Ebright, John R. MD*; Alam, Ejaz MD*; Ahmed, Huma MD*; Tucker, Rennard MD†; Abrams, Judith PhD‡; Levine, Donald MD*
The spleen is located in the left upper and posterior region of the abdominal cavity situated between the fundus of the stomach and the diaphragm. It has an average weight of 150 g in adults and is composed of vascular and lymphoid tissue.1 Although not necessary for survival, it nevertheless has 2 major functions: the filtration from the bloodstream of all foreign matter as well as obsolescent and damaged blood cells and participation in the immune response to blood-borne antigens.1
In the setting of infective endocarditis (IE), the spleen may become enlarged because of a hyperplastic response to persisting bacteremia.1 As a result, splenomegaly may be evident on physical examination in 25% to 60% of all cases, especially in those patients in whom the infection has been present for at least several weeks.2
Two additional abnormalities in the setting of IE may be less common but pose greater risk to the patient and cause uncertainty in management for the clinician. These are splenic infarction and splenic abscess. Of the two, infarcts are the more common, occurring in approximately 40% of left-sided IE.3 They result from embolic occlusion of 1 or more branches of the splenic artery. The spleen is vulnerable to infarction because of the lack of collateral circulation between the branches of the splenic artery and also because sluggish blood flow within its red pulp may become increasingly slow in the setting of splenic hypertrophy.4
Splenic abscesses may occur in at least 3% to 5% of patients with IE but are associated with a higher risk of mortality than infarction.5,6 Abscesses may result directly from septic emboli originating from cardiac vegetations and lodging within the splenic vessels or from bacteremic spread of organisms to areas of bland infarction. Mortality approaches 40% even with treatment and may be preceded by persisting bacteremia, sepsis, secondary infections, and splenic rupture or hemorrhage.5,7 In addition, other complications and morbidity including persisting fever, left-upper quadrant abdominal or pleuritic pain, or recurring bacteremia also may result from abscesses within the spleen. Bacteremia from splenic abscess has the potential for bacterial seeding of prosthetic heart valves placed in patients with valve destruction due to IE and is a common concern for both cardiovascular surgeons and infectious disease specialists.3,8
DETECTION AND MANAGEMENT OF SPLENIC INFARCTS AND ABSCESSES IN PATIENTS WITH IE AT THE DETROIT MEDICAL CENTER
The medical records of 166 patients with IE and who were cared for at the Detroit Medical Center from 1996 through 2000 were retrospectively reviewed. One hundred eighteen had definite, and 48 had possible IE according to the Duke criteria.9
Either an abdominal computed tomography (CT) scan or ultrasound (US) (or both) was performed on 59% (98/166) of patients during their hospitalization. Of the 98 patients who had imaging studies performed, 27 (nearly 27%) were found to have splenic defects which were compatible with infarcts or abscesses (Fig. 1).
When the 27 patients with abnormal splenic imaging studies were compared with 71 patients with negative studies, no statistically significant differences were found in demographics (age, sex, and history of injection drug use) or in the specific microorganisms responsible for IE. However, there was a significant difference between the 2 groups when compared in terms of certain clinical or laboratory findings. The group with imaging studies compatible with infarcts or abscesses were more likely to have at least one of the following: fever or bacteremia persisting for longer than 7 days after hospital admission or abdominal pain or tenderness (P = 0.04 by χ2 test).
Fourteen patients with splenic defects compatible with infarct or abscess underwent splenectomy. Nine were found to have splenic abscesses based on gross or histological findings; 4 had splenic infarction; and 1 splenic hemorrhage. The presence of any clinical or laboratory findings before splenectomy was not useful in differentiating abscess from infarction.
Finally, the appearance of the abnormal CT scans in terms of shape, size, location, or numbers of defects did not prove useful in differentiating abscess for infarction.
PROBLEMS IN DECISION-MAKING FOR PHYSICIANS CARING FOR PATIENTS WITH IE AND SPLENIC INFARCTS OR ABSCESSES
Three basic questions continue to challenge physicians managing patients with IE and complications involving the spleen:
1. Under what circumstances should abdominal CT and/or US be performed to investigate the possibility of splenic infarct or abscess?
2. How are the 2 lesions differentiated clinically, radiographically, and pathologically?
3. How are abscesses in the spleen best managed? Can splenectomy be avoided? Are antibiotics alone ever sufficient?
Clinical Features of Splenic Infarcts and Abscesses
The clinical features of splenic infarct and abscess overlap and consist of abdominal (especially left-upper quadrant), back or pleuritic pain, tenderness, and fever. The frequency of these findings is difficult to establish from the literature due, in large part, to inconsistent evaluations of patients with IE for infarct or abscess and also because of the problem in distinguishing infarct from abscess. Nevertheless, in 1 large retrospective study of patients with splenic infarcts due to IE as well as other causes, approximately 30% had left-upper quadrant abdominal pain as their principle symptom.10 In another retrospective review, a case was presented in which a patient with splenic infarction presented with fever, left pleuritic chest pain, and flank tenderness. However, the authors noted that in an autopsy series of 96 consecutive cases of splenic infarction, only 10% were suspected clinically.11 Two additional retrospective studies of patients with splenic infarcts resulting from multiple underlying illnesses in addition to IE noted abdominal pain and/or fever in 69%, and 61% of 59 patients and 23 patients, respectively.12,13 Interpretation of these data is complicated in the report by Nores et al,12 however, because 10 of the 59 patients were discovered to have developed splenic abscesses at the time of splenectomy.
A prospective study was performed by Haft et al14 in which 25 consecutive patients with IE underwent abdominal CT studies. Six patients (25%) had abnormal spleens with hypodense lesions visualized and thought to represent infarcts. No definitive diagnostic studies (splenectomy, percutaneous aspiration) or long-term follow-up was reported in these patients, making the diagnosis of infarction presumptive, rather than confirmed. Nevertheless, 2 (33%) of the 6 patients were symptomatic with abdominal pain.
In their retrospective review, Ting et al8 reported an incidence of splenic infarcts or abscesses in 20 (19%) of 108 patients undergoing valvular surgery for left-sided endocarditis. This incidence almost certainly was low because CT was performed in only 37% of the total group of 108 patients. Of the patients with abnormal CT, 8 had symptoms. A little more than half (11/20) were asymptomatic and assumed (perhaps reasonably, but without percutaneous aspiration or splenectomy to attempt confirmation) to have infarcts underscoring a common but poorly documented assertion that infarcts are more often asymptomatic as compared with splenic abscesses.
Twenty-seven patients with IE and splenic abscesses confirmed after splenectomy or at autopsy were reported by Robinson et al.15 All patients were reported to have fever due to either endocarditis or abscess. Six had pleuritic chest pain. None had abdominal pain. In contrast, 37 patients (54%) with splenic abscesses and IE reported by Johnson et al5 had abdominal pain as their most common symptom. Five patients (13%) had persistent fever.
A recent multicenter retrospective study performed in France examined 153 of 225 patients with IE using abdominal US, CT, or both and discovered lesions compatible with infarcts or abscesses in 35% of those imaged.16 Differentiation of infarct from abscess was attempted using clinical and radiological criteria. In most cases, the diagnosis was not confirmed because of the infrequency of splenectomy or autopsy. Using clinical assessment, radiological, or sonographic results for presumptive diagnosis in most of the patients, the authors reported fever in approximately 55% and 90% of patients with abscesses and infarcts, respectively. Abdominal pain was reported in approximately 35% of both groups of patients.
Finally, a recent authoritative scientific statement focusing on IE and its complications recommended that persistent or recurrent bacteremia, persistent fever, or other signs of sepsis are suggestive of splenic abscess and should prompt the clinician to obtain an abdominal imaging study.3
In summary, the presence of new-onset abdominal pain or tenderness, especially in the left-upper quadrant, left pleuritic or flank pain, persisting fever, and possibly persisting bacteremia or sepsis in the setting of IE, suggests the possibility of splenic abscess or infarction (with persisting bacteremia, sepsis, or fever probably favoring abscess). Abdominal CT or possibly US is appropriate for further evaluation in such cases. However, a high percentage of patients with splenic abscesses or infarcts lack abdominal pain or tenderness (more than 50% in some studies15,16) making accurate prediction difficult and general recommendations for abdominal CT in the setting of IE controversial.
It may be reasonable to obtain abdominal imaging studies in the subset of patients, who will likely undergo valve replacement surgery regardless of clinical findings. The rationale for this approach is to reduce the potential risk of bacteria from an occult splenic abscess infecting a newly placed prosthetic heart valve. It has been our experience that some clinicians and cardiovascular surgeons favor this approach and seriously consider splenectomy in patients with hypodense splenic lesions before valve replacement surgery. Unfortunately, at this time, there is simply insufficient data to address the appropriateness of this aggressive approach.
Distinguishing Between Splenic Infarcts and Abscesses Using Ultrasonography, Radiological and Radionuclide Procedures
Both ultrasonography and radionuclide studies, such as Technetium Tc 99m liver-spleen scans, gallium scans, and indium 111-labeled white blood cell scans, have largely been supplanted by more sensitive and specific abdominal CT.4 Whenever possible, contrast material should be given during abdominal CT to make areas of low density more clearly defined from surrounding normal splenic parenchyma. This occurs primarily as a result of diffuse perfusion of the contrast agent throughout the normal splenic tissue rather than the uncommon occurrence of rim enhancement along the perimeter of infarct or abscess (Fig. 1).17,18
Splenic infarcts typically have been described as being well defined, often multiple, peripheral wedge-shaped low-density defects with their apices pointed toward the hilum. Splenic abscesses, on the other hand, have been described as more commonly single, centrally located hypodense lesions of round or irregular shape.4
Not surprisingly, such clear distinctions frequently are not possible. Much overlap occurs between the CT patterns of splenic infarction and abscess.4 With rare exceptions,19 the diagnosis of abscess is unequivocal only if gas is present within the splenic defect. This, unfortunately, is an uncommon finding.17 Whereas, splenic infarcts may resolve over time without specific therapy,10 patients with IE are almost always receiving treatment with antibiotics bringing into question the reliability of that observation in the setting of endocarditis.
Distinguishing Between Splenic Infarcts and Abscesses Using US or CT-Guided Percutaneous Aspiration
In recent decades, there has been increasing interest and experience using US or CT-guided percutaneous aspiration, drainage, and biopsy for diagnosis and/or treatment of splenic disorders, including splenic abscesses.20 Nevertheless, concern with complications, especially hemorrhage, in this highly vascular organ prevents uniform acceptance of these procedures.21 Most reports of percutaneous procedures involving the spleen consist of relatively small series, and none represent randomized controlled studies. However, the available information suggests that, with skilled and experienced interventional radiologists, such procedures at least merit consideration. Of the 3 types of procedures, splenic aspiration for diagnostic purposes is the least complicated and possibly the safest. Lucey et al20 recently reported their experience with splenic aspirations in 8 patients. No complications occurred. Purulent material was aspirated from 3 of 8 abscesses in 3 patients. The other 5 patients were diagnosed as having splenic cyst or infarction.20 Diagnosis of splenic abscess requires retrieval of purulent material and/or fluid which is positive for microorganisms on stain and culture. If the fluid is bloody and culture positive while the patient is still bacteremic, differentiation of abscess from bacteremia may not be possible. At this time, there is insufficient experience to determine the sensitivity or specificity of splenic aspiration in confirming the diagnosis of splenic abscess. Possible barriers to accurate diagnosis include sampling errors; sterile aspirate resulting from antibiotic administration; lack of sufficient purulence in an early, evolving abscess; or thick, tenacious pus which cannot be withdrawn through a small-bore needle.
At this time, despite the potential risks and limitations mentioned above, attempting confirmation of splenic abscess using percutaneous aspiration appears to be a potentially helpful addition to our diagnostic methodology in patients with IE and splenic lesions.
Distinguishing Between Splenic Infarcts and Abscesses Pathologically After Splenectomy
In general, the distinction between uninfected intact, but ischemic, splenic parenchyma and culture or stain-positive pus in the case of splenic abscess would seem to be straightforward. However, even with this ultimate diagnostic procedure, differentiation may be confusing in the setting of IE.
It is postulated that many infarcts in patients with IE are either simultaneously infected at the time of embolization to the spleen or become infected at a later time from sustained bacteremia.3 Infected infarcts may vary in the likelihood or rate at which suppurative changes occur, possibly depending upon infecting microorganism, choice and duration of antibiotics, and duration of time from infection to splenectomy. Therefore, it seems likely that a variety of pathological changes may be present at time of splenectomy ranging from bland, uninfected infarction; uninfected ischemic necrosis; infected, but nonsuppurative infarction; or frank abscess.22 Moreover, truly infected spleens may be culture negative at time of splenectomy because of previously administered antibiotics.8
The pathologist can be most confident of the diagnosis of splenic abscess if bacterial stains or cultures are positive, and frank pus is present or an abundance of polymorphonuclear cells are visible on histological examination of necrotic tissue.4,5 However, the possible transition stages in ischemic splenic injury, especially if complicated by bacterial infection, may cause confusion and difficulty in diagnosis. We suspect that cases of culture-negative but infected infarcts which have not progressed to the stage of suppuration may be misdiagnosed as bland infarcts and that uninfected spleens with ischemic necrosis may be misidentified as abscesses. The extent to which these possible errors of diagnosis occur when the spleen is examined after splenectomy is not known at this time.
Treatment of Patients With Splenic Abscess in the Setting of IE
Bland splenic infarcts usually resolve slowly without intervention. Only occasionally does surgery with splenectomy become necessary to manage splenic infarcts complicated by subcapsular hemorrhage, splenic rupture, pending rupture, or significant ischemic necrosis with liquefaction.10
In contrast, patients with splenic abscess have traditionally been managed with splenectomy to reduce the high mortality rate associated with this disease. Although splenectomy in addition to administration of antibiotics remains the most common approach at this time, there is increased interest and experience using percutaneous drainage as an effective and less invasive alternative, allowing for spleen preservation.
Several reports of percutaneous drainage, mostly with relatively small numbers of patients, have appeared since the early 1980s, and although the results have not been uniformly successful, they have at least been encouraging. Care in selecting patients for whom the procedure is appropriate as well as meticulous methodology is crucial for safely performing nonsurgical drainage of splenic abscesses. Coagulation profiles and platelet counts should be normal. Patient cooperation is essential. Some authors recommend limiting the procedure to patients with well-formed, unilocular collections that have discrete walls and no internal septations or to patients with small abscesses less than 3 cm in diameter.23,24 Needle and drainage catheter placement must be guided by CT or real-time sonographic visualization avoiding colon, pleura, lung, and kidney.20
A recent review of splenic abscess drainage using percutaneous technique (not limited to patients with IE) reported a success rate of 60% to 72%, somewhat lower than the 80% to 90% success rates in patients with abscesses located elsewhere in the abdomen and pelvis.20 Potential complications of percutaneous drainage include fistula formation, bacteremia, empyema, lung injury, bowel perforation, intra-abdominal leakage of abscess contents, and bleeding.4
Despite encouraging reports of many authors, a recent review by experienced interventional radiologists expressed great caution in performing percutaneous drainage of splenic abscesses except in the most unusual circumstances. The primary concern expressed by these authors was for serious hemorrhagic complication due to needle and catheter placement.21 Accordingly, the general advisability and widespread acceptance of this procedure in replacing splenectomy awaits further experience and study.
Lastly, there remains the intriguing question of whether some patients with splenic abscesses (or nonsuppurative, infected infarcts) may be managed with antibiotics alone, avoiding percutaneous drainage and splenectomy. A few reports of success with antibiotic treatment alone are available.25,26 The approach would appear to have merit, especially with infected, nonsuppurative infarcts as long as repeated CT studies to monitor resolution are performed.4 Nevertheless, largely because of the difficulty in reliably distinguishing splenic infarct from abscess without (and sometimes with) invasive procedures, a fully satisfactory answer to this question may elude us.
1. Aster JC. Diseases of white blood cells, lymph nodes, spleen and thymus. In: Kumar V, Abbas AK, Fausto N, eds. Robins and Cotan Pathologic Basis of Disease
. 7th ed. Philadelphia, PA: Elsevier Saunders; 2005:702-705.
2. Fowler VG, Scheld WM, Bayer AS. Endocarditis and intravascular infections. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Disease
. 6th ed. Philadelphia, PA: Elsevier Churchill Livingstone; 2005:975-1022.
3. Bayer AS, Bolger AF, Taubert KA, et al. Diagnosis and management of infectious endocarditis and its complications. Circulation
4. Allan JD. Splenic abscess: pathogenesis, clinical features, diagnosis and treatment. Curr Clin Top Infect Dis
5. Johnson JD, Raff MJ, Barnwell PA, et al. Splenic abscess complicating infectious endocarditis. Arch Intern Med
6. Du LTH, Wechsler B, Cabane J, et al. Abscess splenic and infectious endocarditis. Arch Intern Med
7. Chun CH, Raff MJ, Contreras R, et al. Splenic abscess. Medicine
8. Ting W, Silverman NA, Arzouman DA, et al. Splenic septic emboli in endocarditis. Circulation
. 1990;82(suppl IV):105-109.
9. Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis
10. Jaroch MT, Broughan TA, Hermann RE. The natural history of splenic infarction. Surgery
11. O'Keefe JH, Holmes DR, Schaff HV, et al. Thromboembolic splenic infarction. Mayo Clin Proc
12. Nores M, Phillips EH, Morgenstern L, et al. The clinical spectrum of splenic infarction. Am Surg
13. Goerg C, Schwerk WB. Splenic infarction: sonographic patterns, diagnosis, follow-up and complications. Radiology
14. Haft JI, Altieri J, Smith LG, et al. Computed tomography of the abdomen in the diagnosis of splenic emboli. Arch Intern Med
15. Robinson SL, Saxe JM, Lucas CF, et al. Splenic abscess associated with endocarditis. Surgery
16. Trouillet JL, Hoen B, Battik R, et al. Splenic involvement in the course of infective endocarditis. Rev Med Interne
17. Johnson JD, Raff MJ, Drasin GF, et al. Radiology in the diagnosis of splenic abscess. Rev Infect Dis
18. Balcar I, Seltzer SE, Davis S, et al. CT patterns of splenic infarction: a clinical and experimental study. Radiology
19. Barzilai M, Schlag-Eisenberg D, Peled N, et al. Noninfectious gas accumulation in an infarcted spleen. Dig Surg
20. Lucey BC, Boland GW, Maher MM, et al. Percutaneous nonvascular splenic intervention: a 10-year review. AJR Am J Roentgenol
21. Haaga JR, Nakamoto D. Computed tomography-guided drainage of intra-abdominal infections. Curr Infect Dis Rep
22. Balthazar EJ, Hilton S, Naidich D, et al. CT of splenic and perisplenic abnormalities in septic patients. AJR Am J Roentgenol
23. Walia HS, Aman S, Walia HK. Rational approach to therapy in splenic abscess. J R Coll Surg Edinb
24. Gleich S, Wolin DA, Herbsman H. A review of percutaneous drainage in splenic abscess. Surg Gynecol Obstet
25. Jolobe OMP, Melnick SC. Splenic abscess: successful non-surgical therapy. Postgrad Med J
26. Cheesbrough JS, Jones EW, Finch RG. The management of splenic abscess. Q J Med
© 2007 Lippincott Williams & Wilkins, Inc.