Modeling Influencing Factors in B-Cell Reconstitution After Hematopoietic Stem Cell Transplantation in Children

Abstract

Reduced total and memory B-cell numbers in peripheral blood long term after hematopoietic stem cell transplantation (HSCT) are associated with an increased incidence of infections and immune complications. Using novel modelling strategies, baseline factors influencing B-cell reconstitution can be comprehensively studied. This study aims to investigate the numerical total and memory B-cell reconstitution in children and the association with baseline determinants 0.5-2 years after allogeneic HSCT. Eligible for inclusion were children transplanted in our center between 2004-2017 who received a first HSCT for malignant or non-malignant disorders. The continuous absolute counts of total and memory B-cells were evaluated as outcome measure. Exploratory analysis at one year was done to identify possible determinants. Linear mixed effect modelling was used to analyze the association of these determinants with total and memory B-cell reconstitution 0.5-2 years after HSCT. In a cohort of 223 evaluable patients analyzed at 1-year after HSCT donor age, stem cell source, donor type, recipient age and conditioning were identified as significant determinants for total and memory B-cell numbers. Multivariable analysis revealed that both donor and recipient age were inversely correlated with the size of total and memory B-cell reconstitution. In contrast, no correlation was found with stem cell source, donor type and conditioning. Making use of linear mixed modelling both stem cell donor and recipient age were identified as independent determinants of total and memory B-cell reconstitution 0.5-2 years after HSCT.

Keywords: B lymphocyte; allogeneic; hematopoietic stem cell transplantation; immune reconstitution; linear mixed effect modelling; pediatric.

Figure 1

(A–J) Determinant analysis at 1 year. BM, bone marrow; CB, cord blood; PBSC, peripheral blood stem cells; Haplo, haplo-identical; IRD, identical related donor; UD, unrelated donor; MAC, myeloablative; RIC, reduced intensity conditioning. P value from the Wilcoxon signed-rank test or Kruskal-Wallis test comparing categories and total and memory B-cell numbers. Data is presented as median with interquartile ranges.

Figure 2

(A–D) The effect of time on total (A) and memory B-cell (B) numbers after HSCT. The correlation of memory B-cells and total B-cells 1 year (C, n=223) after HSCT and 2 years (D, n=86) after HSCT. Data is presented as individual trajectories and dots in black and as the average reconstitution in red. R is the Pearson’s correlation coefficient.

Figure 3

(A–J) Cell number trajectories of total B-cells and memory B-cells and the effect of baseline factors over time. Yrs, years; BM, bone marrow; CB, cord blood; PBSC, peripheral blood stem cells; Haplo, haplo-identical; IRD, identical related donor; UD, unrelated donor; MAC, myeloablative; RIC, reduced intensity conditioning. Lines represent mean predicted values, shaded areas the 95% CI’s.