key: cord-0784360-lxh0mv3x authors: Quiring, Robert; Burke, Victoria title: Escherichia coli prosthetic valve endocarditis from a non-genitourinary source date: 2021-11-05 journal: IDCases DOI: 10.1016/j.idcr.2021.e01329 sha: f2ffed5e90ace28a150dab1773b140f5022b392d doc_id: 784360 cord_uid: lxh0mv3x Escherichia coli (E. coli) is a rare cause of infective endocarditis due to its lack of traditional virulence factors that promote endocardial adherence. Previous case reports of E. coli infective endocarditis demonstrate specific risk factors to include advanced age over 70, female sex, diabetes, immunosuppression, and intravascular or cardiac devices. Antecedent genitourinary infection is the most common source. We present a case of a 55-year-old Honduran man with a recent bioprosthetic mitral valve replacement and tricuspid valve repair who presented with one month of subjective fevers, night sweats, anorexia, and significant weight loss. After extensive work-up, the patient was diagnosed with E. coli infective endocarditis secondary to E. coli growth in blood cultures and a transesophageal echocardiogram (TEE) revealing a vegetation on his prosthetic mitral valve. An indolent gastrointestinal source was suspected to be the source of infection with imaging only notable for mild periappendiceal stranding concerning for a possible site of antecedent inflammation. He was treated with a 6-week course of ceftriaxone and gentamicin inpatient and then discharged on trimethoprim-sulfamethoxazole suppressive therapy with serial echocardiographic follow-up given the persistent small vegetation on repeat echocardiogram. Our case report and review of ten recent cases of prosthetic valve endocarditis described in the literature illustrates several common features of the epidemiology, presentation, and management of E. coli prosthetic valve endocarditis including more commonly reported non-genitourinary sources of bacteremia, a trend towards more frequent surgical interventions, and a declining mortality rate. Infective endocarditis (IE) is a life-threatening infection of the endocardial surface of the heart or heart valves which can result in severe cardiac dysfunction. Most cases are caused by gram positive organisms or fastidious gram negative oral flora known as the "HACEK organisms" owing to the numerous virulence factors these pathogens possess aiding endocardial adherence [1, 2] . Escherichia coli, a gram-negative bacillus member of the Enterobacteriaceae family, typically lacks these traditional virulence factors and is subsequently an infrequent cause of IE accounting for only 0.5% of cases in a large international cohort [3, 4] . While there are a variety of sources of initial infection, an antecedent infection in the urinary tract is often identified as the source of initial bacteremia that seeds the heart valves [4, 5] . In this report we present a unique case of E.coli prosthetic mitral valve endocarditis in a 55-year-old man seeded from a non-genitourinary source. We subsequently review 10 cases of E.coli prosthetic valve endocarditis (PVE) found in the literature over the past 20 years and provide a characterization of this illness. A 55-year-old Honduran man with a past medical history of nonischemic cardiomyopathy due to valvular heart disease with https associated combined heart failure (ejection fraction 40%, stage III diastolic dysfunction), a recent bioprosthetic mitral valve replacement with tricuspid valve repair (2 months prior to admission), and paroxysmal atrial fibrillation presented to our hospital with 3 weeks of subjective fevers, night sweats, 30-pound weight loss, fatigue, and anorexia. One month prior to this presentation, the patient was briefly admitted following his valve replacement surgery with a day of watery diarrhea and was found to be in septic shock due to E. coli bacteremia of uncertain source. Evaluation during this admission included an unremarkable urinalysis and urine cultures and CT imaging only notable for distention and mild dilation of the small bowel suggestive of possible enteritis. Stool studies were not obtained given the self-limited nature of his antecedent diarrheal illness, but serum Strongyloides serologic testing was notably negative at the time. After a brief parenteral antimicrobials course, he was instructed to complete a two-week course of oral cefpodoxime 200 mg q12h upon discharge for transient bacteremia from a presumed self-limited gastrointestinal source. He did well clinically for one week after discharge before developing subjective fevers with generalized myalgias despite his outpatient cefpodoxime course. Over the following 3 weeks the patient progressively declined with more frequent subjective fevers, night sweats, malaise, anorexia with weight loss, and abdominal fullness further prompting another presentation to the emergency department. He denied sternal wound pain or drainage and a thorough review of systems for other localizing infectious symptoms was otherwise negative. On admission, vital signs were notable for a temperature of 38.8 °C (101.8 °F), blood pressure 91/62 mm Hg, heart rate 84 beats per minute, respiratory rate 21 breaths per minute, and oxygen saturation 100% on room air. Physical examination revealed an irregularly irregular heartbeat without murmur and an unremarkable lung exam. His abdomen was soft, nondistended, nontender to palpation, and no organomegaly was noted. No distal extremity edema or distal embolic phenomena were noted on skin exam. Notably, a stage 2 sacral wound that had developed following his valve surgery had healed completely since his prior admission. Laboratory tests were significant for a leukocytosis with a white blood cell count of 14000 (normal 4500-11000 cells/μl), anemia with a hemoglobin of 9.4 (normal 13.5-17.5 gm/dL), and normal kidney and liver function on complete metabolic panel. Inflammatory markers were elevated with a sedimentation rate of 119 mm/hour (normal 0-20 mm/hr). Urinalysis and culture were again unremarkable. Chest radiograph showed no evidence of acute cardiopulmonary disease, and an initial transthoracic echocardiogram showed no obvious vegetations or compromise of his prosthetic valve. The patient was started on broad spectrum antimicrobials with intravenous vancomycin 1 g q12h and cefepime 2 g IV q8h at the time of admission. Antibiotics were quickly narrowed to cefepime alone after admission blood cultures ultimately grew gram-negative rods in 4/4 collection tubes which later speciated as Escherichia coli. He was transitioned to ceftriaxone 2 g IV q24 based on sensitivity data, but blood cultures remained persistently positive on hospital day 2 and 3 raising concern for an endovascular nidus of infection. Transesophageal echocardiogram was obtained on hospital day 5 which revealed a small 3 mm mobile, filamentous vegetation on the anterior cusp of the prosthetic mitral valve without significant valvular compromise. As a result, he met Duke criteria for diagnosis of infective endocarditis by fulfilling one major criterion (TEE demonstrating a vegetation) and three minor clinical criteria (persistent fevers, predisposing cardiac condition, and persistently positive cultures.) [6] . Antimicrobial therapy was tailored to combination therapy with ongoing ceftriaxone 2 g IV q24h and gentamicin 5 mg/kg IV q24h which was adjusted continually based on levels of accumulation and changes in renal function. He continued to improve clinically, and blood cultures demonstrated complete clearance on hospital day 5. Given that the patient lacked evidence of an antecedent urinary tract infection as the source of his recurrent bacteremia, additional imaging was obtained to exclude a secondary source or nidus of infection. This included CT of the chest, abdomen, and pelvis to exclude an indolent abscess and doppler ultrasounds of the extremities and portal vasculature to exclude septic thrombophlebitis. An MRI of the full spine was also performed to exclude osteomyelitis/discitis given his prior sacral wound during his previous hospitalization. Imaging studies were only notable for mild periappendiceal stranding concerning for a possible site of antecedent inflammation as the source of his recurrent E.coli bacteremia. Given a benign abdominal exam, surgical consultants did not feel he met criteria for acute appendicitis and declined intervention. Neither a colonoscopy nor capsule endoscopy were performed during his inpatient stay owing to limitations on endoscopic procedures performed during a concurrent coronavirus pandemic surge. Given the patient's rapid clinical improvement on tailored antimicrobial coverage, cardiothoracic surgery concluded that the risks of redo-valve replacement outweighed the benefits as there were no significant valvular defects noted on echocardiogram. The patient subsequently completed a six-week course of combination antimicrobial therapy with appropriate levels of gentamicin and no adverse sequelae of therapy. Repeat echocardiogram showed a largely unchanged small mobile vegetation on the anterior cusp of the prosthetic mitral valve without valvular compromise. He was reassessed by cardiothoracic surgery who again recommended conservative management with chronic oral suppression and additional echocardiographic follow-up given his overall clinical improvement without significant valvular compromise on serial echocardiograms. He was subsequently transitioned to suppressive therapy with trimethoprim-sulfamethoxazole given the presence of continued vegetation on his prosthetic valve with plans for withdrawal after a distant repeat echocardiogram at 6 months of followup. The patient was reevaluated several weeks post-discharge and remained clinically well. Unfortunately, he was not seen again in follow-up at our clinic since he returned to his home country of Honduras for continuity of care. While both native and prosthetic valve IE due to E.coli has been described as a rare entity, previous studies have characterized E. coli IE as frequently stemming from nosocomial origin and more likely related to an antecedent genitourinary infection [3, 8] . These studies have also demonstrated a greater association with endovascular devices for cases of E. coli IE with PVE accounting for 8% of all cases before the year 2000 [8] . Since the year 2000, however, the share of E.coli IE cases involving prosthetic valves has significantly increased with PVE now accounting for 33% of 27 reported cases of E.coli IE described in a recent review of the literature [7] . This increase in prosthesis infection is likely a reflection of the increasing incidence of valve surgery and cardiac device implantation over this same time period [1, 2] . Previously characterized risk factors for the development of E.coli IE include advanced age over 70, female sex, diabetes, immunosuppression, and implanted intravascular or cardiac devices [5, [7] [8] . Table 1 illustrates the clinical characteristics of the 10 cases of E.coli PVE described in the literature over the last 20 years which includes our own case [8] [9] [10] [11] . Although antecedent urinary tract infection preceded the development of E.coli PVE in 40% of these reviewed cases, we found that documented gastrointestinal infection or pathology preceded the diagnosis of PVE in 50% of cases. This trend reflects a notable increase in recognition of this organ system as a source of E.coli PVE [7] . Notably the mitral valve remained the most commonly affected valve in PVE cases and distant embolic events were rare occurring in only 10% of cases. Valve ring abscess was a commonly described complication in these PVE cases reported in 50% of patients [8] [9] [10] [11] . Although exact antimicrobial regimens remain heterogeneous among individual cases, combination therapy is the standard treatment for PVE and was utilized in 60% of reviewed cases [8] [9] [10] [11] . Surgical intervention was performed frequently in 40% of cases owing to the increasing frequency of intracardiac abscess in PVE cases [8] [9] [10] [11] . Despite the inherent risks of combination antimicrobial therapy and surgery, early mortality is encouragingly lower at 10% compared to the reported 21% of overall cases of E. coli infective endocarditis reviewed by Akuzawa and collegues [7] . Our case described in this report illustrates several common epidemiological, presentation, and management trends of E.coli PVE. In particular, our patient possessed a major risk factor for the development of E.coli IE by nature of having an implanted prosthetic mitral valve. In contrast to previously described risk factors, our patient did not have an antecedent genitourinary infection as the source of his initial bacteremia. Although there was no concrete evidence for a gastrointestinal source or nidus of infection, serial imaging revealed small bowel and periappendiceal stranding which suggests sites of inflammation that may have resulted in translocation of E.coli into the bloodstream and initial seeding of his valve. Serum Strongyloides serologic testing was negative and unfortunately could not provide a unifying diagnosis to explain his patchy intestinal stranding and transient episodes of enteric bacteremia. Given the constraints of the pandemic on our local medical system and the patient's preordained return to his home country within several months of initial treatment, we could not further evaluate for a nidus of infection within the gastrointestinal tract with endoscopy. Regardless of the exact source of the initial infection, his presentation illustrates the emerging trend that the genitourinary system is no longer the only anticipated source of bacteremia in E.coli PVE cases. The patient did not have the feared complication of valve ring abscess requiring redo-valve replacement given his rather prompt initial presentation to the healthcare system and early, although intermittent, administration of antimicrobials. Consistent with the standard of care, he received combination antimicrobial therapy and did clinically well following his treatment course. Although E. coli endocarditis is a rare condition, it is necessary to ensure a prompt diagnosis and timely intervention due to the high mortality rate. With the relatively high incidence of E.coli bacteremia in hospitalized patients, practitioners should strongly consider including echocardiography in the diagnostic evaluation of those patients with persistent or relapsed E.coli bacteremia -particularly those patients with prosthetic valve replacements. Although urinary tract infections historically have been described as the most common source of E.coli bacteremia, our case and a recent literature review suggests an increasing prevalence of non-genitourinary sources of E.coli PVE. Therefore, clinicians should consider detailed evaluation of systems beyond the urinary tract when evaluating potential E. coli PVE cases. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for profit sectors. Written informed consent was obtained from the patient for publication of this case report. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request. Infective endocarditis Trends in infective endocarditis in California and New York state International Collaboration on Endocarditis Prospective Cohort Study (ICE-PCS) I. Non-HACEK gram-negative bacillus endocarditis Risk factors and outcomes of endocarditis due to non-HACEK gram-negative bacilli: data from the prospective multicenter Italian endocarditis study cohort Escherichia coli native valve endocarditis Infective Endocarditis in Adults: Diagnosis, Antimicrobial Therapy, and Management of Complications: A Scientific Statement for Healthcare Professionals from the Native valve endocarditis due to Escherichia coli infection: a case report and review of the literature Escherichia coli endocarditis: seven new cases in adults and review of the literature Endocarditis due to gram-negative bacilli at a French teaching hospital over a 6-year period: clinical characteristics and outcome A case of Escherichia coli endocarditis after hemorrhoidectomy performed by a herbalist A rare case of prosthetic valve endocarditis caused by extended-spectrum β-lactamase producing Escherichia coli Special thanks to LSUHSC Library and Reference Librarian, Kathryn Kerdolff, for her assistance in conducting a thorough literature review.