Blood culture negative endocarditis: dire straits for the Heart Team

Infective endocarditis is a severe condition, with high morbidity and mortality.^1-3 Blood culture negative endocarditis is a subset of IE which represents an even more defying scenario, as the practicing physician must treat the patient with antimicrobials which are not directed to an identified pathogen, but rather to possible pathogens. Not surprisingly, BCNE has greater mortality, as reported in the recent predominantly European series of patients included between 2016-2018, wherein patients with culture negative IE (defined as patients with IE with negative blood and tissue cultures) were compared with those with BCPE.⁴ In this publication, 30- day mortality was 10.2% for 263/2590 patients with BCPE vs 78/523 (14.9%) for those with CNE, and death at 1 year was still significantly different between groups, being 22.5% for BCPE vs 25.8% for CNE. The latter difference between groups did not hold for the subsets of patients who had undergone surgery.⁴ Other smaller series have also shown greater mortality for BCNE.^5-6 Some series have not shown greater mortality but high morbidity, with complicating heart failure being more frequent in BCNE.^7-8 For all these reasons, the consensus statement provided by DeSimone and expert colleagues is an important contribution to field. ⁹

The frequency of BCNE varies geographically, reported as 10% in Europe, 7% in North America, and 20% in South America in the multinational study prospective cohort study from the International Collaboration on Endocarditis group.¹ In the years 2000 to 2005, 2781 adults with definite IE were identified, according to the modified Duke criteria.¹⁰ More recently, the EURO Endo study reported BCNE in 16.8% of 3116 adult patients;² similarly, a German study of 521 patients with IE who were operated on from 2013-2020 had a 16.3% rate of BCNE. ¹¹ Since these estimates relate to tertiary centers' data, the conundrums are not so much laboratory resources but clinical practice, especially the prescription of antibiotics and the low suspicion of IE as the cause of an infectious syndrome in the first place, resulting in a late diagnosis.

Although blood cultures usually become positive within 5 days for most culturable microorganisms causing true bacteremia or fungemia,¹² incubation times should be prolonged for patients with suspected IE who have negative cultures after 3 days incubation, especially if the patients have prosthetic valve IE or intracardiac devices, as Cutibacterium acnes is a possible agent, and related blood cultures become positive in a median of 7 days (IQR, 6-9 days);¹³ a recent multicenter retrospective study reported that 54.3% of C.acnes IE was initially BCNE.¹⁴ Moreover, if Bartonellae are suspected, based on epidemiological grounds, incubation times need to be extended for up to 3 weeks, though serology will be a quicker means to diagnosis.¹⁵

Approximately half the patients with left-side endocarditis undergo operations.^1.2 This presents an opportunity for establishing diagnosis in those who have negative blood cultures. Valve culture should be requested, but with the knowledge that they have a low yield, of around 15-41%, especially in those patients who have received more than 2 weeks antimicrobial therapy. ^16-19 Furthermore, valve culture results may lead to confusion when organisms such as coagulase negative staphylococci, Acinetobacter, Ralstonia picketti and other non- Pseudomonas aeruginosa non- fermenting rods are grown, as they may represent contaminants picked up during surgery or when manipulating the specimen.^16,17 Cell valve cultures have a higher yield, of 45.7% .¹⁵ Besides valve culture, histopathology is mandatory and may be especially helpful in centers or countries with fewer resources, as it may confirm the diagnosis of IE and provide clues to its etiology by showing morphological features such as cocci, rods or yeasts in the Brown-Brenn or other stains.²⁰ Cocci on histopathology, unfortunately, are the morphological presentation for staphs, streps and enterococci, but finding rods or yeasts is oftentimes helpful.

Molecular diagnosis of excised valves is a useful tool, as discussed in the statement,⁹ but is often inaccessible in low- middle- income countries. Besides, a positive PCR from a valvular specimen may represent prior infection, as DNA may persist for months to years in cardiac valves following an adequately treated IE episode.^21-23 These findings reinforce the importance of correlating valvular molecular biology results with histopathology of valves. On the other extreme of the spectrum of diagnostic possibilities, a recent study from Germany evaluated excised native or prosthetic valves of patients who had IE and who underwent conventional microbiological culture and fluorescence in situ hybridization combined with 16S rRNA-gene polymerase chain reaction and sequencing (FISHseq), finding that while valve cultures alone identified pathogens in 32% of BCNIE patients, the combination with FISHseq confirmed the infective etiology in 98% and identified causative pathogens in 52%.¹¹

The possible causative microorganisms in BCNE depend on epidemiological issues, for example, the site of acquisition for IE. If the infection was acquired in the community, relevant aspects are features of the place where the patient lives, for example, if it is a rural area, where Coxiella burnetii infection may be seen in farm animals and their products, such as milk, butter, cheese and hides, or even in the dust in the surrounding environment; ²⁴ what the patient does for a living, for example, is he/she a veterinarian, a farmer, does he/she have contact with animals , for example, in a PET shop ; does he/she have any pets at home ( cats are particularly important for B.henselae infection),²⁵ and specifically, have mammal pets given birth in the recent days to weeks (infected mammals' placentas are rich in Coxiella which may aerosolize).²⁴ Other important information in the patient's history is how well the patient looks after his/her oral cavity, whether he/she has been submitted to oral, genitourinary or gastrointestinal procedures in recent days or weeks, whether he/she has been bitten by insects such as mosquitoes, or ticks, or has had piercings, that could serve as portals of entry for skin microbes. ²⁶ Empirical treatment regimens should cover oral streptococci, staphylococci, and enterococci, and less often, the HACEK group; vancomycin and ceftriaxone are appropriate,⁹ though synergism for enterococci and for patients who harbor prosthetic material will be lacking. Moreover, if Bartonella and Coxiella are suspected, doxycycline may be added empirically, while waiting for serological or molecular biology results.

In health care settings, causative microorganisms often involve staphylococci, but enterococci, (multidrug resistant) Gram negatives and Candida are progressively more frequently reported.^27,28 Noteworthily, antibiotic use is common in nosocomial, and health care- associated non- nosocomial settings, and IE acquisition in these scenarios often results from bacteremia or fungemia secondary to central line infections or other infections, such as pneumonia or deep wound site infections. ^27,28 These considerations are important regarding empirical antimicrobial therapy for health care-related BCNE. Treating BCNE in patients who have health care-related BCNE will depend on the susceptibility pattern encountered for Gram negatives locally, as ESBL producers are commonly found in this scenario. A frequently used empirical combination therapy is vancomycin and meropenem; the addition of rifampin and/or gentamycin for synergy makes sense for staphylococci, enterococci and Gram negatives. The duration will usually be not less than 6 weeks intravenous therapy for beta lactams and glycopeptides. Fungi should be particularly suspected in patients with large vegetations and frequent recurring embolization despite antibiotic therapy, and antifungal empirical therapy is recommended especially in early prosthetic valve IE.²⁹ In inoperable patients due to unacceptably high surgical risk, oral antimicrobials are often continued for a long period of time, and PET CT scans may be of help in judging when to stop.^30,31 However, deciding on the choice of oral antibiotics for possibly resistant unknown pathogens is a nearly impossible task. Besides imaging, metagenomic sequencing may indicate the fall in the microbial counts or even negativity of counts ³², further helping the clinician to decide when antimicrobials suffice.

In summary, BCNE is a niche of IE that deserves careful consideration regarding diagnosis and treatment. The AHA statement contributes to clarify several points and draw the clinicians' attention to this special subset of patients, who need close monitoring and follow up.