Highlights
-
•
Dialysis patients were younger but had more comorbidities and advanced disease.
-
•
Factors linked to dialysis included surgery, CKD, diabetes, liver disease, and infection with Staphylococcus aureus and Enterococcus spp.
-
•
Dialysis marked a high-risk group with in-hospital mortality near 40%, corresponding to an almost 2 fold increased adjusted hazard compared with patients with no dialysis.
-
•
Among patients surviving the initial admission, initiation of dialysis was also linked to a higher risk of mortality during the early post-discharge period; however, this elevated risk diminished after 3 months.
ABSTRACT
Background
Acute kidney injury (AKI) requiring dialysis is a serious complication of infective endocarditis (IE), yet detailed data on risk factors and outcomes are limited.
Methods
Using the nationwide NIDUS registry, we identified all dialysis-naive patients hospitalized with first-time left-sided IE in Denmark from 2016 to 2021. Patients were grouped by whether they developed dialysis-requiring AKI. Factors associated with dialysis were assessed with multivariable logistic regression. In-hospital and post-discharge mortality were evaluated using Kaplan–Meier methods and multivariable Cox models.
Results
Among 2,738 patients with left-sided IE, 203 (7%) received dialysis for AKI (70% male; median age 70 years) and 2,535 (93%) did not (66% male; median age 75 years). Those requiring dialysis had greater comorbidity, and more frequently underwent valvular surgery (57.0% vs 19.8%, P < .01), with two-thirds initiating dialysis postoperatively. Independent factors associated with initiation of dialysis included chronic kidney disease, diabetes, liver disease, sepsis, valvular surgery, and infection with Staphylococcus aureus or Enterococcus species. In-hospital mortality was substantially higher among patients requiring dialysis (37.9 vs 16.5%; HR 1.86, 95% CI 1.39-2.50). In-hospital mortality was higher with dialysis (37.9% vs 16.5%; HR 1.86, 95% CI 1.39-2.50). Among survivors, associated mortality was higher within 3 months (12.0% vs 9.0%; HR 2.35, 95% CI 1.30-4.27) but similar from 3 to 12 months (6.3% vs 11.3%; HR 0.89, 95% CI: 0.41-1.93).
Conclusion
Patients initiating dialysis during IE had an increased burden of traditional risk factors for AKI—with valve surgery representing the strongest associated risk factor for initiating dialysis. Patients initiating dialysis during IE had a markedly higher in-hospital mortality. Among survivors, higher associated mortality was confined to the early post-discharge period and diminished after 3 months.
Despite a changing epidemiology and advancement in treatment modalities, infective endocarditis (IE) remains a life-threatening disease characterized by severe complications and substantial mortality. , Acute kidney injury (AKI) of any degree is a frequent and serious complication, with reported incidences of up to 25%, and is associated with markedly worsened prognosis and a high risk of long-term deterioration in kidney function. ,,,,
In its most severe form, AKI may necessitate dialysis. Previous studies have reported dialysis-requiring AKI in 5%-14% of IE cases, ,,,, consistently conferring an increased risk of short- and long-term adverse outcomes. , Beyond higher mortality, the need for dialysis reflects severe systemic illness and is linked to prolonged hospitalization, greater rates of heart failure, and persistent renal impairment leading to chronic kidney disease (CKD). ,, While prior studies have analyzed large IE populations, they were limited either by low granularity of available clinical information or by data derived from highly selected tertiary-center cohorts, restricting generalizability. Extending current knowledge with detailed, unselected data may therefore improve risk stratification, guide preventive measures, and provide more comprehensive information to support patient counseling and management.
To address these gaps in knowledge, the present study aimed to investigate factors associated with dialysis-requiring AKI in patients with IE and to assess its impact on in-hospital and 1-year mortality. The associations identified may support further research into optimal strategies for early prevention and management of AKI in this vulnerable population.
Methods
Data sources
The study cohort was derived from the NatIonal Danish endocarditis stUdieS (NIDUS) database, a nationwide Danish registry of patients with IE, as described in detail elsewhere. The database includes consecutive cases classified as definite or possible IE based on the modified Duke/European Society of Cardiology (ESC) 2015 diagnostic criteria. Diagnoses were validated by a multidisciplinary team, and comprehensive clinical data were collected for 3,795 IE episodes and 3,557 individual IE patients diagnosed between January 1, 2016, and December 31, 2021.
Study population
We identified all patients with first-time left-sided IE between January 1st, 2016, and December 31st, 2021. Patients were categorized into those with and without first-time dialysis treatment during admission. Dialysis was defined as any duration of dialysis treatment during IE admission. Patients were excluded if they had right-sided IE, cardiac implantable electronic device–related infection, endocarditis not localized to the heart valves, or were receiving chronic dialysis at the time of admission.
Covariates
All covariates in NIDUS were recorded from electronic medical records by a team of health care professionals and endocarditis experts, as described previously. Baseline comorbidities were defined as known comorbidities or treatment for this condition at the time of IE admission. CKD was defined as any known kidney disease or insufficiency at admission resulting in a baseline eGFR < 60 mL/min. AKI was defined as a 50% ≥ increase in serum creatinine within 48 hours during admission. Sepsis was defined based on the sequential organ failure assessment (SOFA) score. Self-reliance was defined as not receiving any domiciliary care. First-line antibiotic treatment was defined as the antibiotic regimen administered on the first day of IE diagnosis, or, if antibiotic therapy at IE-specific doses was initiated prior to the day of diagnosis, the date of initiation of such treatment was used.
Outcomes and follow-up
The objective was to describe patients’ medical history, IE-specific disease characteristics, clinical management patterns, antibiotic treatment, and all-cause mortality. Antibiotic treatment was defined as the regimen administered on the first day of IE diagnosis or, if antibiotics at IE-specific doses were initiated before that date, the initiation date of such therapy was used. Mortality was assessed during hospitalization, and among hospital survivors, follow-up continued for 1 year after discharge. As the proportional hazards assumption was not met ( P =.04 for proportional hazards assumption), follow-up for mortality was further stratified into 2 periods: from 0 to 3 months after discharge and from 3 to 12 months after discharge. Additionally, we aimed to identify factors associated with the initiation of dialysis treatment during admission. This analysis was repeated after excluding those with CKD.
Handling of missing information
The total fraction of missing data was < 0.1%. Due to the low proportion, missing values were not reported separately in descriptive tables but were highlighted in tables when >5% and >10% were missing. Patients with missing data in 1 or more covariates were excluded from the multivariable Cox regression analyses by complete case analysis.
Statistics
Baseline characteristics were compared between study groups. Categorical variables were reported as frequencies and percentages, with differences calculated with chi-square test. Continuous variables were presented as medians with interquartile ranges (25th-75th percentiles), and differences were calculated with the Wilcoxon rank-sum test. Crude in-hospital all-cause mortality was calculated as the number of patients who died before discharge divided by the total number of patients in the study group and differences were assessed with chi-square test. Adjusted in-hospital mortality rates were estimated using a multivariable Cox proportional hazards model, controlling for age, sex, CKD, diabetes mellitus, liver disease, chronic obstructive pulmonary disease, sepsis at presentation or during admission, previous cancer diagnosis, active cancer, previous stroke, heart failure, prosthetic or native valve IE, valvular heart surgery, and bacterial etiology. Absolute rates of mortality following discharge were assessed based on Kaplan Meier estimates, while adjusted rates were computed with multivariable Cox proportional hazards model, controlling for the same covariates as in the model for in-hospital mortality. Interaction between sex and valve surgery during admission was assessed, with no significant interaction observed ( P =.54).
Factors associated with initiation of dialysis were examined using multivariable logistic regression examining the following covariates: sex, age, CKD, diabetes mellitus, liver disease, chronic obstructive pulmonary disease, sepsis at presentation or during admission, chronic heart failure, prosthetic or native valve IE, bacterial etiology, and valvular heart surgery. The same analysis was repeated after excluding patients with CKD before admission.
All statistical analyses were performed using the SAS enterprise statistical software (version 8.3, Cary, NC) and Rstudio (version 2025.09.1 + 401). Level of statistical significance will be recognized by a P -value <.05.
Supplementary analysis
To test the robustness of our data, 3 supplementary analyses were performed. First, we examined outcomes after stratifying the population into those with and without surgical treatment. Second, we examined the association between S. aureus and dialysis initiation, by comparing S. aureus with all other microbiological etiologies compiled. Third, antibiotic treatment was examined after excluding patients with known CKD at admission.
Results
Patients’ medical history
We identified 2,737 patients with first-time left-sided infective endocarditis (IE), of whom 509 (18.6%) developed AKI during admission. Among these 509 patients, 203 (40.0%) required first-time dialysis treatment (70.4% male), corresponding to 7.4% of the overall cohort, while 2,534 (92.6%) did not receive dialysis (66.1% male). The patient selection process is illustrated in Figure 1 . Baseline medical history is presented in Table 1 .
Population selection process.
This flow chart illustrates the population selection process.
Abbreviations: AKI, acute kidney injury; IE, infective endocarditis.
Table 1
Patients’ medical history before admission.
| No dialysis N = 2,534 | Dialysis N = 203 | P value | |
|---|---|---|---|
| Demography | |||
| Male sex, N (%) | 1,674 (66.1) | 143 (70.4) | .20 |
| Median age, years (25-75 percentiles) | 74.8 (66.0-81.6) | 70.1 (62.5-75.6) | <.01 |
| Employment status | |||
| Student | 13 (0.5) | 3 (1.5) | .15 |
| Employed | 390 (15.6) | 33 (16.6) | – |
| Unemployed | 122 (4.9) | 14 (7.0) | – |
| Retired | 1,982 (79.1) | 149 (74.9) | – |
| Medical history before admission, N (%) | |||
| Nephropathy | 274 (10.9) | 54 (26.6) | <.01 |
| Diabetes mellitus | 532 (21.1) | 63 (31.0) | <.01 |
| Liver disease | 118 (4.7) | 17 (8.4) | .02 |
| Chronic obstructive pulmonary disease | 372 (14.8) | 33 (16.3) | .78 |
| Prior cancer | 374 (14.8) | 25 (12.3) | .15 |
| Active cancer | 247 (9.8) | 10 (4.9) | .02 |
| Ischemic or hemorrhagic stroke | 369 (14.6) | 28 (13.8) | .88 |
| Chronic heart failure | 374 (14.8) | 32 (15.8) | .21 |
| Native heart valve disease | 351 (13.9) | 24 (11.9) | .44 |
| Congenital heart disease | 93 (3.7) | 11 (5.5) | .10 |
| Cardiac implantable electronic device | 325 (12.9) | 23 (11.3) | .78 |
| Previous valvular heart surgery | 688 (27.1) | 65 (32.0) | .14 |
| Coronary artery bypass graft | 325 (12.9) | 19 (9.4) | .53 |
| Alcohol abuse | 430 (17.0) | 39 (19.2) | .41 |
| Females (>7 units/wk) | 95 (16.1) | 12 (27.3) | – |
| Males (>14 units/wk) | 335 (24.2) | 27 (23.3) | – |
| Active smoker | 472 (18.8) | 44 (22.0) | .29 |
| Intravenous drug use | 41 (1.6) | 6 (3.0) | .15 |
| Self-reliant in activities of daily living | 1,762 (69.8) | 159 (78.3) | .02 |
| Pharmacotherapy before admission | |||
| Oral anticoagulants | 347 (25.6) | 36 (31.9) | .12 |
| Aspirin | 454 (33.5) | 31 (27.2) | .21 |
| Adenosine diphosphate inhibitors | 221 (16.4) | 17 (14.9) | .74 |
| Immune suppressants | 276 (11.1) | 23 (11.3) | .81 |
| Recent dentist visit | 116 (4.6) | 8 (4.0) | .82 |
Patients initiating dialysis during admission were significantly younger (median age 70.1 years vs 74.8 years) and were more self-reliant in activities of daily living before admission (78.3% vs 69.8%). Compared with patients without dialysis treatment, patients who received dialysis were more comorbid carrying higher proportions of known CKD (26.5% vs 10.9%), diabetes mellitus (31.0% vs 21.1%), liver disease (8.4% vs 4.7%) but less often had active cancer (4.9% vs 9.8%).
Clinical presentation and disease characteristics
Table 2 summarizes disease characteristics and IE related complications. At hospital presentation, patients who subsequently required dialysis generally exhibited more severe disease manifestations, including heart failure (7.9% vs 2.3%), severe valvular insufficiency (14.8% vs 7.3%), sepsis (35.5% vs 22.1%), and embolism (16.8% vs 11.8%).
Table 2
Disease characteristics and IE related complications during admission.
| No dialysis N = 2,534 | Dialysis N = 203 | P value | |
|---|---|---|---|
| Complications at presentation, N (%) | |||
| Heart failure | 57 (2.3) | 16 (7.9) | <.01 |
| Conduction abnormalities | 118 (4.7) | 9 (4.4) | .88 |
| Valvular insufficiency | 185 (7.3) | 30 (14.8) | <.01 |
| Sepsis | 560 (22.1) | 72 (35.5) | <.01 |
| Embolism | 299 (11.8) | 34 (16.8) | <.01 |
| IE type, N (%) | |||
| Prosthetic valve endocarditis | 591 (23.3) | 57 (28.1) | .12 |
| Native valve endocarditis | 1,943 (76.7) | 146 (71.9) | – |
| Aortic valve endocarditis | 1,425 (56.2) | 119 (58.6) | .16 |
| Mitral valve endocarditis | 862 (34.0) | 58 (28.6) | – |
| Aortic and mitral valve endocarditis | 247 (9.8) | 26 (12.8) | – |
| Bacterial etiology, N (%) | |||
| Culture negative endocarditis | 197 (7.8) | 8 (4.0) | <.01 |
| Streptococcus spp. | 898 (35.4) | 49 (24.1) | – |
| Staphylococcus aureus | 708 (27.9) | 84 (41.4) | – |
| Coagulase negative staphylococci | 123 (4.9) | 14 (6.9) | – |
| Enterococcus spp. | 428 (16.9) | 36 (17.7) | – |
| Others (including HACEK and fungi) | 180 (7.1) | 12 (5.9) | – |
| Median vegetation size, mm (25-75 percentiles) | 10 (6.0-14.0) | 13 (8.5-17.5) | <.01 |
| Median length of stay, days (25-75 percentiles) | 35 (25-46) | 50 (30-68) | <.01 |
| Complications during admission, N (%) | |||
| Acute kidney injury | 306 (12.1) | 203 (100.0) | <.01 |
| Embolism | 371 (14.7) | 51 (25.1) | <.01 |
| Heart failure | 135 (5.3) | 41 (20.2) | <.01 |
| Conduction abnormalities | 277 (10.9) | 45 (22.2) | <.01 |
| Persisting bacteremia | 108 (4.3) | 22 (10.8) | <.01 |
| Severe or worsening valve regurgitation | 239 (9.4) | 36 (17.8) | <.01 |
| Sepsis | 240 (9.5) | 62 (30.5) | <.01 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
