A 10-year institutional review of surgery for structural valve dysfunction in the developing world.

Abstract

Background Prosthetic heart valves do not fulfil the requirements for an ideal valve, resulting in the development of prosthetic dysfunction or complications over time. Structural valve dysfunction may be influenced by multiple components which include patient’s factors, valve related factors and intraoperative factors. The inter-relation of these factors has a significant impact on morbidity and mortality associated with reoperative surgery for prosthetic valve dysfunction, particularly in a developing world where a large burden of communicable diseases together with lack of health care resources affect surgical outcome. In this study we examined the clinical records of the patients who underwent reoperative valve surgery to evaluate the clinical profile and factors that affect the surgical outcome after reoperation at a large tertiary referral center in a developing country. Objectives 1) To describe the demographic profile of patients with malfunctioning prosthetic heart valves and define their clinical presentation 2) To describe the clinical presentation of valve dysfunction 3) To determine the possible mechanisms of mechanical and bioprosthetic valve failure 4) To determine the factors affecting the immediate surgical outcomes in subjects undergoing redo cardiac prosthetic valve surgery. Materials and methods A retrospective analysis of the clinical, perioperative and follow-up data of patients who underwent redo cardiac valve surgery for structural valve dysfunction between January 2005 and December 2014 at Inkosi Albert Luthuli Central Hospital, Durban, South Africa was undertaken. Patients were identified using the Speedminer software program which is a Data Warehouse software package used to store data collected on the hospital Medicom database. The file of each of patient who underwent redo cardiac prosthetic valve replacement for structural valve dysfunction was accessed and data were extracted on age, gender, potential risk factors for valve thrombosis, symptomatology, investigations including International normalized Ratio (INR) status and follow up. All patients were evaluated preoperatively by the cardiologist and the cardiothoracic surgical team and submitted for either an elective or emergency valve replacement. Excluded from the study were those patients who underwent cardiothoracic surgery for nonvalvular reasons, i.e. coronary artery bypass surgery and congenital heart disease. Results During the ten year period (2005 to 2014) 2618 valve replacement operations were performed. During the same period 128 reoperations (4.9%) were performed in 113 patients (mean age 35.59 (SD±16.66) years). The majority of the patients were Black (72.6%) and female (75%). Fifteen patients (13.3%) were HIV infected and nine were pregnant. Acute dyspnoea (NYHA class III 34.37% and class IV 21.88%) was the presenting feature in 72 patients (56.25%). Clinical presenting features of an obstructed valve (flash pulmonary oedema with or without clinically audible prosthetic valve clicks) were documented in the clinical records of 44 of the 128 (34.4%) reoperations. In seventeen instances subjects presented with acute onset of cardiac failure (13.3%) and in eleven the presentation was characterised by signs of low cardiac output state (5.3%). There were no clinical indicators of an obstructed valve in the remaining 56 (43.8%). Of these 56 patients: 38 presented with change in effort tolerance and 18 where asymptomatic. Valve dysfunction was detected by echocardiography and confirmed fluoroscopically in 71/128 cases (55.47%). In the remaining patients the diagnosis was made either at fluoroscopy (11.72%) or on echocardiography (32.81%). The ejection fraction (EF) was severely impaired (EF<40%) in 7.08% of patients. The mean left atrial size was 52.28mm and mean pulmonary systolic pressure 45mmHg (range 26-104). Mechanical valve dysfunction was documented in 110/128 reoperations (obstructed valve (100) and prosthetic infective endocarditis (10). In almost two thirds of instances with obstructed mechanical prostheses levels of anticoagulation achieved were poor (INR<2.0); 30/110 (27.27%) were within therapeutic ranges of 2-4 and 9/110 (8.18%) was >4.0. HIV status did not influence the outcome of surgery and did not appear to be the main mechanism of valve obstruction. The bioprosthetic valve group comprised the remaining 18 of 128 reoperations. In this group 13/18 patients had structural valve deterioration with periprosthetic leaks, and remaining 5 had prosthetic infective endocarditis (aortic root abscess (1) and annular dehiscence (4).Emergency surgery was performed in 54.7% of the study population, of which 60.2% were in the mitral position. There was a total of 13 early in-hospital deaths (11.5%) of which one “on table” death was due to a low cardiac output state (LCOS). Postoperative mortality was related to prosthetic endocarditis (5/13) and high grade dyspnoea at presentation (7/13). Multivariate analysis revealed that bypass time >3.5 hours (HR 5.58, 95%CI 1.24-24.95), cross clamp time >120 minutes (HR 4.48, 95%CI 1.25-18.73), and third time redo operations (HR 4.26, 95%CI 1.23-14.75) were the independent predictors for early in-hospital mortality. Conclusion Our study shows a 4.9% reoperation rate after the previous valve replacement surgery with 11.5% perioperative mortality. Our results confirm that reoperative surgery is associated with significant morbidity and mortality. More than half the patients presented acutely for mechanical valve obstruction which was due to inadequate levels of anticoagulation and required emergency surgery. Early mortality was related to poor NYHA class at presentation and to the presence of infective endocarditis. An important finding of this study was the high rate of valve obstruction associated with poor anticoagulation in patients who received the Cryolife On–X valve. They had a shorter interval to valve obstruction requiring redo valve replacement compared to the other mechanical prostheses.