Plasmodium falciparum infection in pregnant women attending antenatal care in Luanda , Angola Infecção por Plasmodium falciparum em mulheres grávidas em consultas pré-natal em Luanda , Angola

Introduction: Malaria during pregnancy remains a serious public health problem. The aim of this study was to establish the prevalence and possible risk factors for malaria in pregnant women attending antenatal care at Augusto Ngangula Specialized General Hospital in Luanda, Angola. Methods: Pregnant women (679 total) who attended antenatal care from April to September 2008 were included in the study after signing informed consent. For each participant, the social-demographic profile and malaria and obstetric histories were investigated via a questionnaire. Diagnosis was made by optic microscopy, and hemoglobin concentration measured. The associations between age, parity, gestational age, residence, schooling, malaria during gravity, anemia and treatment with incidence of Plasmodium falciparum infection were analyzed through logistic regression. Results: During the period of study, 74 (10.9%) out of 679 women were infected by P. falciparum. The average concentration of hemoglobin was 11.1 ± 0.07g/dL, and there were significant associations between the history of malaria during pregnancy, P. falciparum infection (p<0.01) and anemia at the time of observation (p<0.001). Conclusions: Previous history of malaria during pregnancy represents a risk factor for current infection and anemia was an important complication associated with malaria, even in women who were treated with sulfadoxine-pyrimethamine during pregnancy.

Malaria from Plasmodium falciparum during pregnancy is a serious public health problem in Sub-Saharan Africa.More than 50 million women residing in malaria-endemic areas become pregnant every year [1][2][3] .Although malaria during pregnancy might be asymptomatic due to the development of some degree of immunity by mothers residing in areas with stable transmission, it is still associated with unfavorable effects on the mother or the child [4][5][6][7][8] .For this reason, abortions, stillbirths, premature deliveries and maternal deaths are common phenomena 9,10 .Many, if not most, cases of maternal and perinatal death can be prevented 11 .
The impact of malaria-related anemia on pregnant women (id est., the risk of death) and fetuses (i.e., low birthweight) has been increasingly acknowledged during the last decades 12,13 , keeping in mind that anemia does not necessarily have to be symptomatic to become a risk factor during pregnancy.Low birth weight in newborns from mothers who had malaria during pregnancy is common; it is estimated that it causes 5% of perinatal deaths in areas with poor resources 14 .The prevalence of infection and the density of parasites reach their peak during the first half of pregnancy and progressively decrease until delivery 15 .
Factors that influence the prevalence of malaria in pregnant women were identified and include maternal age, parity, use of prophylaxis, nutrition, genetics of the host, level of anti-parasitic immunity and genetics of the parasite 16 .
During the last decade, several studies were performed in African countries with high indices of transmission [17][18][19][20][21][22][23][24] ; however, few data were published for Angola, which is one of the five countries with the lowest available amounts of information regarding malaria during pregnancy, a fact that makes decisionmaking for prevention and control difficult 25 .The aim of this study was to assess the prevalence of P. falciparum and its possible associated risk factors in pregnant women who attended antenatal care in Luanda.

Setting of the study and participants
A transversal study was performed at the antenatal care (ANC) facility at Augusto Ngangula Specialized General Hospital (Hospital Geral Especializado Augusto Ngangula -HGEAN) in Luanda, Angola.The study was performed during the high-transmission season, from April to September 2008.Pregnant women were recruited during routine antenatal visits.All patients were invited to participate in the study.Only those who gave informed consent were included; women with other diseases, particularly HIV carriers, were excluded.Each participant was evaluated only once.

Data collection
Data on malaria in pregnant women who attended ANC at HGEAN were obtained via a questionnaire including the following information: social-demographic profile, malaria history and obstetric history, including age, parity, gestational age and place of residence.Gestational age was calculated from the last menstruation date and via ultrasound when there was uncertainty.The obstetric physical exam helped to complete the required data (blood pressure, paleness of conjunctiva, level of the uterine fundus and fetal heart rate).

Blood collection and microscopic diagnosis
Blood was collected through digital or venous puncture.Samples were prepared for parasite detection and estimation of parasitemia.Blood smear and thick drop assays were prepared according to the protocol described by the World Health Organization (WHO) 26 .Blood smear and thick drop assays were stained with 10% Giemsa dye.Smears were fixed with methanol for one minute before staining.Parasite density was assessed by counting parasites in 200 leukocytes, assuming that each woman had 8,000 leukocytes/μL of blood.Smears were rated negative when there were no parasites in 200 fields of a thick drop smear in immersion oil 27 .

Hemoglobin determination
Hemoglobin concentrations were determined by the hematic acid method using a hemoglobinometer (HaemoCue, Switzerland) in all 679 cases in the study.

Statistical analysis of data
Data were stored in a Windows Microsoft Excel® spreadsheet and were revised for verification.The software SPSS v.18.0 for Windows (SPSS, Inc.) was used to analyze data.
An exploratory and descriptive analysis of variables (some of which were re-codified) was performed.For continuous variables, comparisons between two or more independent groups were analyzed using Mann-Whitney U and Kruskal-Wallis tests, after the failures of the presuppositions of the parametric tests (Student's t-test and ANOVA) were verified.For qualitative variables, the Pearson Chi-Square test was used.A value of p<0.05 was considered significant.For estimation of the prevalence of P. falciparum infection with a 95% confidence interval (CI), the Agresti-Coull, Wilson or Jeffreys methods were used, as recommended 28 .As described below, simple and multiple logistic regression models were also used 29,30 .To identify a simpler model, a Backward/Wald procedure was also used to automatically discard non-significant variables.
Simple and multiple logistic regressions were used to analyze potential risk factors associated with the presence of ongoing P. falciparum infection.To investigate the association between the several independent variables and malaria, we began by performing simple logistic regressions with each independent variable.
Next, we applied multiple regression models to control possible confusion.Variables exhibiting statistically significant associations (p<0.05) or with important epidemiological meanings were included.These co-variables were kept in models, independent of their significance, in univariate analysis due to their possible relevance in the final results; thus, we could analyze their possible influence when considered together with the other variables.The first model included the independent variables treatment, age, anemia (with and without infection), parity and malaria during pregnancy; the second model also included gestational age.

Ethical considerations
This study was approved by the Ethics Committee of the Faculty of Medicine of Agostinho Neto University, Luanda, Angola.

Population study
Data were collected from a total of 679 pregnant women who attended ANC at HGEAN.The demographic and obstetric profiles and the parasitological-hematological data were collected from the participants.These profiles are summarized in Table 1.Briefly, the ages of the women ranged between 12 and 42 years (mean= 25.7 years and median= 25 years); most women (66.1%) had a low schooling level, most were multiparous (74.1%), and more were in the second trimester of pregnancy (46.2%) than were in either the first or third trimesters.
Optical microscopy performed at the time of the study showed mono-infection by P. falciparum in 10.9% (n=74) of women (n=679), with a 95%CI between 8.8% and 13.5%.Distribution according to gestational age was as follows: 1 st trimester=9.7%(n=18), 2 nd trimester = 11.5% (n=36%), and 3 rd trimester = 11.1% (n=20).Details on parasitemia and the stratification of parity and gestational age are summarized in Table 2.The prevalences of P. falciparum (expressed in percentages) in primiparous (11.9%) and multiparous (10.5%) women were not significantly different.Additionally, the densities of parasites between primiparous and multiparous women were not significantly different.The rates of infection according to pregnancy trimester (expressed in percentages) were higher in the second and third trimesters (though these rates were not statistically significant).The mean and standard deviation of hemoglobin concentration in the studied samples was 11.1 ± 0.07g/dL, and the median was 10g/ dL.When comparing the median values of hemoglobin concentration between P. falciparum-infected (9.9g/dL) and non-infected (11.1g/dL) pregnant women,a highly significant difference was identified (p<0.001).It is notable that the values of serum hemoglobin in noninfected women had a tendency for greater dispersion (Figure 1).
A significant association (p<0.001) was found between anemia (defined as a hemoglobin concentration < 11g/dL) and the presence/ absence of P. falciparum; among the 339 anemic women, 56 (16.5%) were infected with P. falciparum, whereas among the 337 non-anemic women, only 17 (5%) were infected (Table 3).Only eight (1.2%) women fell into the category of severe anemia (< 7g/dL), and none had malaria.For this reason, logistic regression recorded some numerical instabilities; in later analysis, we only considered categories with or without anemia, independent of its severity.We also observed that anemia was associated with gestational age and treatment.
When women were classified as primiparous or multiparous, both primiparous and multiparous women infected with P. falciparum showed a (non-significant) tendency for lower values DISCUSSION   of circulating hemoglobin than did women who were not infected with P. falciparum.When comparing the median values of hemoglobin levels and types of treatment, significant differences were identified (p<0.001,Kruskal-Wallis test).Multiple comparisons showed that hemoglobin levels in women treated with sulfadoxine-pyrimethamine differed significantly from both untreated women and women treated by other means.When classified according to their current state of infection, this pattern of differences between treatments was maintained in non-infected women (p<0.001)(Figure 2; blue).In infected women, there were still significant differences (p=0.023) between treatments (Figure 2; green); there were differences in hemoglobin concentrations only between sulfadoxine-pyrimethamine-treated and untreated women.
When a multiple logistic regression model was applied that included the variables treatment, age, anemia, parity, malaria during pregnancy and gestational age, only anemia and malaria during pregnancy showed significant associations with current malaria.
Infection was significantly more common in women with anemia than in those without [Relative risk estimated by odds ratio (OR) = 4.03, 95% CI 2.23-7.27].Women who had previous malaria during pregnancy also had a higher risk of current malaria [OR = 4.86, 95% CI 1.45-16.25].
Finally, when a Backward/Wald-like procedure was used, we ended in stage 5, including only anemia and malaria during pregnancy in the models, thus strengthening the association of these variables with current malaria.
Malaria is an avoidable cause of maternal and fetal mortality and morbidity; however, the international community is far from implementing the necessary interventions to drastically reduce the impact of this disease in the countries where it is endemic.Early detection and timely and adequate treatment are essential requirements for reducing the incidence of deaths and must be the main components of anti-malaria intervention in underdeveloped countries [6][7][8] .This study sought to characterize the pattern of malariarelated morbidity during pregnancy at Augusto Ngangula Specialized General Hospital in Luanda, Angola, an urban area rated as having stable transmission by the Department of Endemics Control.The prevalence of malaria during pregnancy was 10.9% (95% CI 8.8%-13.5%),which agrees with the prevalence in African countries that are characterized as having intense transmission of malaria 8,17,21 .This study showed that there was no association between malaria and parity, as has also been reported by other studies performed in areas with intense transmission of malaria 31 .In areas where transmission is high, the level of acquired immunity is significant, and primiparous women are the most affected 17,21 ; however, our study found a larger affected number among multiparous women.In this study, the ages of infected and non-infected women were not significantly different.
This agrees with some data published in the literature 18 ; however, other studies performed in Sub-Saharan Africa have reported a significant association between maternal age and malaria during pregnancy 17,32 .Women who attended antenatal care during the second and third trimesters of pregnancy exhibited higher prevalences of malaria infection.This agrees with several studies that have reported the second trimester, along with the beginning of the

FIGURE 2 -
FIGURE 2 -Concentrations of serum hemoglobin in pregnant women as a function of treatment and presence of malaria.Star: possible extreme outliers; open circle: possible outliers.

FIGURE 1 -
FIGURE 1 -Concentrations of serum hemoglobin in pregnant women with and without malaria.Star: possible extreme outliers; open circle: possible outliers.

Campos
PA et al -Plasmodium falciparum infection in pregnant womenHemoglobin g

TABLE 2 -Rate of Plasmodium falciparum prevalence in women who attended the Augusto Ngangula SpecializedGeneral Hospital (HGEAN), according to the demographic and obstetric variable categories.
*Value of p according to the Pearson Chi-Square test.

TABLE 3 -Anemia among women as a function of gestational age and treatment.
*Value of p according to the Pearson Chi-Square test.