Infective endocarditis (IE) is a highly morbid disease, for which most outcomes data come from patients with left-sided valvular lesions. Echocardiographic findings such as vegetation size and prosthetic valve involvement have been identified as important predictors of mortality in left-sided IE, but predictors of outcomes in right-sided IE are less well characterized. Therefore, the aim of this study was to identify clinical and echocardiographic findings predictive of mortality in tricuspid valve (TV) IE. We retrospectively reviewed all echocardiograms showing TV vegetations that were performed at the Massachusetts General Hospital from January 1, 2003, to December 31, 2013. We identified 105 patients who had echocardiographic evidence of TV vegetations and a definite clinical diagnosis of IE based on the modified Duke’s criteria but did not have intracardiac device–associated vegetations. Of the 105 patients, 88 survived until discharge. Clinical and echocardiographic factors that positively correlated with in-hospital mortality included age (p = 0.002), immunosuppression status (p = 0.016), blood urea nitrogen level (p = 0.029), Candida causative organism (p = 0.025), left ventricular ejection fraction <40% (p = 0.027), right ventricular (RV) systolic dysfunction (p = 0.009), and estimated RV systolic pressure >40 mm Hg (p = 0.040). Of these factors, immunosuppression status, blood urea nitrogen level, and RV systolic dysfunction were independently associated with increased in-hospital mortality. In conclusion, RV systolic dysfunction may serve as an echocardiographic marker to aid clinicians in identifying high-risk patients with right-sided IE for more aggressive therapy.
Although right-sided infective endocarditis (IE) accounts for only 5% to 10% of IE cases, right-sided valvular IE remains a major source of morbidity and mortality for affected patients. To date, few studies have evaluated prognostic factors in right-sided IE. Those that have been performed suggest that vegetation size and fungal etiology are predictors of poor outcome. Whether other clinical and echocardiographic factors can identify high-risk patients with right-sided IE remains incompletely understood. We hypothesized that factors specific to right-sided heart disease may impact clinical outcomes in patients with tricuspid valve (TV) IE. Given the limited studies on right-sided IE and the lack of more contemporary data, we conducted this study to identify clinical and echocardiographic findings predictive of mortality in TV IE.
Methods
We retrospectively reviewed all echocardiograms showing TV vegetations that were performed at the Massachusetts General Hospital (MGH) Cardiac Ultrasound Laboratory (Boston, Massachusetts) from January 1, 2003, to December 31, 2013. The MGH echocardiography laboratory performs nearly 30,000 echocardiograms per year. Inclusion criteria included echocardiographic evidence of TV vegetation and a definite clinical diagnosis of IE based on the modified Duke’s criteria as determined through electronic medical record review. Patients were excluded if they had intracardiac device–associated vegetation with no subsequent TV lesion seen on echocardiogram after device removal, if the patient died before a diagnosis of IE was made, if the patient left the hospital against medical advice, or if infectious disease consultation deemed that IE was unlikely. A total of 105 patients met the inclusion criteria and none of the exclusion criteria. Of the 105 patients, 7 had repeat hospitalizations for IE during the aforementioned period, but only the first episode was included for analysis. The Partners Healthcare Institutional Review Board approved the study before initiation.
Clinical and echocardiographic data were obtained from review of the electronic medical record (MYM) and review of the echocardiographic images when available (MYM, RBW). These data included demographic information, past medical history and co-morbidities, laboratory values, chest imaging results, blood culture data, and echocardiogram reports. All laboratory and imaging data were collected as early in the course of illness as possible. Many patients were transferred to MGH after initial admission to another facility, and laboratory and imaging data were obtained on transfer rather than on outside admission. Clinical co-morbidities were acquired from the admission note or the first infectious disease consultation note and were recorded as absent if not specifically identified. Immunosuppression was defined as taking any known immunosuppressing drugs within 3 months of admission and/or receiving cancer chemotherapy within 6 months of admission. The outcome measure was in-hospital mortality because many patients were lost to follow-up after discharge.
Specific echocardiographic information included size of vegetation in the longest dimension measured, degree of tricuspid regurgitation (TR), native or prosthetic valve, concomitant involvement of the mitral or aortic valves, presence of an intracardiac device, right ventricular (RV) dilation, RV systolic dysfunction, left ventricular ejection fraction (LVEF), estimated RV systolic pressure (RVSP), and the presence of a patent foramen ovale. All information was obtained from the first echocardiogram performed during the index admission that identified a vegetation, but if information was missing from the first echocardiogram, then subsequent echocardiograms were reviewed. The presence of RV systolic dysfunction was determined by an integrated qualitative and quantitative approach as recommended by the American Society of Echocardiography (ASE). Specifically, the quantitative parameter used was tricuspid annular plane systolic excursion (TAPSE). This M-mode–based parameter was a routine part of the MGH transthoracic echocardiogram protocol throughout the duration of this study period (2003 to 2013) and was available/reported for all subjects in the study. A value ≤16 mm indicated RV systolic dysfunction, in accordance with ASE guidelines. In addition to TAPSE, visual estimation of RV systolic function was performed. The presence of TAPSE ≤16 mm and/or visual assessment of RV systolic dysfunction were therefore used to define RV systolic dysfunction. Tissue Doppler peak systolic velocity (S′) was not a routine part of the echocardiogram protocol during most of the study period. Therefore, this parameter was not routinely used in the assessment of RV systolic function.
Statistical analysis was performed using R version 3.2.2 ( https://www.r-project.org/ ). Fisher’s exact test and Student t test were used to test univariate association between predictors and outcomes. Statistical significance was defined as p value <0.05 from a 2-sided test. Multivariate analysis was performed using logistic regression with statistically significant variables from univariate analysis. Variables that had significant missing data were excluded from multivariate analysis to minimize bias. Survival analysis was performed using Cox proportional hazard models. Kaplan–Meier curves measured survival from the time of transfer and not from initial hospital admission at a pretransfer facility.
Results
Table 1 lists the clinical and echocardiographic characteristics of the patients (n = 105). All patients had native TV IE. Significant risk factors for in-hospital mortality included age (p = 0.002), immunosuppression status (p = 0.016), blood urea nitrogen level (p = 0.029), Candida organism (p = 0.025), LVEF <40% (p = 0.027), RV systolic dysfunction (p = 0.009), and estimated RVSP >40 mm Hg (p = 0.040). Intravenous drug user (IVDU) status was correlated with improved survival (p = 0.033). Of the patients who were immunosuppressed, 5 received chemotherapy within 6 months of admission; one was a kidney transplant recipient; one was a lung transplant recipient; and 4 were taking prednisone and/or azathioprine. Notably, transfer status from another facility, diabetes mellitus, persistently positive blood cultures >72 hours, Staphylococcus aureus organism (both methicillin-sensitive and methicillin-resistant species), size of TV vegetation, and other valve involvement were not significantly correlated with worse prognosis.
| Variable | Total (N = 105) | In-Hospital Mortality (N = 17) | Survival to Discharge (N = 88) | OR | p value |
|---|---|---|---|---|---|
| Age (years) | 45.0 ± 18.8 | 58.1 ± 17.1 | 42.5 ± 18.1 | 0.002 | |
| Male | 45 (43%) | 6 (35%) | 39 (44%) | 0.688 | 0.597 |
| White race | 86 (82%) | 11 (65%) | 75 (85%) | 0.322 | 0.078 |
| Patient transferred from another hospital | 50 (48%) | 8 (47%) | 42 (48%) | 0.974 | 1 |
| Intravenous drug user | 51 (49%) | 4 (24%) | 47 (53%) | 0.272 | 0.033 |
| Prior IE episode of any valve | 14 (13%) | 2 (12%) | 12 (17%) | 0.846 | 1 |
| Hepatitis C | 42 (40%) | 4 (24%) | 38 (43%) | 0.408 | 0.178 |
| End stage kidney disease on hemodialysis | 6 (6%) | 1 (6%) | 5 (6%) | 1.037 | 1 |
| Diabetes | 20 (19%) | 5 (29%) | 15 (17%) | 1.776 | 0.337 |
| Prior heart failure | 12 (11%) | 3 (18%) | 9 (10%) | 1.868 | 0.407 |
| HIV | 3 (3%) | 2 (12%) | 1 (1%) | 11.163 | 0.067 |
| Cancer | 10 (10%) | 6 (35%) | 12 (14%) | 3.402 | 0.071 |
| Immunosuppression | 11 (11%) | 5 (29%) | 6 (7%) | 5.555 | 0.016 |
| Blood urea nitrogen (mg/dL) | 32.2 ± 25.8 | 49.9 ± 35.5 | 28.7 ± 22.2 | 0.029 | |
| Creatinine (mg/dL) | 1.7 ± 1.6 | 2.1 ± 1.2 | 1.6 ± 1.6 | 0.222 | |
| White blood cell count (K/μL) | 15.0 ± 8.9 | 18.3 ± 9.4 | 14.3 ± 8.7 | 0.124 | |
| Hematocrit (%) | 31.0 ± 6.7 | 31.8 ± 7.3 | 30.9 ± 6.6 | 0.626 | |
| Platelet count (K/μL) | 194 ± 122 | 149 ± 128 | 202 ± 120 | 0.131 | |
| Infiltrates on chest x-ray or chest CT | 70 (67%) | 10 (59%) | 60 (68%) | 0.669 | 0.575 |
| Persistent blood cultures > 72 hours | 26/84 (31%) | 2/10 (20%) | 24/74 (32%) | 0.524 | 0.717 |
| Staphylococcus aureus | 71 (68%) | 11 (65%) | 60 (68%) | 0.857 | 0.783 |
| Methicillin-resistant S. aureus | 27 (26%) | 3 (18%) | 24 (27%) | 0.574 | 0.550 |
| Streptococcus species | 13 (12%) | 2 (12%) | 11 (13%) | 0.934 | 1 |
| Enterococcus species | 11 (11%) | 3 (18%) | 8 (9%) | 2.124 | 0.380 |
| Vancomycin-resistant Enterococcus species | 4 (4%) | 2 (12%) | 2 (2%) | 5.589 | 0.122 |
| Coagulase-negative Staphylococcus species | 6 (6%) | 1 (6%) | 5 (6%) | 1.037 | 1 |
| Candida species | 2 (2%) | 2 (12%) | 0 (0%) | – | 0.025 |
| Other organisms | 7 (7%) | 1 (6%) | 6 (7%) | 0.855 | 1 |
| More than one organism isolated | 11 (11%) | 4 (24%) | 7 (8%) | 3.502 | 0.076 |
| TV vegetation size ≥ 1.0 cm | 70/94 (75%) | 10/14 (71%) | 60/80 (75%) | 0.835 | 0.749 |
| TV vegetation size ≥ 2.0 cm | 32/94 (34%) | 6/14 (43%) | 26/80 (33%) | 1.550 | 0.544 |
| TV vegetation size ≥ 3.0 cm | 13/94 (14%) | 3/14 (21%) | 10/80 (13%) | 1.894 | 0.404 |
| Moderate or severe tricuspid regurgitation | 46 (44%) | 7 (41%) | 39 (44%) | 0.881 | 1 |
| Concurrent aortic or mitral valve vegetation | 19 (18%) | 2 (12%) | 17 (19%) | 0.560 | 0.732 |
| Presence of intracardiac device | 25 (24%) | 5 (29%) | 20 (23%) | 1.391 | 0.549 |
| LVEF (%) | 64.1 ± 12.1 | 58.6 ± 19.9 | 65.1 ± 9.9 | 0.216 | |
| LVEF < 40% | 5/102 (5%) | 3/16 (19%) | 2/86 (2%) | 9.336 | 0.027 |
| RV dilation | 20 (19%) | 4 (24%) | 16 (18%) | 1.380 | 0.736 |
| RV systolic dysfunction | 10 (10%) | 5 (29%) | 5 (6%) | 6.720 | 0.009 |
| Estimated RVSP (mmHg) | 43.9 ± 11.3 | 52.2 ± 11.7 | 42.5 ± 10.7 | 0.010 | |
| Estimated RVSP > 40 mmHg | 57/95 (60%) | 12/14 (86%) | 45/81 (56%) | 4.734 | 0.040 |
| Confirmed presence of patent foramen ovale | 21 (20%) | 2 (12%) | 19 (22%) | 0.487 | 0.551 |
| Surgery within hospitalization | 11 (11%) | 0 (0%) | 11 (13%) | 0 | 0.205 |
| Need for intensive care unit | 47 (45%) | 14 (82%) | 32 (36%) | 7.629 | 0.001 |
| Mechanical ventilation | 34 (32%) | 13 (76%) | 21 (24%) | 10.088 | <0.001 |
| New need for hemodialysis | 12 (11%) | 5 (29%) | 7 (8%) | 4.719 | 0.033 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
