Exploring the association of hemoglobin level and adverse events in children with cancer presenting with fever in neutropenia

Abstract

Background: In children and adolescents with fever in neutropenia (FN) during chemotherapy for cancer, hemoglobin ≥90 g/L at presentation with FN had been associated with adverse events (AE). This analysis explored three hypothetical pathophysiological mechanisms potentially explaining this counterintuitive finding, and further analyzed the statistical association between hemoglobin and AE.

Methods: Two of 8 centers, reporting on 311 of 421 FN episodes in 138 of 215 patients participated in this retrospective analysis based on prospectively collected data from three databases (SPOG 2003 FN, transfusion and hematology laboratories). Associations with AE were analyzed using mixed logistic regression.

Results: Hemoglobin was ≥90 g/L in 141 (45%) of 311 FN episodes, specifically in 59/103 (57%) episodes with AE, and in 82/208 (39%) without (OR, 2.3; 99%CI, 1.1-4.9; P = 0.004). In FN with AE, hemoglobin was bimodally distributed with a dip around 85 g/L. There were no significant interactions for center, age and sex. In multivariate mixed logistic regression, AE was significantly and independently associated with leukopenia (leukocytes <0.3 G/L; OR, 3.3; 99%CI, 1.1-99; P = 0.004), dehydration (hemoglobinPresentation/hemoglobin8-72 hours ≥1.10 in untransfused patients; OR, 3.5; 99%CI, 1.1-11.4; P = 0.006) and non-moderate anemia (difference from 85 g/L; 1.6 per 10 g/L; 1.0-2.6; P = 0.005), but not with recent transfusion of packed red blood cells (pRBC), very recent transfusion of pRBC or platelets, or with hemoglobin ≥90 g/L as such.

Conclusions: Non-moderate anemia and dehydration were significantly and relevantly associated with the risk of AE in children with cancer and FN. These results need validation in prospective cohorts before clinical implementation.

Conflict of interest statement

Competing Interests: The authors have the following interests: This study has received unrestricted research grants by two commercial funders - Bayer AG (Switzerland), and GSK AG (Switzerland). This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.

Figures

Figure 1. Adverse events versus hemoglobin at presentation.

Frequency histogram of episodes of fever in neutropenia (FN; left vertical axis). Proportions of FN episodes with any adverse event (AE) measured, and by 2nd and 8th order regression smoothing (right vertical axis). SMC, serious medical complication.

Figure 2. Course of hemoglobin used to calculate Hb-ratio.

Hemoglobin at presentation, and within 72; AE, adverse event; SMC, serious medical complication.

Figure 3. Adverse events versus dehydration, estimated by Hb-ratio.

Frequency histogram of FN episodes (left vertical axis). Proportions of FN episodes with any AE measured, and by 2nd and 8th order regression smoothing (right vertical axis).

Figure 4. Adverse events versus the difference from moderate anemia.

Frequency histogram of FN episodes (left vertical axis). Proportions of FN episodes with any AE measured, and by 2nd and 8th order regression smoothing (right vertical axis).