OSA-induced multi-organ dysfunction after elective coronary artery bypass surgery in coronary heart disease patients

Background The aim of this study was to explore the underlying impact of obstructive sleep apnea (OSA) on postoperative parameters of multi-organ function, including cardiac and cerebral vascular, respiratory as well as renal postoperative complications, among coronary heart disease (CHD) patients following elective coronary artery bypass grafting (CABG). Methods Electronic literature databases, including PubMed, ISI Web of Science, Directory of Open Access Journals, and the Cochrane Library electronic databases, were searched manually and automatically for relevant English articles. All of the selected articles focused on a comparison of the incidence of primary and secondary outcomes in CHD patients undergoing elective CABG with and without OSA. Results A total of 13 articles met our inclusion criteria. The current study demonstrated OSA signicantly increased the incidence of major adverse cardiac and cerebrovascular events (MACCEs) in CHD patients undergoing elective CABG compared with the controls (odds risk (OR), 1.97; 95% condence interval (CI), 1.50 to 2.59, p <0.0001). In addition, OSA was associated with an increased risk of new revascularization in CHD patients undergoing elective CABG (OR, 9.47; 95% CI, 2.69 to 33.33, p <0.0001). Moreover, reintubation and tracheostomy in the OSA group was increased (OR, 3.43; 95% CI, 1.35 to 8.71; p =0.009) and 372% (OR, 4.72; 95% CI, 1.23 to 18.13; p=0.024), respectively, compared with the control group. Besides, we also conrmed OSA signicantly increased the acute kidney injury (AKI) incidence by 124% (OR, 2.24; 95% CI, 1.07 to 4.72; P < 0.0001). Finally, our results demonstrated that OSA increased medical resource utilization including length of postoperative hospital stay and ICU stay. Conclusions OSA may contribute to postoperative multi-organ dysfunction among CHD patients undergoing elective CABG by increasing the incidence of MACCEs, especially new revascularization, as well as respiratory, and renal complications. The current study demonstrated OSA signicantly increased the incidence of MACCEs in CHD patients undergoing elective CABG compared with the controls (OR, 1.97; 95% CI, 1.50 to 2.59, p < 0.0001; Fig. 1A and Table 3) without signicant heterogeneity (I 2 = 9.3%) and publication bias (p = 876). The funnel plot also showed low publication bias (Fig. ii online). No signicant effects were detected in meta-regression analysis with respect to the examined covariate (p = 0.665 Table 3); namely, the valvular surgery. In addition, OSA was associated with an increased risk of new revascularization, including CABG and percutaneous coronary intervention (PCI), in CHD patients undergoing elective CABG (OR, 9.47; 95% CI, 2.69 to 33.33, p < 0.0001; Fig. 1D and Table 3) without signicant heterogeneity (I 2 = 0.0%). Since all the included patients in this endpoint underwent CABG alone, the confounding factor had no effects on the pooled outcome. Furthermore, the current meta-analysis conrmed OSA did not affect the incidence of postoperative MI and CVA or TIA (Fig. 1B, 1C and Table 3). and postoperative management of CHD patients undergoing elective CABG. Previous studies investigating the effects of OSA on postoperative cardiac events among CHD patients following CABG show controversial in follow-up. sleep-disordered of hospital stay ICU stay. length of stay (LOS) in for with ventilatory CABG, of multi-organ dysfunction. the and of peri-operative phase therapy effective therapy for OSA, CPAP may optimize the condition of surgical patients with CI:Condence Interval; MACCEs:Major Adverse Cardiovascular and Cerebral Events; MI:Myocardial infarction; TIA:Transient Ischemic Attack; CVA:Cerebral Vascular Accident; PCI:Percutaneous Coronary Intervention; AKI:Acute Kidney Injury; CHD, Coronary Heart Disease; POAF:Postoperative Atrial Fibrillation; STS, Society of Thoracic Surgeons; ARDS:Acute Respiratory Distress Syndrome.


Background
Obstructive sleep apnea (OSA), a syndrome caused by repeated narrowing of the throat during sleep, is a common breathing-associated sleep disorder with recent estimates of prevalence in the general adult population ranging from 9 to 38%. 1, 2 In coronary heart disease (CHD) patients undergoing elective coronary artery bypass grafting (CABG), the incidence of mild OSA (5 ≤ apnea-hypopnea index [AHI] < 15) and moderate to severe OSA (AHI ≥ 15) is 74% and 48%, respectively. 3 Despite the high prevalence of OSA among CHD patients following CABG, the impact of OSA in the postoperative multi-organ function is not well established. Most of the evidence is focused on arrhythmia recurrence after cardiac surgery, 4,5 the effects of OSA on respiratory as well as renal parameters still need to be investigated further. In addition, as for the cardio-cerebrovascular complications, recent studies are based on patients diagnosed with different cardiac diseases undergoing various types of elective cardiac surgeries. 6−8 Data on CHD patients developing postoperative complications after CABG are sparse. As far as now, no meta-analyses have focused on documenting a direct correlation between CHD patients following elective CABG with OSA and higher risk of adverse cardiac and cerebral vascular events.
This current meta-analysis was initiated to ll the above knowledge gaps. We particularly focused on CHD patients with or without OSA. The aim of this study was to explore the underlying impact of OSA on postoperative parameters of multi-organ function, including cardiac and cerebral vascular, respiratory as well as renal postoperative complications, among CHD patients following elective CABG. Recognition of the association and determinants of OSA on postoperative multiple organ dysfunction should lead to strategies to improve the prognosis of CHD patients undergoing elective CABG.

Literature search and selection criteria
This systematic review and meta-analysis were conducted and reported in adherence to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 9 Electronic literature databases, including PubMed, ISI Web of Science, Directory of Open Access Journals, and the Cochrane Library electronic databases, were searched manually and automatically for relevant English articles. All of the selected articles focused on a comparison of the incidence of primary and secondary outcomes in CHD patients undergoing elective CABG with and without OSA.
As for further full-text assessment. The references of the retrieved articles were also reviewed to identify additional eligible studies. The study selection program was shown in Figure i online. Studies meeting the following criteria were included: The following search terms were used: ("sleep apnea" [All Fields] OR "sleep apnea" [All Fields]) AND ("CABG" [All Fields] OR "coronary artery bypass grafting" [All Fields]) AND ("major adverse cardiac and cerebrovascular events" [All Fields] OR "cardiovascular outcomes" [All Fields] OR "major cardiovascular and cerebrovascular adverse events" [All Fields]).
The last search was conducted on July 1, 2019 Two investigators (J.W. and W.Y.) performed the initial search separately, deleted duplicate records, screened titles and abstracts for relevance, and identi ed relevant article 1. Patients: subjects were CHD patients following elective CABG. 5. Study design: observational studies (prospective or retrospective cohort studies) were included in this meta-analysis.

Data Extraction and Quality Assessment
Two reviewers (J.W. and W.Y.) independently extracted the following information from the eligible studies: rst author, year of publication, country, study design, patient characteristics, number of patients enrolled, intervention, and outcome data. When the same patients were reported in several publications, only the largest study was retained to avoid duplication of information. The current study used the Downs and Black score system to evaluate the quality of each study. The quality of studies with a score≥20 was considered good and a score < 20 was considered poor.

Study End Points and OSA De nition
The AHI was de ned as the average number of episodes of apneas and hypopneas per hour of sleep. OSA must be diagnosed by polysomnography (PSG), which was de ned in events per hour as AHI ≥5 or screened as "high-risk OSA" by screening questionnaire.
Diagnosis of OSA and pre-/postoperative continuous positive airway pressure (CPAP) use in patients with OSA were shown in Table 1

Statistical Analysis
Differences are expressed as odds risk (OR) or standardized mean difference (SMD) with 95% con dence interval (CI). Study heterogeneity was tested using the I 2 statistic. Heterogeneity was considered to be low in studies with an I 2 between 25% and 50%, moderate in studies with an I 2 between 50% and 75%, and high in studies with an I 2 exceeding 75%. I 2 exceeding 50% represents signi cant heterogeneity. A xed-effects model was used when study heterogeneity was not signi cant; a random effects model was used when study heterogeneity was signi cant. Funnel plots and the Egger linear regression test were used to estimate publication bias. Meta-regression analysis was performed to examine whether the pooled outcomes could be modi ed by the confounding factor. A Q model statistic was used in the meta-regression. All p values are two-sided.

Study Selection, Characteristics and Quality Assessment
The selection process is illustrated in Fig. i online. A total of 13 articles met our inclusion criteria. 3, 11−22 The main characteristics of the included studies are described in Table 2. Total primary and secondary pooled outcomes in patients with and without OSA were demonstrated in Table 3. Of the 13 reports, three contained CHD patients following elective CABG + valvular surgery. 3,12,19 Meta-regression analysis was performed to examine whether the above pooled outcomes could be modi ed by the valvular surgery. The results were shown in Table 3. In addition, Downs and Black Score analysis revealed that all of the 13 articles had good quality (score ≥ 20, Table 2).  The impact of OSA on postoperative multi-organ function in CHD patients following elective CABG

Postoperative cardiac and cerebrovascular parameters
The current study demonstrated OSA signi cantly increased the incidence of MACCEs in CHD patients undergoing elective CABG compared with the controls (OR, 1.97; 95% CI, 1.50 to 2.59, p < 0.0001; Fig. 1A and Table 3) without signi cant heterogeneity (I 2 = 9.3%) and publication bias (p = 876). The funnel plot also showed low publication bias (Fig. ii online). No signi cant effects were detected in meta-regression analysis with respect to the examined covariate (p = 0.665 Table 3); namely, the valvular surgery. In addition, OSA was associated with an increased risk of new revascularization, including CABG and percutaneous coronary intervention (PCI), in CHD patients undergoing elective CABG (OR, 9.47; 95% CI, 2.69 to 33.33, p < 0.0001; Fig. 1D and Table 3) without signi cant heterogeneity (I 2 = 0.0%). Since all the included patients in this endpoint underwent CABG alone, the confounding factor had no effects on the pooled outcome. Furthermore, the current meta-analysis con rmed OSA did not affect the incidence of postoperative MI and CVA or TIA (Fig. 1B, 1C and Table 3).

Postoperative respiratory parameters
The current study demonstrated that reintubation in the OSA group was increased by 243% compared with the control group (OR, 3.43; 95% CI, 1.35 to 8.71; p = 0.009; Fig. 2B and Table 3) but with signi cant heterogeneity (I 2 = 50.0%). Meanwhile, patients with OSA group showed a 372% increase in tracheostomy compared with the control group (OR, 4.72; 95% CI, 1.23 to 18.13; p = 0.024; Fig. 2C and Table 3) without signi cant study heterogeneity (I 2 = 0.0%). Furthermore, we found OSA did not affect the incidence of postoperative major pulmonary complications and pulmonary edema ( Fig. 2A, 2D and Table 3). In meta-regression analysis, we found the examined covariate, namely, the valvular surgery, had no signi cant effect on the above pooled outcomes.

Postoperative renal parameters
We also con rmed OSA signi cantly increased the AKI incidence by 124% (OR, 2.24; 95% CI, 1.07 to 4.72; P < 0.0001; Fig. 3 and Table 3) without signi cant study heterogeneity (I 2 = 0.0%). In meta-regression analysis, we found the examined covariate, namely, the valvular surgery, had no signi cant effect on the above pooled outcomes.
The impact of OSA on postoperative exploratory parameters in CHD patients undergoing elective CABG Our results demonstrated that OSA increased medical resource utilization including length of postoperative hospital stay (SMD, 0.30; 95% CI, 0.12 to 0.47; p < 0.0001; Table 3) and ICU stay (SMD, 0.30; 95% CI, 0.16 to 0.44; p < 0.0001; Table 3) without signi cant study heterogeneity. On the contrary, OSA did not affect the incidence of postoperative in-hospital death, infection and ICU-readmission (Table 3). In meta-regression analysis, we found the examined covariate, namely, the valvular surgery, had no signi cant effect on the above pooled outcomes.

Meta-regression Analysis of the Potential Modi ers
Meta-regression analysis demonstrated the above pooled outcomes could not be modi ed by the valvular surgery (Table 3). Furthermore, the results also shown the above pooled outcomes were not affected by the following modi ers, including con rmation of OSA (PSG or not); quality of study; loss of patients due to follow-up and OSA with CPAP treatment (Table 4).

Discussion
To the best of our knowledge, this is the rst meta-analysis to comprehensively examine the association between OSA and postoperative multiorgan dysfunction in CHD patients undergoing elective CABG. Our key ndings are: 1) compared with the control group, OSA signi cantly increased the incidence of MACCEs in CHD patients undergoing elective CABG. In addition, OSA was associated with an increased risk of new revascularization, including CABG and PCI, in CHD patients undergoing elective CABG; 2) reintubation in the OSA group was increased compared with the control group. Meanwhile, patients with OSA group showed an increase in tracheostomy compared with the control group; 3) OSA signi cantly increased the AKI incidence compared with the control group; 4) we also con rmed the association between OSA and increased medical resource utilization. Meta-regression showed the examined covariates had no signi cant effect on the above pooled outcomes.
In the past 10 years, only one meta-analysis had investigated the effects of preoperative OSA on postoperative MACCEs among patients following cardiac surgery. 6 However, the heterogeneity of the results cannot be ignored (I 2 = 64%). Previous meta-analysis focused on patients with varying cardiovascular diseases, including CHD, valvular heart disease and aortic disease, which might account for the source of the huge heterogeneity and led to inaccurate pooled results. In this current meta-analysis, we reduced the heterogeneity by including only CHD patients alone. Previous studies investigating the effects of OSA on postoperative cardiac events among CHD patients following CABG show controversial results.
For instance, Uchôa and colleagues 14 did not nd a signi cant increase in MACCEs (including POAF) among CHD patients undergoing CABG with OSA in the short-term follow-up. On the contrary, Tafelmeier and colleagues 13 demonstrated sleep-disordered breathing, particularly OSA, is associated with adverse cardiac events after CABG, independent of known confounders. The ndings of our study are in accordance with previous studies that found OSA signi cantly increased the incidence of MACCEs in CHD patients undergoing elective CABG. Notably, the current metaanalysis rstly con rmed OSA was associated with an increased risk of new revascularization. The mechanisms underlying OSA-associated adverse cardiac events among CHD patients undergoing CABG may be correlated to the following aspects: endothelial function, arterial stiffening, as well as systemic in ammation. Our previous meta-analysis of 18 studies aimed to investigate the mechanisms underlying OSA-associated cardiovascular events. 1 We demonstrated OSA, particularly moderate-severe OSA, was signi cantly associated with impaired endothelial function (measured by ow mediated dilation and nitroglycerin-induced dilation), increased arterial stiffness (determined by carotid-femoral pulse wave velocity and indicates augmentation index), and elevated serum levels of in ammatory markers (such as high-sensitivity C-reactive protein and Creactive protein), independent of known confounders. 1 The above factors were often used to predict cardiovascular events in general and morbid populations, especially in CHD patients undergoing CABG. 23−25 Besides, postoperative troponin T levels, correlated with end-organ damages, suggesting the primary role of an ischemic trigger of end-organ damages. 26 Although the recent study demonstrated no differences in the myocardial infarction occurrence in CHD patients undergoing CABG with and without OSA, previous studies indicated the potential role of OSAassociated myocardial ischemia in triggering adverse cardiac events (including POAF) among cardiac surgical patients. 14,26 High-quality randomized controlled trials are still needed to con rm. Finally, limited by the original literature, the current meta-analysis con rmed OSA did not affect the incidence of postoperative CVA or TIA. However, in an observational study of 392 men and women with CHD referred for coronary angiography, Valham et al. demonstrated OSA is an independent predictor for stroke among CHD patients. 27 Possible factors accounting for OSAassociated stroke among CHD patients include apnea-induced nocturnal cerebral ischemia, hypertension, as well as an increased risk of arteriosclerosis, 28−30 which are further correlated to stroke. It is worth mentioning that, the independence relationship between sleep-disordered breathing and postoperative heart failure has been con rmed, which contributes to the increased risk of MACCEs. 13 Although no differences were observed in the occurrence of major pulmonary complications among CHD patients with and without OSA, reintubation and tracheostomy in the OSA group was increased by 243% and 372%, respectively, compared with the control group. Respiratory failure is a common indication for re-admission to ICU and in-hospital mortality in CHD patients undergoing CABG. 8 Both delayed and premature extubation may be correlated to an increased risk of various complications, 31,32 and the need for reintubation or tracheostomy was viewed as a potential adverse event after CABG. 31−33 Previous studies found that requiring reintubation or tracheostomy was associated with organ dysfunction, especially respiratory failure, secondary to physiologic stress, increased mortality, higher hospital costs and longer length of ICU stay. 34−36 Among CHD patients following CABG, OSA was correlated to a longer intubation time. 3,21 The current research further con rmed that OSA may contribute to respiratory dysfunction and other adverse events by increasing the incidence of postoperative reintubation and tracheostomy in CHD patients undergoing CABG.
Currently, no evidence-based studies investigating the effects of OSA on postoperative renal function among CHD patients following CABG. Available reports on the association between OSA and the incidence of AKI after cardiac surgery show con icting results. In an observational study, Wong et al. found that OSA increased the incidence of postoperative AKI and required renal replacement therapy during hospitalization. 16 In contrast, KAW et al. found that OSA exerted no adverse effects on renal function after cardiac surgery. 19 We rstly con rmed OSA signi cantly increased the AKI incidence (by 124%) compared with the control group. The mechanism underlying the adverse effects of OSA on renal outcomes among cardiac surgical populations still need to be investigated further. An animal model of sleep apnea demonstrated brotic changes and in ammatory in the kidney that were attributed to the impacts of intermittent hypoxia. 37 In addition, previous study also reported nocturnal hypoxemia in OSA patients is correlated to accelerated renal dysfunction, possibly because of increased activity of the renal renin-angiotensin system (RAS) as well as glomerular hypertension. 38 Moreover, CHD patients following CABG with OSA are also at increased risk of developing preoperative hypertension and diabetes, both of which are closely associated with AKI. 39 Moreover, in an observational study, Tafelmeier et al.
assessed the incidence of AKI after elective CABG in CHD patients without sleep-disordered breathing (SDB), with central sleep apnea (CSA), or with OSA. They demonstrated OSA was correlated to impaired renal function as observed by signi cantly higher creatinine values and signi cantly lower glomerular ltration rates than those of patients in the other two groups. 13 Our results also demonstrated that OSA increased medical resource utilization including length of postoperative hospital stay and ICU stay. An increased length of stay (LOS) in ICU for OSA group suggests that CHD patients with OSA may require greater ventilatory support and longer monitoring after CABG, because of OSA-induced postoperative multi-organ dysfunction. Based on this knowledge, peri-and postoperative management of CHD patients with OSA may be optimized to minimize the rate of postoperative cardio-cerebrovascular, respiratory, as well as renal complications, further reducing the consumption of limited medical resources.
Since OSA may contribute to postoperative multi-organ dysfunction among CHD patients undergoing elective CABG, routine sleep apnea screening before CABG need to be conducted for a diagnosis to be rmly established and to determine the type and severity of OSA that will dictate treatment alternatives. PSG is the gold standard for con rming the presence of OSA. 3 However, PSG is costly and time-consuming. 26 This justi ed the development of ambulatory abbreviated cardiorespiratory recording techniques and establishment of guidelines for the use of these devices. 27 In addition, sleep apnea treatment initiation in the peri-operative phase is also recommended. CPAP therapy is the most effective therapy for OSA, and perioperative CPAP use may optimize the condition of surgical patients with OSA. 6 There are some limitations to the ndings of this current research since the absence of randomized controlled trials. Selection bias and treatment bias, inherent to the observational studies, may exist. Moreover, in the current study, heterogeneity was reduced by including only CHD patients alone. Limited by the original literature, we also included CHD patients undergoing CABG + the valvular surgery. Although meta-regression showed the valvular surgery did not affect on our pooled outcomes, high-quality randomized controlled trials focusing on CHD patients undergoing isolated CABG are still needed in the future. In addition, since the sample size of each original studies limitations, currently we cannot further address the effects of mild, moderate as well as severe OSA on multiple organ dysfunction, respectively, in CHD patients undergoing CABG.

Conclusions
In conclusion, the current evidence-based research demonstrated OSA may contribute to postoperative multi-organ dysfunction among CHD patients undergoing elective CABG by increasing the incidence of MACCEs, especially new revascularization (including CABG and PCI), as well as respiratory, and renal complications. Furthermore, our results also demonstrated that OSA increased medical resource utilization including length of postoperative hospital stay and ICU stay. Based on this knowledge, peri-and postoperative management of CHD patients with OSA may be optimized to minimize the rate of postoperative parameters of multi-organ dysfunction, further reducing the consumption of limited medical resources.  CHD, coronary heart disease; CABG, coronary artery bypass grafting; MACCEs, major adverse cardiac and cerebrovascular events; MI, myocardial infarction; CVA, cerebral vascular accident; TIA, transient ischemic attack; PCI, percutaneous coronary intervention; OR, odds risk; CI, con dence interval.