Placental but Not Peripheral Plasmodium falciparum Infection During Pregnancy Is Associated With Increased Risk of Malaria in Infancy

Summary

Malaria during pregnancy causes anemia and low birthweight, but in utero exposure to malaria antigens may also have long-lasting impact on infant immunity. We demonstrate that placental malaria infection is associated with increased risk of malaria infection in infancy.

Abstract

Pregnancy-associated Plasmodium falciparum infection impacts the health of mothers and newborns, but little is known about the effects of these infections on infant susceptibility to malaria. We followed 473 mother-infant pairs during pregnancy and through 2 years of age. We observed that children born to mothers with placental malaria, but not those born to mothers with peripheral infection without evidence of placental sequestration, had increased risk of malaria during the first year of life compared with children born to mothers with no malaria during pregnancy. Malaria infections with placental sequestration have long-lasting impact on infant susceptibility to malaria infection.

It is estimated that 125 million pregnancies occur each year in areas of malaria transmission [1] and that pregnancy-associated malaria is responsible for up to 200000 infant deaths per year [2]. However, the impact of pregnancy-associated malaria on infant risk of malaria has not been clearly characterized.

Recently, observational studies suggest that infants born to mothers who have malaria infection during pregnancy are at an increased risk of clinical malaria and Plasmodium falciparum infection early in life [3–6]. Conclusions from these observational studies of infant malaria and fetal malaria exposure are limited, most notably because they only evaluate malaria exposure at delivery and cannot control for shared risk factors. Increased risk of malaria infection in the infant may be due to environmental factors that also cause the increased risk in the mother or a common genetic basis for altered susceptibility. In a recent evaluation of a large cohort of mothers and infants, maternal malaria infection during pregnancy was not associated with infant risk of malaria after controlling for exposure to malaria-infected Anopheles bites [7].

Malaria in pregnant women may be restricted to infection in the periphery if parasites do not sequester or treatment rapidly clears placental infection, or it may result in prolonged placental infection that is detectable in placental tissue at delivery. Although women with peripheral malaria infection likely have similar exposure to malaria infection to women who have placental infection, the exposure of the fetus to malaria antigens will be higher in the presence of placental infection compared with infection that is restricted to the periphery. We used detailed antenatal data collected as part of a clinical trial to identify women with peripheral or placental malaria infection. We hypothesized that placental infection, but not peripheral malaria infection, would be associated with increased risk of malaria in the infant.

METHODS

Population

The data presented here are from a clinical trial of preventing malaria in pregnancy (ClinicalTrials.gov identifier: NCT01443130) and a longitudinal infant follow-up substudy of the clinical trial. These studies took place in Ndirande township in Blantyre, Malawi. All women in the clinical trial were human immunodeficiency virus negative, in their first or second pregnancy, and followed from 20 to 26 weeks’ gestation until delivery. As part of the clinical trial, the participants were randomized to receive sulfadoxine-pyrimethamine intermittent preventive therapy (IPT), chloroquine IPT, or chloroquine prophylaxis during pregnancy.

Study Procedures

Women had regularly scheduled examinations (monthly increasing to weekly by the expected delivery date) and were also encouraged to visit the clinic whenever ill. Infants had quarterly examinations and were encouraged to visit the clinic whenever ill. Medical history and bed net use were recorded, and physical examination, including axillary temperature, was performed at each maternal and infant visit. Finger-prick filter papers were collected at every visit regardless of symptoms, from the cord blood and from the placenta. Dried blood spots underwent quantitative real-time polymerase chain reaction (qPCR) for P. falciparum 18S ribosomal RNA to detect malaria infection. Extraction and PCR protocols are described online [8]. If the subject had signs or symptoms suggestive of malaria, thick and thin blood smears were prepared and read in real time as previously described [9]. At delivery, a full-thickness biopsy of the placenta was collected for histological examination.

Ethical approval was obtained from the University of Malawi College of Medicine’s Research Ethics Committee and the University of Maryland Baltimore Institutional Review Board. Written informed consent was obtained from all participants before conducting any study-related activities.

Definitions

Malaria infection was defined as parasites detected by qPCR, regardless of symptoms. A new malaria infection was defined as either the first detected malaria infection or a malaria infection following a previously negative malaria test. Only new infections were included in the analyses. Placental malaria was defined as the presence of malaria hemozoin pigment in the placenta or parasites detected by qPCR in the placenta. Maternal peripheral malaria was defined as a qPCR-positive filter paper collected at any point during pregnancy. Clinical malaria was defined as parasites detected on blood smear by light microscopy accompanied by symptoms including at least 1 of the following: axillary temperature ≥37.5°C measured at the clinic, history of symptoms of malaria in the previous 48 hours including fever, headache, myalgia, vomiting, or weakness. A new febrile episode was defined as either the first febrile episode or a febrile episode following a visit with a normal temperature and without a history of fever.

Statistical Analyses

Data analysis was performed using Stata version 12.0 software (StataCorp, College Station, Texas). Graphs were produced using GraphPad Prism version 5.01 for Windows (GraphPad Software, San Diego, California). To measure the association between maternal malaria status and infection, clinical malaria, or fever, odds ratios were calculated using logistic regression. Multivariate logistic regression model controlled for the covariates of maternal age, gestational age at delivery, and clinical trial arm. To measure the association between maternal malaria status and cumulative incidence of infection, clinical malaria, or fever in infancy, we calculated the relative rate ratio. Time to malaria or fever in infancy was modeled using a Nelson-Aalen cumulative hazard estimate and groups were compared using a log-rank test. Children were censored from time-to-event analysis if they missed 2 quarterly visits.

RESULTS

Demographics

Enrollment occurred between February 2012 and April 2014. The infant follow-up concluded in August 2016. We enrolled 473 mother-infant pairs for infant follow-up. Among the mothers, 67 (14.2%) had placental malaria, 63 (13.3%) had a peripheral infection detected during pregnancy but no placental malaria, and the remaining 343 (72.5%) had no evidence of malaria at any point during pregnancy. Among the placentas that were examined, 55 had hemozoin pigment only, 9 had both hemozoin pigment and parasites, and 3 had parasites only. Among those women with placental malaria, 38 (56.7%) had peripheral malaria at first antenatal visit. Among those with peripheral infection only, 33 (52.4%) had peripheral infection at the time of first antenatal visit. There was no significant difference in maternal age, gravidity, gestational age at delivery, or clinical trial treatment arm between the 3 in utero exposure groups. The infants were followed for a median duration of 415 days, including 241 infants followed for at least a year and 168 for a full 2 years.

Children Born to Mothers With Placental Malaria Have Increased Risk of Malaria During Infancy

Children born to mothers with placental malaria had significantly higher odds of malaria infection (odds ratio [OR], 2.7; 95% confidence interval [CI], 1.1–6.7) and clinical malaria (OR, 4.1; 95% CI, 1.3–13.1) in the first year of life (Table 1) and throughout the 2-year study period (Supplementary Table 1) compared to infants born to mothers without malaria infection. The associations did not change after adjustment for potential confounders (Table 1 and Supplementary Table 1). Children born to mothers with placental malaria also experienced a higher relative rate ratio of infection (1.6; 95% CI, 1.0–2.6) and clinical malaria (2.3; 95% CI, 1.1–4.8) (Supplementary Table 2) and had a significantly shorter time to first malaria infection and first clinical malaria episode in infancy compared to children born to mothers with no malaria during pregnancy (Figure 1).

Table 1.

Odds of Malaria During the First Year of Life

	Placental Malaria (n = 29)			Peripheral Malaria (n = 28)			No Maternal Malaria (n = 184)
	No. (%)	OR (95% CI)	P Value	No. (%)	OR (95% CI)	P Value	No. (%)	OR
Infection	8 (28)	2.7 (1.1–6.7)	.04	5 (18)	1.5 (.5–4.4)	.44	23 (13)	ref
Infection adjusted		2.5 (1.0–6.3)	.06		1.5 (.5–4.4)	.45		ref
Clinical malaria	5 (17)	4.1 (1.3–13.1)	.02	2 (7)	1.5 (.3–7.3)	.62	9 (5)	ref
Clinical malaria adjusted		3.9 (1.2–13.0)	.03		1.4 (.3–7.0)	.66		ref
Nonmalarial fever	17 (59)	1.1 (.5–2.4)	.83	12 (43)	0.6 (.3–1.3)	.18	104 (57)	ref
Nonmalarial fever adjusted		1.2 (.5–2.6)	.72		0.6 (.3–1.3)	.17		ref

Abbreviations: CI, confidence interval; OR, odds ratio. Bolded values indicate statistical significance

Figure 1.

Children born to mothers with placental but not peripheral malaria had a significantly shorter time to malaria infection and clinical malaria in infancy. Time to event analysis was modeled using Nelson-Aalen hazard estimator. Differences between groups are calculated using log-rank test. Children born to mothers with placental malaria are depicted in red. Children born to mothers with peripheral malaria are depicted in green. Children born to mothers with no malaria are depicted in blue. All comparisons are made to children born to mothers with no malaria during pregnancy. Time to malaria infection analysis: P < .01 for placental malaria and P = .54 for peripheral malaria. Time to clinical malaria analysis: P < .01 for placental malaria and P = .76 for peripheral malaria. Time to nonmalarial fever analysis: P = .70 for placental malaria and P = .07 for peripheral malaria.

Children Born to Mothers With Peripheral Malaria During Pregnancy Do Not Have an Increased Risk of Malaria During Infancy

Children born to mothers with peripheral malaria during pregnancy had similar odds of malaria infection and clinical malaria during the first year of life and over the course of the study as children born to mothers with no malaria. Similar results were obtained in the multivariate model. They also had similar time to first infection or clinical episode (Table 1). However, children born to mothers with peripheral malaria did have an increased relative rate ratio of clinical malaria during infancy (2.1; 95% CI, 1.0–4.5) (Supplementary Table 2).

Nonmalarial Fever in Infancy

Children born to mothers with placental malaria or peripheral malaria had the same odds of nonmalarial fever during the first year of life as children born to mothers with no malaria during pregnancy (Table 1, Figure 1).

DISCUSSION

Placental malaria, but not peripheral malaria infection during pregnancy, is associated with a statistically significant increased risk of malaria infection and disease during infancy. To our knowledge, this is the first report in the literature that has distinguished placental from peripheral infection and identifies placental sequestration as the key risk factor associated with increased risk of malaria infection in infancy. Our results corroborate the association between placental malaria and increased risk of malaria in infants found in several studies using microscopy-detectable malaria as the primary outcome [4, 10–12], but somewhat conflict with work by Mutabingwa et al [13] in which children born to primigravid women with placental malaria were protected from malaria during infancy. That study limited the definition of placental malaria to active infection, whereas the majority of the mothers in our study had past placental infection. A combination of timing of infection during gestation, duration of infection, and antigen load modifying fetal immunity likely explains our different observations.

We saw no association between placental malaria and time to first nonmalarial fever, suggesting that immune alterations from placental infection are malaria specific. Rachas et al [14] have demonstrated an association between malaria during pregnancy and nonmalarial febrile illness, using more-intensive active case detection of fever and a less-sensitive method for detecting parasites. The impact of in utero malaria exposure on susceptibility to nonmalarial infections requires further investigation.

One limitation of our study is that almost all cases of peripheral infection were present at the time of enrollment; few occurred after mothers began their antenatal care, and few cases of active placental infection were present at the time of delivery. Thus, we could not examine the effect of timing of maternal peripheral or placental infection on infant risk of malaria. Another limitation is the low number of malaria infections seen in our study, which potentially limited our ability to detect differences in this group.

Our observations suggest that malaria antigen load in utero may be responsible for increased susceptibility. It is possible that infection limited to the peripheral blood represents lower exposure to infected Anopheles bites than placental infection. However, this has never been biologically proven. That there were similar rates of peripheral infection at enrollment between the peripheral only compared to placental malaria infection also suggests that the groups are similar with respect to exposure. Further studies, especially focusing on comparing immunological outcome of infants born to mothers with similar risk of malaria infection but with different in utero fetal exposures, are warranted.

In summary, we have demonstrated a significant association between placental, but not peripheral malaria, and malaria in infancy using molecular techniques. We observed this significantly increased risk of malaria during infancy in children born to mothers with malaria, despite the fact that the vast majority of placental malaria cases in our cohort were past infections without evidence of parasites at the time of delivery. To maximize infant health, it will be critical to prevent all cases of placental malaria. Moreover, treating an infection during pregnancy is not sufficient to prevent adverse outcomes during infancy. As we have previously argued [9, 15], public health interventions need to reach women early in pregnancy or even prior to conception. Interventions to prevent pregnancy-associated malaria should include evaluation of infant risk of malaria in key outcome assessments.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We are grateful to the women who volunteered to participate and to the nurse-midwives of the Ndirande Health Centre maternity ward and antenatal clinic who supported this study. We thank Terrie Taylor, DO, for logistical support during the study and feedback. We also thank the Blantyre Malaria Project-Ndirande Clinic team members whose dedication made this study possible and who are committed to research to improve the health of Malawians.

Financial support. This work was supported by the US National Institutes of Health (grant numbers U01AI087624 and K24AI114996 to M. K. L. and T32GM092237, T32AI007540, and F30AI114195 to S. B.); the American Society of Tropical Medicine and Hygiene (Benjamin H. Kean Travel Fellowship to S. B.); the Infectious Diseases Society of America Medical Scholars Program (to S. B.); and the University of Maryland Global Health Interprofessional Council (to S. B.).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.