Introduction
Multiple myeloma (MM) is a hematologic malignancy characterized by abnormal proliferation of clonal plasma cells in the bone marrow, which may be accompanied by extramedullary infiltration. Recent therapeutic advances have contributed to the development of novel agent-based induction therapies, including proteasome inhibitors, immunomodulatory drugs, and monoclonal antibodies. However, high-dose chemotherapy followed by autologous stem cell transplantation (ASCT) is the standard consolidation therapy for eligible, newly diagnosed patients, significantly improving responses and extending progression-free survival [ 1 , 2 ]. The ASCT procedure, which involves high-dose conditioning chemotherapy (such as melphalan) and the subsequent hematopoietic reconstitution phase, presents a physiological challenge to patients and is associated with various complications. In particular, anemia adversely affects early post-transplant quality of life, recovery course, and long-term prognosis in patients with MM undergoing ASCT [ 3 , 4 ].
The complex etiology of transplantation-related anemia has been attributed to the myelotoxicity of the conditioning regimen, the quality and quantity of stem cell infusion, and the delayed recovery of the post-transplant hematopoietic microenvironment [ 5 , 6 ]. However, recent studies have revealed that systemic chronic inflammation mediates the pathophysiology and treatment-related complications of MM, a chronic inflammatory disease [ 7 ]. Tumor cells interact with the bone marrow microenvironment, promoting the production of several pro-inflammatory cytokines (such as interleukin (IL)-6 and tumor necrosis factor-alpha (TNF-α)), establishing a microenvironment conducive to tumor growth, and suppressing hematopoietic function [ 7 ]. Furthermore, this inflammatory state may be further modulated or exacerbated during treatment.
The systemic immune-inflammation index (SII) is a composite indicator derived from peripheral blood neutrophil, lymphocyte, and platelet counts. The levels of SII indicate the pro-inflammatory state (neutrophils), immune function (lymphocytes), and stress response (platelets). Recently, SII has been validated as a robust marker for assessing systemic inflammatory burden and predicting prognosis in various solid tumors and hematologic malignancies [ 8 ]. Inflammation and anemia exhibit bidirectional interaction. Chronic inflammation can lead to “anemia of inflammation” through multiple mechanisms, including the direct inhibition of hematopoietic progenitor cell (especially erythroid progenitor) proliferation and differentiation by pro-inflammatory cytokines, the upregulation of hepcidin that results in iron metabolism dysregulation (functional iron deficiency), and the decreased responsiveness to erythropoietin [ 9 ]. In patients with MM, the inflammatory marker levels are positively correlated with the severity of anemia [ 10 ]. It is worth noting that there is a profound bidirectional interaction between inflammation and anemia. On one hand, chronic inflammation can lead to “anemia of inflammation” through multiple mechanisms. On the other hand, tissue hypoxia associated with anemia can exacerbate systemic inflammatory responses, forming a “inflammation-anemia” vicious cycle [ 11 ]. This association has received preliminary supportive evidence in the general population and patients with chronic diseases [ 12 , 13 ].
However, current studies have predominantly focused on unidirectional investigations, examining the impact of baseline inflammatory markers on MM prognosis or analyzing pre-treatment anemia as an independent prognostic factor [ 14 ]. To our knowledge, there is still a lack of systematic studies investigating the bidirectional association between SII and post-transplant anemia specifically in the context of ASCT. Therefore, this study aimed to explore the bidirectional relationship between pre-ASCT SII and post-transplant anemia in patients with MM through a retrospective cohort study. The findings of this study are expected to provide new clinical clues for identifying high-risk patients and optimizing peri-transplant supportive care.
Methods
Study design and participants
This single-center, retrospective cohort study recruited patients diagnosed with MM who underwent their first ASCT in the Department, the First Affiliated Hospital of Bengbu Medical University between January 2020 and December 2025. The study protocol was reviewed and approved by the Medical Ethics Committee of The First Affiliated Hospital of Bengbu Medical University (Ethics Approval Number: 2024034). As this study only examined the anonymized clinical data, the ethics committee granted a waiver for obtaining individual patient informed consent.
Patient selection
Potential cases were initially screened by querying the hospital medical record management system using “multiple myeloma” and “autologous stem cell transplantation” as the primary diagnosis and procedure keywords, respectively. The data of the final study cohort were retrieved through sequential filtering.
The inclusion criteria were as follows: patients aged ≥ 18 years; patients in whom diagnosis was confirmed according to International Myeloma Working Group (IMWG) criteria via bone marrow aspiration cytology, serum/urine immunofixation electrophoresis, and imaging studies; patients who received high-dose melphalan-based conditioning chemotherapy and completed the first ASCT; patients who had a complete blood count result 1 week before ASCT (with neutrophil counts > 1.0 × 10 9 /L; non-neutropenic phase) for calculating the SII; patients who had complete and traceable clinical follow-up and laboratory examination records for at least 6 months post-transplant (with a focus on hemoglobin monitoring data).
The exclusion criteria were as follows: patients with concurrent active malignancies, uncontrolled severe infections (such as sepsis and pneumonia, requiring intravenous antibiotic therapy 1 week before transplantation), or known autoimmune diseases; patients with other well-defined causes of anemia, including iron deficiency anemia (ferritin < 30 ng/mL and transferrin saturation < 20%), vitamin B12 or folate deficiency, active bleeding, or severe renal insufficiency (estimated glomerular filtration rate (eGFR) < 15 mL/min/1.73 m² or dialysis-dependent); patients who received immunomodulatory therapies other than CD38 monoclonal antibodies (such as daratumumab) before transplantation that can significantly affect lymphocyte or neutrophil counts (including immune checkpoint inhibitors and chimeric antigen receptor T-cell therapy); patients for whom the rate of missing data for variables (including blood count parameters required for SII calculation and baseline Hb value) was > 5%. See Fig. 1 . Flowchart of patient selection.
Fig. 1 Flowchart of patient selection
Data collection and management
Data were extracted independently by two researchers from the electronic medical record system and the hematopoietic stem cell transplantation database. A dual-entry method was employed for data input. Cross-checking and logical validation were performed after establishing an independent database. The multiple imputation by chained equations method was used for imputation for key continuous variables with a missing rate of < 5%, generating 5 complete datasets for subsequent analysis. For variables with a missing rate of > 5% or where non-random missingness was possible, the missing pattern was reported in the Results section, and its potential impact on conclusions was discussed.
During the study period, our hospital implemented a unified red blood cell transfusion criterion (Hb < 80 g/L or presence of obvious hypoxia symptoms), and the transfusion protocol was relatively standardized.
Definitions of study variables
SII
Measurement and Calculation : SII was calculated based on the most recent complete blood count results before ASCT using the following formula: SII = platelet count (PLT, ×10⁹/L) × neutrophil count (NEUT, ×10⁹/L) / lymphocyte count (LYMPH, ×10⁹/L). This integrated index reflects systemic inflammatory activity, immune status and platelet-mediated stress response simultaneously.
Analytical forms: SII was used in two forms in the analysis: (1) as a continuous variable, log10-transformed (log₁₀SII) to improve its distribution; (2) as a binary variable. The optimal predictive cutoff value was determined by receiver operating characteristic (ROC) curve analysis using “occurrence of anemia within 30 days post-transplant” as the endpoint. The cutoff value was 245.5 × 10⁹/L, with an area under the curve (AUC) of 0.72 (95% CI: 0.65–0.79), a maximum Youden index of 0.38, and corresponding sensitivity and specificity of 68.0% and 70.0%, respectively. Based on this cutoff value, patients were divided into high SII group (SII ≥ 245.5 × 10⁹/L) and low SII group (SII < 245.5 × 10⁹/L).
Anemia definition and assessment
Diagnostic Criteria : Anemia was defined according to the World Health Organization criteria (adult males: hemoglobin (Hb) < 130 g/L; adult non-pregnant females: Hb < 120 g/L). Anemia severity was graded as follows based on Hb levels: Mild: 90–129 g/L in males and 90–119 g/L in females; moderate (60–89 g/L); severe (< 60 g/L).
Baseline Anemia: When the Hb level measured at the same time point used for SII calculation met the anemia criteria, it was considered baseline anemia.
Post-transplant anemia outcomes
Incidence: The primary outcome was the occurrence of anemia within 30 days post-ASCT. Post-transplant day 30 was selected as the observation time point because this early phase of hematopoietic reconstitution after high-dose chemotherapy has the highest incidence of anemia and the strongest inflammatory stress, directly affecting early recovery and in-hospital management, making it the most clinically relevant short-term outcome.
Severity: This secondary outcome was defined as the worst anemia grade during the observation period.
Duration: This secondary outcome was defined as the number of days from the first diagnosis of anemia post-transplant to the date when Hb levels returned to non-anemic levels on two consecutive measurements (≥7 days apart). Data were censored if anemia had not recovered by the study cutoff date or if the patient started new treatment due to disease progression or died.
Transfusion indication : During the study period, our hospital uniformly implemented the following red blood cell transfusion criteria: (1) hemoglobin (Hb) < 80 g/L; or (2) Hb ≥ 80 g/L with presence of obvious hypoxia symptoms (e.g., chest tightness, shortness of breath, fatigue, decreased exercise tolerance). The transfusion regimen was determined by the attending physician based on the patient’s clinical status, with a routine single transfusion of 2 units of packed red blood cells.
Covariates (Potential Confounders)
The following covariates were examined: demographic data (age and sex); disease characteristics (International Staging System (ISS)/R-ISS stage, cytogenetic risk, and presence of extramedullary disease at diagnosis); treatment information (induction regimen pre-transplant and line of therapy and infused CD34⁺ cell count); transplant-related events (time to neutrophil/platelet engraftment and incidence of infection within 100 days post-transplant); laboratory indicators (baseline C-reactive protein (CRP), albumin, lactate dehydrogenase (LDH), and renal function); and best treatment response assessed at 100 days (± 28 days) post-ASCT (according to IMWG criteria).
Sample size estimation
This was an exploratory retrospective analysis. Based on previous data from our center, the incidence rate of anemia within 30 days post-ASCT in patients with MM was approximately 50%. To examine the association between high SII levels and anemia, the PASS 2021 software estimated a minimum required sample size of 162 under the following conditions: α = 0.05 (two-sided); power 1-β = 0.80; expected odds ratio (OR) = 2.0; estimated proportion in the high-SII group = 40%. Based on the potential data missingness and screening exclusions, consecutive cases were planned for inclusion with a final target sample size of approximately 200.
Statistical analysis
Statistical analyses were performed using SPSS 26.0 and R 4.2.1 software. Continuous variables were described as mean ± standard deviation or median (interquartile range), and group comparisons were made using the t-test or Mann-Whitney U test. Categorical variables were described as frequency (percentage), and group comparisons were made using the chi-square test or Fisher’s exact test.
Using “occurrence of anemia within 30 days post-transplant” as the outcome variable, a receiver operating characteristic (ROC) curve was plotted to determine the optimal predictive cutoff value of SII by maximizing the Youden index. To assess potential optimistic bias in the cutoff determination process, 1000 bootstrap resampling iterations were performed for internal validation of the AUC, and the bias-corrected AUC was calculated.
The primary analysis included bidirectional association tests. For the forward association (SII → anemia), multivariable logistic regression (for the occurrence of anemia) and Cox proportional hazards regression (for the duration of anemia) were used. The models forced the inclusion of a set of pre-defined core confounding variables (age, sex, ISS stage, baseline Hb, whether the induction regimen contained lenalidomide, and post-transplant infection), and additionally included neutrophil engraftment time and platelet engraftment time as covariates. Other candidate variables with P < 0.10 in univariate analysis were selected using a stepwise regression method based on AIC. SII was included both as a categorical variable (high/low) and as a continuous variable (log₁₀SII). For the reverse association (baseline anemia → SII), a multivariable linear regression model was used, with log₁₀SII as the dependent variable, baseline anemia status as the primary independent variable, and adjustment for confounding factors including age, sex, CRP, ISS stage, LDH, albumin, and eGFR.
To assess the robustness of the results, subgroup analyses (stratified by ISS stage, age, baseline anemia, and post-transplant infection) were performed. Interaction tests were conducted using logistic regression models by adding a product interaction term between SII (binary) and the subgroup variable. An interaction term with P < 0.05 was considered statistically significant.
Using SII, NLR, PLR, and CRP as test variables and “occurrence of anemia within 30 days post-transplant” as the state variable, ROC curves were plotted, and the AUC, optimal cutoff value (based on the Youden index), sensitivity, and specificity were calculated for each indicator. The DeLong test was used to compare the differences in AUC between SII and each single indicator, with P < 0.05 considered statistically significant.
A series of sensitivity analyses were also performed, including: changing the SII grouping criterion (using the median), excluding patients with post-transplant infection, using propensity score matching (PSM, 1:1 nearest neighbor matching, caliper value 0.02) to balance intergroup differences, restricting the analysis to patients who achieved complete response or very good partial response before transplantation, and comparing complete case analysis with multiple imputation results.
All statistical tests were two-sided, and a P value < 0.05 was considered statistically significant.
Results
Patient baseline characteristics
This study enrolled 200 patients with MM who underwent ASCT. The patients were divided equally into the high-SII and low-SII groups ( n = 100) based on baseline SII levels.
As shown in Table 1 , the age (56.3 ± 7.2 years vs. 57.0 ± 7.6 years; P = 0.548) and sex distribution (male proportion: 58.0% vs. 52.0%, P = 0.477) were not significantly different between the high-SII and low-SII groups. Similarly, the ISS stage, cytogenetic risk stratification, and incidence of extramedullary disease at diagnosis were not significantly different between the high-SII and low-SII groups (all P > 0.05), suggesting good comparability of disease stage and biological risk between groups.
Table 1 Comparison of Baseline Data Between Patients in the High and Low SII Groups Baseline Characteristics High SII Group ( n = 100) Low SII Group ( n = 100) Test Statistic P Value Demographic Characteristics Age, , years 56.3 ± 7.2 57.0 ± 7.6 t=-0.602 0.548 Gender, Male, n (%) 58 (58.0) 52 (52.0) = 0.727 0.394 Disease Characteristics ISS Stage, Stage Ⅰ-Ⅱ / Ⅲ, n (%) 71 (71.0) / 29 (29.0) 62 (62.0) / 38 (38.0) = 1.818 0.178 Cytogenetic Risk, Standard / High, n (%) 77 (77.0) / 23 (23.0) 73 (73.0) / 27 (27.0) = 0.427 0.514 Extramedullary Lesion at Diagnosis, Absent / Present, n (%) 82 (82.0) / 18 (18.0) 85 (85.0) / 15 (15.0) = 0.327 0.568 Baseline Anemia Status, Absent / Present, n (%) 43 (43.0) / 57 (57.0) 58 (58.0) / 42 (42.0) = 4.500 0.034 Baseline Hb Level, Median (IQR), g/L 111.0 (98.0 ~ 122.0) 129.5 (118.0 ~ 135.0) Z=-3.022 0.003 Treatment-related Characteristics Lenalidomide-containing Induction Regimen Before Transplantation, Absent / Present, n (%) 36 (36.0) / 64 (64.0) 39 (39.0) / 61 (61.0) = 0.085 0.770 Lines of Treatment Before Transplantation, 1 / ≥2, n (%) 86 (86.0) / 14 (14.0) 88 (88.0) / 12 (12.0) = 0.177 0.674 Infused CD34⁺ Cell Count, Median (IQR), ×10⁶/kg 4.5 (4.1 ~ 4.9) 4.7 (4.1 ~ 5.1) Z=-1.745 0.081 Depth of Response Before Transplantation, CR/VGPR/PR and Below, n (%) 41 (41.0) / 45 (45.0) / 14 (14.0) 43 (43.0) / 36 (36.0) / 21 (21.0) = 2.448 0.294 Laboratory Indicators Baseline CRP, , mg/L 12.5 ± 3.9 8.3 ± 3.5 t = 7.947 <0.001 NLR, median (IQR) 3.8 (2.8–4.9) 2.5 (1.9–3.2) Z=-6.234 <0.001 PLR, median (IQR) 145.2 (120.5-175.6) 110.3 (95.2-130.8) Z=-5.876 <0.001 Baseline Albumin, , g/L 35.3 ± 2.7 37.5 ± 2.8 t=-5.864 <0.001 Baseline LDH, , U/L 226.3 ± 42.7 188.5 ± 27.1 t = 7.478 <0.001 Baseline eGFR, , mL/min/1.73 m² 74.5 ± 10.9 80.4 ± 9.8 t=-3.986 <0.001 Baseline SII, Median (IQR), ×10⁹/L 305.5 (270.8 ~ 345.0) 188.5 (166.5 ~ 211.2) Z = 11.927 <0.001 log₁₀SII, Median (IQR) 2.5 (2.4 ~ 2.5) 2.3 (2.2 ~ 2.3) Z = 11.927 <0.001 Neutrophil engraftment time, median (IQR), days 11.0(10.0–12.0) 11.0(10.0–12.0) Z=-0.821 0.412 Platelet engraftment time, median (IQR), days 13.0(12.0–15.0) 13.0(11.0–14.0) Z=-1.204 0.229 Statistical measures: Z values (Mann-Whitney U test) were used for continuous variables with non-normal distributions; χ² values were used for categorical variables High-risk cytogenetics definition: t(4;14), t(14;16), del(17p). CR: complete response; VGPR: very good partial response; PR: partial response Depth of response pre-transplant: Assessed according to IMWG criteria. This variable was balanced between the two groups ( P = 0.294) and therefore did not require additional adjustment Neutrophil engraftment was defined as a neutrophil count > 0.5 × 10⁹/L for 2 consecutive days Platelet engraftment was defined as a platelet count > 20 × 10⁹/L for 7 consecutive days without transfusion Abbreviations: IQR : interquartile range; SII : Systemic Immune-Inflammation Index; Hb : hemoglobin; CRP : C-reactive protein; LDH : lactate dehydrogenase; eGFR : estimated glomerular filtration rate
Table 1
Comparison of Baseline Data Between Patients in the High and Low SII Groups
Statistical measures: Z values (Mann-Whitney U test) were used for continuous variables with non-normal distributions; χ² values were used for categorical variables High-risk cytogenetics definition: t(4;14), t(14;16), del(17p). CR: complete response; VGPR: very good partial response; PR: partial response Depth of response pre-transplant: Assessed according to IMWG criteria. This variable was balanced between the two groups ( P = 0.294) and therefore did not require additional adjustment Neutrophil engraftment was defined as a neutrophil count > 0.5 × 10⁹/L for 2 consecutive days Platelet engraftment was defined as a platelet count > 20 × 10⁹/L for 7 consecutive days without transfusion Abbreviations: IQR : interquartile range; SII : Systemic Immune-Inflammation Index; Hb : hemoglobin; CRP : C-reactive protein; LDH : lactate dehydrogenase; eGFR : estimated glomerular filtration rate
| Baseline Characteristics | High SII Group ( n = 100) | Low SII Group ( n = 100) | Test Statistic | P Value |
| Demographic Characteristics | ||||
| Age, , years | 56.3 ± 7.2 | 57.0 ± 7.6 | t=-0.602 | 0.548 |
| Gender, Male, n (%) | 58 (58.0) | 52 (52.0) | = 0.727 | 0.394 |
| Disease Characteristics | ||||
| ISS Stage, Stage Ⅰ-Ⅱ / Ⅲ, n (%) | 71 (71.0) / 29 (29.0) | 62 (62.0) / 38 (38.0) | = 1.818 | 0.178 |
| Cytogenetic Risk, Standard / High, n (%) | 77 (77.0) / 23 (23.0) | 73 (73.0) / 27 (27.0) | = 0.427 | 0.514 |
| Extramedullary Lesion at Diagnosis, Absent / Present, n (%) | 82 (82.0) / 18 (18.0) | 85 (85.0) / 15 (15.0) | = 0.327 | 0.568 |
| Baseline Anemia Status, Absent / Present, n (%) | 43 (43.0) / 57 (57.0) | 58 (58.0) / 42 (42.0) | = 4.500 | 0.034 |
| Baseline Hb Level, Median (IQR), g/L | 111.0 (98.0 ~ 122.0) | 129.5 (118.0 ~ 135.0) | Z=-3.022 | 0.003 |
| Treatment-related Characteristics | ||||
| Lenalidomide-containing Induction Regimen Before Transplantation, Absent / Present, n (%) | 36 (36.0) / 64 (64.0) | 39 (39.0) / 61 (61.0) | = 0.085 | 0.770 |
| Lines of Treatment Before Transplantation, 1 / ≥2, n (%) | 86 (86.0) / 14 (14.0) | 88 (88.0) / 12 (12.0) | = 0.177 | 0.674 |
| Infused CD34⁺ Cell Count, Median (IQR), ×10⁶/kg | 4.5 (4.1 ~ 4.9) | 4.7 (4.1 ~ 5.1) | Z=-1.745 | 0.081 |
| Depth of Response Before Transplantation, CR/VGPR/PR and Below, n (%) | 41 (41.0) / 45 (45.0) / 14 (14.0) | 43 (43.0) / 36 (36.0) / 21 (21.0) | = 2.448 | 0.294 |
| Laboratory Indicators | ||||
| Baseline CRP, , mg/L | 12.5 ± 3.9 | 8.3 ± 3.5 | t = 7.947 | <0.001 |
| NLR, median (IQR) | 3.8 (2.8–4.9) | 2.5 (1.9–3.2) | Z=-6.234 | <0.001 |
| PLR, median (IQR) | 145.2 (120.5-175.6) | 110.3 (95.2-130.8) | Z=-5.876 | <0.001 |
| Baseline Albumin, , g/L | 35.3 ± 2.7 | 37.5 ± 2.8 | t=-5.864 | <0.001 |
| Baseline LDH, , U/L | 226.3 ± 42.7 | 188.5 ± 27.1 | t = 7.478 | <0.001 |
| Baseline eGFR, , mL/min/1.73 m² | 74.5 ± 10.9 | 80.4 ± 9.8 | t=-3.986 | <0.001 |
| Baseline SII, Median (IQR), ×10⁹/L | 305.5 (270.8 ~ 345.0) | 188.5 (166.5 ~ 211.2) | Z = 11.927 | <0.001 |
| log₁₀SII, Median (IQR) | 2.5 (2.4 ~ 2.5) | 2.3 (2.2 ~ 2.3) | Z = 11.927 | <0.001 |
| Neutrophil engraftment time, median (IQR), days | 11.0(10.0–12.0) | 11.0(10.0–12.0) | Z=-0.821 | 0.412 |
| Platelet engraftment time, median (IQR), days | 13.0(12.0–15.0) | 13.0(11.0–14.0) | Z=-1.204 | 0.229 |
However, patients in the high-SII group exhibited a less favorable baseline profile for hematological and inflammation-related indicators. Compared with those in the low-SII group, the incidence of anemia at baseline was significantly higher (57.0% vs. 42.0%; P = 0.034) and the baseline Hb level was significantly lower in the high-SII group [median (IQR): 111.0 (98.0–122.0) g/L vs. 129.5 (118.0–135.0) g/L; P = 0.003].
The high-SII group exhibited significantly higher inflammation-related and tumor burden-related markers, including CRP, LDH, and SII (all P < 0.001). The log₁₀SII levels in the high-SII group were significantly higher than those in the low-SII group ( P < 0.001). Indicators related to nutrition and organ function were significantly downregulated in the high-SII group. In particular, the serum albumin levels and eGFR in the high-SII group were significantly lower than those in the low-SII group (both P < 0.001).
The pre-transplant induction regimen composition, line of therapy, infused CD34⁺ cell count, and depth of response pre-transplant were not significantly different between the high-SII and low-SII groups (all P > 0.05), indicating balanced distribution of transplant strategy and treatment-related factors.
Thus, the high-SII group exhibited enhanced inflammatory responses, severe anemia, and impaired organ function at baseline. However, the demographic characteristics, disease stage, and treatment-related factors in the high-SII group were similar to those in the low-SII group.
Forward association between sii and post-ASCT anemia outcomes
Univariate analysis
As shown in Table 2 , the incidence of anemia within 30 days post-ASCT in the high-SII group was significantly higher than that in the low-SII group (64.0% vs. 36.0%; χ² = 14.580; P < 0.001). In addition to increased risk of anemia occurrence, high SII values were correlated with severe anemia and delayed recovery.
Table 2 Univariate Comparison of Anemia-related Outcomes After ASCT Between Patients in the High and Low SII Groups Outcome Indicators High SII Group ( n = 100) Low SII Group ( n = 100) Test Statistic P Value Incidence of Anemia Within 30 Days After ASCT, n (%) 64 (64.0) 36 (36.0) = 14.580 <0.001 Distribution of Anemia Severity, n (%) = 6.893 0.032 - Mild Anemia 18 (18.0) 24 (24.0) - - - Moderate Anemia 20 (20.0) 11 (11.0) - - - Severe Anemia 19 (19.0) 7 (7.0) - - Incidence of Moderate or Severe Anemia, n (%) 39 (39.0) 18 (18.0) = 5.464 0.019 Duration of Anemia, median (IQR), days 34.0(18.0–56.0) 16.4(7.0–32.0) Z = 4.629 <0.001 Unrecovered Anemia at the End of Follow-up, n (%) 15 (15.0) 5 (5.0) = 4.500 0.034 Anemia diagnosis and grading were based on WHO criteria (Hb < 130 g/L for men, < 120 g/L for non-pregnant women) Inter-group comparisons for mild/moderate/severe anemia were performed using Fisher’s exact test (due to low expected cell counts), with P-values reported separately to reflect the gradient of effect ASCT : autologous stem cell transplantation
Table 2
Univariate Comparison of Anemia-related Outcomes After ASCT Between Patients in the High and Low SII Groups
Anemia diagnosis and grading were based on WHO criteria (Hb < 130 g/L for men, < 120 g/L for non-pregnant women) Inter-group comparisons for mild/moderate/severe anemia were performed using Fisher’s exact test (due to low expected cell counts), with P-values reported separately to reflect the gradient of effect ASCT : autologous stem cell transplantation
| Outcome Indicators | High SII Group ( n = 100) | Low SII Group ( n = 100) | Test Statistic | P Value |
| Incidence of Anemia Within 30 Days After ASCT, n (%) | 64 (64.0) | 36 (36.0) | = 14.580 | <0.001 |
| Distribution of Anemia Severity, n (%) | = 6.893 | 0.032 | ||
| - Mild Anemia | 18 (18.0) | 24 (24.0) | - | - |
| - Moderate Anemia | 20 (20.0) | 11 (11.0) | - | - |
| - Severe Anemia | 19 (19.0) | 7 (7.0) | - | - |
| Incidence of Moderate or Severe Anemia, n (%) | 39 (39.0) | 18 (18.0) | = 5.464 | 0.019 |
| Duration of Anemia, median (IQR), days | 34.0(18.0–56.0) | 16.4(7.0–32.0) | Z = 4.629 | <0.001 |
| Unrecovered Anemia at the End of Follow-up, n (%) | 15 (15.0) | 5 (5.0) | = 4.500 | 0.034 |
The distribution of anemia severity significantly varied between the high-SII and low-SII groups (χ² = 6.893; P = 0.0319). Compared with the low-SII group, the high-SII group exhibited a markedly higher proportion of patients with moderate or severe anemia (39.0% vs. 18.0%; P = 0.019).
The duration of anemia in the high-SII group was longer than that in the low-SII group (34.0(18.0–56.0)days vs. 16.4༈7.0–32.0༉days, P <0.001). Furthermore, the proportion of patients whose anemia had not resolved by the end of follow-up in the high-SII group was significantly higher than that in the low-SII group (15.0% vs. 5.0%; P = 0.034).
These results indicate that a high baseline SII level is significantly associated with the occurrence, worsening, and delayed recovery of post-ASCT anemia.
Multivariate regression analysis
Multivariate regression models were constructed to clarify the independent predictive value of SII for post-ASCT anemia (Table 3 ). The potential confounding factors (age, sex, ISS stage, baseline Hb, induction regimen containing lenalidomide, post-transplant infection, neutrophil engraftment time, platelet engraftment time and depth of pre-transplant response) were incorporated into the model.
Table 3 Multivariate Regression Analysis of SII and Anemia Outcomes After ASCT Independent Variable Outcome 1: Anemia Occurrence Within 30 Days After ASCT (Logistic Regression) Outcome 2: Duration of Anemia (Cox Regression) Unadjusted OR (95% CI) Adjusted OR (95% CI) P Value Unadjusted HR (95% CI) Adjusted HR (95% CI) P Value SII (Dichotomous: High vs. Low) 2.25 (1.28 ~ 3.95) 2.86 (1.52 ~ 5.38) 0.001 0.62 (0.48 ~ 0.80) 0.65 (0.50 ~ 0.85) 0.002 log₁₀SII (Continuous Variable) 1.28 (1.09 ~ 1.51) 1.30 (1.10 ~ 1.50) 0.002 0.70 (0.58 ~ 0.84) 0.72 (0.60 ~ 0.87) 0.001 Adjusted variables (total 9): Age, sex, ISS stage, baseline Hb level, whether the induction regimen contained lenalidomide, post-transplant infection, and depth of pre-transplant response (CR/VGPR vs. PR or below), neutrophil engraftment time, platelet engraftment time OR: Odds ratio (OR > 1 indicates increased risk of anemia occurrence). HR: Hazard ratio (HR < 1 indicates decreased risk of anemia recovery/prolonged duration) Interpretation for continuous variable: For each 1-unit increase in log₁₀SII, the risk of anemia occurrence increases by 30%, and the risk of anemia recovery decreases by 28% The Cox regression models satisfied the proportional hazards assumption based on Schoenfeld residual tests ( P > 0.05)
Table 3
Multivariate Regression Analysis of SII and Anemia Outcomes After ASCT
Adjusted variables (total 9): Age, sex, ISS stage, baseline Hb level, whether the induction regimen contained lenalidomide, post-transplant infection, and depth of pre-transplant response (CR/VGPR vs. PR or below), neutrophil engraftment time, platelet engraftment time OR: Odds ratio (OR > 1 indicates increased risk of anemia occurrence). HR: Hazard ratio (HR < 1 indicates decreased risk of anemia recovery/prolonged duration) Interpretation for continuous variable: For each 1-unit increase in log₁₀SII, the risk of anemia occurrence increases by 30%, and the risk of anemia recovery decreases by 28% The Cox regression models satisfied the proportional hazards assumption based on Schoenfeld residual tests ( P > 0.05)
| Independent Variable | Outcome 1: Anemia Occurrence Within 30 Days After ASCT (Logistic Regression) | Outcome 2: Duration of Anemia (Cox Regression) | ||||
| Unadjusted OR (95% CI) | Adjusted OR (95% CI) | P Value | Unadjusted HR (95% CI) | Adjusted HR (95% CI) | P Value | |
| SII (Dichotomous: High vs. Low) | 2.25 (1.28 ~ 3.95) | 2.86 (1.52 ~ 5.38) | 0.001 | 0.62 (0.48 ~ 0.80) | 0.65 (0.50 ~ 0.85) | 0.002 |
| log₁₀SII (Continuous Variable) | 1.28 (1.09 ~ 1.51) | 1.30 (1.10 ~ 1.50) | 0.002 | 0.70 (0.58 ~ 0.84) | 0.72 (0.60 ~ 0.87) | 0.001 |
When SII was considered a dichotomous variable (high vs. low), a high SII value was an independent risk factor for anemia occurrence within 30 days post-ASCT (adjusted odds ratio [aOR = 2.86; 95% CI: 1.52–5.38; P = 0.001) and was significantly associated with delayed anemia recovery (adjusted hazard ratio [aHR = 0.65; 95% CI: 0.50–0.85; P = 0.002).
When SII was considered a continuous variable (log₁₀SII), each unit increase in log₁₀SII value was associated with a 30% increase in the risk of anemia occurrence (aOR = 1.30; P = 0.002) and a 28% decrease in the risk of anemia recovery (aHR = 0.72; P = 0.001).
This dose-response relationship further supports a robust and independent association between SII and post-transplant anemia outcomes.
Kaplan-Meier survival analysis (Fig. 2 ) revealed that the cumulative probability of anemia recovery throughout the follow-up period in the high-SII group was consistently lower than that in the low-SII group. The two curves significantly diverged from an early stage (log-rank test; P < 0.05). The curves provided visual confirmation of the association between high SII levels and delayed anemia recovery.
Fig. 2 Kaplan–Meier Survival Analysis
Comparison of predictive performance between SII and single inflammatory indicators
Using “occurrence of anemia within 30 days post-transplant” as the outcome variable, an ROC curve was plotted to determine the optimal cutoff value of SII. The original AUC of SII for predicting post-transplant anemia at 30 days was 0.72 (95% CI: 0.65–0.79). After internal validation using 1000 bootstrap resampling iterations, the bias-corrected AUC was 0.70 (95% CI: 0.62–0.77), indicating a low risk of model overfitting.
To further evaluate the incremental predictive value of SII compared with traditional single inflammatory indicators, we compared the predictive performance of SII, NLR, PLR, and CRP for anemia within 30 days post-transplant (Table 4 ; Fig. 3 ).
Table 4 Comparison of ROC curves of SII, NLR, PLR, and CRP for predicting post-transplant anemia Indicator AUC 95% CI Optimal cutoff value Sensitivity (%) Specificity (%) Youden index SII 0.72 0.65–0.79 245.5 68.0 70.0 0.38 NLR 0.65 0.58–0.72 3.2 60.0 65.0 0.25 PLR 0.63 0.56–0.70 130.5 58.0 66.0 0.24 CRP 0.62 0.55–0.69 10.5 mg/L 55.0 68.0 0.23 DeLong test results: SII vs. NLR: P = 0.09 SII vs. PLR: P = 0.04 SII vs. CRP: P = 0.03 ROC curves were plotted with "occurrence of anemia within 30 days post-transplant" as the outcome indicator. AUC, area under the curve; CI, confidence interval; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; CRP, C-reactive protein
Table 4
Comparison of ROC curves of SII, NLR, PLR, and CRP for predicting post-transplant anemia
DeLong test results: SII vs. NLR: P = 0.09 SII vs. PLR: P = 0.04 SII vs. CRP: P = 0.03 ROC curves were plotted with "occurrence of anemia within 30 days post-transplant" as the outcome indicator. AUC, area under the curve; CI, confidence interval; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; CRP, C-reactive protein
| Indicator | AUC | 95% CI | Optimal cutoff value | Sensitivity (%) | Specificity (%) | Youden index |
| SII | 0.72 | 0.65–0.79 | 245.5 | 68.0 | 70.0 | 0.38 |
| NLR | 0.65 | 0.58–0.72 | 3.2 | 60.0 | 65.0 | 0.25 |
| PLR | 0.63 | 0.56–0.70 | 130.5 | 58.0 | 66.0 | 0.24 |
| CRP | 0.62 | 0.55–0.69 | 10.5 mg/L | 55.0 | 68.0 | 0.23 |
Fig. 3 Comparison of ROC curves among SII, NLR, PLR, and CRP
ROC curve analysis showed that the AUC of SII was 0.72 (95% CI: 0.65–0.79), which was higher than that of NLR (AUC = 0.65, 95% CI: 0.58–0.72), PLR (AUC = 0.63, 95% CI: 0.56–0.70), and CRP (AUC = 0.62, 95% CI: 0.55–0.69). The DeLong test showed that the AUC of SII was significantly superior to PLR ( P = 0.04) and CRP ( P = 0.03), but the difference compared with NLR did not reach statistical significance ( P = 0.09).
At the optimal cutoff value, SII had a sensitivity of 68.0%, specificity of 70.0%, and Youden index of 0.38, all higher than those of the single indicators. These results suggest that SII, which integrates the three dimensions of neutrophils, lymphocytes, and platelets, has better discriminative ability than most single inflammatory indicators in predicting post-ASCT anemia.
Reverse association between baseline anemia status and SII levels
To explore the bidirectional relationship between inflammation and anemia, the association between baseline anemia status and SII levels was analyzed (Table 5 ). Univariate analysis revealed that the median SII value in patients with anemia at baseline was significantly higher than that in patients without anemia at baseline (200.0 × 10⁹/L vs. 150.0 × 10⁹/L; P = 0.019).
Table 5 Analysis of the Reverse Association Between Baseline Anemia and SII Levels Analysis Type Result Description Test Statistic P Value Univariate Analysis (Mann-Whitney U test) Median SII in the baseline anemia group = 200.0 (IQR: 160.0 ~ 250.0), non-anemia group = 150.0 (IQR: 120.0 ~ 180.0) Z = 2.345 0.019 Multiple Linear Regression Analysis (Dependent variable: log₁₀SII) After adjusting for age, sex, CRP, ISS stage, LDH, serum albumin, and eGFR, the regression coefficient for baseline anemia (yes vs. no) was β = 0.16 (95% CI: -0.02 to 0.34). t = 1.732 0.085 Spearman Correlation Analysis Baseline Hb level was negatively correlated with log₁₀SII ( r =-0.22, 95% CI: -0.35~-0.08) r =-0.22 0.008 Dose-Response Relationship (Multiple Linear Regression) For every 10 g/L decrease in baseline Hb, log₁₀SII increased by an average of 0.08 (95% CI: 0.01 ~ 0.15) t = 2.426 0.025 r: Spearman’s correlation coefficient (a negative correlation indicates that lower Hb is associated with higher log₁₀SII) The multiple linear regression model was adjusted for potential confounders, confirming the independent association between baseline anemia and elevated SII
Table 5
Analysis of the Reverse Association Between Baseline Anemia and SII Levels
r: Spearman’s correlation coefficient (a negative correlation indicates that lower Hb is associated with higher log₁₀SII) The multiple linear regression model was adjusted for potential confounders, confirming the independent association between baseline anemia and elevated SII
| Analysis Type | Result Description | Test Statistic | P Value |
| Univariate Analysis (Mann-Whitney U test) | Median SII in the baseline anemia group = 200.0 (IQR: 160.0 ~ 250.0), non-anemia group = 150.0 (IQR: 120.0 ~ 180.0) | Z = 2.345 | 0.019 |
| Multiple Linear Regression Analysis (Dependent variable: log₁₀SII) | After adjusting for age, sex, CRP, ISS stage, LDH, serum albumin, and eGFR, the regression coefficient for baseline anemia (yes vs. no) was β = 0.16 (95% CI: -0.02 to 0.34). | t = 1.732 | 0.085 |
| Spearman Correlation Analysis | Baseline Hb level was negatively correlated with log₁₀SII ( r =-0.22, 95% CI: -0.35~-0.08) | r =-0.22 | 0.008 |
| Dose-Response Relationship (Multiple Linear Regression) | For every 10 g/L decrease in baseline Hb, log₁₀SII increased by an average of 0.08 (95% CI: 0.01 ~ 0.15) | t = 2.426 | 0.025 |
In the multivariable linear regression model, after adjusting for relevant confounding factors, baseline anemia showed a trend toward association with increased log₁₀SII (β = 0.16, P = 0.085). Spearman correlation analysis further confirmed a significant negative correlation between baseline hemoglobin level and log₁₀SII ( r = -0.22, P = 0.008). Dose-response analysis showed that for every 10 g/L decrease in baseline hemoglobin, log₁₀SII increased by an average of 0.08 ( P = 0.025).
A smooth scatter plot (Fig. 4 ) visually demonstrated this reverse association, suggesting a potential continuous gradient relationship between the severity of anemia and the inflammatory burden.
Fig. 4 Smoothed Scatter Plot
Subgroup analysis and interaction effects
The results of subgroup analysis are presented in Table 6 . The positive correlation between SII and the risk of post-ASCT anemia remained consistent across most clinical subgroups, demonstrating good robustness.
Table 6 Subgroup Analysis and Interaction Test Results (Association between SII and Incidence of Post-ASCT Anemia) Subgroup Variable Subgroup Category Anemia Incidence in High SII Group (%) Anemia Incidence in Low SII Group (%) Adjusted OR (95% CI) P Value (Within Subgroup) ISS_stage - Stage Ⅰ-Ⅱ 46.5 (33/71) 32.3 (20/62) 3.55 (1.28 ~ 9.78) 0.015 ISS_stage - Stage Ⅲ 82.8 (24/29) 57.9 (22/38) 2.91 (1.44 ~ 5.90) 0.003 age_group - < 60 years 62.7 (47/75) 32.9 (23/70) 3.43 (1.73 ~ 6.80) <0.001 age_group - ≥ 60 years 68.0 (17/25) 43.3 (13/30) 2.78 (0.92 ~ 8.42) 0.103 anemia_baseline - Yes 73.7 (42/57) 35.7 (15/42) 5.04 (2.12 ~ 11.95) <0.001 anemia_baseline - No 51.2 (22/43) 36.2 (21/58) 1.85 (0.83 ~ 4.12) 0.157 post_transplant_infection - Yes 65.5 (19/29) 36.4 (12/33) 3.33 (1.17 ~ 9.44) 0.041 post_transplant_infection - No 63.4 (45/71) 35.8 (24/67) 3.10 (1.55 ~ 6.21) 0.002 All subgroup analyses were adjusted for the 7 core confounding variables listed in Table 3 Interaction test: Significant interaction effects were observed between SII and ISS stage ( P = 0.042) and between SII and post-transplant infection ( P = 0.028), indicating that the effect of SII on anemia is stronger in patients with ISS stage III and in those with post-transplant infections
Table 6
Subgroup Analysis and Interaction Test Results (Association between SII and Incidence of Post-ASCT Anemia)
All subgroup analyses were adjusted for the 7 core confounding variables listed in Table 3 Interaction test: Significant interaction effects were observed between SII and ISS stage ( P = 0.042) and between SII and post-transplant infection ( P = 0.028), indicating that the effect of SII on anemia is stronger in patients with ISS stage III and in those with post-transplant infections
| Subgroup Variable | Subgroup Category | Anemia Incidence in High SII Group (%) | Anemia Incidence in Low SII Group (%) | Adjusted OR (95% CI) | P Value (Within Subgroup) |
| ISS_stage | - Stage Ⅰ-Ⅱ | 46.5 (33/71) | 32.3 (20/62) | 3.55 (1.28 ~ 9.78) | 0.015 |
| ISS_stage | - Stage Ⅲ | 82.8 (24/29) | 57.9 (22/38) | 2.91 (1.44 ~ 5.90) | 0.003 |
| age_group | - < 60 years | 62.7 (47/75) | 32.9 (23/70) | 3.43 (1.73 ~ 6.80) | <0.001 |
| age_group | - ≥ 60 years | 68.0 (17/25) | 43.3 (13/30) | 2.78 (0.92 ~ 8.42) | 0.103 |
| anemia_baseline | - Yes | 73.7 (42/57) | 35.7 (15/42) | 5.04 (2.12 ~ 11.95) | <0.001 |
| anemia_baseline | - No | 51.2 (22/43) | 36.2 (21/58) | 1.85 (0.83 ~ 4.12) | 0.157 |
| post_transplant_infection | - Yes | 65.5 (19/29) | 36.4 (12/33) | 3.33 (1.17 ~ 9.44) | 0.041 |
| post_transplant_infection | - No | 63.4 (45/71) | 35.8 (24/67) | 3.10 (1.55 ~ 6.21) | 0.002 |
Interaction analysis revealed significant interaction effects of SII with both ISS stage and post-transplant infection status. The association between high SII levels and anemia risk was pronounced in patients with ISS stage III (interaction P = 0.042). Similarly, the predictive effect of SII on anemia risk was significantly strong in patients who experienced post-transplant infections (interaction P = 0.028).
Among patients with anemia at baseline, high SII levels exhibited a strong correlation with post-transplant anemia. In contrast, this correlation was not significant in patients without anemia at baseline, suggesting that the underlying hematological status influences the predictive performance of SII.
Sensitivity analysis
The results of multiple sensitivity analyses are summarized in Table 7 . The significant association between SII and the occurrence of post-ASCT anemia was consistent irrespective of the SII grouping criteria, the exclusion of patients with post-transplant infections, the application of PSM, the restriction of the analysis to patients with well-controlled disease, or the usage of different methods for handling missing data. This confirms the high robustness of the conclusions.
Table 7 Summary of Sensitivity Analysis Results Sensitivity Analysis Scenario Analysis Content Main Results (Adjusted) Impact on the Core Association 1. Changing SII Grouping Criterion Repeating the Logistic regression in Table 2 after grouping by the median SII of the entire cohort (220 × 10⁹/L) aOR = 2.40 (1.30 ~ 4.43) for the high SII group, P = 0.005 Conclusion unchanged; the association remains significant 2. Excluding Confounding Population Repeating the analysis in Table 2 after excluding patients with post-transplant infection ( n = 47) aOR = 2.65 (1.40 ~ 5.02), P = 0.003 Conclusion unchanged; the effect size is stable, indicating the association is independent of infection 3. Controlling Selection Bias Propensity score matching (1:1, caliper = 0.02); matching variables: age, gender, ISS stage, baseline Hb, depth of response before transplantation; n = 148 after matching aOR = 2.91 (1.50 ~ 5.64) for the high SII group, P = 0.001 Conclusion strengthened; the association is not caused by baseline differences 4. Controlling Disease Burden Repeating the analysis in Table 2 with only patients who achieved CR/VGPR before transplantation ( n = 168) aOR = 2.70 (1.25 ~ 5.85), P = 0.012 Conclusion unchanged; the association still exists in patients with well-controlled disease 5. Handling Missing Data Complete case analysis ( n = 182) vs. Multiple imputation analysis ( n = 200) aOR = 2.78 (1.46 ~ 5.29) for complete cases, aOR = 2.86 (1.52 ~ 5.38) for multiple imputation Conclusion is robust; not affected by missing data handling methods 6. Changing Outcome Definition Repeating the analysis in Table 2 with “anemia occurrence within 100 days after ASCT” as the outcome aOR = 2.95 (1.58 ~ 5.52), P < 0.001 Conclusion consistent; the association still exists in long-term outcomes • aOR: adjusted Odds Ratio; CI: Confidence Interval;• All sensitivity analyses were adjusted for the 7 core confounding variables in Table 2 , and the results were consistent with the main conclusions, confirming that the association between SII and anemia after ASCT is robust and reliable This study consecutively enrolled all MM patients who received their first ASCT at our institution between January 2020 and December 2025. A three-step sequential screening strategy was applied: Step 1, inclusion criteria screening; Step 2, exclusion criteria screening; Step 3, grouping based on the optimal SII cutoff value (245.5) determined by ROC curve analysis. A total of 200 patients were included in the final analysis Abbreviations: ASCT , autologous stem cell transplantation; MM , multiple myeloma; SII , systemic immune-inflammation index; IMWG , International Myeloma Working Group
Table 7
Summary of Sensitivity Analysis Results
• aOR: adjusted Odds Ratio; CI: Confidence Interval;• All sensitivity analyses were adjusted for the 7 core confounding variables in Table 2 , and the results were consistent with the main conclusions, confirming that the association between SII and anemia after ASCT is robust and reliable This study consecutively enrolled all MM patients who received their first ASCT at our institution between January 2020 and December 2025. A three-step sequential screening strategy was applied: Step 1, inclusion criteria screening; Step 2, exclusion criteria screening; Step 3, grouping based on the optimal SII cutoff value (245.5) determined by ROC curve analysis. A total of 200 patients were included in the final analysis Abbreviations: ASCT , autologous stem cell transplantation; MM , multiple myeloma; SII , systemic immune-inflammation index; IMWG , International Myeloma Working Group
| Sensitivity Analysis Scenario | Analysis Content | Main Results (Adjusted) | Impact on the Core Association |
| 1. Changing SII Grouping Criterion | Repeating the Logistic regression in Table 2 after grouping by the median SII of the entire cohort (220 × 10⁹/L) | aOR = 2.40 (1.30 ~ 4.43) for the high SII group, P = 0.005 | Conclusion unchanged; the association remains significant |
| 2. Excluding Confounding Population | Repeating the analysis in Table 2 after excluding patients with post-transplant infection ( n = 47) | aOR = 2.65 (1.40 ~ 5.02), P = 0.003 | Conclusion unchanged; the effect size is stable, indicating the association is independent of infection |
| 3. Controlling Selection Bias | Propensity score matching (1:1, caliper = 0.02); matching variables: age, gender, ISS stage, baseline Hb, depth of response before transplantation; n = 148 after matching | aOR = 2.91 (1.50 ~ 5.64) for the high SII group, P = 0.001 | Conclusion strengthened; the association is not caused by baseline differences |
| 4. Controlling Disease Burden | Repeating the analysis in Table 2 with only patients who achieved CR/VGPR before transplantation ( n = 168) | aOR = 2.70 (1.25 ~ 5.85), P = 0.012 | Conclusion unchanged; the association still exists in patients with well-controlled disease |
| 5. Handling Missing Data | Complete case analysis ( n = 182) vs. Multiple imputation analysis ( n = 200) | aOR = 2.78 (1.46 ~ 5.29) for complete cases, aOR = 2.86 (1.52 ~ 5.38) for multiple imputation | Conclusion is robust; not affected by missing data handling methods |
| 6. Changing Outcome Definition | Repeating the analysis in Table 2 with “anemia occurrence within 100 days after ASCT” as the outcome | aOR = 2.95 (1.58 ~ 5.52), P < 0.001 | Conclusion consistent; the association still exists in long-term outcomes |
Discussion
This study systematically evaluated the association between the SII and post-transplant anemia outcomes in patients with MM undergoing ASCT. The bidirectional relationship between baseline anemia and inflammatory burden was examined. The results showed that: (1) a higher pre-transplant SII level was independently associated with the occurrence of early post-ASCT anemia, increased severity of anemia, and delayed recovery; (2) after adjusting for relevant confounding factors, baseline anemia status showed a trend toward association with higher SII, with a dose-response relationship; (3) the adverse effect of SII was further enhanced in patients with ISS stage III disease and those who developed post-transplant infection, demonstrating significant interaction effects.
These findings suggest that the systemic inflammatory state during the peri-transplant period regulates hematopoietic reconstitution after ASCT and reveal the presence of a potential vicious cycle between inflammation and anemia.
This study demonstrated that high SII levels predict adverse post-ASCT anemia outcomes, which is consistent with the classic theory of anemia of inflammation (AI) [ 6 , 15 ]. Chronic or acute inflammation is reported to suppress erythropoiesis through multiple mechanisms, including cytokine-mediated bone marrow suppression and iron metabolism dysregulation [ 16 ]. High levels of SII, a composite index integrating neutrophil, lymphocyte, and platelet counts, indicate a pro-inflammatory state, impaired immune function, and reactive platelet activation. SII is reported to predict outcomes in various cancers and transplantation settings owing to its good clinical accessibility [ 17 ].In this study, the AUC of SII for predicting post-transplant anemia was 0.72, which was significantly superior to that of PLR (0.63) and CRP (0.62) (DeLong test, both P < 0.05), and was numerically higher than that of NLR (0.65). This suggests that SII, which integrates the three dimensions of neutrophils, lymphocytes, and platelets, may have advantages over single inflammatory indicators in reflecting systemic inflammatory burden and predicting post-transplant anemia. However, the difference in AUC between SII and NLR did not reach statistical significance ( P = 0.09), indicating that their predictive performance was comparable, and further validation with larger sample sizes is needed in future studies. The superior predictive performance of SII may be attributed to its ability to reflect the underlying pathophysiological mechanisms of anemia in this context. Clinically, inflammatory markers such as the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) are often used to aid in distinguishing functional iron deficiency (FID) from absolute iron deficiency due to dietary lack. In FID, inflammation-driven upregulation of hepcidin restricts iron availability for erythropoiesis, a process frequently accompanied by elevated NLR and PLR [ 13 ]. In contrast, absolute iron deficiency without a significant inflammatory background typically does not elevate these ratios. Our finding that SII outperformed PLR and CRP suggests that the integration of neutrophils, lymphocytes, and platelets in the SII formula captures the systemic inflammatory burden mediating anemia more comprehensively than these individual markers. Future studies should directly compare the diagnostic efficacy of SII, NLR, and PLR in differentiating iron deficiency subtypes among MM patients. During ASCT, high-dose conditioning chemotherapy (such as melphalan) can promote extensive tissue damage and disrupt mucosal barriers, inducing the release of pro-inflammatory cytokines, such as IL-6, TNF-α, and IL-1β [ 18 ]. These inflammatory mediators may directly inhibit the proliferation and differentiation of erythroid progenitor cells (burst-forming unit E and colony-forming unit E). Additionally, inflammatory mediators upregulate the expression of hepatic hepcidin, leading to functional iron deficiency and impaired erythropoiesis [ 19 ]. Furthermore, inflammation can diminish bone marrow responsiveness to erythropoietin, exacerbating post-transplant anemia [ 20 ]. In this study, patients in the high-SII group exhibited upregulated levels of CRP and LDH and downregulated levels of albumin levels at baseline, indicating a pathophysiological state of “high inflammation, high tumor burden, and low nutrition.” This is consistent with the inflammation-mediated hematopoietic suppression mechanisms. The increased incidence of anemia, high severity of anemia (proportion of moderate or worse anemia: 39.0% vs. 18.0%), and delayed recovery speed (median duration: 34.0 days vs. 16.4 days) in the high-SII group post-transplantation are clinical manifestations of the inhibitory effect that this persistent and intense inflammatory microenvironment exerts on hematopoietic reconstitution. It is worth noting that there was no significant difference in neutrophil or platelet engraftment time between the high and low SII groups, suggesting that the kinetics of hematopoietic stem cell reconstitution were comparable between the two groups. Therefore, the delayed anemia recovery observed in the high SII group is more likely attributable to the selective inhibition of erythropoiesis by systemic inflammation represented by SII, rather than non-specific overall hematopoietic delay.
This study also demonstrated the independent reverse association between baseline anemia status and elevated SII levels. Thus, anemia may not be a passive consequence of chronic inflammation but can actively exacerbate the systemic inflammation in patients with MM. Several mechanisms may potentially underlie this process. The first mechanism involves hypoxia-induced inflammatory activation. Tissue hypoxia resulting from anemia can stabilize hypoxia-inducible factors, upregulating the expression of pro-inflammatory and pro-angiogenic factors (such as VEGF, IL-6, and TNF-α) and promoting aberrant myeloid cell proliferation [ 21 ]. Meanwhile, the second mechanism involves enhanced oxidative stress. Hypoxia/reperfusion injury associated with anemia can increase the production of reactive oxygen species, activating core inflammatory pathways, such as NF-κB [ 22 ]. Finally, the third mechanism involves compensatory neuroendocrine regulation. To cope with reduced oxygen-carrying capacity, stress-like responses are activated through the activation of the sympathetic nervous system, indirectly affecting the inflammatory state. In this study, baseline Hb was significantly and negatively correlated with log₁₀SII ( r = − 0.22; P = 0.008). For every 10 g/L decrease in Hb, log₁₀SII increased by an average of 0.08 ( P = 0.025). This suggests a continuous gradient relationship between anemia and systemic inflammation through a dose-response pattern. Based on this finding, we propose a hypothesis of an “inflammation-anemia vicious cycle” in patients with MM undergoing ASCT. Baseline inflammation (high SII levels) increases the risk and severity of post-transplant anemia, while pre-existing anemia before transplantation may “pre-activate” or exacerbate systemic inflammation levels. Thus, the inflammatory state is unfavorable even before the transplantation, adversely affecting transplant-related stress tolerance and recovery capacity. Similar inflammation-anemia positive feedback patterns have been proposed in chronic kidney disease, malignancies, and inflammatory diseases [ 10 ]. This study offers preliminary clinical evidence that may be consistent with this model, which requires prospective validation.
Subgroup and interaction analyses further revealed the heterogeneous clinical predictive value of SII. The predictive value of high SII levels in anemia was high in patients with stage III (ISS) tumors (interaction P = 0.042). This may be because patients with advanced-stage tumors exhibit enhanced secretion of inflammatory factors from tumor cells, increased tumor burden, and an inflammatory bone marrow microenvironment [ 23 ]. The predictive value of SII was also significantly enhanced in patients with post-transplant infections (interaction P = 0.028). This indicates that acute infection (a potent “second inflammatory hit”) and pre-existing inflammatory status (high SII levels) exert synergistic inhibitory effects on the fragile post-transplant hematopoietic function [ 24 ]. In patients with anemia at baseline, a high SII level was strongly associated with post-transplant anemia (aOR = 5.04). However, this association was not significant in patients without anemia at baseline. Thus, assessing the inflammatory status (SII) in patients with MM presenting with anemia at baseline is critical for predicting post-transplant hematological recovery. This cohort may benefit the most from targeted anti-inflammatory or hematopoietic support interventions.
The robustness of the findings of this study was confirmed through multiple sensitivity analyses. However, this study has several limitations. First, the retrospective study design cannot completely avoid residual confounding and cannot establish a strict causal relationship. Furthermore, the study enrollment period spanned from 2020 to 2025 (including the COVID-19 pandemic period), during which clinical supportive care, infection control strategies, and transfusion indications may have changed, constituting a potential source of temporal confounding bias. Second, SII is a composite index, and the specific contributions and mechanisms of neutrophilia, lymphopenia, and thrombocytosis cannot be resolved. Third, data on iron metabolism indicators (such as ferritin, transferrin saturation) and cytokine profiles (such as IL-6, hepcidin) are lacking. Although absolute iron deficiency was controlled for in the exclusion criteria, we cannot directly verify the core mechanistic hypothesis that “high SII leads to functional iron deficiency through upregulation of hepcidin.” Fourth, detailed data on peri-transplant red blood cell transfusion and erythropoiesis-stimulating agent (ESA) use were not collected. Because patients in the high SII group had more severe inflammation and poorer hematopoietic recovery, they were clinically more likely to receive transfusions or ESA therapy (treatment allocation imbalance). This may lead to an underestimation of the effect estimates for adverse anemia outcomes in the high SII group (e.g., aOR = 2.86), meaning that the actual interventions partially offset the harm caused by high inflammation, and the observed differences represent the “net effect.” Therefore, the conclusions of this study should be interpreted as associations observed without adjusting for these two key clinical interventions. Fifth, the negative correlation between baseline Hb and log₁₀SII may be mediated by MM tumor burden and shared systemic inflammatory pathways. Moreover, due to the lack of longitudinal post-transplant SII data (days 7, 14, and 30), this study cannot rigorously establish the temporal causality of the “anemia → inflammation” direction, and the bidirectional association should be interpreted as a cross-sectional exploratory finding. Future prospective studies should dynamically monitor iron metabolism parameters, cytokine profiles, and longitudinal SII changes to elucidate the specific mechanisms of the “inflammation-anemia” axis and validate the causal relationship.
Conclusions
This study demonstrated a significant bidirectional association between pre-transplant SII levels and post-ASCT anemia outcomes in patients with MM undergoing ASCT. High SII levels are independent predictors for the occurrence, severity, and delayed recovery of post-ASCT anemia. Meanwhile, baseline anemia status showed a marginal trend toward elevated inflammatory burden. The findings of this study improve the clinical understanding of the critical role of inflammation in post-transplant hematopoietic recovery and support the use of SII as a simple, accessible tool for risk stratification to optimize peri-transplant anemia and inflammation management strategies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Lin Liu, Min Ruan and Meng Wang contributed equally to this work.
Acknowledgements
The authors thank Jiajia Li from Department of Hematology, the first affiliated hospital of Bengbu Medical University for experimental assistance.
Authors’ contributions
Concept and design: Lin Liu, Min Ruan and Meng Wang. Methodology: Lin Liu, Min Ruan, Meng Wang, and Lianfang Pu. Acquisition, analysis, or interpretation of data: Lianfang Pu, Zihan Ling and Jiachang Bo. Drafting of the manuscript: Lin Liu, Min Ruan, Meng Wang and Kai Zhu. Critical revision of the manuscript for important intellectual content: Lin Liu, Min Ruan, Meng Wang and Kai Zhu. Funding acquisition: Lin Liu. Administrative, technical, or material support: Kai Zhu. Supervision: Lin Liu. All authors read and approved the final manuscript.
Funding
This study was supported by the Key Scientific Research Project of Department of Education of Anhui Province (2023AH052011); Bengbu Medical University-Qilu Pharmaceutical Co.Ltd.2025 Clinical Research Collaboration Special Project (BY-QLZDE016); Bengbu Medical University Provincial College Students’ Innovation & Entrepreneurship Training Program (S202510367137); Bengbu Medical University Provincial College Students’ Innovation & Entrepreneurship Training Program (S202510367009).
Data availability
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Clinical Medical Research Ethics Committee of Bengbu Medical University (Ethics Approval Number: 2024034). As this study only examined the anonymized clinical data, the ethics committee granted a waiver for obtaining individual patient informed consent.
Consent for publication
Not Applicable.
Competing interests
The authors declare no competing interests.
Abbreviations
Anemia of inflammation
Autologous stem cell transplantation
C-reactive protein
Hemoglobin
International Staging System
Interquartile range
International Myeloma Working Group
multiple myeloma
Lactate dehydrogenase
Odds ratio
Systemic immune-inflammation index
Tumor necrosis factor-alpha
References
1. Gulla A Anderson KC Multiple myeloma: the (r)evolution of current therapy and a glance into future Haematologica 2020 105 10 2358 67 10.3324/haematol.2020.247015 33054076 PMC7556665 Gulla A, Anderson KC. Multiple myeloma: the (r)evolution of current therapy and a glance into future. Haematologica. 2020;105(10):2358–67. 33054076 10.3324/haematol.2020.247015 PMC7556665
2. Cavo M Gay F Beksac M Pantani L Petrucci MT Dimopoulos MA Autologous haematopoietic stem-cell transplantation versus bortezomib-melphalan-prednisone, with or without bortezomib-lenalidomide-dexamethasone consolidation therapy, and lenalidomide maintenance for newly diagnosed multiple myeloma (EMN02/HO95): a multicentre, randomised, open-label, phase 3 study Lancet Haematol 2020 7 6 e456 68 10.1016/S2352-3026(20)30099-5 32359506 Cavo M, Gay F, Beksac M, Pantani L, Petrucci MT, Dimopoulos MA, et al. Autologous haematopoietic stem-cell transplantation versus bortezomib-melphalan-prednisone, with or without bortezomib-lenalidomide-dexamethasone consolidation therapy, and lenalidomide maintenance for newly diagnosed multiple myeloma (EMN02/HO95): a multicentre, randomised, open-label, phase 3 study. Lancet Haematol. 2020;7(6):e456–68. 32359506 10.1016/S2352-3026(20)30099-5
3. Giralt S, Garderet L, Durie B, Cook G, Gahrton G, Bruno B et al. American Society of Blood and Marrow Transplantation, European Society of Blood and Marrow Transplantation, Blood and Marrow Transplant Clinical Trials Network, and International Myeloma Working Group Consensus Conference on Salvage Hematopoietic Cell Transplantation in Patients with Relapsed Multiple Myeloma. Biol Blood Marrow Transplant, 2015;21(12):2039–2051. 10.1016/j.bbmt.2015.09.016 PMC4757494 26428082
4. Jeevanantham D Rajendran V Tremblay L Larivière C Knight A Evidence-based guidelines for physiotherapy management of patients with multiple myeloma: study protocol Syst Rev 2018 7 1 118 10.1186/s13643-018-0785-7 30111363 PMC6094887 Jeevanantham D, Rajendran V, Tremblay L, Larivière C, Knight A. Evidence-based guidelines for physiotherapy management of patients with multiple myeloma: study protocol. Syst Rev. 2018;7(1):118. 30111363 10.1186/s13643-018-0785-7 PMC6094887
5. van de Donk N Pawlyn C Yong KL Multiple myeloma Lancet 2021 397 10272 410 27 10.1016/S0140-6736(21)00135-5 33516340 van de Donk N, Pawlyn C, Yong KL. Multiple myeloma. Lancet. 2021;397(10272):410–27. 33516340 10.1016/S0140-6736(21)00135-5
6. de Allende LC Díaz de RL Entresotos Conde S Díez Anaemia of chronic diseases: Pathophysiology, diagnosis and treatment Med Clin (Barc) 2021 156 5 235 42 33358297 10.1016/j.medcli.2020.07.035 de Allende LC, Díaz de RL, Entresotos, Conde S, Díez. Anaemia of chronic diseases: Pathophysiology, diagnosis and treatment. Med Clin (Barc). 2021;156(5):235–42. 33358297 10.1016/j.medcli.2020.07.035
7. Lanser L Weiss G Anemia of Inflammation Adv Exp Med Biol 2025 1480 179 95 10.1007/978-3-031-92033-2_13 40603792 Lanser L, Weiss G. Anemia of Inflammation. Adv Exp Med Biol. 2025;1480:179–95. 40603792 10.1007/978-3-031-92033-2_13
8. Li Z, Luo J, Peng L, Wu B, Yu Z. Prognostic significance of systemic immune-inflammation index in hepatocellular carcinoma: a meta-analysis. Clin Transl Oncol. 2026;28(2):601–16. 10.1007/s12094-025-04028-3 40828354
9. Ignatz-Hoover JJ Driscoll JJ Therapeutics to harness the immune microenvironment in multiple myeloma Cancer Drug Resist 2022 5 3 647 61 10.20517/cdr.2022.23 36176763 PMC9511806 Ignatz-Hoover JJ, Driscoll JJ. Therapeutics to harness the immune microenvironment in multiple myeloma. Cancer Drug Resist. 2022;5(3):647–61. 36176763 10.20517/cdr.2022.23 PMC9511806
10. Weiss G Ganz T Goodnough LT Anemia of inflammation Blood 2019 133 1 40 50 10.1182/blood-2018-06-856500 30401705 PMC6536698 Weiss G, Ganz T, Goodnough LT. Anemia of inflammation. Blood. 2019;133(1):40–50. 30401705 10.1182/blood-2018-06-856500 PMC6536698
11. Wang Z Zhang J Luo S Zhao X Prognostic Significance of Systemic Immune-Inflammation Index in Patients With Diffuse Large B-Cell Lymphoma Front Oncol 2021 11 655259 10.3389/fonc.2021.655259 34123808 PMC8187786 Wang Z, Zhang J, Luo S, Zhao X. Prognostic Significance of Systemic Immune-Inflammation Index in Patients With Diffuse Large B-Cell Lymphoma. Front Oncol. 2021;11:655259. 34123808 10.3389/fonc.2021.655259 PMC8187786
12. Nemeth E Ganz T Hepcidin and Iron in Health and Disease Annu Rev Med 2023 74 261 77 10.1146/annurev-med-043021-032816 35905974 PMC9943683 Nemeth E, Ganz T. Hepcidin and Iron in Health and Disease. Annu Rev Med. 2023;74:261–77. 35905974 10.1146/annurev-med-043021-032816 PMC9943683
13. Chen S Xiao J Cai W Lu X Liu C Dong Y Association of the systemic immune-inflammation index with anemia: a population-based study Front Immunol 2024 15 1391573 10.3389/fimmu.2024.1391573 38799419 PMC11116595 Chen S, Xiao J, Cai W, Lu X, Liu C, Dong Y, et al. Association of the systemic immune-inflammation index with anemia: a population-based study. Front Immunol. 2024;15:1391573. 38799419 10.3389/fimmu.2024.1391573 PMC11116595
14. Jaime-Pérez JC Hernández-Coronado M Picón-Galindo E Salazar-Cavazos L Gutiérrez-Aguirre CH Gómez-Almaguer D Results of a completely outpatient autologous stem cell transplant program for lymphoma patients receiving reduced-intensity conditioning Leuk Lymphoma 2021 62 7 1619 28 10.1080/10428194.2021.1876870 33491518 Jaime-Pérez JC, Hernández-Coronado M, Picón-Galindo E, Salazar-Cavazos L, Gutiérrez-Aguirre CH, Gómez-Almaguer D. Results of a completely outpatient autologous stem cell transplant program for lymphoma patients receiving reduced-intensity conditioning. Leuk Lymphoma. 2021;62(7):1619–28. 33491518 10.1080/10428194.2021.1876870
15. Crisell D Talbot N Drakesmith H Iron, Hepcidin, and Immunity Adv Exp Med Biol 2025 1480 197 215 10.1007/978-3-031-92033-2_14 40603793 Crisell D, Talbot N, Drakesmith H. Iron, Hepcidin, and Immunity. Adv Exp Med Biol. 2025;1480:197–215. 40603793 10.1007/978-3-031-92033-2_14
16. Li X Finberg KE Iron Deficiency Anemia Adv Exp Med Biol 2025 1480 163 78 10.1007/978-3-031-92033-2_12 40603791 Li X, Finberg KE. Iron Deficiency Anemia. Adv Exp Med Biol. 2025;1480:163–78. 40603791 10.1007/978-3-031-92033-2_12
17. Fest J Ruiter R Mulder M Groot Koerkamp B Ikram MA Stricker BH The systemic immune-inflammation index is associated with an increased risk of incident cancer-A population-based cohort study Int J Cancer 2020 146 3 692 8 10.1002/ijc.32303 30924141 PMC6916270 Fest J, Ruiter R, Mulder M, Groot Koerkamp B, Ikram MA, Stricker BH, et al. The systemic immune-inflammation index is associated with an increased risk of incident cancer-A population-based cohort study. Int J Cancer. 2020;146(3):692–8. 30924141 10.1002/ijc.32303 PMC6916270
18. Horio T, Morishita E, Mizuno S, Uchino K, Hanamura I, Espinoza JL, et al. Donor Heme Oxygenase-1 Promoter Gene Polymorphism Predicts Survival after Unrelated Bone Marrow Transplantation for High-Risk Patients. Cancers (Basel). 2020;12(2):424. 10.3390/cancers12020424 PMC7072481 32059452
19. Jalouli M. Emerging Role of Hypoxia-Inducible Factors (HIFs) in Modulating Autophagy: Perspectives on Cancer Therapy. Int J Mol Sci. 2025;26(4):1752. 10.3390/ijms26041752 PMC11855875 40004215
20. Taylor CT Colgan SP Regulation of immunity and inflammation by hypoxia in immunological niches Nat Rev Immunol 2017 17 12 774 85 10.1038/nri.2017.103 28972206 PMC5799081 Taylor CT, Colgan SP. Regulation of immunity and inflammation by hypoxia in immunological niches. Nat Rev Immunol. 2017;17(12):774–85. 28972206 10.1038/nri.2017.103 PMC5799081
21. Desantis V Borrelli P Panebianco T Fusillo A Bochicchio D Solito A Comprehensive analysis of clinical outcomes, infectious complications and microbiological data in newly diagnosed multiple myeloma patients: a retrospective observational study of 92 subjects Clin Exp Med 2024 24 1 137 10.1007/s10238-024-01411-2 38937383 PMC11211138 Desantis V, Borrelli P, Panebianco T, Fusillo A, Bochicchio D, Solito A, et al. Comprehensive analysis of clinical outcomes, infectious complications and microbiological data in newly diagnosed multiple myeloma patients: a retrospective observational study of 92 subjects. Clin Exp Med. 2024;24(1):137. 38937383 10.1007/s10238-024-01411-2 PMC11211138
22. Girmenia C Bertaina A Piciocchi A Perruccio K Algarotti A Busca A Incidence, Risk Factors and Outcome of Pre-engraftment Gram-Negative Bacteremia After Allogeneic and Autologous Hematopoietic Stem Cell Transplantation: An Italian Prospective Multicenter Survey Clin Infect Dis 2017 65 11 1884 96 10.1093/cid/cix690 29020286 Girmenia C, Bertaina A, Piciocchi A, Perruccio K, Algarotti A, Busca A, et al. Incidence, Risk Factors and Outcome of Pre-engraftment Gram-Negative Bacteremia After Allogeneic and Autologous Hematopoietic Stem Cell Transplantation: An Italian Prospective Multicenter Survey. Clin Infect Dis. 2017;65(11):1884–96. 29020286 10.1093/cid/cix690
23. Sonneveld P Dimopoulos MA Boccadoro M Quach H Ho PJ Beksac M Daratumumab, Bortezomib, Lenalidomide, and Dexamethasone for Multiple Myeloma N Engl J Med 2024 390 4 301 13 10.1056/NEJMoa2312054 38084760 Sonneveld P, Dimopoulos MA, Boccadoro M, Quach H, Ho PJ, Beksac M, et al. Daratumumab, Bortezomib, Lenalidomide, and Dexamethasone for Multiple Myeloma. N Engl J Med. 2024;390(4):301–13. 38084760 10.1056/NEJMoa2312054
24. Gay F Oliva S Petrucci MT Conticello C Catalano L Corradini P Chemotherapy plus lenalidomide versus autologous transplantation, followed by lenalidomide plus prednisone versus lenalidomide maintenance, in patients with multiple myeloma: a randomised, multicentre, phase 3 trial Lancet Oncol 2015 16 16 1617 29 10.1016/S1470-2045(15)00389-7 26596670 Gay F, Oliva S, Petrucci MT, Conticello C, Catalano L, Corradini P, et al. Chemotherapy plus lenalidomide versus autologous transplantation, followed by lenalidomide plus prednisone versus lenalidomide maintenance, in patients with multiple myeloma: a randomised, multicentre, phase 3 trial. Lancet Oncol. 2015;16(16):1617–29. 26596670 10.1016/S1470-2045(15)00389-7