INTRODUCTION Anemia is a public health problem worldwide and is most prevalent in preschool children, for whom it is the most frequent cause of nutritional defi cits. In turn, iron defi ciency is the main cause of anemia, affecting 43% of children globally. Previous studies in Cuba show rates of iron defi ciency in preschool children between 38.6% and 57.6%, higher in infants (71.2% to 81.1%). WHO recommends using serum ferritin as an indicator of iron defi ciency accompanied by acute (C-reactive protein) and chronic (α1-acid glycoprotein) infl ammation biomarkers. OBJECTIVE Assess how infl ammation affects measuring and reporting of iron-defi ciency anemia rates in Cuban preschool children.METHODS Data were obtained from serum samples contained in the National Anemia and Iron Defi ciency Survey, and included presumably healthy preschool Cuban children (aged 6-59 months). Serum samples were collected from 1375 children from randomly selected provinces in 4 regions of the country from 2014 through 2018. We examined the association between ferritin and two infl ammatory biomarkers: C-reactive protein and α1-acid glycoprotein. Individual infl ammation-adjusted ferritin concentrations were calculated using four approaches: 1) a higher ferritin cut-off point (<30 g/L); 2) exclusion of subjects showing infl ammation (C-reactive protein >5 mg/L or α1-acid glycoprotein >1 g/L); 3) mathematical correction factor based on C-reactive protein or α1-acid glycoprotein; and 4) correction by regression with the method proposed by the Biomarkers Refl ecting Infl ammation and Nutritional Determinants of Anemia Group. We estimated confi dence intervals of differences between unadjusted prevalence and prevalence adjusted for infl ammation by each method. RESULTSThe proportion of children with infl ammation according to C-reactive protein concentrations >5 mg/L was lower (11.1%, 153/1375) than the proportion measured according to the concentrations of α1-acid glycoprotein, at >1 g/L (30.8%, 424/1375). The percentage of children with high concentrations of at least one of the aforementioned biomarkers was 32.7% (450/1375). Thus, each correction method increased the observed prevalence of iron defi ciency compared to unadjusted estimates (23%, 316/1375). This increase was more pronounced when using the internal regression correction method (based only on C-reactive protein) or the method based on a higher cut-off point. Adjustment using all four methods changed estimated iron defi ciency prevalence, increasing it from 0.1% to 8.8%, compared to unadjusted values. CONCLUSION One-third of preschool children had biomarkers indicating elevated infl ammation levels. Without adjusting for infl ammation, iron defi ciency prevalence was underestimated. The signifi cant disparity between unadjusted and infl ammationadjusted ferritin when using some approaches highlights the importance of selecting the right approach for accurate, corrected measurement. The internal regression correction approach is appropriate for epidemiological studies because...

INTRODUCTION Ferritin is the best biomarker for assessing iron defi ciency, but ferritin concentrations increase with infl ammation. Several adjustment methods have been proposed to account for infl ammation's eff ect on iron biomarker interpretation. The most recent and highly recommended method uses linear regression models, but more research is needed on other models that may better defi ne iron status in children, particularly when distributions are heterogenous and in contexts where the eff ect of infl ammation on ferritin is not linear.OBJECTIVES Assess the utility and relevance of quadratic regression models and quantile quadratic regression models in adjusting ferritin concentration in the presence of infl ammation. METHODSWe used data from children aged under fi ve years, taken from Cuba's national anemia and iron defi ciency survey, which was carried out from 2015-2018 by the National Hygiene, Epidemiology and Microbiology Institute. We included data from 1375 children aged 6 to 59 months and collected ferritin concentrations and two biomarkers for infl ammation: C-reactive protein and α-1 acid glycoprotein. Quadratic regression and quantile regression models were used to adjust for changes in ferritin concentration in the presence of infl ammation. RESULTSUnadjusted iron defi ciency prevalence was 23% (316/1375). Infl ammation-adjusted ferritin values increased iron-defi ciency prevalence by 2.6-4.5 percentage points when quadratic regression correction model was used, and by 2.8-6.2 when quantile regression was used. The increase when using the quantile regression correction model was more pronounced and statistically signifi cant when both infl ammation biomarkers were considered, but adjusted prevalence was similar between the two correction methods when only one biomarker was analyzed.CONCLUSIONS The use of quadratic regression and quantile quadratic regression models is a complementary strategy in adjusting ferritin for infl ammation, and is preferable to standard regression analysis when the linear model's basic assumptions are not met, or when it can be assumed that ferritin-infl ammation relationships within a subpopulation may deviate from average trends.KEYWORDS Alpha-1-acid glycoprotein, C-reactive protein, anemia, iron defi ciency, ferritin, acute phase protein, Cuba