Date Published: September 17, 2008
Publisher: Public Library of Science
Author(s): Jennifer F. Friedman, Luz P. Acosta, Charles H. King
Partial Text: In lesser-developed countries (LDCs), the causes of anaemia during pregnancy are multi-factorial, yet much of the aetiological fraction of disease is attributable to a few entities. Iron deficiency is the most common cause of anaemia among pregnant women, resulting from both dietary insufficiency of iron as well as losses through the gastrointestinal tract. These losses are largely due to hookworm infection, but schistosomiasis at higher intensities of infection may also lead to blood loss .
Brooker and colleagues have conducted a timely and informative meta-analysis examining the burden of hookworm among pregnant women. In this meta-analysis, they have included cross-sectional, observational, and randomized controlled trials to estimate the contribution of hookworm infection to maternal anaemia. All studies that provided quantitative data on both hookworm intensity of infection and hemoglobin were included, yielding 19 for inclusion. Overall, hookworm infection during pregnancy was related to a standardized mean difference in hemoglobin of −0.24 g/dL in comparing uninfected to lightly infected women, and −0.57 g/dL in comparing lightly infected to heavily infected women.
The limitations of this study are common to many meta-analyses, whereby the quality of summary estimates are driven largely by the quality of the studies included. This is particularly challenging in the case of meta-analyses of cross-sectional and observational studies where bias and confounding are more likely to play a role than in well-executed randomized controlled trials . Approaches to this issue are to either judge the quality of studies and then exclude based on a particular quality score, or include all studies with careful consideration of the potential for confounding or bias in interpretation of results. The latter approach, taken by the authors, is potentially problematic in this study given that a host of factors related to poverty may be related to both hookworm and anaemia, thus confounding this relationship. These potential confounders include an iron-deficient diet, access to iron supplementation during pregnancy, schistosomiasis, HIV, and malaria infections. If diseases of poverty related to both hookworm infection and anemia are not adjusted for in analyses, this variance may be wrongly attributed to hookworm, overestimating its effect.
The treatment trials discussed in this manuscript provide the clearest insight into implications of this meta-analysis. It is beyond the scope of this commentary to fully review those studies, but the implications include the following:
Much has been learned in the past two decades with respect to the impact of specific diseases of LDCs and their impact on human pregnancy. Next steps will require scientists with expertise in a range of diseases to evaluate more complete and integrated interventions to improve the health of both women of reproductive age and their newborns. This includes the evaluation of optimal timing of interventions. Although pregnancy remains a favorable time to capture women of childbearing age, an approach whereby women enter pregnancy in good health is ideal given the minimization of risks of interventions to the fetus, the limited period of recovery and benefit possible for interventions during pregnancy, and the greater risk for interactions with interventions that must be given during pregnancy such as therapeutics for malaria and HIV. In addition, the examination of multiple simultaneous interventions during pregnancy will be of increasing importance as more drugs become available, most of which have been examined in isolation. It is possible that drug combinations confer a greater than additive risk of side effects. Further need to simultaneously assess interventions is supported by a recent example that raised the specter that iron supplementation during pregnancy may increase malaria risk . These concerns notwithstanding, there is great potential to improve the health of pregnant women in the coming decades if industrialized nations provide much needed support, scientists from a range of backgrounds collaborate to evaluate multiple concurrent interventions, and women’s socio-political status independent of their role as mothers is improved.