IncIdence and factors assocIated wIth perIcardIal effusIon after cardIac valve surgery

Methods: We retrospectively reviewed medical records of all consecutive patients submitted to cardiac valve surgery in a tertiary teaching hospital from January 2012 to July 2014, where echocardiography was routinely performed before patient discharge. Moderate-to-severe PE was defined as ≥ 10 mm of thickness, or signs of cardiac tamponade on echocardiography. Additional clinical and perioperative data were extracted from medical records using a standardized protocol.


Pericardial effusion (PE) is a postoperative complication of open-heart surgery.
It may lead to cardiac tamponade, early readmissions and reinterventions, and increased mortality rates 1 .The reported incidence of postoperative PE varies from 1% to 77% 2,3 depending on the type of surgery, clinical characteristics, and the criteria used for identification and quantification.Cardiac valve surgery is one of the procedures most related to the development of PE 3 .A higher risk for postoperative PE has been associated with higher body surface area, pulmonary thromboembolism, hypertension, renal failure, immunosuppression, emergency surgery status, type of cardiac operation other than coronary artery bypass grafting, and prolonged cardiopulmonary bypass 4,5 .As a consequence, identifying predictors for the development of PE after cardiac valve surgery may help to prevent or reduce its occurrence.

Pericardial effusion after cardiac valve surgery
In this study, we aimed to determine the incidence of moderate-to-severe PE after cardiac valve surgery and the risk factors associated with this condition in a tertiary teaching hospital.

METHODS
In this cross-sectional analysis, we studied patients submitted to cardiac valve surgery who underwent predischarge echocardiographic examination between January 2012 and July 2014, in the Hospital de Clínicas de Porto Alegre.The institutional review board approved the study and waived written informed consent.
Data were extracted from medical records and the Department of Cardiac Surgery administrative registry using a standardized form.We collected preoperative data about previous comorbidities, medication use, blood tests, imaging exams, and surgical indication; intraoperative data about surgery time and surgical technique; and postoperative data about thoracic drainage, medication use, atrial fibrillation, stroke, severe bleeding, reinterventions, readmissions, and mortality.
Of the 353 identified patients, 335 were included in this study (16 patients died and two patients did not undergo a predischarge echocardiography).Patients were categorized into two groups: moderate-to-severe PE and absent or small PE (figure 1).Moderate-to-severe PE was defined as ≥ 10 mm of thickness 6 , or signs of cardiac tamponade on echocardiography.Patients who did not match these criteria were classified as absent of small PE.

Statistical Analysis
Data were presented as mean and standard deviation or median and interquartile range (IQR) (continuous data) or as count and proportion (categorical and ordinal data).Continuous variables were compared using the independent-samples t test and categorical variables were compared using the chi-square test.After univariate screening, multivariable linear regression was used to adjust for selected clinically and statistically significant pre-and intraoperative covariates (age, sex, history of stroke, preoperative international normalized ratio (INR) levels, and surgery time).
All statistical analyses were performed with PASW 18.0 (SPSS Inc, Chicago).All tests were two-sided and p-values of < 0.05 were considered statistically significant.

Patients
The majority of echocardiographic examinations were conducted up to 6 days of after surgery (median: 6.0 days; IQR: 5.0-7.0).From January 2012 to July 2014, moderate-to-severe PE occurred in 27 (8%) of 335 patients submitted to cardiac valve surgery (mean age: 62; standard deviation 12 years; 70% male).Five of those presented with echocardiographic signs of increased intrapericardial pressure or clinical cardiac tamponade.Patients with moderate-to-severe PE had more history of stroke and asthma than absent or small PE patients, but were similar to those with respect to other preoperative clinical and echocardiographic characteristics (table 1).The group with moderate-to-severe PE had lower hemoglobin and hematocrit levels, higher INR, and a higher prevalence of INR greater than 2 compared to the absent or small PE group (table 2).The association between moderate-to-severe PE and preoperative INR levels remained statistically significant after adjustments.

Surgery
The most performed surgery was aortic valve replacement (70%) and the most common indication for the procedure was aortic valve stenosis (55%).Ninety-three (28%) patients were submitted to multiple procedures (for example, aortic valve replacement and coronary artery bypass graft surgery) during the surgery.Patients with moderate-to-severe PE had longer surgeries, as well as longer periods of ischemia and cardiopulmonary bypass (table 3).Surgical status (elective or urgent/emergent surgery) did not differ between groups, nor did the proportion of patients submitted to multiple procedures.

Postoperative Parameters
Postoperative data are presented in Table 4. Patients with moderate-to-severe PE had longer periods with thoracic drain than patients with absent or small PE, with a mean difference of 5.3 hours (33.6 hours vs. 28.3hours; p = 0.001), though the total drained volume

DISCUSSION
In this study, the incidence of moderate-to-severe PE following cardiac valve surgery was 8%.We found that patients who develop moderate-to-severe PE had a higher prevalence of previous stroke, higher levels of preoperative INR, longer surgery times, and higher prevalence of new and recurrent postoperative atrial fibrillation than patients with absent or small PE.Moreover, the group with moderate-to-severe PE underwent more reinterventions (for bleeding and for other causes) and had a greater number of deaths.
The high variability of the reported PE incidence after cardiac surgery is related to the different definitions for this condition 3,7,8 .Our study used only echocardiographic parameters to define PE, while others used clinical presentation and echocardiographic data 7 ; however, our incidence of PE is in the lower range of published incidence rates.This may be attributed to the fact that echocardiograms were systematically performed in all patients, regardless of symptoms.

Preoperative Considerations
We identified that history of stroke was a risk factor for the development of moderate-to-severe PE.This association could be attributed to a higher use of antiplatelet agents and anticoagulants aiming at secondary prevention.Although not statistically significant, the prevalence of anticoagulated patients was almost twice as high in the group with moderate-to-severe PE, which is likely to be clinically relevant.Besides that, higher levels of preoperative INR was the most important factor associated with moderate-to-severe PE.Some studies have found association between the previous use of immunosuppressive therapy and the development of postoperative PE 4,9 , although we only had two patients taking these drugs.The protective association between history of previous cardiac surgery and PE reported in other studies 4,10,11 was not present in our analysis.

Intraoperative Considerations
Patients with moderate-to-severe PE were more likely to have longer surgeries, with longer periods of ischemia and cardiopulmonary bypass.This finding has already demonstrated to be an important risk factor for PE 3,12 , which may be related to changes in coagulation parameters during these periods.We did not find differences in the types of cardiac valve replacement and the type of prosthesis (mechanic or biologic).Some authors have shown that aortic root and aortic aneurysm surgery are independent risk factors for the development of significant postoperative PE due to prolonged cardiopulmonary bypass and extensive dissection of the heart and the aorta 4,13,14 .In this study, we analyzed the combination of aortic root with cardiac valve replacement, but we did not find a higher risk to develop moderate-to-severe PE in this subgroup.

Postoperative Considerations
Patients with moderate-to-severe PE presented higher rate of postoperative atrial fibrillation than absent or small PE patients, similarly to the results described by Ikäheimo et al. 8 Arrhythmia can be a marker of higher risk of adverse outcomes in these patients 15,16 which may have mediated a higher incidence of clinical endpoints regardless the presence of PE.
Aspirin reduces the risk of death and ischemic complications 12 after coronary artery bypass grafting, but it is associated with a higher risk of postoperative PE 3,17 .In our analysis, patients with moderate-to-severe PE used less aspirin and had a similar use of anticoagulants compared to absent or small PE patients.This finding reinforces the safety of early postoperative aspirin administration regarding PE risk.
The efficacy of anticoagulation after cardiac valve surgery depends on a delicate balance between the risk of thromboembolic events and risk of bleeding, and the appropriate time to start these drugs is not clear 18,19 .Although PE is one of the main concerns regarding proper anticoagulation initiation, our data Pericardial effusion after cardiac valve surgery are not enough to suggest a specific recommendation in this aspect.
The association between the time of chest tube removal and PE is a controversial topic.Most of our patients had their chest drainage removed on the second postoperative day.Some authors showed that there is no difference between early (< 24 hours) and late (≥ 24 hours) chest tube removal 20,21 .Chest tube withdrawal delay would increase the risk of infection and would cause more discomfort to patients 22,23 , but its early removal could predispose to the development of PE 24 .In our study, we did not find significant difference between the groups with respect to time of chest tube removal.
The development of moderate-to-severe PE was related to higher risk of reinterventions and death.The causes of deaths were related to cardiogenic shock and septic shock.Only patients from the group of moderate-to-severe PE presented cardiac tamponade.Although, Kuvin et al. 3 reported a high rate of cardiac tamponade after postoperative moderate-to-severe PE (74%), we found a rate of 11% among patients with moderate-to-severe PE.This discrepancy may be due to differences in the definition of cardiac tamponade.

Study Limitations
Some limitations of this analysis should be noted.Although most of our results are consistent with what was reported in the literature, our sample of moderate-to-severe PE patients was small and our study was based on a retrospective analysis of medical records limited to one study center.We analyzed only cardiac valve surgeries, since all these patients should have undergone an echocardiographic examination before hospital discharge as a routine of our service.Additionally, we lack information about the high risk group, since we excluded patients who died before performing the hospital discharge routine.

CONCLUSION
In our study, the incidence of moderate-to-severe PE in patients submitted to cardiac valve surgery is not low, but still in the lower range of previous reports.This condition may be associated with preoperative coagulation state, and also with prolonged surgery and postoperative higher morbidity and mortality.Identifying risk factors associated with moderate-to-severe PE may help to better prevent or reduce the occurrence of this condition.
CPB: Cardiopulmonary Bypass.Data are presented as mean ± SD or n (percentage).