| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
Interact CardioVasc Thorac Surg 2009;9:241-245. doi:10.1510/icvts.2008.197715 © 2009 European Association of Cardio-Thoracic Surgery
Short- and long-term outcomes of surgery for active infective endocarditis: a Tunisian experience
a Department of Cardiology, University Hospital Hédi Chaker, Sfax, Tunisia Received 28 October 2008; received in revised form 8 April 2009; accepted 14 April 2009
*Corresponding author. Tel.: +21674292165; fax: +21674243992.
From January 1997 to December 2006, all patients with a Duke criteria-based definite diagnosis of infective endocarditis (IE) operated on during the active phase in a Tunisian high volume tertiary-care centre were included. Among the 186 patients with IE identified during the study period, 88 (48.35%) required surgery in the active phase. Mean age was 34.9 years, 54 (61.4%) were men. The infected valve was native in 70 cases (79.5%) and prosthetic in 18 (20.5%). Streptococcus sp. were the most common causative microorganisms. The most frequent indication for operation was congestive heart failure. There were 24 in-hospital deaths (27.27% early mortality). By multivariate analysis, severe congestive heart failure (HR=13.82, 95% CI [3.38–38.15], P<0.001) and large >15 mm vegetations (HR=6.02, 95% CI [1.48–18.52], P=0.03) were predictive of in-hospital mortality. Survivors were followed-up from 3 to 120 months, mean of 28.6. Actuarial 5- and 10-year survivals free from the combined endpoint of recurrent IE, cardiovascular death and late surgery in survivors were 69±5% and 63±7%, respectively. In conclusion, despite medical progress, surgery for endocarditis in Tunisia remains challenging and yields high mortality rates. Severe heart failure is the most powerful predictor of mortality. Long-term outcome is, however, satisfactory.
Key Words: Infective endocarditis; Mortality; Predictors
Despite significant progress made in diagnostic and therapeutic modalities, infective endocarditis (IE) remains a serious affection associated with considerable mortality and morbidity [1]. Although antibiotics still represent the cornerstone of management of patients with IE, surgery could prove mandatory especially in those with uncontrollable infection and those developing heart failure or other complications related to the disease progress [2]. Numerous papers from developed countries have already addressed the issue of surgical management in the active phase of IE [3–6]; however, there has been no meaningful publication from developing countries. In this study, we present a series of 88 patients operated on for definite IE in a 10-year period in a Tunisian high volume tertiary-care centre (Sfax, Tunisia, North Africa) and we investigate their profile and outcome.
2.1. Population All the patients who were admitted to the cardiology and infectious disease departments of a single high volume tertiary-care medical centre (the Hedi Chaker University Hospital of Sfax, Tunisia) with a Duke criteria-based diagnosis of IE from January 1997 to December 2006 were retrospectively evaluated. The hospital database and a systematic review of all reports from the echocardiography laboratory were used to identify eligible patients. One hundred and eighty-six patients with definite IE according to the modified Duke's criteria [7] were identified. Among them, 88 needed surgical intervention in the active phase and were included in the present study. IE was considered active if the surgery was deemed necessary before completion of a standard course of antibiotic treatment, irrespective of whether there were ongoing signs of sepsis or whether the blood cultures, surgical specimens, or both were positive for the infecting organism. Prosthetic valve endocarditis (PVE) was defined as infection occurring on any type of tissue or mechanical valve device. PVE were called early PVE when occurring within 12 months after surgery, beyond, they were called late PVE. Early operation was defined as surgical intervention within seven days after diagnosis. Blood cultures were systematically collected at admission (3–6 sets of aerobic and anaerobic blood cultures were typically performed), and serological assessment was completed for Coxiella burnetii, Bartonella sp., Chlamydiae sp., Legionella, Brucella, Mycoplasma pneumoniae, Candida sp., and Aspergillus sp. for episodes with blood cultures that remained negative. In addition, since 2001, valve cultures and polymerase chain reaction were performed on explanted valves. Effective treatment included medical and surgical treatment; indications for urgent surgery were severe valvular dysfunction with heart failure, abscess or perivalvular extension, large vegetations with high risk of embolization, or recurrent embolization during antibiotic treatment, failure of conservative medical treatment. For each patient, a detailed description of antibiotic regimen was obtained. Likewise, surgical management was detailed, as well as the short- (in-hospital mortality) and long-term outcome. Late outcomes observed were recurrences, need for surgery and deaths. Categorical variables are reported as percentages, and continuous variables are reported as means±standard deviations (S.D.). Statistical comparison between subgroups was tested with unpaired t-test for continuous variables and 2-test for categorical variables. Fisher's exact test (two-tailed) was used if the expected count in any cell was <5. Multivariate stepwise logistic regression analysis was used to identify independent variables predicting per and postoperative in-hospital mortality. Only variables significantly related to death on univariate analysis were taken into account for the stepwise forward multiple regression analysis (the Hosmer–Lemeshow goodness-of-fit statistic test was used to assess the model fit). Cumulative survival and event-free periods were estimated by the Kaplan–Meier method. Differences in survival at follow-up were estimated by log-rank tests. The statistical analysis was performed with the SPSS 15.0 statistical package (SPSS Inc, Chicago, Illinois, USA). The significance level used in univariable and multivariable analysis was set at P<0.05. All P-values were two-sided.
3.1. Patients During the study period, 186 patients were admitted with Duke criteria-based diagnosis of IE, 88 (48.3%) underwent surgery during the active phase of the disease and constituted the study population. Mean age was 34.9 years (range 16–68). There were 54 males (61.4%) and 34 females (38.6%). IE occurred on native valves in 70 cases (79.5%) and on prosthetic valves in 18 cases (20.5%), all of them were mechanical prostheses. Rheumatic valvular disease was the most common underlying heart condition among patients with native valve IE (n=43, 48.9%). Detailed features of the population are presented in Table 1.
3.2. Surgery Indications for surgery during active IE were one or more of the following: congestive heart failure in 46 patients (52.3% of the whole), persistent signs of septicaemia despite antibiotic treatment in 11 patients (12.5%), recurrent embolisation in 7 patients (8%), vegetation size >15 mm in 18 patients (20.4%), and annular abscess in 11 patients (12.5%). Total percentages are >100% due to combined indications. Mean delay from admission to surgery was 16.2±12.4 days (range 0–52 days). Total duration of antibiotic therapy was 45.1±21.2 days. Blood cultures were positive in 45 episodes (51.1%). Patients with negative blood culture results received prior antibiotic treatment in 25 cases (55.5%). In 17 cases, causative microorganisms were identified either by serology (13 cases) or leaflet culture or polymerase chain reaction (4 cases). Thus, infective agent was identified in 62 cases (70.4%). Streptococci (n=23, 26.1%) were the most frequent microbiological finding followed by Staphylococci (n=19, 21.6%) and Bartonella (n=6, 6.8%). Detailed microbiological data are presented in Table 2.
3.4. In-hospital mortality Twenty-four patients died perioperatively before hospital dismissal (overall in-hospital mortality was 27.27%). The causes of death were cardiogenic shock in five patients, perioperative death in seven, septic shock in four, prosthesis dysfunction in six and multiorgan failure in two. Factors affecting morality, on univariate analysis are displayed in Table 3.
By multivariate analysis, severe NYHA IV congestive heart failure (HR=13.82, 95% CI [3.38–38.15], P<0.001) and large >15 mm vegetations (HR=6.02, 95% CI [1.48–18.52], P=0.03) were predictive of in-hospital mortality (operative and early postoperative mortality). (The Hosmer–Lemeshow test significance value=0.83.) Complete follow-up was possible in 61 of the 64 surviving patients (95.3%). Mean follow-up was 28.6±31.2 months (range 3–120 months). Recurrences occurred in 7 among the 64 survivors with a mean delay of 37 months, three of them within the first postoperative year (early PVE) and were all successfully managed medically. Late surgery was needed in four patients for prosthetic valve dysfunction, none of these died at re-operation. Four late cardiovascular deaths (6.25% late mortality rate) were recorded (2 arrythmias and 2 severe heart failures). The combined endpoint of recurrent IE, cardiovascular death and late surgery occurred in 15 patients (23.45% of the survivors). Actuarial 5- and 10-year survivals free from the endpoint in patients surviving the postoperative period were 69±5% and 63±7%, respectively (Fig. 1).
Although none of the studied factors proved to be significantly related to the occurrence of the composite end-point, a trend (log-rank P=0.066) towards a higher occurrence was noted among patients operated on within seven days (early surgery group).
To our knowledge, our data represent the largest series describing characteristics and outcome of surgery for active IE using modern diagnosis and therapeutic tools in a developing country. We found an important early postoperative mortality rate exceeding 27% but a satisfactory long-term outcome for the survivors, indeed, we registered a low late mortality rate (6.25%) and an excellent 63% event-free actuarial survival at 10 years. It is noteworthy that all our operated patients were implanted mechanical prosthetic valves. The exclusive use of mechanical prosthesis was imposed by constraints of homografts availability in our centre, and generally in our country, rather than a deliberate institutional choice. Nonetheless, we do not believe that such an attitude negatively affected our results, as superiority of homografts over conventional prosthesis in the setting of active IE has never really been established. Indeed, although several publications [8] reported promising results with homografts fulfilling all theoretical requirements in the particular context of IE, namely, ability of restoring the aortic root anatomy and of providing best haemodynamic profile with no need for anticoagulation, intrinsic resistance to infection, and better penetration of antibiotics, more recent investigations [9] failed to show any superiority for homografts. Moreover, uncertainties remain mainly concerning their durability, the reality of their resistance to IE recurrences or their viability raising serious concerns about their place [9]. Consequently, as aforementioned, we do believe that our exclusive use of mechanical prosthesis cannot be considered as a limitation to our work. 4.2. Early mortality and predictors As previously stated, we registered a considerable early mortality rate. Our results are, however, similar to those in a recent American series by Tleyjeh et al. [10] reporting a 27% mortality rate in the surgical group of their population and to those in a series by Guerra et al. [11] who found a 30.5% operative mortality but are higher than the majority of other Western series generally reporting rates below 20% [3, 5, 12].Our study showed that independent predictors of in-hospital mortality were congestive heart failure and >15 mm vegetations, results that are similar to those we previously reported in native valve IE [13]. Heart failure was comprehensively shown to be a major determinant of mortality in the surgical management of IE [4–6] with up to 6-fold increase in mortality in patients with cardiogenic shock in a series by David et al. [5]. The impact of timing of surgery on outcome is an intensely debated issue with conflicting evidence [14]. In our study, we found that early surgical intervention within seven days of the diagnosis was significantly associated with worse outcome on univariate analysis, with almost two-fold higher mortality but not on multivariate analysis. Our data join those reported by Gammie et al. [15] reporting 14.9% vs. 4.9% for active vs. healed IE. Those findings are, however, to be cautiously interpreted given that authors compared active vs. treated IE which, we assume, is at least slightly different from comparing two subgroups of active IE. Hill et al. [12] using our seven days' cut-off for early surgery found a four-fold higher mortality for early surgery despite statistically non-significant differences. When compared with data from developed countries, long-term outcomes, in our series, seem satisfactory with limited recurrences and few late cardiovascular deaths. In their series of mechanical prostheses for treatment of active IE, Guerra et al. [11] reported an actuarial survival free from death, re-operation, and re-infection of 73.1% at 5 years and of 59.8% at 10 years. D'Udekem et al. [4] noted a 61±6% actuarial survival at 10 years, and at 15 years, survival in a series by David et al. [5] was 59%.Generally, it appears that, despite the persisting early problems in treating patients with IE with the perioperative period constituting a critical high-risk phase, the long-term survival is satisfactory. We demonstrate that this paradigm, developed throughout numerous Western studies is actually universal, as it also applies to the totally different context of our developing African country.
Several limitations of this study are to be highlighted. Firstly, it was retrospective and lacked a randomized controlled treatment strategy, so we were unable to control for selection bias in medical vs. surgical treatment. Secondly, this study was conducted in one tertiary-care institution, resulting not only in limitations to the generalization of the results but also in unavoidable referral bias. Owing to the relative low number of patients included in this study, some risk factors might not have reached statistical significance.
We believe that our study provides a meaningful insight to the surgical management of IE in developing countries with inherent material limitations. It underlies several discrepancies with Western series, mainly regarding epidemiological features but also reveals a lot of similarities. Several conclusions can be drawn: despite undeniable progress in antibiotic therapy and cardiosurgical management, in our region, the operative mortality remains high. Severe congestive heart failure and large vegetations constitute independent predictors of early mortality. For patients surviving the critical perioperative period, long-term outcomes are very satisfactory.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
Top
Abstract
1. Introduction