Infectious endocarditis on percutaneous aortic valve prosthesis: comparison with surgical bioprostheses

INTRODUCTION

Transcatheter aortic valve implantation (TAVI) revolutionized the treatment of severe aortic valve stenosis over the past decade.1,2 Furthermore, in recent years, there has been a clear preference for TAVI over surgical techniques regarding the treatment of this valvular heart disease, with an increasing number of patients, including low surgical risk ones.3 Although post-TAVI infective endocarditis (IE) is a rare complication reported in 1% up to 6% of patients, it is often associated with grim clinical outcomes and high mortality rates despite diagnostic1,2 and therapeutic4 advances. Complications are expected to rise exponentially as the number of TAVIs continues to rise at a constant rate in our setting as well.5

Few studies have compared the incidence of IE on TAVI and surgical aortic valve replacement (SAVR). However, randomized clinical trials have shown similar annual incidences of IE after SAVR and TAVI.6,7 Due to the strict patient selection of such studies, the results obtained in each center during the routine clinical practice may vary, and few studies like this using dedicated databases have been conducted. A multicenter study8 revealed the characteristics of IE on TAVI vs SAVR and the outcomes of a large patient cohort, yet it did not analyze the incidence of IE. Assessing the outcomes of this severe complication at each center should be a priority due to its potential impact on decision-making for the heart team.3

The aim of the present study was to analyze both the incidence and mortality rates of IE on TAVI vs SAVR patients from a tertiary teaching hospital in a long series of patients with severe aortic stenosis.

METHODS

Study design and population

We conducted an observational, single-center, retrospective study of a prospective cohort including all cases of IE in symptomatic patients with severe aortic valve stenosis treated with TAVI or SAVR using a biological prosthetic valve, diagnosed, and followed at a tertiary teaching referral hospital by an endocarditis team from January 2008 through December 2022.

Endpoints and definitions

The primary endpoint of the study was to analyze the overall incidence and mortality rates of IE on TAVI at our center. Secondary endpoint was to compare both rates in patients with IE on TAVI and SAVR during the same period. Another secondary endpoint was to study the number of patients who underwent surgery for IE in both groups. IE was diagnosed using the Duke criteria9 or the 2015 European Society of Cardiology modified criteria10 depending on the time of diagnosis in the study period. Cases of IE in TAVI patients during this period were identified, and their characteristics were compared with those of IE in SAVR patients.

Follow-up

Follow-up events were defined based on the criteria established by the Valve Academic Research Consortium-2.11 All complications, need for surgery, and mortality at the follow-up were recorded. IE-related cardiac complications included decompensated heart failure, fistula, prosthetic dehiscence, abscess, and complete atrioventricular block. Systemic complications included acute kidney failure, sepsis, embolism, and disorders of the central nervous system. Early IE was defined as that occurring within the first year after TAVI or surgery based on European recommendations.10 Since 1987, a prospective protocol for inclusion and follow-up of all IE cases has been in place at our center, with a systematic registry including in-person visits, at least, annually for all patients, as well as phone consultations when needed.

Statistical analysis

Qualitative data are expressed as percentages, and continuous data as mean and standard deviation or median [interquartile range], depending on whether they follow a normal distribution. Inter-group comparisons were drawn using the chi-square test or Fisher’s exact test for qualitative variables, and the Student’s t-test or Mann-Whitney U test for continuous variables, as appropriate. Time-to-event analyses for all-cause mortality were conducted using Kaplan-Meier curves. All tests were two-sided, and P values < .05 were considered statistically significant. Statistical analyses were performed with SPSS software (version 24; IBM Corp, Armonk, NY, United States).

RESULTS

Study population and incidence rate

During the study period, a total of 778 patients successfully underwent TAVI, 70% of them with self-expanding valves. After a median follow-up of 49 months (p25-p75, 29–108 months), 10 cases of IE were eventually diagnosed, which amounts to an overall incidence rate of 1.29% (an incidence rate of 0.09/100 patient-years). Twenty-four of the 1457 patients treated with surgical bioprostheses were diagnosed with IE for an overall incidence rate of 1.64% (an incidence rate of 0.12/100 patient-years). The hazard ratio for the IE incidence rate between the 2 groups was 0.75 (95% confidence interval, 0.36–1.57). The incidence rate of IE on TAVI remained stable throughout the study period, as did the incidence of IE on surgical bioprostheses. Four of the 10 IE cases in TAVI occurred between 2008 and 2015—incidence of 1.21%—and 6 between 2016 and 2022 (incidence of 1.33%). The incidence rates of IE on surgical bioprostheses for the same periods were 1.62% and 1.67%, respectively. Clinical characteristics, treatment, and mortality rates were also similar across periods in both IE groups.

Characteristics of IE in TAVI and SAVR

Approximately half of the cases reported in both groups turned out to be early IE: 5 of the 10 TAVI IE cases and 11 of the 24 cases described on surgical bioprostheses occurred within the first year after implantation (50% TAVI and 46% SAVR; P = .56). The remaining cases were late prosthetic IE (table 1). Among TAVI IE cases, the 5 early ones were diagnosed 2, 4, 6, 8, and 11 months after implantation, while the 5 late ones were diagnosed within years 2 (3 cases) and 3 (2 cases). Among the IE reported on surgical bioprostheses, early cases were diagnosed within the first 2 months (2 cases), between months 3 and 6 (3 cases), and between months 6 and 12 (5 cases) after surgery. Regarding late cases, 5 occurred in year 2, 4 in year 3, and 4 > 3 years after surgery.

Table 1. Baseline characteristics of patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

A possible source of infection different from implantation or surgery was identified in 2 of the 10 TAVI IE cases (20%) and in 2 of the 24 surgical bioprostheses. IE cases (8.3%). Only 1 of the early IE cases was associated with a possible infection source: a colonoscopy in a TAVI IE patient due to Enterococcus. None of the SAVR IE cases had an identified source, with implantation or the perioperative period being regarded as the probable sources of infection in the remaining patients. Among late IE cases, 1 TAVI IE case was associated with a dental procedure—despite proper antibiotic prophylaxis—and 2 SAVR IE cases with an upper gastrointestinal endoscopy and a dental visit, respectively. These patients’ clinical characteristics are shown in table 1. TAVI patients were older, with a median [interquartile range] age of 76 years [67–85] vs 63 years [49–77] of those treated with surgical bioprostheses (P < .001) and had higher rates of diabetes (70% vs 29%; P = .034). However, there were no significant differences in sex, other comorbidities, or symptoms. As expected, the TAVI group had a significantly higher EuroSCORE II vs the SAVR group (table 1). The profile of causative pathogens was very similar between the 2 groups (table 2). Enterococci were the pathogens most widely identified, followed by coagulase-negative staphylococci and Staphylococcus aureus, also with no significant differences being reported between the 2 groups. In 17% of SAVR patients and 20% of TAVI patients, the causative agent could not be identified (figure 1). Diagnostic echocardiographic findings of IE were also very similar between the 2 groups: transthoracic echocardiography identified IE only in half of the TAVI cases vs 37% of the cases treated with surgical bioprostheses (table 2). Transesophageal echocardiography diagnosed all IE cases in the TAVI group and slightly fewer in the SAVR group (table 2 and figure 1). In TAVI IE patients, the echocardiographic lesions identified included 9 endocardial vegetations—all affecting the leaflets—with 3 stent expansions (figure 2), 1 periaortic abscess, and 1 right ventricular perforation.

Table 2. Microbiological profile of the most common microorganisms and diagnostic lesions in the echocardiogram of patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

Disease progression

The course of the disease was similar in the 2 groups (table 3). Half of the TAVI group experienced cardiac complications, as did one-third of the SAVR group, without any significant differences being reported. Half of the patients had a surgical indication due to IE (40% in the TAVI group vs 50% in the SAVR group; P = .49), but among those, only 20% of the TAVI group and 25% of the SAVR group eventually underwent surgery (P = .93) (table 3). The in- hospital mortality rate was similar (20% in the TAVI group vs 25% in the SAVR group; P = .51), without any significant differences being reported in the 1-year mortality rate, which remained high—at approximately 30%—in both groups (table 3, figure 1, and figure 3).

Table 3. Complications, rate of surgical procedures, and 1-year mortality rate in patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

Figure 1. Infective endocarditis (IE) after transcatheter aortic valve implantation in a cohort of 778 patients vs a cohort of patients with surgical bioprostheses-related IE. SAVR, surgical aortic valve replacement; TAVI, transcatheter aortic valve implantation; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography.

Figure 2. Transesophageal echocardiogram showing a large vegetation on the leaflets of a transcatheter aortic valve (arrows).

Figure 3. Kaplan-Meier curves showing the 1-year mortality rate in patients with infective endocarditis after transcatheter aortic valve implantation (TAVI) or surgical aortic valve replacement (SAVR).

DISCUSSION

The main finding of our study was that the incidence rate of IE on TAVI is low in our setting and similar to that of surgical aortic bioprostheses, despite being patients with higher surgical risk, older age, and more comorbidities. These results are similar to those reported in the literature (1% up to 6%);4 however, more recent large TAVI trials suggest lower rates. The PARTNER 3 trial reported annual rates of 0.2%,1 and similarly, the low-risk Evolut study2 found incidence rates of 0.1% and 0.2% at 30 days and 1 year, respectivelu, which is more consistent with our rates. Studies directly comparing the incidence of IE after SAVR and TAVI are scarce, and some have produced contradictory results.7,12 However, most observational studies based on large national databases and randomized clinical trials have found no statistically significant differences on this regard,6,13,14 even though TAVI patients are older and generally have more comorbidities, which is consistent with our results. There were also no differences in our study regarding early IE incidence (50%), which was similar in both groups, although slightly lower for IE on TAVI vs what has been reported by the literature (rates up to 64%.)15 However, a multicenter study conducted in Spain found a higher rate of early IE in TAVI vs SAVR patients (78.1% vs 39.3%; P = .001), which is significantly higher than that described in our series for the TAVI group.16 This could be explained by the different antibiotic prophylaxis regimens and procedures used at each center.

Secondly, another relevant finding from our study is that enterococci were the main cause of IE in both patient groups, which is consistent with the literature on TAVI-related IE17,18 but not on surgical bioprostheses-related IE, of which S. aureus is usually the main microorganism involved. These differences are difficult to explain, as the increase in enterococcal incidence as a causal agent in TAVI is primarily associated with older patient age and transfemoral access but still would not explain why it is also the most common pathogen in SAVR-related IE.4 Compared with a multicenter Spanish series of similar characteristics to ours,16 there are some differences: the most common causative microorganism of IE—in both TAVI and SAVR— was Staphylococcus epidermidis, followed by enterococci— less frequent than in our series—and thirdly, S. aureus. These differences could be explained by varying antibiotic prophylaxis regimens used in different settings, underscoring the critical importance of understanding the most common microorganisms involved in prosthetic IE, in general, to apply the most appropriate and effective prophylaxis regimen.

Another noteworthy aspect of this study is that echocardiographic lesions diagnostic of IE could be identified in 100% of TAVI- related IE cases vs 67% of SAVR-related IE cases in surgical pros­theses. This contrasts sharply with most reports, which indicate that the combined sensitivity rate of transthoracic and transesophageal echocardiography was 67.8% in TAVI patients, 73% in SAVR patients, and nearly 90% in native valves.19 However, a very recent study comparing patients with IE after TAVI or SAVR confirmed that vegetations were identified via echocardiogram in up to 82% of the TAVI group, more in line with our findings and significantly higher than the diagnostic rate of IE reported in surgical prostheses (62.5%; P < .001).8 Our results also differ from those reported in the literature in that most lesions found via echocardiography were vegetations, whereas other authors, such as Salaun et al.,20 reported vegetations in only 5 out of 11 cases diagnosed with TAVI-related IE, with the remaining cases being atypical lesions.

Lastly, this study also highlights the course of the disease, with a 1-year mortality rate of 30%, which is lower than that reported in other published studies (between 33% and 66%).4,12,21-23 The in-hospital mortality rate was also lower than that reported in other studies, such as the international multicenter registry of post-TAVI endocarditis21 (36% vs 20% in our series) and much lower than the Spanish multicenter trial (35%).16 These differences may be due to the wide variability in patient characteristics across studies. Of note, there were no significant differences in mortality when comparing surgical bioprostheses-related IE, unlike other series reporting lower mortality for the latter vs TAVI-related IE.23 However, a study published by Panagides et al.,8 comparing early and 1-year mortality in IE after TAVI and SAVR using propensity score matching found no significant differences between the 2. In our series, only 20% of TAVI patients with a surgical indication underwent surgery, which is consistent with other studies where surgical rates were similar to ours, with optimal medical therapy being the most common strategy.4,12,16,22,23 Some studies found no improvement in prognosis for these patients undergoing surgery, with similar mortality rates associated with optimal medical therapy,24 while a meta-analysis published by Tinica et al.25 showed that the surgical strategy was significantly superior to conservative therapy. In any case, given the lack of strong evidence, the treatment strategy for TAVI-related IE remains unclear, even if the presence of complications warrants surgical indication, and decisions should be based on local expertise.

Study limitations

Our study has the inherent limitations of its observational design, with data collected over many years, during which diagnostic criteria for IE, valve types—towards better technical models—and procedural approaches—towards less invasive methods—have evolved. The small number of IE cases in the 2 groups means results should be interpreted with caution. As this is a crude analysis, the presence of confounding bias cannot be ruled out; nevertheless, data show real-life outcomes. Additionally, patients with mechanical valves were not included, so results cannot be extrapolated to this group.

CONCLUSIONS

In our setting, TAVI-related IE has a low overall incidence rate, with no significant differences vs SAVR-related IE, despite involving older and more comorbid patients. Regarding the causative microorganism of IE—both in surgical and percutaneous surgical bioprostheses—enterococci were the most common pathogen. There were no differences in mortality between the 2 types of aortic valves, and treatment for IE was predominantly conservative.

FUNDING

None declared.

ETHICAL CONSIDERATIONS

The study was conducted in full compliance with the Declaration of Helsinki and approved by the Clinical Research Ethics Committee at the beginning of the registry in 1987. However, for patients with TAVI or SAVR, all participants signed an informed consent form authorizing the collection and analysis of their data for research purposes. Since this was an observational study, patient treatment was not impacted. For this new registry analysis, approval was obtained from our center ethics committee. Data were handled completely anonymously in full compliance with the provisions of Organic Law 3/2018 of the Spanish Data Protection Authority.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence was used in the development of this work.

AUTHORS’ CONTRIBUTIONS

A. Roldán participated in data collection and analysis. C. Urbano participated in data collection and analysis. N. Aguayo participated in patient identification, data collection, and analysis. M. Crespín, J. López, and J.C. Castillo participated in the conception of the article and its interpretation. R. González contributed to statistical analysis and result interpretation. D. Mesa and M. Ruiz contributed to data processing, analysis and result interpretation, and collaborated in the revision and preparation of the manuscript for publication. J. Perea, I. Gallo, J. Suárez de Lezo, and S. Ojeda participated in the conception of the work, helped gather patient information, and provided guidance on literature review and manuscript drafting. M. Pan and M. Anguita supervised all stages of manuscript drafting, from conception, data collection, and result interpretation, to revision, correction, and preparation of the article for submission.

CONFLICTS OF INTEREST

S. Ojeda is an associate editor of REC: Interventional Cardiology. The journal’s editorial procedure to ensure impartial handling of the manuscript has been followed. The remaining authors declared no conflicts of interest whatsoever.

REFERENCES