Treatment and Prophylaxis of Post-pericardiotomy Syndrome in Cardiac Surgery Patients: A Meta-analysis of Main Trials


 Purpose: Post-pericardiotomy syndrome (PPS) is a common complication of cardiac surgery. This study represents a meta-analysis of trials on treatment and prophylaxis of PPS with colchicine, indomethacin and dexamethasone. The primary endpoint was to investigate the efficacy of these drugs, while the secondary endpoints were the efficacy in reducing hospital readmission, cardiac tamponade, symptom persistence, atrial fibrillation as well as their safety and adverse effects. Methods: Literature research was carried out using PubMed. Inclusion criteria were: studies investigating ≥10 patients with clinically PPS in which colchicine, dexamethasone and indomethacin were compared with placebo. We excluded publications referring only to animal experiments or in vitro experiments, studies on < 10 patients, case reports, congress reports, review articles, editorial letters. Results: Among studies with post-operative colchicine treatment, two of them demonstrated a significant reduction of PPS [95% CI -0,86 (-1,96, 0,25)], [95% CI 0,24 (0,03,0,44)]. In the single pre-surgery colchicine administration study, a reduction of PPS was obtained [95% CI -0,55 (-1,04, -0,06)]. In the indomethacin study, pre-surgery administration was linked to a reduction of PPS [95% CI -2,18 (-4,31, -0,06)]. This result was not reported with pre-operative dexamethasone intake [95% CI 0,50 (0,12, 0,88)]. Conclusion: The overall results show a better outcome when colchicine and indomethacin were administered as primary prophylactic agents in order to prevent PPS and PE. Further studies are needed to confirm these results.


Introduction
Post-pericardiotomy syndrome (PPS) is a common complication affecting the pericardium and pleurae following cardiac surgery, associated with signi cant morbidity and prolonged inhospital stay. Although the pathophysiology remains unclear, surgical trauma and cardiopulmonary bypass (CPB) are thought to play an important role by triggering a systemic in ammatory response syndrome (SIRS) [1]. For this reason, nonsteroidal antiin ammatory drugs (NSAIDs), steroids and colchicine are the main pharmacological agents used to treat this condition, despite their side effects. Currently, there is no scienti c evidence of resolution or prevention of this condition with a speci c drug.
PPS has a variable clinical presentation and is de ned by at least 2 of the following 5 signs and symptoms: new-onset pericardial effusion, new-onset pleural effusion, fever, pleuritic chest pain, and pericardial or pleural rubbing on physical examination [2]. Due to the similarities in their anatomical features and the relationship of the structures, we should also consider a similar etiopathogenesis for both pleural and pericardial effusions. This also re ects on the surgical technique, which often involves both structures: after incising the pericardium, the pleurae could be opened and drainage tubes positioned.
Pleuro-pericardial effusion may be detected starting a few days after surgery until the follow-up period, often complicating the post-operative course. In most cases, the accumulated uid may cause respiratory distress (atelectasis) and infections, leading to a prolonged hospital stay or intensive care unit (ICU) admission.
Diagnosis of pericardial effusion relies on postoperative echocardiography, while pleural effusion (PE) is usually diagnosed through chest x-ray. A radiological classi cation of pleural effusion using a semi-quantitative scale has been proposed [3] and is shown in Table 1. Thoracic ultrasound, on the other hand, has the advantage of being less invasive, readily available and repeatable. Furthermore, chest ultrasonography has more sensibility and reliability than radiography [4,5]. Also in this case, a classi cation is used to assess the amount of pleural effusion [6] ( Table 2).
Pericardiocentesis is performed when the increased pressure in the pericardial cavity leads to echocardiographic signs of cardiac chambers compression (right and left atrial collapse, ventricular collapse, swinging heart etc.) and subsequent haemodynamic instability.
The best management of pleural effusion is controversial. Clear indications on the right timing to perform a thoracentesis are lacking. Usually, a pleural effusion is considered complicated when thoracentesis is needed and is accompanied by clinical and Arterial Blood Gas (ABG) parameters alterations.
Nonsteroidal anti-in ammatory drugs, colchicine and steroids can be employed. Thoracentesis is typically performed as a nal resort, considering that this procedure is not free from complications such as haemorrhage and pneumothorax.
In order to attenuate the in ammatory response, a single dose of long-acting corticosteroids is given during heart surgery in many European countries, although de nitive results of its e cacy are lacking [7]. The DECS study published by Bunge J.J.H et al. [3] is the largest one among those on corticosteroids usage. It is a multicenter randomized trial that compares the use of dexamethasone with placebo in primary prophylaxis of PPS, giving to selected patients 1 mg/kg of dexamethasone at the beginning of the anesthetic procedures. Statistical evaluation testi ed that there were no differences between the two groups.
Colchicine and Indomethacin can be also employed in primary prophylaxis of PPS. Several randomized controlled trials (RCT) were published, especially on colchicine. In our study we combined Finkelstein Y et al. study [8], COPPS study [9], COPPS-2 study [10], POPE study [11] and Izadi AA et al. [12] for evaluating colchicine e cacy. Regarding indomethacin, we investigated the work of Inan MB et al. [13].

Methods
Literature research was carried out using PubMed. We combined the following key words to identify the publications in the following query: "pericarditis OR pericardial effusion OR pleuro-pericardial syndrome AND colchicine OR dexamethasone OR indomethacine AND heart surgery". We applied the following inclusion criteria: studies on ≥10 patients with clinical pericarditis or post-pericardiotomy syndrome in which colchicine, dexamethasone and indomethacin were compared with placebo. We excluded publications referring only to animal experiments or in vitro experiments, human studies on < 10 patients, case reports, congress reports, review articles, editorial letters. The full publications that met the criteria listed before were included in the analysis.
The primary end-point of the study was to investigate the e cacy of colchicine, dexamethasone and indomethacin in the treatment and prevention of PPS and pleural effusion. The secondary end-points were to assess the e cacy of the three drugs in reducing hospital readmission, cardiac tamponade, symptom persistence, atrial brillation after 72 h, as well as the safety and adverse effects of pharmacological treatment.
We collected data regarding study characteristics, publication year, randomization, blinding, intention-to-treat analysis, patients clinical characteristics, treatment indication and duration, pharmacological dosage, follow-up, clinical outcomes. Clinical features and study methods are summarized in Table 3.
The assessment of study quality was obtained by Jadad scale [14] and included randomization, blinding of participants and dropouts. Upon the obtained value, Jadad score allows to establish the probability of bias (Table 4).

Statistical analysis
Metanalysis was performed by random effect model (Der Simonian -Laird model). We considered risk ratios related to each single study with con dence interval percentage weight on the Overall Risk ratio.
Hypothesis testing for equivalence was set at the two-tailed 0.05 level. Heterogeneity was based on the Cochran Q test, with I2 values.
Funnel plots by graphical inspection and Egger regression test were used for assessment of publication bias.
Univariable meta-regression was performed to explore a signi cant increase or decrease in the number of patients in the treatment group who had the event.
For each group some independent variables were omitted for reasons of lack of data or due to multicollinearity.

Results
In this review we examined an amount of 361 studies from 1977 to 2016. Of these studies we considered only 7 randomized controlled trials, including trials regarding PPS and pleural effusion after heart surgery. 5 of them were double blind multicenter randomized trials [8, 9, 10, 11], 1 was a triple blind multicenter randomized trial [12], 1 was a single center randomized trial. Follow up period was from 1 to 12 months. Only ve patients got lost during the follow up of all studies. The groups did not differ in clinical characteristics and type of surgery.
The indication for pharmacological therapy was dependent upon PPS and postoperative pericardial effusion.
The seven prospective studies from 2002 to 2015 included 1677 patients (mean age 60,8, male 67,4 %). Control group received placebo. The weight-adjusted colchicine dosage was 0.5 mg in three studies and 1 mg in two studies. The duration of colchicine treatment was between 14 days and 30 days. It was started 48-72 hours before surgery [10], on 3rd postoperative day [8,9], on day 7, 30 [11] and 3 weeks after heart surgery [12].
Indomethacin dosage was 25 mg 3 times/day for 3 days before surgery. Treatment duration lasted for six weeks after operation.
Dexamethasone dosage was 1 mg/Kg i.v. as a single intraoperative bolus.

Main Outcomes
Primary endpoints were related to reduction of pericardial and pleural effusion in all the studies examined, while there was no homogeneity with secondary endpoints in the study groups.
In Forest and Funnel plot statistical analysis upon primary endpoints was reported an overall effect size only in two studies (Fig. 1A, Fig. 1B).
Forest and Funnel plot for secondary endpoints demonstrated non signi cant effect size depending also on the heterogeneity of the pathologies involved ( Fig. 2A, Fig. 2B).
Inan et al. explored indomethacin effect on pericardial effusion when administered 7 days before surgery. Medication continued for six weeks after operation. We quote this study for its statistical features similar to the other in this review, but its low number of patients is a limit for a statistical comparison with other studies with a high number of patients enrolled. It is worth mentioning that during hospitalization and follow-up a signi cant positive difference was registered between the Indomethacin group and control group. In Table 5 are reported characteristics of population study in order to potentially explore factors reducing or increasing the number of events. The analysis failed to show any signi cant triggering factor.
Instead, pre-operative pericardial effusion was identi ed as a factor able to impact on studies outcome. Indeed, in Table 6, between the different characteristics analyzed and reported, none of the other factors, despite pre-operative PE, have achieved an association with the heterogeneity between studies.
There are several factors that may cause a PPS in response to cardiac surgery. Cardio-pulmonary by-pass (CPB) plays a major role due to the in ammatory activation process. Prolonged time of CPB may enhance the in ammatory response. Surgical trauma may cause an autoimmune-based in ammatory syndrome [20,21]. Duration of the operation is also considered a determinant part of the in ammatory response. The association between the levels of circulating cytokines and incidence of PPS has been evaluated in patients who underwent coronary artery bypass grafting (CABG). A low preoperative IL-8 level was identi ed as a high risk marker for development of PPS. It is thought that IL-8 may have a protective effect on endothelial and mesothelial cells and therefore patients with high preoperative levels of IL-8 may be less prone to PPS. The limited number of studies that have evaluated particular serum biomarkers and histological markers underlines the need for a more systematic examination on cellular and genomic expression of plasmatic factors which confer an increased susceptibility to PPS. PE may be localized in the entire pericardial space but regional effusions may be detected.
The Post-pericardiotomy syndrome may occur days or months after surgery. Generally it occurs with nonspeci c signs and symptoms, such as fever and exhaustion, but it may also present with chest pain and breathing di culties. Symptomatology depends on the in ammatory reaction triggered after surgical trauma.
Complications related to PPS after heart surgery may cause slow recovery and prolonged hospitalization.
Among PPS complications, pleural effusion can cause dyspnea, peripheral desaturation up to a real respiratory fatigue. Pericardial effusion is another relevant complication, which can lead to chest pain, asthenia, until the onset of cardiac arrhythmias or cardiac tamponade. These complications may lead to greater morbidity as well as to a greater risk of mortality and a progressive delay in the rehabilitation process and reintegration to normal life. This syndrome presents with pleural or pericardial reaction, characterized by fever, chest pain, and a friction rub. Although symptoms often are speci c to PPS, diagnosis may be aided by the use of radiography, echocardiography, electrocardiography, and laboratory analysis. Chest radiology can highlight the presence of pleural effusion (Table 1) whose degree may indicate a possible treatment. In addition to radiological evaluation, pleural ultrasound has a great speci city in diagnosing a pleural effusion ( Table 2). Pericardial effusion is instead evaluated by echocardiography, highlighting both the effusion rate and any interaction with cardiac contractility.
Several studies reported various treatments, in particular with non-steroidal anti-in ammatory drugs, colchicine and corticosteroids. These drugs are especially used for PPS and PE treatment. Only few studies are reported for primary prophylaxis of PPS.
In this review we investigated 7 randomized trials on colchicine, indomethacin and dexamethasone used for primary prophylaxis purposes. All trials have an homogenous number of enrolled patients, except the work on indomethacin with a lower number of patients in comparison to the other studies. This represents a bias on statistical analysis. We quote this work since it is the only one regarding indomethacin to have a similar feature to the other ones we have studied.
The rst 5 studies reported in this review concerned colchicine treatment in cardiac surgery patients. Finkelstein et al. [8] performed a prospective, randomized double blind study. They found a statistical signi cance in their data concerning the colchicine treatment group. No signi cant data were reported on speci c secondary end points and adverse effects.
Imazio et al (COPPS − 2) [10] attested in their multicenter study that preoperative administration of colchicine signi cantly reduced postoperative in ammation and its complication, especially PPS. As suggested by a subgroup analysis, colchicine was more e cacious against C-reactive protein elevation. They avoided a loading dose and used weight-adjusted doses in order to obtain a higher degree of treatment adherence. High rates of adverse events were registered such as gastrointestinal intolerance and drug discontinuation. These ndings suggested that colchicine must be employed only in well selected patients.
Meurin et al (POPE-2 study) [11], assessing the effectiveness of colchicine to treat asymptomatic postoperative pericardial effusion, found that the use of the drug did not signi cantly reduce the volume of PE. Authors justi ed the absence of e cacy considering the in ammatory etiology not being the main trigger of PE, considering more plausible etiologies postoperative hemorrhagic effusion and PE secondary to heart failure. They concluded that in patients with a true PPS colchicine may be e cacious as described in the ICAP study [22].
Izadi Amoli et al. performed a triple blind trial comparing 2 weeks treatment group with placebo group. Authors found that there was no difference in the two groups regarding postoperative mild to moderate PE. It was a study conducted in a high turnover specialized department with various cardiac pathologies. According to randomization rules there were no signi cant differences between the 2 study groups in baseline and clinical characteristics. Forest plot, used in this analysis, con rmed the main ndings related to the no signi cance of colchicine use on reducing PE compared with placebo group.
Inan et al. [13] investigated the prophylactic indomethacin effect on postoperative PE. Within the limitations of the low number of enrolled patients, results of this study suggest that preoperative indomethacin intake may have a bene cial role on the outcome and incidence of postoperative PE. Conclusions of the authors indicated indomethacin an alternative protocol to other drugs to avoid postoperative PE.
Bunge et al. [3] found no protective effect of a single high intraoperative dose of dexamethasone on PPS or complicated PPS in a cohort of patients undergoing valvular surgery.

Conclusion
PPS and PE are the epiphenomena of the in ammatory status secondary to heart surgery, especially when CPB is adopted. Review of the main studies about PPS and PE treatment on cardiac surgery represents a main concern in avoiding complications such as cardiac tamponade or infections in the rst postoperative days.
Limits of this study may be found in the heterogeneity of the population and in the timing of drug administration. In the studies with postoperative colchicine administration, 2 of them demonstrated a signi cant reduction of PE. In the single pre-surgery colchicine administration study, a reduction of PE was as well obtained. In the indomethacin study, pre-surgery administration was linked to a reduction of PE. This result was not reported with pre-operatively dexamethasone intake. The aforementioned studies also demonstrated no reduction in PE of non-in ammatory etiology. In future studies it could be useful that this feature is excluded from primary endpoints.       Forest Plot of Adverse effects.