In this cohort of 259 patients undergoing haplo-SCT using ATG-based protocol, we demonstrated that CRS is a common phenomenon after allograft infusion with an incidence of 39.8% and most of them are grade 1 CRS (80.6%). We did not observe any severe CRS. Older recipient age, PBSC as a sole stem cell source, and higher elevated CRP levels during conditioning could predict the occurrence of CRS. CRS does not affect long-term survival.
The incidence rate of CRS can vary according to different diagnostic criteria. There were several diagnostic criteria for CRS according to published literature(5, 22–24). The vast majority of studies follow the diagnostic criteria of Lee et al.(5). However, there was no consensus on the diagnostic time of CRS. In a previous study, the diagnostic time of CRS was divided into two categories: one was the day from graft infusion day to 5 to 7 days after SCT(9, 11, 25), and the other was defined as extended to 14 days after SCT(6, 8, 10, 26). We defined the highest CRS grading diagnostic time as the day from graft infusion day to five days after SCT for the following reasons. Firstly, the patients with fever body temperature elevated immediately after allograft infusion, and the patients who had fevers after two days of stem cell infusion were all relieved through antibiotics. Secondly, the median time to diarrhea was one day after graft infusion in the current study. Finally, the median time to neutrophil engraftment was 12 days post-graft infusion in our study to exclude CRS in the peri-engraftment period. Furthermore, diarrhea and elevated transaminase are very common adverse effects of ATG administration, it was difficult to distinguish between the toxicity of ATG and CRS of allograft infusion. Therefore, whether the diagnostic criteria of CRS proposed by Lee et al. and the diagnostic timing of CRS are appropriate for ATG-based haplo-SCT needs further investigation.
Literature on the association between CRS with transplant outcome is variable(6–13). We found NRM was 20% for grade 2 CRS and 11% for grade 0–1 CRS, but they were not statistically significant. Severe CRS has been reported to increase NRM and decrease OS(8–11). We did not observe any severe CRS in the current study. There were several reasons for this finding. Firstly, in the current cohort, the median age of patients was relatively younger. It was reported that recipient age ≥ 60 was associated with higher NRM and grade 3–4 CRS (9, 11). Secondly, probably owing to the immune-based mechanism, there were various immunosuppressants including ATG, Cy, CSA, MMF, and Flu were administrated before allograft infusion so that minimized allo-reaction of infused donor T cells against the patient to the maximum extent. Finally, recipient with complete remission status before SCT was over 80% of all cohort in our study. Similar to the process of chimeric antigen receptor T cell therapy(27), pre-transplant disease status with a higher tumor burden was known to be the predictor of severe CRS(6).
CRS was reported to decrease disease relapse and increase risk of chronic GVHD by increasing quantity of T-cell subpopulations, particularly Tregs (15). In the current study, CIR was slightly lower in grade 1–2 CRS than in grade 0 CRS (14% vs. 19%), however, grade III-IV aGVHD and severe cGVHD were vaguely higher in grade 1–2 CRS than grade 0 CRS (14% vs. 11% and 12% vs. 9%). As a result, GRFS were comparable between patients without CRS and grade 1–2 CRS (66% vs. 70%, p > 0.05). In the meantime, there are also reports with opposite conclusions that severe CRS could reduce the risk of chronic GVHD (26). Concordant with this finding, in our study, the incidence of cGVHD for grade 2 CRS was 33% and 48% for grade 0–1 CRS. On the contrary, grade II-IV aGVHD was higher for grade 2 CRS than for grade 0–1 CRS (45% vs. 33%). We hypothesize that the effect of early cytokine release on recipients was transient and endothelial damage or macrophage activation etc. were also played a role in aGVHD. Further studies are needed to investigate the mechanism of CRS effect both on aGVHD and cGVHD.
PBSC has been reported to increase the risk of CRS(6, 7, 28). Our study demonstrated that patients who received PBSC solely were more likely to cause CRS than those who received PBSC combined with bone marrow graft. Previous studies demonstrated that CRS rate after bone marrow graft infusion was 33 ~ 79%(6, 7, 28). Interestingly, in our cohort, we did not observe any patient had fever after bone marrow graft infusion and all CRS began within the first day after PBSC infusion except one patient who developed diarrhea after bone marrow infusion. The possible reasons are as follows. First of all, the works of literature were based on the PT/Cy conditioning regimen(6, 7, 28). In addition, patients who were infused with bone marrow allograft received PBSCs immediately the next day. Thus, we can only observe the response within 24 hours of bone marrow infusion, and cannot detect the subsequent response. Finally, Otoukesh et al. found that a higher CD3+ cell dose can predict a higher rate of grade 2–4 CRS (11). The median CD3+ T-Cell was 0.12×108/L in bone marrow was almost 30 times lower than PBSCs in our study (data not shown).
There are two major limitations in this study that could be addressed in future research. First, owing to the nature of the retrospective study, levels of cytokines were not determined in this cohort. Abboud et al. found that fever and CRP level are associated with the level of interleukin-6 (IL-6), nonetheless, there was no relationship between the level of IL-6 and CRS severity in Pt/Cy setting(8). Tocilizumab is a therapeutic approach to decrease IL-6 in plasma, thereby reducing organ toxicity. However, it is of uncertain benefit, and there was not any severe CRS developed in our study. Second, for the low incidence of grade 2 CRS, the ability to explore the risk factors of grade 2 CRS was limited by the size of this part of the cohort. In the future, we will increase the sample size, register a clinical trial, and launch multicenter research that could provide more power to find potentially significant risk factors for grade 2 or higher grades of CRS.
In conclusion, CRS is a common complication after stem cell infusion with ATG-based haplo-SCT. Fever occurs nearly in 90% of the patients with CRS, which provides support for differential diagnosis. The ATG based haplo-SCT is safe, for severe CRS was absent in our experience. The 3-year NRM for grade 2 CRS was up to 20%, a large-scale prospective study is warranted to identify both host and graft factors that are susceptible to the higher grade of CRS for reducing NRM.