Infective Endocarditis Antibiotic Prophylaxis: Practice-Changing Data for Patients Undergoing Non-Dental Invasive Procedures

In his landmark 1885 Gulstonian Lecture, Sir William Osler reviewed the clinical features of endocarditis, including the "probable connection of the disease with the presence of micro-organisms" ⁽¹⁾. In the intervening ~140 years, the disease itself has fundamentally changed, as rheumatic heart disease has receded and novel risk factors have emerged, including an ever-expanding array of invasive medical and surgical procedures in patients with underlying structural heart disease ⁽²⁾.

In response to this shifting epidemiology, multiple iterations of society guidelines have attempted to define which patients are most likely to benefit from antibiotic prophylaxis to prevent IE. Though incidence is increasing, IE remains sufficiently rare that prospective randomized trials have not been conducted. Thus, guidelines have, by necessity, relied upon observational data and expert opinion. Recent versions, including the 2021 American Heart Association/American College of Cardiology (AHA/ACC) and the 2015 European Society of Cardiology (ESC) guidelines, have focused on prevention of transient bacteremia associated with invasive dental procedures among high-risk patients (those with prosthetic valves or valve repair, cyanotic congenital heart disease, heart transplant with dysfunctional valves, or with a prior history of IE) for which the supporting data is most robust ^{(3, 4)}. Conversely, little is known about the association of non-dental invasive procedures (NDIP) and the development of IE. In early iterations of published guidelines, antibiotic prophylaxis prior to NDIPs was recommended in moderate- and high-risk patients. With evolving understanding of the adverse effects of unnecessary use of antibiotics, including the development of antimicrobial resistance, and the lack of robust evidence to support the association of NDIPs with development of IE in these patients, the updated guidelines do not recommend this practice. However, two recently published nationwide studies and several smaller cohort, case-control, and cross-over studies have prompted Larry Baddour and colleagues to publish an AHA science advisory calling for the matter to be revisited ⁽⁵⁾.

To contextualize current recommendations with the available data on the risk of developing IE after NDIPs, the authors conducted a structured literature search and identified five case control, two case-crossover, and one cohort study of interest. The five case control studies included a total of 1,276 patients with IE that were evaluated to establish the odds of having undergone a procedure and the development of IE ^(6-10). These studies each assessed a different population and various NDIPs as risk factors for IE. The NDIPs included cutaneous and mucosal surface procedures, cardiovascular, gastrointestinal, genitourinary, and surgical procedures. Some of the NDIPs that were found to be associated with IE were barium enema, peripheral or central venous catheter placement, urological, and cardiovascular procedures. However, the odds ratios attributed to these NDIPs had wide confidence intervals across all the studies such that precise point estimates cannot be made. The cohort study utilized a random sample of Medicare beneficiaries undergoing colonoscopy ⁽¹¹⁾ and evaluated the three-month post-procedural risk of developing IE. In this cohort, the patients defined by the authors as high-risk for IE (those with valve disease, structural heart disease, intravascular devices, or end-stage renal disease) were more likely to develop IE after colonoscopy. However, it was unclear if this association was a result of the presence of any underlying colonic lesions or the colonoscopy itself.

Finally, and most importantly, Baddour and colleagues highlight two more recently published large case-crossover studies. These studies included over 7,000 patients from a nationwide Swedish registry, and over 14,000 hospitalized patients from national admissions data in England ^{(12, 13)}. Both studies found that among patients hospitalized for treatment of IE, patients were more likely to have an NDIP in the three months prior to developing IE compared to a control period prior to admission. The NDIPs included diagnostic and therapeutic procedures such as bone marrow biopsy, coronary angiography, endoscopy, and cardiac device implantation, among others.

While these case-crossover studies were more robust than the case control studies, the authors acknowledge their limitations. These include the absence of microbiological data to allow a more robust assessment of the association between the procedure and the development of IE. Additionally, while guidelines do not recommend prophylaxis prior to NDIPs for the prevention of IE, antibiotics may have been administered to these patients for other reasons. Lastly, some of these procedures may have been performed for the diagnosis of IE, which could create a temporal but not causal association between the procedure and diagnosis of IE. Future studies evaluating the causal relationship between NDIPs and IE in certain high-risk populations would benefit from prospective and systematic collection of information regarding indications for procedures, existing patterns of use of preprocedural antibiotic prophylaxis, and microbiological data.

These limitations notwithstanding, the best available data now suggests that some NDIPs are associated with IE, and that specific subgroups of patients are at higher risk than others for developing post-NDIP IE. The potential benefits for antibiotic prophylaxis likely outweigh the risks associated with increasing antibiotic use in these at-risk patient subgroups. As Osler said in 1885, "It is of use, from time to time, to take stock, so to speak, of our knowledge of a particular disease, to see exactly where we stand in regard to it, to inquire to what conclusions the accumulated facts seem to point, and to ascertain in what direction we may look for fruitful investigations in the future". On this basis, the guideline recommendations for antibiotic prophylaxis prior to NDIPs in high-risk patients must be re-evaluated and take into consideration the rapidly evolving epidemiology of IE.