LAUSR

To the Editor: We read with interest the article by Tonial et al (1) published in the recent issue of Pediatric Critical Care Medicine. This retrospective study was conducted over a period of 31⁄2 years (July 2013 and January 2017) involving 350 children 6 months to 18 years old with diagnosis of sepsis (suspected or proven) admitted to a tertiary care PICU of Brazil. The serum ferritin was measured within 48 hours of PICU admission among 312 children. The authors demonstrated that the prevalence of iron deficiency anemia in this cohort was high (40.3%). The median of highest serum ferritin level within 48 hours of PICU admission was 150.5 ng/ mL (interquartile range [IQR], 82.25–382 ng/mL) and ferritin levels were significantly higher among nonsurvivors than survivors (586 ng/mL [226–1,093 ng/mL] vs 138 ng/mL [80.5– 287 ng/mL], respectively; p = 0.001). Authors also noted that a 10-fold increase in serum ferritin level was associated with a five-fold increase in mortality and there was monotonic increase in mortality with increase in ferritin levels (p < 0.05). We wish to raise few important concerns regarding the results of the study by Tonial et al (1). The hyperferritinemic sepsis and hemophagocytic lymphohistiocytosis (HLH) have many overlapping features. Sepsis may be associated with secondary HLH and HLH is an independent predictor of mortality (2–4). Authors have not commented on the evaluation of these patients for the presence of HLH. Also, authors could have incorporated routine biomarkers of sepsis (C-reactive protein and procalcitonin) to predict the severity of disease and outcome as well as could have calculated discriminatory power for mortality using area under the receiver operating characteristic curve. Results of the study by Tonial et al (1) are impressive and aptly applicable for the low-middle income countries (LMICs), like ours. The usual cutoff value of serum ferritin in cases with sepsis to call it as hyperferritinemic sepsis is greater than 500 ng/mL and levels of greater than 1,000 ng/mL are associated with mortality (4, 5). Few studies from developing countries also suggested that hyperferritinemia is associated with poor outcome in critically ill children (6, 7). But in LMICs, where the prevalence of iron deficiency anemia is high, the threshold levels of serum ferritin in children with hyperferritinemic sepsis may be lower than the suggested levels. This fact is highlighted by the result of the study by Tonial et al (1) where authors noted that lower threshold of serum ferritin predicted mortality in children with high prevalence of iron deficiency anemia and showed a linear relation between ferritin level and mortality even at serum ferritin levels less than 500 ng/ mL. There are only a few studies that highlighted the need for setting a lower threshold of serum ferritin to define hyperferritinemia and its association with mortality in critically ill children with sepsis in LMICs (1, 7). A study form author’s unit (7) enrolled 42 children 6 months to 12 years old with septic shock and measured serum ferritin at enrollment. One-third of cases were malnourished, 86% had anemia, and two-third had microcytic hypochromic RBCs. Ferritin at admission among survivors and nonsurvivors was 415 ng/mL (262–852 ng/mL) and 763 ng/mL (480–1,820 ng/mL), respectively (p = 0.11). The serum ferritin greater than 500 ng/mL (relative risk, 2.48; 95% CI, 0.95–6.43) and greater than 1,000 ng/mL (relative risk, 1.94; 95% CI, 0.94–4.02) were associated with higher mortality. The admission C-reactive protein and Pediatric Logistic Organ Dysfunction score correlated positively with plasma ferritin. The authors have disclosed that they do not have any potential conflicts of interest.