Risk factors for hospital mortality in valve replacement with mechanical prosthesis

Objective: Identification of risk factors for cardiac surgery can improve surgical results. Our aim is to identify factors related to increased hospital mortality for patients who underwent mechanical cardiac prosthesis implant. Methods: Prospective study with retrospective data acquirement study including 335 consecutive patients who underwent at least one implant of St. Jude Medical mechanical prosthesis between December 1994 and September 2005 at the Cardiology Institute of RS. Valve implants were 158 (47.1%) in aortic position, 146 (43.6%) in mitral and 31 (9.3%) in aortic and mitral. The following characteristics were analyzed in relation to hospital death: gender, age, body mass index, NYHA functional class, ejection fraction, type of valve lesion, hypertension, diabetes mellitus, serum creatinine, preoperative arrhythmias, prior heart surgery, CABG surgery, concomitant tricuspid valve surgery and operative priority (elective, urgent or emergent). Logistical regression was used to analyze data and odds-ratio was calculated for individual factors. Results: During the follow-up there were 13 (3.88%) deaths. In-hospital mortality risk was associated with serum creatinine (P<0.05), ejection fraction < 30% (P<0.001), mitral valve lesion (P<0.05), concomitant CABG surgery (P<0.01), prior cardiac surgery (P<0.01) and reoperation (P<0.01). Increased odd-ratio were related to previous cardiac surgery (5.36; IC95% 0.94-30.56), combined revascularization (5.28; IC95% 1.51-18.36), valvar reoperation (4.69; IC95% 0.93-23.57) and concomitant tricuspid annulosplasty (3.72; IC95% 0.7518.30). Conclusion: The mortality rate is within the parameters found in the literature, identifying recognized factors which neutralization by changes in surgical indication and medical management may enable risk reduction. Descriptors: Cardiac surgical procedures. Risk factors. Prostheses and implants. Heart valve prosthesis. Mortality.


INTRODUCTION
The valve implant surgery represents approximately 20% of all heart surgeries and because of the risk factors found in this population, accounts for 30% of the total mortality rate [1].Mortality documented in the literature for this type of surgery varies according to the hospital, surgical technique of the surgeon and the characteristics of the population, ranging between 1% and 15%, regardless of the type of prosthesis implanted [2][3][4][5].
The need for better assessment of patients undergoing heart valve surgery and its outcome led to numerous prospective and retrospective studies performed in order to estimate the in-hospital mortality, based on preoperative risk factors [6][7][8].Retrospective studies with large numbers of patients are able to better identify the characteristics that might affect the surgery outcome and create models of risk stratification for different institutions [2][3][4][5].

Valve surgery
All patients underwent surgery with cardiopulmonary bypass using a membrane oxygenator and varying levels of hemodilution and hypothermia.Myocardial preservation was performed by using hypothermic St. Thomas II cardioplegia solution.Postoperative care were previously described [9].
From a total of 335 procedures, 216 patients underwent first heart surgery, 76 had previously undergone heart surgery and 43 had undergone two or more heart surgeries.146 (43.6%) isolated implants of mitral mechanical prostheses were performed, as well as 158 (47.1%) of the aortic model and associated implant of mitral and aortic prostheses occurred in 31 (9.3%) procedures.The valve replacement surgeries were combined with 29 (8.7%) coronary artery bypass grafting (CABG) or correction of mechanical defects secondary to myocardial ischemia and 17 (5%) tricuspid annuloplasty surgery.As the type of surgery, 329 were elective and six urgent ones, and 14 patients underwent reoperation during the hospital stay for a valve replacement surgical procedure.

Outcomes and definition of risk factors
It was considered as the primary outcome the death during surgical hospital stay for implantation of mechanical heart valve prosthesis, and the preoperative and surgical variables associated with an increased hospital mortality were considered predictors of risk factors.
The demographic characteristics assessed for mortality during hospital stay were gender, age, atrial fibrillation, ejection fraction, systemic arterial hypertension (SAH), diabetes mellitus (DM), body mass index, serum creatinine, NYHA functional class (according to the model proposed by the NYHA), type of valve lesion (aortic, mitral or mitralaortic) and previous heart surgery.The surgical variables assessed were associated surgery (CABG and tricuspid valve surgery), valve surgery reoperation and type of surgery (elective or urgent).These characteristics are similar to the study by Ambler et al. [2] and were previously used by the authors of this study in studies with valve bioprostheses [10].

Ethical considerations
The research project for this study was submitted to the Research Unit of the Cardiology Institute of Rio Grande do Sul and approved for implementation by the Research Ethics Committee of the institution, and recorded under the protocol No. 4022/2007.During implementation of this project, norms related to patient privacy and confidentiality in the handling of medical information were followed.Postoperative history was obtained from the database of the Surgery Service and hospital records.

Data Analysis
We used univariate and multivariable statistical analysis using SPSS software for Windows, version 14.0, to determine prevalent and independent predictors of hospital mortality risk.For this, we used the chi-square test, Student's t test and logistic regression.In multivariate analysis, the variables were used in the form that presented a higher discriminatory power.All significant characteristics (P<0.05) in univariate analysis were considered for multivariate analysis.We considered risk characteristics those with significant association with hospital mortality for an alpha level of 0.05.
The odds ratio with confidence interval of 95% (OR, 95%) was obtained by logistic regression analysis, to estimate the relative risk of each characteristic assessed.

Hospital mortality
The hospital mortality was 3.88%.Of the 13 deaths, two (15.4%) were secondary to bleeding with cardiac tamponade, six (46.1%) for left ventricle failure, four cases (30.7%) for sepsis (leading to multiple organ failure) and one (7.7%)for renal failure.

Risk factors
The significant demographic characteristics related to increased mortality were ejection fraction below 30% (P<0.001),mitral valve lesion (P<0.041) and previous heart surgery (P<0.01).The serum creatinine was significantly higher in patients who died than in survivors (respectively, 1.15 and 0.97 g/dL, P=0.010).
The surgical characteristics statistically significant correlated with high hospital mortality were coronary artery bypass grafting concomitantly (P<0.01) and reoperation (P<0.05).The hospital mortality was 3.9% for aortic valve replacement, 6.8% for mitral valve replacement and 3.2% for mitral-aortic valve replacement.Table 1 shows the risk for demographic and operative characteristics.

Odds ratio
Odds ratio of the characteristics assessed are shown in Figure 1 and detailed in decreasing order as follows: prior

DISCUSSION
The possibility to quantify the risk factors for patients undergoing valve replacement surgery is studied and known for more than 20 years [11].However, due to advances in surgery and constant progress in noninvasive methods to diagnose this disease, the surgical risk has been decreasing [12].Some of the characteristics of patients which in the past increased the risk of surgery, can now be safely managed.Therefore, as a result of improved perioperative management and high life expectancy, risk factors have been changed [13].
Over the years, the surgical experience has shown that factors such as age, low body mass index, renal failure, low left ventricular ejection fraction, indication for emergency surgery and valve reoperation, among others, contribute to an increased hospital mortality and greater attention from physicians involved in clinical and surgical care of the patients has been given [6][7][8]14].
The accomplishment of heart surgery, regardless of type, was also associated with increased surgical risk, according to Ambler et al. [2].In their study, the authors developed a model of risk stratification for heart valve surgery taking into account characteristics associated with high hospital mortality.The following characteristics were also statistically significant: emergency surgery, age over 79 years and renal failure with need for dialysis.
For Nowicki et al. [15], in a study on independent risk factors for surgical aortic valve replacement, previous heart surgery was also confirmed as a risk factor associated with age over 70 years, small body surface, elevated creatinine, NYHA class IV, previous cardiac arrest, congestive heart failure, atrial fibrillation, emergency surgery and associated CABG.On the other hand, in mitral valve surgery, the following characteristics were statistically significant: female gender, advanced age, diabetes mellitus, CABG, prior stroke, elevated creatinine, NYHA class IV, emergency surgery and congestive heart failure.
Roques et al. [16], in the EuroSCORE study, a predictor of hospital mortality, found that heart surgery and concomitant CABG were associated with high surgical risks.Other variables significantly associated with high mortality were advanced age, creatinine, low left ventricular function, congestive heart failure, pulmonary hypertension, emergency surgery, multiple valve replacement or tricuspid procedure.
The need for reoperation was also described as a factor able to increase the surgical risk and therefore increase the in-hospital mortality.It has been identified previously by Jamieson et al. [3] and by Edwards et al. [17].For the former, other factors also were associated with high mortality rate, such as emergency surgery, renal failure (whether or not on dialysis), low ejection fraction and functional class IV (NYHA).On the other hand, the latter identified as independent risk factors for isolated surgical valve replacement, in addition to reoperation, emergency surgery, renal failure and cardiac arrest.
Brandão et al. [18] presented a study of risk factors for a number of patients who underwent implantation of bileaflet mechanical prostheses with a mitral mortality rate of 13.5% and 7.5% aortic, justified as superior to that found in this study and dued to the inclusion of the outcome of surgeries for repair of aneurysms and dissections of the ascending aorta -known as surgeries of greatest risk.But the increased mortality observed in the group of mitral valve replacement by the authors and also observed in this sample is consistent with the results found in the literature [19,20].
Patients with less severe valve diseases have been accepted due to the tendency to recommend surgery in earlier states of the disease, which can determine a lower prevalence of risk factors and result in lower mortality [21].Possibly, this population -which has younger patientsperforms implant of mechanical prostheses, and the inclusion of patients with severe cardiac or systemic involvement is reduced, resulting in hospital mortality supposedly lower than that observed in the population for the implantation of bioprosthesis [22 ].
To determine the factors that contributed to hospital mortality, some of the risk predictors identified in the studies already cited in the literature [2][3][4][5]14] were considered, focusing on those presented by Amber et al. [2].This attitude is justified by the ready availability of medical information considered by these authors, whose record is part of the hospital routine.We should consider that this approach excludes some known risk factors such as pulmonary arterial hypertension, chronic obstructive pulmonary disease and peripheral vascular disease [3], because such factors are not found in the records of most patients assessed.
The use of odds ratios as a resource for statistical analysis made it possible to estimate the surgical risk separately determined by one of the characteristics assessed without interference from the others [23].The predictors of higher risk in this study were: previous heart surgery (OR 5.36), associated CABG (OR 5.28), the patient required reoperation during the same hospital stay after another valve surgery (OR4, 69), associated tricuspid valve repair (OR 3.72), combined mitral and aortic valve replacement (OR 3.14), diabetes mellitus (OR 2.6), hypertension (OR 2.34), atrial fibrillation (OR 2.14), females (OR 1.96) and LVEF between 30 and 50% (OR 1.61).Of the characteristics mentioned above, the OR was increased in study by Hellgren et al. [21] for atrial fibrillation in procedures for combined aortic or mitral valve replacement (OR 4.1).However, other factors mentioned by the authors, such as valve replacement over the age of 70 years (OR 2.1) and NYHA functional class IV (OR 2.2) were not identified in our study, possibly due to differences in the populations approached, especially with regard to age.
In a study performed in this institution [10], with patients who had undergone implant of bioprosthesis, increased odds ratio was observed for associated tricuspid surgery (OR 3.71), combined mitral and aortic surgery (OR 2.86) and females (OR 2.43).Other characteristics with higher OR, such as emergency surgery, dialysis, age over 80 years and two or more previous surgeries were not highlighted in this study.Also, it is possible that the difference in age of patients included in the two studies has influenced in the findings.A study comparing results with implantation of bioprostheses and mechanical prostheses in the same institution may be useful in clarifying the differences in mortality and risk factors associated with different valve replacement implants.

CONCLUSION
It is expected that operative risk factors identified in the study, which were ejection fraction, mitral valve disease, high serum creatinine, previous heart surgery and surgery during hospital stay for heart valve surgery -also identified by other authors and already emphasized in this discussion [2][3][4][5], have their role minimized in the risk, either by changing the criteria of indication (as higher precocity), better clinical compensation or change in operative routines.

Table 2 .
Hospital mortality according demographic characteristics of continuous value DE BACCO, MW ET AL -Risk factors for hospital mortality in valve replacement with mechanical prosthesis