Micronutrient Deficiencies among Breastfeeding Infants in Tanzania

Infant mortality accounts for the majority of child deaths in Tanzania, and malnutrition is an important underlying cause. The objectives of this cross-sectional study were to describe the micronutrient status of infants in Tanzania and assess predictors of infant micronutrient deficiency. We analyzed serum vitamin D, vitamin B12, folate, and ferritin levels from 446 infants at two weeks of age, 408 infants at three months of age, and 427 mothers three months post-partum. We used log-Poisson regression to estimate relative risk of being deficient in vitamin D and vitamin B12 for infants in each age group. The prevalence of vitamin D and vitamin B12 deficiency decreased from 60% and 30% at two weeks to 9% and 13% at three months respectively. Yet, the prevalence of insufficiency at three months was 49% for vitamin D and 17% for vitamin B12. Predictors of infant vitamin D deficiency were low birthweight, urban residence, maternal education, and maternal vitamin D status. Maternal vitamin B12 status was the main predictor for infant vitamin B12 deficiency. The majority of infants had sufficient levels of folate or ferritin. Further research is necessary to examine the potential benefits of improving infants’ nutritional status through vitamin D and B12 supplements.


Introduction
In 2015, the global under-five mortality rate was estimated to be 41 deaths per 1000 livebirths and the neonatal mortality rate was estimated to be 18 deaths per 1000 livebirths; however, many low-and low-middle income countries have significantly higher rates [1]. In order to meet the new Sustainable Development Goals (SDGs), significant reductions in under-five mortality and neonatal mortality must be achieved [2]. Despite progress in reducing child mortality in Tanzania, neonatal and infant mortality still accounts for the majority of child deaths [3,4]. Tanzania's under-five mortality was 60 deaths per 1000 livebirths and neonatal mortality rate was 22 deaths per 1000 livebirths in 2015 [1]. Understanding the relative importance of underlying causes of neonatal and post neonatal death is important for identifying high impact interventions.

Serum Ferritin Iron Deficiency
Adults [35] <15 ng/mL Infants 3 months [35] <12 ng/mL Infants 2 weeks [36] <30 ng/mL We used SAS PROC GENMOD's log-Poisson regression with robust variances [37] to estimate the relative risk of being deficient for vitamin D and vitamin B12 based on both infant and maternal predictors for infants at two weeks and three months old. Outcomes of interest were infant vitamin D deficiency (<50 nmol/L) and infant vitamin B12 deficiency (<203 pg/mL). Risk factors included in the multivariate model were low birthweight (less than 2500 g), preterm birth (less than 37 weeks of gestation), sex, urban residence, number of antenatal care visits (<4 visits, ≥4 visits), maternal education (none, 1-7 years, >7 years), maternal age (<20 years, 20-25 years, 25-30 years, >30 years), and maternal micronutrient status. Urban residence was defined as those recruited from Dar es Salaam study sites. For the model predicting infant vitamin D deficiency at three months, maternal vitamin D status was defined as either sufficient (≥50 nmol/L), deficient (<50 nmol/L), or missing [34]. For the Nutrients 2017, 9, 1258 4 of 12 model predicting infant vitamin B12 deficiency at three months, maternal vitamin B12 status was defined as sufficient (>271 pg/mL), marginally deficient (203-271 pg/mL), deficient (<203 pg/mL), or missing [18]. Univariate analyses were conducted for variables that were considered a priori as potential predictors, and all variables were included in the multivariate model. Preterm and low birthweight were not included in the same model due to collinearity. For all covariates besides preterm, reported relative risks, confidence intervals, and p values are from a multivariate model that includes low birthweight, but not preterm. p for trend values were reported for maternal age and maternal education categorical variables. Models for infants at two weeks did not have maternal vitamin deficiency because maternal status was not collected in this age group. Statistical analysis was conducted using SAS, Version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results
We processed 1281 blood samples that measured at least one of the following micronutrients: vitamin D, vitamin B12, folate, ferritin, and hsCRP. A total of 446 infants were sampled at two weeks old, 408 infants sampled were at three months old, and 427 mothers were sampled at three months postpartum. Of the infants sampled at three months and mothers sampled at three months postpartum, 364 were infant-mother pairs.
Baseline characteristics for infants with samples at two weeks and infants with samples at three months are displayed in Table 2. A total of 13% of infants sampled at two weeks and 17% of infants sampled at three months were born preterm (before 37 weeks of gestation). The majority of infants (97.8%) in our sample received colostrum at birth and 90% of infants were breastfed within one hour of birth. From the sample, 43% of mothers reported exclusive breastfeeding at three months and 46% reported partial or predominate breastfeeding at three months. For both infant groups, only a quarter of mothers reported having four or more antenatal care visits (ANC). Descriptive statistics for infants' and mothers' micronutrient statuses are displayed in Table 3. On average, infants at three months had higher vitamin D levels (nmol/L), vitamin B12 (pg/mL), folate (ng/mL) levels, and lower ferritin (ng/mL) levels than infants two weeks old. For infants two weeks old, 62.6% were categorized as deficient in vitamin D and 26.4% were deficient in vitamin B12. No infants at two weeks of age had deficiencies in folate or iron. For infants at three months, 8.6% were categorized as deficient in vitamin D and 48.4% were categorized as insufficient deficient in vitamin D. 13.3% of infants at three months were categorized as vitamin B12 deficient and 16.7% were categorized as insufficient in vitamin B12. Less than 1% of infants at three months were categorized as deficient in folate or iron ( Figure 1). Descriptive statistics for infants' and mothers' micronutrient statuses are displayed in Table 3. On average, infants at three months had higher vitamin D levels (nmol/L), vitamin B12 (pg/mL), folate (ng/mL) levels, and lower ferritin (ng/mL) levels than infants two weeks old. For infants two weeks old, 62.6% were categorized as deficient in vitamin D and 26.4% were deficient in vitamin B12. No infants at two weeks of age had deficiencies in folate or iron. For infants at three months, 8.6% were categorized as deficient in vitamin D and 48.4% were categorized as insufficient deficient in vitamin D. 13.3% of infants at three months were categorized as vitamin B12 deficient and 16.7% were categorized as insufficient in vitamin B12. Less than 1% of infants at three months were categorized as deficient in folate or iron ( Figure 1). * Q1 = 25th percentile, Q3 = 75th percentile. For mothers at three months postpartum, 14.8% were deficient in vitamin D and 67.2% were categorized as insufficient; 5.2% of mothers were vitamin B12 deficient, while 25.6% of mothers were categorized as insufficient for vitamin B12; 19.0% of mothers were deficient in folate and 26.0% of mothers were deficient in ferritin ( Figure 2).  For mothers at three months postpartum, 14.8% were deficient in vitamin D and 67.2% were categorized as insufficient; 5.2% of mothers were vitamin B12 deficient, while 25.6% of mothers were categorized as insufficient for vitamin B12; 19.0% of mothers were deficient in folate and 26.0% of mothers were deficient in ferritin ( Figure 2).
We identified several predictors of infant vitamin D deficiency at two weeks and three months (Table 4). For infants at two weeks, maternal education level was the main predictor of Vitamin D deficiency. Infants at two weeks whose mothers received at least a secondary education were at twice Nutrients 2017, 9, 1258 6 of 12 the risk of being deficient in vitamin D than those whose mothers received no education. For infants at three months, infants who weighed less than 2500 g (low birthweight) at birth had 4.4 times the risk of being deficient in vitamin D compared to infants who weighed more than 2500 g at birth (multivariate Relative Risk (RR): 4.37, 95% Confidence Interval (CI): 1.95, 9.77). Infants at three months living in an urban area had nearly three times the risk of being deficient in vitamin D compared to infants living in rural area (multivariate RR: 2.95, 95% CI: 1.46, 5.97). In addition, infants at three months born to mothers who had some primary education had 5.2 times the risk of being deficient in vitamin D, and infants born to mothers with some secondary education had 8.5 times the risk of being deficient in vitamin D compared to infants whose mothers received no education (p for trend: 0.03). Finally, infants at three months whose mothers were vitamin D deficient (<50 nmol/L) had 2.8 times the risk of being deficient in vitamin D compared to infants whose mothers were not deficient in vitamin D (multivariate RR: 2.83, 95% CI: 1.45, 5.52).
Nutrients 2017, 9, 1258 6 of 12 We identified several predictors of infant vitamin D deficiency at two weeks and three months (Table 4). For infants at two weeks, maternal education level was the main predictor of Vitamin D deficiency. Infants at two weeks whose mothers received at least a secondary education were at twice the risk of being deficient in vitamin D than those whose mothers received no education. For infants at three months, infants who weighed less than 2500 g (low birthweight) at birth had 4.4 times the risk of being deficient in vitamin D compared to infants who weighed more than 2500 g at birth (multivariate Relative Risk (RR): 4.37, 95% Confidence Interval (CI): 1.95, 9.77). Infants at three months living in an urban area had nearly three times the risk of being deficient in vitamin D compared to infants living in rural area (multivariate RR: 2.95, 95% CI: 1.46, 5.97). In addition, infants at three months born to mothers who had some primary education had 5.2 times the risk of being deficient in vitamin D, and infants born to mothers with some secondary education had 8.5 times the risk of being deficient in vitamin D compared to infants whose mothers received no education (p for trend: 0.03). Finally, infants at three months whose mothers were vitamin D deficient (<50 nmol/L) had 2.8 times the risk of being deficient in vitamin D compared to infants whose mothers were not deficient in vitamin D (multivariate RR: 2.83, 95% CI: 1.45, 5.52). We also found several predictors for infant vitamin B12 deficiency for infants at two weeks and three months of age (Table 5). For infants at two weeks, maternal age was the main predictor of infant vitamin B12 status. As maternal age increased the relative risk of an infant at two weeks being deficient in vitamin B12 increases (p for trend = 0.0006), yet for infants at three months increased maternal age potentially decreased risk of deficiency but not a significant predictor of infant vitamin B12 status. For infants at three months, infants whose mothers were marginally vitamin B12 deficient (203-271 pg/mL) had four times the risk of being deficient in vitamin B12 compared to infants whose mothers were not deficient in vitamin B12 (multivariate RR: 4.06, 95% CI:1.88, 8.70). Infants at three months whose mothers were vitamin B12 deficient (<203 pg/mL) had 9.6 times the risk of being deficient in vitamin B12 compared to infants whose mothers were not deficient in vitamin B12 (multivariate RR: 9.63, 95% CI: 4.83, 19.17). We also found several predictors for infant vitamin B12 deficiency for infants at two weeks and three months of age (Table 5). For infants at two weeks, maternal age was the main predictor of infant vitamin B12 status. As maternal age increased the relative risk of an infant at two weeks being deficient in vitamin B12 increases (p for trend = 0.0006), yet for infants at three months increased maternal age potentially decreased risk of deficiency but not a significant predictor of infant vitamin B12 status. For infants at three months, infants whose mothers were marginally vitamin B12 deficient (203-271 pg/mL) had four times the risk of being deficient in vitamin B12 compared to infants whose mothers were not deficient in vitamin B12 (multivariate RR: 4.06, 95% CI:1.88, 8.70). Infants at three months whose mothers were vitamin B12 deficient (<203 pg/mL) had 9.6 times the risk of being deficient in vitamin B12 compared to infants whose mothers were not deficient in vitamin B12 (multivariate RR: 9.63, 95% CI: 4.83, 19.17).

Discussion
This study characterized the micronutrient status and identified predictors of micronutrient deficiency for young infants at two weeks and three months. In this Tanzanian cohort of infants and mothers, we found that some degree of maternal micronutrient deficiency was prevalent for all micronutrients assessed in this study. For infants at two weeks and three months, micronutrient deficiency was prevalent only for vitamin D and vitamin B12, but not for folate or iron. Our study found that over 60% of infants at two weeks were deficient in vitamin D. By the age of three months prevalence of vitamin D deficiency drops significantly to approximately 9%, but still greater than 40% of infants at three months were insufficient in vitamin D. Results from the three month cohort are consistent with other studies in Kenya and Uganda that found that a significant proportion of infants and children were at least marginally vitamin D deficient [38,39]. Our study also found that at two weeks of age, nearly 30% of infants were deficient and 20% were insufficient in vitamin B12, while 13% and 17% of infants at three months were vitamin B12 deficient and insufficient, respectively. While there is limited evidence from other studies conducted in East Africa regarding infant vitamin B12 status, these results are consistent with a study in Nepal that found that 17% of infants were vitamin B12 deficient and 40% were marginally deficient [11].
Few epidemiological studies have examined micronutrient status of infants less than six months; therefore, little is known about how micronutrient status changes over time during infancy. Although our data is not longitudinal, our results suggest that on average vitamin B12, folate, and vitamin D are lower in the first few weeks of life than at three months of age, while ferritin is higher in the first few weeks of life and then declines. A randomized longitudinal study in New Zealand also found that infant serum vitamin D increases steadily from birth to six months of age in the placebo group [40]. At birth and in early infancy, sources of infant serum vitamin D are primarily from maternal sources of vitamin D and as the child gets older, there is greater opportunity for the infant to synthesize vitamin D from ultraviolet light. The changes in ferritin levels during infancy are well documented in the literature and indicate that ferritin may not be the best measure of iron deficiency during this time period of life unless more precise cutoff points for deficiency are established [41]. Further longitudinal research is necessary to determine normal micronutrient levels in infancy and if low levels of micronutrients at infancy are predictors of deficiency later in life.
This study reports that infants at three months who were less than 2500 g at birth, residing in an urban area, whose mothers had attained higher levels of education, or whose mothers were vitamin D deficient were more likely to be vitamin D deficient. These results are consistent with other studies that show that infant vitamin D deficiency is associated with low birthweight and maternal vitamin D status [11]. Infants born to mothers with adequate vitamin D levels should have adequate vitamin D levels for their first eight weeks of life [42]. If the mother's stores are insufficient during pregnancy, infants may be born with low vitamin D levels and vitamin D concentration in breastmilk may be low [40]. The association between infant vitamin D deficiency and higher maternal education and urban residence may be explained by the amount of direct sunlight an infant receives. Unfortunately, data on the amount of direct sunlight was not available for this population, but we hypothesize that infants born to mothers with higher levels of education may be less likely to spend long periods outdoors. Similarly, infants in urban areas may spend less time outdoors than their rural peers.
Additionally, we found that the main predictor of infant vitamin B12 deficiency was the maternal vitamin B12 status. Infants whose mothers were vitamin B12 deficient had more than nine times the risk of being deficient in vitamin B12, and infants whose mothers were marginally deficient in vitamin B12 had four times the risk of being deficient than infants whose mothers were vitamin B12 sufficient, indicating a dose-response relationship. The only source of vitamin B12 for infants is breastmilk. Vitamin B12 concentrations in breastmilk are dependent on maternal vitamin B12 status [23]. Few studies have examined the effect of vitamin B12 supplementation in pregnancy, but a randomized trial in India found that supplementing women during pregnancy and early lactation significantly increased maternal and infant vitamin B12 serum concentrations [43].
Interestingly, while there is significant vitamin D and vitamin B12 deficiency in infants, there are very few cases of folate or iron deficiency in our study cohort. This may be due to the fact that antenatal care recommendations in Tanzania per WHO guidelines provide prenatal supplements with iron and folic acid, but do not provide supplements for vitamin B12 or vitamin D [44]. In 2016, the WHO updated their antenatal care recommendations and now recommend eight antenatal care visits [45]. These new guidelines still recommend routine supplementation for iron folic-acid, but still do not recommend routine supplementation for vitamin D and vitamin B12 during antenatal care visits due to limited evidence currently available.
This study has a few limitations. First, study participants came from a large randomized trial recruiting from Morogoro and Dar es Salaam study areas, hence the results may not be generalizable to all regions of Tanzania. In addition, micronutrient cutoffs for vitamin D, vitamin B12, and folate are based on cutoff points for adults. These cutoff points may not be most appropriate for young infants. One of the main challenges of our analysis was the lack of well-defined cutoff points for infants at two weeks, and we did not have information about maternal vitamin status at two weeks. Furthermore, only 25% of infants and mothers had hsCRP measured, therefore we could not adjust ferritin levels for acute phase infection. Additionally, we chose to measure serum folate because it is easier to measure in the field, but it is not as robust as red cell folate. In addition, serum folate fluctuates with intake therefore is a better measurement of short-term deficiency, rather than long-term status. Lastly, this was not a longitudinal study. Infants measured at two weeks were not the same infants measured at three months and thus we could not assess changes in micronutrient status over time. Our study has several strengths. Few studies have previously examined multiple micronutrient deficiencies in young infants. In addition, this is one of the few studies that has data on both infant and maternal micronutrient status for infants at three months.

Conclusions
Overall, we found a significant proportion of infants were either deficient or marginally deficient in vitamin D and vitamin B12, particularly among infants at two weeks, and that maternal status for vitamin D and vitamin B12 were significant predictors of infant micronutrient deficiency for infants at three months. Further research is necessary to determine the burden of micronutrient deficiencies in infancy in other settings, and to examine the potential benefits of improving infants' status through incorporating vitamin D and vitamin B12 supplements into the currently provided routine prenatal supplements or through direct infant supplementation.