Splenectomy outcomes in immune cytopenias: Treatment outcomes and determinants of response

Abstract

Background.

Splenectomy is commonly used to treat refractory immune–mediated cytopenia, but there are no established factors that are associated with response to the procedure.

Objectives.

A cohort study was conducted to evaluate the hematologic and surgical outcomes of splenectomy in adult patients with immune cytopenias and identify preoperative factors associated with response.

Methods.

Data from the Cleveland Clinic Foundation for 1824 patients aged over 18 who underwent splenectomy from 2002 to 2020 were analyzed.

Results.

The study found that the most common indications for splenectomy were immune thrombocytopenic purpura (ITP) and autoimmune hemolytic anemia, with a median age of 55 years and median time from diagnosis to splenectomy of 11 months. Hematologic response rates were 74% overall, with relapse in 12% of cases. Postsplenectomy discordant diagnoses were present in 13% of patients, associated with higher relapse rates. Surgery-related complications occurred in 12% of cases, whereas only 3% of patients died from disease complications. On univariate analysis, preoperative factors associated with splenectomy treatment failure were ≥3 lines of pharmacologic treatment, whereas isolated thrombocytopenia, primary ITP, and age ≤40 years had a strong association with response. The multivariable regression confirmed that treatment failure with multiple lines of medical therapy was associated with the failure to respond to splenectomy.

Conclusion.

Overall, the study demonstrates that splenectomy is an effective treatment option for immune-mediated cytopenias with a low complication rate.

Introduction

Approximately 22,000 splenectomy procedures are conducted annually in the United States [1]. Although the spectrum of indications for splenectomy has changed over time, besides trauma and major abdominal surgeries, common current indications include hereditary spherocytosis or immune cytopenias [2]. Classic examples of immune-mediated cytopenias are immune thrombocytopenic purpura (ITP), autoimmune hemolytic anemia (AIHA), autoimmune or idiopathic neutropenia (AIN), and combinations of these conditions such as Evans syndrome (ES). These cytopenias can also occur in the context of clonal lymphoproliferative processes such as T-cell large granular lymphocytic leukemia (T-LGL) or chronic lymphocytic leukemia [3, 4]. In some instances, splenectomy as a procedure for refractory immune cytopenias helps in establishing a correct histopathologic diagnosis [5]. The outcomes of splenectomy have been reported in a few retrospective case series for individual conditions such as ES [6], AIN [7], and AIHA [8]. Less commonly reported outcomes include Felty’s syndrome (FS), wherein splenectomy was historically used prior to the advent of disease-modifying agents with immediate albeit transient resolution of neutropenia and T-LGL with an overall response rate (ORR) of 56% [4]. Despite these reports, prospective randomized trials are currently not available to guide therapeutic decisions.

Over the past two decades, guidelines for ITP treatment have shifted significantly. Initially considered a second-line option, splenectomy is now frequently delayed, becoming a third-line choice after the failure of rituximab and/or thrombopoietin-receptor agonists (TPORa) [9]. The discovery of TPORas in the 2000s marked a significant turning point, and their subsequent integration into medical practice in the 2010s was driven by their dual purpose: either to achieve stable responses or to serve as a bridge to splenectomy [10, 11]. These changes are clearly reflected in the latest American Society of Hematology guidelines and an updated international consensus report [12, 13]. In the context of AIHA, both splenectomy and rituximab have traditionally been considered second-line treatments, with initial expert preference for splenectomy due to its ability to provide a durable response lasting beyond 2 years [14]. However, over the past decade, the use of splenectomy has waned. This decline can be attributed to challenges in identifying individual likelihood to respond to this invasive procedure and concerns related to infection risks [14, 15] to the extent that British recommendations have now positioned splenectomy as a third-line treatment option [16]. Conversely, there are no clear guidelines regarding the timing of splenectomy in AIN [17].

Concomitantly, advances in surgical laparoscopic techniques and mandatory systematic administration of modern vaccines have resulted in shorter hospital length of stays (LOS) with fewer complications, rapid convalescence, and lower costs compared to the previous decades [18]. Moreover, patient preferences may favor long-term durable response over medical management [19].

These considerations alongside the lack of evidenced-based recommendations regarding the timing of splenectomy with respect to other therapies provided us an impetus to compile this study and update older and smaller series. Therefore, the primary aim of this study was to determine the clinical outcomes of adult patients who underwent splenectomy for immune cytopenias following the failure of conventional standard of care therapies to identify the most suitable indications for this procedure and the factors associated with clinical response.

Methods

We retrospectively reviewed a cohort of patients who underwent splenectomy at the Cleveland Clinic between 2002 and 2020 (n = 1824) in accordance with the IRB-approved retrospective data analysis registry protocol. Written informed consent for publication was obtained from patients. Patients were referred to surgery department based on indication from our program as well as directly from other physicians outside our institution.

After excluding cases with established B-cell lymphoproliferative disorders, myeloproliferative neoplasms, and other non-autoimmune hemolytic disorders (Fig. 1), we identified 494 patients with cytopenias, among whom 339 underwent splenectomy for immune-mediated cytopenias as an indication for the procedure. According to the compromised cell lineages and reported splenectomy indications, patients were then categorized into three main groups: ITP (n = 235), AIHA (n = 68), and AIN (n = 10). We also studied outcome of bicytopenias, for example, ES (n = 26) and co-associated conditions such as T-LGL (n = 10). The diagnosis of ITP was made based on the International Working Group criteria [20]. The diagnosis of AIHA was made based on laboratory evidence of hemolytic anemia with a positive direct antiglobulin test (DAT) [21, 22], whereas AIN was diagnosed in the presence of autoantibodies against neutrophils, as per established criteria [23] (for definitions, see Table S1). Hematologic parameters (hemoglobin, platelet, and absolute neutrophil counts) prior to splenectomy, at time of first response, at 26 and 52 weeks, were recorded to document quality and durability of response. Response criteria, including complete response (CR) and partial response (PR) are defined in Table S2. Vaccinations and perioperative antibiotics were administered as per our institutional protocol, encompassing immunizations two weeks before surgery, followed by booster shots 8 weeks afterward. Morbidity was classified into categories, namely intraoperative, short-term (≤30 days), and long-term (>30 days). Mortality from infections was ascertained from medical records.

Fig. 1.

Study population and patients’ selection: A flow diagram showing the selection of the patients included in our study cohort and exclusion criteria. In total 1824 splenectomy patients were identified, among whom 494 had cytopenia. We then excluded 155 patients with already established B-cell lymphoproliferative disorders, myeloproliferative neoplasms, and other non-autoimmune hemolytic disorders. Based on splenectomy indications, we grouped patients into three diagnoses: immune thrombocytopenic purpura (ITP) (n = 235), autoimmune hemolytic anemia (AIHA) (n = 68), Evans syndrome (ES) (n = 26), and autoimmune or idiopathic neutropenia (AIN) (n = 10).

Statistical analysis

We reported categorical variables of response in percentages and counts. Continuous variables were expressed using mean, median, interquartile range (IQR), and range. The association with response to splenectomy was determined using univariate analysis by calculating odds ratios (OR) for response. Subsequently, we performed a multivariable regression analysis as well to quantify the effect of multiple lines of treatment accounting for age and sex. The long-term relapse-free survival (RFS) rates were calculated using the Kaplan–Meier (KM) method. All statistical analyses were two-sided, with a cutoff p-value of 0.05.

Results

Characteristics of the cohort study

We identified 339 patients who underwent elective splenectomy for immune cytopenias. The median age among these patients was 55 years (IQR: 36–68), and the M:F ratio was 0.98. Significantly lower median age (49 years) and higher percentage of female patients (60%) were observed in patients with AIN (Table 1). Primary ITP was indication for splenectomy in 235 patients, of whom 15% had an associated diagnosis of systemic autoimmune disease. DAT-positive AIHA was a reason for splenectomy in 68 patients, of whom 68% had primary AIHA, whereas 32% had AIHA, which was secondary to an autoimmune disease. Splenectomy was performed in 39 patients with idiopathic neutropenia, including 10 with autoimmune neutropenia positive for neutrophil antibody. Two patients had been diagnosed with FS and 10 patients had coinciding T-LGL. A summary of associated autoimmune conditions is tabulated in Table S3. A significant number of patients had bicytopenia (n = 88, 26%), including 26 presenting with ES. The overall median time from diagnosis to splenectomy was 11 months (IQR: 4–35; Table 1), which was 17 (IQR: 3–22) months in ES compared to 12 months (IQR: 4–43) in ITP, 9 months (IQR: 3–25) in AIHA, and 14 months (IQR: 5–23) in AIN. Vaccination status per Centers for Disease Control (CDC) recommendations for patients is tabulated in Table S4. Overall, the mean spleen weight was 477 g (SD ± 687; Table 1).

Table 1.

Baseline characteristics of patients undergoing splenectomy for immune cytopenia.

Variable	All n = 339	ITP n = 235	ES n = 26	AIHA n = 68	AIN n = 10	p-Value
Male gender (%)	168 (50%)	113 (48%)	14 (54%)	37 (54%)	4 (40%)	0.7052
Median age (Years)	55	51	58	63	49	<0.0001
Age ≤40 years (%)	31%	39%	15%	10%	30%	0.000028
Median BMI (kg/m2)	29	30	28	28	27	0.6330
Median time to splenectomy in months (Range)	11 (0–231)	12 (0–231)	17 (0–73)	9 (0–190)	14 (1–89)	0.2335
Laparoscopic splenectomy n (%)	290 (86%)	208 (89%)	17 (65%)	56 (82%)	9 (90%)	0.0007
Conversions (laparoscopic to open splenectomy), n	16	9	3	4	0	0.3024
Procedural conversion rate	5%	4%	12%	6%	0%	0.3024
Mean spleen length (± SD) (cm)	16 (6)	15 (6)	21 (5)	18 (5)	19 (6)	<0.0001
Mean spleen volume (±SD) (cm3)	1013 (1449)	772 (1410)	2146 (1691)	1383 (1288)	1101 (902)	<0.0001
Mean spleen weight (± SD) (g)	477 (687)	350 (631)	1165 (938)	637 (556)	600 (757)	<0.0001

Abbreviations: AIHA, autoimmune hemolytic anemia; AIN, autoimmune or idiopathic neutropenia; ALC, absolute lymphocyte count; ANC, absolute neutrophil count; BMI, body mass index; ES, Evans syndrome; Hb, hemoglobin; ITP, immune thrombocytopenic purpura; Plt, platelets; pre-op, preoperative; SD, standard deviation.

Pharmacologic therapy prior to splenectomy

Most of the patients (98%) received pharmacologic therapy prior to splenectomy and were either refractory to treatment (12%) or relapsed with cytopenia (86%) prior to the procedure. Herein, we list all prior lines. Sixty-eight (20%) patients received one line of therapy, 33% received two lines, and the remainder received three or more lines of therapy prior to splenectomy. The most common treatments included steroids in 282 patients (83%), IVIG in 189 (56%), and rituximab in 133 patients (39%). Other adjunctive treatments included TPORa (n = 86) and various forms of immunosuppressive therapy (IST) (n = 19) as listed in Table S5. Patients with FS (n = 2) were treated with steroids and rituximab, whereas T-LGL patients received IST with oral cyclophosphamide, cyclosporine, and methotrexate.

Surgical outcomes and short-term morbidity/mortality (≤30 days)

Of all patients, 306 underwent laparoscopic splenectomy, among whom n = 16 were converted to open laparotomic splenectomy. The indications for conversions included difficulty maintaining intraoperative hemostasis in n = 7 patients with ITP (median preoperative platelet count of 53 × 10⁹), massive splenomegaly (n = 6), and peritoneal adhesions (n = 3). The remaining 33 patients had an open laparotomic approach (Table 1). Average operative time was 2.3 h (Table S6A), with 4 (range: 1–50) days median hospital LOS. Patients who underwent open laparotomy had significantly higher median hospital LOS (8 vs. 2 days; p = 0.02). The overall surgical complication rate reached 12% (n = 40/339). Direct surgical complications are tabulated in Table S6B. Among them, three patients had non-major intraoperative bleeding from hilar staple line, which was mitigated intraoperatively. The most common 30-day postoperative complication was hemorrhage (n = 4/339). There were four cases of deep vein thrombosis (DVT) with two in the portal venous system (n = 3 ITP cases and n = 1 ES). Other short-term morbidities included respiratory failure (n = 3) and pneumonia (n = 2). Hospital readmission rate was 0.1%, without any significant differences considering 30-day readmission rate from complications between both approaches (p-value: 0.97) (Table S6A).

Histomorphologic diagnosis

Splenic histopathologic reports for all cases were reviewed. Tissue discordance was defined as spleen histology that establishes a secondary etiology for immune cytopenia.

The rate of discordant diagnosis on histology was 13% (n = 45). Such cases were excluded from the calculation of response for individual diseases. The corresponding outcomes are likely due to primary diagnosis, and thus, the response rate for discordant cases was not calculated. This discordant diagnosis (Table S7) included FS (n = 2), B-cell neoplasm (n = 27), and T-cell lymphoma (n = 6). In five patients, a previously unrecognized T-LGL diagnosis was found. We further characterized patients with a new or “discordant” histomorphologic diagnosis postsplenectomy. Compared to the general cohort, splenomegaly and bicytopenia/pancytopenia were more frequent in this subgroup [60% vs. 10%, p < 0.0001, and 51% vs. 9%, p < 0.0001, respectively].

Hematologic outcomes

The ORR (defined as the proportion of patients with a PR or CR to splenectomy) was 74%, of which 86% had CR and 14% had PR. ORR in ITP, AIHA, and AIN were 81%, 82%, and 80%, respectively, compared to 69% in ES (p = 0.52). CR rates were 67% for primary ITP (n = 200) and 51% for secondary ITP (n = 35; p = 0.28). Compared to baseline, the mean 1-year postsplenectomy platelet count was significantly higher (45 [SD ± 36] vs. 252 [SD ± 172] × 10⁹/L) (p < 0.0001; Fig. 2). CR rates were 58% for primary AIHA (n = 42) and 52% for secondary AIHA (n = 26; p = 0.36). Compared to baseline, the mean 1-year hemoglobin level was significantly higher (9 [SD ± 2] vs. 13 [SD ± 2] g/dL) (p < 0.0001). Mean 1-year absolute neutrophil count (4 [SD ± 4] × 10⁹/L) was similar compared to baseline count (p = 0.07). In T-LGL, the response rate with respect to cytopenia indication was 80% versus 100% in FS.

Fig. 2.

Outcomes with splenectomy—Hematological response and relapse rates: Hematologic response is represented by first column of pie charts indicating respective types of response—complete (dark blue), partial (light blue), or none (white). Relapse percentage per year is represented by the numeric value. Response based on count recovery is represented by bar histograms (black) showing the mean hematologic values, in weeks, for the various etiologies of cytopenia based on original indication for splenectomy. AIHA, autoimmune hemolytic anemia; AIN, autoimmune or idiopathic neutropenia; ES, Evans syndrome; HCT, hematocrit; ITP, immune thrombocytopenic purpura; ns, nonsignificant statistical difference based on p-value >0.05; T-LGL, T-cell large granular lymphocytic leukemia.

Relapse

At a median follow-up of 62 months, the overall rate of cytopenia relapse was 12%. No significant differences were observed in relapse rates: 12%, 12%, and 10% in ITP, AIHA, and AIN, respectively versus 8% in ES (p-value: 0.93). In subgroup analysis, primary ITP showed a relapse rate of 11% compared to 17% in secondary ITP (p = 0.3), whereas primary versus secondary AIHA had relapse rates of 10% versus 15%, respectively (p = 0.47). Furthermore, we observed a 20% relapse rate in T-LGL (Fig. 2). KM analysis showed no difference in long-term RFS among the groups of ITP, AIHA, AIN, ES, and T-LGL (p = 0.57; Fig. S1A). However, significant differences in RFS were observed between patients registering concordant versus discordant postsplenectomy diagnoses (p = 0.026; Fig. 3; see also Fig. S1B for details).

Fig. 3.

Long-term relapse-free survival (RFS) using Kaplan–Meier method. RFS comparison by subgroups: (A) primary (blue) versus secondary (yellow) immune thrombocytopenic purpura (ITP); (B) primary (blue) versus secondary (yellow) autoimmune hemolytic anemia (AIHA); (C) concordant (blue) versus discordant (yellow) postsplenectomy tissue diagnosis; (D) female (blue) versus male (yellow).

Long-term sequelae after splenectomy (>30 days)

Rate of severe infections.

During the follow-up period, there were a total of 76 cases of severe infections requiring hospital admission occurring in n = 57 (17%) of splenectomised patients. Of these admissions, n = 22 (30%) required intensive care for septic shock. The overall incidence of a severe infection was 4.5 per 100 person-years (95% CI 3.4–5.5). The incidence of severe infections did not differ significantly based on indication for splenectomy (ITP: 5, AIHA: 3, ES: 6, AIN: 4, p = 0.1).

Type of infections.

Respiratory infections and bacteremia were the most common category of severe infections (Table 2). Streptococcus pneumoniae represented the most frequent causative organism accounting for 11% of infections. Other gram-positive organisms included Staphylococcus spp. (8%), Enterococcus (4%), and Streptococcus (3%). Gram-negative organisms included Escherichia Coli (7%), Klebsiella (5%), Pseudomonas (3%), Legionella (1%), and Proteus (1%).

Table 2.

Categories of severe infections requiring admission.

Site	No of events (%)	Rate per 100 person-years	95% CI
Respiratory	29 (38%)	1.7	1.1–2.5
Bacteremia	26 (34%)	1.5	1.0–2.2
Skin and soft tissue	10 (13%)	0.5	0.2–1.1
Urinary tract infection	8 (11%)	0.4	0.2–0.9
Endocarditis	3 (4%)	0.1	0.04–0.5

Coli (7%), Klebsiella (5%), Pseudomonas (3%), Legionella (1%), and Proteus (1%).

Thrombosis.

There were n = 61 cases (17%) of DVT beyond the immediate 30 day follow-up period. These included n = 18 cases of pulmonary embolism, n = 8 splenic vein thrombosis, n = 2 portal vein thrombosis, and n = 1 cerebral thrombosis. The remainder (n = 32) solely involved the extremities. The incidence of DVT was 3.6 per 100 person-years (95% CI 2.8–4.6). No significant difference in incidence was observed based on indication (ITP: 4, AIHA: 3, ES: 4, AIN: 6, p = 0.7).

Mortality

When we assessed long-term mortality, only 3% of patients eventually died from cytopenias complications at data cut-off, 14% died from other comorbidities neither related to procedure or cytopenias, including six patients who expired from sepsis and two patients having discordant histomorphologic diagnosis of lymphoma dying from the progression of the underlying malignancy. The identified etiologies of septic deaths included clostridium difficile colitis (n = 1), liver abscess (n = 1), cholecystitis (n = 1), and neutropenic sepsis (n = 1). In 2/6 septic deaths, no organisms were isolated. These patients were adequately vaccinated per CDC recommendations.

There were 11 cases of severe infections resulting in death during the study period. The cumulative incidence of severe infections leading to death was 1 case per 100 person-years (ITP: 0.3, AIHA: 1.5, ES: 1.5, AIN: 1.4, p = 0.07).

Time to severe infection, thrombosis, and mortality after splenectomy

The risk of severe infections persisted over time with only 26% occurring within first year, 24% between years 2–4, and 50% at 5 years and beyond. Conversely, the highest risk of death from severe infections was seen within the first 2 years after splenectomy, with 55% of cases occurring within this period and the incidence declining subsequently over time. Of the reportedly septic deaths, the median interval from splenectomy to death was 40 months.

This risk of thrombosis was greatest within the first-year postsplenectomy with 51% of cases occurring during the first year, 28% between years 2–4, and 25% at 5 years and beyond.

Preoperative factors associated with response to splenectomy

In univariate analysis, factors associated with failure to respond included treatment failure with ≥3 lines of treatment prior to splenectomy (steroids, IVIG/rituximab, and TPORa) (OR 0.22, 95% CI 0.13–0.35), primary refractoriness or treatment failure to rituximab (OR 0.16, 95% CI 0.08–0.32), and IVIG (OR 0.17, 95% CI 0.08–0.40). Other favorable preoperative factors included isolated thrombocytopenia (OR 2.30, 95% CI 1.35–3.92), ITP (OR 2.98, 95% CI 1.79, 4.94), primary ITP (OR 2.70, 95% CI 1.65–4.48), and age ≤40 years (OR 2.45, 95% CI 1.30–4.64; Fig. 4). Based on affected cell lineages, the presence of anemia (OR 0.47, 95% CI 0.28, 0.78) and bicytopenia (OR 0.42, 95% CI 0.24–0.71) were associated with treatment failure. However, on multivariable regression, only treatment failure with multiple lines of pharmacologic therapy (steroids, IVIG, and rituximab) prior to splenectomy was most significantly associated with a failure to respond to the procedure, whereas response to either first- or second-line medical treatment was positively associated with splenectomy response (coefficient: 1.26, 95% CI 0.84–1.69). Female sex and age did not retain significance after accounting for bi/pancytopenia and the presence of a discordant histologic diagnosis. No significant differences in spleen weight or size between responders and non-responders were observed (Fig. 5).

Fig. 4.

Factors associated with splenectomy response by univariate analysis. Forest plot indicating significant factors associated with splenectomy response (blue) and splenectomy treatment failure (red).

Fig. 5.

Factors associated with splenectomy response by multivariable regression analysis.

Discussion

For decades, splenectomy has been practiced as treatment for refractory cytopenias due to extravascular destruction. However, this indication underwent major modifications due to the progress in pharmacologic therapies as reflected by the most recent guidelines [12]. Regardless, splenectomy remains a consideration for patients placing a high value on avoiding long-term medications [19]. Prospective or randomized trials would be difficult to perform due to the rarity of indications and the practicality of trial design. Therefore, clinical decision-making must rely on case series and retrospective analyses with consideration of surgical risk versus those of persistent cytopenias and the tested lack of rationally selected salvage therapies. Herein, we described durable response with splenectomy in referred patients with immune cytopenia who were refractory or relapsed on medical therapy with low periprocedural or long-term morbidity at the median follow-up of 62 months.

Treatment of immune cytopenia has evolved in the last couple of decades, partially due to advances in the management of the underlying diseases, such as infections, autoimmune processes, and hematological malignancies [24]. However, with the recent advances in minimally invasive surgical techniques and improved surgical outcomes, a reappraisal of the role of splenectomy becomes necessary. Due to the heterogeneity of these diseases, a careful assessment of the response following this invasive procedure is needed, considering the prior lines of therapies. Therefore, evaluating the patients for the presence of favorable preoperative factors is instrumental in guiding the decision-making process to proceed with splenectomy.

A major limitation of prior studies reporting outcomes with splenectomy is the inconsistency in describing preoperative clinical factors, especially considering prior medical therapies, such as steroids, IVIG, and rituximab. For instance, although some studies had reported a positive correlation between response in ITP and younger age population [25, 26], others have shown no differences [27, 28]. Smaller studies have analyzed responses to prior medical therapy such as steroids [28] and IVIG [29–31] with conflicting results. A recent analysis showed variability in real-life practice [32], and as such, there is currently no definitive consensus regarding the influence of medical treatment response on the decision to undergo splenectomy [33].

Moreover, patients with autoimmune cytopenia often display bicytopenia or pancytopenia, and data on the value of the number of affected lineages is lacking. Our study addresses some of the short-comings of prior reports by taking advantage of a relatively large, heterogeneous, and well-annotated cohort of patients with immune cytopenias that were either refractory or had relapsed on medical therapy. Although bicytopenia was present in some patients, more severely affected lineage constituted therapy indication and was used for outcome analysis.

The observed long-term remission rate of 74% was similar to prior studies performed within the last decade in the era of newer medical therapies [11, 34]. We also observed that the best results with splenectomy were achieved in females, younger patients, and individuals who responded to medical therapy on univariate analysis. Recognizing that bi/pancytopenia was most often associated with a discordant diagnosis, a multivariable regression analysis was performed. This analysis showed refractoriness or relapse on multiple lines of pharmacologic therapy prior to splenectomy retained significance in correlation with splenectomy treatment failure. This significance persisted even after adjusting for potential confounding factors such as age and sex. Surprisingly, splenomegaly and splenic weight had no association with response, as these are often cited as indications for splenectomy [35].

In addition to therapeutic benefit, splenectomy gives an advantage of tissue examination to evaluate for postsplenectomy diagnosis concordance, which might be utilized in complicated cases. Discovering additional or conflicting diagnoses should be expected due to the extensive range of diseases, the absence of established diagnostic criteria, and the inherent challenges associated with making accurate diagnoses. Indeed, the presence of discordant diagnoses in histology provides further evidence supporting the potential usefulness of splenectomy in specific cases. It is important to note that our results aim to align with real-life scenarios, and as a result, our retrospective analysis may not be as rigorous as a clinical trial, which, for obvious reasons, may never be conducted.

We found that the severity of cytopenias and reticulocytosis (in hemolytic anemia) did not differ significantly between groups based on concordance. The discordant subgroup of patients had more significant splenomegaly and bi/pancytopenia, which could be of diagnostic utility. More importantly, although a concordant splenic tissue diagnosis was not significantly associated with response, the discordant subgroup had significantly higher rates of relapse.

Another major consideration for splenectomy remains the risk of direct surgical complications and long-term morbidity, including overwhelming postsplenectomy infection and DVT. In the immediate postoperative period, we observed a complication rate of 12% in our series, including bleeding requiring transfusion (1%), infections (1%), and portal/DVT (1%). The major reasons for procedural conversion from laparoscopic to open splenectomy included intraoperative difficulties in maintaining hemostasis as the majority of cases had low platelet counts. The reported rates of intraoperative conversion from laparoscopic to open splenectomy may vary depending on the study population and surgeon experience [36], and there are conditional expert recommendations for intraoperative platelet transfusions in ITP patients undergoing splenectomy [36]. We found laparoscopic splenectomy to be significantly associated with shorter hospital LOS. Improvements in complication rates have been ascribed to better patient selection and implementation of interventions, such as appropriate immunization, perioperative antibiotics, thrombophylaxis, and early-catheter removal [37].

Our cohort received appropriate immunization per institution policy with a severe infection rate of 17% during the study period. The most frequent were respiratory infections similar to other registry studies [38, 39]. There were 47 deaths at the end of the study period with a cumulative incidence of death from infection being 3.2%. These mortality outcomes are more reflective of the comorbidities in our cohort rather than complications from splenectomy. With regards to DVT, intraabdominal cases were most prevalent within 90 days following splenectomy.

Numerous studies have consistently underscored the effectiveness of splenectomy as a therapeutic option for immune cytopenias. Our own study has reinforced this notion by demonstrating the attainment of a durable response. With a better description and understanding of the long-term sequalae after splenectomy, clinicians are now well versed in engaging patients with informed discussions concerning the prioritization of a durable response over extended medical therapy. This consideration gains added importance in the view of recent economic data, which highlights the substantial financial burden associated with refractory immune-mediated cytopenias. Notably, the implementation of early splenectomy has been identified as a cost-effective strategy, particularly among adults with chronic ITP [40].

Conclusion

Splenectomy is a rational therapeutic option for immune cytopenia and is associated with a low operative complication rate, mortality, and durable response at median follow-up of 5 years. Our findings highlight a contribution of splenectomy to the establishment of correct diagnosis. The discordance between preoperative clinical and histopathologic diagnoses contributes to the relapse rates. Controlling for other variables, failure with multiple lines of medical therapy is associated with splenectomy treatment failure. Evaluation of preoperative factors is crucial to identify patients likely to benefit from this invasive procedure.

Data availability statement

Requests for additional information not provided in the main text or the Supporting Information should be sent to the corresponding author.