This is the first analysis of ZPP as a potential iron biomarker of malaria risk in pregnancy. Use of ZPP transition categories allowed malaria risk to be assessed concurrently with longitudinal changes in this iron biomarker. Gestational iron status is influenced by antenatal haematinic use, iron bioavailability [18
] and enteric iron losses. Under the high malaria exposure in which these women were living, a higher malaria prevalence in primigravidae than multigravidae would be expected [19
], and was observed across all ZPP transition categories, indicating the substantial influence of parity on malaria susceptibility under these malaria-endemic conditions. Almost two thirds of multigravidae and a quarter of primigravidae would be classified as iron deficient, based on either the ZPP measurement alone, and approximately 90
% were anaemic.
ZPP reflects bone marrow iron requirements [20
], and as ZPP and haemoglobin are diluted equally during pregnancy due to plasma volume expansion, the use of a ZPP/haemoglobin index avoids misinterpretation that may result from plasma volume changes. In a non-malarious area the diagnostic sensitivity and predictive value of ZPP for evaluating iron depletion and risk of anaemia in pregnancy compared favourably to those of ferritin and transferrin saturation measurements.[21
]. Expansion of plasma volume is greater than that of red cell mass and leads to a decline in haemoglobin and haematocrit [18
], with little change in average red-cell haemoglobin concentration, or MCHC [16
]. MCHC is, therefore, a useful indicator of iron deficiency although by definition it can only diagnose iron deficiency at the stage of anaemia. Based on a MCHC <32
g/dl, approximately one in five primigravidae and three in five multigravidae would be classified as iron deficient. MCHC may be a poor biomarker when calculated from microhaematocrit without frequent calibration.
In the cross-sectional analysis at first antenatal visit, malaria prevalence was higher in women with raised ZPP values for all gravidae classes except secundigravidae. At delivery, malaria prevalence (peripheral or placental blood) was also higher in woman with raised ZPP concentrations. This association in the cross-sectional study of higher malaria prevalence in women with elevated ZPP levels is the statistically most significant finding as in the longitudinal sub-group analyses sample sizes are much smaller. The association may result from inflammation increasing ZPP concentration, rather than indicating an association of increased malaria risk with maternal iron deficiency. This is because, based on MCHC, multigravidae at first antenatal visit who had low MCHC values also had lower malaria prevalence compared to those with normal MCHC, suggesting that iron deficiency may reduce malaria risk. This suggests ZPP is not a good biomarker for iron status in this population. A limitation of ZPP as an indicator of pregnancy iron status relates to factors known to elevate its concentration, including infection [22
], impaired iron utilization in chronic inflammation [23
], chronic malarial haemolysis sufficient to result in iron demand in excess of the iron transport system’s ability to deliver iron to the red cell, and drug intake, for example with cotrimoxazole [20
]. All of these factors are common in women living under conditions of high malaria exposure and some of these factors may explain the association in the present study of raised ZPP values with increased malaria prevalence. Raised bilirubin concentration, secondary to haemolysis could also interfere with haematoflourometer readings when whole blood samples were used [24
], and use of washed red cells would have been preferable. Very high levels of zinc protoporphyrin may be toxic to the malaria parasite [26
], and in very anaemic Tanzanian children reduced malaria risk was reported with raised ZPP concentration [11
In the present study gravidae who maintained normal ZPP concentrations at first antenatal visit and at delivery (i e, possibly iron replete throughout pregnancy), had higher delivery malaria prevalence than those who transitioned from raised to normal ZPP values (i e, became iron replete in late pregnancy). This might suggest that continuous iron repletion during gestation was associated with higher delivery malaria risk, and not late gestational iron repletion. But women with continuous iron repletion did not have higher delivery malaria prevalence than women who transitioned to raised ZPP values at delivery (i e, were iron deficient late in pregnancy, or throughout pregnancy), suggesting that if ZPP is accurately indicating maternal iron status then it is not discriminating malaria risk, as both chronically iron-replete and chronically iron-deficient women were at equivalent risk of malaria at delivery. However, if malaria-related inflammation raises ZPP concentration, independent of maternal iron status, then gestational transition to lower ZPP values would be associated with a reduction in malaria prevalence at delivery compared to women who show no change in ZPP concentration, as shown in this study.
In regression analysis, peripheral or placental malaria was positively associated with primigravidae and HIV infection, and lower malaria risk was associated with uptake of one or more doses of IPTp-SP. These associations would be expected and their magnitude provides a useful gauge against which to assess malaria risk estimates related to ZPP categories. Transition to normal ZPP concentration was associated with reduced risk for malaria parasitaemia in peripheral or placental blood (Table ). These findings are consistent with data from the cross-sectional analysis at first antenatal visit and delivery and suggest that malaria parasitaemia elevates ZPP concentration. If antiretroviral therapy were available, as was not the case at the time of this study, then the host inflammatory response may be less marked, and as a consequence ZPP as a biomarker of iron status might have better validity.
Multivariate analysis of malaria at delivery
A proportion of women in this population will have raised ZPP concentrations solely related to maternal iron deficiency and not related to inflammation. A limitation of this analysis is that without additional sensitive serum biomarkers of iron status it is not possible to discriminate this group from those with raised ZPP values due to inflammation alone. For this reason it is not possible to determine from ZPP measurements alone whether iron deficiency or repletion alters malaria susceptibility in these pregnant women. The proportions of women with raised ZPP values due to iron deficiency, or due to inflammation will also vary between gravidae as iron deficiency is generally more frequent in multigravidae [27