Four electronic databases (Cochran library, Medline via PubMed, SCOPUS, and Web of Science) were searched for relevant citations using terms and keywords referring to sinus node dysfunction and cardiac surgical procedures up to 6 June 2022, with no starting date restrictions. No limits were applied to document type, language, or date. Utilized keywords are presented in Table 1, and adopted search strategies for each database are enclosed in Supplementary Table S1.

All retrieved citations were first screened for eligibility criteria based on their title and abstracts, and if potentially relevant, their full text was assessed. The selection process for each article was performed by two independent reviewers (R. H, A. E, or M. R), and the status of each citation (removed/entered) and the reason for rejection was entered in a spreadsheet. In cases of discrepancy between investigators, a third reviewer assessed the citation (K. H). The eligibility criteria for studies are reported in Table 2. Our primary endpoint consisted of the occurrence of SND and PPM implantation due to SND. We Defined SND as “the occurrence of symptomatic sinus bradycardia with a rate less than 50 bpm, sinus arrest or sinus pauses longer than 3 s, symptomatic sick sinus syndrome, tachy-brady syndrome,” atrial fibrillation/flutter with a rate less than 60 bpm, or a junctional rhythm when the temporary pacing was removed (11). All-cause mortality and death due to major cardiovascular events were our secondary endpoints. The essential patient characteristics for included studies consisted of age, gender, and race.

The full text of all selected studies was obtained and data regarding study characteristics, subject characteristics (age, gender, race, reported comorbidities, reported baseline characteristics), performed procedures (use of cardiopulmonary bypass, cardiopulmonary bypass time, aortic cross-clamp time), and incidence of desired outcomes (Number of early post-op sinus node dysfunction (in-hospital), late post-op sinus node dysfunction (post-discharge), post-operation PPM implantation, and all-cause mortality) was extracted into an Excel sheet. Data were extracted from tables or figures when possible, or authors were contacted If required data was not reported in the full text. The data extraction process was carried out by two reviewers (A. E and M. R) and was checked for accuracy by another reviewer (R. H). Disagreements were resolved with the consultation of a third reviewer (K. H).

RoB 2: a revised tool for assessing the risk of bias in randomized trials, ROBINS-I: a tool for assessing the risk of bias in non-randomized studies of interventions, the Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomized (cohort, case–control, cross-sectional) studies, and JBI critical appraisal tool for case-series were used for the quality appraisal of selected studies (12–16).

Mean ± SD and 95% CI for baseline characteristics and outcomes of all included studies have been reported with analytic weights. Due to high heterogeneity and variation in methods of SND measurement, pooled estimates were only calculated for PPM implantations. The total prevalence was estimated in each study using data from the most extended follow-up available. The pooled prevalence values and associated 95% confidence intervals were computed using random effects models through Freeman-Tukey double arcsine transformation and back transformation using Restricted maximum-likelihood estimator for between-study variance(τ²) (17, 18). Both Cochran’s Q test and the I2 statistic were employed to assess the heterogeneity of the studies. We also used τ² statistics estimated with the restricted maximum likelihood method to evaluate the heterogeneity of studies. A τ of 0.04 and lower was interpreted as indicating low heterogeneity, whereas a value higher than 0.36 was considered highly heterogeneous. Other intermediate values were interpreted as medium heterogeneity (19).

To determine the impact of the participant’s age, gender, left ventricle ejection fraction (LVEF), body mass index (BMI), CBP time, and aortic XCL time, we conducted a sensitivity analysis by meta-regression analyses. Subgroup analyses were conducted for different intervention groups (valve, maze, valve-maze, combined surgeries, other). Additionally, we calculated prediction ranges to give a range of anticipated prevalence for PPM implantation among people having cardiac operations. Although there is no agreed-upon definition of what constitutes a positive result in a proportional meta-analysis and the assumption that positive results are more frequently published is not necessarily valid for proportional studies, we evaluated publication bias using funnel plots and The Begg and Mazumdar test (20, 21). All statistical analyses were performed using R (version 4.2.1), packages meta version 5.5.0, metafor 3.4.0, and STATA software (version 16.0, STATA Corp., College Station, Texas).

Template data collection forms, data extracted from included studies, data used for all analyses, analytic code and any other materials used in the review is available for readers upon request from the corresponding author.

Our initial search yielded 2012 individual results, 1830 of which were excluded through the screening phase. From the remaining 182 studies, the full text of only 153 articles was available. Two independent reviewers thoroughly examined these articles and checked for inclusion and exclusion criteria, excluding 66 articles. The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram is presented in Figure 1.

The mean age of study participants ranged from 22.40 to 77 years, with a weighted average of 61.95 ± 7.67 years. Regarding the gender composition of studies, 36.92% ± 12.72% of patients were female. Hypertension (HTN) accounted for the most prevalent comorbidities, with a mean of 53.15% reported by 31 studies (Figure 2). It should be noted that the prevalence of heart failure (HF) was remarkedly high, 37.12%, whereas it has been reported by a few studies with all 3,246 patients. Peripheral vascular disease (PVD) was reported in 10 studies with 4,553 patients, with the least prevalence of 5.18%. Summary statistics for baseline patient and intervention characteristics and comorbidities are shown in Table 3, furthermore, a detailed description of all patient comorbidities is presented in Supplementary Table S2.

The studies that were included were highly qualified overall. The RoB-2 tool used five separate bias assessments: reporting, performance, attrition, detection, and selection. The only RCT included had a low risk for reporting, attribution, and selection, but there were few concerns about performance and detection using the RoB-2 tool. According to the ROBINS-I tool, all studies had a low to moderate risk of bias, except for one non-RCT research with a serious risk of bias. Also, in the case series group judged by the JBI critical appraisal tool, all the studies achieved an acceptable reporting quality except for one that needed further information (22). High-quality observational studies were included, according to the Newcastle-Ottawa Scale (score > 6) used to evaluate non-randomized research. Detailed information about the quality of all included studies is presented in Supplementary Tables S3–S7.

Forty-two studies reported the occurrence of SND after cardiac surgery, forming 48 intervention arms in total. Detailed regarding study design, interventions, follow-up, sample size, age, gender composition, cardiopulmonary bypass time, aortic cross-clamp time, and outcome measures are available in Supplementary Table S8. The weighted mean age of participants equaled 56.95 ± 11.18, ranging from 22.4 to 73.23; on average, 44.48% ± 11.70% were female. The pooled prevalence of SND among studies was 13.71% (95% CI 8.13–20.33).

From 87 studies that entered the review, 78 were eligible for meta-analysis, constituting a total number of 83 intervention groups. The results of this study suggest that 2.874% (95% CI 2.09–3.76) of patients undergo PPM implantation due to SND after cardiac surgeries, with significant variation between different interventions (p-value = 0.0023). The results of the meta-regression suggested no significant relationship between PPM implantation and CPB time (p = 0.21) or XCL time (p = 0.35) which indicates that this variation has not been a result of differences in CPB time or XCL time. Moreover, no significant correlation was found between PPM implantation patients’ age, gender, LVEF, and BMI. Since this meta-regression was conducted on population levels because these factors had not been separately reported for participants with and without PPM, a more robust analysis was not feasible.

The incidence of PPM implantation due to SND in the first month after surgery 2.707% (95% CI [1.65; 3.95]), which does not significantly differ from the overall prevalence of PPM implantation (2.874% [95% CI 2.09–3.76]). This may indicate the higher risk of PPM implantation in first post-surgical month. However, this interpretation may be compromised by the fact that many of included studies had only reported PPMs implanted in first month and did not further follow-up patients. This was the case in 32 of 78 studies. Comparing different surgery types, patients who underwent lone valve surgeries had the lowest incidence (0.789% [0.332; 1.384]), whereas combined valve and maze surgeries had the highest incidence of PPM implantation (6.626% [3.497; 10.472]; Figure 5).

Among the four main intervention groups, lone maze surgery was associated with the highest prevalence of post-operation PPM implantation (4.93% CI [3.24; 6.92], Figure 3) (29–33). This might be due to the nature of the maze surgery, in which lesions are imposed on atria that lead to conduction disorders in tissues surrounding the sinus node, hence sinus node dysfunction (34). A higher incidence of PPM implantation in combined surgeries that include maze procedures in comparison to those without maze also favors this interpretation (5.174% vs. 2.415%, p-value = 0.0594, Figure 4). This observed insignificant p-value may be due the diluted effect of the maze procedure in studies that report combined cardiovascular surgeries, because in these studies all participants are not subjected to the maze procedure.

Furthermore, the pooled population of “lone maze” studies are significantly more male dominant in comparison to “valve,” “valve-maze,” and “combination” groups (84.08% CI [69.39; 94.72] vs. 48.40% CI [41.92; 54.91], 48.27% CI [43.26; 53.30], and 54.95% CI [50.66; 59.21], respectively). Findings of other meta-analyses regarding pacemaker implantation after transcatheter aortic valve replacement suggest that men are more likely to receive a PPM, and such a trend may also exist in maze procedures or cardiac surgeries (35, 36). However, to the best of our knowledge, such analysis has not been performed on PPM implantation after the maze procedure. As for the incidence of sinus node dysfunction, this outcome was only reported in 2 studies with values of 4.32% and 14.63% (29, 37).

The pooled prevalence of PPM implantation in 2,582 patients undergoing lone valve surgery from 13 studies equaled 1.01% CI [0.23; 2.14], making it the lowest prevalence among all intervention groups. The primary intervention in these studies was aortic valve replacement in 3, mitral valve surgeries in 6, and multiple valve procedures in 4 studies. The largest study in this group was Salmi 2020, with 704 participants, which focused on isolated aortic valve replacement with bioprostheses (38). Cleveland et al. reported the highest PPM implantation prevalence among these studies, with 6 of 50 patients receiving PPM (39). Seven studies have reported the prevalence of SND with a weighted average of 7.69 ± 10.42, ranging from 0 in studies by Guiraudon et al. and Tuinenburg et al. to 31.46% in a study by Kumar et al. (40–42). The largest study in this group used Rapid-deployment Intuity® and conventional bioprostheses for aortic valve replacement in 924 patients, leading to 0.86% incidence of SND (43).

Thirteen studies reported the incidence of PPM implantation following operations with concomitant valve and maze procedures, establishing an aggregate prevalence of 2.90% CI [0.61; 6.28], with prevalence ranging between 0% and 14.28% (44–46). Among included studies, 12% of the weight was appointed to a study with a prevalence of 0.49% for PPM implantation (47). Based on reported data from 8 studies and 360 participants, a weighted average of the incidence of SND in the “valve-maze” group equaled 10.55% ± 19.87% (41, 44–46, 48–51).

Combined cardiac surgeries were the most common intervention among all entered studies, with 50 intervention groups reporting PPM implantation in 34535patients and 29 studies reporting SND. The pooled estimate for the prevalence of post-operative PPM implantation was 3.308% CI [2.28; 4.489], with a range of 0% to 27.78% (44, 45, 52). One study by Feldman et al. reported a prevalence of 27.78%, which is an outlier compared to studies with a similar population. This was probably because it only reported the characteristics of patients that had received a PPM post-operatively since it aimed to evaluate pacemaker dependency. This magnifies the prevalence in comparison to the regular target population (52). However, the sensitivity analysis suggests that omitting this study would not significantly influence the pooled estimate of the analysis (Supplementary Figure S8).

Four of these studies had populations of more than 1,000 patients. Waddingham et al. screened 5,950 patients for Cardiac implantable electronic device (CIED) implantation after cardiac surgery, during the same admission, from 2015 to 2018 and reported 250 implants, 30 of which were due to sinus node dysfunction (PPM prevalence 0.54% CI [0.34–0.72]) (53). Rizzo et al. followed 3,493 patients for 33 months on average and identified that 9 of 45 patients received post-operative PPM due to Sick sinus syndrome (54). From 1,234 patients who underwent concomitant CABG and valve replacement surgery, Al-Ghamdi and colleagues identified 20 patients who received PPM (55). The indication was sick sinus syndrome with symptomatic bradycardia in three (15%) and atrial fibrillation (AF) with a slow ventricular rate in four (20%), which, based on our definition, are classified as sinus node dysfunction (55). Finally, Bis et al. reported 185 cases of PPM implantation following cardiac surgeries in 15,902 patients, over 11 years. However, only 48 of these cases fitted our definition of SND. Unfortunately, no report of timing for PPM implantation was provided in their paper (56).

The subgroup analysis of combined cardiac surgeries based on their inclusion of the maze procedure demonstrates a lower prevalence of PPM implantation in the patients spared from the maze procedure; however, this was not statistically significant (combined with maze: 4.05% CI [2.74; 5.58]; combined without maze: 2.41% CI [0.97; 4.31]; p-value = 0.2514). As mentioned before, in combined surgeries, all the participants in a study may not have undergone maze surgery along with other procedures; the observed effect is diluted.

The weighted average of SND’s prevalence in the combined surgeries was 9.34% ± 14.10% with a range of 0 to 80. From these 29 intervention groups, four groups were outliers on the higher end of the range. The detailed results of these studies are further explained. From 15 patients followed by Pasic et al., 12 manifested sinus node dysfunction in the form of severe sinus bradycardia, sinus pauses or sinus arrest, sinoatrial exit block, atrial tachyarrhythmias, alternating periods of atrial bradyarrhythmias and tachyarrhythmias during the first 3 months after the surgery (57).

In another study from Pasic et al., 36 patients demonstrated sinus node dysfunction post-operatively, receiving temporary pacing using epicardial wires. Furthermore, five of these patients received a PPM afterward (58). Szalay et al. identified 38 out of 52 patients with SND in the form of junctional rhythm, 8 of which were classified as PPM dependent during the follow-up (59). In another study, 141 patients with valvular heart disease and coronary heart disease complicated with atrial fibrillation under combined intervention: using penetrating technique radiofrequency exposure to treat their atrial fibrillation (60). Forty-one percent of these patients showed signs of SND early after the operation (60).

Comparing combined interventions with and without the maze procedure, the weighted average of SND prevalence in the former group was higher than in the latter (12.32 ± 17.86 and 6.73 ± 9.46, respectively); however, this was not statistically significant (p-value = 0.303).

The limitations of the selected research limit our investigation. To limit the domination of larger studies on overall results, we used random effect models to weight the studies. The results of the current study might have been influenced by significant clinical and statistical heterogeneity. Factors such as various methodological designs, clinical settings, and patient characteristics may have contributed to this outcome. However, such factors are inseparable components of the nature of meta-analysis of proportions (61). In the current study, heterogeneity may lie in the methodological diversity of the studies and great variety in performed interventions, which is expected in prevalence data. Nevertheless, the results of the leave-one-out analysis suggest that no single study influenced the overall results of the meta-analysis. The bias from observational studies and the lack of a control group might have been another source of bias in our review. Moreover, as identification of SND and PPM implantation had not been the primary outcome of many studies, and it had only been reported as a secondary or tertiary outcome, we faced many missing data that limited the robustness of our interpretation. No evidence of publication bias was observed. As a result, the findings should be carefully considered and need additional research.

The guideline is silent about the incidence rate of SND and related PPM implementation after major cardiovascular surgeries. Based on the present report, patients undergoing the maze and maze-valve procedures are at higher risk of post-op SND. We did not find a statistically significant relationship between age, gender, CBP time, and XCL time.