|Proxies and prevention of malaria in pregnancy|
||McGready R, Nosten F.
||Lancet Infectious Diseases, 2012 Sep 17. pii: S1473-3099(12)70240-2. doi: 10.1016/S1473-3099(12)70240-2.
Multiple African Countries
||There exists an overwhelming consensus that pregnant women need to be protected against malaria. Chemoprophylaxis was recommended in the past, but no safe and effective drug is available at present. In Africa, malaria prevention in pregnancy relies largely on insecticide-treated bednets (ITNs) and the partial protection provided by intermittent preventive therapy (IPTp) with sulfadoxine—pyrimethamine. In The Lancet Infectious Diseases, Thomas P Eisele and colleagues1 report on malaria prevention effectiveness during pregnancy across Africa, using 27 Demographic and Health Surveys (DHS), four Malaria Indicator Surveys, and one AIDS Indicator Survey from 25 countries. The authors conclude that the policy at present of IPTp with sulfadoxine—pyrimethamine and ITN is effective in the reduction of low birthweight and neonatal mortality. Should we be encouraged by these results?
In the analysis, full malaria prevention protection was defined as being provided by two doses or more of sulfadoxine—pyrimethamine, a household possessing one ITN or more, or both. When fully implemented, some protection against low birthweight and neonatal mortality was better than no protection: protective efficacy (PE) was 21% (95% CI 14—28) for low birthweight and 18% (4—30) for neonatal mortality. A disappointing result not highlighted by the authors, was the overall low uptake of prevention measures: two doses or more of IPTp (15•1%, IQR 5•0—43•5; in the 19 countries that had implemented it at least the year before the survey and had not abandoned the policy, as Rwanda had) and ITN (19•8%, IQR 5•0—34•6). We are clearly a very long way from full implementation.
Why is there such a mismatch between the purported overall effectiveness of these interventions and their uptake? Efforts to scale up IPTp with sulfadoxine—pyrimethamine in pregnancy have coincided with efforts to scale down the use of sulfadoxine—pyrimethamine treatment in non-pregnant populations and replace it with highly effective artemisinin combination therapy. Sulfadoxine—pyrimethamine is a failing drug combination and its efficacy in malaria prevention in pregnancy must be compromised by increasing levels of resistance2 as shown in Malawi3 and Tanzania.4 Continued use in the presence of resistance has been suggested to be harmful.5 Although prevention might need less drug effect than cure, any maternal parasitaemia will increase the risk of anaemia, the main contributor of malaria-related maternal mortality in regions of intense transmission. Other factors that could have contributed to this mismatch include the absence of pharmacokinetic data to support present dosing regimens, the initial confusion about how IPT worked, the antagonistic effects of folate supplementation, the different dosing needed for HIV-positive and HIV-negative women, and the difficulties of delivering IPTp in a timely way with a small number of antenatal care visits.6 Furthermore, IPTp was always an imperfect intervention that was not designed to provide complete maternal protection against malaria, which is surely the objective of malaria prevention interventions.
In any analysis of pregnancy outcomes, effect modifiers are likely to have a major role. The researchers acknowledge the risk that women who receive full protection against malaria might differ from those with no protection, hence the emphasis on matching when such data were available or the use of proxies when exact data were not available. Matching women for one dose or more of tetanus vaccination and iron supplementation captures women who attended antenatal care at least once. As a proxy of antenatal care, this matching will not show the frequency of visits, which are necessary to provide at least two doses of IPTp with sulfadoxine—pyrimethamine. A woman who delivers with the assistance of a health professional is more likely to have her baby's weight measured at birth than is one who has a home delivery. Yet delivery with a health professional was not included in low birthweight analysis. The proxy for accurate birthweight measurement7 was the less precise and more subjective estimated weight category by the mother,8 which in the present analysis had somewhat poor agreement (kappa statistic 0•43) with the actual weight of the baby.1 Unexpectedly, intensity of malaria transmission measured by a proxy of the parasite rate in children aged 2—10 years, was not an effect modifier on the relation between malaria prevention in pregnancy and newborn health outcomes.9 The proxy for ITN use was net availability at the household level, not whether the woman slept under the bednet while pregnant. In the analysis, ITN ownership in the absence of IPTp with sulfadoxine—pyrimethamine was not significantly associated with protection against neonatal mortality, by contrast with previous findings.10 These observations raise questions about whether the data and the interpretation will stand up to closer scrutiny.
We agree with Eisele and colleagues that real maternal protection against malaria is needed. The question remains: how?