http://orcid.org/0000-0002-7602-483X
Abstract
Introduction: There are insufficient data on the aetiologic factors underlying splenic infarction (SI). Therefore, there is no consensus regarding the appropriate diagnostic approach.
Methods: We conducted a retrospective analysis of all patients admitted with SI from January 2004 to December 2014. Medical records were screened for the clinical presentation, underlying causes, associated medical conditions and methods of patient evaluation.
Results: We found 89 subjects with 90 episodes of SI. Presentation of SI was characterized by abdominal, flank and chest pain (82.2%, 18.9%, 7.8%, respectively); leukocytosis (in 67% of tested subjects); elevated LDH (72%), CRP (97.5%) and D-Dimer (100%). The main underlying mechanisms were cardioembolic (54.4%), vascular (20%), haematologic disorders (15.6%) and multiple causes (21.1%). Atrial fibrillation and atherosclerosis were common in older patients (age > 70 years) while antiphospholipid syndrome occurred exclusively in younger individuals. SI was the presentation of previously unknown medical conditions in 38% of patients. Abdominal CT, ECG, echocardiography and blood cultures demonstrated the highest diagnostic yield.
Conclusions: Contributing factors are identified in the majority of SI patients. We recommend CT, ECG, echocardiography and blood cultures in all cases. Atrial fibrillation should be sought in older patients, while APLS and haematologic disorders should be suspected in younger ones.
There is no consensus regarding the diagnostic approach and management of splenic infarction.
Cardiovascular disease and atrial fibrillation are the main causes for SI in elderly subjects while hematological, infectious and other causes are more prevalent in younger ones.
Our data strongly suggests a high diagnostic yield for CT scan, ECG, blood culture and echocardiogram in every patient with SI.
KEY MESSAGES
Introduction
Infarction of the spleen is an uncommon clinical entity associated with variable symptoms and signs that are typically non-specific. They consist of abdominal or flank pain, left upper quadrant tenderness and fever that may be interpreted as an abdominal emergency. Laboratory abnormalities that support splenic infarction (SI) include leukocytosis, thrombocytosis, elevated levels of lactate dehydrogenase (LDH), and increased inflammatory markers [Citation1–3]. Imaging modalities, especially computerized tomography (CT), have been instrumental in the detection of SI. CT is considered as the best noninvasive study to confirm SI or to exclude alternative diagnoses [Citation1].
Previous studies have reported that the two main mechanisms that account for SI are thromboembolism and a rapidly growing spleen [Citation1,Citation4,Citation5], due to diverse aetiologies [Citation6]. In some cases, SI may represent the first manifestation of an underlying disease [Citation1,Citation4].
There are limited data regarding the causes and possible aetiologic factors that lead to SI. The information is mostly provided by case reports and retrospective case series. The relatively larger series, with more than 20 subjects, were published in the 1980s and 90s [Citation2,Citation3,Citation5,Citation7,Citation8]. These studies either investigate selected populations [Citation2,Citation5,Citation8,Citation9], or report limited patient data [Citation7,Citation10]. Furthermore, aetiologic factors such as the sickle cell trait or infectious diseases vary significantly among reports, thus reflecting heterogeneity in population characteristics. Only two studies refer to SI as the presenting finding of other diseases. However, conclusions are limited by the fact that they report variable data and include 26 [Citation4] and 32 [Citation6] patients. Moreover, there is no consensus for the management of SI, due to lack of knowledge on the yield of the available diagnostic tests.
In this communication we present a cohort of 89 subjects diagnosed with SI. We provide current data regarding their clinical characteristics at the time of presentation, aetiologic factors, yield of diagnostic modalities and accompanying diseases.
Methods
Patients
In this retrospective study we evaluated 89 adult patients (age ≥ 18 years) admitted to Meir Medical Center (Kfar Saba, Israel) from 1 January 2004 to 31 December 2014.
Medical records of all patients discharged with ICD9 diagnosis code 289.59 (other diseases of the spleen, which includes SI) were evaluated. All cases were confirmed by imaging studies. An infarct was considered as acute based on assessment of patient medical records, when associated with new symptoms or signs, laboratory results, or new imaging findings suggesting an acute event. CT criteria for SI included: typical peripheral, wedge shaped hypo-enhancing lesions; multiple heterogeneous lesions of patchy enhancement; massive hypo-dense lesions and entire hypo-enhancing spleen indicating global splenic infarction [Citation11].
Data collection
Data was retrieved from inpatient and outpatient computerized medical records. Patient medical records were reviewed for baseline characteristics, past medical history, clinical findings, medications, laboratory tests, imaging studies, in-hospital evaluation and treatment, as well as outpatient evaluation when relevant. Causes and possible aetiologic factors of SI were determined by the authors after thorough evaluation of medical records, discharge letters and outpatient follow-up visits up to 1 May 2015.
Statistical analysis
Descriptive variables are presented as mean or median ± standard deviation, or as rate of positivity according to variables. We compared variables between groups of patients using t test, Mann-Whitney test or Chi-square test according to scale measured variables. Difference between groups was considered significant when p < .05 (one sided).
The study was approved by the ethics committee of the Meir Medical Centre (application number 0305-14-MMC). The study was performed in accordance with GCP guidelines.
Results
Population characteristics
The study group included 89 subjects who experienced a total of 90 episodes of acute SI (). All patients were admitted through the Emergency Department (ED) after presenting with acute symptoms. The most frequent presenting symptom was pain in the abdomen (82.2%), flank (18.9%) or chest (7.8%). However, a distinct group of patients had no pain (8.9%) but rather reported having fever, vomiting or dyspnea. In terms of medical history, the majority of patients had risk factors for cardiovascular disease with hypertension (63.3%), dyslipidemia (46.7%) and diabetes mellitus (34.4) being the most prevalent. This was followed by cardiac abnormalities: atrial fibrillation (37.8%), ischemic heart disease (33.3%) and heart failure (25.6%), as well as stroke (21.1%), kidney disease (16.7%; including 2 hemodialysis patients) and peripheral vascular disease (5.6%). The least common were haematological and solid neoplasms and hypercoagulability states (1.1%-5.6%). In accordance with this range of medical conditions, 51% of the subjects had already been receiving antiaggregant or anticoagulation prior to admission.
Table 1. Clinical and laboratory characteristics of subjects at presentation.
Initial laboratory investigation revealed leukocytosis in 67% of the subjects and increased lactate dehydrogenase (LDH) in 72% (). Of note, C-reactive protein (CRP) was increased in 97.5% of the 40 patients in whom it was evaluated and D-dimer in all 14 tested subjects. Hypoalbuminemia was detected in 47% of subjects, while thrombocytosis was found in only 7.8%.
Outcome of investigation for SI: newly diagnosed underlying diseases and associated findings
At this point we evaluated SI as a presenting manifestation of underlying diseases (). Infectious diseases (endocarditis and non-endocarditis) were found in 12.2% of patients. It was also found that 11.1% had vascular diseases, unrecognized previously, including significant atherosclerosis, mural thrombi of the aorta, and dissection of the celiac artery. New cardiac abnormalities were observed in 14.5% of the patients. This group comprised severe left ventricular dysfunction, mural thrombi of the left ventricle and AF. Other findings were severe mitral valve insufficiency, severe mitral stenosis, mobile calcification on the mitral valve, and spontaneous echo contrast (“smoke”) in the left atrium as a sign of a hypercoagulable state. Interestingly, half of the ten patients in this study who had a myeloproliferative neoplasm with JAK2 mutation (JAK2-MPN) were newly diagnosed ( and ). Similarly, three of the five cases of antiphospholipid syndrome (APLS) and three of the solid malignancies were newly diagnosed ( and ). Thus, investigation for an SI aetiology yielded recognition of previously unknown underlying medical conditions in 38% of the study group.
Table 2. Newly diagnosed diseases and related findings in patients admitted with SI.
CT scan imaging, which was used to diagnose all SI cases, disclosed various additional findings (). The average time to the first CT study was 2.7 ± 4.1 days from admission. Only 29 patients had undergone imaging in the emergency room and, therefore, SI was diagnosed post-admission in the remainder two thirds of the subjects. Multiple splenic infarcts were observed in 32% of subjects while subtotal or total infarcts were detected in 9% (8 patients). Associated radiographic findings, observed in these imaging studies, are outlined in .
Table 3. Findings on abdominal CT in addition to splenic infarcts.
It should be mentioned that AF was confirmed in all patients with CT findings of intestinal ischemia (6/6) and in 85% of patients with renal infarcts (17/20). Interestingly, solitary renal infarcts were demonstrated in 8 cases while multiple lesions were noted in 12, of which 4 were bilateral. However, there was no association between the number of infarct lesions in the spleen and specific aetiologies.
Aetiologic factors for SI
Causative or precipitating factors for SI were derived from the past medical history of the patients as well as from findings that were newly disclosed by specific investigation upon admission (). A cardioembolic source for SI was identified in 54.4% of the patients, mostly due to atrial fibrillation (AF), with or without associated anatomic cardiac abnormalities. Infectious endocarditis was diagnosed in only 2.2% of the cases. Vascular disease constituted the second most prevalent aetiology, mainly attributed to atherosclerosis of the aorta, celiac or splenic arteries. JAK2-MPN was present in 11.1% of subjects. Infectious diseases other than endocarditis were diagnosed in 10%, albeit usually accompanied by another precipitating factor. These included bacteremia with no identifiable source, bacteremia associated with pyelonephritis or cholangitis, pyelonephritis and cholangitis without bacteremia and systemic cytomegalovirus infection.
Table 4. Causes and etiologic factors of splenic infarction.
Solid tumors (5.6%) included metastatic melanoma, advanced pancreatic cancer, and metastatic transitional cell carcinoma of the urinary bladder. Antiphospholipid syndrome (APLS) was found in 5.6% of the patients, and there was no convincing aetiology in another 5.6% (cryptogenic SI). Uncommon causes, designated as miscellaneous and comprising 3.3% of cases, included splenomegaly secondary to sarcoidosis, nephrotic syndrome due to idiopathic membranous glomerulonephritis, and Felty’s syndrome (associated with bacteremia). Of note, in 21.1% of the subjects, more than one precipitating factor was identified.
Interestingly, the types of aetiologies for SI varied between young and old patients, that is below and above the median age of 70 years (). Most remarkably, AF and atherosclerotic disease were prevalent in older individuals, while APLS was observed exclusively in younger ones. Infections other than endocarditis were significantly more common in the younger age group as well (18%).
Figure 1. Distribution of SI etiologies with patient age. AF, atrial fibrillation; athero, atherosclerosis; APLS, anti phospholipid syndrome; JAK2, MPN with JAK2. V617F mutation; infect, infections other than endocarditis; cancer, solid malignancies; vasc, vascular diseases other than atherosclerosis; cardiac, cardiac diseases without AF. *, p < .05.
Intriguingly, most cases of thrombocytosis were detected in JAK2-positive MPN patients (5/7 vs. 2/83, p < .0001). However, no other clinical or laboratory abnormality was distinct in these patients, including the absence of polycythemia.
Yield of diagnostic tests in revealing aetiologic factors for SI
outlines the utilization of various diagnostic modalities and their yield in uncovering the aetiology for SI. We found that abdominal CT and ECG were done in all cases. Together with echocardiography (TTE, n = 63; TEE, n = 14) and blood cultures (n = 52), these are the only modalities employed in the majority of subjects. Echocardiography yielded a variety of diagnoses including left ventricular dysfunction, aortic and mitral valvular disease, infective endocarditis, aortic atherosclerotic plaques and left atrial thrombus. The haematological and serological screening tests for APLS (n = 33), thrombophilia (n = 24), JAK2 mutation (n = 13) and collagen disease (n = 12) were less preferred. The tests that were mostly employed, CT scan, ECG, echocardiography and blood cultures, were also the most diagnostically informative in terms of providing an aetiology for SI, as detailed in . In terms of blood tests, JAK2 mutation and APLS evaluation led to positive results while screening for congenital thrombophilia and for collagen vascular diseases were negative in all cases.
Table 5. Yield of diagnostic tests for newly diagnosed etiologies.
Prognosis
In-hospital mortality was documented in 10 subjects. Of note, six of these patients had atrial fibrillation alone or in combination with other entities, two were hemodialysis patients with infection, one with a solid tumor and one with MPN and infection. All five patients with solid tumors died within 1 year following diagnosis of SI while all of the patients with APLS, cryptogenic or iatrogenic cases survived the first year.
Discussion
In this study, we report data that outline the presentation, causes, investigation modalities and associated findings in individuals with SI. In comparison with previous reports, this communication includes a considerably large cohort and provides up-to-date data. Similar to previous reports, clinical signs and symptoms may be indistinguishable from other abdominal emergencies and are non-specific [Citation4]. Elevated levels of CRP and D-Dimers are very common and are suggestive of SI. Thrombocytosis, which is often regarded as a typical finding in SI, was generally absent in our study, with the exception of subjects with JAK2-MPN.
Contributing factors and associated illnesses can be identified in almost all cases of SI, and have an impact on management. Our work shows that one in five patients has more than one plausible cause for SI, a finding not previously reported. Most causes for SI can be categorized into one of seven major aetiologic groups: (1) cardioembolic disease, mostly AF with or without additional cardiac abnormalities; (2) large artery disease due mostly to advanced atherosclerosis; (3) haematological malignancies, particularly JAK2-MPN; (4) acute and subacute infectious diseases commonly associated with bacteremia; (5) APLS detected exclusively in patients younger than 50; (6) advanced solid malignancies, identified by CT, usually in older patients; (7) miscellaneous causes.
Previously unsuspected medical entities can be identified in a significant proportion of patients presenting with SI. A recent study demonstrated new diagnoses in 25% of patients with SI, particularly mitral valve disorders and APLS [Citation6]. However, another report claimed that this is true in the majority of cases [Citation4]. A hypercoagulable state was identified in 45% of the patients [Citation1], and the rate of cardiac abnormalities and aortic atherosclerosis was deemed high enough to justify TEE in all patients [Citation7]. Our findings indicate newly diagnosed significant medical conditions in 38% of the cases, consistent with previous studies.
In addition, we also provide novel data regarding the yield of various diagnostic modalities which may be used to identify such entities. This evaluation can support the planning of a practical diagnostic scheme. We found that, although SI can be diagnosed by several imaging approaches, a formal abdominal CT is imperative. CT scan not only excludes alternative diagnoses, but also identifies significant associated findings and underlying causes.
Similarly, ECG and echocardiography studies emerge as powerful diagnostic tools due to the fact that embolism of cardiac origin is the most common cause of SI. In older patients, AF is a highly prevalent aetiology and therefore should be carefully sought, even when other causative factors are evident. Its presence is highly suggested in the presence of concomitant intestinal ischemia or renal infarcts. In patients with AF, an episode of SI probably implies a high risk for future stroke, which can be effectively prevented by anticoagulation.
Splenic infarction is a known embolic complication of endocarditis, but the prevalence of other associated infections has not been previously appreciated. Apart from septic emboli, severe infection may promote SI by activation of the coagulation system [Citation12] and rapid splenic enlargement. Blood cultures are often taken, given the high rate of endocarditis described in some previous studies [Citation1,Citation2,Citation4,Citation5,Citation8]. We found a low incidence of endocarditis, in line with recent studies [Citation6,Citation10], but also a considerable number of other infections some of which associated with bacteremias. In patients with SI, focal and systemic signs of infection such as fever or leukocytosis may be absent, masked or attributed to SI per se. We suggest that a meticulous investigation for active infection, including routine blood cultures, is mandatory in all patients with SI.
Thrombotic events are frequent in patients with MPN and account for mortality and morbidity in most of these individuals. A major risk factor for thrombotic events in patients with MPN is the presence of the V617F mutation in the JAK2 gene [Citation13], which is identified in the majority of patients and can establish diagnosis [Citation14]. SI is a known complication of MPN, and was noted as the presenting symptom in previous case reports and case series [Citation1,Citation4,Citation6]. In the present study, a significant number of patients had MPN, mostly JAK2-positive, many of whom were newly diagnosed following the evaluation of SI. Notably, these subjects did not exhibit typical features of MPN, including erythorocytosis and splenomegaly. Our findings suggest that screening for JAK2 mutation should be considered in cases with otherwise unexplained SI, specifically in the presence of thrombocytosis. Another haematological condition which should be considered is APLS, specifically in patients younger than 50 years of age, in whom it was found in 39% of the subjects.
Interestingly, the subjects in our study were somewhat older than those in previous studies, a fact that may explain the relatively large proportion of cardioembolic sources (mostly AF) and atherosclerotic disease. It is, therefore, not surprising that a considerable number of patients had experienced prior thromboembolic episodes due to the presence of significant risk factors. These include coronary and cerebrovascular diseases (33.3% and 21.1%, respectively), peripheral vascular disease (5.6%), as well as previous venous thrombosis (4%) ().
In previous series, the incidence of endocarditis was 6–26%, and MPN was noted in 3–39% of patients [Citation1–9], the latter more commonly reported in studies from the 1990s [Citation3,Citation7,Citation8]. These rates are higher than in the present communication, probably reflecting differences in patient selection and in diagnostic modalities, with splenectomy [Citation1,Citation4,Citation6] and autopsy becoming less common over the years. The advent of technologically advanced diagnostic tools may account for the relatively low rate of cryptogenic SI in our study.
Our work underscores the importance of investigation for an underlying cause and associated radiographic findings in subjects with SI. We found that prognosis is associated with the aetiology of SI. The findings that were most associated with in-hospital mortality were atrial fibrillation, alone or in combination with other diseases, while solid malignancy entailed a high rate of 1-year mortality. Conversely, APLS, cryptogenic or iatrogenic cases exhibited a favorable outcome. It is conceivable that the high proportion of kidney disease (16.7%) may also compromise prognostic outcomes, as we have recently reported in patients discharged from internal medicine wards [Citation15]. The two hemodialysis patients in this study died during their stay in the ward, and the rate of mortality among kidney disease subjects appeared to be high, although not reaching statistical significance when compared to other subjects. Finally, our data is not sufficient to support the efficacy of anti-thrombotic treatment. Although 4 of the 10 who died during their stay in the hospital did not receive any anticoagulant or anti-platelet therapy, the numbers were too small to draw a conclusion.
Although the retrospective design of this study presents methodological limitations, such as the lack of a standardized protocol, a prospective study does not seem feasible due to the exceedingly low prevalence of SI. We included all cases admitted with SI over a considerable period of time (11 years) thus allowing a comprehensive collection and analysis of SI data. However, it is possible that some patients were not included in our study due to under-reporting
Conclusions
Our data demonstrate that factors contributing to the development of SI can be identified in most cases, and that the presence of multiple aetiologies is likely. Aetiologic factors and associated diseases vary with age, and should be taken into account when planning patient management. On the basis of the data presented herein regarding the yield of various diagnostic modalities, we propose several recommendations. It is advisable to order CT, ECG and echocardiography studies as well as blood cultures in all cases. The indication for investigation of AF increases with age as does the likelihood of atheroscelrosis and malignancy. Conversely, testing for APLS is relevant in young patients. Routine evaluation for MPN, including testing for the JAK V617F mutation should be considered, especially in the presence of thrombocytosis. Given the expected morbidity and mortality rates following SI, in patients with cardiovascular risk factors, regular long-term follow-up is recommended.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- Antopolsky M, Hiller N, Salameh S, et al. Splenic infarction: 10 years of experience. Am J Emerg Med. 2009;27:262–265.
- Jaroch MT, Broughan TA, Hermann RE. The natural history of splenic infarction. Surgery. 1986;100:743–750.
- Nores M, Phillips EH, Morgenstern L, Hiatt JR. The clinical spectrum of splenic infarction. Am Surg. 1998;64:182–188.
- Lawrence YR, Pokroy R, Berlowitz D, et al. Splenic infarction: an update on William Osler’s observations. Isr Med Assoc J. 2010;12:362–365.
- O’Keefe JH Jr, Holmes DR Jr, Schaff HV, et al. Thromboembolic splenic infarction. Mayo Clin Proc. 1986;61:967–972.
- Schattner A, Adi M, Kitroser E, et al. Acute splenic infarction in an academic general hospital over 10 years. Medicine (Baltimore). 2015;94:e1363.
- Frippiat F, Donckier J, Vandenbossche P, et al. Splenic infarction: report of three cases of atherosclerotic embolization origination from the aorta and retrospective study of 64 cases. Acta Clin Belg. 1996;51:395–402.
- Goerg C, Schwerk WB. Splenic infarction: sonographic patterns, diagnosis, follow-up, and complications. Radiology. 1990;174:803–807.
- Cox M, Li Z, Desai V, et al. Acute nontraumatic infarctions in a tertiary-care center: causes and predisposing factors in 123 patients. Emerg Radiol. 2016;23:155–160.
- Romano S, Scaglione M, Gatta G, et al. Association of splenic and renal infarctions in acute abdominal emergencies. Eur J Radiol. 2004;50:48–58.
- Balcar I, Seltzer SE, Davis S, et al. CT patterns of splenic infarction: a clinical and experimental study. Radiology. 1984;151:723–729.
- Levi M, van der Poll T, Schultz M. Infection and inflammation as risk factors for thrombosis and atherosclerosis. Semin Thromb Hemost. 2012;38:506–514.
- Falanga A, Marchetti M. Thrombosis in myeloproliferative neoplasms. Semin Thromb Hemost. 2014;40:348–358.
- Spivak JL. Myeloproliferative neoplasms. N Engl J Med. 2017;376:2168–2181.
- Topaz G, Gharra W, Eisen A, et al. Impaired renal function is associated with adverse outcomes in patients with chest pain discharged from internal medicine wards. Eur J Int Med. 2018;53:57–61.