Increased risk of venous thromboembolism in splenectomized patients with sickle cell disease

Sickle cell disease (SCD) is the main cause of functional asplenia, due to splenic atrophy that is generally complete by five years of age in patients with severe genotypes (i.e., homozygous SS and Sβ0-thalassaemia).1 In addition, unpredictable complications, such as acute splenic sequestration (ASS) and hypersplenism, may require surgical splenectomy in up to 30% of patients.2 Some authors have reported an increase in acute chest syndrome (ACS) and vaso-occlusive crisis (VOC) in the few years following the procedure in children,3 but data on the long-term consequences of splenectomy in SCD adults are scarce. In conditions other than SCD, increases in venous thromboembolic events (VTEs) and pulmonary hypertension (PHT) have been reported following splenectomy.4, 5 The aim of this study was to compare the characteristics of adult SCD patients with a history of surgical splenectomy to those of non-splenectomized SCD patients. We conducted a retrospective, observational study in the adult SCD Referral Centre of the European Georges Pompidou Hospital in Paris. We included all SS or Sβ0 SCD patients who received regular care between October 2015 and May 2021 and had a history of splenectomy (yes/no) clearly specified. The patient characteristics were collected from their electronic medical records using a standardized form. Demographic data, ongoing and past treatments, baseline biological parameters and prevalence of various acute and chronic complications of SCD (including VOC, ACS, PHT, VTEs) were compared between the splenectomized and non-splenectomized patients. VTEs included deep vein thrombosis (DVT) and pulmonary embolism (PE) while central venous catheter (CVC)-related thromboses were excluded. Moreover, because of the distinct pathophysiology, PEs occurring concomitantly with an ACS were excluded from the evaluation of the overall VTE risk. Characteristics associated with splenectomy were investigated by multivariate logistic regression. This non-interventional retrospective study using data collected for routine clinical practice was conducted in compliance with the Helsinki Declaration and the MR004 French legislation. Full details of the methods are available in the Supporting Information (Data S1). Of the 360 SCD patients included, 42 (11.7%) had undergone splenectomy at a median age of 9.5 years [7.2–15.5]. Surgery was performed for ASSs in 14/23 (60.9%) patients, hypersplenism in 7/23 (30.4%), and SCD-unrelated causes in two patients (pancreatic tumour surgery and liver transplantation). The median duration of postsplenectomy follow-up was 12.9 years [9.5–17.4]. Detailed comparisons between the splenectomized and non-splenectomized patients are presented in Table 1. In comparison to non-splenectomized patients, the splenectomized were younger (median age 23.2 years vs 29.2 years; p = 0.006) and had more often the Sβ0 genotype (21.4% vs 2.2%; p < 0.001). After adjustment for age and SS/Sβ0 genotype, the α-thalassaemic trait was significantly associated with splenectomy [odds ratio (OR) 3.8; 95% confidence interval (CI) (1.3–11.3); p = 0.015] (Figure S1). Hydroxyurea (hydroxycarbamide) exposure was more frequent (90.2% vs 74.5%; p = 0.042) and was initiated earlier in life (15.8 years vs 21.2 years; p = 0.001) in the splenectomized patients. There was no difference in the number of VOCs in the year preceding the last visit or in the number of ACSs. It was not possible to accurately compare the prevalence of all bacterial infections between the groups, as only osteomyelitis and CVC-related infections were systematically collected in our database (no difference for these infections). Finally, the prevalence of chronic complications in the groups was similar in univariate and multivariate analysis; notably, there was no difference in terms of PHT prevalence. All patients n = 360 Splenectomized n = 42 Non splenectomized n = 318 Fourteen percent of our cohort (n = 44) had experienced at least one episode of VTE until the last visit, with a median age at first thrombosis of 27.5 years (21.8–31.4). The median time to first thrombosis after splenectomy was 12.3 years (2.2–21.6). Prevalence of VTEs was higher amongst splenectomized patients (25% vs 12.5%; p = 0.059) (Table 2). A history of PE unrelated to an ACS was significantly more frequent in the splenectomized patients (12.8% vs 3.7%, p = 0.029). After adjustment for age and SS/Sβ0 genotype (Figure S1), the prevalence of VTEs remained significantly higher in the splenectomized patients, who had a 3.8-fold increased risk of developing non-ACS-related PE or DVT compared with the non-splenectomized patients (95% CI 1.2–11.2; p = 0.015). The risk of non-ACS-related PEs was also independently associated with splenectomy with an OR of 4.55 (95% CI 1.1–16.6; p = 0.025). The prevalence of clinical ischaemic strokes was similar in both groups. No other arterial thrombosis was observed. This study showed that SCD patients of Sβ0 genotype or with an associated α-thalassaemia trait were more likely to be splenectomized than SS patients. Few reports have compared the SS and Sβ0 phenotypes, but Sβ0 patients might be more prone to persistent splenomegaly6, 7 and ASS8; characteristics that may result in a higher rate of splenectomy, as observed as early as 19799 and in a recent Italian cohort (n = 534, 187 SS and 176 Sβ0) showing that 53% of Sβ0 patients had a history of splenectomy versus 9.6% of SS patients.2 Coexistent α-thalassaemia may also cause persistent splenomegaly.10 Serjeant et al. suggested that microcytosis allows for easier extraction of sickle cells from small vessels, thereby decreasing the obstruction of the splenic microcirculation that causes early infarcts.9 Our results also suggest that splenectomy may predispose SCD patients to VTEs. It is consistent with rising evidence of a hypercoagulable state following surgical splenectomy in various conditions, such as β-thalassaemia intermedia.11 The underlying pathophysiological process is not fully understood. However, causality cannot be established by this retrospective study, especially because VTEs were very far out from the splenectomy and classical VTE risk factors were not included in adjusted analyses. In SCD, VTEs occur more frequently than in the general population,12, 13 but the association with splenectomy had never been reported until a recent US study including 233 adults (all genotypes, 15% splenectomized), where VTEs were significantly associated with splenectomy in a multivariate analysis (OR 3.41; 95% CI 1.43–8.16).14 The singularity of this association is that splenectomized patients are herein compared with patients with functional hyposplenism or asplenia, suggesting that residual splenic tissue in non-splenectomized SCD patients retains minimal splenic function that reduces thrombotic risk. Thus, although the spleen is often overlooked in SCD because its function is supposed to be lost early in the course of the disease, preserving splenic tissue might be more important than previously thought. In conclusion, this study showed that SCD adult patients with a history of splenectomy had more often the Sβ0 genotype or an associated α-thalassaemia trait and may have an increased risk of VTE. Juliette Tennenbaum performed the research. Jean-Benoît Arlet, Geoffroy Volle, Brigitte Ranque and Jacques Pouchot designed the research study. Juliette Tennenbaum, Brigitte Ranque, Jean-Benoît Arlet, Laure Joseph, Djamal Khimoud and Geoffroy Volle contributed to the acquisition of data. Juliette Tennenbaum, Brigitte Ranque, Jean-Benoît Arlet and Geoffroy Volle performed the analysis and interpretation of data. Juliette Tennenbaum, Jean-Benoît Arlet and Geoffroy Volle wrote the paper. All authors reviewed and approved the final version of the paper. The authors have no competing interests. No conflicts of interest declared. We have obtained verbal consent from the patient or patient's parent/guardian. n/a. n/a. Data S1: Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.