INTRODUCTION
Coronary artery bypass grafting (CABG) is a critical intervention for managing severe coronary artery disease, significantly improving patient survival and quality of life[1]. However, postoperative complications, including postoperative atrial fibrillation (POAF), pericardial effusion, pleural effusion, and cardiac tamponade, remain significant challenges[2-4]. POAF affects 20%-40% of CABG patients and is associated with increased risks of stroke, heart failure, renal dysfunction, prolonged hospitalization, and elevated mortality and healthcare costs[4-7]. Pericardial and pleural effusions further contribute to morbidity, potentially leading to cardiac tamponade or respiratory complications[2,8].
Traditional management of post-CABG complications involves pharmacological interventions, such as antiarrhythmic drugs and anticoagulants, and mechanical approaches like pacemaker placement[9,10]. These strategies often address symptoms rather than underlying causes and may introduce risks such as bradycardia, hypotension, or bleeding[9,10]. Posterior pericardiotomy (PP), a surgical technique involving a longitudinal incision in the posterior pericardium to facilitate drainage of pericardial fluid into the pleural space, has emerged as a promising preventive approach[3,8,11,12]. Evidence suggests that PP reduces the incidence of POAF and pericardial effusion by preventing fluid accumulation and mitigating inflammatory responses[3,8,11,12]. Despite these benefits, the long-term outcomes and potential risks of PP, such as pleural effusion or infection, remain under debate due to limited data[6,12].
This meta-analysis synthesizes evidence from randomized controlled trials (RCTs) to evaluate the efficacy and safety of PP in reducing postoperative complications in CABG patients compared to standard care. By providing comprehensive evidence, this study aims to inform clinical decision-making and contribute to evidence-based guidelines for optimizing patient outcomes.
MATERIALS AND METHODS
Study design
This study is a systematic review and meta-analysis conducted to analyze the outcomes of pericardiotomy after CABG. The review is in accordance with the PRISMA guidelines (Prospero ID: CRD42025645108)[13].
Information sources and search strategy
A systematic literature search was conducted across PubMed, Cochrane Library, ClinicalTrials.gov, and Ovid to identify studies published between January 1980 and January 2025, focusing on cardiac surgery, pericardiotomy, pericardiectomy, atrial fibrillation (AF), and pericardial effusion. The search strategy integrated Medical Subject Headings (MeSH) and free-text terms to ensure comprehensive retrieval of relevant studies while minimizing irrelevant results. MeSH terms included “Cardiac Surgical Procedures” [MeSH], “Thoracic Surgery” [MeSH], “Coronary Artery Bypass” [MeSH], “Pericardiectomy” [MeSH], “Atrial Fibrillation” [MeSH], “Arrhythmias, Cardiac” [MeSH], and “Pericardial Effusion” [MeSH]. Free-text terms comprised “cardiac surgery”, “cardiothoracic surgery”, “CABG”, “pericardiotomy”, “pericardiectomy” “atrial fibrillation”, “arrhythmia”, and “pericardial effusion”. These terms were carefully selected to maximize study retrieval and capture all relevant literature.
Eligibility criteria
This meta-analysis included RCTs published in English that investigated adult patients (≥ 18 years) undergoing isolated CABG. The intervention of interest was prophylactic posterior left pericardiotomy, compared to a control group that did not undergo the procedure. The primary outcome assessed was the incidence of POAF.
Studies were excluded if they involved pediatric or adolescent patients, were observational in nature (including cohort studies, case reports, case series, reviews, and abstracts), or were conducted on animals. Additionally, non-English publications were not considered.
Study selection
Studies were selected through a systematic screening process based on predefined inclusion and exclusion criteria. Two independent reviewers assessed the titles and abstracts of all retrieved studies to identify those meeting the eligibility criteria. Full-text articles of potentially relevant studies were then reviewed to confirm their inclusions. Any discrepancies between the reviewers were resolved through discussion and consensus among the research team. This rigorous selection process ensured the inclusion of only high-quality, relevant studies in the meta-analysis.
Data extraction
Data extraction was performed independently by two authors to ensure accuracy and consistency. A standardized template was used to collect relevant information from each included study. Key details extracted included study characteristics such as the authors, publication year, country, study design, and participant numbers, demographics and comorbidities. Additionally, outcome data related to postoperative cardiovascular complications such as AF, early and late pericardial effusion, pericardial tamponade, hospital stay time, mortality were recorded. In cases of disagreement between the reviewers, consensus discussions were held to resolve discrepancies. This process helped maintain the reliability and thoroughness of the data collection.
Quality assessment
The risk of bias assessment for the included RCTs was conducted using the Revised Cochrane Risk of Bias tool (RoB 2.0), which evaluates potential biases across five key domains: (1) Bias arising from the randomization process; (2) Bias due to deviations from intended interventions; (3) Bias due to missing outcome data; (4) Bias in measurement of the outcome; and (5) Bias in selection of the reported result. Each study was independently assessed by two reviewers, with discrepancies resolved through discussion or consultation with a third reviewer. The risk of bias for each domain was categorized as low risk, some concerns, or high risk of bias[14].
Statistical analysis
Data were synthesized using random-effects meta-analysis to account for the expected heterogeneity across studies. The following statistical measures were calculated: Risk ratio (RR) for dichotomous outcomes (e.g., AF, mortality, etc). Mean difference (MD) for continuous outcomes (e.g., length of hospital stays, etc). Both RRs and MDs were accompanied by 95% confidence intervals (CI). Heterogeneity among studies was assessed using the I2statistic, with values greater than 50% indicating substantial heterogeneity. To assess potential publication bias, we performed the Egger's Regression test for studies that include at least 10 studies. This test evaluates the asymmetry of the funnel plot for each key outcome and was used to determine if there was any evidence of publication bias. All statistical analyses were conducted using RevMan 5.3 (Cochrane Collaboration). The Egger’s Regression test for publication bias was performed using R version 4.4.1.
RESULTS
Study selection
The systematic search of PubMed, Cochrane, ClinicalTrials.gov, and Ovid, conducted up to July 2025, yielded 1324 records. After removing 128 duplicates, 1106 records underwent title and abstract screening, of which 960 were excluded due to irrelevance, non-human studies, or inappropriate study designs. The full-text review was performed for 46 articles, with 26 excluding due to lack of comparative data, insufficient outcome reporting, single-arm studies, and duplicate or overlapping data. Ultimately, 20 studies[15-34] were included in the meta-analysis (Figure 1).
Figure 1
PRISMA flowchart depicting the literature screening process.
Study characteristics
The meta-analysis included 20 RCTs conducted across nine countries, with the majority from Egypt (5 studies) and Turkey (6 studies), along with contributions from Iran, Pakistan, the United Kingdom, Thailand, Australia, China, and Greece. The mean age of participants ranged from 54 years to 67 years, with comparable distributions between intervention (PP) and control groups. The percentage of male participants varied widely across studies, ranging from 32% to 96.6%, with the highest male representation reported in Elkhouly et al[21] and Kaya et al[26] (above 85%). The prevalence of diabetes and hypertension also varied significantly, with diabetes rates ranging from 10% to 65% and hypertension rates ranging from 10% to 79%, with Rathnayake et al[32] reporting the highest prevalence of hypertension (79%). Several studies (e.g., Ahmad et al[15], Kuralay et al[31]) did not report comorbidity data. Overall, the included studies represented a diverse population with varying baseline characteristics, ensuring a broad and representative analysis of the intervention’s effects (Supplementary Table 1).
Outcomes
AF: AF was reported by 19 studies. The pooled RR of AF for pericardiotomy is 0.48 (95%CI: 0.36-0.65, P < 0.00001), with heterogeneity of I2 = 70% that dropped significantly upon removal of studies like Rathnayake et al[32] and Kongmalai et al[30] (I2 = 42%) (Figure 2A).
Figure 2 Forest plots.
A: Forest comparing atrial fibrillation between post-pericardiotomy after coronary artery bypass grafting (CABG) and no posterior pericardiotomy (PP); B: Forest comparing cardiac tamponade between post-pericardiotomy after CABG and no PP; C: Forest comparing early pericardial effusion between post-pericardiotomy after CABG and no PP; D: Forest comparing hospital length of stay between post-pericardiotomy after CABG and no PP. CI: Confidence interval.
Cardiac tamponade: A total of 12 studies reported cardiac tamponade. PP was associated with significantly lesser rate of cardiac tamponade with the estimated pooled RR for cardiac tamponade is 0.16 (95%CI: 0.08-0.34, P < 0.00001), and heterogeneity I2 = 0% (Figure 2B).
Early pericardial effusion: 15 studies reported this outcome. The estimated pooled RR for early pericardial effusion is 0.31 (95%CI: 0.14-0.68, P = 0.004) with a heterogeneity I2 value of 96% that dropped significantly upon removal of studies like Rathnayake et al[32] and Kongmalai et al[30] (I2 = 54%) (Figure 2C).
Hospital length of stay: Hospital length of stay was reported by a total of 11 studies. The combined MD for hospital length of stay was estimated to be -1.23 (95%CI: -1.87 to 0.59, P = 0.0002), and a heterogeneity I2 = 85% that dropped significantly upon removal of studies like Rathnayake et al[32] and Sadeghpour et al[33] (I2 = 48%) (Figure 2D).
Mortality: A total of 7 studies reported about the outcome. No significant difference was observed between the groups. The estimated pooled RR for Mortality is 0.92 (95%CI: 0.48-1.76, P = 0.80) and no heterogeneity I2 = 0% (Supplementary Figure 1).
Late pericardial effusion: Late pericardial effusion was reported by 9 studies. The estimated pooled RR for late pericardial effusion resulting from the studies is 0.11 (95%CI: 0.05-0.21, P < 0.00001) with a heterogeneity I2 of 0% (Supplementary Figure 2).
Pleural effusion: 12 studies reported this outcome. The combined RR for pleural effusion is estimated to be 1.46 (95%CI: 1.21-1.76, P < 0.0001) with no heterogeneity (I2 = 0%) (Supplementary Figure 3).
Revision surgery for bleeding: The outcome was reported by a total of 15 studies. No significant difference was observed. The estimated pooled RR for revision surgery for bleeding is 0.77 (95%CI: 0.53-1.12, P = 0.17) and heterogeneity I2 = 2% (Supplementary Figure 4).
Quality assessment: Among the 20 RCTs assessed using RoB 2.0, 15 studies (75%) were classified as having a high risk of bias, 3 studies (15%) had some concerns, and only 2 studies (10%) were deemed to have a low risk of bias. This indicates that the majority of included studies had methodological limitations that could influence the overall findings of the meta-analysis (Supplementary Figures 5 and 6).
DISCUSSION
Postoperative complications following CABG, such as AF, cardiac tamponade, pericardial effusion, pleural effusion, and the need for revision surgery, contribute significantly to morbidity and prolonged hospital stays[14]. Our meta-analysis of 20 RCTs involving 5331 CABG patients demonstrates that PP significantly reduces the incidence of postoperative AF (RR = 0.48, 95%CI: 0.36-0.65, P < 0.00001), early pericardial effusion (RR = 0.31, 95%CI: 0.13-0.51, P < 0.0001), late pericardial effusion (RR = 0.11, 95%CI: 0.04-0.15, P < 0.00001), and cardiac tamponade, despite significant heterogeneity (I2 = 71%-83%). These findings align with prior studies[35-38] and support PP’s efficacy in mitigating complications by facilitating pericardial fluid drainage into the pleural space, reducing mechanical irritation and electrical instability that trigger AF[16]. However, PP was associated with a higher incidence of pleural effusion (RR = 1.46, 95%CI: 1.27-2.04, P < 0.0001, I2 = 0%) and no significant reduction in revision surgery for bleeding or intensive care unit stay.
Postoperative AF, affecting 5%-40% of CABG patients, peaks on postoperative day two and is driven by multifactorial etiologies, including pericardial inflammation, oxidative stress, and autonomic dysregulation[14,36]. PP, involving a 4-5 cm longitudinal incision posterior to the phrenic nerve, addresses these by draining posterior pericardial effusions that are difficult to manage with substernal chest drains due to their proximity to the heart and bypass grafts[37]. This drainage reduces localized pressure on the left atrium and ventricle, which can precipitate AF or tamponade[39,40]. Our results confirm PP’s antiarrhythmic effect, consistent with Mulay et al’s study[11] and subsequent meta-analyses[1].
The observed heterogeneity (I2 = 71%-85%) likely stems from variations in surgical techniques, patient characteristics, and perioperative management. Moreover, the high risk of bias in 15 of the 20 included studies, primarily due to lack of blinding and deviations from intended interventions, may have influenced our findings, particularly for subjective outcomes like AF detection[41-44]. Lack of blinding in these trials could introduce detection bias, as clinicians aware of the intervention status might have been more vigilant in monitoring for AF in the control group, potentially overestimating its incidence. This is particularly relevant for AF, which often requires electrocardiographic confirmation and clinical judgment, both susceptible to subjective interpretation. For instance, unblinded clinicians might prioritize more frequent electrocardiogram monitoring or interpret borderline arrhythmic events more stringently in the control arm, inflating the apparent benefit of PP. Similarly, the assessment of pericardial effusion via echocardiography could be subject to observer bias in unblinded settings, potentially exaggerating the reported reduction in effusion rates. While objective outcomes, such as cardiac tamponade, are less prone to detection bias due to their clear clinical presentation, the overall reliability of our pooled estimates may be compromised by these biases. Sensitivity analyses excluding high-risk-of-bias studies could help quantify this impact, but only two studies provided low-risk, unbiased data, limiting such analyses[20,41].
Despite these limitations, PP significantly reduced both early and late pericardial effusions, likely due to effective drainage and reduced pericardial constraint[45-47]. The increased pleural effusion incidence is an expected consequence of PP, as pericardial fluid drains into the pleural cavity, but this rarely leads to respiratory complications[11,36]. Hospital stay was reduced (MD = -1.23 days, 95%CI: -1.87 to -0.59, P = 0.0002), though high heterogeneity (I2 = 85%) suggests variability in institutional protocols. No significant reduction in ICU stay was observed (MD = -0.34 days, 95%CI: -0.83 to 0.14, P = 0.17), possibly due to factors beyond PP, such as postoperative care practices.
Despite these promising results, several limitations temper the strength of our conclusions. The high heterogeneity (I2 = 71%-85%) across outcomes likely stems from variations in surgical techniques, patient characteristics, and perioperative medication use. For example, differences in beta-blocker or amiodarone use could confound AF outcomes, yet most studies did not adequately report or adjust for these factors, limiting our ability to assess their impact. Similarly, variations in surgical expertise or postoperative monitoring protocols may influence effusion detection and hospital stay duration, but these were inconsistently documented across trials.
A critical limitation is the high risk of bias in 15 of the 20 included studies[3-7,9,10,12-14,36-40], primarily due to lack of blinding, deviations from intended interventions, and inconsistent randomization criteria. Only two studies were deemed low risk for bias[20,41]. Lack of blinding introduces detection bias, particularly for subjective outcomes like AF, which relies on electrocardiographic confirmation and clinical judgment. Unblinded clinicians may have monitored control groups more vigilantly, potentially overestimating AF incidence and inflating PP’s apparent benefit. Similarly, echocardiographic assessment of pericardial effusions is susceptible to observer bias in unblinded settings. Objective outcomes, such as cardiac tamponade, are less affected due to their clear clinical presentation, but the overall reliability of pooled estimates is compromised. Additionally, the absence of patient-level data prevented subgroup analyses to explore how factors like age, comorbidities, or surgical approach modify PP’s effects, further limiting our ability to account for confounding variables.
Missing outcome data in several trials introduces attrition bias, potentially underestimating event rates for AF and effusions. The geographic concentration of studies, with eight conducted in Turkey[6,8-9,11-14,36] and five in Egypt[4,7,10,38,39], raises concerns about generalizability. For instance, differences in patient demographics, such as prevalence of diabetes or hypertension, or variations in surgical practices between these regions and global populations, were not reported in most studies, precluding a robust assessment of regional bias. This geographic limitation suggests caution in extrapolating our findings to diverse populations.
Given these limitations, particularly the high risk of bias and lack of patient-level data, strong recommendations for PP’s routine adoption in CABG are premature. While PP appears effective in reducing AF and pericardial effusions, the increased risk of pleural effusion and the lack of impact on ICU stay or revision surgery highlight the need for a balanced risk-benefit assessment. Future research should prioritize blinded, multicenter trials with standardized protocols and diverse populations to confirm PP’s efficacy and address confounding factors. Such studies could also explore non-invasive predictors of AF, such as left atrial volume index or P-wave dispersion, to refine patient selection for PP.
