This study found a high prevalence of placental malaria and low birth weight despite anti-folate use among women living in Tororo, a region of very high malaria transmission. Indeed, a very high proportion of women delivering at Tororo District Hospital received appropriate anti-malarial prophylaxis and, given the sampling technique, this is likely representative of women who experience hospital births in this region. As has been seen in other studies, a high proportion of LBW was attributable to placental malaria [1
]. Among HIV-uninfected women, placental malaria was most prevalent among primigravidae, and multigravidae were at decreased risk, as has been seen in many prior studies [25
]. This pattern was not observed in HIV-infected women, for whom gravidity did not significantly alter the risk of placental malaria. This difference in risk between HIV-infected and uninfected women has also been described previously [2
The protective efficacy of daily TS against malaria among pregnant women has not been established. As daily TS is the standard of care for all those with HIV infection in Uganda, it is not possible to randomize HIV-infected pregnant women to daily TS versus IPT-SP. Rather, the prevalence of placental malaria was compared between HIV-uninfected women on IPT-SP and HIV-infected women on daily TS to estimate efficacy. The prevalence of placental malaria was similar between HIV-infected women on daily TS and HIV-uninfected women on IPT-SP. Numerous studies have demonstrated that HIV infection nearly doubles the risk of placental malaria [5
]. Observational studies and trials suggest that monthly IPT-SP regimens can decrease this risk to levels seen among HIV-uninfected women receiving 2-dose IPT [30
]. These data suggest that daily TS can, similarly, decrease the risk of placental malaria in the setting of HIV infection.
Placental malaria infection serves as the primary surrogate outcome for evaluating malaria in pregnancy because it is associated with adverse maternal and fetal outcomes [32
]. One of the most important outcomes of placental malaria is LBW. The most common method of diagnosing placental malaria remains thick blood smear of placental blood, but many studies have used PCR of placental blood. Prior studies correlating PCR+ and blood smear+ placental malaria with LBW have produced conflicting results [33
]. Although PCR is more sensitive than placental blood smear for malaria detection, the clinical predictive value of placental malaria diagnosed by PCR remains unclear, especially in the setting of HIV [35
]. In this study, placental malaria determined by either PCR or smear of placental blood was associated with LBW. Further, these data suggest that placental malaria detected by blood smear may be the most clinically relevant definition. Specifically, a statistically significant association between PCR+/smear- placental malaria and LBW was found only among HIV-uninfected women, although this study was underpowered to detect such an association among HIV-infected women. Indeed, the underlying mechanism leading to PCR+/smear- placental malaria in this population receiving anti-folate prophylaxis remains unknown. PCR+/smear- placental malaria may be seen simply due to increased sensitivity of PCR compared to blood smear for identification of low-level infections. A more intriguing explanation is that PCR+/smear- placental malaria may represent past P. falciparum
infection of the placenta with successful clearance of the parasite by either IPT-SP or daily TS. In this scenario, the placenta is exposed to active infection for a shorter period of time and, therefore, the harmful effects on the fetus are mitigated. If this explanation is true, the data here suggest that HIV-uninfected women on IPT-SP were more able to clear P. falciparum
infection of the placenta as their gravidity increased. On the other hand, the HIV-infected multigravidae in this study were no more likely to clear infection than primigravidae, likely due to altered immune-recognition in the setting of HIV.
One concern regarding use of IPT or daily TS is the selection of drug-resistant infections. In this study, the prevalence of markers of anti-folate resistance did not differ between HIV-infected women on TS and HIV-uninfected women on SP, although baseline prevalence of these polymorphisms is very high in Tororo [36
], limiting the ability to recognize selection of the polymorphisms by anti-folate therapy. These results suggest that the comparison of HIV-infected and uninfected women was not confounded by differences in resistance profiles in parasites infecting the two groups.
This study had several limitations. The comparable prevalence of placental malaria between HIV-infected and HIV-uninfected women may have been biased by unmeasured confounders including insecticide-treated bed net use or other protective factors associated with both risk of placental malaria and differential antenatal care between HIV-infected and HIV-uninfected women. Moreover, while available antenatal cards recorded provision of IPT and daily TS as appropriate based on HIV status, this study was unable to confirm compliance with this intervention. HIV-infected and HIV-uninfected groups were recruited at slightly different times of the year, potentially introducing sampling bias. Nonetheless, the holoendemic transmission of malaria in Tororo [37
] and the long gestational period during which placental malaria can be acquired make this sampling bias unlikely. Similarly, logistic regression models including or excluding season at the time of delivery did not alter the findings. Although the study population was restricted to women delivering in the hospital, about two-thirds of Ugandan women overall give birth outside of health facilities [38
], and 45% of HIV-infected women engaged in the CDC-run HIV treatment program in Tororo deliver at home. (Jaco Homsy, personal communication) As such, these findings may not be generalizable to those women delivering outside the hospital setting. Lastly, data on antiretroviral use were available for approximately 60% of the HIV-infected women in this study and, thus, this study could not systematically evaluate the association between immune reconstitution and risk of placental malaria.
Despite these limitations, this study provides important data on the epidemiology and clinical implications of placental malaria among HIV-infected women on daily TS delivering at a district hospital. As daily TS is increasingly becoming the standard of care for HIV-infected pregnant women in sub-Saharan Africa, these data offer a baseline against which to compare future anti-malarial interventions. Furthermore, these data suggest that placental malaria detected by blood smear is an appropriate surrogate outcome for future studies involving these populations. Larger studies investigating the clinical relevance of other measures of placental malaria among HIV-infected women on daily TS, including PCR and histopathology, are warranted. Given the time and resource-intensive nature of histopathology, it would particularly interesting to assess whether PCR+/smear- placental malaria correlates with past or chronic P. falciparum infection on histopathology in this population. Future studies should also include assessments of other sequelae of malaria in pregnancy, including maternal anaemia, late spontaneous abortion and infant mortality.