reduced educational level and rural residence. Occupation also varies significantly between the groups, but no difference was found regarding the number of previous pregnancies. The aforementioned characteristics were used as explanatory variables in the following multiple logistic and linear regressions analysis due to reported heterogeneity and their possible impact as confounding effects.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of the mothers participating in the study, according to infection and gestational trimester of first infection.

https://doi.org/10.1371/journal.pntd.0009390.t001

P. vivax infections during the first trimester are associated with MiP poor pregnancy outcomes

Malaria in pregnancy is known to impact proper fetal development [2]. We assessed pregnancy outcomes and newborns’ anthropometric measurements to ascertain if P. vivax infections at a given gestational trimester are associated with impaired fetal development and poor outcomes. Multiple logistic regression analysis revealed that P. vivax infections during the first trimester of pregnancy increase the odds of premature birth (aOR 8.12, 95%CI 2.69–24.54, p < 0.0001), LBW (aOR 4.34, 95%CI 1.49–12.62, p = 0.007), as well as LBW in those born at term (aOR 4.75, 95%CI 1.06–21.35, p = 0.04). First trimester infections were also associated with reduced head circumference (aOR 3.58, 95%CI 1.29–9.97, p = 0.01) and newborn’s length (aOR 4.48, 95%CI 1.66–12.06, p = 0.003) (Fig 2). No differences were observed for the overall birth weight and length (Fig 3A and 3D and S1 Table). Though, in line with our findings, we observed statistically significant differences in head and chest circumference of children born from women infected with P. vivax during the first trimester and those born from non-infected women (median [IQR], head–Pv-1^st tri 34.0 cm [32.0–34.0] vs. NI 34.0 cm [33.0–35.0], p = 0.007; chest–Pv-1^st tri 33.0 cm [32.0–34.0] vs NI 34.0 cm [33.0–35.0], p = 0.006) (Fig 3B and 3C and S1 Table). Our results suggest that P. vivax infections during the first gestational trimester promote preterm birth and impair fetal development, possibly due to clinical manifestations associated with MiP.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Association between adverse pregnancy outcomes and P. vivax infections according to the trimester of first infection.

Forest plot of odds ratios of adverse events in newborns from women infected during pregnancy compared with newborns from non-infected women is shown. n/N—number of events by total number of individuals in each group; aOdds Ratio—adjusted Odds Ratio; CI—confidence interval; tri—gestational trimester. Odds ratios were adjusted (aOR) for maternal age, gravidity status, residence, education level and occupation. Preterm birth–birth < 37^th week of gestation; Low birth weight–birth weight < 2500 g; Term low birth weight–birth weight < 2500 g from 37^th week of gestation; Small for gestational age, Reduced head circumference and Reduced body length–birth weight, head circumference and length (respectively) < 10^th centile for sex-specific gestational age. P-values were estimated through multiple logistic regression methods.

https://doi.org/10.1371/journal.pntd.0009390.g002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Anthropometric measurements of newborns according to maternal infection and trimester of first infection.

(A) Newborn weight; (B) Head circumference; (C) Chest circumference; and, (D) Length. NI–non-infected (N = 160–169); Pv–P. vivax-infected pregnant women (N = 130–150); Pv-1^st tri–P. vivax infection 1^st trimester (N = 38–46); Pv-2^nd tri–P. vivax infection 2^nd trimester (N = 43–54); Pv-3^rd tri–P. vivax infection 3^rd trimester (N = 49–50). Weight is given in grams (g). Head and chest circumferences, and body length are informed in centimeters (cm). Data are represented as Tukey boxplots, with the bottom and top of each box representing the first and third quartiles; the line inside the box is the median, and the whiskers represent the lowest and the highest data within 1.5 IQR of the first and upper quartiles; and the circles, outliers. Differences between groups were determined by multiple liner regression (adjusted for maternal age, gravidity status, residence, education level, and occupation).

https://doi.org/10.1371/journal.pntd.0009390.g003

Clinical and infection characteristics of pregnant women with P. vivax MiP

As for the aforementioned outcomes, pregnant women infected with P. vivax had some clinical characteristics that are worth mentioning (Table 2). Women infected during the first trimester had reduced gestational age at delivery when compared with non-infected counterparts (median [IQR], Pv-1^st tri 39.0 weeks [37.0–40.0] vs. NI 40.0 weeks [39.0–40.0], p = 0.006). These results are in line with our observations that suggest an association between first trimester infections and pregnancy outcomes such as LBW and preterm delivery (Fig 2 and Table 2). Reduced maternal weight gain was also observed in this group (median [IQR], Pv-1^st tri 10.0 Kg [7.0–12.0] vs. NI 13.5 Kg [10.0–16.8], p = 0.004). Anemia is frequently found in pregnant women infected with P. vivax, which has been suggested to be one of the leading causes of impaired fetal development during P. vivax MiP [16,35]. However, a statistically significant reduction in hematocrit was only observed for the overall Pv-infected group (median [IQR], Pv 34.8% [32.4–37.5] vs. NI 36.6% [34.5–38.6], p = 0.008) and increased clinical anemia in those infected during the third trimester (Pv-3^rd tri 28.3% vs. NI 14.1%, p = 0.02). Additionally, women in the Pv-infected group reported more malaria episodes before current pregnancy than their non-infected counterparts regardless the infection trimester (Table 2). Pregnant women infected during the first trimester had higher chances of having three or more infections than infected women from other groups. Parasitemia tend to be similar for the first infection across gestational trimesters; yet, it seems to decrease on average from the second infection onward (Table 3). 8.1% of women infected in the first trimester had past PM, while 4.6 and 4.3% of women infected during the second and third trimesters had active acute PM, respectively. This was established by the presence of hemozoin in milder or moderate amounts. However, it was absent in more than 90% of the examined placentas, regardless of the infection’s gestational trimester. Our observations support the idea that systemic effects rather than PM are linked to the outcomes observed during P. vivax MiP in our cohort.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Clinical characteristics of the mothers participating in the study, according to infection and gestational trimester of first infection.

https://doi.org/10.1371/journal.pntd.0009390.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of the infection during pregnancy.

https://doi.org/10.1371/journal.pntd.0009390.t003

P. vivax infections across gestational trimesters lead to diverse placental histologic alterations and angiogenic and cytokine imbalance

Reduced PM prevalence is frequently observed in P. vivax infections because, contrary to P. falciparum infections, P. vivax sequestration in the placenta is controversial [14,15] and studies that report this event frequently have reduced sample sizes. Nevertheless, placental histologic changes had been observed [9–11]. To determine whether Pv-infections across the gestational trimesters result in considerable histologic and cytological alterations, we have evaluated several histologic parameters in more than 100 placentas from Pv-infected women (Fig 4 and S2 Table). We observed that placentas from Pv-infected women had a significant increase in syncytial nuclear aggregates (SNA), fibrin deposition, monocytes and leukocytes infiltrate (median [IQR], SNA–Pv 15.0 [11.0–21.0] vs. NI 13.0 [10.0–17.0], p = 0.001; fibrin–Pv 2.5 [1.9–2.8] vs. NI 1.9 [1.9–2.8], p = 0.001; monocytes–Pv 8.0 [6.0–13.0] vs NI 4.0 [2.0–6.0], p < 0.0001; leucocytes–Pv 21.5 [14.0–29.0] vs. NI 15.0 [9.0–21.0], p = 0.004) (Fig 4 and S2 Table). Differences between placentas from non-infected women and those from women infected in the first trimester were observed regarding SNA and monocyte infiltrate (median [IQR], SNA–Pv-1^st tri 19.0 [11.0–22.0] vs. NI 13.0 [10.0–17.0], p = 0.02; monocytes–Pv-1^st tri 7.0 [5.0–9.5] vs NI 4.0 [2.0–6.0], p = 0.01) (Fig 4A and 4C and S2 Table). Interestingly, infections closer to term tend to yield higher levels of fibrin deposition and monocytes infiltrate at delivery (Fig 4B and 4C and S2 Table). These findings are representative of placental anomalies that can be extended to angiogenic and inflammatory mediators in placental plasma (S3 and S4 Tables). First and third trimester infections led to reduced levels of angiopoietin-2 (ANG-2) and increased levels of soluble TEK receptor tyrosine kinase (sTIE-2), respectively (median [IQR], ANG-2 –Pv-1^st tri 4.4 ng/mL [1.0–6.6] vs. NI 6.8 ng/mL [3.5–12.4], p = 0.03; sTIE-2 –Pv-3^rd tri 18.5 ng/mL [11.7–24.7] vs. NI 13.9 ng/mL [9.1–19.0], p = 0.008). Overall, factors such as sFlt-1, sVEGFR-2, and the hormone leptin were also altered in placentas from Pv-infected women (S3 Table). Regarding inflammatory mediators, only IL-12 and C5a showed more consistent and significant alterations, as they were reduced at delivery due to first and third trimester infections (median [IQR], IL-12 –Pv-1^st tri 3.3 pg/mL [2.6–3.8] vs. NI 3.6 pg/mL [3.0–4.2], p = 0.009; C5a –Pv-1^st tri 807.1 pg/mL [534.9–913.9] vs. NI 1117.2 pg/mL [814.8–1475.5], p = 0.03; Pv-3^rd tri 673.0 pg/mL [445.9–871.3] vs. NI 1117.2 pg/mL [814.8–1475.5], p = 0.006) (S4 Table). Altogether, our observations suggest that poor pregnancy outcomes associated with first trimester infections may be linked to placental inflammation, structural abnormalities and angiogenic imbalance during P. vivax MiP.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Evaluation of placental histological parameters according to maternal infection and trimester of first infection.

(A) Syncytial nuclear aggregates; (B) Fibrin deposition score; (C) Monocytes infiltrate (CD68+); (D) Leukocytes infiltrate (CD45+). Histologic parameters were evaluated by microscopy through H&E (syncytial nuclear aggregates and fibrin deposition) or immunohistochemistry (monocytes and leukocytes). NI–non-infected (N = 153–155); Pv–P. vivax-infected pregnant women (N = 118–128); Pv-1^st tri–P. vivax infection 1^st trimester (N = 33–37); Pv-2^nd tri–P. vivax infection 2^nd trimester (N = 42–44); Pv-3^rd tri–P. vivax infection 3^rd trimester (N = 43–47). Data are represented as Tukey boxplots, with the bottom and top of each box representing the first and third quartiles; the line inside the box is the median, and the whiskers represent the lowest and the highest data within 1.5 IQR of the first and upper quartiles; and the circles, outliers. Differences between groups were determined by multiple liner regression (adjusted for maternal age, gravidity status, residence, education level, and occupation).

https://doi.org/10.1371/journal.pntd.0009390.g004

Antibodies against P. vivax MSP1₁₉ are not associated with protection from poor pregnancy outcomes

The impact of gravidity on P. vivax MiP outcomes is still inconclusive [16,35,36]. Possibly the absence of pregnancy-specific antigens in P. vivax and its restricted distribution to areas of low malaria transmission may impair immunity acquisition across pregnancies [8,37]. As such, other antigens might be targeted by the immune system to confer some degree of protection from P. vivax MiP. Therefore, we measured antibodies against the C-terminal fragment of the P. vivax merozoite surface protein 1 (PvMSP1₁₉). We used it as a surrogate marker of previous exposures (Fig 5A and S5 Table) and sought possible associations between antibody titers and improved pregnancy outcomes. Infected pregnant women who were recruited with more than one malaria episode during their current pregnancy were removed from the analysis, to target antibodies acquired upon any malaria episode occurred before pregnancy. We observed that Pv-infected women previously exposed to the parasite (Pv-PvMSP1^Ab+) had higher titers of IgG when compared with non-infected pregnant women with prior exposure (NI-PvMSP1^Ab+) (median [IQR], Pv-PvMSP1^Ab+ 65.1 [14.1–91.6] vs. NI-PvMSP1^Ab+ 4.0 [2.1–9.8], p < 0.0001) (Fig 5A and S5 Table). However, no associations were observed between IgG titers and outcomes, suggesting that the presence of these antibodies do not protect pregnant women from the adverse outcomes of P. vivax MiP (Fig 5B). Accordingly, no statistically significant differences were observed for birth weight, head and chest circumference, and body length, between P. vivax-infected women who had PvMSP1₁₉-specific antibodies and those who did not have them (Fig 5C–5F and S5 Table).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Anthropometric measurements of newborn, according to previous exposure to P. vivax (PvMSP1).

(A) Antibody profile among non-infected and P. vivax-infected pregnant women; (B) Forest plot of odds ratios of adverse pregnancy outcomes in infected women with anti-PvMSP1 IgG compared with infected women without anti-PvMSP1 IgG; (C) Newborn weight; (D) Head circumference; (E) Chest circumference; and, (F) Length. NI-PvMSP1^Ab-—non-infected non-exposed (N = 106–112); NI- PvMSP1^Ab+—non-infected exposed (N = 54–58); Pv- PvMSP1^Ab-—P. vivax-infected non-exposed (N = 19); Pv- PvMSP1^Ab+—P. vivax-infected exposed (N = 80–95). PvMSP1- Plasmodium vivax merozoite surface protein 1; Ab—Antibody; LBW—low birth weight; SGA—small for gestational age; HC—head circumference. RI—Reactivity indices were calculated as the ratio between each test sample’s corrected absorbance values and a cut-off value for antigen, corresponding to the average corrected absorbance of samples from 10 malaria-naïve blood donors plus 3 standard deviations. Weight in grams (g). Head and chest circumferences, and body length in centimeters (cm). Data are represented as Tukey boxplots, with the bottom and top of each box representing the first and third quartiles; the line inside the box is the median, and the whiskers represent the lowest and the highest data within 1.5 IQR of the first and upper quartiles; and the circles, outliers. Odds ratios were adjusted (aOR) for maternal age, gravidity status, residence, education level and occupation using multiple logistic regression. Differences among the groups were determined by multiple linear regression (adjusted for maternal age, gravidity status, residence, education level and occupation).

https://doi.org/10.1371/journal.pntd.0009390.g005

Overall, our results suggest that P. vivax infections during the first gestational trimester are associated with poor pregnancy outcomes most likely due to placental histologic and physiologic abnormalities, which are unlikely to be prevented by the presence of IgG against PvMSP1_19, acquired during previous P. vivax-malaria episodes.