Risk of Infective Endocarditis in Streptococcus mitis Bloodstream Infections Among Patients with Neutropenia from Hematologic Malignancies

Abstract Streptococcus mitis commonly causes bloodstream infections (BSIs) in neutropenic patients but infrequently results in infective endocarditis (IE) in this population. Among 210 patients with neutropenia and S. mitis BSI, 55% underwent cardiac imaging. None were diagnosed with S. mitis IE; 3 had recurrent S. mitis BSI within 12 weeks.

Streptococcus mitis, an alpha hemolytic Streptococcus sp., is a relatively common cause of streptococcal bloodstream infection (BSI) associated with a subacute presentation of infective endocarditis (IE) in immunocompetent patients.In a recent health registry study, S. mitis/oralis accounted for ∼6% of monomicrobial streptococcal BSIs, but caused 19% of IE cases [1].In a US study of streptococcal IE, alpha hemolytic Streptococcus was responsible for over three-quarters of cases, of which S. mitis accounted for 61% of alpha hemolytic Streptococcus IE cases [2].
Alpha hemolytic streptococci and specifically S. mitis are also a common cause of BSI during chemotherapy-induced neutropenia, where they cause an acute clinical syndrome, not a subacute illness.While historically BSI during neutropenia has predominately been caused by gram-negative pathogens, decades ago there was an epidemiologic shift toward grampositive organisms including alpha hemolytic streptococci [3].Studies of alpha hemolytic streptococcal BSI during neutropenia in the 1980s-1990s defined risk factors including mucositis, use of cytarabine chemotherapy, and use of certain antimicrobial prophylaxis agents including fluroquinolones [3,4].S. mitis is the predominate cause of alpha hemolytic streptococcal BSI during neutropenia and has been linked to an acute alpha streptococcal shock syndrome with rash, prolonged fevers, and acute respiratory distress syndrome [4,5].
Unlike S. mitis BSI in immunocompetent patients where a subacute presentation of IE is a major clinical manifestation, in neutropenic hosts this pathogen behaves differently.Limited data suggest that S. mitis does not cause IE in neutropenic patients with hematologic malignancy (HM).In a study of 121 episodes of alpha hemolytic streptococcal BSI, 30 of 47 episodes in the general population fulfilled guideline-defined criteria for definite IE, while none of 74 episodes in patients with HM (69% of which were due to S. mitis) fulfilled definite IE criteria [5].More recently, a metanalysis of patients with neutropenia and alpha hemolytic streptococcal BSI that included 35 studies (943 patients) suggested that the rate of endocarditis in the population is <1%.However, this metanalysis was limited by a lack of systematic assessment for IE (ie, cardiac imaging) in all included studies [6].The authors suggest that routine screening for IE with either transthoracic (TTE) or transesophageal echocardiogram (TEE) may be unnecessary in this population in the absence of a new murmur.
We undertook this study to describe the incidence of IE in a contemporary cohort of neutropenic HM patients with S. mitis BSI at 3 institutions where cardiac imaging is pursued about half of the time.A goal of this analysis was to assess the clinical impact of cardiac imaging in this patient population.

METHODS
We performed a retrospective cohort study of all adults with HM and S. mitis BSI during a concurrent period of neutropenia Patients were identified by electronic medical record search for blood cultures positive for S. mitis during the study period.Hematologic malignancy included patients with acute and chronic leukemia, lymphoma, and multiple myeloma.Neutropenia was defined as absolute neutrophil count [ANC] <500 K/µL; patients were included if they had an ANC <500 within 48 hours of first positive blood culture.
Medical records were reviewed for covariates of interest including demographics, HM type, previous hematopoietic cell transplant, use of fluoroquinolone prophylaxis within 24 hours of bacteremia, baseline cardiac valvular conditions or cardiac devices, personal history of IE, presence and management of indwelling central venous catheters (CVCs), relevant microbiological information, and cardiac imaging results including TTE at baseline and TTE, TEE, and cardiac computed tomography (CT) within 7 days of BSI.
The primary outcome was number of patients diagnosed with IE based on updated Duke International Society for Cardiovascular Infectious Diseases (ISCVID) criteria [7].Secondary outcomes included number of patients with findings on cardiac imaging, duration of therapy, and BSI recurrence within 12 weeks.

RESULTS
Among 210 patients with neutropenia and concurrent S. mitis BSI, 123 (59%) were male, and the median age (interquartile range [IQR]) was 59 (43-66) years.Baseline demographics, microbiological information, and outcomes are shown in Table 1.Most patients had an acute febrile illness-only 3 (1%) had fever for >1 day before blood cultures were positive.All patients had native cardiac valves, and 3 (1%) had cardiac devices.Of 177 patients who had a baseline echocardiogram, 5 had moderate or severe aortic or mitral valve abnormalities (stenosis or regurgitation).Eighty-three patients (40%) had a levofloxacinintermediate or resistant S. mitis isolate, of whom 53 (64%) received a fluoroquinolone within 24 hours before positive blood culture.Additionally, 62 patients (30%) had polymicrobial BSI; common co-pathogens are shown in Table 1.Most patients (n = 115, 59%) had and retained a CVC.The median duration of antimicrobial therapy for S. mitis BSI (IQR) was 15 (14-19) days, and 16 patients (8%) died during index hospital admission.• Prosthetic valve 0 (0) • Mitral valve (regurgitation, n = 1) 1 (0.5) Pacemaker or intracardiac device 3 No patients in the cohort developed definite IE, and 1 patient with polymicrobial bacteremia had possible endocarditis due to coagulase-negative staphylococci.One hundred fifteen patients had a TTE (55%) within 7 days of BSI.Of these 115 patients, 1 had suspected valvular vegetations on TTE, but confirmatory TEE did not show any vegetations.Two patients had negative TTEs, but given a high clinical suspicion a subsequent cardiac CT was performed, which showed no valvular vegetations.No patient underwent positron emission tomography-CT or single-photon emission-CT imaging.Despite negative cardiac imaging (TTE followed by cardiac CT), 1 patient completed empiric treatment for possible IE by Duke-ISCVID criteria [7].The patient had 1 major criterion: Staphylococcus schleiferi bacteremia in 3 sets of blood cultures (drawn on 2 days); and 2 minor criteria: fever and vascular phenomena (possible splenic abscesses).The patient had a single blood culture bottle that also grew S. mitis concurrent with S. schleiferi.Therefore, the patient received a 6-week course of vancomycin therapy.
Recurrence of S. mitis BSI within 12 weeks of first positive blood culture occurred in only 3 patients.All had negative cardiac imaging upon documentation of recurrent infection.One patient recurred at 6.5 weeks after initial positive blood culture, presumptively from gastrointestinal translocation.A second patient with AML and metastatic colon cancer recurred 10.5 weeks after initial positive blood culture and was treated with 6 weeks of ertapenem for presumed liver abscesses related to colon cancer.The third patient recurred at 11 weeks and had concomitant pulmonary nodules and was treated with a course of empirical antifungal therapy and 4 weeks of IV ceftriaxone.

DISCUSSION
Previous studies [5,6] suggest that alpha hemolytic streptococcal BSI is rarely linked to IE in patients with neutropenia, but these were limited by lack of both cardiac imaging and description of the clinical syndrome.Others have hypothesized that the risk for endocarditis is lower in this population due to thrombocytopenia, which is often concurrent with neutropenia in patients with HM, as platelets play a key role in the pathogenesis of vegetation formation [6].In this cohort of patients with concurrent neutropenia and S. mitis BSI, most patients had an acute febrile illness without longer-term fevers that might suggest subacute IE.Fifty-five percent of patients underwent cardiac imaging, and no patients developed definitive IE based on cardiac imaging and application of updated Duke-ISCVID criteria.These results suggest that obtaining cardiac imaging in this patient population solely based on the presence of S. mitis BSI is likely unnecessary.
Ninety-five patients did not undergo cardiac imaging.To ensure that a diagnosis of IE was not inadvertently missed, we assessed for recurrent S. mitis BSI within 12 weeks.Three patients had recurrence after initial therapy, 2 of whom did not undergo TTE initially.However, cardiac imaging with subsequent BSI was not consistent with IE, suggesting that the recurrence within a 12-week period was likely not related to undiagnosed IE.Despite negative TTE and cardiac CT, 1 patient received 6 weeks of empiric vancomycin therapy for concurrent S. schleiferi and possible IE.Not surprisingly, many patients with a fluoroquinolone-resistant S. mitis isolate were exposed to a fluoroquinolone within the previous 90 days.Although use of fluoroquinolone prophylaxis during neutropenia is recommended by guidelines [9,10], prior exposure can limit oral treatment options for patients who develop S. mitis BSI.
This study is limited by the low number of patients with underlying cardiac valvular pathology, previous IE, mechanical valves, or cardiac devices that might predispose to endocarditis.It is not clear based on our study whether neutropenic patients with S. mitis BSI with underlying valvular disease/pathology are at greater risk for IE and therefore may need cardiac imaging performed.Additionally, because the median duration of therapy was about 2 weeks, it is possible that there were some undiagnosed cases of IE, particularly right-sided IE, cured by 2 weeks of antibiotics.
In conclusion, S. mitis BSI is a relatively common acute febrile illness in chemotherapy-induced neutropenia, likely the result of transient translocation of organisms from damaged mucosal surfaces, and was not a significant cause of IE in our cohort.Routine use of cardiac imaging to rule out IE appears unnecessary for many patients in this population.