Infective endocarditis after transcatheter aortic valve implantation: risk factors and mortality impact in a case-control study

Abstract

Aims

Infective endocarditis following transcatheter aortic valve implantation (TAVI-IE) is an uncommon but clinically devastating complication. We aimed to identify risk factors for TAVI-IE and to estimate its association with all-cause mortality.

Methods and results

We conducted a case-control study including patients who underwent TAVI at Haukeland University Hospital, Norway, between 2012 and 2023. Patients who developed TAVI-IE (n=71) were compared with age-matched and sex-matched controls without IE (n=213; 1:3 ratio). Death was treated as a competing event in analyses of IE, and we estimated the subdistribution HRs (SHR) for IE using Fine-Gray competing risk regression. Cox regression models with IE as a time-dependent covariate assessed the impact of infection on mortality.

The incidence of TAVI-IE was 1.2% per patient-year with a median time from TAVI to infection of 13 months (IQR 4–29). In multivariable competing risk analysis, diabetes mellitus remained an independent predictor of TAVI-IE (SHR 2.08, 95% CI 1.19 to 3.65, p=0.010). Obesity (27% vs 15%, p=0.019) and balloon-expandable valve use (28% vs 13%, p=0.003) were more often observed in patients with TAVI-IE. Enterococcus faecalis was the most frequent pathogen (30%). TAVI-IE was associated with an approximately twofold increase in all-cause mortality (adjusted HR 2.13, 95% CI 1.48 to 3.07, p<0.001) with the highest risk in early infections.

Conclusion

TAVI-IE is an infrequent but severe complication associated with excess mortality. Diabetes mellitus was the dominant independent risk factor and E. faecalis the leading pathogen. These findings may help target monitoring and prevention in patients at the highest risk.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• In this case-control study using competing-risk analysis, diabetes mellitus was identified as an independent predictor of TAVI-IE. Furthermore, TAVI-IE was associated with an approximately twofold increase in all-cause mortality, even after multivariable adjustment.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings suggest that metabolic risk assessment should be integrated into TAVI programmes. Patients with diabetes mellitus may warrant closer surveillance, particularly during the first post-implantation year, when mortality risk associated with TAVI-IE is the greatest.

Introduction

Severe symptomatic aortic stenosis is the main indication for surgical intervention on heart valves.^{1 2} Since its introduction in 2002, transcatheter aortic valve implantation (TAVI) has become a well-established alternative to surgical aortic valve replacement (SAVR). Current European guidelines recommend TAVI above SAVR in patients 70 years or older, those at high surgical risk or in younger patients unsuitable for open-heart surgery.² TAVI has thus enabled aortic valve replacement in older, frailer and more multimorbid patients, previously considered inoperable.

Several studies have demonstrated comparable outcomes after TAVI and SAVR, including similar mortality rates.³,⁵ The reported incidence of infective endocarditis following TAVI (TAVI-IE) ranges between 0.3% and 2.0% per patient-year,⁶,⁹ with mortality rates as high as 30–60%.⁸,¹⁰

While several registry-based studies have examined TAVI-IE,^{11 12} detailed case-level analyses incorporating competing-risk methodology remain limited,^{13 14} particularly as TAVI expands to lower-risk populations across Europe. Given the high background mortality in the TAVI population, traditional Cox models may overestimate infection risk. Our study therefore provides a detailed risk factor assessment using competing-risk analyses within a well-defined regional TAVI cohort. As TAVI technology has advanced and indications have expanded to include younger and lower-risk patients, understanding and reducing the risk for TAVI-IE has become increasingly important.

In our recent investigation on the epidemiology of IE in Western Norway, TAVI-IE accounted for 10% of all cases. Both 30-day (8%) mortality and 90-day (14%) mortality were low in the group with prosthetic valve endocarditis including TAVI-IE.¹⁵

In the present study, we compared patients who developed TAVI-IE with age-matched and sex-matched controls who underwent TAVI without developing subsequent IE. The primary objective was to identify preventable risk factors for TAVI-IE in an already old and high-risk patient population. Secondary objectives included assessment of bacterial pathogens and the impact of IE on mortality.

Methods

Study setting and population

Haukeland University Hospital (HUH) is a tertiary care centre in Bergen, Norway, and has the regional responsibility for all TAVI procedures performed in Western Norway, serving a population of approximately 1.1 million inhabitants.¹⁶ Between 2012 and 2023, 1576 TAVI procedures were performed at HUH.^{17 18} During this period, 71 patients were treated for TAVI-IE at either HUH or other secondary or primary hospitals in the region. All included cases were 18 years or older. The main inclusion criterion was treatment for TAVI-IE documented in the medical records. Cases were identified through International Classification of Diseases V.10 codes (I33.0, I38 and I39.0 with subgroups) or prospectively included in a local IE registry. All cases were verified through review of medical records and linkage to the mandatory Norwegian Registry of Invasive Cardiology through the Transcatheter Aortic Valve Implantation in Western Norway (TAVI-NOR) study, ensuring complete capture of all TAVI procedures in the region.^{17 18} Cases were classified as definite or possible IE according to the modified Duke criteria and/or the 2015 European Society of Cardiology guidelines and/or the 2023 Duke-International Society of Cardio-Vascular Infectious Diseases criteria, as applicable.¹⁹,²¹ Patients fulfilling criteria for definite or possible IE received similar treatment in accordance with the national guidelines for antimicrobial therapy of endocarditis in Norway.²² Antibiotic treatment and prophylactic practices followed the national guidelines throughout the study period.²² Sample collection procedures remained unchanged. For patients with recurrent IE, only the first episode was included.

The 71 TAVI-IE patients were included as cases. For each case, three controls were randomly selected from the source population of patients who underwent TAVI but who did not develop IE during the observation period. Matching was performed in a 1:3 ratio based on age (±2 years), sex and date of TAVI (±2 years) using incidence density sampling. No additional variables were used for matching, as differences in comorbidities, procedural characteristics and clinical profiles between cases and controls were themselves study objectives. Baseline characteristics, echocardiographic findings, microbiological results and outcomes were compared between the groups. As this was an observational study, blinding was not applicable. Interobserver and intraobserver variability were not formally assessed, as analyses were based on routinely collected clinical data.

The index date of the study was the date of TAVI (figure 1). The observation period was from 2012 to 2023, with censoring done on 31 December 2024.

Figure 1. Flowchart for recruitment of transcatheter aortic valve implantation associated infective endocarditis (TAVI-IE) cases and controls. Flowchart illustrating the selection of the study population. A total of 1576 patients underwent TAVI at Haukeland University Hospital, Bergen, between 2012 and 2023. Among these, 71 patients developed TAVI-IE.

Microbial isolates

All microbial isolates were obtained from blood cultures and identified using matrix-assisted laser desorption/ionisation-time of flight mass spectrometry. Excised valves (n=3) were investigated using broad-range amplification of the bacterial 16S rRNA directly from the sample DNA, followed by Sanger sequencing (direct 16S rRNA sequencing) in line with standard laboratory procedures.

Statistics

Data were analysed using Stata (V.18; StataCorp, College Station, Texas, USA). Continuous variables are reported as mean±SD and categorical variables as proportions. Groups were compared by Student’s t-test for continuous variables and χ² tests for categorical variables.

Death was treated as a competing event in analyses of IE. Competing risk regression according to Fine and Gray was applied to estimate subdistribution HRs (SHR) with 95% CIs, and analyses were performed in Stata using stcrreg. Covariates for the adjusted model (age, sex, diabetes mellitus (DM), obesity, cardiovascular disease, atrial fibrillation, femoral access and valve type) were selected a priori based on clinical relevance and previous literature. Results are presented as SHR with 95% CIs.

Time-to-event analyses were conducted using Kaplan-Meier estimates and Simon-Makuch plots to account for the time-dependent onset of IE. Survival between groups was compared using the Mantel-Byar test. Patients were censored at death or end of follow-up, with follow-up time before IE diagnosis contributing to the no-infection group and time after diagnosis to the IE group.

Survival was analysed by using univariable and multivariable Cox proportional hazards regression. We first fit a Cox proportional hazards model with TAVI-IE modelled as a time-dependent exposure since follow-up time runs from TAVI procedure (0 before IE; one from the time of IE onward), and then used an adjusted model that included prespecified covariates (age, sex, left-ventricular ejection fraction <50%, atrial fibrillation, DM, obesity and cardiovascular disease) identified a priori based on clinical relevance and previous findings on risk factors for TAVI-IE. To assess the effect of early infection, we repeated the time-dependent Cox analysis restricted to the first 12 months after TAVI implantation. Finally, to mitigate bias from unequal follow-up, we additionally performed a 12-month landmark Cox analysis (including only patients alive at 12 months) and a truncated follow-up analysis in which controls were administratively censored at the IE median follow-up (57 months); events occurring exactly at the cut were treated as censoring.

Results are presented as HRs with 95% CIs. A two-sided p value ≤0.05 was considered statistically significant.

Results

We first describe the cases with TAVI-IE, followed by a comparison of clinical characteristics of cases and controls at baseline. We continue to present analyses of risk factors for IE and conclude with analyses on mortality.

TAVI-IE cases

To characterise the clinical presentation and microbiology of IE after TAVI, we examined detailed features of the 71 TAVI-IE cases (table 1). Of the 71 cases who were treated for TAVI-IE, 46 (65%) were definite IE, whereas the remaining 25 (35%) were possible IE according to established diagnostic criteria prior to treatment. The mean age at TAVI-IE was 82 (±6) and the median time from the TAVI procedure to IE diagnosis was 13 months (IQR 4–29 months).

Table 1. Clinical and microbiological characteristics of patients with TAVI-associated infective endocarditis.

Microbiology	All TAVI-IE n (%)	Very early ≤1 month	Early 1–≤12 months	Late >12 months
Total cases	71 (100)	8 (11)	22 (31)	41 (58)
Enterococci*	24 (34)	1 (1)	10 (14)	13 (18)
Viridans group streptococci†	18 (25)	2 (3)	4 (6)	12 (17)
Staphylococcus aureus	14 (20)	1 (1)	6 (8)	7 (10)
Other organisms‡	15 (21)	4 (6)	2 (3)	9 (13)

Distribution of causative micro-organisms in 71 patients with TAVI-IE, stratified by time from implantation: very early (≤1 month), early (1–≤ 12 months) and late (>12 months). All percentages by the total (71). Categorical variables are expressed as number (percentage).

^*Enterococcus faecalis (n=21), E. faecium (n=3).

^†Streptococcus mitis (n=8), Str. gordonii (n=2), Str. sanguinis (n=4), Str. salivarius (n=2), Str. pasteuralis (n=1), Str. cristatus (n=1).

^‡Str. dysgalacticae (n=3), Sta. epidermidis (n=2), Sta. lugdunensis (n=2), Aggregatibacter spp (n=1), Bacillus cereus (n=1), Cutibacterium acnes (n=1), Aerococcus urinae (n=1), Coxiella burnetii (n=1), Gemella hemolysans (n=1), Candida albicans (n=1), no growth (n=1).

IE, infective endocarditis; TAVI, transcatheter aortic valve implantation.

The infections were classified by time of onset as very early (≤1 month, n=8, 11%), early (1–≤12 months, n=22, 31%) and late (>12 months, n=41, 58%) with late infections being the most common. In six patients (8%), the infection also affected a pacemaker or other cardiac implantable electronic devices. No cases of double-valve IE with simultaneous mitral or tricuspid valve involvement were observed. Three patients (4%) were operated due to heart failure.

Enterococcus species were the most frequently pathogens, identified in 24 (34%) cases, with Enterococcus faecalis accounting for 21 cases (30%).

Clinical characteristics for cases and controls at baseline

The mean age at TAVI implantation was 80±6 years with males representing 84% of both cases and controls (table 2). The annual incidence of TAVI-IE ranged from 8.0 to 22.7 per 1000 patient-years, with the highest rate in 2016 and a declining trend over time. The overall incidence of TAVI-IE was 1.2% per patient-year with a mean follow-up time of 4.5±2.4 years in the IE group and 5.8±2.3 years in the control group (p<0.001). DM was present in 22 (31%) IE cases and in 39 (18%) controls (p=0.024). Obesity (body mass index ≥30 kg/m²) was documented in 19 (27%) IE cases and 31 (15%) controls (p=0.019). Balloon-expandable valves were used in 20 (28%) cases and 28 (13%) controls (p=0.003).

Table 2. Baseline characteristics at the time of TAVI procedure of patients with and without TAVI-associated infective endocarditis.

Variable	TAVI-IE n=71	Controls n=213	P value
General factors
Female sex	11 (16)	33 (16)	1.000
Age (years)	80±6	81±7	0.958
Follow-up (years)	4.5±2.4	5.8±2.3	< 0.001
Total mortality	42 (59)	135 (63)	0.071
Early mortality (≤1 year)	5 (7)	3 (1)	0.019
Late mortality (>1 year)	37 (52)	132 (62)	0.019
Non-cardiovascular factors
Diabetes mellitus	22 (31)	39 (18)	0.024
Smoking history	43 (61)	113 (53)	0.271
COPD	12 (17)	49 (23)	0.278
Chronic kidney disease	23 (32)	69 (32)	1.000
BMI (kg/m²)	27±4	26±4	0.183
Obesity (BMI ≥30 kg/m²)	19 (27)	31 (15)	0.019
Underweight (BMI <20 kg/m²)	2 (3)	13 (6)	0.284
Cardiovascular factors
Hypertension	54 (76)	146 (69)	0.230
Cardiovascular disease*	48 (68)	174 (82)	0.013
Atrial fibrillation	32 (45)	85 (40)	0.444
Heart rate (beats per min)	72±14	71±13	0.719
Mean arterial pressure (mm Hg)	97±11	97±10	0.907
Procedure-related factors
Femoral access	67 (94)	184 (86)	0.069
Postoperative pacemaker	27 (38)	79 (37)	0.887
Self-expanding valve	42 (59)	148 (70)	0.109
Mechanically expanding valve	9 (13)	37 (17)	0.352
Balloon-expandable valve	20 (28)	28 (13)	0.003
Transthoracic echocardiography
AS Vmax (m/s)	4.24 (0.54)	4.34 (0.56)	0.184
AS effective orifice area (cm2)	0.79±0.21	0.73±0.20	0.065
Ejection fraction (%)	56.3±10.5	55.2±11.7	0.519
Ejection fraction <50%	15 (21)	52 (24)	0.572

Comparison of demographic, clinical and procedural variables between 71 patients with TAVI-IE and 213 age-matched and sex-matched TAVI controls without infection. Continuous variables are expressed as mean±SD; categorical variables as number (percentage). Group comparisons were performed using Student’s t-test or χ2/Fisher’s exact test as appropriate.

^*Cardiovascular disease includes coronary artery disease and/or stroke.

AS, aortic stenosis; AS Vmax, aortic stenosis maximum Doppler derived blood flow velocity; BMI, body mass index; COPD, chronic obstructive pulmonary disease; TAVI-IE, transcatheter aortic valve implantation associated infective endocarditis.

Risk factors for TAVI-IE

The high mortality in the TAVI population, largely driven by advanced age, represents a major competing event that may preclude the occurrence of IE. To account for this, we applied a Fine-Gray subdistribution hazard model, treating death as a competing event. This approach provides a more accurate estimate of infection risk in the presence of substantial mortality in both cases and controls. The results are shown in table 3, complementing the descriptive comparisons in table 2, and providing a refined assessment of independent risk factors for TAVI-IE.

Table 3. Univariable and multivariable Fine-Gray competing risk regression for development of TAVI-associated infective endocarditis.

	Unadjusted			Adjusted
Variables	SHR	95% CI	P value	SHR	95% CI	P value
Age, per year	0.98	0.95 to 1.01	0.173	0.99	0.95 to 1.03	0.627
Male sex	0.98	0.52 to 1.88	0.961	1.06	0.45 to 2.53	0.883
Diabetes mellitus	1.76	1.06 to 2.92	0.028	2.08	1.19 to 3.65	0.010
Obesity (BMI ≥30 kg/m²)	1.92	1.14 to 3.23	0.014	1.46	0.75 to 2.84	0.267
Cardiovascular disease*	0.77	0.43 to 1.40	0.399	0.75	0.38 to 1.48	0.408
Atrial fibrillation	1.25	0.79 to 2.00	0.339	1.28	0.73 to 2.24	0.391
Femoral access	2.43	0.88 to 6.73	0.086	2.86	0.86 to 9.49	0.086
Balloon-expandable valve	2.58	1.54 to 4.31	<0.001	1.62	0.78 to 3.35	0.195

Results of Fine-Gray subdistribution hazard models estimating the risk of TAVI-IE with death treated as a competing event. The unadjusted model includes single predictors, while the multivariable model adjusts for relevant baseline covariates (age, sex, diabetes mellitus, obesity, cardiovascular disease, atrial fibrillation, femoral access and valve type). SHR are presented with 95% CI and p values.

Event of interest: infective endocarditis (n=71); competing event: death from other causes (n=177).

^*Cardiovascular disease includes coronary artery disease and/or stroke.

BMI, body mass index; IE, infective endocarditis; SHR, subdistribution HR; TAVI, transcatheter aortic valve implantation.

In the competing risk regression analysis, balloon-expandable valves (SHR 2.58, 95% CI 1.54 to 4.31, p<0.001), obesity (SHR 1.92, 95% CI 1.14 to 3.23, p=0.014) and DM (SHR 1.76, 95% CI 1.06 to 2.92, p=0.028) were associated with a higher subdistribution hazard of TAVI-IE in univariable analysis (table 3). Following multivariable adjustment, DM remained an independent predictor for TAVI-IE (SHR 2.08, 95% CI 1.19 to 3.65, p=0.010).

To facilitate comprehension of the competing-risk analysis, we additionally performed a standard Cox regression analysis for the risk of developing IE. This analysis also identified DM as the main risk factor for IE in the adjusted model (HR 2.16, 95% CI 1.24 to 3.74, p=0.006) consistent with the findings from the competing-risk model (online supplemental table 1).

Online supplemental table 1 shows the Cox regression analysis for the risk of developing IE.

Outcomes

We estimated the absolute time-dependent risk of death following TAVI-IE, with cumulative mortality of 13% at 30 days, 17% at 90 days, 30% at 1 year and 59% during total follow-up.

We further assessed the impact of IE on all-cause mortality following TAVI. Survival time was calculated from the index TAVI procedure until death or end of follow-up with patients alive at study completion censored. To account for the time-dependent occurrence of IE, follow-up time prior to diagnosis contributed to the no-infection group, whereas time after diagnosis contributed to the TAVI-IE group.

Cumulative survival probabilities were illustrated using Simon-Makuch plots (online supplemental figure 1).

To further quantify the impact of IE on mortality risk beyond descriptive survival analyses, we performed Cox proportional hazards regression with IE modelled as a time-dependent covariate (table 4). This approach accounts for the fact that IE occurs during follow-up rather than at baseline. Infective endocarditis was associated with an approximately twofold higher risk of death in both the unadjusted model (HR 1.98, 95% CI 1.40 to 2.81, p<0.001) and the adjusted model (HR 2.13, 95% CI 1.48 to 3.07, p<0.001).

Table 4. Cox-regression analysis of mortality with IE as a time-dependent covariate.

	Unadjusted			Adjusted
Variables	HR	95% CI	P value	HR	95% CI	P value
Infective endocarditis	1.98	1.40 to 2.81	<0.001	2.13	1.48 to 3.07	<0.001
Age, per year	1.05	1.02 to 1.07	0.001	1.06	1.03 to 1.09	<0.001
Male sex	1.33	0.87 to 2.05	0.193	1.18	0.74 to 1.88	0.493
LVEF <50%	2.21	1.60 to 3.05	<0.001	1.98	1.41 to 2.78	<0.001
Atrial fibrillation	1.73	1.28 to 2.34	<0.001	1.45	1.05 to 1.99	0.023
Diabetes mellitus	1.22	0.86 to 1.72	0.265	1.20	0.83 to 1.72	0.329
Obesity (BMI ≥30 kg/m²)	1.02	0.69 to 1.49	0.936	1.16	0.77 to 1.76	0.484
Cardiovascular disease	1.11	0.76 to 1.63	0.583	1.18	0.78 to 1.77	0.441

Cox proportional hazards models estimating the effect of TAVI associated IE and clinical covariates on all-cause mortality during follow-up. IE was included as a time-dependent covariate to account for its occurrence after TAVI. Results are presented as HRs with 95% CI and p values for both unadjusted and adjusted models (adjusted for age, sex, LVEF <50%, atrial fibrillation, diabetes mellitus, obesity and cardiovascular disease).

BMI, body mass index; IE, infective endocarditis; LVEF, left-ventricular ejection fraction; TAVI, transcatheter aortic valve implantation.

Finally, to address potential bias related to unequal follow-up time and the time-dependent occurrence of IE, we performed additional sensitivity analyses (online supplemental table 2). Within the first 12 months post TAVI, IE was associated with markedly increased mortality risk (HR 30.56, 95% CI 4.40 to 212.11, p=0.001). In a 12‑month landmark analysis restricted to patients alive at 12 months, IE remained associated with subsequent mortality (HR 1.73, 95% CI 1.13 to 2.63, p=0.011). After administratively censoring controls at the median follow‑up time of IE cases (57 months), the association persisted and strengthened (HR 2.54, 95% CI 1.54 to 4.19, p<0.001).

Discussion

In this single-centre case-control study from Western Norway, IE after TAVI was associated with an approximately twofold increase in all-cause mortality. DM was the most important risk factor for TAVI-IE and remained an independent predictor after adjustment for relevant clinical variables. These findings add to the limited evidence on predictors of TAVI-IE in contemporary cohorts of older patients undergoing TAVI, a population characterised by both high background mortality and infection risk.

DM and obesity were significantly more frequent among patients who developed TAVI-IE, underscoring the role of metabolic and inflammatory factors in infection susceptibility. Previous studies have consistently identified DM as an independent risk factor for native and prosthetic valve endocarditis, particularly in older, male patients with multiple comorbidities such as hypertension, chronic kidney disease and coronary artery disease.²³,²⁶ Hyperglycaemia, impaired immune response and chronic low-grade inflammation predispose individuals with DM to infection.²⁷ As our groups were matched for age and sex, the observed association between DM and TAVI-IE appears independent of these major demographic factors. Obesity, although less consistently reported as a risk factor for TAVI-IE, is well known to increase infection risk through mechanisms including chronic inflammation, greater procedural complexity and delayed wound healing.^{610 28},³¹

Considering our findings, integrating metabolic evaluation, including evaluation of glycaemic control, into the pre-TAVI work-up may help identify patients at highest risk for TAVI-IE. As such, optimisation of DM management may represent an important component of preventive strategies both before and after TAVI.

Balloon-expandable valves were associated with a higher tendency of infection in the univariable competing risk analysis, although this association was attenuated after multivariable adjustment. Previous registry studies have presented conflicting data regarding a possible association between valve type and an increased risk of endocarditis. While some studies report higher infection rates with self-expanding prostheses,^{6 10} others find no significant difference after adjustment.^{4 8} The mechanisms underlying a potential vulnerability for IE after balloon-expandable valve implantation remain uncertain but hypothetically involve valve design, implantation characteristics, greater surface area or differences in endothelialisation and flow turbulence. Although our numbers are limited, the consistent association warrants further investigation in larger cohorts. Larger registry studies are needed to determine whether the observed correlation reflects a true device effect or underlying procedural and/or patient selection factors.

Pacemaker implantation was frequent in both groups but was not associated with IE in our study. This contrasts with findings from previous large registry studies.^{26 32}

In our cohort, E. faecalis was the most common causative pathogen, followed by Staphylococcus aureus. This contrasts with several large international registries in which S. aureus and coagulase-negative Staphylococci were the most prevalent, with the former being particularly frequent in early-onset TAVI-IE.^{10 12 33} Nevertheless, other studies have also identified Enterococcus species as the leading pathogens, accounting for approximately 20–25% of cases.^{26 32} In our cohort, the proportion was substantially higher, with E. faecalis alone responsible for 30% of all TAVI-IE cases. This distribution may reflect regional microbiological patterns and/or specific population characteristics as 84% of the TAVI-IE cases are males.

Several studies have suggested that transfemoral procedures may be associated with an increased risk of enterococcal infection, as these interventions involve vascular manipulation through skin areas potentially colonised by gastrointestinal and genitourinary bacterial flora.^{21 34} In our cohort, however, we observed only one very early infection with E. faecalis and found no significant differences in microbial distribution across very early, early and late infections. Given that the median time from TAVI to infection was approximately 1 year, endogenous rather than periprocedural factors are likely to play a more prominent role in the pathogenesis of TAVI-IE. Malignancy has also been proposed as a risk factor for enterococcal IE.³¹,³³ Although neither this study nor data from our previous study¹⁵ are directly supporting this association, we feel that occult malignancy should still be actively considered as a potential source of bacteraemia in older adults.

Our IE cohort showed a pronounced male predominance with only 16% female patients, although 39% of all TAVI procedures performed nationally in the recent years are on females.¹⁸ This finding suggests that TAVI-IE may be even more male-predominant than previously reported.^{12 35} The over-representation of males among TAVI-IE patients, relative to both the general TAVI population and our overall IE cohort,^{15 36} may reflect sex-related differences in cardiovascular risk profiles, valve pathology or microbial colonisation patterns, particularly with regard to E. faecalis. Further research is warranted to explore this sex-linked difference in susceptibility to TAVI-IE.

The concept of futility has become increasingly relevant in interventional cardiology, particularly in older and multimorbid patients considered for TAVI.³⁷ A procedure is generally regarded as futile when it fails to improve survival or quality of life despite technical success. In our cohort, only 7% of TAVI-IE cases died within the first year after TAVI, with a median time from TAVI to IE of 13 months. However, given a mean age of 82 years at IE diagnosis and a subsequent median survival of 24 months, TAVI-IE represents a dramatic, although rare, complication. Management typically requires prolonged hospitalisation and approximately 6 weeks of intravenous antimicrobial therapy.^{21 22 38} As such, TAVI-IE should be considered a major adverse event with substantial implications for patients’ quality of life, particularly in the later stages of life.

The incidence of TAVI-IE was only 1.2% per patient-year. By comparison, the incidence of IE following SAVR was approximately 4.3% per patient-year in our recent study.¹⁵ Within this context, the incidence observed after TAVI appears clinically acceptable when balanced against the survival benefits of TAVI in older, high-risk patients who otherwise face poor outcomes from severe aortic stenosis.

This study has several limitations. Although the case-control design is a strength, it allows only for the identification of associations rather than causal relationships. The single-centre design may limit generalisability of our findings. However, HUH serves as the sole tertiary referral centre in our region, thereby reducing referral bias and enhancing population representativeness. Data on glycaemic control (eg, Hemoglobin A1C test) were not systematically available, precluding analysis of whether infection risk is driven by disease severity or metabolic control. The proportion of definite IE was lower compared with some previous reports, which may reflect less extensive diagnostic work-up in older and frailer patients, in whom advanced imaging or invasive investigations are not always performed. As a result, some cases may have been classified as possible rather than definite IE. However, treatment decisions were similar for both groups, and analyses restricted to definite IE yielded comparable results. Despite the relatively low number of IE cases, the incidence observed aligns with previously reported rates, and the competing-risk approach strengthens statistical robustness in the context of substantial mortality.

Conclusion

In conclusion, TAVI-IE is an uncommon but severe complication associated with substantially increased mortality. DM emerged as the dominant independent predictor of TAVI-IE in competing risk analysis. E. faecalis was the most frequently identified pathogen and may contribute to the observed male predominance of TAVI-IE in our cohort.

These findings support the integration of metabolic risk assessment into TAVI programmes and suggest that patients with DM may benefit from targeted surveillance, particularly during the first post-implantation year, when the mortality risk associated with TAVI-IE is the highest.

Data availability statement

Data are available upon reasonable request.