PATIENTS AND METHODS

We reviewed the records of 23 malfunctioning mitral valve patients who underwent surgery via right thoracotomy from July 15, 2000, to August 30, 2012 at Heart Hospital, in Tehran, Iran. They represented 18% of 125 operations for prosthetic valve malfunction performed in the same period.
The clinical characteristics of these patients, consisting of 13 women and 10 men are presented in Table 1. The mean age of patients was 43.8 years (median 33; range 20-55 years). All the patients had associated diseases. On admission, no patient was moribund and none required external chest compression for cardiopulmonary resuscitation, although 67% of the patients had hemodynamic instability. One patient was unconscious on admission, and 2 had a preoperative stroke. Renal and respiratory insufficiency was present before admission in 29% and 70% of the patients, respectively. A total of 57% of the patients had a reduction in renal function preoperatively, which was largely attributable to hemodynamic instability.

Table 1: Clinical data

Ischemia was encountered in only one patient. Moderate to severe chronic cardiac failure was present before admission in 29% of the subjects.

No patient was refused surgery because of age or concomitant disease. On admission, all the patients had severe coagulation disorders. Platelet aggregation (aggregometry) and homeostasis (thromboelastography) were markedly reduced in all the patients. In addition, variable degrees of thrombocytopenia, low levels of the thrombin inhibitor angiotensin III, and coagulation factors (fibrinogen, V, VU, \mi, and DC), and high levels of plasmin degradation of cross-linked fibrin (D-dimer) were measured. All procedures were performed with emergency status within 24 hours after the initial onset of dyspnoea on exertion and chest pain, and all the patients were operated on within 4 hours of their arrival. The operative techniques used are summarized in Table 2.

Click here for Table 2: Operations

Operative Technique
All the operations in this series of 23 cases were performed through right thoracotomy incision according to previously described standard surgical techniques. In the last case, ascending aortic cannulation was done instead of femoral cannulation, which was due to a history of recent bilateral femoral artery embolectomy procedure. Preoperative coagulation disorders were aggressively treated from the time of admission with the aid of blood bank products, antithrombin III, aprotinin and coagulation factor concentrates.

Continuous ultra filtration during CPB was occasionally used as was a cell saver device (Haemonetics Corp., Braintree, Mass). Arterial pressure monitoring lines were routinely placed in radial arteries. Systemic anticoagulation for CPB was accomplished with heparin at an initial dose of 3 mg/kg body weight. Adequate heparinization for CPB was assessed with the activated clotting time (Hemochron FTK-ACT; International Technique Corp., Edison NJ), with further dose of heparin administered as required so as to maintain an activated clotting time greater than 750 seconds. At the end of the procedures, heparin was neutralized with protamine (0.8 mg of protamine per milligram of heparin). CPB was established with cardiac indices of 2.0, 2.5 L. Min-1 m-2. Cooling was stopped at 33° C in patients operated on without the cross-clamp technique under mild hypothermia. Arterial inflow adequacy was assessed clinically. Acid-base management during mild hypothermia was via an alpha-stat approach. Before opening the left atrium, the head-down position (Trendelenberg) was established and immediately after opening the left atrium, provision was made to render the mitral valve incompetent so that the left ventricle would pump into the left atrium instead of the aortic root.

The postoperative homeostasis protocol used was derived from that proposed by Szefner, and included tests for platelet function, thrombin formation and its regulatory pathways, and for the fibrinolytic system. Treatment included the administration of small doses of protamine, modulated doses of aprotinin, fresh frozen plasma as well as fibrinogen. Dosages were adapted to each patient’s clinical profile as well as to test interpretation criteria in order to provide personalized treatment.

Statistical Methods
Medical records of all the 125 patients operated on for malfunctioning mitral valve between July 18, 2000, and August 30, 2012, were reviewed and preoperative variables that were believed to have an impact on outcome were obtained (Table 1). Data were studied by descriptive and statistical analyses. The SPSS application software version 10.0 (SPSS Inc., Chicago, ILL) was used for statistical analyses.

Significant differences between the two procedure groups (i.e. median sternotomy and right thoracotomy without aortic cross-clamping) were assessed with univariate analysis (Tables 1 to 3): categoric data were compared by means of the x2 test or the Fisher exact test (as appropriate) and continuous variables with the student t test. Data were further analyzed by univariate and multivariate logistic regression with the response variable on the one hand being hospital mortality (30 days mortality), and on the other hand, morbidity. All variables with a zero or near zero cell count were treated as continuous so that convergence could be obtained. Multivariate models were applied to isolate the effect of each factor adjusted for all other factors: the 0.25 level (P< 0.25) was used as a screening criterion for the selection of candidate variables. In order that problems created by multi collinearity could be avoided, variables that were too highly correlated among them were included at different times in different models. A background procedure was used. A Wald x2 test was used for testing the significance of individual coefficients. The results are shown as odds ratios with 95% confidence intervals.

RESULTS

Descriptive Analysis

Mortality: Overall hospital mortality was 13% (three patients, Table 3).

Click here for Table 3: Complications

ntra-operative mortality was 9% (12 patients). Intra-operative deaths included all 9 patients with bleeding from coagulation disorders. The bleeding was due to the poor quality of the ventricular tissue at the site of adhesional release in 9 patients and to the technique of repair in the other patients.

Mortality from cardiac causes was 12% (15 cases); 12 patients could not be weaned off the CPB because of intractable cardiac dysfunction and 2 died postoperatively of low cardiac output on postoperative days 8 and 15 respectively. Permanent neurologic damage with cerebral death occurred in 9 (7%) patients. Sepsis complicated by multi-organ system failure (specifically renal and respiratory failures), was the cause of death in 10 (8%) patients. Multiple reoperation for bleeding were required in 9 (8%) other patients. Intravascular disseminated coagulation (DIC) in its different phases was present in all patients from the time of admission. Despite aggressive management and routine blood coagulation tests, coagulation factor levels and fibrinol: vsin remained abnormal for the duration of CPB. Factor V, factor VIII, Fibrin degradation products (FDP), antithrombin III, D-dimer levels, and platelet counts were the most evident defects before, during, and after CPB and in the postoperative period.

Prolonged postoperative inotropic support for low cardiac output was required in 34 patients. Respiratory failure necessitating tracheostomy for prolonged assisted ventilation occurred in 39 patients, of whom 9 died of sepsis before hospital discharge. Stroke with neurologic damage complicated the postoperative course of 8 (7%) patients.

Five patients required hemodialysis treatment. The presence of renal failure in the postoperative period was associated with death in all patients with this complication. The mean duration of hospital stay was 38 days with an average of 10 days in ICU.

Late Mortality:
Six patients who had preoperative moderate COPD required prolonged ventilator weaning and were discharged from ICU from 32 days to 44 days afterwards. Of these, 3 died of pneumonia later on and the remaining succumbed to malignant arrhythmias during re-hospitalization later on for respiratory failure. One patient who had post operative renal failure died of cardiac arrest during dialysis 5 months after the operation.

Statistical analysis
The two procedure groups examined (right thoracotomy vs. median sternotomy) were not significantly different from each other with respect to the majority of perioperative characteristics that are believed to have an impact on outcome in patients with malfunctioning prosthetic mitral valve (Tables 1 and 2). As indicated by univariate analysis, perioperative patient variables correlated to operative mortality were median sternotomy (P<0.001), bleeding (P<0.001), and cardiac ischemia (P=0.02). For morbidity, univariate analysis indicated the following significant variables: median sternotomy (P<0.0001), and preoperative myocardial ischemia (P=0.05). Mortality and morbidity models based on multivariate logistic regression analysis confirmed that median sternotomy (as a continuous variable) is by far the most important risk variable.