The reported prevalence of ZD is prone to seasonality bias as the study is cross-sectional. The study was carried out in January and February which are locally considered as transitional months to household food insufficiency. If it had been conducted in food insecurity prone months (April to July), a higher prevalence would have been expected.
According to IZiNCG, the risk of ZD is considered to be of public health concern when the prevalence of low serum zinc concentrations is greater than 20% . Hence, the study signals the public health significance of ZD in the area .
Till date two pocket studies in Ethiopia attempted to determine the prevalence prenatal ZD. Both were conducted in Sidama zone [6, 13]. According to Abebe et al., in 2004 among 99 pregnant women in third trimester, the prevalence was 72% and the mean plasma zinc was 45.6 μg/dl . According to Hambidge et al., in 2006 among 17 pregnant women in third trimester, the mean plasma zinc was 44.1 μg/dl . The studies reported severe deficiency situation compared to the current study. The variation can be explained by the time gap and seasonal difference in data collection. Further, the studies may have overestimated the problem as they only included pregnant women in the third trimester.
IZiNCG suggested similar cutoffs point (50 μg/dl) for defining ZD during the second and third trimesters . Nonetheless in the current study, parallel to studies conducted elsewhere [10, 27], the zinc level in the third trimester was significantly lower than that of the second trimester. This is also consistent with the understanding that hemodilution continues into the third trimester [7, 28]. The finding may indicate the need of having different cutoff point for the second and third trimesters.
According to IZiNCG, serum zinc level is artificially affected by CRP status, the time of day of blood sample collection and the fasting status the study subjects [2, 22]. However, in current study fasting status was not associated to serum zinc level. Similarly, the serum zinc level differences between CRP positive and negative samples and samples collected in the morning and afternoon were not as huge as observed in NHANES II . The finding signifies that in predominantly zinc deficient community, the effect of the aforementioned three factors may not be as prominent as expected in affluent community.
In the linear regression model the unexplained 46.5% of the variability in serum zinc level might be attributable to range of factors like serum copper level, serum albumin, intestinal and hemoparasites etc. which were not measured in the study.
Pregnant women from enset staple diet category were better-off in their zinc status as compared to women from maize staple diet category. The association could not be explained by dietary diversity and agro-ecological factors as these potential confounders were statistically controlled. The difference can be due to the better bioavailability of zinc in enset based diets .
Superior household economic standing enhances maternal zinc status [8, 11]. However, in this study household wealth index was not associated to zinc status. Nevertheless involvement of women in IGAs is found to have positive influence. This might be due to the reason that in Ethiopia maternal income is usually directly spent to cover household food expenditures.
Two previous studies in Nigeria [8, 30] failed to witness any definite trend on the effect of maternal education on prenatal zinc status. However, in the current study maternal education showed positive influence. Higher education status might have contributed to superior zinc status through enhancing good nutritional awareness and practice prior to and during pregnancy.
Maternal age was inversely associated to zinc status. The finding is consistent with the understanding that serum zinc level reaches peak during adolescence and young adulthood, and then declines . Other studies also reported more or less parallel finding [8, 10].
The study witnessed the deleterious effect of too many and too close births on zinc status. The finding is parallel to the knowledge that repeated pregnancies deplete maternal store. Previous studies conducted in Malawi , Nigeria  and USA  also supported the finding.
Laboratory and animal model studies indicated that zinc and iron compete for absorption in the intestinal lumen as they have similar physicochemical properties [33–35]. However, many community based studies concluded divergently [10, 36–40]. In our study, parallel to studies conducted in Nigeria , Iran  and UK [39, 40], daily iron-folate supplementation was not associated to maternal zinc status. This might be due to the reason that in the study community the intake of iron and zinc rich foods was low. Hence, competitive absorption which happens at higher concentration of the nutrients might not have taken place.
Hemoglobin and serum zinc levels were positively correlated. The association persisted after adjustments were made for potential nutritional and non-nutritional confounders. Many previous studies concluded likewise [29, 41–43]. As the study is cross-sectional, it is not viable to exclude "the chicken or the egg" causality dilemma. However, as zinc is known to participate in multiple metabolic pathways, it might have causal role in anemia.
Chronic overexertion is a predisposing factor to maternal nutritional depletion . However in current study maternal workload was not associated to zinc status. The finding might not be conclusive as level of maternal workload was measured using a relative rather than an absolute scale.
The study found negative association between frequency of coffee intake and zinc status. Coffee is known to contain tannin which can potentially inhibit zinc absorption . However, empirical evidences are lacking. Few available animal model studies concluded divergently [46, 47]. Further studies should be conducted in this direction.
Health care service related factors like distance from nearby health facility, frequency of ANC and nutrition education during pregnancy were not related to zinc status. This might be due to the reason that nutritional care is not well integrated into maternity services. In addition, the provision of nutrition education might not be effective in the absence of concurrent livelihood promotion strategies.