Sarcopenia-associated CD8⁺ T-cell reconstitution predicts poor outcomes in severe aplastic anemia after hematopoietic stem cell transplantation

Abstract

Sarcopenia has been established as an adverse prognostic factor in severe aplastic anemia (SAA) patients treated with allogeneic hematopoietic stem cell transplantation (HSCT), which is a complex and important process of hematopoietic and immune reconstitution (IR). Our study aimed to investigate the effect of sarcopenia on the longitudinal IR dynamics in SAA patients following HSCT. We retrospectively analyzed lymphocyte subsets, including CD3⁺, CD4⁺, CD8⁺ T cells, CD19⁺ B cells and natural killer (NK) cells, in 279 SAA patients during the first year post-HSCT. Sarcopenia was diagnosed using sex-specific pectoral muscle index (PMI) thresholds calculated from baseline chest CT. Early IR for each cell subset was defined as cell counts at or above the median value at 3 months post-HSCT. The association between early IR and HSCT outcome, as well as factors influencing early IR were evaluated. The cohort (median age of 31 years, range, 24–39 years, 56.3% male) had a baseline sarcopenia prevalence of 49.8% (139/279). Compared with the non-IR group, early IR of CD3⁺, CD4⁺ and CD8⁺ T cells was associated with superior 5-year OS and FFS rates (P < 0.05). Multivariable analysis confirmed early IR of CD8⁺ T cells as an independent OS predictor (HR = 0.193, 95% CI: 0.043–0.873; P = 0.033). However, baseline sarcopenia was independently associated with impaired early IR of CD8⁺ T cells (OR = 1.863, 95% CI: 1.110–3.126; P = 0.019). Addressing baseline sarcopenia may promote early IR of CD8⁺ T cells following HSCT in SAA patients, potentially improving posttransplant clinical outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00277-026-07093-y.

Introduction

Severe aplastic anemia (SAA), characterized by pancytopenia and hypocellular bone marrow (BM), is a life-threatening disorder in which allogeneic hematopoietic stem cell transplantation (HSCT) is the most effective treatment for reconstituting the hematopoietic and immunologic systems [1]. The recovery of the immune system is crucial to the success of HSCT and represents a complex and important process following the procedure. Following HSCT, immunity is deficient due to underlying disease, pretransplant conditioning, or both, and it takes weeks to several months or years for the immune system to reconstitute. Several studies [2, 3] have shown that the rapidity and degree of immune reconstitution (IR) are closely related to the occurrence, development and outcome of transplantation complications, and ultimately affect overall survival (OS). Early and effective IR after HSCT is essential to promote donor cell engraftment, protect against opportunistic infections, and ultimately improve OS. Instead, delayed IR is considered a barrier to the wider application of HSCT in hematologic diseases [4, 5]. These findings suggest that the assessment of early IR may provide tools to better predict and modulate adverse outcomes and subsequently improve survival after transplantation. Notably, current IR paradigms are derived primarily from hematologic malignancies [6–8], whereas SAA is a rare nonmalignant hematologic disorder characterized by autoimmune-mediated destruction of hematopoietic stem cells. Given this fundamental difference, SAA patients typically receive intensive immunosuppressive therapy (e.g., anti-thymocyte globulin/cyclosporine) rather than profound chemotherapy exposure prior to transplantation, potentially resulting in a distinct IR profile post-HSCT [9]. Thus, there is an urgent need to develop specific biological markers to identify characteristics associated with early IR and to predict individual reconstitution trajectories, thereby improving transplant outcomes for patients with SAA.

The kinetics of post-HSCT IR are modulated by transplant-related variables, including the stem cell source, donor type, conditioning regimen, and pharmacologic prophylaxis, and the development and treatment of graft-versus-host disease (GVHD). Furthermore, emerging evidence highlights modifiable peri-transplant parameters, particularly nutritional status and body composition as critical determinants of IR efficiency. Disease-related factors such as psychological stress, physical inactivity, or changes in food intake, may lead to a deterioration in lymphocytes and more specifically T-cell counts and function [10, 11]. Sarcopenia, an objective manifestation of general physical status, has received increasing attention as an early sensitive marker of therapeutic effects on hematological diseases [12, 13] and is a prevalent complication in patients undergoing HSCT due to long-term chemotherapy or immunosuppressive therapy [14, 15]. Moreover, sarcopenia has been recommended as a key factor in determining the eligibility of elderly patients or patients otherwise medically unfit for HSCT [16]. Due to the lack of universal cut-offs for this population, we defined sarcopenia as a pectoral muscle index (PMI) below the sex-specific median on chest-CT, a condition that was closely associated with poor OS and failure-free survival (FFS) in SAA patients undergoing HSCT [17, 18]. Nevertheless, the mechanism by which sarcopenia contributes to the adverse outcomes of SAA patients undergoing HSCT remains unclear. It is even uncertain whether baseline sarcopenia affects the process of IR in SAA patients treated with HSCT. The underlying pathophysiology may involve the interaction between skeletal muscle and immune system. Myokines, such as interleukin (IL)-6, have been demonstrated to play a vital role in regulating immune cell migration, inflammatory responses, and metabolism [19], suggesting that inadequate muscle reserve may impair immunometabolism.

To address these knowledge gaps, this study conducted longitudinal profiling of lymphocyte subset dynamics in SAA patients during the first post-HSCT year. Specifically, we investigated temporal patterns of the reconstitution of lymphocyte subsets, and the prognostic implications of achieving early IR. More importantly, the role of baseline sarcopenia in modulating IR trajectories was investigated. By elucidating muscle mass–immune recovery crosstalk, our findings aim to establish sarcopenia as a predictive biomarker for IR stratification and a potential target for pre-HSCT nutritional interventions.

Methods

Patient selection

This retrospective study included 320 consecutive adult (≥ 18 years old) SAA patients who received their first allogeneic HSCT between January 2017 and March 2022 at our institution. The exclusion criteria were as follows (Fig. 1): (a) association with other tumors, (b) without clinical or imaging data at baseline, (c) without myeloid engraftment or died by day 30, and (d) without immune cell subsets at any post-transplantation time point, specifically including five deaths prior to the 3-month post-HSCT, and seven cases with missing immune cell profiling data. The follow-up data were updated on December 21, 2022. The study was performed according to the declaration of Helsinki. Approval was granted by the Ethics Committee of the Guangzhou First People’s Hospital, and the requirement for written informed consent was waived. The overall study design is outlined in Fig. 2.

Fig. 1.

Flow diagram of the participants in the study. Abbreviations: SAA, severe aplastic anemia; HSCT, hematopoietic stem cell transplantation

Fig. 2.

Study design for biomarker analysis of immune reconstitution in SAA patients undergoing HSCT. Abbreviations: SAA, severe aplastic anemia; HSCT, hematopoietic stem cell transplantation; IR, immune reconstitution; OS, overall survival; FFS, failure-free survival

Immune reconstitution monitoring

At our medical institution, transplant recipients are closely monitored during hospitalization until they achieve clinical stability. As part of routine clinical care, laboratory parameters for assessing IR were monitored twice weekly during inpatient stays and at least weekly during outpatient follow-up. Absolute counts of lymphocyte series in peripheral blood were longitudinally quantified using flow cytometry (Becton Dickinson Biosciences, San Jose, CA, USA) with automated lymphocyte gating, including APC- labelled anti-CD3 monoclonal antibody, BV510- labelled anti-CD4 monoclonal antibody, APC-Cy7-labelled anti-CD8 monoclonal antibody, PE-labelled anti-CD19 monoclonal antibody and APC-labelled anti-CD56 monoclonal antibody. Measurements for T lymphocytes (CD3, CD4, and CD8), B lymphocytes (CD19), and natural killer (NK) cells were obtained at 1, 2, 3, 6 and 12 months post-HSCT in this study. The standard reference ranges used in our medical test laboratory were as follows: CD3⁺T cells, 603–2990 cells/µL; CD4⁺T cells, 441–2156 cells/µL; CD8⁺T, 125–1312 cells/µL; CD19⁺B, 107–698 cells/µL; and NK cells 95–640 cells/µL.

Definitions and evaluation

All non-contrast chest CT scans were performed on a 320 multi-row detector CT scanner (Aquilion ONE; Toshiba Medical Systems, Tokyo, Japan). The scanning protocol for unenhanced chest CT baseline and the corresponding measurements of the pectoral muscle area (PMA) have been described in our previous reports [17]. Sarcopenia is defined as a pectoral muscle index (PMI) that is lower than the sex-specific median at baseline, where the PMI is calculated by dividing the PMA (in squared centimeters, cm²) by the square of height (in meters, m) [17, 18]. Clinical terms related to transplantation were defined according to our previously published studies [18]. All transplant recipients received cyclosporine A, mycophenolate mofetil, and short-term methotrexate for GVHD prophylaxis. The diagnosis and grading of acute GVHD (aGVHD) and chronic GVHD (cGVHD) were performed according to the standard criteria of the transplant center [20, 21]. Neutrophil engraftment was defined as the first day of absolute neutrophil count ≥ 0.5 × 10⁹/L for 3 consecutive days. Platelet engraftment was defined as recovery of a platelet count > 20 × 10⁹/L for the first day of 7 consecutive days without platelet transfusion support. For antiviral prophylaxis, acyclovir was administered to all patients. Viral DNA levels of cytomegalovirus (CMV) and Epstein-Barr virus (EBV) in the peripheral blood were routinely monitored by real-time RT-PCR twice weekly for at least 3 months after transplantation, though quantitative data were not uniformly available for this analysis.

Outcomes of interest

The primary outcome of interest is the impact of early IR on OS and FFS in SAA patients undergoing HSCT, along with an exploration of the key factors influencing early IR, particularly the effect of baseline sarcopenia on early IR. Due to the lack of standardized cut-offs for SAA patients, early IR was defined using median-based thresholds for each lymphocyte subsets (CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells, and NK cells) at 3 months post-HSCT in our study [22]. Accordingly, patients who had a higher cell counts (≥ median) were classified as the early IR group, whereas those with lower counts (< median) comprised the non-IR group. This definition effectively distinguished patients with relatively rapid and slow IR. OS was defined as the time from transplantation to death or the last follow-up. FFS was defined as survival with a response, excluding primary or secondary graft failure, relapse, and death.

Statistical analysis

Clinical and transplant characteristics were summarized via standard descriptive statistics. The Mann–Whitney U test for continuous data, and χ² test or Kruskal–Wallis H test for categorical variables was performed to compare the clinical and transplant characteristics between the sarcopenic and non-sarcopenic group. The immune cell counts and proportions are summarized as medians (ranges, 25th–75th percentiles). The Mann–Whitney U test was used to compare the differences in immune cell counts between two adjacent time points within 1 year after HSCT. Logistic regression analysis was further performed to assess the significant factors associated with early IR in immune cells. Additionally, Kaplan–Meier curves were used to evaluate the cumulative rates of OS and FFS between the early IR and non-IR groups for every immune cell. The independent prognostic factors of OS and FFS were assessed by Cox proportional hazards regression analysis, and the factors with P < 0.1 in the univariate analysis were included in the multivariate analysis to determine the hazard ratio (HR) and 95% confidence interval (CI). A backward stepwise selection algorithm and only the statistically significant variables were retained in the final mode. Finally, the prognostic accuracy of the model was further evaluated via receiver operating characteristic (ROC) curves. Statistical analyses were performed with SPSS software (ver. 21.0, SPSS, Inc., Chicago, IL, USA) and R software (R Foundation for Statistical Computing, Vienna, Austria). Two-sided P values < 0.05 were considered indicative of statistical significance.

Results

Patient characteristics and transplant outcomes

In total, 279 adult SAA patients were enrolled at baseline and patients’ characteristics were compared between sarcopenic and non-sarcopenic group (Table 1). Nearly one-fourth (66/279, 23.7%) of the patients were diagnosed with very severe aplastic anemia (VSAA). The median age at HSCT was 31 years (range, 24–39 years). Additionally, the median PMI was 7.6 cm²/m²(range, 6.4–9.3) for all patients at baseline, 8.9 cm²/m²(range, 7.4–10.1) for males, and 6.6 cm²/m²(range, 5.5–7.4) for females. The median times required for neutrophil and platelet engraftment were 12.0 (range, 10.0–14.0) days and 12.0 (range, 10.0–14.0) days, respectively. The 100-day cumulative incidences of grades 0-I and grades II-IV aGVHD were 73.1% and 26.9%, respectively. Fifty-seven patients developed cGVHD (32 limited and 25 extensive). Twenty-seven (27/279, 9.7%) patients died after transplantation during a median follow-up of 33.2 months (range, 14.6–51.5 months). The estimated 1-year, 3-year and 5-year OS rates of all patients were 93.9%, 91.0% and 91.0%, and the estimated 1-year, 3-year and 5-year FFS rates were 93.2%, 89.2%, and 88.5%, respectively. Ultimately, 139 (49.8%) patients were diagnosed with sarcopenia, and 140 (50.2%) patients were considered non-sarcopenic. SAA patients in the sarcopenic group were characterized by a significantly lower PMI, BMI and a higher incidence of grade III-IV aGVHD compared to the non-sarcopenic group (P < 0.05).

Table 1.

SAA patients’ characteristics and transplant outcomes

Characteristics	Total cohort (n = 279)	Sarcopenia (n = 139)	Non-sarcopenia (n = 140)	P
Clinical data
Age, median (range), years	31.0 (24.0–39.0)	29.0 (23.0–38.0)	32.5 (24.3–39.8)	0.162
Sex, n (%)				0.958
Male	157 (56.3)	78 (56.1)	79 (56.4)
Female	122 (43.7)	61 (43.9)	61 (43.6)
BMI, n (%)				< 0.001
≥ 18.5 Kg/m2	242 (86.7)	108 (77.7)	134 (95.7)
< 18.5 Kg/m2	37 (13.3)	31 (22.3)	6 (4.3)
PMI, median(range), cm2/m2	7.6 (6.4–9.3)	6.4 (5.4–7.6)	9.3 (7.7–10.6)	< 0.001
Transplant data
Diagnosis, n (%)				0.753
SAA	213 (76.3)	108 (77.1)	105 (75.5)
VSAA	66 (23.7)	32 (22.9)	34 (24.5)
Donor type, n (%)				0.847
MSD	90 (32.3)	47 (33.8)	43 (30.7)
MUD	79 (28.3)	39 (28.1)	40 (28.6)
HID	110 (39.4)	53 (38.1)	57 (40.7)
Stem cell source, n (%)				0.864
PB	79 (28.3)	39 (28.1)	40 (29.0)
BM + PB	198 (71.0)	100 (71.9)	98 (71.0)
Data not available	2 (0.7)
Sex of donor to recipient, n (%)				0.921
Female to female	32 (11.5)	17 (12.3)	15 (10.8)
Female to male	49 (17.6)	26 (18.8)	23 (16.5)
Male to male	103 (36.9)	50 (36.2)	53 (38.1)
Male to female	93 (33.3)	45 (32.6)	48 (34.5)
Data not available	2 (0.7)	1 (0.1)	1(0.1)
Conditioning regimen, n (%)				0.713
CY + BU+ATG	129 (46.2)	69 (49.6)	60 (42.9)
FLU + CY+ATG	62 (22.2)	29 (20.9)	33 (23.6)
TBI + CY+ATG + FLU	74 (26.5)	34 (24.5)	40 (28.6)
Others	14 (5.0)	7 (5.0)	7 (5.0)
Median CD34+T in graft, ×106 (range)	3.6 (2.4–5.0)	3.4 (2.1–4.8)	3.7 (2.5-5.0)	0.148
Transplant outcomes
aGVHD, n (%)				0.002
0–I°	204 (73.1)	90 (64.7)	114 (81.4)
II–IV°	75 (26.9)	49 (35.3)	26 (18.6)
cGVHD, n(%)				0.160
None	222 (79.6)	111 (79.9)	111 (79.3)
Limited	32 (11.5)	13 (9.3)	19 (13.6)
Extensive	25 (8.9)	15 (10.8)	10 (7.1)
Neutrophil engraftment, median, days (range)	12.0 (10.0–14.0)	12.0 (10.0–14.0)	12.0 (11.0–14.0)	0.583
Platelet engraftment, median, days (range)	12.0 (10.0–14.0)	12.0 (10.0–14.0)	11.0 (10.0–13.0)	0.069
Follow–up of alive patients, median, months (range)	33.2 (14.6–51.5)	29.4 (13.4–51.7)	33.6 (15.4–50.4)	0.281

PMI pectoralis muscle index, SAA severe aplastic anemia, VSAA very SAA, MSD matched sibling donor, MUD matched unrelated donor, HID haploidentical donor, CY cyclophosphamide, BU busulfan, ATG antithymocyte immunoglobulin, FLU fludarabine, TBI total body irradiation, BM bone marrow, PB peripheral blood, aGVHD acute graft versus host disease, cGVHD chronic graft versus host disease. The bold values were statistically significant

Immune reconstitution

The counts of immune cell subsets, including CD3⁺ T, CD4⁺ T, and CD8⁺ T lymphocytes, CD19⁺ B lymphocytes, and NK cells, within the first year after HSCT were monitored. The immune cells that were present at 1month (n = 244), 2 months (n = 250), 3 months (n = 248), 6 months (n = 229), and 12 months (n = 184) after HSCT were used for the analysis. The median and 25th–75th percentiles of all the immune cell subsets are shown in Fig. 3.

Fig. 3.

Median counts of immune cell and subsets in SAA patients at different time points within the first year after HSCT. (a) CD3⁺ T cells; (b) CD4⁺ T cells; (c) CD8⁺ T cells; (d) CD4⁺/CD8⁺ T cells; (e) CD19⁺ B cells; and (f) NK cells. Error bars indicate the 25th to 75th percentiles. n, represented the number of cases analyzed at each time point. **P values <0.01, ***P values<0.001

Among the adaptive immune T-cell subsets, CD3⁺ T-cell recovery mirrored CD8⁺ T-cell dynamics, reflecting CD8⁺ dominance (Fig. 3a). Notably, CD8⁺ cytotoxic T-cell counts recovered to within the normal reference range (125–1312 cells/µL) by 2 months post-HSCT, showing a significant increase from the 1-month (median: 357 vs. 34 cells/µL; P < 0.001; Fig. 3c). In contrast, CD4⁺ helper T cells exhibited progressive but incomplete recovery, with only 18.5% of patients reaching normal thresholds (> 441 cells/µL) by 12 months (P < 0.01 for monthly increments; Fig. 3b). This disparity drove sustained CD4⁺/CD8⁺ ratio inversion throughout the first year (median 0.33 at 12 months; Fig. 3d).

Compared with T lymphocytes, B lymphocytes demonstrated delayed kinetics: CD19⁺ B lymphocytes were virtually undetectable (median 1 cells/µL) until 3 months, followed by gradual recovery to normal levels (median 180 cells/µL) at 12 months (Fig. 3e). Innate immunity, represented by NK cells, rapidly returned to physiological ranges (median 132.5 cells/µL) within 1 month, after which stable levels were maintained (Fig. 3f).

Impact of early IR on transplant outcomes

In this study, 248 SAA patients were monitored for IR after 3 months. According to the median immune cell count at 3 months, the thresholds defining early IR for CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells and NK cells were 639.0 cells/µL, 73.0 cells/µL, 488.5 cells/µL,1.0 cell/µL, and 165.5 cells/µL, respectively. Among them, 124 patients (50.0%) achieved CD3⁺ T-cell IR, 125 patients (50.4%) achieved CD4⁺ T-cell IR, 124 patients (50.0%) achieved CD8⁺ T-cell IR, 164 patients (66.1%) achieved CD19⁺ B-cell IR, and 124 patients (50.0%) achieved NK-cell IR.

Compared with those in the non-IR group, the 5-year OS and 5-year FFS rates were significantly superior in the early IR group of CD3⁺ T cells (98.4% vs. 91.1%, P = 0.012; Fig. 4a and 97.6% vs. 87.0%, P = 0.002; Fig. 5a), CD4⁺ T cells (97.6% vs. 91.9%, P = 0.046; Fig. 4b and 96.8% vs. 87.8%, P = 0.009; Fig. 5b) and CD8⁺ T cells (98.4% vs. 91.1%, P = 0.013; Figs. 4c and 96.8% vs. 87.9%, P = 0.010; Fig. 5c), respectively. This translated to a significantly lower mortality risk in patients with early CD8⁺ T-cell recovery (HR value was 0.19 in multivariable analysis). However, neither early IR in CD19⁺ B cells (P = 0.130; Fig. 4d and P = 0.450; Fig. 5d) nor early IR in NK cells (P = 0.440; Fig. 4e and P = 0.830; Fig. 5e) was significantly associated with 5-year OS or FFS.

Fig. 4.

Association between the early IR of immune cells and OS: (a) OS according to IR of CD3⁺ T cells. (b) OS according to IR of CD4⁺ T cells. (c) OS according to IR of CD8⁺ T cells. (d) OS according to IR of CD19⁺ B cells. (e) OS according to IR of NK cells. Abbreviations: IR, immune reconstitution; OS, overall survival

Fig. 5.

Association between the early IR of immune cells and FFS: (a) FFS according to IR of CD3⁺ T cells. (b) FFS according to IR of CD4⁺ T cells. (c) FFS according to IR of CD8⁺ T cells. (d) FFS according to CD19⁺ B cells. (e) FFS according to IR of NK cells. Abbreviations: IR, immune reconstitution; FFS, failure-free survival

In the univariable Cox proportional hazards regression analysis (Table 2), clinical and transplant factors (including age, sex, diagnosis, donor type, sex of donor to recipient, conditioning regimen, stem cell source, aGVHD and cGVHD) and immune factors (including CD4⁺ T-cell IR, CD8⁺ T-cell IR, CD19⁺ B-cell IR, and NK-cell IR) were included as candidate variables in the multivariable model. According to the final multivariable analysis, early IR of CD4⁺ T cell (HR, 0.259, 95% CI: 0.086–0.781; P = 0.016) was significantly associated with increased FFS, together with grade II-IV aGVHD (HR, 3.305, 95% CI: 1.340–8.150, P = 0.009). Additionally, only early IR of CD8⁺ T cells (HR, 0.193, 95% CI: 0.043–0.873, P = 0.033) and grade II-IV aGVHD (HR, 9.551, 95% CI: 2.625–34.750; P = 0.001) were independent prognostic factors for 5-year OS in multivariable analysis. Finally, the model was constructed on the basis of independent factors affecting OS, including early IR of CD8⁺ T cells and grade II-IV aGVHD. The area under the curve (AUC) of the ROC curve for assessing the accuracy in SAA patients after HSCT with respect to 1-year, 3-year and 5-year OS was 0.810 (95% CI: 0.668–0.952), 0.822 (95% CI: 0.706–0.938), and 0.838 (95% CI: 0.703–0.972), respectively (Figure S1).

Table 2.

Multivariable Cox regression analysis of OS and FFS in SAA patients undergoing HSCT

Characteristics	OS				FFS
	Univariable		Multivariable		Univariable		Multivariable
	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
Clinical and transplant factors
Age	0.984 (0.946–1.023)	0.416			0.988 (0.953–1.023)	0.495
Sex (Female vs. male)	0.870 (0.404–1.874)	0.722			0.984 (0.490–1.979)	0.965
Diagnosis (VSAA vs. SAA)	1.153 (0.487–2.727)	0.746			0.748 (0.308–1.818)	0.522
Donor type
MSD	Ref		Ref		Ref		Ref
MUD	1.867 (0.446–7.813)	0.393	2.919 (0.322–26.451)	0.341	0.693 (0.227–2.120)	0.521	0.655 (0.180–2.383)	0.521
HID	5.659 (1.674–19.130)	0.005	5.671 (0.702–45.833)	0.104	2.067 (0.904–4.722)	0.085	1.044 (0.366–2.979)	0.936
Sex of donor to recipient
Female to female	Ref				Ref
Female to male	0.956 (0.160–5.722)	0.961			1.292 (0.237–7.053)	0.767
Male to male	1.950 (0.436–8.713)	0.382			2.296 (0.522–10.104)	0.272
Male to female	1.715 (0.376–7.826)	0.486			2.095 (0.469–9.360)	0.333
Conditioning regimen
CY+BU+ATG	Ref				Ref
FLU+CY+ATG	0.546 (0.181–1.644)	0.282			1.095 (0.464–2.584)	0.835
TBI+CY+ATG+FLU	0.834 (0.340–2.046)	0.691			0.949 (0.402–2.238)	0.904
Others	0.761 (0.100-5.791)	0.792			0.704 (0.093–5.347)	0.734
Stem cell source (BM+PB vs. PB)	1.469 (0.593–3.640)	0.406			2.179 (0.837–5.673)	0.111
aGVHD (II-IV vs. 0-I)	5.084 (2.327–11.108)	<0.001	9.551 (2.625–34.750)	0.001	3.448 (1.721–6.908)	<0.001	3.305 (1.340–8.150)	0.009
cGVHD (extensive vs. limited)	3.840 (0.399–36.920)	0.244			3.875 (0.403–37.260)	0.241
Immunologic factors
CD4+ T cells IR (high)	0.292 (0.080–1.062)	0.062	0.719 (0.135–3.838)	0.700	0.256(0.085–0.772)	0.016	0.259 (0.086–0.781)	0.016
CD8+ T cells IR (high)	0.184 (0.041–0.828)	0.027	0.193 (0.043–0.873)	0.033	0.260(0.086–0.783)	0.017	0.440 (0.119–1.619)	0.217
CD19+ B cells IR (high)	0.442 (0.148–1.314)	0.142			0.702(0.282–1.745)	0.446
NK cells (high)	0.643 (0.210–1.968)	0.439			0.911(0.370–2.242)	0.839

SAA severe aplastic anemia, VSAA very SAA, MSD matched sibling donor, MUD matched unrelated donor, HID haploidentical donor, CY cyclophosphamide, BU busulfan, ATG antithymocyte immunoglobulin, FLU fludarabine, TBI total body irradiation, BM bone marrow, PB peripheral blood, aGVHD acute graft versus host disease, cGVHD chronic graft versus host disease, IR immune reconstruction. The bold values were statistically significant

Factors affecting the early IR of CD8⁺ T cells

Factors that may have affected the early IR of CD8⁺ T cells were evaluated via multivariable logistic regression analysis, by adjusted for age, sex, BMI, sarcopenia status, diagnosis, conditioning regimen, donor type, stem cell source, sex of donor to recipient, and aGVHD (Table S1). The results indicated that sarcopenia was an independent risk factor of early IR of CD8⁺ T cells (OR = 1.863, 95% CI: 1.110–3.126, P = 0.019) (Fig. 6).

Fig. 6.

Forest plots showed the independent factors affecting the early IR of CD8⁺ T cells from multivariate logistic regression analysis. Abbreviations: IR, immune reconstitution

Discussion

Our longitudinal investigation delineates characteristic IR patterns in SAA patients post-HSCT, revealing three pivotal findings: (1) The sequential recovery of lymphocyte subsets follows an innate to adaptive hierarchy, progressing through NK cells, CD8⁺ T cells, and CD19⁺ B cells, whereas CD4⁺ T cells exhibit delayed reconstitution beyond 12 months post-transplantation; (2) early IR of CD8⁺ T cells independently predicts superior OS; (3) baseline sarcopenia, quantified via the PMI of chest CT, emerges as a novel modifiable predictor of impaired CD8⁺ T-cell recovery. These insights extend the current understanding of IR dynamics in SAA patients undergoing HSCT and establish muscle mass as a modifiable prognostic factor.

Early and rapid reconstitution of the immune system is essential for improving posttransplant complications and mortality. The current understanding of IR kinetics after HSCT is predominantly derived from patients with malignant hematological diseases, with SAA patients, particularly adults, remaining underrepresented. Heining et al. [23] reported IR in 148 adult patients with hematological diseases who underwent HSCT, including only two SAA patients. The CD3⁺ and CD8⁺ T-cell counts did not normalize until 6 months after HSCT, and the CD4⁺ T-cell and CD19⁺ B-cell counts did not recover during the first year. The IR sequence in adult SAA patients in our study tended to be faster. Among the adaptive T cells, CD8⁺ T cells were the most rapidly reconstructed after 2 months, followed by the return of CD3⁺ T cells after 3 months following HSCT. Despite approximately one-third (33.3%) of patients had undetectable CD19⁺ B cells (0 cells/µL) at 3 months post-HSCT, the cell counts increased significantly from 6 to 12 months and returned to normal within the first year. Notably, the patterns of IR in our study were broadly similar to those reported in other studies involving SAA patients, including 61 pediatric SAA patients aged 14–201 months [2] and 81 adult SAA patients treated with haploidentical stem cell transplantation [24]. These findings may suggest that the type of donor and the chronological age of the recipient may have a limited influence on IR, as was demonstrated in our study.

The distinct IR kinetics in SAA patients necessitate disease-specific thresholds, diverging from those of patients with malignant hematological disorders [25]. Using the 3-month median counts as the threshold for early IR, we found that patients achieving early IR had significantly superior 5-year OS and FFS rates. Furthermore, the central finding was that early IR of CD8⁺ T cells was an independent protective factor for OS (HR = 0.193, P = 0.033), even after adjusting for clinical, transplant and immunologic factors. The impact of early reconstitution of CD8⁺ T cells on transplant outcomes has also been documented in several studies [26–28]. By releasing cytotoxic molecules and cytokines, CD8⁺ T cells play a crucial role in clearing virally infected and tumor cells. The correlation between CD8⁺ T cell reconstitution and protection against relapse, as well as improved disease-free survival and OS, highlights the importance of adequate reconstitution of this subset [26–29]. Given the rapid reconstitution of CD8⁺ T cells in SAA patients at 2 months after HSCT, this may serve as a significant predictive variable for the timely identification of adverse events and the need for therapeutic intervention. Additionally, in our transplantation model, we noted that early IR of CD4⁺ T-cell was associated with improved FFS but not OS. The effects of CD4⁺ helper T cells on outcomes have been controversial across various studies [2, 6, 26, 30]. Therefore, it is essential to design prospective, larger-scale and multicenter studies to further confirm the predictive value of early IR in terms of HSCT outcomes.

Early CD8⁺ T-cell IR is critical for improving OS in SAA patients undergoing HSCT. Therefore, it is essential to evaluate the factors affecting the early IR of CD8⁺ T cells to improve patient outcomes. Previous studies have identified age as an important factor influencing the process of IR, with younger recipients showing a more rapid increase in immune cell counts [24, 31]. Aging of the immune system significantly impacts the kinetics of IR, often resulting in suboptimal clinical outcomes in older patients following HSCT [32]. While immunosenescence theoretically impedes lymphoid recovery, our data corroborate those of Heining et al. [23] that lymphocyte subset trajectories remain comparable across age strata (19–42 years vs. 43–68 years). These contradictory findings regarding age-dependent IR kinetics underscore the need for objective biomarkers transcending chronological parameters.

Crucially, we identified CT-quantified sarcopenia as a critical barrier to CD8⁺ T-cell IR (OR = 1.863, P = 0.019), bridging nutritional status with adaptive immunity. This discovery extends our previous sarcopenia-survival correlations by elucidating its immunomodulatory role. Accumulating evidence indicated a potential link between sarcopenia and impaired immune competence. Previous studies have reported that nutrition and exercise can improve CD8⁺ lymphocyte levels in animal models, suggesting that sarcopenia may contribute to local immune suppression [33]. Mechanistically, skeletal muscle interacts with the immune system partly through the secretion of myokines, such as IL-6, IL-7, and IL-15. These myokines exhibit pleiotropic immunoregulatory functions that extend beyond anti-inflammatory response. For example, IL-6 stimulates B cells differentiation into antibody-producing plasma cells and promotes the growth and differentiation of T cells, especially CD4⁺ T cells, by determining their effector functions [34]. However, SAA patients with sarcopenia might have a milieu of inflammatory or metabolic changes similar to cancer cachexia [35], thereby collectively increase energy expenditure and create a state of metabolic dysregulation, which in turn negatively affects T-cell quantity and function. Hayes et al. [11] reported that factors related to cancer treatment and diagnosis, such as treatment-related inactivity and dietary changes, lead to a deterioration in lymphocytes, specifically T-cell counts and functions, which aligns with our findings. Kim et al. [36] confirmed that exercise programs positively affect the recovery of lymphocyte counts in allogeneic HSCT patients. In the exercise groups, the lymphocyte count increased by an average of 40.9 cells/µL post-HSCT, whereas in the control groups, an average decrease of 640.7 cells/µL was observed. These findings highlight sarcopenia as a modifiable determinant of IR. Sarcopenia, which frequently develops as a consequence of prolonged chemotherapy or immunosuppressive therapy in HSCT patients, is increasingly recognized as an independent risk factor for poor clinical outcomes [15, 37]. Therefore, we suggested that nutritional optimization and guided exercise protocols prior to HSCT may be promising strategies to mitigate sarcopenia and improve immune recovery.

Some limitations of this study should be noted. First, despite enrolling the largest adult SAA cohort to date, our single-center retrospective cohort design introduces potential selection bias. Second, some clinical information, such as CMV reactivation rates, antibiotic exposure histories, and detailed infection severity grading, was not systematically captured. Third, the exclusion of SAA patients who lacked detectable immune cell subsets at all post-transplant time points, may induce an overestimation of OS rates in the analyzed cohort. Finally, the lack of consensus in defining early IR thresholds, particularly for nonmalignant HSCT recipients reflects interstudy variability in different patient populations and monitoring timepoints. Prospective multicenter studies integrating multiomics analyses and serial functional immune monitoring are essential.

In conclusion, adult SAA patients demonstrate sequential IR following an innate-to-adaptive immunological hierarchy during the first year post-HSCT. Notably, early IR of CD8⁺ T cells emerges as an independent prognostic indicator for improved OS, whereas baseline sarcopenia is an adverse factor that negatively impacts the reconstitution of CD8⁺ T cells. This link positions sarcopenia as a modifiable predictor of IR impairment. Consequently, pretransplant interventions combining structured physical exercise and protein-targeted nutritional support may be considered to preserve musculoskeletal integrity, thereby promoting CD8⁺ T cells reconstitution and improving long-term clinical outcomes in SAA patients undergoing HSCT.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Dandan Chen, Xinhua Wei and Zhaohu Yuan performed the research, analyzed the data, and wrote the paper. Wenjian Mo, Weifeng Liu, Xinhua Wei and Zhaohu Yuan designed the study, performed the statistical analysis and analyzed the results. Luyang Gao, Yuan Guo, Yan Zhang, and Mengfan Wang conducted the literature review, collected patients’ data and reviewed the manuscript. All authors gave final approval for the manuscript.

Funding

This work was supported by the Guangdong Basic and Applied Basic Research Foundation (Grant No. 2023A1515220190), and Guangzhou Regional Major Clinical Technology Program (Grant No. 2026 C-ZD046).

Data availability

On a reasonable request, the data supporting study’s findings are available from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.