Influence of the number and timing of malaria episodes during pregnancy on prematurity and small-for-gestational-age in an area of low transmission

Abstract

Background: Most evidence on the association between malaria in pregnancy and adverse pregnancy outcomes focuses on falciparum malaria detected at birth. We assessed the association between the number and timing of falciparum and vivax malaria episodes during pregnancy on small-for-gestational-age (SGA) and preterm birth.

Methods: We analysed observational data collected from antenatal clinics on the Thailand-Myanmar border (1986-2015). We assessed the effects of the total number of malaria episodes in pregnancy on SGA and the effects of malaria in pregnancy on SGA, very preterm birth, and late preterm birth, by the gestational age at malaria detection and treatment using logistic regression models with time-dependent malaria variables (monthly intervals). World Health Organisation definitions of very preterm birth (≥28 and <32 weeks) and late preterm birth (≥32 and <37 weeks) and international SGA standards were used.

Results: Of 50,060 pregnant women followed, 8221 (16%) had malaria during their pregnancy. Of the 50,060 newborns, 10,005 (21%) were SGA, 540 (1%) were very preterm, and 4331 (9%) were late preterm. The rates of falciparum and vivax malaria were highest at 6 and 5 weeks' gestation, respectively. The odds of SGA increased linearly by 1.13-fold (95% confidence interval: 1.09, 1.17) and 1.27-fold (1.21, 1.33) per episode of falciparum and vivax malaria, respectively. Falciparum malaria at any gestation period after 12-16 weeks and vivax malaria after 20-24 weeks were associated with SGA (falciparum odds ratio, OR range: 1.15-1.63 [p range: <0.001-0.094]; vivax OR range: 1.12-1.54 [p range: <0.001-0.138]). Falciparum malaria at any gestation period after 24-28 weeks was associated with either very or late preterm birth (OR range: 1.44-2.53; p range: <0.001-0.001). Vivax malaria at 24-28 weeks was associated with very preterm birth (OR: 1.79 [1.11, 2.90]), and vivax malaria at 28-32 weeks was associated with late preterm birth (OR: 1.23 [1.01, 1.50]). Many of these associations held for asymptomatic malaria.

Conclusions: Protection against malaria should be started as early as possible in pregnancy. Malaria control and elimination efforts in the general population can avert the adverse consequences associated with treated asymptomatic malaria in pregnancy.

Keywords: Gestation; Malaria in pregnancy; Preterm birth; Small-for-gestational-age; Timing.

Fig. 1

Study profile. ^*These women had malaria episodes entered retrospectively, so were not undergoing antenatal screening from ≤10 weeks’ gestation. ^**These newborns were live born, and are in a ‘grey zone’ between miscarriage and extreme preterm birth (birth at 24–28 weeks’ gestation) [53]

Fig. 2

Rates of falciparum and vivax malaria over gestational age in women who started antenatal screening before 10 weeks’ gestation (N = 19,768). Vertical dashed lines are at 14 weeks’ and 28 weeks’ gestation, indicating the end of the first and second trimesters, respectively. Hazard rates were calculated per 1000 pregnancy weeks. For the hazard of initial malaria, women were censored at the gestational age of their first episode, birth, or time last seen. For recurrent malaria, women became at risk at the gestational age of their initial episode and were censored at the gestational age of their first recurrent episode, birth, or time last seen; we did not attempt to account for a post-treatment prophylactic effect because of the strong assumptions that this would require

Fig. 3

Descriptive statistics for parasitaemia and symptoms by the gestational age at malaria detection. Orange: falciparum malaria. Blue: vivax malaria. CI confidence interval

Fig. 4

The association between the number of malaria episodes in pregnancy and small-for-gestational-age (SGA). Numbers are odds ratios with 95% confidence intervals. See Additional file 4 for a table version of this figure. The associations between the number of episodes and log_e(odds) were linear for falciparum malaria (p = 0.819) and vivax malaria (p = 0.118). SGA was missing in 2636 [5%] (no malaria 2078 [5%]; malaria 558 [7%])

Fig. 5

The association between the gestational age at falciparum or vivax malaria detection and treatment and small-for-gestational-age (SGA). See Additional file 5 for a table version of this figure, including differentiation between symptomatic and asymptomatic malaria. The reference group within each time interval is women with no malaria detected within each respective time interval. SGA was missing in 2636 [5%] (no malaria 2078 [5%]; malaria 558 [7%])

Fig. 6

The association between the gestational age at falciparum or vivax malaria detection and treatment and preterm birth. The association between malaria at 0–4 and 4–8 weeks’ gestation and very preterm birth was not estimated due to zero events in the malaria groups. See Additional file 6 for a table version of this figure, including differentiation between symptomatic and asymptomatic malaria