Splenectomy reappraised: Bridging the gap in immune cytopenia treatment paradigms

The spleen is often the primary site of autoantibody production and autoantibody-coated clearance of hematologic cells in several immune-mediated hematologic rare diseases collectively referred to as autoimmune cytopenias. The affected hematologic cell lines in autoimmune cytopenia disorders most commonly include platelets, red cells, and neutrophils, with the respective conditions referred to as immune thrombocytopenic purpura (ITP), warm autoimmune hemolytic anemia (wAIHA), and autoimmune neutropenia (AIN). When two of these conditions co-occur, the immune-mediated hematologic process is referred to as Evans Syndrome (ES), named after Dr. Robert Evans who in 1951 first described the association of the former two diagnoses [1]. The eponym was then extended to include the possibility of AIN in the 1990s, a diagnosis with pathophysiology suggestive of the causal involvement of antineutrophil antibodies in 1975 [2].

The 2023 treatment considerations for AIN, when treatment is needed, include granulocyte-colony stimulating factor and/or immunosuppression, whereas the recommended first- and second-line treatment paradigms for ITP, wAIHA, and ES include immunosuppression anchored on corticosteroids and the monoclonal anti-CD20 rituximab, with additional consideration in the case of ITP for thrombopoietic agents referred to as thrombopoietin receptor agonists [3-6]. Third-line treatment includes additional immunosuppressant considerations, such as azathioprine and cyclosporine, and most recently, splenectomy, as a now-downgraded (i.e., third-line) therapeutic modality [4-6].

For ITP and wAIHA, and by extension their overlap known as ES, splenectomy was the preferred second-line treatment for over 50 years through the 2010s, being recommended by leading experts as recently as 2010 and 2011 for AIHA and ITP, respectively [7, 8]. As experience with the off-label use of rituximab grew, splenectomy increasingly drifted toward third-line treatment status due to concern of the tripartite risk of infection, thrombosis, and perioperative mortality [4, 6]. The point estimates for these risks, however, date to the 1990s and 2000s, and so predate advances in preoperative vaccination, thromboprophylaxis, and the laparoscopic approach to splenectomy. Nevertheless, and without the quantification of total splenectomy risk through the 2010s, splenectomy-associated risks have been qualitatively noted as likely too great. In our experience at the bedside, this has led to questions of appropriate patient counseling and informed consent. This is especially the case for individuals whose values and preferences align with pursuing a potentially one-time, definitive treatment, but who are dissuaded from pursuing splenectomy because the “risks are too great.” The issue does not exist on a population level, where we know that splenectomy continues to be the therapeutic modality with the highest probability of documented, long-term effectiveness without need for continuous therapy. Rather, the issue exists at the crucial individual level, where clinicians worry that we often cannot adequately adjust the pretest probability for the risk of disease recurrence after splenectomy for individual patients. Due to both ethical and practical considerations, we will never have prospective, randomized trial data to make these decisions. And yet these decisions need to be made, ideally with consent that is as informed as data limitations ultimately allow.

It is in this context of the absence of planned prospective data collection in this arena that, in this issue of the Journal of Internal Medicine, Ogbue et al. provide crucial insights of their institutional experience with splenectomy in the treatment of patients with a variety of autoimmune cytopenias at the Cleveland Clinic, an institution in the United States with a deep commitment to clinical excellence [9]. The authors meticulously applied thoughtful exclusion criteria through a retrospective evaluation of 1824 consecutive splenectomy cases across the years 2002–2020 to properly isolate 339 patients who underwent splenectomy for one of the following autoimmune cytopenia diagnoses with a subsequent median follow-up of 5 years: ITP (n = 235), AIHA (n = 68), ES (n = 26), and AIN (n = 10). Of particular importance, they report (1) appropriately adjudicated (i.e., disease-definition-specific) response and relapse rates at each yearly interval of follow-up for each autoimmune cytopenia; (2) complication rates accrued in per person-years for infection, thrombosis, and mortality; (3) patient vaccination status; (4) postoperative complications; and (5) pathology-proven discordant diagnoses appropriately removed from effectiveness metrics (Fig. 1). The authors additionally utilized their sample size to perform logistic regression across an appropriate maximum amount of risk factors to identify covariates that may be associated with favorable and unfavorable response to splenectomy.

Specifically, the authors report a robust 74% overall hematologic response rate with a 12% relapse rate; rates that have historically been expected with splenectomy across cohorts of individuals with the considered autoimmune cytopenias. Additionally, surgical complication rates and long-term morbidity were low, and this is despite a patient cohort that appears to be notably under-vaccinated for splenectomy, an issue that has been noted previously in modern medicine [10]. The authors attempted to identify key clinical covariates that are associated with both favorable and unfavorable responses to splenectomy. Although several were identified on univariable analyses, multivariable regression identified preoperative refractoriness or relapse after multiple lines of therapy as associated with splenectomy treatment failure. Arguably the most striking aspect of their study lies in their reporting of the value added of splenectomy as a diagnostic tool, in addition to its therapeutic potency. In 13% of all patients who underwent splenectomy, tissue diagnosis yielded a diagnosis that was discordant from the working diagnosis prior to splenectomy. The correct diagnoses largely included a variety of malignancies that have entirely different treatment paradigms, with most representing different types of lymphoma, which are notorious for difficulty of diagnosis when confined to the spleen and are known to drive immune-mediated hematologic findings that can resemble some autoimmune cytopenias.

Ogbue et al. are to be commended for conducting a rigorous retrospective analysis across heterogeneous conditions in a way that has added value to the field for patients and physicians. These findings buttress an increasing base of information within an important knowledge gap regarding splenectomy risks, the presence of which over the years has convinced some of our patients to stay away from splenectomy without properly informed consent. The authors also showcase the additional value of splenectomy for patients in making the correct diagnosis, even in the context of an institution known globally for clinical excellence. As important, properly reported rates in per-person-years of this well-annotated cohort provide a bevy of information that will be useful to researchers in health decision science and separately for researchers who may opt to pursue the creation of a patient decision aid in this space. Most important, it provides a well-done study beyond the ITP arena, with an additional focus on wAIHA and ES, that will incrementally help improve patient–physician shared decision-making in these treatment arenas.

As with all retrospective analyses, however, there are notable limitations. The work, in its nature, cannot be predictive. The sample sizes, although superb for rare disease work, are also relatively small, and the diseases under study are notably heterogeneous. Similar work at other institutions, particularly for diseases other than primary ITP, where there is substantially more data than in wAIHA and ES, would be helpful to confirm the effectiveness data. Most important, our patients are making a lifelong, unalterable decision when it comes to pursuing splenectomy, and so data well beyond a median follow-up of 5 years is needed. Thankfully, we now have data out to 20 years in ITP in at least one report, but similar data is lacking for wAIHA and ES. In addition, no generalizable conclusions should be drawn on the effectiveness of splenectomy for individuals with AIN given the very small sample size. Finally, the identified covariates associated with response, or lack thereof, to splenectomy are important confirmations and are not novel for practicing clinicians: The issue regarding uncertainty for patient selection for the splenectomy modality will continue.

Nevertheless, despite the limitations and even in the context of wonderful non-surgical therapeutic options, Ogbue et al. add sufficient data to remind us that splenectomy as a third-line option is inappropriate for all patients with autoimmune cytopenias. They additionally provide new data as to the risk of postsplenectomy infection while noting that the period under study does not exclusively span the modern vaccination era and that many patients may have been notably under-vaccinated. These are opportunities for the field and for our patients. As we continue to re-appraise the value of splenectomy, we need to ensure that all patients opting for the possibility of definitive, one-time treatment receive all recommended vaccinations and undergo appropriate discussion and risk stratification that considers their individual-specific covariates when it comes to thromboprophylaxis and the surgical approach. This would allow a true alignment of clinical care with patient values and preferences, in the context of known uncertainty and expected comparative data limitations when it comes to splenectomy consideration for all individuals living with autoimmune cytopenias.

The authors have no conflicts of interest to declare.