One-year outcome following biological or mechanical valve replacement for infective endocarditis.

Nearly half of patients require cardiac surgery during the acute phase of infective endocarditis (IE). We describe the characteristics of patients according to the type of valve replacement (mechanical or biological), and examine whether the type of prosthesis was associated with in-hospital and 1-year mortality


Background
Nearly half of patients require cardiac surgery during the acute phase of infective endocarditis (IE).We describe the characteristics of patients according to the type of valve replacement (mechanical or biological), and examine whether the type of prosthesis was associated with in-hospital and 1-year mortality.

Methods and Results
Among 5,591 patients included in the International Collaboration on Endocarditis Prospective Cohort Study, 1,467 patients with definite IE were operated on during the active phase and had a biological (37%) or mechanical (63%) valve replacement.
In multivariable analysis, mechanical prostheses were less commonly implanted in older patients (odds ratio: 0.64 for every 10 years), and in patients with a history of cancer (0.72), but were more commonly implanted in mitral position (1.60).

Conclusions
Patients with IE who receive a biologic valve replacement have significant differences in clinical characteristics compared to patients who receive a mechanical prosthesis.Biologic valve replacement is independently associated with a higher in-hospital and 1-year mortality, a result which is possibly related to patient characteristics rather than valve dysfunction.or older for both the aortic and the mitral positions, but do not provide specific recommendations for surgery in IE. [5] There are limited data to support the choice of either type of prosthesis in IE. [6] The characteristics of patients receiving biological or mechanical prosthesis and the association between type of valve prosthesis and outcome are not clearly defined.Thus, the objectives of this observational study were to describe the characteristics of patients according to the type of prosthesis and to examine the relationship between prosthesis type and 1-year mortality.intravenous drug user; patients treated with valve repair rather than replacement or who received a homograft or an autograft; patients receiving both a mechanical prosthesis and a bioprosthesis and patients whose 1-year survival data were missing.

International Collaboration on Endocarditis -Prospective Cohort Study
A standard case report form was used at all sites to collect data.The case report form included 275 variables and was developed by ICE according to standard definitions.[7] Data were collected during the index hospitalization and then entered at the coordinating centre or by site investigators using an Internet-based data entry system.Clinical characteristics regarding the current episode of IE (including source of acquisition, [10,11] microbiology and echocardiography findings, complications, management, and outcome) were collected.All sites were queried to obtain 1-year outcome data for survival, with use of national death indices, medical records, or patient contact, as available.

Statistical analysis
The outcomes of interest in this study were in-hospital and 1-year mortality.Data are presented as means (standard deviations) for continuous variables and as frequencies (percentages) for categorical variables.Simple comparisons were made with the Wilcoxon rank-sum test or the Chi-square test as appropriate.
A generalized estimating equation method was used to determine factors that predicted implantation of a biological or a mechanical prosthesis.Variables found to have an association with the outcome of interest (p<0.05) on univariable analysis were considered for the final model in a backwards stepwise fashion.The final parameter estimates were converted to odds ratios (OR) with corresponding 95% confidence intervals (CI).
A proportional hazards regression model was used to determine if prosthesis type was associated with 1-year mortality.Variables that differed significantly (p<0.05) between the two prosthesis groups in univariable analysis and clinically sound variables were considered for the final model.Survival times were censored at 1 year or date of last contact.Risk estimates are presented as hazard ratios and 95% CI.Survival curves were produced by plotting the estimated survival distribution obtained from the proportional hazards regression model, stratified by type of prosthesis.Influence of age was studied both per ten-year intervals and with a cutoff of 65 years according to the ACC-AHA valvular disease guidelines.

A C C E P T E D M A N U S C R I P T
All tests were 2-sided, and statistical significance was determined at the 0.05 level.All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).

RESULTS
There were 5,668 patients with definite and possible IE enrolled in the ICE-PCS.
Based on pre-specified inclusion and exclusion criteria for this study, 1,467 patients, including 917 (63%) who received mechanical prostheses only and 550 (37%) who received bioprostheses only, were included in this study (Figure 1).2).

DISCUSSION
In the present study, 1,467 patients received valve prostheses during the acute phase of IE with 37% receiving biologic valve replacement and 63% a mechanical prosthesis.Both inhospital and one-year mortality were higher in the bioprosthesis group.The higher mortality associated with bioprosthesis extended beyond the in-hospital acute phase of IE, and was independently associated with 1-year mortality in multivariable analysis.These results have relevance to current clinical practice, as biologic valve replacements were used in approximately 60% of valve replacement surgeries for IE in the United States from 2002 to

A C C E P T E D M A N U S C R I P T
prostheses, but none have included patients with IE. [13][14][15] In a Veterans' Administration trial involving 575 patients undergoing single aortic or mitral valve replacement randomized to receive a biological or mechanical valve, the 15-year mortality after aortic valve replacement was higher with a bioprosthesis than mechanical prosthesis, but not after mitral valve replacement.[14] Bloomfield et al. randomized 533 patients to biological or mechanical prosthesis, and there was a non significant trend toward higher mortality after 12 years with the bioprosthesis.[13] However, in a meta-analysis of three trials, 5-year and 11-year mortality were not statistically different between the two types of prosthetic valves.[16] Other observational studies have compared the results of biologic or mechanical valve prosthesis for IE.In a previous study of 185 patients who received a valve prosthesis during the acute phase of IE, the 4-year mortality was higher in the bioprosthesis group.[17] In a small study of patients undergoing aortic valve replacement for aortic valve IE, 5-year mortality of patients who received biologic replacements (bioprostheses or homografts) was two-fold higher than for patients who received mechanical valve replacement, yet the increased mortality was evident only in patients less than 65 years of age.[18] Other studies have found no significant difference in mortality for biologic compared to mechanical valve replacement, but a higher rate of reoperation in younger patients who received biologic prosthesis.[19,20] In a recent, retrospective study of patients on dialysis with IE who underwent valve surgery, no difference in longer term mortality was evident between type of valve prosthesis.[21] However, this cohort included patients treated with surgery beyond the acute phase of IE and the very high one-year mortality rate may have overshadowed any valve-related effect.[21] In the current study, the increased in-hospital and 1-year mortality associated with biologic valve replacement was evident only in patients younger than 65 years of age.This

A C C E P T E D M A N U S C R I P T
early, increased mortality in younger patients was a surprising finding.Furthermore, the odds ratio associated with bioprosthetic valve type was modest relative to other variables related to 1-year mortality.Although it is unlikely that biologic prostheses had valve degeneration or failure within 1-year follow up, data regarding post-operative echocardiographic assessment of the prostheses were not available.The selection of mechanical or biologic prosthetic valve in the setting of IE involves multiple considerations, including surgeon's preference and experience, size and expected hemodynamics of the prosthetic valve, patient's predicted longevity and valve durability, and risk of bleeding complications related to long-term anticoagulation.Although biologic valve replacement remained statistically associated with higher mortality after adjustment for certain chronic medical conditions, other variables which may have influenced type of valve prosthesis were not available for analysis in this study.This study has several other limitations.Since this is an observational study, the results are subject to selection bias such that unidentified variables may have influenced surgical decision-making regarding the type of prosthesis implanted.We could not ascertain whether in-hospital or 1-year mortality was due to a mechanical cardiac, infectious, or unrelated cause.Data regarding the use of anticoagulation after valve replacement and relapse of IE were not collected in this study, yet may have influenced outcome.

A C C E P T E D M A N U S C R I P T
In conclusion, in a large, contemporary cohort of patients undergoing valve replacement surgery for active IE, bioprosthetic valve replacement was associated with higher in-hospital and 1-year mortality, particularly in patients younger than 65 years of age.Further a younger patient may reflect other comorbid condition with reduced expected survival.On the other hand, among patients > 65 years of age, other medical conditions may be a greater determinant of mortality than the type of prosthesis implanted.The low C-statistic for the survival model may also indicate that baseline clinical characteristics associated with the acute IE episode are NOT strongly associated with 1 year survival in patients treated with surgery.Previous survival analyses have focused largely on in-hospital or shorter term mortality, but intermediate term mortality may be related to other factors not captured at baseline.

Figure 1 -Figure 2 -Fig. 1 Patients
Figure 1 -Disposition of subjects enrolled in the ICE-PCS cohort

Patient Selection, Data Collection and Outcomes
or mechanical valve replacement were included.Exclusion criteria were: age <18 years old;

Table 2 .
Bioprosthesis use was independently associated with 1-year mortality; the risk of death was The clinical characteristics of patients receiving biologic or mechanical prostheses are presented in Table1.Compared to patients who received mechanical prostheses, those who received bioprostheses were older (61.6 SD 15.2 vs 53.6 SD 15.2 years; p<.0001), more often had a history of cancer (9% vs 6%; p=0.009) and moderate or severe renal disease (9% vs 4%; p<0.001).A higher proportion of bioprostheses were used in North and South America whereas in other regions of the world, mechanical prostheses were more frequently implanted.increased by 30% (hazard ratio: 1.298 [1.011 -1.665]; p = 0.0406).The hazard ratio was significantly higher in patients < 65 years of age (1.620 [1.123-2.339])but not in patients ≥ 65 years of age (0.845 [0.596-1.199]).Kaplan-Meier 1-year mortality estimates were 28.4% in the bioprosthesis group and 19.7% in the mechanical prosthesis group (p < 0.001) (Figure