Page 138 - 2022_01-Haematologica-web

Y. Zaimoku et al.
the group with ARC >30×109/L, P=0.024) (Figure 2B). There was no such clear difference of response curve between the IST groups according to ANC or ALC (Online Supplementary Figure S3).
Prediction of complete response
Pretreatment ALC was especially elevated in patients who achieved a complete response in both treatment groups, IST plus eltrombopag and IST alone (P-values, 0.065 and 0.095, compared to no complete response) (Figure 3; Table 2), which was statistically significant in all patients combined (P=0.020). Other variables, including the strong predictors of overall response, were unexpect- edly not discriminatory for complete response.
A large retrospective study of 312 pediatric AA patients found significantly reduced pretreatment ALC in IST responders compared to non-responders (median of 1.6×109/L vs. 2.0×109/L, respectively; P=0.006);21 and other studies in children also showed similar trends, while the correlation of higher ALC with better outcomes was mostly described in adult patients (Online Supplementary Table S2). We therefore hypothesized that the impact of ALC on IST responsiveness differed by age. We did observe a significant inverse correlation of ALC with com- plete response in 34 younger children aged <10 years for the two IST groups; six complete responders compared to the other 28 patients had significantly reduced ALC at a median (IQR) of 0.41 (0.31-1.25)×109/L versus 1.50 (1.07- 2.13)×109/L (P=0.037) (Online Supplementary Figure S4A). ALC retained a significant positive correlation with com- plete response in adolescents aged 10-19 years (P=0.012) and in adults aged 20 years or older (P=0.029). Patient age of 10 years was the best cutoff to discriminate the inverse relationship of ALC with complete response in younger and older individuals (Online Supplementary Figure S4B-D). Therefore, exclusion of the young children aged <10 years from the statistical analyses led to a significant correlation of ALC with complete response (P=0.010 in the IST plus eltrombopag group and P=0.012 in the group treated with
IST alone) as well as with overall response (P=0.016 in the IST plus eltrombopag group and P=0.00011 in the IST alonegroup).Completeresponsetendedtoassociatewith higher ALC in all the age groups over 10 years old, but the association was especially obvious in adolescents and young adults, aged 10-40 years (Online Supplementary Figure S4A). The complete response rate to IST plus eltrombopag in the adolescents and young adults reached 60% (28 of 47) if the ALC was more than 1.3×109/L before therapy, which was significantly higher than the 32% (15 of 47, P=0.013) in the other half of patients with lower ALC; this cut-off of ALC could also stratify complete response probability in the group treated with IST alone (21% [14 of 68] vs. 3% [2 of 66]; P=0.0024).
We did not observe such an age-dependent inverse cor- relation in other predictors of overall response (Online Supplementary Figure S5).
Rabbit antithymocyte globulin-based immunosuppres- sive therapy
Because the correlation of ALC with complete response has not been reported, we additionally assessed this rela- tionship in 54 patients, aged 10 years or older, treated with rabbit ATG and cyclosporine (NCT00260689).32 Four of the 54 patients achieved a complete response; they had significantly higher ALC compared to the remaining 50 patients at a median (IQR) of 1.97 (1.74-2.41)×109/L versus 1.14 (0.81-1.36)×109/L (P=0.0045) (Online Supplementary Figure S6). None of six children aged <10 years treated with rabbit ATG-based IST achieved a complete response. ARC and ANC did not correlate with complete response, but as expected they did with overall response, similar to hATG-based IST.
Discussion
Eltrombopag combined with standard IST is now the best nontransplant therapy for SAA, but little has been
Figure 3. Blood counts and complete response. Pretreatment absolute reticulocyte count, absolute neutrophil count, absolute lymphocyte count and plasma throm- bopoietin level of patients in the groups treated with immunosuppressive therapy (IST) plus eltrombopag (EPAG) and IST alone are shown according to responses (complete, partial and none) at 6 months. Initial absolute lymphocyte count tended to be elevated in patients who achieved complete response after IST both with and without EPAG, but absolute reticulocyte count, absolute neutrophil count and thrombopoietin, could not discriminate who would go on to have a complete response. ARC: absolute reticulocyte count; ANC: absolute neutrophil count; ALC: absolute lymphocyte count; TPO: plasma thrombopoietin; NR: no response; PR: par- tial response; CR: complete response.
130
haematologica | 2022; 107(1)