Study participants
Of the 300 HIV-negative women enrolled in the parent trial, 289 (96.3%) were followed to delivery. Of these, nine (3.1%) were excluded for incomplete analyses for placental malaria including seven who did not have placental biopsies taken for histopathology and an additional two with missing results for placental blood analyses. Five women with placental malaria did not have documented evidence of parasitaemia prior to delivery and, given the inability to determine timing or number of infections, they were excluded from the analysis. Of the remaining 275 women, 228 (82.9%) had evidence of infection with malaria parasites during pregnancy and were included in the analyses. Of these, 127/228 (55.7%) did not have evidence of placental malaria and 101/228 (44.3%) had placental malaria (Fig. 1). The number of participants diagnosed with placental malaria via each of the three methods used are displayed in Figure 2. Histopathology, which also detects past placental malaria, was the most sensitive of the three methods.
The median age of study participants was 21.1 years, and 38.2 percent were primigravid. Demographic characteristics and obstetrical outcomes of the study participants stratified by placental malaria status are presented in Table 1. Women with placental malaria were younger (19.0 years (17.9-21.3) vs. 22.7 (19.9-25.7); p<0.01) (median (IQR)) and more likely to be primigravid (63.4% vs. 18.1%, p<0.01). Those with placental malaria also enrolled (and therefore initiated intermittent preventive therapy) at a slightly later GA (15.3 weeks (14.0-17.6) vs. 14.7 (13.6-15.9) wks, p<0.01) and were significantly more likely to have received sulfadoxine-pyrimethamine treatment as compared to dihydroartemisinin-piperaquine (48.5% versus 31.5%, p≤0.01). All of the included twin gestations (n=4) had evidence of placental malaria in at least one placenta (Table 1). As previously reported [10], women with placental malaria delivered at an earlier GA (39.3 weeks (38.0-40.3) vs. 40.0 (39.0-40.9), p<0.01) and were more likely to deliver preterm (<37 weeks GA) (15.0% vs. 4.7%, p<0.01) with correspondingly low birth weight (Table 1). Additionally, infants from pregnancies affected by placental malaria were more likely to be SGA (25.7% vs. 15.0%, p=0.04). There were no significant differences in fetal or neonatal survival between the groups.
Table 1. Characteristics of study participants by placental malaria status
Variable
|
Placental malaria status in pregnancy
|
No placental malaria
(n=127)
|
Placental malaria
(n=101)
|
P value
|
|
Maternal age, yr
|
22.7 (19.9-25.7)
|
19.0 (17.9-21.3)
|
<0.001
|
|
Primigravid
|
23 (18.1)
|
64 (63.4)
|
<0.001
|
|
BMI, kg/m2
|
20.6 (19.0-22.2)
|
21.0 (19.7-22.7)
|
0.08
|
|
Twin gestation*
|
0 (0.0)
|
4 (4.0)
|
0.04
|
|
GA at enrollment, wks
|
14.7 (13.6-15.9)
|
15.3 (14.0-17.6)
|
0.003
|
|
Household wealth index
Lowest tertile
Middle tertile
Highest tertile
|
46 (36.2)
42 (33.1)
39 (30.7)
|
37 (36.6)
37 (36.6)
27 (26.7)
|
0.80
|
|
Intermittent preventive therapy drug
sulfadoxine-pyrimethamine
dihydroartemisinin-piperaquine
|
40 (31.5)
87 (68.5)
|
49 (48.5)
52 (51.5)
|
0.01
|
|
GA at delivery, wk
|
40.0 (39.0-40.9)
|
39.3 (38.0-40.3)
|
<0.001
|
|
Preterm birth <37 wks GA
Very preterm birth <32 wks GA
|
6 (4.7)
1 (0.8)
|
15 (14.9)
4 (4.0)
|
0.01
0.17
|
|
Low birth weight**
|
10 (7.9)
|
19 (18.8)
|
0.01
|
|
Small for gestational age†
|
19 (15.0)
|
26 (25.7)
|
0.04
|
|
Stillbirth
|
0 (0.0)
|
2 (2.0)
|
0.09
|
|
Neonatal demise‡ (N=224)
|
4 (3.2)
|
0 (0.0)
|
0.13
|
|
BMI = body mass index, GA = gestational age, SGA = small for gestational age
* Information missing for 1 patient
**Defined as birthweight <2500g.
† Defined as birthweight <10%ile for GA based on East African growth standards.15
‡ Data unavailable for four patients including two in the peripheral malaria group and 2 in the placental malaria group.
Data are presented as median (interquartile range) or n (%). Wilcoxon Rank Sum and Chi-squared or Fischer Exact tests were used to compare nonparametric continuous variables and proportions, respectively. Pregnancies with peripheral malaria only (without placental malaria) and pregnancies with placental malaria are the comparison groups for the P values.
Longitudinal measures of P. falciparum infection
Timing of initial documented infection and first episode of symptomatic malaria in women with and without placental malaria are depicted graphically in Fig. 3. Characteristics of infection are compared by placental malaria status in Table 2. Any symptomatic malaria at any time during pregnancy was more common in patients with placental malaria than those without placental malaria (29.7% vs. 18.1%, p=0.04). Of those with at least one episode of symptomatic malaria, those with placental malaria had their first episode of symptomatic malaria at a later GA as compared to those without placental malaria (24.0 weeks (18.9-28.3) vs. 18.3 (16.1-20.0), p≤0.01). GA at first documented infection (including symptomatic and asymptomatic) was not statistically different between the two groups (p=0.15) (Table 2). Primigravidity, gestational age at enrollment, any symptomatic malaria during pregnancy and total number of times parasitemia was detected in pregnancy were all controlled for in multivariate analyses because they were significant in univariate analyese. Gestational age at first documented infection was included in multivariate modeling rather than GA at first symptomatic infection to allow for examination of timing of infection without overfitting for symptomatology. In multivariate analyses, primigravidity, later GA at enrollment and increasing frequency of infections detected during pregnancy remained strongly associated with development of placental malaria (aOR 9.06, 95% CI 4.38-18.71, p<0.001; aOR 1.24, 95% CI 1.02-1.50, p=0.03; aOR 2.21, 95% CI 1.67-2.93, p<0.001, respectively) (Table 3).
Table 2. Measures of malaria infection, by placental malaria status
Variable
|
Placental malaria status in pregnancy
|
No placental malaria
(n=127)
|
Placental malaria
(n=101)
|
P value
|
Positive LAMP at enrollment*
|
84 (66.1)
|
76 (75.3)
|
0.14
|
Number of times parasitaemia detected per pregnancy
|
2.0 (2.0-3.0)
|
3.0 (2.0-5.0)
|
<0.001
|
Any symptomatic malaria during pregnancy
|
23 (18.1)
|
30 (29.7)
|
0.04
|
GA at first detection of symptomatic malaria, wk
|
N=23
18.3 (16.1-20.0)
|
N=30
24.0 (18.9-28.3)
|
0.01
|
GA at initial documented parasitaemia, wk
|
15.9 (14.3-17.9)
|
16.1 (14.7-18.6)
|
0.15
|
GA = gestational age, LAMP = loop mediated isothermal amplification. Data are presented as median (interquartile range) or n (%). Wilcoxon Rank Sum and Chi-squared tests were used to compare nonparametric continuous variables and proportions, respectively.
*No patients had symptomatic malaria at the time of enrollment.
Table 3. Predictors of placental malaria
Variable
|
aOR
|
95% CI
|
P value
|
Primigravidity
|
9.06
|
4.39-18.71
|
<0.001
|
GA at enrollment, wks
|
1.24
|
1.02-1.50
|
0.03
|
Any symptomatic malaria during pregnancy
|
1.88
|
0.82-4.32
|
0.14
|
GA of first documented infection, wks
|
1.06
|
0.97-1.16
|
0.19
|
Total number of times parasitaemia detected in pregnancy
|
2.21
|
1.67-2.93
|
<0.001
|
GA = gestational age. Multivariate binary logistic regression using placental malaria as the main outcome were performed and were adjusted for primigravidity, GA at study enrollment, any symptomatic malaria during pregnancy, GA of first documented infection, and total number of malaria infections (symptomatic and/or asymptomatic) in pregnancy. Variables were kept continuous where possible.
Subgroup analyses
In primigravidas, each additional week of GA at enrollment was associated with an increase in the odds of placental malaria (aOR 1.84, 95% CI 1.21-2.78, p<0.01), but this association was not seen in multigravidas. In multigravid women, each additional week of gestation at initial detected infection was associated with a 1.12-fold increase in odds of placental malaria (aOR 1.12, 95% CI 1.02-1.24, p=0.02). This trend was not seen in primigravidas. Total number of times infection was detected in pregnancy remained strongly associated with placental malaria in both primigravid and multigravid women (Table 4 and Fig. 4).
Table 4. Predictors of placental malaria stratified by gravidity
Variable
|
Primigravid (N=87)
|
Multigravid (N=141)
|
aOR
|
95% CI
|
P value
|
aOR
|
95% CI
|
P value
|
GA at enrollment, wks
|
1.84
|
1.21-2.78
|
<0.01
|
1.12
|
0.88-1.44
|
0.36
|
Any symptomatic malaria in pregnancy
|
0.57
|
0..14-2.38
|
0.44
|
2.89
|
1.05-7.95
|
0.04
|
GA of first documented infection, wks
|
0.82
|
0.67-1.01
|
0.06
|
1.12
|
1.02-1.24
|
0.02
|
Total number of times parasitaemia detected in pregnancy
|
2.70
|
1.44-5.06
|
<0.01
|
2.30
|
1.65-3.23
|
<0.001
|
GA = gestational age. Multivariate binary logistic regression using placental malaria as the main outcome were performed and were adjusted for GA at study enrollment, GA of first documented infection, and total number of malaria infections (symptomatic and/or asymptomatic) in pregnancy. Variables were kept continuous where possible.
Multivariate analyses stratified by presence of any symptomatic malaria or only asymptomatic parasitemia was also performed. In women with only asymptomatic parasitemia, primigravidity was associated with over a 12-fold increase in odds of developing placental malaria (aOR 12.07, 95% CI 5.24-27.81, p<0.001), while there was no statistically significant association between primigravidity and placental malaria among those with any symptomatic malaria. In both groups, total number of infections in pregnancy remained highly associated with placental malaria (Table 5).
Table 5. Predictors of placental malaria stratified by presence of symptoms
Variable
|
Any Symptomatic Malaria (N=53)
|
Only Asymptomatic Parasitemia (N=175)
|
aOR
|
95% CI
|
P value
|
aOR
|
95% CI
|
P value
|
Primigravid
|
2.63
|
0.54-12.76
|
0.23
|
12.07
|
5.24-27.81
|
<0.001
|
GA at enrollment, wks
|
1.36
|
0.89-2.08
|
0.15
|
1.26
|
1.00-1.58
|
0.05
|
GA of first documented infection, wks
|
1.13
|
0.99-1.23
|
0.07
|
1.01
|
0.90-1.13
|
0.83
|
Total number of times parasitemia detected in pregnancy
|
2.43
|
1.37-4.30
|
<0.01
|
2.13
|
1.53-2.97
|
<0.001
|
GA = gestational age. Multivariate binary logistic regression using placental malaria as the main outcome were performed and were adjusted for GA at study enrollment, GA of first documented infection, and total number of malaria infections (symptomatic and/or asymptomatic) in pregnancy. Variables were kept continuous where possible.
Lastly, multivariate analyses were stratified by GA at initial documented infection. These categories were selected based on timing of enrollment in the study and subsequent testing: <14 weeks (enrollment in first trimester), 14-19.9 weeks (enrollment in second trimester), and ≥20 weeks gestation (infection after study enrollment). Primigravidity was strongly associated with placental malaria in women whose initial documented infection occurred prior to 14 weeks (aOR 29.30, 95% CI 3.95-217.5, p=0.001), as well as in women whose first infection occurred between 14 and 19.9 weeks (aOR 10.63, 95% CI 4.12-27.42, p<0.001). Primigravidity was not a significant predictor of placental malaria in women whose initial infection occurred at or beyond 20 weeks gestation. Total number of times infection was detected in pregnancy, however, was strongly associated with placental malaria in women whose first infection occurred at or beyond 20 weeks (aOR 5.61, 95% CI 1.59-19.78, p<0.01) as well as in women whose first infection occurred between 14 and 20 weeks gestation (aOR 2.63, 95% CI 1.80-3.87, p<0.001) (Table 6).
Table 6. Predictors of placental malaria stratified by gestational age at initial documented infection
Variable
|
<14wks (N=39)
|
14-19.9wks (N=148)
|
>=20wks (N=41)
|
aOR
|
95% CI
|
P value
|
aOR
|
95% CI
|
P value
|
aOR
|
95% CI
|
P value
|
Primigravid
|
29.30
|
3.95-217.5
|
0.001
|
10.63
|
4.12-27.42
|
<0.001
|
6.02
|
0.79-46.10
|
0.08
|
GA at enrollment, wks
|
0.78
|
0.14-4.33
|
0.78
|
1.58
|
1.19-2.09
|
0.002
|
1.14
|
0.77-1.67
|
0.51
|
Any symptomatic malaria in pregnancy
|
1.44
|
0.12-16.71
|
0.77
|
1.92
|
0.67-5.52
|
0.23
|
9.90
|
0.96-102.48
|
0.06
|
Total number of times parasitaemia detected in pregnancy
|
1.22
|
0.65-2.30
|
0.54
|
2.63
|
1.80-3.87
|
<0.001
|
5.61
|
1.59-19.78
|
<0.01
|
GA = gestational age. Multivariate binary logistic regression using placental malaria as the main outcome were performed and were adjusted for GA at study enrollment, GA of first documented infection, and total number of malaria infections (symptomatic and/or asymptomatic) in pregnancy. Variables were kept continuous where possible.