Anne M Williams

Adjusting ferritin concentrations for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project

Background: The accurate estimation of iron deficiency is important in planning and implementing interventions. Ferritin is recommended as the primary measure of iron status, but interpretability is challenging in settings with infection and inflammation. Objective: We assessed the relation between ferritin concentrations and inflammation and malaria in preschool children (PSC) (age range: 6–59 mo) and women of reproductive age (WRA) (age range: 15–49 y) and investigated adjustment algorithms to account for these effects. Design: Cross-sectional data from 15 surveys for PSC (n = 27,865) and 8 surveys for WRA (24,844), from the Biomarkers Reflecting the Inflammation and Nutritional Determinants of Anemia (BRINDA) project were analyzed individually and combined with the use of a meta-analysis. Several approaches were explored to estimate depleted iron stores (ferritin concentration <12 μg/L in PSC and <15 μg/L in WRA) in inflammation and malaria settings as follows: 1) increase ferritin-concentration cutoff to <30 μg/L; 2) exclude individuals with C-reactive protein (CRP) concentrations >5 mg/L or α-1-acid glycoprotein (AGP) concentrations >1 g/L; 3) apply arithmetic correction factors; and 4) use a regression correction approach. Results: Depleted iron-store estimates incrementally increased as CRP and AGP deciles decreased (4% compared with 30%, and 6% compared with 29% from highest compared with lowest CRP deciles for pooled PSC and WRA, respectively, with similar results for AGP). Depending on the approach used to adjust for inflammation (CRP plus AGP), the estimated prevalence of depleted iron stores increased by 7–25 and 2–8 absolute median percentage points for PSC and WRA, respectively, compared with unadjusted values. Adjustment for malaria in addition to CRP and AGP did not substantially change the estimated prevalence of depleted iron stores. Conclusions: Our results lend support for the use of internal regression correction to estimate the prevalence of depleted iron stores in regions with inflammation. This approach appears to mathematically reflect the linear relation of ferritin concentrations with acute-phase proteins. More research is warranted to validate the proposed approaches, but this study contributes to the evidence base to guide decisions about how and when to adjust ferritin for inflammation.

Adjusting soluble transferrin receptor concentrations for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project

Background: Iron deficiency is thought to be one of the most prevalent micronutrient deficiencies globally, but an accurate assessment in populations who are frequently exposed to infections is impeded by the inflammatory response, which causes iron-biomarker alterations. Objectives: We assessed the relation between soluble transferrin receptor (sTfR) concentrations and inflammation and malaria in preschool children (PSC) (age range: 6–59 mo) and women of reproductive age (WRA) (age range: 15–49 y) and investigated adjustment algorithms to account for these effects. Design: Cross-sectional data from the Biomarkers Reflecting the Inflammation and Nutritional Determinants of Anemia (BRINDA) project from 11,913 PSC in 11 surveys and from 11,173 WRA in 7 surveys were analyzed individually and combined with the use of a meta-analysis. The following 3 adjustment approaches were compared with estimated iron-deficient erythropoiesis (sTfR concentration >8.3 mg/L): 1) the exclusion of individuals with C-reactive protein (CRP) concentrations >5 mg/L or α-1-acid glycoprotein (AGP) concentrations >1 g/L, 2) the application of arithmetic correction factors, and 3) the use of regression approaches. Results: The prevalence of elevated sTfR concentrations incrementally decreased as CRP and AGP deciles decreased for PSC and WRA, but the effect was more pronounced for AGP than for CRP. Depending on the approach used to adjust for inflammation, the estimated prevalence of iron-deficient erythropoiesis decreased by 4.4–14.6 and 0.3–9.5 percentage points in PSC and WRA, respectively, compared with unadjusted values. The correction-factor approach yielded a more modest reduction in the estimated prevalence of iron-deficient erythropoiesis than did the regression approach. Mostly, adjustment for malaria in addition to AGP did not significantly change the estimated prevalence of iron-deficient erythropoiesis. Conclusions: sTfR may be useful to assess iron-deficient erythropoiesis, but inflammation influences its interpretation, and adjustment of sTfR for inflammation and malaria should be considered. More research is warranted to evaluate the proposed approaches in different settings, but this study contributes to the evidence on how and when to adjust sTfR for inflammation and malaria.

Vitamin B-12 concentrations in breast milk are low and are not associated with reported household hunger, recent animal-source food, or Vitamin B-12 intake in women in rural Kenya

BACKGROUND Breast milk vitamin B-12 concentration may be inadequate in regions in which animal-source food consumption is low or infrequent. Vitamin B-12 deficiency causes megaloblastic anemia and impairs growth and development in children. OBJECTIVE We measured vitamin B-12 in breast milk and examined its associations with household hunger, recent animal-source food consumption, and vitamin B-12 intake. METHODS In a cross-sectional substudy nested within a cluster-randomized trial assessing water, sanitation, hygiene, and nutrition interventions in Kenya, we sampled 286 women 1-6 mo postpartum. Mothers hand-expressed breast milk 1 min into a feeding after 90 min observed nonbreastfeeding. The Household Hunger Scale was used to measure hunger, food intake in the previous week was measured with the use of a food-frequency questionnaire (FFQ), and vitamin B-12 intake was estimated by using 24-h dietary recall. An animal-source food score was based on 10 items from the FFQ (range: 0-70). Breast milk vitamin B-12 concentration was measured with the use of a solid-phase competitive chemiluminescent enzyme immunoassay and was modeled with linear regression. Generalized estimating equations were used to account for correlated observations at the cluster level. RESULTS Median (IQR) vitamin B-12 intake was 1.5 μg/d (0.3, 9.7 μg/d), and 60% of women consumed <2.4 μg/d, the estimated average requirement during lactation. Median (IQR) breast milk vitamin B-12 concentration was 113 pmol/L (61, 199 pmol/L); 89% had concentrations <310 pmol/L, the estimated adequate concentration. Moderate or severe hunger prevalence was 27%; the animal-source food score ranged from 0 to 30 item-d/wk. Hunger and recent animal-source food and vitamin B-12 intake were not associated with breast milk vitamin B-12 concentrations. Maternal age was negatively associated with breast milk vitamin B-12 concentrations. CONCLUSION Most lactating Kenyan women consumed less than the estimated average requirement of vitamin B-12 and had low breast milk vitamin B-12 concentrations. We recommend interventions that improve vitamin B-12 intake in lactating Kenyan women to foster maternal health and child development. The main trial was registered at clinicaltrials.gov as NCT01704105.

Assessment of iron status in settings of inflammation: challenges and potential approaches

The determination of iron status is challenging when concomitant infection and inflammation are present because of confounding effects of the acute-phase response on the interpretation of most iron indicators. This review summarizes the effects of inflammation on indicators of iron status and assesses the impact of a regression analysis to adjust for inflammation on estimates of iron deficiency (ID) in low– and high–infection-burden settings. We overviewed cross-sectional data from 16 surveys for preschool children (PSC) (n = 29,765) and from 10 surveys for nonpregnant women of reproductive age (WRA) (n = 25,731) from the Biomarkers Reflecting the Inflammation and Nutritional Determinants of Anemia (BRINDA) project. Effects of C-reactive protein (CRP) and α1-acid glycoprotein (AGP) concentrations on estimates of ID according to serum ferritin (SF) (used generically to include plasma ferritin), soluble transferrin receptor (sTfR), and total body iron (TBI) were summarized in relation to infection burden (in the United States compared with other countries) and population group (PSC compared with WRA). Effects of the concentrations of CRP and AGP on SF, sTfR, and TBI were generally linear, especially in PSC. Overall, regression correction changed the estimated prevalence of ID in PSC by a median of +25 percentage points (pps) when SF concentrations were used, by −15 pps when sTfR concentrations were used, and by +14 pps when TBI was used; the estimated prevalence of ID in WRA changed by a median of +8 pps when SF concentrations were used, by −10 pps when sTfR concentrations were used, and by +3 pps when TBI was used. In the United States, inflammation correction was done only for CRP concentrations because AGP concentrations were not measured; regression correction for CRP concentrations increased the estimated prevalence of ID when SF concentrations were used by 3 pps in PSC and by 7 pps in WRA. The correction of iron-status indicators for inflammation with the use of regression correction appears to substantially change estimates of ID prevalence in low– and high–infection-burden countries. More research is needed to determine the validity of inflammation-corrected estimates, their dependence on the etiology of inflammation, and their applicability to individual iron-status assessment in clinical settings.

Adjusting retinol-binding protein concentrations for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project

Background: The accurate estimation of the prevalence of vitamin A deficiency (VAD) is important in planning and implementing interventions. Retinol-binding protein (RBP) is often used in population surveys to measure vitamin A status, but its interpretation is challenging in settings where inflammation is common because RBP concentrations decrease during the acute-phase response. Objectives: We aimed to assess the relation between RBP concentrations and inflammation and malaria in preschool children (PSC) (age range: 6–59 mo) and women of reproductive age (WRA) (age range: 15–49 y) and to investigate adjustment algorithms to account for these effects. Design: Cross-sectional data from 8 surveys for PSC (n = 8803) and 4 surveys for WRA (n = 4191) from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project were analyzed individually and combined with the use of a meta-analysis. Several approaches were explored to adjust RBP concentrations in PSC in inflammation and malaria settings as follows: 1) the exclusion of subjects with C-reactive protein (CRP) concentrations >5 mg/L or α-1-acid glycoprotein (AGP) concentrations >1 g/L, 2) the application of arithmetic correction factors, and 3) the use of a regression correction approach. The impact of adjustment on the estimated prevalence of VAD (defined as <0.7 μmol/L) was examined in PSC. Results: The relation between estimated VAD and CRP and AGP deciles followed a linear pattern in PSC. In women, the correlations between RBP and CRP and AGP were too weak to justify adjustments for inflammation. Depending on the approach used to adjust for inflammation (CRP+AGP), the estimated prevalence of VAD decreased by a median of 11–18 percentage points in PSC compared with unadjusted values. There was no added effect of adjusting for malaria on the estimated VAD after adjusting for CRP and AGP. Conclusions: The use of regression correction (derived from internal data), which accounts for the severity of inflammation, to estimate the prevalence of VAD in PSC in regions with inflammation and malaria is supported by the analysis of the BRINDA data. These findings contribute to the evidence on adjusting for inflammation when estimating VAD with the use of RBP.

Predictors of anemia in preschool children: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project

Background: A lack of information on the etiology of anemia has hampered the design and monitoring of anemia-control efforts. Objective: We aimed to evaluate predictors of anemia in preschool children (PSC) (age range: 6–59 mo) by country and infection-burden category. Design: Cross-sectional data from 16 surveys (n = 29,293) from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project were analyzed separately and pooled by category of infection burden. We assessed relations between anemia (hemoglobin concentration <110 g/L) and severe anemia (hemoglobin concentration <70 g/L) and individual-level (age, anthropometric measures, micronutrient deficiencies, malaria, and inflammation) and household-level predictors; we also examined the proportion of anemia with concomitant iron deficiency (defined as an inflammation-adjusted ferritin concentration <12 μg/L). Countries were grouped into 4 categories on the basis of risk and burden of infectious disease, and a pooled multivariable logistic regression analysis was conducted for each group. Results: Iron deficiency, malaria, breastfeeding, stunting, underweight, inflammation, low socioeconomic status, and poor sanitation were each associated with anemia in >50% of surveys. Associations between breastfeeding and anemia were attenuated by controlling for child age, which was negatively associated with anemia. The most consistent predictors of severe anemia were malaria, poor sanitation, and underweight. In multivariable pooled models, child age, iron deficiency, and stunting independently predicted anemia and severe anemia. Inflammation was generally associated with anemia in the high- and very high–infection groups but not in the low- and medium-infection groups. In PSC with anemia, 50%, 30%, 55%, and 58% of children had concomitant iron deficiency in low-, medium-, high-, and very high–infection categories, respectively. Conclusions: Although causal inference is limited by cross-sectional survey data, results suggest anemia-control programs should address both iron deficiency and infections. The relative importance of factors that are associated with anemia varies by setting, and thus, country-specific data are needed to guide programs.

Breastfeeding and Complementary Feeding Practices among HIV-Exposed Infants in Coastal Tanzania

Background: Appropriate infant feeding is a persistent challenge for human immunodeficiency virus (HIV)-infected mothers in sub-Saharan Africa. Objective: This study aimed to describe correlates of infant feeding among HIV-infected mothers in coastal Tanzania. Methods: HIV-infected women (n = 400) with infants younger than 18 months were enrolled from June to November 2011 from 3 public health facilities in Pwani, Tanzania: Tumbi Regional Hospital (TRH), Chalinze Health Center (CHC), and Bagamoyo District Hospital (BDH). Participants were surveyed about sociodemographics and infant feeding behavior at enrollment; infant feeding data were collected prospectively and retrospectively in the month of study follow-up. Results: Statistically significant correlates of exclusive breastfeeding (EBF) were infant age (months) (adjusted odds ratio [AOR] = 0.6; 95% confidence interval [CI], 0.5-0.9), enrollment facility (TRH: reference; CHC: AOR = 5.0, 95% CI, 1.2-20.8; BDH: AOR = 11.6, 95% CI, 2.3-59.9), and HIV disclosure to one’s mother (AOR = 0.2; 95% CI, 0.1-0.6). Exclusive breastfeeding prevalence among infants younger than 6 months was 77%, but 50% of infants older than 6 months no longer receiving breast milk did not receive animal source foods (ASF) daily. Enrollment facility (TRH: reference; CHC: AOR = 0.2, 95% CI, 0.1-1.0; BDH: AOR = 0.1, 95% CI, 0.01-0.4) and HIV disclosure (to mother-in-law: AOR = 0.2, 95% CI, 0.1-0.8; to brother: AOR = 0.3, 95% CI, 0.1-0.8) were negatively associated with ASF provision. Conclusion: High prevalence of EBF suggests that it is an attainable behavior, whereas low prevalence of daily ASF provision suggests that adequate diets are difficult to achieve after breastfeeding cessation. These findings support current recommendations for HIV-infected mothers in resource-poor regions to continue breastfeeding for at least 1 year and suggest the need for greater support with complementary feeding. Associations between HIV disclosure and infant feeding merit further exploration, and correlations between enrollment facility and infant feeding highlight the potential influence of clinics on achieving infant feeding recommendations.

Approaches to assess vitamin a status in settings of inflammation: Biomarkers reflecting inflammation and nutritional determinants of anemia (BRINDA) project

The accurate estimation of vitamin A deficiency (VAD) is critical to informing programmatic and policy decisions that could have important public health implications. However, serum retinol and retinol binding protein (RBP) concentrations, two biomarkers often used to estimate VAD, are temporarily altered during the acute phase response, potentially overestimating the prevalence of VAD in populations with high levels of inflammation. In 22 nationally-representative surveys, we examined (1) the association between C-reactive protein (CRP) or α1-acid glycoprotein (AGP) and retinol or RBP, and (2) how different adjustment approaches for correcting for inflammation compare with one another. In preschool age children (PSC) and school age children (SAC), the association between inflammation and retinol and RBP was largely statistically significant; using the regression approach, adjustments for inflammation decreased the estimated prevalence of VAD compared to unadjusted VAD (range: −22.1 to −6.0 percentage points). In non-pregnant women of reproductive age (WRA), the association between inflammation and vitamin A biomarkers was inconsistent, precluding adjustments for inflammation. The burden of VAD can be overestimated if inflammation is not accounted for, and the regression approach provides a method for adjusting retinol and RBP for inflammation across the full range of concentrations in PSC and SAC.

HIV status disclosure among postpartum women in rural Tanzania: predictors, experiences and uptake of a nurse-facilitated disclosure intervention

ABSTRACT HIV status disclosure is a key support strategy to start and maintain HIV care and treatment and to reduce HIV transmission. We explored the patterns and correlates of disclosure and described the effectiveness of nurse-facilitated disclosure among HIV-infected mothers of infants in coastal Tanzania. We enrolled 400 HIV positive women in an observational longitudinal study in 2011, interviewed them about maternal sociodemographic and economic characteristics, maternal and child health and history of HIV disclosure experiences and offered nurse-facilitated HIV disclosure at enrolment or at follow-up 1 month later. Mothers frequently disclosed their status to husbands and/or female relatives and experienced predominantly positive reactions. Economically vulnerable women disclosed more often to elderly female relatives, indicating that Infant and Young Child Feeding counseling given to HIV positive women should garner the support of elderly female relatives for implementing appropriate feeding practices. Nurse-facilitated disclosure was feasible in this low resource setting and was used by patients to help them with both first-time disclosure and disclosure to new persons.

Infant Serum and Maternal Milk Vitamin B-12 Are Positively Correlated in Kenyan Infant-Mother Dyads at 1–6 Months Postpartum, Irrespective of Infant Feeding Practice

Abstract Background Vitamin B-12 is an essential nutrient required for many functions including DNA synthesis, erythropoiesis, and brain development. If maternal milk vitamin B-12 concentrations are low, infants may face elevated risks of deficiency when exclusively breastfed. Objective We evaluated cross-sectional associations between infant serum vitamin B-12 concentrations and maternal milk vitamin B-12 concentrations at 1–6 mo postpartum among an unsupplemented population in rural western Kenya, and assessed biological demographic, and dietary characteristics associated with adequate infant serum vitamin B-12. Methods We modeled 1) infant serum vitamin B-12 using maternal milk vitamin B-12 concentration with linear regression; and 2) adequate (>220 pmol/L) infant serum vitamin B-12 using hypothesized biological, demographic, and dietary predictors with logistic regression. In both models, we used generalized estimating equations to account for correlated observations at the cluster-level. Results The median (quartile 1, quartile 3) infant serum vitamin B-12 concentration was 276 pmol/L (193, 399 pmol/L) and approximately one-third of infants had serum vitamin B-12 ≤220 pmol/L, indicating that they were vitamin B-12 depleted or deficient. There was a positive correlation between maternal milk and infant serum vitamin B-12 (r = 0.36, P < 0.001) and in multivariable analyses, maternal milk vitamin B-12 concentration was significantly associated with infant serum vitamin B-12 adequacy (P-trend = 0.03). Conclusions Despite a high prevalence (90%) of maternal milk vitamin B-12 concentrations below the level used to establish the Adequate Intake (<310 pmol/L), there was a low prevalence of infant vitamin B-12 deficiency. We found few factors that were associated with infant vitamin B-12 adequacy in this population, including infant feeding practices, although maternal vitamin B-12 status was not measured. The contribution of maternal milk to infant vitamin B-12 status remains important to quantify across populations, given that maternal milk vitamin B-12 concentration is modifiable with supplementation. This trial was registered at clinicaltrials.gov as NCT01704105.