Preoperative Antihypertensive Medication in Relation to Postoperative Atrial Fibrillation in Patients Undergoing Cardiac Surgery: A Meta-Analysis

Background. We undertake a systematic review and meta-analysis to evaluate the effect of preoperative hypertension and preoperative antihypertensive medication to postoperative atrial fibrillation (POAF) in patients undergoing cardiac surgery. Methods. We searched PubMed, Embase, and Cochrane Library (from inception to March 2016) for eligible studies. The outcomes were the effects of preoperative hypertension, preoperative calcium antagonists regimen, preoperative ACE inhibitors regimen, and preoperative beta blocking agents regimen with POAF. We calculated pooled risk ratios (OR) and 95% CIs using random- or fixed-effects models. Results. Twenty-five trials involving 130087 patients were listed. Meta-analysis showed that the number of preoperative hypertension patients in POAF group was significantly higher (P < 0.05), while we found that there are no significant differences between two groups in Asia patients by subgroup analysis, which is in contrast to other outcomes. Compared with the Non-POAF group, the number of patients who used calcium antagonists and ACE inhibitors preoperatively in POAF group was significantly higher (P < 0.05). And we found that there were no significant differences between two groups of preoperative beta blocking agents used (P = 0.08). Conclusions. Preoperative hypertension and preoperative antihypertensive medication in patients undergoing cardiac operations seem to be associated with higher risk of POAF.


Introduction
Atrial fibrillation (AF) is a supraventricular tachyarrhythmia characterized by uncoordinated atrial activation with deterioration of mechanical function [1]. Postoperative atrial fibrillation (POAF) is the most common perioperative cardiac arrhythmia [2]. Although many studies have tried to assess risk factors for POAF, it remains incomplete and unclear. Moreover, patients with POAF have increased risk of stroke, other arrhythmias, cardiovascular mortality, and higher health-care costs compared with patients without POAF [3][4][5]. Constantly ascertaining the risk factors and prevention of POAF is of great importance for the physician.
Patient characteristics resulting in POAF are many. Some studies have confirmed that preoperative hypertension is a danger factor to POAF in patients with cardiac surgery [6][7][8], while a lot of studies suggestion that the number of preoperative hypertension patients has not significantly statistical differences between POAF group and Non-POAF group [9,10]. So we undertook a systematic review and metaanalysis to evaluate the effect of preoperative hypertension to POAF in patients with cardiac surgery.

Methods
This systematic review was performed according to the guidelines of the preferred reporting items for systematic reviews and meta-analyses (PRISMA) [6]. We prospectively registered our system review at PROSPERO (Registration number: CRD42016038101). The proposed study will utilize published data; as such there is no need for ethical approval.

Data Sources and Search
Strategy. The PubMed, Embase, and Cochrane Library databases were searched from inception to February 2016 for relevant studies investigating the association between preoperation hypertension and POAF in cardiac surgical patients. The following search terms were used: Atrial fibrillation, "Fibrillation, atrial", Familial atrial fibrillation, Auricular fibrillation, "Fibrillation, auricular", Postoperative period, "Period, postoperative", Cardiac surgical procedures, "Procedure, cardiac surgical", "Surgical procedure, cardiac", "Surgical procedures, heart", Cardiac surgical procedure, Heart surgical procedures, "Procedure, heart surgical", "Surgical procedure, heart". A manual search of the reference sections of included trials, published metaanalyses, and relevant review articles was conducted to identify additional articles. If duplicated data were shown in several studies, only the most recent, largest, or most complete study was included.
Original studies included in our meta-analysis had to meet the following criteria: (1), an observation human study; (2) investigating the association between preoperative hypertension and POAF in cardiac surgical patients; (3) providing sufficient data to calculate them. Only English language studies were chosen.

Data Extraction and Assessment of Study Quality.
Patient characteristics (authors, number of patients, year of publication, ASA rating, age, gender, type of surgery and anaesthesia, and endpoint) were recorded. If the data mentioned above were unavailable in the article, the corresponding authors were called upon for missing information. All of the data were independently extracted using a standard data collection form by both authors, and then the collected data were checked and entered into Review Manager analysis software (RevMan) Version 5.3. All discrepancies were checked, and a consensus was achieved by discussion. A record of reasons for excluding studies was kept.
The order from higher to lower in the quality of studies was the following: (1) prospective cohort study, (2) retrospective cohort study, and (3) case-control study. We evaluate the quality of the studies by Newcastle-Ottawa Scale (NOS) [11]: a maximum of nine points to each cohort study (four for quality of selection, two for comparability, and three for quality of outcome and adequacy of follow-up) and a score of nine points to every case-control study (four for quality of selection, two for comparability, and three for quality of exposure). The score of each study less than 6 was regarded as a low-quality study; otherwise, it was a high-quality study.

Statistical Analysis.
The risk ratio (RR) with 95% CI was used as a common measure of the effect between the two groups. The meta-analysis was carried out using Review Manager, version 5.3 (The Cochrane Collaboration, Software Update, Oxford, UK). Statistical heterogeneity across studies was usually investigated using the 2 statistic. When 2 values of less than 50% were determined, heterogeneity could be accepted, and the fixed-effects model was expected to be adopted. Otherwise, the randomized-effects model was adopted, and we investigated the influence of a single study on the overall pooled estimate by omitting one study in each turn. A value of <0.05 was considered statistically significant.
To explore potential sources of heterogeneity among studies, we performed four sets of subgroup analysis: by Non-English (n = 1) No available data for outcomes of interest (n = 9) Heart transplantation (n = 1) Not cardiac surgery (n = 1) Not atrial fibrillation (n = 2) 25 studies included in qualitative synthesis 25 studies included in meta-analysis study design (cohort versus case-control studies), by type of cardiac operation performed (coronary artery bypass grafting (CABG) only versus aortic valve replacement (AVR) only), by NOS scores (less than 6 versus with or higher 6), and by different region (Asia versus Europe versus America versus Oceania). We also performed a sensitivity analysis by excluding studies where the association between hypertension and POAF was opposite to the one from others.

Identification of Eligible Studies.
In total, 252 potentially relevant abstracts were identified. After duplicates were removed, 247 unique abstracts remained. After examining the abstracts, 39 publications seemed to meet the inclusion criteria. Of these, 14 were excluded for the following reasons: Non-English language [32], no available data on the outcome of interest in [33][34][35][36][37][38][39][40][41], heart transplantation [42], no cardiac surgery [43], and no atrial fibrillation [44,45]. Finally, the remaining 25 studies [6-10, 12-31] to existing data met our selection criteria and were included in the systematic review. A flow diagram of the search strategy and study selection is illustrated in Figure 1.

Study Characteristics.
The characteristics of all included studies were presented in Table 1. Most patients underwent CABG only; three studies reported Aortic valve replacements alone [14,24,30]. The type of cardiac operation in five studies contained CABG and valve operation [6,13,15,22,23]. These studies were published between 1997 and 2014. All studies clearly indicated the study population and defined the outcome. Sample size of included studies varied from 53 to 49264.

The Effect of Preoperative Hypertension to POAF.
The aggregated results were studied in 25 trials [6-10, 12-31] and illustrated in Figure 2. Heterogeneity was noted among the studies ( 2 = 54%; = 0.0008), and a randomized-effects model was chosen. The results indicate that the number of preoperative hypertension patients in POAF group was significantly higher than Non-POAF group (RR = 1.07, 95% CI: 1.05-1.09, < 0.00001). After investigating the influence of a single study on the overall pooled estimate by omitting one study in each turn, we found that the 2 is still higher than 50 except for one study [8]. When we omit the study of Almassi's [8], the 2 drop to 38% and a fixed-effects model was selected, and the outcome between two groups has a significant difference (RR = 1.06, 95% CI: 1.05-1.07, < 0.00001) (Figure 3).

Discussion
This is the first time meta-analysis to discuss the effect of preoperative hypertension to POAF. The pooled meta-analysis of 25 studies suggested that patients who have hypertension before operation were easier to develop AF postoperative. Preoperative ACE inhibitors and calcium antagonists regimen may be risk factors for POAF in patients undergoing cardiac surgery. In addition, patients with preoperative beta blocking agents were not linked with POAF.
People with hypertension are liable to suffer left atrial enlargement by the increased cardiac afterload, which was leading to atrial remodeling following the progression of disease [46]. Hypertension caused left ventricular hypertrophy and increases left ventricular stiffness, decreases coronary flow reserve, wall stress, and filling pressure and increases the activation of the sympathetic nervous system, which are associated with AF occurrence. At the same time, the proliferation and differentiation of fibroblasts into myofibroblasts cause disturbances in extracellular matrix. Studies suggested that the cardiac extracellular matrix remodeling was significantly changed in the hypertensive patients with Heterogeneity: 2 = 10.11, df = 6 (P = 0.12); I 2 = 41% Test for overall effect: Z = 5.46 (P < 0.00001) Almassi et al. 1997  AF [47,48]. This may help explain why we found the numbers of preoperative hypertension patients in POAF group to be larger than Non-POAF group ( < 0.00001). Following consideration of the heterogeneity, we performed sensitivity analysis and subgroup analysis to talk about the potential reasons. We performed four sets of subgroup analysis and found the heterogeneity still exists in some groups (  to POAF in on-pump coronary artery bypass group. Studies have confirmed a lower incidence of POAF in off-pump coronary artery bypass patients [49,50]. Other studies in our meta-analysis did not state the type of CABG similar to Almassi's study, which may be the cause of heterogeneity. Antihypertensive drugs are prescribed mainly to reduce the morbidity and mortality caused by hypertension and its complications, while we observed that preoperative application of calcium antagonists and ACE inhibitors regimens is meaningful risk to POAF ( < 0.05). Read through those articles and attempt to find the confounding factors that influence the outcomes. Most of the studies did not perform logistic regression analysis except three studies [16,27,29]. Although values in three studies are all greater than 0.05, the OR values in two studies [16,29] are greater than 1 ( Table 2). It may be telling us that calcium antagonists and ACE inhibitors regimens have a positive effect on POAF, in spite of no significant difference existing. Several reasons contribute to the outcome: (1) preoperative application of calcium antagonists and ACE inhibitors regimens is significant risk factors to POAF and needs more studies to confirm it; (2) some confounding factors affect the results and need further analysis; (3) the myocardial excitability is higher in patients who apply calcium antagonists and ACE inhibitors preoperatively, for they usually stop these drugs postoperatively.
We recognize several limitations in our analysis. First, we performed an unavailable meta-analysis, and no access to individual patient data from individual studies was available. Second, this was a meta-analysis of observational studies. Subgroup analyses for some potential confounding of the association between preoperative hypertension and POAF were performed, and we did not find any strong subgroup effects. Third, none of the studies reported the association of different level hypertension in POAF in cardiac surgery patients, and therefore, we could not analyze this. Finally, this meta-analysis was based on studies published in the English language and unpublished literature could be missing, which may have generated bias.

Conclusion
Preoperative hypertension in patients undergoing cardiac operations seems to be associated with higher risk of POAF. Considering the limitations of this study, our finding should be reviewed with caution, and large-scale studies are needed to confirm our findings.