Characterizing Immune Mediated Cytopenias following Allogeneic Hematopoietic Cell Transplantation for Pediatric Non-Malignant Disorders

Abstract

Immune mediated cytopenias (IMC) - isolated or combined hemolytic anemia, thrombocytopenia, or neutropenia - are increasingly recognized as serious complications following allogeneic hematopoietic cell transplantation (HCT) for non-malignant disorders (NMD). However, IMC incidence, duration, response to therapy, and risk factors are not well defined. This retrospective chart review identified cases of IMC with serologic confirmation amongst patients who underwent HCT for NMD at a single institution between 2010 and 2017. IMC following HCT for NMD in a large pediatric cohort (n = 271) was common with a cumulative incidence of 18%, identified at a median of 136 days after HCT. Treatment included prolonged immune suppression (>3 months) in 58% of all IMC cases, 91% when multiple cell lines affected. Multiple therapeutic agents were utilized for the majority affected and median time to resolution of IMC was 118 days from diagnosis. Fine-Gray competing risk multivariate regression analysis identified a combined risk factor of younger age (<3 years) and inherited metabolic disorder, as well as hemoglobinopathy (at any age) associated with one-year incidence of IMC (p < 0.01). We expand these findings with the observation of declining donor T-lymphoid chimerism from day +60 to +100 and lower absolute CD4+ counts at day +100 (p < 0.01), prior to median onset of IMC, for patients with IMC compared to those without. In this cohort, four deaths (8%) were associated with IMC, including 2 requiring second transplant for secondary graft failure. While the pathogenesis of IMC post-HCT for NMD remains elusive, further research may identify approaches to prevent and better treat this HCT complication.

INTRODUCTION

Immune mediated cytopenias (IMC) - isolated or combined hemolytic anemia, thrombocytopenia, and neutropenia - are recognized as a potentially serious complication of hematopoietic cell transplantation (HCT). In contrast to the general pediatric population where autoimmune hemolytic anemia and immune thrombocytopenic purpura have an incidence of <0.05%¹, the reported incidence of IMC following HCT ranges from ~2–10%². Complications of IMC in the transplant population include graft failure and death, with reports of mortality ranging from 9 – 50%^3,4. Previous reports identify HCT for non-malignant disorders (NMD) as one of the greatest risks for IMC development⁴.

There is great interest in identifying post-HCT IMC risk factors. Transplant type, hematopoietic stem cell source, conditioning regimen, graft-vs-host disease (GvHD), and cytomegalovirus (CMV) reactivation are examples of factors with various reported degrees of association². Use of alemtuzumab as part of the preparative regimen results in prolonged lymphopenia with additional effects on regulatory T cell function and monocyte function⁵. Cyclosporine also disrupts immune reconstitution via impairment of thymic-dependent clonal deletion⁶. Therefore, agents used during conditioning and following transplant have been linked to IMC development. It has been hypothesized that the preferred use of these agents in individuals receiving alternate donor HCT explains the increased incidence of IMC associated with unrelated donors⁷. Interestingly, in the solid organ transplant population, patients receiving intensive immunosuppression with alemtuzumab, daclizumab, and mycophenolate mofetil demonstrated improvement of hemolytic anemia after dose reduction⁸. Finally, infectious etiologies including CMV and Epstein-Barr Virus (EBV) have been associated with IMC development, theoretically secondary to molecular mimicry or B-cell expansion⁹. NMD as an HCT indication has been identified as a risk factor for IMC in pediatric cohorts^4,10–12. Lack of prior chemotherapy and use of non-myeloablative conditioning regimens increase the risk for mixed donor and recipient chimerism after donor engraftment. It has been hypothesized that the resulting immune dysregulation contributes to the risk of IMC development¹³. Unfortunately, many of these risk factors co-occur in individual patients making identification of independent contribution to IMC challenging.

Once identified, the treatment of IMC is controversial. Cytopenias are often clinically significant and refractory to immune suppression. Proposed treatment strategies are not informed by prospective studies^14–17. Various agents have been employed in the treatment of IMC following HCT. Treatment modalities can be broadly conceptualized as systemic immune suppression, targeting antibody producing cells, increasing bone marrow production of relevant cell lines, and finally, direct antibody removal from circulation (Table S1). This retrospective study aimed to investigate the incidence and duration of IMC in a large cohort of pediatric NMD patients and to determine pre- and post-transplant IMC risk factors.

MATERIALS AND METHODS

Patient selection

We identified patients aged 0 to 25 years who underwent allogeneic HCT for NMD between 2010 and 2017 at the University of Minnesota. In accordance with the Declaration of Helsinki, all parents/guardians provided Institutional Review Board (IRB)-approved informed consent for HCT related research. A prospectively-maintained, institutional database provided patient sex, age at HCT, indication for HCT, conditioning agents, donor and graft sources, GvHD prophylaxis, donor/recipient CMV status, and GvHD grade if present. Conditioning regimens were classified as myeloablative (a combination of two alkylators or use of full dose irradiation), reduced toxicity (myeloablative but less toxic by use of a single alkylator in combination with fludarabine and serotherapy), reduced intensity (lower dose alkylator and/or irradiation exposure resulting in incomplete myeloablation and mixed donor chimerism), or none¹⁸.

Characterizing immune mediated cytopenias

Cases of IMC were defined by positive direct Coombs test (DAT), anti-platelet antibody, and/or anti-granulocyte antibody. Patients with cytopenias but negative serologies were not classified as IMC due to the poor reliability of distinguishing competing causes of cytopenias in a retrospective review. The time of diagnosis was defined by the time of positive laboratory testing, sooner if data demonstrated cell-line decline prior to obtaining diagnostic testing. IMC duration was defined by months to resolution: Brief (0–3), Intermediate (3–6), or Prolonged (>6). We use duration of IMC to capture clinical morbidity in one measure, irrespective of number of cell lines affected, need for multiple agents or therapies, or in-versus out-patient management which can vary between providers. Disease resolution was defined as the time that therapy was discontinued, or for patients treated supportively with blood product, by the date of the last transfusion. With some patient-to-patient variability, our standard of care approach to single cell line cytopenia included steroids + IVIG, with rituximab often added upfront when multiple cell lines are affected. Sequential addition to steroids of IVIG, then rituximab, then bortezomib following 2–4 weeks of observation for response was typical. Alternative approaches and agents as shown in Table S1 were often employed in refractory IMC.

Risk factor analysis

Baseline patient and transplant characteristics, post-transplantation complications, and serially documented outcomes including peripheral blood donor chimerism and immune reconstitution were prospectively collected for all HCT patients. A subset of patients had consented for co-enrollment in an IRB-approved investigation of immune reconstitution, yielding extended lymphocyte subset peripheral blood analyses.

Statistical analysis

Demographic, transplant characteristics and immune reconstitution were summarized by standard descriptive statistical methods. Statistical comparison of categorical variables was performed by Chi-square test and Kruskal-Wallis (Wilcoxon) rank-sum test was used for comparison of continuous variables between patients with IMC (n = 50) and patients without IMC (n = 221). This descriptive analysis excluded patients experiencing graft failure or death prior to day 100 after HCT.

The Kaplan-Meier method¹⁹ was used to estimate the probability of 5-year overall survival (OS) with 95% confidence interval (CI) and survival curves were compared using log-rank test between patients with (n = 50) and without (n = 221) IMC. The cumulative incidence of 1 year IMC was calculated reflecting death as a competing risk for total of 297 subjects including patients with IMC (n = 50), without IMC (n = 221) and those experiencing early death or graft failure before day 100 after HCT (n = 26). Patients in the latter group were censored at time of death or graft failure. Gray’s test was used to compare the cumulative curves. Fine and Gray regression²⁰ was used to look at the association between risk factors and IMC. Univariate competing risk showed association of IMC with indication for HCT (Inherited metabolic disorder vs. hemoglobinopathy vs. other disease), corticosteroid use for GvHD prophylaxis (yes vs. no), conditioning intensity (myeloablative vs. reduced toxicity vs. reduced intensity or none), and age at transplant (< 3 years vs. ≥ 3 years). Due to high inter-variable correlations, a new combined variable was created: inherited metabolic disorder at age younger than 3 years, inherited metabolic disorder at age younger of 3 years or older, hemoglobinopathy at any age, other diseases at age younger than 3 years, and other diseases at age of 3 years or older. Fine-Gray regression with the new variable was used to report a multivariate model.

All statistical analyses were implemented using Statistical Analysis System statistical software version 9.4 (SAS Institute Inc., Cary, NC) and R 3.6.2 (The R Foundation for Statistical Computing). P-values were two-tailed with a value of 0.05 deemed significant.

RESULTS

IMC in NMD is common, often prolonged, and can be lethal

Between 2010 and 2017, 297 pediatric patients underwent HCT for NMD. Twenty-six experienced graft failure or death prior to day +100. Of the 271 patients who were at risk, 50 developed IMC (cumulative incidence of 18.4%; Table 1). The median time to diagnosis was 136 days post-HCT (IQR 92–196 days). Eighty percent of patients with IMC had a positive DAT, 44% had an anti-platelet antibody, and 30% an anti-neutrophil antibody. Two or more cell lines (e.g. Evans syndrome or pancytopenia) were affected in 44%.

Table 1:

Descriptive analysis of demographic and HCT characteristics

Transplant Characteristic	All Groups n=271	IMC n=50	No IMC n=221	P-value
Age in years, median (range)	6.5 (0.1 – 24.1)	4.8 (0.1 – 20)	6.9 (0.1 – 24.1)	<0.01
Indication for HCT, No. (%)
Inherited Metabolic Disorder	98	26 (27%)	72 (73%)	<0.01
Hemoglobinopathy	12	6 (50%)	6 (50%)
Idiopathic Severe Aplastic Anemia	98	8 (8%)	90 (92%)
Primary Immune Deficiency	31	6 (19%)	25 (81%)
Epidermolysis Bullosa	32	4 (13%)	28 (87%)
Sex, No. (%)
Male	166	35 (21%)	131 (79%)	0.16
Female	105	15 (14%)	90 (86%)
Conditioning Intensity, No. (%)
Myeloablative	116	16 (14%)	100 (86%)	0.05
Reduced Toxicity	55	17 (31%)	38 (69%)
Reduced Intensity	98	17 (17%)	81 (83%)
None	2	0	2 (1%)
Lymphodepletion, No. (%)
Anti-thymocyte globulin	118	20 (17%)	98 (83%)	0.09
Alemtuzumab	103	25 (24%)	78 (76%)
Neither	50	5 (10%)	45 (90%)
Donor Source, No. (%)
Matched related	63	7 (11%)	56 (89%)	0.54
Mismatched related	8	2 (25%)	6 (75%)
Matched unrelated	123	25 (20%)	98 (80%)
Mismatched unrelated	73	15 (21%)	58 (79%)
Haploidentical	4	1 (25%)	3 (75%)
Graft Source, No. (%)
Bone marrow	171	26 (15%)	145 (85%)	0.20
Umbilical Cord	96	23 (24%)	73 (76%)
Peripheral Blood	4	1 (25%)	3 (75%
Sex Matched Donor, No. (%)
Yes	130	26 (20%)	104 (80%)	0.58
No	138	24 (17%)	114 (83%)
GvHD Prophylaxis, No. (%)
With corticosteroid	30	13 (43%)	17 (57%)	<0.01
Without corticosteroid	241	37 (15%)	204 (85%)
Acute GVHD, No. (%)
Grade 0 – 1	213	41 (19%)	172 (81%)	0.95
Grade 2 – 4	32	6 (19%)	26 (81%)
CMV Status, No. (%)
Recipient (R) positive	120	22 (18%)	98 (82%)	0.92
R negative, Donor positive	26	4 (15%)	22 (85%)
R negative, Donor negative	122	23 (19%)	99 (81%)
Graft Composition, Median (range)
CD34 cell dose (x106)	3.1 (0 – 33.7)	3.2 (0.2 – 24.8)	2.9 (0 – 33.7)	0.94
Total nucleated cell dose (x108)	2.4 (0 – 27.4)	2.4 (0.1 – 10.0)	2.3 (0 – 27.4)	0.42

Forty-two percent of IMC courses were Brief (0–3 months), 26% Intermediate (3–6 months) and 32% Prolonged (>6 months). Most patients with isolated anemia or neutropenia had Brief disease, while most patients with thrombocytopenia also had multiple cytopenias and had Intermediate or Prolonged disease (Figure 1). Overall, 90% of cases required at least supportive care, including blood product transfusions for anemia or thrombocytopenia or granulocyte colony-stimulating factor for neutropenia. For those with Intermediate or Prolonged IMC, all received corticosteroid therapy, with a response rate of 34%. The remainder required additional treatment. Overall, an average of 2 unique agents were used for resolution of IMC, achieved at a median of 118 days (IQR 39–207) after diagnosis. For those with Prolonged IMC, an average of 4.4 agents (range 1 – 8) were utilized.

FIGURE 1: Incidence of immune mediated anemia, thrombocytopenia, neutropenia, and multiple cytopenias and associated clinical severity.

The two most common IMC presentations are isolated anemia and multiple cytopenias. Isolated immune thrombocytopenia was rare. More commonly immune thrombocytopenia occurred as a component of Evans syndrome or pancytopenia. With more cell lines involved, IMC courses were >3 months duration (Intermediate/Prolonged). This figure includes patients at risk for IMC, with early deaths and graft failure excluded.

In the 50 patients with IMC, 6 died, four attributable to IMC (8%). One patient with immune mediated pancytopenia died from infectious complications while on immune suppressive therapy and another with immune thrombocytopenia died from recurrent pulmonary hemorrhage. Two additional patients, one with idiopathic severe aplastic anemia and one with severe congenital neutropenia, developed secondary graft failure associated with refractory trilineage IMC. Each ultimately underwent a second HCT and died from multiorgan dysfunction. There was no difference in overall survival between those with IMC and without (Figure 2).

FIGURE 2: Kaplan-Meier survival curve to assess impact of IMC on outcomes for pediatric NMD patients who survived HSCT for a minimum of 100 days.

While deaths were attributable to IMC, there is no observed difference in overall 5-year survival between groups.

Multiple demographic and transplant characteristics associated with IMC

In a univariate analysis excluding early deaths and graft failure as being non-representative of the control no IMC group, multiple factors were associated with IMC (Table 1). A Gray test was performed with IMC as a time dependent variable, including early deaths and graft failure, with equivalent risk factors identified (Table S2). Those include younger age at HCT (median 4.8 vs. 6.9 years, p<0.01), inherited metabolic disorder or hemoglobinopathy as indication for HCT (p<0.01), reduced toxicity conditioning (p=0.05) and GvHD prophylaxis with steroids (p<0.01). The use of serotherapy with anti-thymocyte globulin or alemtuzumab trended towards significance (p = 0.09). Neither recipient sex, donor source, graft source, graft composition (CD34⁺ and total nucleated cell doses, CD3⁺ cell dose unavailable), recipient/donor sex match, presence of acute GvHD, nor CMV status of the recipient or donor were associated. We tested CMV infection, EBV infection, and grade 2–4 acute GVHD as time dependent covariates, and none were significant (hazard ratio (95% CI) 0.69 (0.28 – 1.23), 2.4 (0.6 – 9.4), and 1.0 (0.36 – 2.8) respectively). Following transplant, the average infection density for viremia of any type (1000 * documented viral infections / survival days) was similar between the IMC group at 1.4 (range 0.5 – 12.3) and the unaffected group at 1.5 (0.0 – 13.9; p = 0.48).

Multiple factors potentially associated with IMC post-HCT are highly correlated. Among transplant indication, conditioning intensity, age, lymphodepletion, and use of steroid for GVHD prophylaxis, only lymphodepletion was not correlated with transplant indication. A new variable combining transplant indication and age was used for the Fine-Gray multivariate regression model (Figure 3). Since only one hemoglobinopathy patient was under 3 years old, this transplant indication was not stratified by age. Lymphodepletion was non-significant (p = 0.14), so the final model is reduced to the transplant indication and age variable (Table 2). We found significantly increased 1 year cumulative incidence of IMC in children with inherited metabolic disorder under 3 years (4.8, 2.3–10.3 95% CI) and for those with hemoglobinopathy transplant indication (7.8, 2.8 – 21.4).

FIGURE 3:

One year cumulative incidence of IMC by transplant indication and age variable

Table 2:

Fine-Gray competing risk multivariate regression model of one year IMC incidence, with early death and graft failure censored at time of event.

Potential Risk Factor	Hazard Ratio (95% CI)	P value
Inherited metabolic disorder Age < 3 years	4.8 (2.3 – 10.3)	<0.01
Inherited metabolic disorder Age ≥ 3 years	2.1 (0.86 – 5.0)
Hemoglobinopathy	7.8 (2.8 – 21.4)
Other transplant indication Age < 3 years	2.3 (0.88 – 5.9)
Other transplant indication Age ≥ 3 years	1.00

Decreased donor contribution to and absolute T-lymphocyte recovery associated with IMC

Following HCT, flow cytometry of peripheral blood to assess immune reconstitution of lymphocyte subsets is routinely measured at days +60, +100, +180, and +365 (Table 3). No association was found between IMC and absolute CD19+ or CD8+ lymphocyte counts, but there was an association between lower absolute CD4+ count and IMC at days +100 (98 vs 158, p < 0.01) and +180 (170 vs 276, p < 0.01). To expand on the later finding, the difference between CD4+ counts amongst IMC patients treated by observation and supportive care alone were not statistically different from those treated with immunosuppressive agents at days +100 and +180 (p = 0.96 and p = 0.21, respectively), indicating that the decrease in CD4+ count, particularly at day +100, is not solely reflective of therapy administered to IMC patients. Donor chimerism, as evidenced by the percent of donor peripheral blood CD3+ chimerism was obtained at day +60, day +100, and day +180 (Figure 4). Patients with NMD who developed IMC demonstrated statistically significantly lower CD3+ T lymphocyte donor contribution at day +180 post-HCT (p < 0.01). When looking at individual patient kinetics, more patients with IMC demonstrated a decline by ≥10% in CD3+ T lymphocyte donor chimerism between days 60–100 (24% vs 12%, trend toward significance with p = 0.06) and 100–180 (18% vs 5%, p < 0.01). Nonetheless, the majority of patients in each group had no change or in their donor chimerism percentage (56% vs 67% from day 60–100 and 52% vs 54% from day 100–180).

Table 3:

Association of post-HCT lymphocyte recovery with IMC

Lymphocyte population	Total number of patients with data in n=271 cohort	Absolute cells/uL median (range)		P-value
		With IMC	Without IMC
CD4+ T lymphocytes
day +60	213	107 (0 – 427)	108 (0 – 1023)	0.83
day +100	238	98 (3 – 409)	158 (2 – 1213)	<0.01
day +180	217	170 (10 – 1467)	276 (8 – 2718)	<0.01
day +365	189	528 (33 – 2217)	588 (45 – 2378)	0.06
CD8+ T lymphocytes
day +60	213	50 (2 – 896)	58 (0 – 1118)	0.85
day +100	238	73 (0 – 1036)	91 (0 – 4213)	0.46
day +180	217	117 (0 – 1838)	158 (8 – 4940)	0.40
day +365	189	304 (18 – 2820)	405 (0 – 4460)	0.29
CD19+ B lymphocytes
day +60	206	103 (0 – 2357)	98 (0 – 3006)	0.92
day +100	233	119 (0 – 2217)	184 (0 – 5009)	0.43
day +180	216	208 (0 – 1529)	234 (0 – 9338)	0.35
day +365	184	366 (0 – 3377)	552 (0 – 5786)	0.03

FIGURE 4:

Median donor contribution to CD3+ peripheral blood lymphocytes declines over time post-HCT in association with IMC.

Patients with NMD who developed IMC, as compared to peers who did not, demonstrated a negative rate of change in peripheral blood CD3+ T lymphocyte donor chimerism between days 60–100 (p = 0.06) and 100–180 (p < 0.01), and the absolute percent donor chimerism at day +180 was also significantly lower (*p < 0.01).

A subset of NMD patients had extended peripheral blood lymphocyte subset data available (n = 27 with IMC, n = 103 without IMC), which included quantification via flow cytometry of T regulatory (T_reg), naïve CD4+ and CD8+, memory CD4+ and CD8+, effector CD4+ and CD8+, RA+ CD4+ and CD8+, natural killer, and B cells at days +100, +180 and +365. Comparing IMC to non-IMC cases, median absolute memory CD4+ and effector CD4+ cell counts were decreased at day +100 (38.4 vs 88.6, p < 0.01; 14.9 vs 30.8, p < 0.01), naïve CD4+ cells were decreased at day +180 (20.1 vs 51.7, p = 0.05), and T_reg cells were decreased (22.5 vs 43.3, p = 0.03) on day +365.

DISCUSSION

As the indications for HCT in non-malignant disorders broadens, and as the safety of transplantation for these patients improves, there is great interest in preventing late complications that contribute to morbidity and mortality. Unfortunately, immune mediated cytopenias are one late complication whose pathogenesis remains elusive and which requires focused efforts at prevention and treatment. The pathogenesis of IMC after HCT is controversial and likely multi-factorial. Alloimmune cytopenias caused by donor antibodies against recipient antigens or vice versa may result from complications including passenger lymphocyte syndrome and ABO blood group incompatibility²¹. True autoimmune cytopenias, i.e. donor antibodies against donor antigens or recipient antibodies against recipient antigens, may result from impaired regulatory T cell function following transplantation and impaired mechanisms of tolerance^14,22. A genetic susceptibility to autoimmunity may also contribute²³.

The hypothesis that mixed chimerism contributes to immune dysregulation is supported in this study where a higher percent of the IMC group demonstrated statistically significant decreases in donor CD3+ lymphoid chimerism. Further immune dysregulation in this population is suggested by the relative suppression of CD4+ cells during immune reconstitution compared to CD8+ cells in those with IMC. Deambrosis, et al, observed an increased incidence of IMC in patients conditioned with busulfan and fludarabine (equivalent to our reduced toxicity conditioning) compared with myeloablative busulfan and cyclophosphamide, with associated increased absolute lymphocyte count on day of transplant in the busulfan and fludarabine group²⁴. These observations may offer a potential avenue of IMC prevention, but increased lymphoablation must be balanced by increased morbidity of treatment in this unique population.

The association of GvHD with IMC is inconsistent in the literature, but some authors propose that the two entities lie on a spectrum of alloimmune reaction of donor lymphocytes to recipient tissue²⁵. IMC has been associated with both acute²⁶ and chronic GvHD²⁵. Retrospective analyses examining this relationship are limited by the numerous confounding factors that occur simultaneously. Additionally, data not typically captured include the acceleration or suspension of GvHD prophylaxis weans in response to IMC. This study of NMD recipients did not reveal an association between acute GvHD and IMC, but the choice of GvHD prophylaxis may be important. Most GvHD prophylaxis protocols included the use of cyclosporine, which has been independently associated with IMC. However, an interesting observation in this study is that the use of corticosteroids for GvHD prophylaxis is associated with increased risk of IMC. It is possible that supratherapeutic corticosteroid levels may prolong the lymphopenic effect of conditioning and contribute to immune dysregulation.

Other factors have been implicated in IMC pathophysiology but were not found to be associated in our analysis. These include infection with EBV and CMV, umbilical cord blood as a graft source, and donor/recipient sex mismatch. Our institution robustly collects infectious data up to day +100 post-HCT, but ability to capture events after this varies depending on where patients continue care. This may be a source of bias when assessing risk factors. Another possibility is that viral infections may contribute to antibody negative IMC and therefore not detected in this study. Many factors associated with IMC are inconsistently replicated in the literature. Our analysis revealed that many previously identified factors associated with IMC correlate with each other. The presence of confounding variables will always be a limitation in interpreting retrospective analyses of IMC.

The clinical morbidity of IMC in this cohort is comparable with prior reports. Of those classified with Intermediate or Prolonged disease, corticosteroid therapy is effective in the minority and often numerous agents are tried before disease control is achieved. In recent years, new strategies to treat refractory IMC are emerging in addition to B-cell depletion. Specifically, targeting plasma cells with bortezomib, a proteasome inhibitor, is becoming more common, as are strategies to target the complement pathway^27,28. The use of daratumumab, a monoclonal antibody directed against CD38 on plasma cells, is also emerging as a treatment option for refractory disease. Multiple small case series^29–32 have documented successful response with this approach, and daratumumab is reported to have tolerable safety profiles in this setting, with prolonged hypogammaglobulinemia being a common adverse effect^30,32–33. Unfortunately, for more refractory disease, the evidence for treatment is limited by small retrospective series. Since agents are usually employed simultaneously, the benefit of each agent individually is difficult to assess, which is true in this study as well.

In this study, more than 90% of patients with immune thrombocytopenia have additional cell lines affected. Immune thrombocytopenia may be the largest contributor to treatment refractoriness given that patients with isolated anemia or neutropenia usually have Brief disease (Figure 1). This observation may offer direction to explore the use of novel therapeutics targeting platelet destruction, e.g. SYK inhibitors³⁴, with the goal to ameliorate Intermediate or Prolonged IMC using less immunosuppressive agents. Along these lines, since Intermediate and Prolonged IMC is usually steroid non-responsive, perhaps the first line use of corticosteroids for these patients needs to be examined. Ultimately, prospective studies with consensus IMC screening and treatment protocols are needed to assess treatment efficacy of proposed regimens.

Our cohort supports the observation that patients undergoing HCT for non-malignant disease are at high risk for developing IMC. However, the limitations of this retrospective study reflect the challenges in studying IMC. At present, there is no consensus on the precise definition of IMC, how to define IMC severity, or how to screen for IMC. These barriers undoubtedly contribute to the lack of prospective IMC studies in the literature, even though the weaknesses of analyses such as this emphasize the need for such research. In this retrospective analysis, we may underreport IMC and underpower our ability to detect group differences by limiting diagnosis to those with positive antibody tests, although we judge based on our manual review of charts that this is not a common issue and most likely results in the brief cases being missed. Further, the many possible, potentially confounding, variables contributing to IMC require a much larger cohort to identify truly independent risk factors.

In conclusion, our data point to future research opportunities aimed at prevention, such as modifications of conditioning and GvHD prophylaxis protocols. Our observation that declining donor lymphocyte contribution is associated with IMC development and that patients with inherited metabolic disorders and hemoglobinopathies are at most increased risk help identify groups that warrant closer surveillance.

Highlights.

• Immune mediated cytopenias complicate stem cell transplant for non-malignant disease

• Pre- and post-transplant factors may increase risk of immune mediated cytopenias

• Mixed donor chimerism and unbalanced CD4/CD8 immune reconstitution increase risk

• Consensus screening and treatment protocols are needed for prospective studies