Abstract
Dialysis patients are at high risk for infective endocarditis (IE); however, no large contemporary data exist on this issue. We examined outcomes of 44 816 patients with IE on dialysis and 202 547 patients with IE not on dialysis from the Nationwide Inpatient Sample database from 2006 thorough 2011. Dialysis patients were younger (59 ± 15 years vs 62 ± 18 years) and more likely to be female (47% vs 40%) and African‐American (47% vs 40%; all P < 0.001). Hospitalizations for IE in the dialysis group increased from 175 to 222 per 10 000 patients (P trend = 0.04). Staphylococcus aureus was the most common microorganism isolated in both dialysis (61%) and nondialysis (45%) groups. IE due to S aureus (adjusted odds ratio [aOR]: 1.79, 95% confidence interval [CI]: 1.73‐1.84), non‐aureus staphylococcus (aOR: 1.72, 95% CI: 1.64‐1.80), and fungi (aOR: 1.4, 95% CI: 1.12‐1.78) were more likely in the dialysis group, whereas infection due to gram‐negative bacteria (aOR: 0.85, 95% CI: 0.81‐0.89), streptococci (aOR: 0.38, 95% CI: 0.36‐0.39), and enterococci (aOR: 0.78, 95% CI: 0.74‐0.82) were less likely (all P < 0.001). Dialysis patients had higher in‐hospital mortality (aOR: 2.13, 95% CI: 2.04‐2.21), lower likelihood of valve‐replacement surgery (aOR: 0.82, 95% CI: 0.76‐0.86), and higher incidence of stroke (aOR: 1.08, 95% CI: 1.03‐1.12; all P < 0.001). We demonstrate rising incidence of IE‐related hospitalizations in dialysis patients, highlight significant differences in baseline comorbidities and microbiology of IE compared with the general population, and validate the association of long‐term dialysis with worse in‐hospital outcomes.
Keywords: Infective Endocarditis, End‐Stage Renal Disease, Outcomes
1. INTRODUCTION
At an incidence rate of 267 per 100 000 person‐years, 2% to 5% of patients with end‐stage renal disease (ESRD) on dialysis develop infective endocarditis (IE).1 Patients with ESRD have a significantly higher risk of developing IE than does the general population, due to higher prevalence of degenerative valve disease and calcification, bacteremia during repeated vascular access, and uremia‐related immune‐system deficiencies.2, 3, 4 Together, these result in an 18× higher age‐adjusted risk for IE in dialysis patients compared with the general population.5 Short‐ and long‐term outcomes of IE in dialysis patients are particularly poor, with in‐hospital and 1‐year mortality rates estimated at 23.5% and 61.6%, respectively.6, 7, 8 Considering that >460 000 patients in the United States are currently on dialysis, and that this number continues to rise each year,9 the healthcare burden of IE in this population is expected to be immense. Current literature on this issue, however, remains largely limited to noncontemporary or small, single‐center retrospective studies. Further, the epidemiology of IE has changed significantly in recent years, due in part to alterations in patient risk factors, increasing use of invasive procedures, and revisions to guidelines for antibiotic prophylaxis10; whether dialysis patients have experienced a similar phenomenon is not clearly known.
The aims of our study were to examine secular trends of IE and related in‐hospital outcomes among patients on long‐term dialysis in the United States and to compare with IE in the general population using a large contemporary national database.
2. METHODS
2.1. Data source
We analyzed data from the Nationwide Inpatient Sample (NIS) for years 2006 through 2011. The NIS contains data on inpatient hospital stays from states participating in the Healthcare Cost and Utilization Project (N = 46 in 2011) and provides data on roughly 8 million hospitalizations from about 1000 hospitals each year.11 The NIS is designed to approximate a 20% stratified sample of US community hospitals, defined as “all non‐federal, short‐term, general, and other specialty hospitals, excluding hospital units of institutions,” representing >95% of the US population. Criteria used for stratified sampling of hospitals include ownership, bed size, teaching status, urban/rural location, and US region. All discharges from sampled hospitals are included in the NIS database. The NIS is an all‐payer database that covers all patients, including those covered by Medicare, Medicaid, or private insurance, and those who are uninsured. Inpatient stay records in the NIS include clinical and resource‐use information available from discharge abstracts derived from state‐mandated hospital discharge reports. Discharge weights provided by the NIS allow extrapolation to calculate expected national hospitalization rates.11
2.2. Study population
From 2006 to 2011, a total of 47 911 414 hospital records were included in the NIS, corresponding to a national estimate of 239 557 070 hospital discharges. We used the International Classification of Diseases, Ninth Edition, Clinical Modification (ICD‐9‐CM) codes 421.0, 421.1, and 421.9 to identify all adults (age ≥18 years) with the principal diagnosis of IE (N = 284 814). These codes have been validated in previous studies of IE using the NIS.9 We chose the principal diagnosis because it is considered the primary reason for hospital admission. Hospitalizations with missing data on patient age, sex, length of stay, or in‐hospital death were excluded from analysis. We also excluded discharges in which patients were admitted and discharged alive on the same day. Lastly, we excluded discharges in which patients were transferred from other hospitals (to avoid duplication of records), leaving a cohort of 247 363 patients with IE.
We considered patients to have ESRD if they were identified to carry (1) the diagnosis code for chronic kidney disease requiring long‐term dialysis (585.6) or (2) procedure codes for hemodialysis (39.95) or peritoneal dialysis (54.98), excluding those patients who concomitantly had an ICD‐9‐CM code indicating acute renal failure (584.5 to 584.9). This approach has been used in previous studies using administrative databases to accurately identify patients with ESRD.12 ESRD patients on dialysis (N = 44 816) constituted the dialysis group, and the remainder of the cohort (N = 202 547) constituted the control group. ICD‐9‐CM codes used to identify patient comorbidities, in‐hospital procedures, and outcomes can be found in Supporting Information, Table 1, in the online version of this article.
Table 1.
Baseline demographic and hospital characteristics of patients admitted with IE in the United States, 2006–2011
| Variable | Total, N = 247 363 | Dialysis Group, n = 44 816 | Control Group, n = 202 547 | P Value |
|---|---|---|---|---|
| Age, y, mean ± SD | 61.1 ± 17.5 | 59.3 ± 14.6 | 61.5 ± 18.1 | <0.001 |
| Sex | <0.001 | |||
| M | 144 865 (58.6) | 23 877 (53.3) | 120 988 (59.7) | |
| F | 102 498 (41.4) | 20 939 (46.7) | 81 559 (40.3) | |
| Race | <0.001 | |||
| White | 143 987 (69.3) | 16 498 (43.6) | 127 489 (75.0) | |
| African American | 35 449 (17.1) | 14 739 (39.0) | 20 710 (12.2) | |
| Hispanic | 16 440 (7.9) | 4048 (10.7) | 12 392 (7.3) | |
| Asian or Pacific Islander | 4127 (2.0) | 902 (2.4) | 3225 (1.9) | |
| Native American | 1587 (0.8) | 364 (1.0) | 1223 (0.7) | |
| Other | 6125 (2.9) | 1255 (3.3) | 4870 (2.9) | |
| Nonelective hospitalization | 213 093 (86.3) | 38 831 (86.8) | 174 262 (86.2) | 0.003 |
| Primary payer | <0.001 | |||
| Medicare | 136 994 (55.5) | 33 591 (75.2) | 103 403 (51.2) | |
| Medicaid | 32 786 (13.3) | 4819 (10.8) | 27 967 (13.8) | |
| Private insurance | 54 338 (22) | 5213 (11.7) | 49 125 (24.3) | |
| Self‐pay | 13 851 (5.6) | 467 (1.0) | 13 384 (6.6) | |
| No charge | 1637 (0.7) | 54 (0.1) | 1583 (0.8) | |
| Other | 7127 (2.9) | 552 (1.2) | 6575 (3.3) | |
| Median HHI percentile | <0.001 | |||
| 0–25th | 70 261 (29.3) | 16 497 (37.9) | 53 764 (27.3) | |
| 26th–50th | 59 058 (24.6) | 10 740 (24.7) | 48 318 (24.6) | |
| 51st–75th | 56 888 (23.7) | 9488 (21.8) | 47 400 (24.1) | |
| 76th–100th | 53 938 (22.5) | 6838 (15.7) | 47 100 (24) | |
| Bed sizea | <0.001 | |||
| Small | 30 830 (12.6) | 5389 (12.1) | 25 441 (12.7) | |
| Medium | 53 174 (21.7) | 8694 (19.6) | 44 480 (22.1) | |
| Large | 161 444 (65.8) | 30 284 (68.3) | 131 160 (65.2) | |
| Urban location | 227 887 (92.8) | 42 166 (95) | 185 721 (92.4) | <0.001 |
| Teaching hospital | 134 554 (54.8) | 25 552 (57.6) | 109 002 (54.2) | <0.001 |
| Region | <0.001 | |||
| Northeast | 58 060 (23.5) | 8910 (19.9) | 49 150 (24.3) | |
| Midwest | 51 679 (20.9) | 9743 (21.7) | 41 936 (20.7) | |
| South | 90 747 (36.7) | 19 122 (42.7) | 71 625 (35.4) | |
| West | 46 878 (19) | 7042 (15.7) | 39 836 (19.7) | |
| Comorbiditiesb | ||||
| CIEDs | 39 312 (15.9) | 4907 (10.9) | 34.405 (17) | <0.001 |
| Congenital heart disease | 3449 (1.4) | 94 (0.2) | 3.355 (1.7) | <0.001 |
| Rheumatic heart disease | 25 645 (10.4) | 3964 (8.8) | 21 681 (10.7) | <0.001 |
| Hx of drug abuse | 25 394 (10.3) | 1863 (4.2) | 23 531 (11.6) | <0.001 |
| Hx of HIV infection | 4965 (2.0) | 1222 (2.7) | 3743 (1.8) | <0.001 |
| Hx of AIDS | 1167 (0.5) | 476 (1.1) | 691 (0.3) | <0.001 |
| Hx of valve repair/ replacement | 20 521 (8.3) | 1908 (4.3) | 18 613 (9.2) | <0.001 |
| Hx of CHF | 42 098 (17.0) | 9559 (21.3) | 32 539 (16.1) | <0.001 |
| Hx of uncomplicated DM | 47 441 (19.2) | 9677 (21.6) | 37 764 (18.6) | <0.001 |
| Hx of valvular disease | 36 631 (14.8) | 5858 (13.1) | 30 773 (15.3) | <0.001 |
Abbreviations: AIDS, acquired immunodeficiency syndrome; CCS, Clinical Classifications Software; CHF, congestive heart failure; CIED, cardiac implantable electronic device; DM, diabetes mellitus; F, female; HHI, household income; HIV, human immunodeficiency virus; Hx, history; ICD‐9‐CM, International Classification of Diseases, Ninth Edition, Clinical Modification; IE, infective endocarditis; M, male; SD, standard deviation.
Data are presented as n (%) unless otherwise noted.
Number of beds category is specific to hospital location and teaching status; available at http://www.hcup‐us.ahrq.gov/db/vars/hosp_bedsize/nisnote.jsp.
Comorbidities (including the 29 Elixhauser comorbidities) were extracted from the database using ICD‐9‐CM or CCS codes.
2.3. Patient and hospital characteristics
Baseline patient characteristics included demographics (age, sex, race/ethnicity), primary expected payer, median household income for patient's ZIP code, and other clinically relevant comorbidities (diabetes mellitus [DM], congestive heart failure [CHF], valvular heart disease, congenital and rheumatic heart disease, history of in situ cardiovascular implantable electronic devices, history of drug abuse, human immunodeficiency virus [HIV] infection, acquired immunodeficiency syndrome [AIDS], prior valve replacement/repair). We also studied hospital‐level variables, such as teaching status, bed size (small, medium, and large), hospital region (Northeast, Midwest, South, and West), and location (rural or urban).
2.4. Outcomes
Temporal trends of IE‐related hospitalizations of dialysis patients were studied. Microbiology data were subsequently examined and compared between the 2 groups. The primary outcome of interest was all‐cause in‐hospital mortality, defined as “died” during the hospitalization encounter in the NIS database. Incidence of stroke (ischemic or embolic) and rates of valve replacement surgery (VRS) during index hospitalization were identified as secondary outcomes. We also analyzed overall and group‐specific temporal trends in these outcomes.
2.5. Statistical analysis
Weighted data were used for all statistical analyses. Categorical variables are expressed as frequency and continuous variables as mean ± SD or median and interquartile range. Pearson χ2 or Fisher exact tests were used to compare categorical variables, and the t test was used to compare continuous variables. For trend analysis, we used the Cochran‐Mantel‐Haenszel test for categorical variables and linear regression for continuous variables. To determine the associations of dialysis status with outcomes of interest (in‐hospital mortality, VRS, and stroke), multivariable logistic regression models were constructed. Variables included in these models were age, sex, primary expected payer, median household income, clinically relevant comorbidities (DM, CHF, valvular heart disease, congenital and rheumatic heart disease, cardiovascular implantable electronic devices, history of drug abuse, HIV infection, AIDS, and prior valve replacement), and hospital characteristics (region, bed size, location, and teaching status). Race/ethnicity data were missing in ~24% of the study population and were therefore not included in these models. For trend analyses, we included the independent variable “year” as a continuous variable in the regression model to obtain unadjusted and adjusted odds ratio (aOR) per year. Odds ratios (OR) and 95% confidence intervals (CI) were used to report the results of logistic regression. All P values were 2‐sided, with a significance threshold of <0.05. Statistical analyses were performed using SPSS Statistics version 23.0 (IBM Corp., Armonk, NY).
3. RESULTS
3.1. Study population and baseline characteristics
After excluding hospital discharges with missing data on patient age, sex, length of stay, and mortality, we identified 247 363 adults (mean age, 61 ± 18 years; 59% male) who were admitted with a primary diagnosis of IE from 2006 to 2011 (Table 1), of which 44 816 (18%) had ESRD and were on dialysis (dialysis group). Patients in the dialysis group were younger (mean age, 59 ± 15 years, vs 62 ± 18 years for the control group) and more likely to be female (47% vs 40%) and African‐American (39% vs 12%; all P < 0.001). More patients in the dialysis group were admitted to urban (95% vs 92.4%) or teaching hospitals (58% vs 54%) than were patients in the control group (P < 0.001 for both). The proportions of patients in the lowest quartile of median household income (38% vs 27%) and those insured by Medicare (75% vs 51%; P < 0.001 for both) were higher in the dialysis group, as shown in Table 1.
3.2. Incidence of IE, comorbidities, and microbiology
We observed a steady increase in the incidence of hospitalizations for IE in the dialysis group, from 175 per 10 000 ESRD patients in 2006 to 222 per 10 000 ESRD patients in 2011 (P trend = 0.04). Hospital admissions for IE in dialysis patients rose numerically from 6239 (16% of total IE admissions) in 2006 to 9500 (21% of total IE admissions) in 2011 (relative increase: 52%, P trend < 0.001; Table 2).
Table 2.
Trends in hospitalizations for IE over the study period
| 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | Total | P trend | |
|---|---|---|---|---|---|---|---|---|
| All hospitalizations | 39 441 | 37 185 | 40 051 | 42 575 | 42 309 | 45 803 | 247 364 | <0.001 |
| Hospitalizations in dialysis patients | 6239 | 6011 | 7052 | 8192 | 7821 | 9500 | 44 815 | <0.001 |
| Incidence per 10 000 dialysis patients | 175.4 | 162.8 | 183.9 | 205.1 | 188.5 | 222.3 | 0.04 | |
| Hospitalizations in dialysis patients, % of total admissions | 15.8 | 16.2 | 17.6 | 19.2 | 18.5 | 20.7 | 18.1 | <0.001 |
Abbreviations: IE, infective endocarditis.
In the dialysis group, there was a lower prevalence of most IE risk factors including congenital heart disease (0.2% vs 1.7%), rheumatic heart disease (8.8% vs 10.7%), valvular heart disease (13.1% vs 15.3%), prior valve replacement (4.3% vs 9.2%), and history of drug abuse (4.2% vs 11.6%; all P < 0.001). Patients on dialysis were, however, more likely to have either HIV infection (2.7% vs 1.8%) or AIDS (1.1% vs 0.3%), CHF (21.3% vs 16.1%), and DM (21.6% vs 18.6%; all P < 0.001), as shown in Table 1.
Data on microbiology (available for 67.0% of all patients) are shown in Table 3. Staphylococcus aureus infection was most common in the overall population (48%) as well as in the dialysis (61%) and control (45%) groups. Non‐aureus staphylococcus (15%) and streptococcus (33%) were other frequently reported microorganisms in the dialysis and control groups, respectively. IE due to S aureus (aOR: 1.79, 95% CI: 1.73‐1.84, P < 0.001), non‐aureus staphylococcus (aOR: 1.72, 95% CI: 1.64‐1.80, P < 0.001), and fungi (aOR: 1.4, 95% CI: 1.12‐1.78, P < 0.01) were more likely in the dialysis group compared with the control group, whereas infection due to gram‐negative bacteria (aOR: 0.85, 95% CI: 0.81‐0.89), streptococci (aOR: 0.38, 95% CI: 0.36‐0.39), and enterococci (aOR: 0.78, 95% CI: 0.74‐0.82) were less likely (all P < 0.001). A significant increase in the proportion of fungal infection–related IE was seen in dialysis patients, and a reciprocal trend of increasing staphylococcal and decreasing streptococcal infections was noted for the control group (see Supporting Information, Tables 2 and 3, in the online version of this article).
Table 3.
Microbiological spectrum of IE in patients with available data
| Pathogen | Total, N = 165 735 | Dialysis Group, n = 30 995 | Control Group, n = 134 740 | P Value |
|---|---|---|---|---|
| S Aureus | 79 495 (48.0) | 18 860 (60.8) | 60 635 (45.0) | <0.001 |
| Non‐aureus staphylococcus | 16 076 (9.7) | 4541 (14.7) | 11 535 (8.6) | <0.001 |
| Gram‐negative bacilli | 17 504 (10.6) | 3163 (10.2) | 14 341 (10.6) | 0.024 |
| Fungi | 593 (0.4) | 131 (0.4) | 462 (0.3) | 0.038 |
| Streptococcus | 48 738 (29.4) | 4354 (14.0) | 44 384 (32.9) | <0.001 |
| Enterococcus | 18 042 (10.9) | 2734 (8.8) | 15 308 (11.4) | <0.001 |
Abbreviations: IE, infective endocarditis; S Aureus, Staphylococcus aureus.
Data are presented as n (%).
3.3. All‐cause in‐hospital mortality
A higher proportion of dialysis patients with IE experienced in‐hospital death (17% vs 11%; OR: 1.62, 95% CI: 1.58‐1.67, P < 0.001; Table 4). After adjusting for differences in comorbidities and hospital characteristics using a multivariable regression model, dialysis patients were noted to be at an increased risk of in‐hospital mortality (aOR: 2.13, 95% CI: 2.04‐2.21, P < 0.001). A similar increased risk was noted after adjusting for differences in pathogens in the subgroup of patients with such data available (aOR: 1.89, 95% CI: 1.80‐1.99, P < 0.001). A significant decrease in all‐cause in‐hospital mortality was observed for both the dialysis group (19% to 16%; relative decrease, 16%; P trend < 0.001; aOR per year: 0.95, 95% CI: 0.93‐0.96) and the control group (12% to 10%; relative decrease, 17%; P trend < 0.001; aOR per year: 0.95, 95% CI: 0.94‐0.95, P trend < 0.001) from 2006 to 2011 (Table 5).
Table 4.
Inpatient outcomes in patients admitted with IE
| Outcomes | Total, N = 247 363 | Dialysis Group, n = 44 816 | Control Group, n = 202 547 | P Value |
|---|---|---|---|---|
| In‐hospital mortality | 29 813 (12.1) | 7495 (16.7) | 22 318 (11.0) | <0.001 |
| VRS during index hospitalization | 23 627 (9.6) | 3261 (7.3) | 20 366 (10.1) | <0.001 |
| Aortic valve replacement | 15 264 (64.6) | 1802 (55.3) | 13 462 (66.2) | <0.001 |
| Mitral valve replacement | 10 642 (45.0) | 1857 (56.9) | 8785 (43.1) | 0.068 |
| Tricuspid valve replacement | 1283 (5.4) | 121 (3.7) | 1162 (5.7) | <0.001 |
| Pulmonic valve replacement | 210 (0.9) | 10 (0.3) | 200 (1.0) | <0.001 |
| Multiple valve replacement | 3718 (15.7) | 524 (16.1) | 3194 (15.7) | <0.001 |
| Acute stroke | 23 486 (9.5) | 4480 (10) | 19 006 (9.4) | <0.001 |
Abbreviations: IE, infective endocarditis; VRS, valve replacement surgery.
Data are presented as n (%).
Table 5.
Trends in in‐hospital outcomes of patients with IE
| Outcome | Year, n (%) | OR per Year (95% CI) | ||||||
|---|---|---|---|---|---|---|---|---|
| 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | |||
| Unadjusted | Adjusted | |||||||
| In‐hospital mortality | 5072 (12.9) | 4285 (11.5) | 5286 (13.2) | 4968 (11.7) | 5051 (11.9) | 5152 (11.2) | 0.98 (0.97‐0.98) | 0.95 (0.94‐0.95) |
| Dialysis group | 1210 (19.4) | 985 (16.4) | 1273 (18.1) | 1219 (14.9) | 1306 (16.7) | 1502 (15.8) | 0.96 (0.95‐0.98) | 0.95 (0.93‐0.96) |
| Control group | 3861 (11.6) | 3300 (10.6) | 4013 (12.2) | 3749 (10.9) | 3745 (10.9) | 3650 (10.1) | 0.98 (0.97‐0.98) | 0.95 (0.94‐0.95) |
| VRS | 3656 (9.3) | 3332 (9.0) | 4041 (10.1) | 4301 (10.1) | 3890 (9.2) | 4408 (9.6) | 1.01 (1.00‐1.02) | 1.00 (0.99‐1.01) |
| Dialysis group | 481 (7.7) | 446 (7.4) | 462 (6.6) | 648 (7.9) | 503 (6.4) | 721 (7.6) | 0.99 (0.97‐1.02) | 0.99 (0.97‐1.02) |
| Control group | 3175 (9.6) | 2886 (9.3) | 3579 (10.8) | 3653 (10.6) | 3386 (9.8) | 3687 (10.2) | 1.01 (1.01‐1.02) | 1.01 (0.99‐1.02) |
| Acute stroke | 3118 (7.9) | 3236 (8.7) | 3690 (9.2) | 4165 (9.8) | 4427 (10.5) | 4850 (10.6) | 1.07 (1.06‐1.08) | 1.04 (1.03‐1.04) |
| Dialysis group | 551 (8.8) | 587 (9.8) | 640 (9.1) | 862 (10.5) | 763 (9.8) | 1076 (11.3) | 1.05 (1.03‐1.07) | 1.03 (1.01‐1.05) |
| Control group | 2568 (7.7) | 2649 (8.5) | 3049 (9.2) | 3303 (9.6) | 3664 (10.6) | 3774 (10.4) | 1.07 (1.06‐1.08) | 1.04 (1.03‐1.05) |
Abbreviations: CI, confidence interval; IE, infective endocarditis; OR, odds ratio; VRS, valve replacement surgery.
3.4. Valve‐replacement surgery
A significantly lower proportion of the dialysis group (7.3%) underwent VRS during index hospitalization compared with the control group (10.1%; OR: 0.70, 95% CI: 0.66‐0.73; aOR: 0.82, 95% CI: 0.76‐0.86; P < 0.001 for both). Similar results were seen in the subgroup of patients with available microbiology data (aOR: 0.87, 95% CI: 0.82‐0.92, P < 0.001). The proportion of patients undergoing VRS remained unchanged in both groups (dialysis group, 7.7% to 7.6%; aOR per year: 0.99, 95% CI: 0.97‐1.02, P trend = not significant. Control group, 9.6% to 10.2%; aOR per year: 1.01, 95% CI: 0.99‐1.02, P trend < 0.01) over the course of the study (Table 5). As shown in Table 4, left‐sided rather than right‐sided VRS was more commonly performed in both groups. Mitral valve replacements outnumbered aortic valve surgeries in the dialysis group (56.9% vs 55.3%, respectively), with an opposite pattern noted for the controls (43.1% vs 66.2%, respectively).
3.5. Stroke
Incidence of acute stroke (ischemic and hemorrhagic) during index hospitalization was similar in the dialysis and control groups (10% vs 9.4%, respectively; OR: 1.03, 95% CI: 0.99‐1.07, P = 0.07). However, after multivariable logistic regression analysis, patients in the dialysis group had a higher rate of acute stroke (aOR: 1.08, 95% CI: 1.03‐1.12, P < 0.001). A similar observation was made in the subgroup of patients with available microbiology data (aOR: 1.27, 95% CI: 1.21‐1.34, P < 0.001). An increase in incidence of stroke in both the dialysis group (8.8% to 11.3%; relative increase, 28.4%; aOR per year: 1.03, 95% CI: 1.01‐1.05, P trend < 0.01) and the control group (7.7% to 10.4%; relative increase, 35%; aOR per year: 1.04, 95% CI: 1.03‐1.05, P trend < 0.001) was noted from 2006 to 2011 (Table 5).
4. DISCUSSION
Using national, population‐based data over a 6‐year period, this is the largest contemporary study to examine the characteristics and in‐hospital outcomes of IE in dialysis patients. The estimated incidence of IE increased steadily from 175 to 222 per 10 000 dialysis patients from 2006 to 2011, and patients on long‐term dialysis were noted to constitute a significant percentage (average, 18%) and an increasingly larger proportion (increasing from 15% to 21%) of overall admissions for IE.
The latter is consistent with the findings of a large prospective cohort study that found dialysis recipients constituting 21% of all IE patients in North America.13 The increase in incidence of IE among dialysis patients, as shown in our study, has also been observed previously14 and may be related to both an improvement in access to sensitive diagnostic tests and heightened risk of IE due to more frequent healthcare contact, higher rates of invasive procedures, more frequent implantation of cardiac devices, as well as a higher number of risk factors for IE among dialysis patients.15, 16, 17 In addition, an increasing number of patients with ESRD may be maintained on chronic hemodialysis through indwelling central lines, rather than arteriovenous fistulae, the former being associated with a higher risk of bacteremia.18 When compared with the control group, however, the dialysis group had lower prevalence of several “traditional” risk factors for IE, including history of intravenous drug abuse, prior heart valve surgery, and previously implanted cardiac devices. This observation signifies the importance of intrinsic factors, such as premature valvular degeneration and related endocardial damage, in pathogenesis of IE in dialysis patients.2, 19
S aureus was the most frequent pathogen for IE, irrespective of dialysis status in our study, and this predominance of staphylococcal infection among IE patients has been reported previously for dialysis patients1, 8, 20 and for the general population.10, 21 The high prevalence of staphylococcal (both aureus and coagulase negative) IE among dialysis patients, as shown in our study, indicates that the intravascular access site is the most frequent portal of bacteremia in such patients.22
Long‐term dialysis was associated with a 2‐fold increased risk of in‐hospital death in this study, and in‐hospital mortality in dialysis patients with IE was about 17%. Therefore, we confirm the previously known poor short‐term survival of dialysis patients with IE6, 14 in a more contemporary population. The increased risk of in‐hospital death in dialysis patients with IE may be related to a higher incidence of complications such as stroke and acute heart failure, or perhaps it may be related to the higher proportion of staphylococcal infections in these patients (a known independent marker of higher mortality in IE patients).13 Other contributing factors may include impaired overall immunological responses to infection in ESRD, infection with more antibiotic‐resistant microorganisms,23 as well as lower rates of VRS in patients with such indications.14, 24 The latter is primarily related to the high‐risk nature of VRS in patients with advanced renal disease and IE.24, 25 In addition, longer‐term outcomes of valve replacement in dialysis patients are also hampered by a heightened risk of prosthetic‐valve IE.26, 27 Despite the relatively higher risk of death from IE in dialysis patients observed in our study, comparison with older studies indicates an overall improvement in survival of dialysis patients hospitalized for IE over the years.6
In addition to the inherently higher risk of strokes on the basis of higher prevalence of vascular risk factors, dialysis patients also have an increased likelihood of intracranial hemorrhage from undiagnosed mycotic aneurysms, related to staphylococcal and fungal infections, particularly when receiving heparin during hemodialysis.28, 29 As VRS is shown to be protective against embolic cerebrovascular events,30 the lower rates of VRS in dialysis patients with IE may also be a contributing factor to the high prevalence of stroke in this group.
4.1. Study limitations
Despite using a large, nationally representative database, our study has certain limitations. Given the abstracted nature of the database, we relied on administrative data to obtain information on comorbidities. Accuracy of certain variables may be influenced by hospital coding practices. NIS is a discharge‐level database and is unable to distinguish between multiple hospitalizations of the same patient; and mortality, strokes, and valve surgeries during repeat hospitalizations were unable to be studied and could have affected the precision of our estimates. Additionally, the database does not provide information on clinical variables, such as patient presentation, as well as details of medical therapy, such as type and duration of antibiotic use. Lastly, regardless of the use of a robust multivariable regression model, residual unmeasured confounders could remain. Despite these limitations, the examination of a large, nationally representative patient sample over 6 years in our study is likely to overcome most of these limitations, and our findings are very likely reflective of true trends in IE hospitalizations.
5. CONCLUSION
In this large, multicenter, population‐based observational study, we demonstrated a rising incidence of IE‐related hospitalizations in dialysis patients, highlighted significant differences in baseline comorbidities and microbiology of IE compared with the general population, and demonstrated the association of long‐term dialysis with worse in‐hospital outcomes.
Author Contributions
Dr. Bhatia and Dr. Agrawal contributed equally to this article and are co–first authors.
Conflicts of interest
The authors declare no potential conflicts of interest.
Supporting information
Supplementary Table 1. International Classification of Diseases, Ninth Edition, Clinical Modification (ICD‐9‐CM) and Clinical Classification Software (CCS) Codes Used to Identify Comorbidities, In‐hospital Procedures and Complications.
Supplementary Table 2. Trends in microbiology of infective endocarditis in dialysis patients and control group.
Supplementary Table 3. Summary trends of microbiology of infective endocarditis among patients admitted with infective endocarditis.
Bhatia N, Agrawal S, Garg A, Mohananey D, Sharma A, Agarwal M, Garg L, Agrawal N, Singh A, Nanda S and Shirani J. Trends and outcomes of infective endocarditis in patients on dialysis, Clin Cardiol, 2017;40:423–429. 10.1002/clc.22688
Contributor Information
Nirmanmoh Bhatia, Email: nirmanmoh@gmail.com.
Sahil Agrawal, Email: Sahilagrawal124@gmail.com.
REFERENCES
- 1. Leither MD, Shroff GR, Ding S, et al. Long‐term survival of dialysis patients with bacterial endocarditis undergoing valvular replacement surgery in the United States. Circulation. 2013;128:344–351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Roberts WC, Taylor MA, Shirani J. Cardiac findings at necropsy in patients with chronic kidney disease maintained on chronic hemodialysis. Medicine (Baltimore) . 2012;91:165–178. [DOI] [PubMed] [Google Scholar]
- 3. Umana E, Ahmed W, Alpert MA. Valvular and perivalvular abnormalities in end‐stage renal disease. Am J Med Sci. 2003;325:237–242. [DOI] [PubMed] [Google Scholar]
- 4. Powe NR, Jaar B, Furth SL, et al. Septicemia in dialysis patients: incidence, risk factors, and prognosis. Kidney Int. 1999;55:1081–1090. [DOI] [PubMed] [Google Scholar]
- 5. Abbott KC, Agodoa LY. Hospitalizations for bacterial endocarditis after initiation of chronic dialysis in the United States. Nephron. 2002;91:203–209. [DOI] [PubMed] [Google Scholar]
- 6. Shroff GR, Herzog CA, Ma JZ, et al. Long‐term survival of dialysis patients with bacterial endocarditis in the United States. Am J Kidney Dis. 2004;44:1077–1082. [DOI] [PubMed] [Google Scholar]
- 7. Maraj S, Jacobs LE, Kung SC, et al. Epidemiology and outcome of infective endocarditis in hemodialysis patients. Am J Med Sci. 2002;324:254–260. [DOI] [PubMed] [Google Scholar]
- 8. Spies C, Madison JR, Schatz IJ. Infective endocarditis in patients with end‐stage renal disease: clinical presentation and outcome. Arch Intern Med. 2004;164:71–75. [DOI] [PubMed] [Google Scholar]
- 9. US Renal Data System . 2015 Annual Data Report: Atlas of Chronic Kidney Disease and End‐Stage Renal Disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2015. [Google Scholar]
- 10. Pant S, Patel NJ, Deshmukh A, et al. Trends in infective endocarditis incidence, microbiology, and valve replacement in the United States from 2000 to 2011. J Am Coll Cardiol. 2015;65:2070–2076. [DOI] [PubMed] [Google Scholar]
- 11. Overview of the National (Nationwide) Inpatient Sample (NIS) . http://www.hcup‐us.ahrq.gov/nisoverview.jsp. Accessed April 30, 2016.
- 12. Kumar G, Sakhuja A, Taneja A, et al; MICCOR Group of Investigators . Pulmonary embolism in patients with CKD and ESRD. Clin J Am Soc Nephrol . 2012;7:1584–1590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Murdoch DR, Corey GR, Hoen B, et al; ICE‐PCS Investigators . Clinical presentation, etiology, and outcome of infective endocarditis in the 21st century: the International Collaboration on Endocarditis–Prospective Cohort Study. Arch Intern Med . 2009;169:463–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Chou MT, Wang JJ, Wu WS, et al. Epidemiologic features and long‐term outcome of dialysis patients with infective endocarditis in Taiwan. Int J Cardiol. 2015;179:465–469. [DOI] [PubMed] [Google Scholar]
- 15. Benito N, Miró JM, de Lazzari E, et al; ICE‐PCS Investigators. Health care‐associated native valve endocarditis: importance of non‐nosocomial acquisition [published correction appears in Ann Intern Med. 2010;152:480]. Ann Intern Med . 2009;150:586–594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Zhan C, Baine WB, Sedrakyan A, et al. Cardiac device implantation in the United States from 1997 through 2004: a population‐based analysis. J Gen Intern Med . 2008;23(suppl 1):13–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Voigt A, Shalaby A, Saba S. Rising rates of cardiac rhythm management device infections in the United States: 1996 through 2003. J Am Coll Cardiol. 2006;48:590–591. [DOI] [PubMed] [Google Scholar]
- 18. Doulton T, Sabharwal N, Cairns HS, et al. Infective endocarditis in dialysis patients: new challenges and old. Kidney Int. 2003;64:720–727. [DOI] [PubMed] [Google Scholar]
- 19. Madu EC, D'Cruz IA, Wall B, et al. Transesophageal echocardiographic spectrum of calcific mitral abnormalities in patients with end‐stage renal disease. Echocardiography. 2000;17:29–35. [DOI] [PubMed] [Google Scholar]
- 20. Dohmen PM, Binner C, Mende M, et al. Outcome of aortic valve replacement for active infective endocarditis in patients on chronic hemodialysis. Ann Thorac Surg. 2015;99:532–538. [DOI] [PubMed] [Google Scholar]
- 21. Slipczuk L, Codolosa JN, Davila CD, et al. Infective endocarditis epidemiology over five decades: a systematic review [published correction appears in PLoS One. 2014;9:e111564]. PLoS One . 2013;8:e82665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Lentino JR, Baddour LM, Wray M, et al. Staphylococcus aureus and other bacteremias in hemodialysis patients: antibiotic therapy and surgical removal of access site. Infection. 2000;28:355–360. [DOI] [PubMed] [Google Scholar]
- 23. Naimi TS, LeDell KH, Como‐Sabetti K, et al. Comparison of community‐ and health care‐associated methicillin‐resistant Staphylococcus aureus infection. JAMA. 2003;290:2976–2984. [DOI] [PubMed] [Google Scholar]
- 24. Chu VH, Park LP, Athan E, et al; ICE Investigators . Association between surgical indications, operative risk, and clinical outcome in infective endocarditis: a prospective study from the International Collaboration on Endocarditis. Circulation . 2015;131:131–140. [DOI] [PubMed] [Google Scholar]
- 25. Gaca JG, Sheng S, Daneshmand MA, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg . 2011;141:98–106.e1‐2. [DOI] [PubMed] [Google Scholar]
- 26. Thourani VH, Sarin EL, Kilgo PD, et al. Short‐ and long‐term outcomes in patients undergoing valve surgery with end‐stage renal failure receiving chronic hemodialysis. J Thorac Cardiovasc Surg. 2012;144:117–123. [DOI] [PubMed] [Google Scholar]
- 27. Farrington DK, Kilgo PD, Thourani VH, et al. High risk of prosthetic valve endocarditis and death after valve replacement operations in dialysis patients. Ann Thorac Surg. 2016;101:2217–2223. [DOI] [PubMed] [Google Scholar]
- 28. Merkler AE, Chu SY, Lerario MP, et al. Temporal relationship between infective endocarditis and stroke. Neurology. 2015;85:512–516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Vilacosta I, Graupner C, San Román JA, et al. Risk of embolization after institution of antibiotic therapy for infective endocarditis. J Am Coll Cardiol. 2002;39:1489–1495. [DOI] [PubMed] [Google Scholar]
- 30. Kang DH, Kim YJ, Kim SH, et al. Early surgery versus conventional treatment for infective endocarditis. N Engl J Med. 2012;366:2466–2473. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1. International Classification of Diseases, Ninth Edition, Clinical Modification (ICD‐9‐CM) and Clinical Classification Software (CCS) Codes Used to Identify Comorbidities, In‐hospital Procedures and Complications.
Supplementary Table 2. Trends in microbiology of infective endocarditis in dialysis patients and control group.
Supplementary Table 3. Summary trends of microbiology of infective endocarditis among patients admitted with infective endocarditis.