- Email: [email protected]
Risk Factors for Mortality in Patients Undergoing Cardiothoracic Surgery for Infective Endocarditis
ORIGINAL ARTICLES: ADULT CARDIAC
ADULT CARDIAC SURGERY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/ home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
Aashish Kumar, MBBS, Chris Anstey, FANZCA, FCICM, Peter Tesar, MBBS, FRACS, and Kiran Shekar, FCICM, PhD Department of Intensive Care, Mater Hospital, Brisbane; Department of Intensive Care, Sunshine Coast University Hospital, Sunshine Coast; and Departments of Cardiothoracic Surgery and Intensive Care, The Prince Charles Hospital, Brisbane, Queensland, Australia
Background. This study aimed to identify risk factors associated with mortality of patients who undergo cardiac surgery for infective endocarditis. Methods. A retrospective review was performed of patients with infective endocarditis who underwent cardiac surgery at a quaternary Australian hospital between 2004 and 2014. Patient data were collected and prospectively analyzed. Results. In all, 465 patients underwent surgery during the study period, with 30 deaths (6.45%). Factors independently associated with in-hospital mortality were increasing age (odds ratio [OR] 1.04; 95% confidence interval [CI], 1.01 to 1.07; P [ .009), active bacterial endocarditis at time of operation (OR 4.91; 95% CI, 1.01 to 23.8; P [ .048), preoperative invasive positive pressure
ventilation (OR 3.65; 95% CI, 1.18 to 11.27; P [ .025), increasing cardiopulmonary bypass time (OR 1.01; 95% CI, 1.006 to 1.014; P < .001), and increasing European System for Cardiac Operative Risk Evaluation score (OR 21.73; 95% CI, 2.12 to 223.11; P < .01). Conclusions. The in-hospital mortality of patients with infective endocarditis remains significant, with potential risk factors including increasing age, active bacterial endocarditis, preoperative invasive positive pressure ventilation, increasing cardiopulmonary bypass time, and high European System for Cardiac Operative Risk Evaluation score.
I
infected tissue, foreign material and hardware, drain paravalvular infection, remove potential sources of embolism, and restore endocardial and valvular integrity.7 In certain patient groups, cardiac surgery has been shown to decrease mortality.8,9 However, the benefits of surgery are not uniformly distributed among all patients.10 Considerable variability exists between centers regarding selection of patients for surgery,11 and a significant number of patients with guideline indications for intervention do not undergo surgery.7 Surgical management of infective endocarditis can be challenging, with associated perioperative morbidity and mortality,12 which makes it necessary to better characterize potential perioperative risk factors for mortality in patients with infective endocarditis.13 The aim of the present study is to review patients who underwent cardiac surgery at a quaternary center for bacterial endocarditis over a 10-year period to identify significant risk factors for in-hospital mortality. Detailing of risk factors can assist with preoperative prognostication and perioperatively identifying patients who are likely to have a poor outcome.
nfective endocarditis is an infection of the endothelial surface of the heart, including valves, endocardial wall, septum, and intracardiac devices.1 It remains a significant health care issue that can be associated with considerable morbidity and mortality.2 Significant intracardiac and extracardiac complications can ensue, potentially leading to the requirement of invasive support in the intensive care unit, including but not limited to central nervous system embolism, cardiac failure, septic shock, and renal failure.3 In spite of surgery and modern antibiotic therapy, mortality rates of 10% to 30% are reported.4 The major causes of death are central nervous system events and cardiac failure. As many as 50% of patients with infective endocarditis require cardiac surgery for management of the condition and its sequelae,1 most commonly for heart failure and uncontrolled infection.5,6 Cardiac surgery can remove Accepted for publication May 6, 2019. Address correspondence to Dr Kumar, Mater Hospital Brisbane, Raymond Terrace, South Brisbane, Queensland, QLD 4101, Australia; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2019;108:1101-6) Ó 2019 by The Society of Thoracic Surgeons
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.05.029
ADULT CARDIAC
Risk Factors for Mortality in Patients Undergoing Cardiothoracic Surgery for Infective Endocarditis
1102
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Ann Thorac Surg 2019;108:1101-6
ADULT CARDIAC
Patients and Methods The Prince Charles Hospital is a quaternary cardiothoracic referral center in Brisbane, Queensland, Australia. More than 1000 cardiac surgeries are performed at this center every year by a number of operators, including a significant number on patients with bacterial endocarditis. After Institutional Ethics approval (HREC/15/ QPCH/217), a retrospective review was performed of 10 years’ worth of cardiac surgical data of patients undergoing surgery for infective endocarditis at The Prince Charles Hospital between the years 2004 and 2014. Data collection was performed from a systemized database, and included preoperative demographic variables, microbiology, organ dysfunction assessments, and intraoperative and postoperative factors. For statistical analysis, STATA 14.1 (StataCorp, College Station, TX) was used. Continuous normally distributed data were summarized as mean (SD), nonnormal or categoric data were summarized as median (interquartile range), and dichotomous data were summarized numerically (percentage). Binary comparisons were made using either a two-tailed Student’s t test for normally distributed data or a Mann-Whitney U test for nonnormally distributed data. In both cases, all observations were assumed to be independent of one another. Using mortality as the outcome variable, predictive factors were analyzed with multivariable logistic regression. Initial variable selection was based on the results of a visual correlation analysis using scatter plots.14 Where rough linear relationships were detected, a more formal analysis using Pearson’s correlation coefficient was used, Table 1.
with R values in excess of 0.80 identified as indicating potential collinearity. Once independent predictors were identified, multivariable logistic analysis was performed using an initial fully saturated model with serial deletion of nonsignificant predictors. Sequential models were tested for significant difference using a standard likelihood ratio test. Variance inflation was tested after each iteration, with results in excess of 2.5 regarded as being potentially collinear. Final significant predictors were reported using their odds ratio (OR) and 95% confidence interval (CI). Postestimation diagnostics included checking for model specification, model fit, collinearity, and outliers using standard STATA utilities. Unless otherwise stated, in all analyses the level of significance was set at P less than 0.05. Enrolled patients were those who had undergone surgery for infective endocarditis, diagnosed based on blood culture and intraoperative findings. The primary endpoint was in-hospital mortality, which was defined as death at any time after the surgery during the hospital admission.
Results Over the 10 years of the study, 465 patients underwent cardiac surgery for infective endocarditis. Of these, 435 (93.55%) survived to hospital discharge, and the remaining 30 (6.45%) died in hospital. The mean age was 50.9 years for survivors and 62 years for nonsurviving patients. In both groups, there was a strong male preponderance. Baseline characteristics of patients are shown in Table 1.
Baseline Characteristics Alive (n ¼ 435)
Dead (n ¼ 30)
P Value
50.9 16.9 301 (69.2) 90/116 (77.6) 116/219 (52.9) 64/208 (30.8) 58 11 160 (36.8) 90/163 (55.2) 52/164 (31.7) 320 (73.6) 182 (41.8) 100 (23) 23 (5.3) 56 (12.9) 40 (9.2) 117 (76) 20 (4.6) 34 (7.8) 31/435 (7.1) 0.11 (0.05-0.27)
62 16.3 19 (63.3) 13/17 (76.5) 17/23 (73.9) 15/19 (78.9) 54 13 18 (60) 12/18 (66.7) 9/18 (50) 28 (93.3) 10 (33.3) 9 (30) 1 (3.3) 4 (13.3) 7 (23.3) 202 (21) 6 (20) 7 (23.3) 8/30 (26) 0.48 (0.26-0.63)
<.001 .51 .92 .055 <.001 .06 .01 .35 .12 .016 .36 .38 .64 .94 .01 <.001 <.001 .004 .002 <.001
Variable Age, y Male Previous admission for CCF History of left heart failure History of right heart failure Ejection fraction, % Previous cardiac surgery Previous aortic valve surgery Previous mitral surgery BE status active BE Staphylococcus sp BE Streptococcus sp BE gram negative BE unknown Acute renal failure Preoperative creatinine, mmol/L Dialysis Preoperative IPPV Preoperative inotropes Median EuroSCORE (IQR)
Values are mean SD or n (%) unless otherwise indicated. BE, bacterial endocarditis; CCF, congestive cardiac failure; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; IQR, interquartile range.
Ann Thorac Surg 2019;108:1101-6
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Table 2.
CPB time, min, mean (SD) Cross clamp time, min, mean (SD) ICU LOS, h, median (IQR) IPPV time, h, median (IQR) ICU readmission, n (%)
Alive (n ¼ 435) 127 (69) 89 (53) 38 (23-75) 11 (7-19) 35 (8.1)
Dead (n ¼ 30) 251 (141) 150 (75) 81 (2-320) 49 (2-114) 3 (10)
P Value <.001 <.001 .24 .054 .45
CPB, cardiopulmonary bypass; ICU, intensive care unit; IPPV, invasive positive pressure ventilation; IQR, interquartile range; LOS, length of stay.
Multivariable Model
Variable Age, y Male History of right heart Previous cardiac surgery Active bacterial endocarditis status Acute renal failure Preoperative IPPV Cardiopulmonary bypass time
OR
95% CI
P Value
1.04 1.08 0.99 1.64 4.91 1.25 3.65 1.010
1.01-1.07 0.42-2.78 0.98-1.01 0.65-4.12 1.01-23.8 0.36-4.28 1.18-11.27 1.006-1.014
.009 .86 .35 .30 .048 .72 .02 <.001
CI, confidence interval; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; OR, odds ratio.
The EuroSCORE was excluded from the first model because it is a composite score of multiple variables, such as age and critical preoperative state, and these had already been included separately in the first multivariable model. The results are presented in Table 3. The multivariable model demonstrated that increasing age (OR 1.04; 95% CI, 1.01 to 1.07; P ¼ .009), preoperative invasive positive pressure ventilation (OR 3.65; 95% CI, 1.18 to 11.27; P ¼ .025), active bacterial endocarditis at time of operation (OR 4.91; 95% CI, 1.01 to 23.8; P ¼ .048), and increasing cardiopulmonary bypass time (OR 1.01; 95% CI, 1.006 to 1.014; P < .001) were significant predictors of increased in-hospital mortality, with preoperative invasive positive pressure ventilation being the strongest predictor. The results of the second multivariate model, which included the EuroSCORE, are demonstrated in Table 4. When the EuroSCORE was included, the multivariate model demonstrated that increasing CPB time and EuroSCORE were predictors of increased mortality, with the strongest predictor being EuroSCORE (OR 21.73; 95% CI, 2.12 to 223.11; P ¼ .01). The age*EuroSCORE interaction term was nonsignificant (OR 0.98, P ¼ .71) with a variance inflation of 14.5, vs 16 for the EuroSCORE.
Table 4. Multivariable Model With European System for Cardiac Operative Risk Evaluation Score Variable
Perioperative Details
Variable
Table 3.
OR
Age, y 1.02 Male 0.94 History of right heart failure 0.99 Previous cardiac surgery 0.85 Active bacterial endocarditis status 2.83 Acute renal failure 0.87 Preoperative IPPV 2.43 Cardiopulmonary bypass time 1.01 EuroSCORE 21.73
95% CI
P Value
0.99-1.05 0.36-2.46 0.98-1.01 0.29-2.51 0.56-14.26 0.25-3.00 0.76-7.73 1.006-1.014 2.12-223.11
.20 .89 .47 .77 .21 .82 .13 .001 .01
CI, confidence interval; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; OR, odds ratio.
ADULT CARDIAC
Overall, 77.6% of the survivors and 76.5% of the nonsurvivors had a history of previous hospital admission for congestive cardiac failure (P ¼ .93); and 53.9% of the survivors had a history of left heart failure, vs 73.9% of the nonsurvivors (P ¼ .055). A significantly greater percentage of the nonsurvivors (78.9%) than the survivors (30.8%) had a history of right heart failure (P < .001). In addition, a greater proportion of the nonsurvivors group (60%) had undergone previous cardiac surgery, as opposed to the survivors (36.8%, P ¼ .01). On subgroup analysis, no difference between groups was noted regarding previous aortic valve or mitral valve surgery. A larger proportion of nonsurvivors (93.3%) compared with survivors (73.6%) had active bacterial endocarditis at the time of surgery (P ¼ .016); however, no significant difference was found on analysis based on specific organisms, separated into Staphylococcus, Streptococcus, gram negative, or unknown organisms. Preoperative invasive positive pressure ventilation was required in 23.3% of the nonsurvivors vs 7.8% of the survivors (P ¼ .004). In the nonsurvivor group, 23.3% had acute renal failure vs 9.2% of the survivor group (P ¼ .01). In addition, a greater proportion of the nonsurvivors (20%) required dialysis, as opposed to 4.6% of the survivors (P < .001). The European System for Cardiac Operative Risk Evaluation (EuroSCORE) data were not available for 10 patients, all of whom were in the survivors’ group. Analysis showed that the survivors had a significantly lower mean EuroSCORE than nonsurvivors (P < .001). The perioperative details are presented in Table 2. The mean cardiopulmonary bypass (CPB) time in the survivors group was 127 minutes, whereas it was 251 minutes in the nonsurvivors group (P < .001). The mean aortic cross-clamp time was 89 minutes in the survivors group and 150 minutes in the nonsurvivors group (P < .001). Nonsurvivors also tended to have a greater intensive care unit length of stay and require a longer duration of postoperative positive pressure ventilation; however, these variables did not reach statistical significance. Based on the univariate analysis, multivariable logistic regression modeling was performed to evaluate risk factors for in-hospital mortality. Two models were created, one with EuroSCORE and one without.
1103
1104
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Table 5.
Receiver-Operating Characteristics Curve Analysisa
Variable
ADULT CARDIAC
Ann Thorac Surg 2019;108:1101-6
Model 1 Age Preoperative IPPV Active bacterial endocarditis Cardiopulmonary bypass time Model 2 Cardiopulmonary bypass time EuroSCORE
AUC (Robust SE)
P Value
(0.051) (0.051) (0.046) (0.037)
<.001 <.001 <.001 <.001
0.787 (0.037) 0.816 (0.040)
<.001 <.001
0.684 0.621 0.651 0.787
a
Outcome variable is mortality.
AUC, area under the curve; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation.
In general, the second model was better specified with a better overall fit to the data (Hosmer-Lemeshow c2 P ¼ 0.79 vs 0.58, respectively). Neither model had significant issues with collinearity or outliers. The results of receiveroperating characteristics curve analysis are presented in Table 5. All predictors demonstrated significance, with EuroSCORE having the highest area under the curve. Model 1 was found to have a pseudo-R2 statistic of 0.27, and model 2 had a pseudo-R2 statistic of 0.31.
Comment Cardiac surgery is protective against morbidity and mortality in patients with infective endocarditis, achieving control of infection, restoring cardiac morphology, and decreasing embolic events.15 A reported 25% to 30% of patients with infective endocarditis will meet criteria for cardiac surgery; however, operative intervention may be challenging, with a quoted in-hospital mortality of 10% to 30%.12 Patients frequently have severe comorbidities and complications of infective endocarditis, which affect the decision to proceed with surgery.16 Guidelines for surgery for infective endocarditis are not supported by strong clinical evidence, and clinical decision making is a complex, heterogeneous process.11 Greater understanding of perioperative risk factors is therefore necessary. Currently, risk factors for in-hospital mortality are poorly defined, and increased understanding remains important for risk stratification, both in terms of surgical decision making and for counseling patients and their families in the perioperative period.17 This retrospective cohort study aimed to review the demographics and operative factors of patients with infective endocarditis who underwent surgery in a quaternary cardiac surgery institution in Brisbane, Australia, to identify predictors of in-hospital mortality of patients. To date this is the first study the authors are aware of, to investigate this particular geographic population. Our study demonstrated an in-hospital mortality of 6.45%. This is relatively low compared with the quoted in-hospital mortality rates of 8% to 30% in the literature.12 Variability in outcomes between centers may arise owing to a multitude of reasons, including but not limited to
patient selection, timing of surgery, case volume, and center experience as a tertiary referral center.18-21 The present study found that increasing age, preoperative ventilation, active bacterial endocarditis, a long CPB time, and increasing EuroSCORE were predictors of increased in-hospital mortality, with the strongest predictors being preoperative invasive positive pressure ventilation and EuroSCORE. De Feo and colleagues17 investigated patients with left-sided native valve endocarditis, and also found preoperative ventilation to be an important predictor of in-hospital mortality. The requirement for preoperative ventilation in the setting of infective endocarditis is likely to represent increasing complications of the disease process, including pulmonary edema, septic shock, and altered level of consciousness, potentially from cerebral embolism.22 Therefore, the subgroup of patients requiring preoperative ventilation are more likely to have multiple organ dysfunctions, affecting their survival to hospital discharge.3 Increasing CPB time is a significant factor that has been shown to be associated with increased morbidity and mortality after many types of cardiac surgery. In the setting of infective endocarditis, Salsano and associates23 demonstrated that prolonged bypass independently predicted mortality, which is in line with the findings of the current study. Active bacterial endocarditis at time of operation was found to be a risk factor for in-hospital mortality. That is not an unexpected finding as cardiac surgery and application of cardiopulmonary bypass in the setting of uncontrolled infection can be associated with worsening hemodynamic and multiorgan dysfunction.6 Ideally, a period of antibiotic therapy would be administered prior. However, early surgical intervention may be necessary in selected patients, and there may be mortality benefits for selected patients.15,18,19,21,24 Active infective endocarditis was reported by Gaca and associates25 in a retrospective review of The Society of Thoracic Surgeons Adult Cardiac Surgery Database as a predictive factor for mortality; however, subsequent studies have not demonstrated this finding as a predictor. Addition of the EuroSCORE to the multivariable model demonstrated that the EuroSCORE was strongly predictive of in-hospital mortality. Although the EuroSCORE is not specific for infective endocarditis, it has been assessed for mortality in this setting previously and demonstrated moderate predictive power on its own.11,20 Madeira and colleagues20 assessed EuroSCORE I and II in a singlecenter retrospective review for in-hospital mortality in patients with active infective endocarditis and found that both satisfactorily stratify risk in these patients. As an individual variable included as part of a preoperative risk profile, the addition of the EuroSCORE may improve overall predictive value than when applied alone. Some correlation exists between our findings and those previously reported; however, significant differences also permeate the available literature. In a retrospective study on 138 patients, Gatti and colleagues12 found the significant risk factors for in-hospital mortality to be preoperative anemia, New York Heart
Ann Thorac Surg 2019;108:1101-6
1105
In conclusion, patients with infective endocarditis who undergo cardiac surgery do have significant mortality. Significant risk factors for in-hospital mortality include increasing age, active bacterial endocarditis, a long CPB time, increasing EuroSCORE, and preoperative ventilation. Further research is required in this area to create accurate predictive models and systems to guide clinicians with regard to surgical decision making. The authors wish to thank Gai Harris and Chantal Kelly (Data Management for the Cardiac Surgical Database, The Prince Charles Hospital, Brisbane, Queensland, Australia), for their assistance. Approved by The Prince Charles Hospital Ethics Committee (HREC/15/QPCH/217).
References 1. Samol A, Kaese S, Bloch J, et al. Infective endocarditis on ICU: risk factors, outcome and long-term follow-up. Infection. 2015;43:287-295. 2. Malhotra A, Rayner J, Williams TM, Prendergast B. Infective endocarditis: therapeutic options and indications for surgery. Curr Cardiol Rep. 2014;16:464. 3. Wolff M, Mourvillier B, Sonneville R, Timsit J-F. My paper 10 years later: infective endocarditis in the intensive care unit. Intensive Care Med. 2014;40:1843-1852. 4. Mu~ noz P, Kestler M, De Alarcon A, et al. Current epidemiology and outcome of infective endocarditis. Medicine (Baltimore). 2015;94:e1816. 5. G alvez-Acebal J, Almendro-Delia M, Ruiz J, et al. Influence of early surgical treatment on the prognosis of left-sided infective endocarditis: a multicenter cohort study. Mayo Clin Proc. 2014;89:1397-1405. 6. Mirabel M, Sonneville R, Hajage D, et al. Long-term outcomes and cardiac surgery in critically ill patients with infective endocarditis. Eur Heart J. 2014;35:1195-1204. 7. Cahill TJ, Baddour LM, Habib G, et al. Challenges in infective endocarditis. J Am Coll Cardiol. 2017;69:325 LP-344. 8. Neragi-Miandoab S, Skripochnik E, Michler R, D’Alessandro D. Risk factors predicting the postoperative outcome in 134 patients with active endocarditis. Heart Surg Forum. 2014;17:35. 9. Iung B, Doco-Lecompte T, Chocron S, et al. Cardiac surgery during the acute phase of infective endocarditis: discrepancies between European Society of Cardiology guidelines and practices. Eur Heart J. 2016;37:840-848. 10. Lalani T, Cabell CH, Benjamin DK, et al. Analysis of the impact of early surgery on in-hospital mortality of native valve endocarditis: use of propensity score and instrumental variable methods to adjust for treatment-selection bias. Circulation. 2010;121:1005-1013. 11. Patrat-Delon S, Rouxel A, Gacouin A, et al. EuroSCORE II underestimates mortality after cardiac surgery for infective endocarditis. Eur J Cardiothorac Surg. 2016;49:944-951. 12. Gatti G, Benussi B, Gripshi F, et al. A risk factor analysis for in-hospital mortality after surgery for infective endocarditis and a proposal of a new predictive scoring system. Infection. 2017;45:413-423. 13. Olmos C, Vilacosta I, Habib G, et al. Risk score for cardiac surgery in active left-sided infective endocarditis. Heart. 2017;103:1435-1442. 14. Anscombe FJ. Graphs in statistical analysis. Am Stat. 1973;27: 17-21. 15. Kang D-H, Kim Y-J, Kim S-H, et al. Early surgery versus conventional treatment for infective endocarditis. N Engl J Med. 2012;366:2466-2473.
ADULT CARDIAC
Association functional class IV, and complicated infection, particularly for patients requiring aortic surgery. Chu and associates16 demonstrated an increased risk of mortality with Staphylococcus aureus, a finding not observed in our study and not consistently reported in other studies. The Society of Thoracic Surgeons database was examined by Gaca and associates,25 who found the significant risk factors for operative mortality to be emergency surgery, cardiogenic shock, active endocarditis, and renal failure. Previous reviews have attempted to develop a scoring system for mortality. Based on a retrospective review of 437 patients, Martinez-Selles and colleagues26 developed the PALSUSE score, which includes prosthetic valve, age more than 70 years, significant intracardiac destruction, Staphylococcus aureus, emergency surgery, female sex, and EuroSCORE greater than 10. Assessment of the score indicated satisfactory discriminatory ability for inhospital mortality. Again based on a retrospective review of 424 patients, Olmos and associates13 created the RISKE score, which included the variables age, prosthetic infection, periannular complications, Staphylococcus aureus or fungal infection, renal failure, septic shock, cardiogenic shock, and thrombocytopenia. Although the findings of other studies and our study do correlate to a degree, there is considerable heterogeneity in the findings.17 Several potential reasons for this exist. These include variability in inclusion criteria for patients, such as Olmos and colleagues13, who included only patients with active left-sided bacterial endocarditis. There was variable consensus between studies as to whether prosthetic valve infective endocarditis should be viewed as a distinct condition from native valve infective endocarditis, which also affected patient selection in studies.17 Other studies examined variables that were not available to us from the dataset utilized, and could significantly affect multivariable analysis. In addition, most studies were conducted in narrow geographic population groups, which may affect the potential risk factors identified.13,17,25,26 Ultimately, the lack of correlation and consensus reflects the need for further studies of larger magnitude to better define this population and the risk factors associated. Our study has several limitations. It has a retrospective observational design and relied on data from the cardiac surgical database of a single quaternary referral center, limiting the available data. As a result, comparative data for nonoperative intervention were not available nor could the impact of timing of surgery be assessed. As a quaternary referral center, patients are often referred from other institutions, and the overall patient cohort may not reflect the characteristics of the general population. Patients presented at various stages of severity of their illness and spent variable amounts of time in the hospital or intensive care unit before transfer. EuroSCORE data were not available for 10 patients. Finally, only information on in-hospital mortality was available, and long-term outcome remains unclear.
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
ADULT CARDIAC
1106
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
16. Chu VH, Park LP, Athan E, et al. 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. 17. De Feo M, Cotrufo M, Carozza A, et al. The need for a specific risk prediction system in native valve infective endocarditis surgery. Sci World J. 2012;2012:307571. 18. Pericart L, Fauchier L, Bourguignon T, et al. Long-term outcome and valve surgery for infective endocarditis in the systematic analysis of a community study. Ann Thorac Surg. 2016;102:496-504. 19. Dunne B, Marr T, Kim D, et al. Infective endocarditis. Heart Lung Circ. 2014;23:628-635. 20. Leroy O, Georges H, Devos P, et al. Infective endocarditis requiring ICU admission: epidemiology and prognosis. Ann Intensive Care. 2015;5:45. 21. David TE, Gavra G, Feindel CM, Regesta T, Armstrong S, Maganti MD. Surgical treatment of active infective
Ann Thorac Surg 2019;108:1101-6
22. 23.
24. 25. 26.
endocarditis: a continued challenge. J Thorac Cardiovasc Surg. 2007;133:144-149. Rubinovitch B, Pittet D. Infective endocarditis: too ill to be operated? Crit Care. 2002;6:106-107. Salsano A, Giacobbe DR, Sportelli E, et al. Aortic cross-clamp time and cardiopulmonary bypass time: prognostic implications in patients operated on for infective endocarditis. Interact Cardiovasc Thorac Surg. 2018;27:328-335. Meszaros K, Nujic S, Sodeck GH, et al. Long-term results after operations for active infective endocarditis in native and prosthetic valves. Ann Thorac Surg. 2012;94:1204-1210. 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. Martínez-Selles M, Mu~ noz P, Arnaiz A, et al. Valve surgery in active infective endocarditis: a simple score to predict inhospital prognosis. Int J Cardiol. 2014;175:133-137.
ADULT CARDIAC SURGERY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/ home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
Aashish Kumar, MBBS, Chris Anstey, FANZCA, FCICM, Peter Tesar, MBBS, FRACS, and Kiran Shekar, FCICM, PhD Department of Intensive Care, Mater Hospital, Brisbane; Department of Intensive Care, Sunshine Coast University Hospital, Sunshine Coast; and Departments of Cardiothoracic Surgery and Intensive Care, The Prince Charles Hospital, Brisbane, Queensland, Australia
Background. This study aimed to identify risk factors associated with mortality of patients who undergo cardiac surgery for infective endocarditis. Methods. A retrospective review was performed of patients with infective endocarditis who underwent cardiac surgery at a quaternary Australian hospital between 2004 and 2014. Patient data were collected and prospectively analyzed. Results. In all, 465 patients underwent surgery during the study period, with 30 deaths (6.45%). Factors independently associated with in-hospital mortality were increasing age (odds ratio [OR] 1.04; 95% confidence interval [CI], 1.01 to 1.07; P [ .009), active bacterial endocarditis at time of operation (OR 4.91; 95% CI, 1.01 to 23.8; P [ .048), preoperative invasive positive pressure
ventilation (OR 3.65; 95% CI, 1.18 to 11.27; P [ .025), increasing cardiopulmonary bypass time (OR 1.01; 95% CI, 1.006 to 1.014; P < .001), and increasing European System for Cardiac Operative Risk Evaluation score (OR 21.73; 95% CI, 2.12 to 223.11; P < .01). Conclusions. The in-hospital mortality of patients with infective endocarditis remains significant, with potential risk factors including increasing age, active bacterial endocarditis, preoperative invasive positive pressure ventilation, increasing cardiopulmonary bypass time, and high European System for Cardiac Operative Risk Evaluation score.
I
infected tissue, foreign material and hardware, drain paravalvular infection, remove potential sources of embolism, and restore endocardial and valvular integrity.7 In certain patient groups, cardiac surgery has been shown to decrease mortality.8,9 However, the benefits of surgery are not uniformly distributed among all patients.10 Considerable variability exists between centers regarding selection of patients for surgery,11 and a significant number of patients with guideline indications for intervention do not undergo surgery.7 Surgical management of infective endocarditis can be challenging, with associated perioperative morbidity and mortality,12 which makes it necessary to better characterize potential perioperative risk factors for mortality in patients with infective endocarditis.13 The aim of the present study is to review patients who underwent cardiac surgery at a quaternary center for bacterial endocarditis over a 10-year period to identify significant risk factors for in-hospital mortality. Detailing of risk factors can assist with preoperative prognostication and perioperatively identifying patients who are likely to have a poor outcome.
nfective endocarditis is an infection of the endothelial surface of the heart, including valves, endocardial wall, septum, and intracardiac devices.1 It remains a significant health care issue that can be associated with considerable morbidity and mortality.2 Significant intracardiac and extracardiac complications can ensue, potentially leading to the requirement of invasive support in the intensive care unit, including but not limited to central nervous system embolism, cardiac failure, septic shock, and renal failure.3 In spite of surgery and modern antibiotic therapy, mortality rates of 10% to 30% are reported.4 The major causes of death are central nervous system events and cardiac failure. As many as 50% of patients with infective endocarditis require cardiac surgery for management of the condition and its sequelae,1 most commonly for heart failure and uncontrolled infection.5,6 Cardiac surgery can remove Accepted for publication May 6, 2019. Address correspondence to Dr Kumar, Mater Hospital Brisbane, Raymond Terrace, South Brisbane, Queensland, QLD 4101, Australia; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2019;108:1101-6) Ó 2019 by The Society of Thoracic Surgeons
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.05.029
ADULT CARDIAC
Risk Factors for Mortality in Patients Undergoing Cardiothoracic Surgery for Infective Endocarditis
1102
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Ann Thorac Surg 2019;108:1101-6
ADULT CARDIAC
Patients and Methods The Prince Charles Hospital is a quaternary cardiothoracic referral center in Brisbane, Queensland, Australia. More than 1000 cardiac surgeries are performed at this center every year by a number of operators, including a significant number on patients with bacterial endocarditis. After Institutional Ethics approval (HREC/15/ QPCH/217), a retrospective review was performed of 10 years’ worth of cardiac surgical data of patients undergoing surgery for infective endocarditis at The Prince Charles Hospital between the years 2004 and 2014. Data collection was performed from a systemized database, and included preoperative demographic variables, microbiology, organ dysfunction assessments, and intraoperative and postoperative factors. For statistical analysis, STATA 14.1 (StataCorp, College Station, TX) was used. Continuous normally distributed data were summarized as mean (SD), nonnormal or categoric data were summarized as median (interquartile range), and dichotomous data were summarized numerically (percentage). Binary comparisons were made using either a two-tailed Student’s t test for normally distributed data or a Mann-Whitney U test for nonnormally distributed data. In both cases, all observations were assumed to be independent of one another. Using mortality as the outcome variable, predictive factors were analyzed with multivariable logistic regression. Initial variable selection was based on the results of a visual correlation analysis using scatter plots.14 Where rough linear relationships were detected, a more formal analysis using Pearson’s correlation coefficient was used, Table 1.
with R values in excess of 0.80 identified as indicating potential collinearity. Once independent predictors were identified, multivariable logistic analysis was performed using an initial fully saturated model with serial deletion of nonsignificant predictors. Sequential models were tested for significant difference using a standard likelihood ratio test. Variance inflation was tested after each iteration, with results in excess of 2.5 regarded as being potentially collinear. Final significant predictors were reported using their odds ratio (OR) and 95% confidence interval (CI). Postestimation diagnostics included checking for model specification, model fit, collinearity, and outliers using standard STATA utilities. Unless otherwise stated, in all analyses the level of significance was set at P less than 0.05. Enrolled patients were those who had undergone surgery for infective endocarditis, diagnosed based on blood culture and intraoperative findings. The primary endpoint was in-hospital mortality, which was defined as death at any time after the surgery during the hospital admission.
Results Over the 10 years of the study, 465 patients underwent cardiac surgery for infective endocarditis. Of these, 435 (93.55%) survived to hospital discharge, and the remaining 30 (6.45%) died in hospital. The mean age was 50.9 years for survivors and 62 years for nonsurviving patients. In both groups, there was a strong male preponderance. Baseline characteristics of patients are shown in Table 1.
Baseline Characteristics Alive (n ¼ 435)
Dead (n ¼ 30)
P Value
50.9 16.9 301 (69.2) 90/116 (77.6) 116/219 (52.9) 64/208 (30.8) 58 11 160 (36.8) 90/163 (55.2) 52/164 (31.7) 320 (73.6) 182 (41.8) 100 (23) 23 (5.3) 56 (12.9) 40 (9.2) 117 (76) 20 (4.6) 34 (7.8) 31/435 (7.1) 0.11 (0.05-0.27)
62 16.3 19 (63.3) 13/17 (76.5) 17/23 (73.9) 15/19 (78.9) 54 13 18 (60) 12/18 (66.7) 9/18 (50) 28 (93.3) 10 (33.3) 9 (30) 1 (3.3) 4 (13.3) 7 (23.3) 202 (21) 6 (20) 7 (23.3) 8/30 (26) 0.48 (0.26-0.63)
<.001 .51 .92 .055 <.001 .06 .01 .35 .12 .016 .36 .38 .64 .94 .01 <.001 <.001 .004 .002 <.001
Variable Age, y Male Previous admission for CCF History of left heart failure History of right heart failure Ejection fraction, % Previous cardiac surgery Previous aortic valve surgery Previous mitral surgery BE status active BE Staphylococcus sp BE Streptococcus sp BE gram negative BE unknown Acute renal failure Preoperative creatinine, mmol/L Dialysis Preoperative IPPV Preoperative inotropes Median EuroSCORE (IQR)
Values are mean SD or n (%) unless otherwise indicated. BE, bacterial endocarditis; CCF, congestive cardiac failure; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; IQR, interquartile range.
Ann Thorac Surg 2019;108:1101-6
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Table 2.
CPB time, min, mean (SD) Cross clamp time, min, mean (SD) ICU LOS, h, median (IQR) IPPV time, h, median (IQR) ICU readmission, n (%)
Alive (n ¼ 435) 127 (69) 89 (53) 38 (23-75) 11 (7-19) 35 (8.1)
Dead (n ¼ 30) 251 (141) 150 (75) 81 (2-320) 49 (2-114) 3 (10)
P Value <.001 <.001 .24 .054 .45
CPB, cardiopulmonary bypass; ICU, intensive care unit; IPPV, invasive positive pressure ventilation; IQR, interquartile range; LOS, length of stay.
Multivariable Model
Variable Age, y Male History of right heart Previous cardiac surgery Active bacterial endocarditis status Acute renal failure Preoperative IPPV Cardiopulmonary bypass time
OR
95% CI
P Value
1.04 1.08 0.99 1.64 4.91 1.25 3.65 1.010
1.01-1.07 0.42-2.78 0.98-1.01 0.65-4.12 1.01-23.8 0.36-4.28 1.18-11.27 1.006-1.014
.009 .86 .35 .30 .048 .72 .02 <.001
CI, confidence interval; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; OR, odds ratio.
The EuroSCORE was excluded from the first model because it is a composite score of multiple variables, such as age and critical preoperative state, and these had already been included separately in the first multivariable model. The results are presented in Table 3. The multivariable model demonstrated that increasing age (OR 1.04; 95% CI, 1.01 to 1.07; P ¼ .009), preoperative invasive positive pressure ventilation (OR 3.65; 95% CI, 1.18 to 11.27; P ¼ .025), active bacterial endocarditis at time of operation (OR 4.91; 95% CI, 1.01 to 23.8; P ¼ .048), and increasing cardiopulmonary bypass time (OR 1.01; 95% CI, 1.006 to 1.014; P < .001) were significant predictors of increased in-hospital mortality, with preoperative invasive positive pressure ventilation being the strongest predictor. The results of the second multivariate model, which included the EuroSCORE, are demonstrated in Table 4. When the EuroSCORE was included, the multivariate model demonstrated that increasing CPB time and EuroSCORE were predictors of increased mortality, with the strongest predictor being EuroSCORE (OR 21.73; 95% CI, 2.12 to 223.11; P ¼ .01). The age*EuroSCORE interaction term was nonsignificant (OR 0.98, P ¼ .71) with a variance inflation of 14.5, vs 16 for the EuroSCORE.
Table 4. Multivariable Model With European System for Cardiac Operative Risk Evaluation Score Variable
Perioperative Details
Variable
Table 3.
OR
Age, y 1.02 Male 0.94 History of right heart failure 0.99 Previous cardiac surgery 0.85 Active bacterial endocarditis status 2.83 Acute renal failure 0.87 Preoperative IPPV 2.43 Cardiopulmonary bypass time 1.01 EuroSCORE 21.73
95% CI
P Value
0.99-1.05 0.36-2.46 0.98-1.01 0.29-2.51 0.56-14.26 0.25-3.00 0.76-7.73 1.006-1.014 2.12-223.11
.20 .89 .47 .77 .21 .82 .13 .001 .01
CI, confidence interval; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation; OR, odds ratio.
ADULT CARDIAC
Overall, 77.6% of the survivors and 76.5% of the nonsurvivors had a history of previous hospital admission for congestive cardiac failure (P ¼ .93); and 53.9% of the survivors had a history of left heart failure, vs 73.9% of the nonsurvivors (P ¼ .055). A significantly greater percentage of the nonsurvivors (78.9%) than the survivors (30.8%) had a history of right heart failure (P < .001). In addition, a greater proportion of the nonsurvivors group (60%) had undergone previous cardiac surgery, as opposed to the survivors (36.8%, P ¼ .01). On subgroup analysis, no difference between groups was noted regarding previous aortic valve or mitral valve surgery. A larger proportion of nonsurvivors (93.3%) compared with survivors (73.6%) had active bacterial endocarditis at the time of surgery (P ¼ .016); however, no significant difference was found on analysis based on specific organisms, separated into Staphylococcus, Streptococcus, gram negative, or unknown organisms. Preoperative invasive positive pressure ventilation was required in 23.3% of the nonsurvivors vs 7.8% of the survivors (P ¼ .004). In the nonsurvivor group, 23.3% had acute renal failure vs 9.2% of the survivor group (P ¼ .01). In addition, a greater proportion of the nonsurvivors (20%) required dialysis, as opposed to 4.6% of the survivors (P < .001). The European System for Cardiac Operative Risk Evaluation (EuroSCORE) data were not available for 10 patients, all of whom were in the survivors’ group. Analysis showed that the survivors had a significantly lower mean EuroSCORE than nonsurvivors (P < .001). The perioperative details are presented in Table 2. The mean cardiopulmonary bypass (CPB) time in the survivors group was 127 minutes, whereas it was 251 minutes in the nonsurvivors group (P < .001). The mean aortic cross-clamp time was 89 minutes in the survivors group and 150 minutes in the nonsurvivors group (P < .001). Nonsurvivors also tended to have a greater intensive care unit length of stay and require a longer duration of postoperative positive pressure ventilation; however, these variables did not reach statistical significance. Based on the univariate analysis, multivariable logistic regression modeling was performed to evaluate risk factors for in-hospital mortality. Two models were created, one with EuroSCORE and one without.
1103
1104
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
Table 5.
Receiver-Operating Characteristics Curve Analysisa
Variable
ADULT CARDIAC
Ann Thorac Surg 2019;108:1101-6
Model 1 Age Preoperative IPPV Active bacterial endocarditis Cardiopulmonary bypass time Model 2 Cardiopulmonary bypass time EuroSCORE
AUC (Robust SE)
P Value
(0.051) (0.051) (0.046) (0.037)
<.001 <.001 <.001 <.001
0.787 (0.037) 0.816 (0.040)
<.001 <.001
0.684 0.621 0.651 0.787
a
Outcome variable is mortality.
AUC, area under the curve; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IPPV, invasive positive pressure ventilation.
In general, the second model was better specified with a better overall fit to the data (Hosmer-Lemeshow c2 P ¼ 0.79 vs 0.58, respectively). Neither model had significant issues with collinearity or outliers. The results of receiveroperating characteristics curve analysis are presented in Table 5. All predictors demonstrated significance, with EuroSCORE having the highest area under the curve. Model 1 was found to have a pseudo-R2 statistic of 0.27, and model 2 had a pseudo-R2 statistic of 0.31.
Comment Cardiac surgery is protective against morbidity and mortality in patients with infective endocarditis, achieving control of infection, restoring cardiac morphology, and decreasing embolic events.15 A reported 25% to 30% of patients with infective endocarditis will meet criteria for cardiac surgery; however, operative intervention may be challenging, with a quoted in-hospital mortality of 10% to 30%.12 Patients frequently have severe comorbidities and complications of infective endocarditis, which affect the decision to proceed with surgery.16 Guidelines for surgery for infective endocarditis are not supported by strong clinical evidence, and clinical decision making is a complex, heterogeneous process.11 Greater understanding of perioperative risk factors is therefore necessary. Currently, risk factors for in-hospital mortality are poorly defined, and increased understanding remains important for risk stratification, both in terms of surgical decision making and for counseling patients and their families in the perioperative period.17 This retrospective cohort study aimed to review the demographics and operative factors of patients with infective endocarditis who underwent surgery in a quaternary cardiac surgery institution in Brisbane, Australia, to identify predictors of in-hospital mortality of patients. To date this is the first study the authors are aware of, to investigate this particular geographic population. Our study demonstrated an in-hospital mortality of 6.45%. This is relatively low compared with the quoted in-hospital mortality rates of 8% to 30% in the literature.12 Variability in outcomes between centers may arise owing to a multitude of reasons, including but not limited to
patient selection, timing of surgery, case volume, and center experience as a tertiary referral center.18-21 The present study found that increasing age, preoperative ventilation, active bacterial endocarditis, a long CPB time, and increasing EuroSCORE were predictors of increased in-hospital mortality, with the strongest predictors being preoperative invasive positive pressure ventilation and EuroSCORE. De Feo and colleagues17 investigated patients with left-sided native valve endocarditis, and also found preoperative ventilation to be an important predictor of in-hospital mortality. The requirement for preoperative ventilation in the setting of infective endocarditis is likely to represent increasing complications of the disease process, including pulmonary edema, septic shock, and altered level of consciousness, potentially from cerebral embolism.22 Therefore, the subgroup of patients requiring preoperative ventilation are more likely to have multiple organ dysfunctions, affecting their survival to hospital discharge.3 Increasing CPB time is a significant factor that has been shown to be associated with increased morbidity and mortality after many types of cardiac surgery. In the setting of infective endocarditis, Salsano and associates23 demonstrated that prolonged bypass independently predicted mortality, which is in line with the findings of the current study. Active bacterial endocarditis at time of operation was found to be a risk factor for in-hospital mortality. That is not an unexpected finding as cardiac surgery and application of cardiopulmonary bypass in the setting of uncontrolled infection can be associated with worsening hemodynamic and multiorgan dysfunction.6 Ideally, a period of antibiotic therapy would be administered prior. However, early surgical intervention may be necessary in selected patients, and there may be mortality benefits for selected patients.15,18,19,21,24 Active infective endocarditis was reported by Gaca and associates25 in a retrospective review of The Society of Thoracic Surgeons Adult Cardiac Surgery Database as a predictive factor for mortality; however, subsequent studies have not demonstrated this finding as a predictor. Addition of the EuroSCORE to the multivariable model demonstrated that the EuroSCORE was strongly predictive of in-hospital mortality. Although the EuroSCORE is not specific for infective endocarditis, it has been assessed for mortality in this setting previously and demonstrated moderate predictive power on its own.11,20 Madeira and colleagues20 assessed EuroSCORE I and II in a singlecenter retrospective review for in-hospital mortality in patients with active infective endocarditis and found that both satisfactorily stratify risk in these patients. As an individual variable included as part of a preoperative risk profile, the addition of the EuroSCORE may improve overall predictive value than when applied alone. Some correlation exists between our findings and those previously reported; however, significant differences also permeate the available literature. In a retrospective study on 138 patients, Gatti and colleagues12 found the significant risk factors for in-hospital mortality to be preoperative anemia, New York Heart
Ann Thorac Surg 2019;108:1101-6
1105
In conclusion, patients with infective endocarditis who undergo cardiac surgery do have significant mortality. Significant risk factors for in-hospital mortality include increasing age, active bacterial endocarditis, a long CPB time, increasing EuroSCORE, and preoperative ventilation. Further research is required in this area to create accurate predictive models and systems to guide clinicians with regard to surgical decision making. The authors wish to thank Gai Harris and Chantal Kelly (Data Management for the Cardiac Surgical Database, The Prince Charles Hospital, Brisbane, Queensland, Australia), for their assistance. Approved by The Prince Charles Hospital Ethics Committee (HREC/15/QPCH/217).
References 1. Samol A, Kaese S, Bloch J, et al. Infective endocarditis on ICU: risk factors, outcome and long-term follow-up. Infection. 2015;43:287-295. 2. Malhotra A, Rayner J, Williams TM, Prendergast B. Infective endocarditis: therapeutic options and indications for surgery. Curr Cardiol Rep. 2014;16:464. 3. Wolff M, Mourvillier B, Sonneville R, Timsit J-F. My paper 10 years later: infective endocarditis in the intensive care unit. Intensive Care Med. 2014;40:1843-1852. 4. Mu~ noz P, Kestler M, De Alarcon A, et al. Current epidemiology and outcome of infective endocarditis. Medicine (Baltimore). 2015;94:e1816. 5. G alvez-Acebal J, Almendro-Delia M, Ruiz J, et al. Influence of early surgical treatment on the prognosis of left-sided infective endocarditis: a multicenter cohort study. Mayo Clin Proc. 2014;89:1397-1405. 6. Mirabel M, Sonneville R, Hajage D, et al. Long-term outcomes and cardiac surgery in critically ill patients with infective endocarditis. Eur Heart J. 2014;35:1195-1204. 7. Cahill TJ, Baddour LM, Habib G, et al. Challenges in infective endocarditis. J Am Coll Cardiol. 2017;69:325 LP-344. 8. Neragi-Miandoab S, Skripochnik E, Michler R, D’Alessandro D. Risk factors predicting the postoperative outcome in 134 patients with active endocarditis. Heart Surg Forum. 2014;17:35. 9. Iung B, Doco-Lecompte T, Chocron S, et al. Cardiac surgery during the acute phase of infective endocarditis: discrepancies between European Society of Cardiology guidelines and practices. Eur Heart J. 2016;37:840-848. 10. Lalani T, Cabell CH, Benjamin DK, et al. Analysis of the impact of early surgery on in-hospital mortality of native valve endocarditis: use of propensity score and instrumental variable methods to adjust for treatment-selection bias. Circulation. 2010;121:1005-1013. 11. Patrat-Delon S, Rouxel A, Gacouin A, et al. EuroSCORE II underestimates mortality after cardiac surgery for infective endocarditis. Eur J Cardiothorac Surg. 2016;49:944-951. 12. Gatti G, Benussi B, Gripshi F, et al. A risk factor analysis for in-hospital mortality after surgery for infective endocarditis and a proposal of a new predictive scoring system. Infection. 2017;45:413-423. 13. Olmos C, Vilacosta I, Habib G, et al. Risk score for cardiac surgery in active left-sided infective endocarditis. Heart. 2017;103:1435-1442. 14. Anscombe FJ. Graphs in statistical analysis. Am Stat. 1973;27: 17-21. 15. Kang D-H, Kim Y-J, Kim S-H, et al. Early surgery versus conventional treatment for infective endocarditis. N Engl J Med. 2012;366:2466-2473.
ADULT CARDIAC
Association functional class IV, and complicated infection, particularly for patients requiring aortic surgery. Chu and associates16 demonstrated an increased risk of mortality with Staphylococcus aureus, a finding not observed in our study and not consistently reported in other studies. The Society of Thoracic Surgeons database was examined by Gaca and associates,25 who found the significant risk factors for operative mortality to be emergency surgery, cardiogenic shock, active endocarditis, and renal failure. Previous reviews have attempted to develop a scoring system for mortality. Based on a retrospective review of 437 patients, Martinez-Selles and colleagues26 developed the PALSUSE score, which includes prosthetic valve, age more than 70 years, significant intracardiac destruction, Staphylococcus aureus, emergency surgery, female sex, and EuroSCORE greater than 10. Assessment of the score indicated satisfactory discriminatory ability for inhospital mortality. Again based on a retrospective review of 424 patients, Olmos and associates13 created the RISKE score, which included the variables age, prosthetic infection, periannular complications, Staphylococcus aureus or fungal infection, renal failure, septic shock, cardiogenic shock, and thrombocytopenia. Although the findings of other studies and our study do correlate to a degree, there is considerable heterogeneity in the findings.17 Several potential reasons for this exist. These include variability in inclusion criteria for patients, such as Olmos and colleagues13, who included only patients with active left-sided bacterial endocarditis. There was variable consensus between studies as to whether prosthetic valve infective endocarditis should be viewed as a distinct condition from native valve infective endocarditis, which also affected patient selection in studies.17 Other studies examined variables that were not available to us from the dataset utilized, and could significantly affect multivariable analysis. In addition, most studies were conducted in narrow geographic population groups, which may affect the potential risk factors identified.13,17,25,26 Ultimately, the lack of correlation and consensus reflects the need for further studies of larger magnitude to better define this population and the risk factors associated. Our study has several limitations. It has a retrospective observational design and relied on data from the cardiac surgical database of a single quaternary referral center, limiting the available data. As a result, comparative data for nonoperative intervention were not available nor could the impact of timing of surgery be assessed. As a quaternary referral center, patients are often referred from other institutions, and the overall patient cohort may not reflect the characteristics of the general population. Patients presented at various stages of severity of their illness and spent variable amounts of time in the hospital or intensive care unit before transfer. EuroSCORE data were not available for 10 patients. Finally, only information on in-hospital mortality was available, and long-term outcome remains unclear.
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
ADULT CARDIAC
1106
KUMAR ET AL MORTALITY IN INFECTIVE ENDOCARDITIS
16. Chu VH, Park LP, Athan E, et al. 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. 17. De Feo M, Cotrufo M, Carozza A, et al. The need for a specific risk prediction system in native valve infective endocarditis surgery. Sci World J. 2012;2012:307571. 18. Pericart L, Fauchier L, Bourguignon T, et al. Long-term outcome and valve surgery for infective endocarditis in the systematic analysis of a community study. Ann Thorac Surg. 2016;102:496-504. 19. Dunne B, Marr T, Kim D, et al. Infective endocarditis. Heart Lung Circ. 2014;23:628-635. 20. Leroy O, Georges H, Devos P, et al. Infective endocarditis requiring ICU admission: epidemiology and prognosis. Ann Intensive Care. 2015;5:45. 21. David TE, Gavra G, Feindel CM, Regesta T, Armstrong S, Maganti MD. Surgical treatment of active infective
Ann Thorac Surg 2019;108:1101-6
22. 23.
24. 25. 26.
endocarditis: a continued challenge. J Thorac Cardiovasc Surg. 2007;133:144-149. Rubinovitch B, Pittet D. Infective endocarditis: too ill to be operated? Crit Care. 2002;6:106-107. Salsano A, Giacobbe DR, Sportelli E, et al. Aortic cross-clamp time and cardiopulmonary bypass time: prognostic implications in patients operated on for infective endocarditis. Interact Cardiovasc Thorac Surg. 2018;27:328-335. Meszaros K, Nujic S, Sodeck GH, et al. Long-term results after operations for active infective endocarditis in native and prosthetic valves. Ann Thorac Surg. 2012;94:1204-1210. 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. Martínez-Selles M, Mu~ noz P, Arnaiz A, et al. Valve surgery in active infective endocarditis: a simple score to predict inhospital prognosis. Int J Cardiol. 2014;175:133-137.