Endocarditis caused by probiotics and in patients with HHT are both rare conditions and the case we present is likely the result of different pathogenetic mechanisms operating in a single patient.
In cases of sepsis or bacteremia caused by microorganisms considered probiotics, mechanisms like bacterial translocation from the gut or other alterations in the integrity of mucosal surfaces permitting passage of microbes in the bloodstream have been hypothesized. In a series of 89 episodes of bacteremia caused by lactobacilli, the majority of patients had severe underlying diseases, with a predominance of gastrointestinal or hepatic neoplasia, and the mortality at 1 year was high (69%). In this series, L. rhamnosus was the most frequent isolate (53%); no cases of endocarditis were found [27]. In a series of 73 endocarditis caused by lactobacilli, 63% of patients had an underlying structural heart disease and 12% suffered from previous episodes of endocarditis. Dental procedures were reported in 47% of cases, but consumption of probiotics was reported only in 3 cases [28].
In our literature search of endocarditis associated with probiotics use, we noticed that all patients had infections caused by microorganisms belonging to the species Lactobacillus. The majority of patients were not immunosuppressed, but instead had predisposing factors for bacteremia, like alteration in gut mucosal integrity, colonoscopy or teeth extraction. It is well known that in such circumstances patient may develop bacteremia and endocarditis caused by commensal microorganisms like oral streptococci or saprophytic microorganisms of the gut. The preponderance of Lactobacillus among infections caused by probiotics is interesting. It has been shown that Lactobacillus rhamnosus strains isolated from endocarditis possess the ability to aggregate platelets, to bind collagen and fibrinogen, and to produce glycosidases and proteases [29]. In an experimental study using the rat model of experimental endocarditis aimed at assessing the potential pathogenicity of probiotic and clinical isolates of Lactobacillus rhamnosus and Lactobacillus paracasei, it has been showed that at least some of the probiotic strains of L. rhamnosus exibit a 90% infective dose (ID90) that is comparable to that of clinical isolates of Lactobacillus. These probiotic strains share the same fluorescent amplified fragment length polymorphism (FAFLP) as the clinical isolates with the lowest ID90, although they can be differentiated from clinical isolates by pulsed-field gel electrophoresis (PFGE) fingerprinting [30]. Moreover, strains of L. rhamnosus isolated in endocarditis and other serious infections that were indistinguishable from strains of probiotic L. rhamnosus by PFGE have been described [15, 31]. These results suggest that at least in some circumstances probiotic strains of L. rhamnosus can give rise to bacteremia and endocarditis. Lactobacilli are intrinsically resistant to vancomycin, and may be resistant to various other antibiotics [32], so the treatment of infections caused by these microorganisms can be difficult.
In a series of 85 blood isolates of Lactobacillus [33] all species demonstrated low minimum inhibitory concentrations (MICs) of imipenem, piperacillin-tazobactam, netilmicin and clindamycin and high MICs of vancomycin; instead, the MICs of benzilpenicillin, ampicillin, ceftriaxone and other cephalosporins varied widely among species and among different isolates of the same species. Among cephalosporins, ceftriaxone showed the highest MIC values for isolates of L. rhamnosus. Ampicillin, broad-spectrum cephalosporins and vancomycin are often used in empiric treatment of endocarditis, but they could prove ineffective in bloodstream infection caused by lactobacilli.
People with HHT are predisposed to infections, because of the presence of pulmonary, hepatic and gut arterovenous malformation. The fragile mucosa of the nasal cavity is also considered a portal of entry for pathogenic microorganisms in these patients. The majority of infections reported in HHT patients are cerebral abscesses, however other infections like hepatic abscesses, bacteremia, septic arthritis, osteomyelitis, skin infections and infective endocarditis are also reported [10, 34]. In our literature search, we found only 6 cases of infective endocarditis in patients with HHT, caused by coagulase positive and negative staphylococci and S. mitis. The mortality rate for these episodes was 16%. The causative microorganisms were common commensal of the skin, nose, mouth, oropharyngeal tract and gut. The majority of the patients reported frequent epistaxis with nasal packing but no other risk factors for bacteremia. It has been hypothesized that nasal mucosa may be the portal of entry for microorganisms in case of nasal trauma associated with nasal packing; in other cases, proliferation of microorganisms like S. aureus was favored by the presence of foreign material in the nasal cavity [35, 36]. It has been suggested that patients with HHT are considered at risk of bacteremia and endocarditis after dental procedures, and should be therefore treated with prophylactic antibiotics in these circumstances [37]; the same risk could be assumed for nasal procedures involving trauma to the nasal mucosa, and in these circumstances, antibiotic prophylaxis should probably be offered in these patients. It has also been suggested that these patient undergo nasal decontamination with mupirocin before nasal packing [38].
In the case of our patient, probably different factors have concurred to the occurrence of endocarditis. Although the patient’s history of recurrent infections and bacteremia can raise the suspect of an immune depression, this was not apparently the case. The levels of lymphocytes, neutrophils, immunoglobulins and complement were normal during intervals free of disease. Human immunodeficiency virus (HIV) serology was negative; the patient had no signs of neoplasia and was not taking immunosuppressive drugs. In the patient’s previous infective episodes, the isolates (Streptococcus mitis, Staphylococcus aureus) were typical pathogenic strains. Although the portal of entry of these bacteria was not clear, as apparently the patient had not an arteriovenous malformation in the lung, gut or brain, we hypothesize that the portal of entry was one of the numerous arteriovenous malformation in the nose or the oral cavity.
During the last infectious episode, the patient underwent a prolonged hospitalization in an intensive care unit and then in a long-term facility, received many courses of antibiotics and suffered from prolonged episodes of diarrhea; during this period the patient started to take large amounts of probiotics, many of them containing lactobacilli; therefore, it is plausible to assume that an imbalance in the bacterial flora of the gut and/or the oral cavity, and perhaps the nose, with a relative predominance of Lactobacillus could have occurred. We were not able to characterize the lactobacilli contained in the probiotics ingested by the patient, so we can only suppose that this strain of Lactobacillus rhamnosus was of dietary origin (we recognize that this is a limitation of our study). During the months before the last infectious episode, the patient had many episodes of prolonged epistaxis and underwent many traumatic nasal interventions such as nasal packing and cauterization of nasal telangiectasias. In patients with HHT, prolonged epistaxis is associated with extracerebral infections [9] and nasal instrumentation has been suggested as a possible risk factor for bacteremia [35]. Moreover, we cannot exclude other mechanisms and portals of entry of microbes in the bloodstream, such as microbial translocations of bacteria from the gut. Finally, the presence of a valvular aortic prosthesis was a clear predisposing factor for the development of infective endocarditis.