This prospective, single-center, randomized controlled trial was conducted in the Department of Cardiovascular Surgery, Cardiovascular and Kidney Transplantation Center, Taiz, Yemen, between January 2022 and June 2024.

The study followed the principles of the Declaration of Helsinki. Approval was obtained from the ethical approval was obtained from the Research Ethics Committee of the Cardiovascular and Kidney Transplantation Centre, Taiz University, Yemen (Approval ID 0030425, dated December 20, 2021), and written informed consent was collected from all participants prior to enrollment. The trial is registered with ClinicalTrials.gov (Identifier: NCT07266935).

Patients aged $\ge$18 years undergoing elective open-heart surgery (CABG, aortic valve replacement, ascending aortic surgery, or combined procedures) were eligible.

Exclusion criteria included: (1) previous cardiac or thoracic surgery, (2) left-sided pleural adhesions, (3) preoperative atrial fibrillation or other rhythm disorders, (4) hyperthyroidism, (5) renal failure with plasma creatinine $>$2.0 mg/dL, (6) off-pump CABG, (7) mitral or tricuspid valve surgery (excluded due to distinct pathophysiology and POAF risk [1, 13]), and (8) refusal to participate.

Patients were randomized in a 1:1 ratio to the PP group or the control group using sealed, opaque envelopes prepared by an independent staff member not involved in clinical care. A total of 106 patients were assigned to the PP group, and 104 patients to the control group (Fig. 1).

Fig. 1.

Trial CONSORT flow diagram. PP, Posterior pericardiotomy; OPD, outpatient clinic; POD, postoperative day. *Randomization: sealed, opaque envelopes prepared by an independent assistant not involved in patient care.

The trial followed a single-blind design: investigators responsible for data collection and outcome assessment were blinded to group allocation throughout the perioperative and postoperative periods.

All procedures were carried out through a median sternotomy. Prior to initiation of cardiopulmonary bypass (CPB), patients were anticoagulated with intravenous heparin (3 mg/kg) to achieve an activated clotting time (ACT) $>$480 seconds. CPB was managed with moderate hemodilution (hematocrit 20–26%) and systemic hypothermia (28–32 °C), using a roller pump (2.0–2.4 L/m²/min) and a membrane oxygenator. Mean arterial pressure was maintained between 50–70 mmHg. Myocardial protection was achieved with antegrade infusion of cold blood cardioplegia via the aortic root. At the conclusion of CPB, anticoagulation was reversed with protamine sulfate (3.5 mg/kg).

PP was performed according to the operative procedure. In CABG, it was created after completion of distal vein grafts but before the internal mammary artery–LAD anastomosis. In valve or combined surgeries, PP was carried out after valve excision. The incision measured 4–5 cm, oriented longitudinally and positioned parallel and posterior to the left phrenic nerve, extending from the left inferior pulmonary vein to the diaphragm [11, 14]. Any complications during the procedure were recorded.

Chest drainage was standardized in all patients. A 28 Fr tube was inserted into the left pleural cavity and a 32 Fr tube into the anterior mediastinum. The pleural tube was introduced through a separate pleural opening and directed to the lower pleural space. If the right pleura was inadvertently opened, the mediastinal drain was passed through it before being positioned retrosternal. In all cases, the anterior pericardium was left open, and posterior placement of tubes near the heart was avoided to minimize arrhythmic risk.

After surgery, patients were transferred to the intensive care unit (ICU). Chest drains were regularly milked and stripped to maintain patency, with low-intermittent suction (–20 cmH₂O) applied. Drainage was measured hourly, and tubes were removed on the second postoperative day once output had fallen below 100 mL in 24 hours. Standard prophylaxis with $\beta$-blockers was administered to all patients.

Cardiac rhythm was continuously monitored by electrocardiography for the first three postoperative days. Monitoring was resumed whenever arrhythmia was suspected, when heart rate changes were observed, or if patients experienced palpitations. Electrolytes were corrected as needed, with potassium maintained above 4.5 mmol/L and magnesium above 1.0 mmol/L. Two-dimensional echocardiography was performed on days 3 and 5, before discharge, and again during follow-up to evaluate pericardial effusion. An effusion $>$1 cm was considered clinically significant. Additionally, echocardiographic examinations were also performed after discharge to detect late effusions and tamponade.

• Primary endpoint: occurrence of POAF was defined as atrial fibrillation documented by a 12-lead electrocardiogram or continuous telemetry monitoring, lasting more than 30 seconds, and occurring after cardiac surgery in patients with no prior history of atrial fibrillation [15, 16].

• Secondary endpoints: requirement for antiarrhythmic medications, systemic anticoagulation, or cardioversion; development of pericardial or pleural effusion, cardiac tamponade, or need for reexploration; duration of ventilation (hours under mechanical ventilation until extubation); ICU stay (defined as total time in hours from ICU admission until until transfer to the ward); total hospital stay (days from surgery to discharge); in-hospital mortality; and major adverse cardiovascular events (stroke, myocardial infarction, or all-cause mortality).

• Pulmonary complications included pneumonia, prolonged ventilation ($>$24 hours), or reintubation. Sternal complications included superficial or deep sternal wound infection, wound dehiscence, or sternal instability. Renal impairment was defined as a postoperative increase in serum creatinine of $>$50% from baseline or the need for dialysis.

All analyses followed the intention-to-treat (ITT) principle. A per-protocol (PP) analysis was also performed for sensitivity.

Sample size was calculated assuming a 25% incidence of POAF in controls and a 10% incidence in the PP group, based on prior literature [11, 17]. With $\alpha$ = 0.05 and 80% power, 94 patients per group were required. To account for dropouts, we targeted 210 patient’s total.

The baseline variables listed in Table 1 (including demograph​ic characteri​stics, comorbidities‍,‍ and preoperative laboratory and echocardiographic findings). These variables were chosen based on established predictors of POAF in the literature [1, 2, 3]. Posto​perative outc​omes, such as pericardial effusion‌, cardiac ta​mponade, pleural effusion, re-exploration, hospital mortali‍ty, and other adverse events, were scr​eened d​uring hospitalizatio​n and 30 days post operative fo​llow-up. Data collection was performed prospectively, with preoperative variables assessed‍ before surgery, intraoperative variables recorded during surgery, and postoperative variables monit​ored until discharge and at follow-up for 30 days postoperatively.

All analyses were performed using IBM SPSS Statistics, version 24.0 (IBM Corp., Armonk, NY, USA). Continuous variables were summarized as mean $\pm$ standard deviation (SD) or median with interquartile range (IQR), depending on distribution, while categorical variables were expressed as counts and percentages. Group comparisons were conducted with Student’s t-test or the Mann–Whitney U test for continuous data, and with the chi-square test or Fisher’s exact test for categorical data. Effect estimates were presented as risk differences or odds ratios (OR) with corresponding 95% confidence intervals (CI).

Univariate logistic regression was used to identify risk factors for POAF. Multivariable logistic regression adjusted for clinically relevant covariates (e.g., age, sex, hypertension, diabetes, left atrial size, left ventricular ejection fraction (LVEF)), irrespective of univariate significance. A two-tailed p-value 0.05 was considered statistically significant.

The incidence of POAF was significantly reduced in the PP group compared with controls (8.5% vs. 22.1%; OR 0.32, 95% CI 0.14–0.71; p = 0.006) (Table 2).

Among secondary endpoints, the need for antiarrhythmic therapy was lower in the PP group (7.5% vs. 22.1%; OR 0.28, 95% CI 0.11–0.67; p = 0.003), as was the requirement for systemic anticoagulation (5.6% vs. 12.5%; OR 0.41, 95% CI 0.15–0.97; p = 0.040). Electrical cardioversion was infrequent, with no significant difference between groups (0% vs. 0.6%; p = 0.32).

Patients in the PP group also demonstrated shorter ventilation times and ICU stays. Median ventilation time was 3.0 hours (IQR 3.0–4.0) versus 4.0 hours (IQR 3.0–6.0) in controls (p 0.001). Median ICU stay was 23.0 hours (IQR 22.0–24.0) compared with 24.0 hours (IQR 22.2–45.0) in the control group (p = 0.004).

Postoperative chest drainage during ICU stay was significantly lower in the PP group (median 410 mL, IQR 300–550) than in controls (median 490 mL, IQR 340–707; p = 0.014). However, cumulative drainage during hospitalization did not differ significantly (p = 0.165). Median hospital stay was 6 days in both groups (p = 0.099), and readmission rates were comparable (10.4% vs. 11.5%; p = 0.788).

Posterior pericardiotomy was associated with a marked reduction in pericardial complications compared with the control group. The incidence of pericardial effusion was significantly lower in the PP group (9.4% vs. 39.4%; OR 0.17, 95% CI: 0.08–0.36; p 0.001). No cases of cardiac tamponade occurred among patients who underwent PP, whereas 9.6% of controls were affected (p = 0.003). Similarly, the need for surgical re-exploration due to bleeding or tamponade was less common in the PP group (0.9% vs. 6.7%; OR 0.13, 95% CI 0.02–1.03; p = 0.067). This difference was mainly attributable to tamponade events, which were absent in the PP group but observed in 5.8% of controls (p = 0.036) (Table 3).

Hospital mortality was 0% in the PP group and 4.8% in the control group (p = 0.067). This difference was not statistically significant. Postoperative pulmonary complications were lower in the PP group (3.8% vs. 10.6%; p = 0.056), although this trend did not reach statistical significance. Rates of sternal complications and renal impairment were comparable between the two groups.

On univariate analysis, PP was associated with a reduced risk of POAF (OR 0.48, 95% CI 0.25–0.90; p = 0.022). No other baseline or operative characteristics (age, sex, comorbidities, left atrial size, LVEF, operative times) were significantly associated with POAF (Table 4). In multivariable analysis adjusting for clinically relevant covariates (age, sex, hypertension, diabetes, left atrial size, LVEF, surgical type), PP remained an independent protective factor against POAF (adjusted OR 0.194, 95% CI 0.06–0.62; p = 0.006).