SP-IPTp did not improve overall pregnancy outcomes in our cohort from the community of Muheza, Tanzania, which is located in an area of widespread drug resistance. Instead, IPTp was associated with decreased cord hemoglobin levels and increased risk of fetal anemia.
Where parasites are susceptible to SP, IPTp reduces the prevalence of PM, maternal anemia, and LBW [18
] and increases mean maternal hemoglobin level and birth weight [19
]. A 2007 review concluded that IPTp remained beneficial in areas with high levels of SP resistance [20
], but based this conclusion on communities where the parasitologic treatment failure rate in children had reached a maximum of 26% at day 14 [20
]. By comparison, the 14-day treatment failure rate in children around Muheza, Tanzania, increased from 41% [21
] to 68% [7
] during the period of the current study, and parasite SP resistance alleles approached saturation [9
Evidence of IPTp failure exists. We recently observed that placental infections were exacerbated in women who experienced failure of presumptive treatment and remained infected at delivery [9
]. In addition, a 2008 study from Mozambique examined the effect of IPTp in the context of widespread insecticide-treated bed net use and found that IPTp had no effect on the prevalence of PM, maternal anemia, or LBW [22
]. The absence of effect was attributed to the masking benefit of bed nets but may equally have been attributable to loss of SP efficacy in that area. Other randomized controlled studies of IPTp have variably found no reduction in PM [23
], maternal anemia [25
], or LBW [13
]. We hypothesize that the failure of IPTp to improve outcomes in previous trials may be the result of accumulating resistance in various communities.
The observations from the present study suggest that IPTp may have negligible overall benefits for the community of Muheza, Tanzania. In addition, the deleterious effects of IPTp on cord blood hemoglobin levels and fetal anemia were unexpected. Fetal anemia predisposes to infancy anemia [14
], which is a risk factor for infant mortality [27
]. Furthermore, recent work from Malawi has demonstrated an association between fetal anemia and shorter time to first illness during childhood (respiratory infection, malaria, or diarrhea), as well as a higher prevalence of persistent illness and a higher cumulative incidence of morbidity [29
]. Fetal anemia may play a larger role in determining childhood morbidity and mortality than has previously been appreciated.
We found that sulfa was readily detectable in cord plasma and that levels correlated well with maternal plasma sulfa levels, suggesting considerable in utero exposure to SP. The ratio of cord-maternal sulfa that we describe is consistent with a report of SP used to treat congenital toxoplasmosis that found a mean fetal-maternal ratio of 0.97 (range, 0.65–1.16) [30
]. In a second study of pregnant ewes given sulfamethoxazole (a related sulfonamide), sulfa was concentrated in fetal plasma as a result of decreased renal clearance of the primary metabolite by the fetus [31
]. In addition, we observed an inverse relationship between cord sulfa concentration and both cord hemoglobin level and red blood cell count, supporting our hypothesis that in utero
SP exposure may suppress hematopoiesis in the fetus. Future studies are needed to explore this possibility.
This study has several limitations. First, the dhfr
quintuple mutant was ubiquitous, and nearly a third of parasites carried the additional mutation at dhps
codon 581 [9
]. This degree of drug resistance is unusually high, but it may nevertheless presage events elsewhere in Africa as drug-resistant parasites spread. Second, although a randomized placebo-controlled trial would have been the most rigorous study design to address our hypothesis, we used a cross-sectional approach. SP-IPTp is the standard of care in Tanzania, precluding a study that included a placebo control. Accordingly, our control group consisted of women who through their own actions or those of ANC staff failed to receive IPTp. Although we found and adjusted for multiple confounders, our results could be biased by unmeasured variables. Finally, our analysis of primary IPTp exposure was based on self-report. We validated our IPTp data with measured plasma sulfa levels, however, and found self-report to be very reliable.
Our data suggest that IPTp loses its benefits as drug resistance increases in an area. Models of drug resistance predict that benefits decay as drug-resistant parasites spread, eventually approaching zero. In contrast, our data suggest that continued mass administration of IPTp in an area of high resistance may move beyond zero and result in net harm. In addition, our data from previous work [9
], as well as those from mouse models of competitive facilitation [32
], imply that continued application of a failing drug regimen may exacerbate rather than alleviate infection and disease in some individuals. We hypothesize that above a threshold level of drug resistance the detrimental effects of a regimen will outweigh the beneficial effects, resulting in harm to the community. Our data suggest that Muheza is at or near such a threshold and that continued application of SP may be more harmful than beneficial.
Our findings of the potential for harm highlight a fundamental flaw in ongoing efforts to find effective new IPTp regimens, which are compared with SP [33
]. Such trials may find that new IPTp regimens confer benefits when compared with SP, but may still not be able to conclude that they are better than placebo, increasing the risk of Type I error. More generally, many studies of new interventions for the treatment of malaria or other diseases do not include a placebo control group. Instead, new interventions are evaluated against older interventions. Evidence that a new intervention is better than an older, harmful intervention does not justify the conclusion that it is better than placebo. This key implication of our data should inform future prophylaxis and treatment trials for malaria.
Once a policy of mass drug administration has been implemented, clinical equipoise makes placebo-controlled trials problematic, and the opportunity to confirm potential harm in a community is limited. Cross-sectional observational studies similar to our own should be undertaken at other sites that have varying degrees of resistance and PM prevalence. These might yield an improved model of IPTp efficacy that could predict the appropriate time to remove this failing drug regimen. Separately, this study provides the impetus for future work using placebo-controlled trials of IPTp to ensure that mass drug administration is conferring benefits and not harm.
The WHO recently concluded that priorities for future research were the identification of “when SP should be replaced with a more effective antimalarial” (p. 5) and “the efficacy of SP-IPTp against increasing prevalent quintuple pfdhr/dhps mutant P. falciparum
” [4 p. 5
]. In addition, the WHO concluded that increasing resistance demands the “evaluation of the in vivo
therapeutic and protective efficacy of SP in asymptomatic pregnant women, and its correlation with molecular markers of SP resistance, and the continuous monitoring of effectiveness of SP-IPTp” [4 p. 6
]. This study directly addresses the recommendations of the WHO and is the first to examine IPTp efficacy in an area where the SP treatment failure rate in children exceeds 50% [7
] and where molecular markers of resistance approach saturation [9
]. We find strong evidence that IPTp is failing in Muheza, arguing for urgent reevaluation of IPTp in areas of widespread resistance.