Leptin , insulin like growth factor-I levels and histology-diagnosed placental malaria in an area characterized by unstable malaria transmission in central Sudan

Background: There are few published data on the association between leptin, insulin like growth factor-1 (IGF-1) and malaria during pregnancy. This study aimed to investigate maternal and umbilical cord leptin and IGF-1 levels and malaria during pregnancy, and their association if any with birth weight. Methods: A cross-sectional study was conducted at Medani, Sudan. Medical and obstetrics history was gathered from each parturient woman (n=175) and malaria was investigated by blood film and placental histology. Maternal and umbilical cord leptin and IGF-1 levels were measured using ELISA. Results: Upon histological examination, 48 women were infected with placental malaria, and 127 were found free from the disease. Out of the 48, 2 of the patients showed signs of active infection, 3 of chronic infection and 43 of previous infection. Placental malaria and preterm delivery were associated with low birth weight (< 2500 g). Younger mothers and primigravidae had a higher risk for placental malaria infection. There was no significant difference in maternal and umbilical cord leptin and IGF-1 levels between women infected with placental malaria and those free from the disease. Conclusions: The current study showed that low birth weight was significantly associated with placental malaria. Young mothers and primigravidae had a higher risk to develop the infection. There was no significant difference in the levels of maternal and umbilical cord leptin and IGF-1 levels between women infected with placental malaria and those free from the disease. Both the levels of maternal and cord leptin and IGF-1were found not to be associated with birth weight. Abbreviations: IGF-1: Insulin like growth factor-1; LBW: Low birth weight; ELISA: Enzyme-linked immunosorbent assay; PM: Placental malaria. 1 2 3


Introduction
Malaria during pregnancy is a major public health concern, especially in sub-Saharan Africa where there are approximately 125 million pregnant African women living in malaria-endemic regions. Almost one fifth of these pregnant women are at risk of malaria (Dellicour et al., 2010;Desai et al., 2007). Malaria during pregnancy is the main cause of maternal, perinatal and neonatal adverse effects, especially anemia and low birth weight (LBW) (Ahmed et al., 2014;Menendez et al., 2000;Rogerson et al., 2003).
The pathogenesis of placental malaria and LBW is not fully understood. Leptin is a hormone secreted mainly by adipocytes (Zhang et al., 1994) that can potentiate inflammation by enhancing macrophage phagocytosis (Loffreda et al., 1998;Pacifico et al., 2006). Previous reports have shown that leptin levels were decreased during malarial attack in pregnant women (Conroy et al., 2011), and that these decreased leptin levels were associated with placental malaria infection, as well as low birth weight (Kabyemela et al., 2008a;Kabyemela et al., 2008b).
Insulin-like growth factor-I (IGF-1), also called somatomedin C, is a polypeptide with a sequence similar to that of insulin (Rinderknecht & Humbel, 1978). Recently, maternal and umbilical cord blood levels of IGF-1were investigated in malaria during pregnancy as possible determinants of birth weight (Ayoola et al., 2012;Umbers et al., 2011).
Research on malaria during pregnancy and its associated adverse effects e.g. LBW is highly valuable for researchers and clinicians as it can yield basic data needed for the future vaccine.
Pregnant Sudanese women are susceptible to malaria regardless of their age and parity, and malaria is associated with increased maternal mortality, anemia, LBW, and stillbirths (Adam et al., 2005a;Ali et al., 2011;Bader et al., 2010;Mohammed et al., 2013). Central Sudan is characterized by unstable malaria transmission, and P. falciparum is the main malaria parasite species reported in the area (Malik et al., 2004). To add to the research on placental malaria during pregnancy that has been carried out already (Alim et al., 2015;Mostafa et al., 2015;Salih et al., 2011), the current study was conducted in central Sudan to investigate the maternal and umbilical cord levels of leptin and IGF-1in placental malaria infection.

Methods
A cross-sectional study was conducted from August to October 2014 in the labor ward of the Medani Maternity Hospital. After signing an informed consent form, women with singleton pregnancies were approached to participate in the study. Women with twins, hypertension, diabetes mellitus and antepartum hemorrhage were excluded from the study. Socio-demographic data (age, parity, residence and gestational age) and data on obstetric history, medical history, and bed net use were gathered using a structured questionnaire that was completed by a trained medical officer in the local language (Arabic). Maternal weight and height were measured and body mass index (BMI) was calculated and expressed as weight(kg)/height(m) 2 . Maternal hemoglobin concentrations were estimated (HemoCue AB, Angelhom, Sweden). Newborns were weighed immediately following birth using the Salter scale and the sex of each newborn was recorded. The total sample size was calculated assuming that at least 23% of parturient women would have placental malaria infection. To have over 80% power to detect a difference of 5% at α = 0.05, we recruited 175 women. We assumed that 10% of women might not respond or have incomplete data.

Giemsa-stained blood smears for light microscopy
Maternal, placental, and umbilical cord blood films were prepared for testing. Slides were stained by 10% Giemsa. In the slides positive for malaria the number of asexual parasites was counted per 200 leukocytes, assuming a leukocyte count of 8000 leukocytes per μl (for thick films) or per 1000 red blood cells (for thin films). Blood films were considered negative if no parasites were detected in 100 oil immersion fields of a thick blood film.
The maternal and umbilical cord blood was then allowed to clot, centrifuged for 10 minutes at 3000 rpm and the serum separated and stored at -20°C until further analysis.

Placental histology
The details on placental histology have been mentioned previously (Alim et al., 2015;Mostafa et al., 2015;Salih et al., 2011). In summary, a 3cm 3 full thickness sample was obtained from the maternal surface approximately half the distance between the umbilical cord and the edge of the placenta. The placental biopsy samples were immediately placed in 10% neutral buffered formalin. Buffer was used to prevent formation of formalin pigment, which might be difficult to differentiate from malaria pigment (Bulmer et al., 1993a). The placental biopsy samples were then embedded in paraffin wax/sections. In every case, the thick paraffin sections were stained with hematoxylin-eosin and Giemsa stains. Slides were read by a pathologist who remained blind to the clinical characteristics of each of these samples. Placental malaria infection was characterized using parameters previously described by Bulmer et al.: uninfected (no parasites or pigment), acute (parasites in intervillous spaces), chronic (parasites in maternal erythrocytes and pigment in fibrin, or cells within fibrin and/or chorionic villous syncytiotrophoblast or stroma), and previous (no parasites, and pigment confined to fibrin or cells within fibrin) (Bulmer et al., 1993b).

ELISA for measuring leptin and IGF-1 levels
Maternal and umbilical cord serum levels of leptin and IGF-1 were measured using ELISA Kits and the manufacturers' instructions were strictly followed (DRG Diagnostics, Marburg, Germany).

Statistical analysis
The data analyses were performed using SPSS statistical software for windows (version 18.0). Statistical significance was set at P value < 0.05. To compare means and proportions between groups, student's t-test and Chi-square test were used, respectively. For non-parametric data significant differences in means between two groups were calculated using the Mann-Whitney test. Univariate and multivariate analyses were performed with a logistic regression model where placental malaria infection was the dependent variable and expected risk factors (mother's age, parity, mother's weight, mother's haemoglobin, educational level, residence, use of bed net, antenatal care attendance, use of folic acids supplements and mother's serum leptin and IGF-1) were the independent variables.
Odds ratios (OR) and 95% confidence intervals (CI) were calculated. Linear regression models were set to investigate the factors associated with the level of mother's haemoglobin and birth weight. Predictor variables for mother's haemoglobin model were; antenatal care attendance, parity, BMI, maternal serum leptin and use of folic acid supplements. Predictor variables for birth weight were; mother's age, antenatal care attendance, mother's haemoglobin, placental malaria, mother's height, delivery gestational age.

Ethics
The study received ethical clearance from the Research Board at the Faculty of Medicine, University of Khartoum, Sudan. (Approval number: 2-2011).

Selection of participants
Pregnant women who delivered at Medani Maternity hospital from August through to October 2014 were recruited for this study, following written informed consent. All participants finally included in the study had to satisfy the selection criteria and have none of the exclusion criteria.
The mean age (± SD) of PM + patients was 26± 4.8 years and ranged from 17 to 38 years. The mean age (± SD) of PM − patients was 28± 6 years with a range of 18 to 41 years (Table 1). Younger women (25 -30 years) were significantly more often infected with PM (P = 0.02)(Dataset 1 (Elsheikh et al., 2017)).
Moreover, babies born to women with PM tended to be in the LBW (< 2500 g) category more often than those born to non-infected women, but the p-value failed to reach the significance level (p = 0.054). Maternal weight, BMI, gravidity, gestational age at delivery and hemoglobin levels were unchanged significantly between groups (Dataset 1 (Elsheikh et al., 2017)).

Risk factors for placental malaria
Univariate and multivariate analysis demonstrated that only the mother's age and parity were significant risk factors (p-values were 0.008 for mother's age and 0.009 for parity).
Younger mothers and primigravidae had a higher risk for PM. The risk of infection was lower for older mothers, with an odds ratio (OR) of 0.881 (p = 0.008, 95%CI: 0.802 -0.968). For each additional year in age, the odds of getting placental malaria lowered by a factor of 0.881. The OR for parity was 4.3 (p = 0.009, 95%CI: 1.45 -12.998) ( Table 2, Dataset 1 (Elsheikh et al., 2017)).

Serum levels of leptin and IGF-1
The levels of leptin were higher in LBW infants and their mothers, whilst IGF-1levels were higher in normal weight infants and their mothers. However, these differences failed to reach statistical significance (Table 3). Non-infected mothers and their infants showed higher levels of leptin and IGF-1 than infected ones ( Figure 1 and Figure 2), but these differences also failed to reach statistical significance (Table 4, Dataset 1 (Elsheikh et al., 2017)).  Linear regression analysis showed that gestational age had the strongest positive effect on birth weight (β = 0.191, p = 0.01), followed by antenatal care attendance (p= 0.043) and mother's age (p= 0.85, not significant) ( HME has confirmed that all raw data provided with this manuscript has been de-identified.

Discussion
The main findings of the current study are that placental malaria is significantly associated with LBW. Younger mothers and primigravidae had a higher risk for PM infection. There was no significant difference in leptin and IGF-1 levels between PM+ and PM− women and their infants, as well as between LBW infants and their mothers and normal weight infants and their mothers. Maternal and umbilical cord leptin and IGF-1 levels were not associated with birth weight.
Our results coincide with what has been reported previously about LBW being significantly associated with placental malaria (Albitiet et al., 2010;Aribodor et al., 2009;Menendez et al., 2000). Although some studies conducted in different areas in Sudan did not report  Mann-Whitney test was used to compare the levels of maternal and umbilical cordIGF-1 between the two groups. PM + women showed lower levels of maternal and cord IGF-I but these differences were not statistically significant (Dataset 1).  et al., 2013) found that placental infections had no effect on LBW or anemia.
The study showed that leptin levels were higher in non-infected mothers and their infants than in infected ones, however these differences failed to reach statistical significance. This concurs with a study conducted in Malawi which found a significant reduction of leptin levels in mothers infected with PM, and according to this the authors suggested leptin to be an informative biomarker for diagnosis of PM (Conroy et al., 2011). Another study in Tanzania also reported the same finding (Kabyemela et al., 2008a). Kabyemela and his colleges (Kabyemela et al., 2008b) also investigated the effect of PM in the relationship between cord leptin levels and birth weight. They found that cord leptin had a strong positive relationship with birth weight in offspring of PM − women (P = 0.02 to P < 0.0001) but not in offspring of PM + women, however the differences did not reach significance level.
Although the current study failed to detect a significant differences in IGF-1 levels between PM+ and PM− women, one study have shown that placental malaria-associated inflammation disturbs maternal and fetal levels of IGFs, which regulate fetal growth (Umbers et al., 2011). This may be one mechanism by which placental malaria leads to fetal growth restriction, but this study did not report the effect size of PM on IGF-1 levels.
It is worth mentioning that the differences between the current study and later ones that reported low leptin (Conroy et al., 2011;Kabyemela et al., 2008a;Kabyemela et al., 2008b) and IGF-1 levels in maternal and umbilical cord serum (Umbers et al., 2011) might be due to the duration of malaria infection itself. While the majority of malaria infections in the current study were past placental infection (4% of PM + patients had active infection, 6% had chronic and 90% past infection), the later studies reported results from active placental infections. Furthermore, submicroscopic placental malaria infection using PCR (Polymerase Chain Reaction) was not investigated in the current study.
We have recently reported that women with submicroscopic malaria were at higher risk to have LBW (Mohammed et al., 2013). Likewise Adegnika et al. have reported that microscopic and submicroscopic P. falciparum infection, but not inflammation (C-reactive protein) caused by infection, is associated with low birth weight (Adegnika et al., 2006).
The main limitation of this study was that we relied on a single measure of leptin and IGF-1 levels at delivery; however it was not feasible to obtain the levels of these infants before birth. Most studies relating umbilical cord blood IGF-1levels and birth weight reported single measurements at birth (Ong et al., 2000;Yang & Yu, 2000), and findings are similar. Another limitation was that we did not measure the concentrations of any other components of the IGF axis, including growth hormone, insulin, IGF binding proteins (IGFBP1-5), and receptors (IGF-1R and 2R) to further establish the potential implication of the IGF system in fetal growth. More limitation is that although we are interested only in the biologically active IGF-1 (free form), the ELISA technique used in this study measures the total amount of IGF-1 (free and protein-bound) in serum.

Conclusions
The current study shows that there is no statistically significant difference in the levels of maternal and cord leptin and in the levels of IGF-1 between PM+ women and PM -women, and between women who delivered LBW infants and those who delivered normal weight ones. The main finding is that placental malaria is significantly associated with LBW. Neither maternal and umbilical cord leptin levels nor IGF-1 levels were associated with birth weight.

Data availability
Dataset 1. The file contains data on socio-demographics (age, parity, residence and gestational age), obstetric and medical history, bed net use, maternal weight, height and BMI, maternal hemoglobin, infant birth weights and maternal and umbilical cord leptin and IGF-1 levels for each participant. HME has confirmed that all raw data provided with this manuscript has been deidentified. DOI, 10.5256/f1000research.10641.d158697 (Elsheikh et al., 2017) Author contributions HME and IA designed the experiments. IA conceived the study and participated in study coordination. EME conducted the clinical work. AAM performed the pathological analysis. HME carried out the laboratory work, statistical analysis and study coordination. HME, IA and MIE prepared the first draft of the manuscript. MIE contributed to the experimental design. AHK contributed in statistical analysis. All authors were involved in the revision of the draft manuscript and have agreed to the final content.

Competing interests
No competing interests were disclosed.

Grant information
This work was partially funded by the Ministry of Higher Education (Khartoum, Sudan).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Introduction
The reference to Dellicour et al. 2010 and 125 million pregnancies includes geographies outside of sub-Saharan Africa. The first two sentences are a bit confusing and might be amended to read "number of pregnancies at risk, number in areas endemic for p. falciparum" since that's the focus of the study. ○ I would also amend "the main cause" to "a main cause" since malaria is an important contributor but other factors (nutrition, lack of access to antenatal care etc.) are also principal contributors to adverse outcomes. ○ I think the introduction could be expanded to include more background around the selected biomarkers and the rationale for selecting them. e.g. relationship between Leptin and metabolic function (BMI etc.) and pregnancy, perhaps in relation to the nutritional ○ status of women in the cohort. The authors state that "the pathogenesis of placental malaria and LBW is not fully understood" and then move into Leptin. It might be useful top provide some rationale around how the authors proposed Leptin and IGF-1 play a role in the pathogenesis of LBW resulting from placental malaria. I found the paragraph about vaccine development a bit confusing. I would provide more information around how the results of this study (or research on the effects of placental malaria) might contribute to development of a vaccine or potentially remove that paragraph Provide additional information around the variable "antenatal care attendance" (e.g. is this number of visits total, number of women reaching 3 or 4 visits, number of women beginning visits in the first or second trimester, was this data collected by questionnaire or from the clinics?) ○ All variables presented in Table 2 should have more information (in table and in methods) e.g. maternal age (years), weight (kg), education level (years), residence (type?), hemoglobin (g/dL) use of bed net (yes/no or months during pregnancy?), maternal leptn (ng/mL) etc.) ○ Missing data should be reported (e.g. are they missing data on birth weight (175-166 = 9 data points on birth weight are missing?)) If these missing cases are home deliveries and/or missing because of very early delivery that could confound the results. Also in Table 4 where n = 47 "infected" but the authors report n = 48 cases of PM+ in the results, the missing data is reported in the figure legend but should also be included in the tables/methods.

○
Are the authors confident that -20 is sufficient? Depending on how long the samples were stored for that might not be sufficient (compared with -80) to preserve the samples.

Results
Age and parity are likely highly correlated. That should be mentioned in the results (and potentially effects the analysis, for example including co-variates that are highly correlated in a multivariate model can influence the results)

○
The authors present the breakdown of their cohort by blood group, they may want to ○ provide a rationale as to why they present those results or some interpretation of the results in the discussion (did they examine blood group in relation to infection?) Since the results are presented with LBW as an primary outcome (in association with malaria) the authors may want to acknowledge (or examine if they have the relevant data) the relative contribution of preterm birth and small-for-gestational age outcomes to LBW (which will be made up of both PTB and SGA babies)

○
The ELISA results should include the inter-assay co-efficient of variability (CV) and intraassay.

○
The authors should state "maternal" or "cord blood" (or both) when presenting their results (e.g. in paragraph 1 under the title "Serum levels of leptin and IGF-1" as this is how the results are presented in the tables (they may also want to emphasize in the results that the maternal blood samples were collected at delivery)

Discussion
Perhaps provide a discussion around the levels in serum vs. in plasma.
○ I would avoid the statement "failed to reach statistical significance" (also stated in results sections) and say instead (e.g.) "The study showed that leptin…., however, the difference was not statistically significant". ○ I'm not sure the authors can say that their results "concur with a study conducted in Malawi" if they're not reporting statistical significance. Perhaps they could state instead that the study in Malawi also observed lower serum levels of leptin in women with placental malaria infection, but reported a statistically significant difference. ○ I would avoid the term "normal weight ones" (last paragraph of the discussion) and say instead "those who delivered at normal birth weight (>2500 g)".

Is the work clearly and accurately presented and does it cite the current literature? Partly
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes leptin and IGF -I were included as independent variables for placental malaria. It will be useful for readers to know how these variables increase/decrease risk of PM. Any literature relating these to risk of PM is needed. If not available then the analysis may need to be revisited. Actually if they have data they may consider Iron supplementation as an independent variable in their analysis (studies relating Fe supplementation and risk of PM are available) Were these mothers of SP IPTp? Probably this justifies the inclusion of folic acid supplementation as an independent variable.
Similarly is not clear how the independent variables for maternal hemoglobin levels were identified. Why mother's leptin and IGF-I levels were included as independent variables for maternal hemoglobin levels. It will be useful for readers to know how these variables increase/decrease the risk of anemia. Any literature relating these to risk of anemia needed.
If not available then the analysis may need to be revisited.

Results
Readers may want to know what were the results of malaria microscopy studies. 1.
There are discrepancies in the numbers in the text and tables. Maternal age: Uninfected:

2.
Some of the reported results are not shown. It is stated that "Moreover, babies born to women with PM tended to be in the LBW (<2500 g) category more often than those born to non-infected women, but the p-value failed to reach significance level (p = 0.054)". What was the proportion of LBW infants in PM+ versus proportion of LBW infants in the PM-mothers?

3.
It is stated that PM was associated with LBW: "LBW (< 2500 g) was significantly associated with placental malaria (N= 172, p =0.009)" What is this N = 172 ?

4.
In table 2 univariate and multivariate analyses are shown. In the univariate analysis none of the variables investigated was associated with placental malaria. In the multivariate analysis Maternal age and parity are associated with PM. Can the authors explain this paradox? 5. Table 3 shows that neither leptin nor IGF-I levels differ between LBW babies of PM + and PM -mothers. The sample size is now 166 instead of 175. Can the authors indicate why 9 subjects are missing ?

6.
It is known that cord blood leptin levels correlate with fetal size. It was previously shown that cord leptin levels are lower in LBW babies born to PM-but not in offspring of PM+ mothers probably indicating that PM disrupts the normal relationship between leptin and fetal growth. 7.