Infective endocarditis (IE) is a life-threatening infection affecting a native or prosthetic heart valve, the endocardial surface, or an indwelling cardiac device. Even with contemporary diagnostic and therapeutic technology, its 1-year mortality remains high at approximately 30%1 and significant morbidity with a prolonged hospital stay, reduced quality of life, and an increased risk of reinfection.2,3

Management of IE often requires difficult clinical decisions. Our population is aging, and patients with endocarditis often present with frailty and multiple comorbidities. An increase in surgical and transcatheter of valve replacement, as well as intracardiac device implantation, is all associated with a higher rate of IE and make management more complicated. Most studies of IE have been reported from large tertiary care centers. Many of the patients in the studies have been referred to the large hospitals because of complications, and significant numbers of these patients require surgery, including valve replacement. In contrast, the few studies done in community hospitals have confirmed the rarity of this infection while demonstrating differences in microbiology, etiology, and outcomes.

We aim to describe the clinical characteristics, microbiology, and outcomes in a small community teaching hospital in Massachusetts that does not have cardiothoracic surgery.

METHODS

This is a retrospective observational descriptive study performed at MetroWest Medical Center in Framingham, a teaching community hospital without cardiothoracic surgical facilities. The medical records of all adults who were admitted over the past 5 years from February 2016 to September 2021 with a potential diagnosis of infectious endocarditis were reviewed. Patients classified as possible or definite IE based on Modified Duke criteria4 were included in the study. Patients with no bacteria growth in blood cultures were excluded. This study was approved by the institutional review board of MetroWest Medical Center.

Each patient's demographic, clinical presentation, comorbidities, echocardiogram findings, microbiology with antibiotic treatment, surgical indication, discharge or transfer, and mortality data were retrospectively reviewed and analyzed. The Charlson Comorbidity Index (CCI)5 was calculated.

Any invasive procedure, vascular or urinary catheter in place, or surgery noted within 6 months before admission, any admission or antibiotic use within a month, and any history of intravenous drug use and immunocompromised conditions were recorded as preadmission conditions. Predisposing cardiac conditions included prior IE, mitral valve prolapse or regurgitation, bicuspid aortic valve, aortic stenosis or regurgitation, prosthetic valve, and intracardiac device. Refractory heart failure, aortic regurgitation/mitral regurgitation with shock, multidrug-resistant pathogen, persistent bacteremia, paravalvular abscess, or vegetation >10 mm was included to determine surgical indication. The final antibiotic choices based on the isolated organism in blood culture were noted unless empirically treated before transfer or death. The outcomes we included were transfer to a tertiary care center, in-hospital, 30-day, and 1-year mortality.

Continuous variables were expressed as median (interquartile range) according to the population distribution, and categorical variables were summarized as frequency and percentages. For comparison, Student t test was used for continuous variables, and χ2 test for categorical variables. Fisher exact test was used for categorical variables as a nonparametric test whenever necessary. Receiver operating characteristic analysis allowed to determine the optimal cutoff value of variables that best predicted the outcome. Logistic regression and multivariate analysis using the Cox regression were performed to identify variables associated with in-hospital, 30-day, and 1-year mortality. Variables used for multivariate analysis included clinically relevant factors and all significant variables with P < 0.3 in the bivariate analysis. A cumulative survival curve for the occurrence of mortality was performed by Kaplan-Meier method and compared with log-rank test. P value less than 0.05 was considered significant. All statistical analyses were performed using SPSS statistical software (version 28; SPSS, Inc, Chicago, Ill).

RESULTS Patient Baseline and Clinical Characteristics

A total of 99 patients with definite or possible IE were included in the study. The patient age distribution was bimodal with a smaller peak in decade 30 (10.4%) and a bigger peak in decade 80 (23%) (Fig. 1). The median age was 71 years and 50 (51%) were men. The most common comorbidities included end-stage renal disease on dialysis in 12%, history of congestive heart failure in 13%, history of stroke in 13%, history of cancer in 17%, chronic kidney disease in 23%, history of myocardial infarct in 24%, and diabetes mellitus in 27%. The median CCI was 5 (2–6) (Table 1).

FIGURE 1:

A, Age distribution in IE. X axis = age group, expressed in decades; Y axis = percent (%). B, Flow chart showing survival during follow up of the 99 patients included in the study. *Thirty-day death does not include patients who had in-hospital death. **One-year death does not include patients who died within 30 days.

TABLE 1 - Baseline Characteristics of Patients Admitted With IE Characteristic Median age, y 71 Patients (N = 99) Male 50 Female 49 ESRD on dialysis 12 History of congestive heart failure 13 History of stroke 13 Chronic kidney disease 23 History of myocardial infarction 24 Diabetes mellitus 27 History of any cancer 17 Chronic obstructive pulmonary disease 7 Charlson Comorbidity Index 5 IDU 18 Predisposing cardiac condition 55 Prior IE 9

ESRD indicates end-stage renal disease.

More than half of patients (56%) had preexisting conditions including 4 with recent intensive care unit stay, 8 with a recent central vascular catheter, 9 with a recent invasive procedure, 14 with previous antibiotic use within 30 days, and 21 with previous hospital admission within 30 days (not shown). Eighteen of patients (18%) were injecting drug users (IDUs). Predisposing cardiac conditions were present in 55 of patients (56%) including 9% with prior IE, 17% bicuspid aortic valve, aortic stenosis, or regurgitation, 20% mitral regurgitation or mitral valve prolapse, and 29% with a prosthetic valve or intracardiac device (not shown).

At presentation, 30% of the patients had hypotension, 34% had altered mental status, and half of patients a temperature ≤36°C or ≥39°C. The intensive care unit level of care was needed at the time of bacteremia in 20% of patients (not shown). In 70 of patients (71%), the definite diagnosis of IE by Duke criteria was met, and the remaining 29 (29%) met possible IE by these criteria (Table 2).

TABLE 2 - Clinical Characteristics of Patients Admitted With IE Diagnosis of IE by Duke Criteria Patients (N = 99) Possible IE 29 Definite IE 70 Echocardiography 96 TEE only 12 TTE only 50 Both TEE and TTE 34 Vegetation on echocardiography 67 Prosthesis or device 11 Tricuspid valve 12 Mitral valve 22 Aortic valve 28 Vegetation size >10 mm 23

Echocardiography (Echo) was performed in 97% of patients. Transesophageal echocardiography (TEE) was only done in 12 of patients (13%), transthoracic echocardiography (TTE) only was done in 50 of patients (52%) and in 33 of patients (34%), and both TEE and TTE were done. Vegetations were seen on echo in 67 of patients (68%), on prosthesis or devices in 11, on tricuspid valve in 12, on mitral valve in 22, and on aortic valve in 28. In 23 (23%), vegetations had size larger than 10 mm (Table 2).

Comparisons by Age, Surgical Indication, or No Surgical Indication and Transfer to Tertiary Centers

Not surprisingly, patients who were older than 75 years had more comorbidities (6.8 vs 2.7, P < 0.001), more predisposing cardiac conditions (65% vs 48%, P < 0.001), and a higher CCI (65,6 vs 21,5) compared with those who were younger than 75 years. Patients younger than 75 years were more likely to be an IDU (32% vs 0%, P < 0.001) and had higher surgical indication (55% vs 30%, P = 0.013) but had significantly less mortality than those older than 75 years (0% vs 17% in in-hospital mortality, P = 0.002; 5% vs 35% in 30-day mortality, P < 0.001; 18% vs 53% in 1-year mortality, P < 0.001). The causative microorganisms were similar between individuals younger and older than 75 years (Table 3).

TABLE 3 - Comparison by Age ≤75 Versus >75 Characteristic Age ≤75 (n = 56) Age >75 (n = 43) P Age 59 85 * Male 26 24 Charlson Comorbidity Index 2 6 * Predisposing cardiac condition 27 28 * IV drug use 18 0 * Indication for surgery 31 13 ‡ Vascular phenomenon 12 17 † Vegetation size >10 mm 18 0 * Staphylococci 30 23 MSSA 21 14 MRSA 9 9 Streptococci 11 11 Enterococci 5 8 In-hospital death 0 7 † 30-Day mortality 3 15 * 1-Year mortality 10 23 *

For continuous variables, a t test was done, and for categorical variables, a χ2 test was used.

*P < 0.001.

†P < 0.01.

‡P < 0.05.

IV indicates intravenous.

Surgical indications were present in 44 of 99 patients (44%) (Table 4). The group with surgical indications was significantly younger than those without surgical indications and was more likely to be infected with S. aureus, including methicillin-resistant S. aureus (MRSA). Patients with IE caused by streptococci were less likely to have surgical indications. However, there were no significant differences in in-hospital and 1-year mortality between patients with and without surgical indications. Patients with surgical indication were further divided between those who were transferred and those who remained at the MetroWest Medical Center. Twenty four of 44 patients (55%) with surgical indication were transferred and 20 patients (45%) remained in our institution. For patients who had surgical indications but who remained at our institution, there was a trend for older age, more comorbidities, and higher death rates at all 3 time points, but these differences did not reach statistical significance (Table 4).

TABLE 4 - Comparison by Surgical Indication and Transfer Characteristic No Surgical Indication (n = 55) Surgical Indication (n = 44) P Surgical Indication and Transfer (n = 24) Surgical Indication and No Transfer (n = 20) P Median age, y 76 65 * 60 73 Male, n (%) 35 (64) 15 (34) * 8 (33) 7 (35) Charlson Comorbidity Index 5 4 3 5 Predisposing cardiac condition, n (%) 30 (55) 25 (57) 13 (54) 12 (60) IDU, n (%) 6 (11) 12 (27) † 7 (29) 5 (25) Vascular phenomenon, n (%) 16 (29) 15 (34) 8 (33) 7 (35) Vegetation size >10 mm, n (%) 0 23 (52) ‡ 12 (50) 11 (55) Staphylococci, n (%) 23 (24) 30 (77) * 16 (67) 14 (70) MSSA, n (%) 18 (33) 17 (39) 12 (50) 5 (25) † MRSA, n (%) 5 (9) 13 (24) * 4 (17) 9 (45) Streptococci, n (%) 17 (31) 5 (13) † 3 (13) 2 (10) Enterococci, n (%) 9 (16) 4 (10) 1 (5) 3 (15) In-hospital death, n (%) 5 (9) 2 (5) 0 2 (10) 30-Day mortality, n (%) 10 (18) 8 (18) 3 (13) 5 (25) 1-Year mortality, n (%) 18 (33) 15 (34) 6 (25) 9 (45)

For continuous variables, a t test was done, and for categorical variables, a χ2 test was used.

*P < 0.01.

†P < 0.05.

‡P <0.001.

Comparison Between Microbiology Groups

Overall, S. aureus was the most common microorganism (53/88 [60%]), followed by streptococci (22/88 [25%]), enterococci (13/88 [15%]), and coagulase-negative staphylococci (Table 5). We examined the association of baseline characteristics and infecting pathogen with outcomes in 88 patients with IE. In general, there were few statistically important differences, but Table 6 lists some of the differences that reached significance (P < 0.05). Patients with streptococcal IE were older than those infected by staphylococci and, compared with other bacteria, were more likely to have underlying cardiac conditions, to have an infected prosthetic valve, and least likely to be an IDU. Patients with MRSA IE were more likely than patients with streptococcal or enterococcal IE to have indications for surgery and had the highest death rates at all 3 time points. The 30-day mortality in the streptococcal group was significantly lower than in MRSA group (5% vs 44%, P = 0.008).

TABLE 5 - Microbiology of Patients Admitted With IE Microbiology Patients (N = 99) Staphylococci 58 MSSA 35 MRSA 18 Coagulase-negative staphylococci 5 Streptococci 22 α-Hemolytic streptococci 14 β-Hemolytic streptococci 6 Streptococcus bovis 2 Enterococci 13 Gram-negative rods* 2 HACEK† 2 Others‡ 2

*Includes Pseudomonas aeruginosa (n = 1) and Klebsiella pneumoniae (n = 1).

†Includes Cardiobacterium hominis (n = 1) and Aggregatibacter aphrophilus (n = 1).

‡Includes Aerococcus urinae (n = 1) and Lactococcus garvieae (n = 1).

TABLE 6 - Baseline Characteristics and Outcome of 88 Patients With IE According to the Causative Agent Characteristic MSSA (n = 35) MRSA (n = 18) Streptococci (n = 22) Enterococci (n = 13) Median age, y 64 66 76 80 Charlson Comorbidity Index 5 5 5 5 Predisposing cardiac condition 43% 44% 86% 46% Prosthesis or intracardiac device 26% 11% 55% 23% IV drug use 23% 33% 9% 23% Indication for surgery 49% 72% 23% 31% In-hospital death 6% 22% 0 8% 30-Day mortality 20% 44% 5% 15% 1-Year mortality 37% 61% 14% 31%

IV indicates intravenous.

Complications, Outcomes, and Comparison by Mortality Groups

Major complications noted included paravalvular abscess 5/99 (5%), refractory heart failure 9/99 (9%), persistent bacteremia 18/99 (18%), and systemic embolism 31 (31%) (not shown).

Overall, in-hospital mortality was 7% and cumulative 30-day and 1-year mortality were 18% and 33%, respectively (Table 7). Patients who survived were significantly younger (68 years old) than those who died during index admission (82 years old) or who died in 30 days (79 years old), P < 0.05. Survivors had significantly less comorbidities (CCI 4) compared with those with in-hospital death (CCI 9), death within 30 days (CCI 6), and 1-year mortality from discharge (CCI 5) (Table 8 significant data in bold). Survivors were also less likely to have MRSA infection when compared with those with in-hospital, 30-day, and 1-year mortality (10% vs 57% vs 36% vs 23%, P = 0.014).

IE Outcome Patients (N = 99), % Transfer for higher level of care 32 In-hospital mortality 7 30-Day mortality 18 1-Year mortality 33 Met surgery indications 44
TABLE 8 - Comparison by Outcomes Survivors (n = 63) In-Hospital Death (n = 7) Death After Discharge P 30-Day (n = 11) 1-Year (n = 15) Age, y 68 82 79 81 0.008 Male 30 4 7 9 0.673 Charlson Comorbidity Index 4 9 6 5 <0.001 Any preadmission condition 36 4 7 9 NS Predisposing cardiac condition 32 5 8 10 NS IV drug use 14 0 0 2 NS Systemic embolism 16 2 6 5 NS Vegetation size >10 mm 13 0 4 4 NS Indication for surgery 27 2 6 7 NS Major microbiology Staphylococci 27 6 9 9 NS  MSSA 21 2 5 6 NS  MRSA 6 4 4 3 0.014 Streptococci 18 0 1 2 NS Enterococci 9 1 1 2 NS

Analysis of variance test was used for comparison of each characteristic among all groups.

IV indicates intravenous.

Predictors for Mortality

By univariate analysis (Supplementary Data Table 1, https://links.lww.com/IDCP/A60), COPD, MRSA endocarditis, CCI, age, history of myocardial infarction, and congestive heart failure were associated with increased mortality. After multivariate analysis (Supplementary Table 2, https://links.lww.com/IDCP/A60), CCI, and infection caused by MRSA remained significant for mortality.

Predictive Values of Charlson Comorbidity Index, Age, MRSA, and Composite Simple Score

Clinical predictors were evaluated by receiver operating characteristic analysis to assess the predictive values on mortality (Fig. 2). Charlson Comorbidity Index and age showed moderate predictive value for in-hospital mortality (area under curve [AUC], 0.86 vs 0.78, respectively), 30-day mortality (AUC, 0.79 vs 0.72, respectively), and 1-year mortality (AUC, 0.75 vs 0.73, respectively). Charlson Comorbidity Index of 6.5 had sensitivity of 71.4% and specificity of 80.5% for in-hospital death, sensitivity 55.6%, specificity 85.7% for 30-day mortality, as well as 41.9% sensitivity and specificity of 87.7% for 1 year mortality. Age of 75 years had sensitivity of 100%, specificity of 55.8% in predicting in-hospital death, sensitivity of 83.3%, specificity 61.4% for 30-day mortality, as well as sensitivity 74.2% and specificity of 66.7% for 1 year mortality.

FIGURE 2:

Predictive value of age and Charlson Comorbidity Index for in-hospital mortality, 30-day mortality, and 1-year mortality. A, AUC of simple score, 0.892; CCI, 0.862; age, 0.775; MRSA, 0.698. B, AUC of simple score, 0.831; CCI, 0.786; age, 0.717; MRSA 0.650, C, AUC of simple score, 0.783; CCI, 0.753; age, 0.726; MRSA, 0.615.

The cumulative mortality rate was significantly higher in patients who were older than 75 years, with Charlson Comorbidity Index greater than 6.5 and with MRSA. Simple score was derived based on the clinical predictors of mortalities in our patient cohort including age of 75 years, Charlson Comorbidity Index of 6.5, and MRSA and showed the greatest predictive value for in-patient, 30-day, and 1-year mortality (AUC, 0.89 vs 0.83 vs 0.78, respectively) (Figs. 2, 3).

FIGURE 3:

Cumulative survival compared between age ≤75 versus >75, CCI of ≤6.5 versus >6.5, and MRSA versus no MRSA groups (P < 0.001).

DISCUSSION

Our study provides a unique opportunity to assess the current characteristics of IE in a community hospital. We encountered the following findings. First, our study population had a high proportion of elderly patients with staphylococcal endocarditis. Second, CCI and MRSA infection were independent predictors for mortality and a combination of age >75 years, MRSA, and CCI > 6.5 provided excellent predictive value for mortality. Third, we found that the presence of surgical indications did not impact survival.

In our study, 87% of individuals with IE were 70 years or older, similar to other studies done in high-income countries where populations are older, have more comorbidities and more predisposing factors for IE.1,6,7 Framingham is a suburban community with an aging population but at the same time has been affected by the epidemic of IDUs. We compared our data with a previous study of IE in a community hospital done by Venezio et al.8 The authors reported 40 episodes of IE in 37 patients 40 years ago. Only 43% of the patients were older than 70 and only 6 patients had major underlying diseases. There were only 5 patients with prosthetic valve endocarditis; of the patients with native valve IE, nearly a third had underlying valvular heart disease, most commonly rheumatic (not seen in our study). There were no patients with IDU and only 3 had tricuspid valve infection. Most strikingly, 73% of the patients had penicillin-sensitive streptococcal infection and only 17% had IE caused by S. aureus (all methicillin-sensitive S. aureus [MSSA]). The mortality in patients with native valve infection was 18%. Although this study was done 4 decades ago, the comparison illustrates the rise in importance of MSSA and MRSA, the role of IDU and prosthetic devices, and increasingly older and medically complex patients.

Independent predictors of mortality include MRSA specie