The effects of obstructive sleep apnea and high-sensitivity C-reactive protein on clinical outcome in off-pump coronary artery bypass grafting

This study aimed to investigate the relationship between obstructive sleep apnea (OSA) and high-sensitivity C-reactive protein (hs-CRP) to determine their effects on postoperative complications and clinical outcomes during hospitalization in patients undergoing off-pump cardiac artery bypass grafting (OPCABG). This prospective, single-center study enrolled patients who underwent OPCABG. OSA was evaluated using a portable sleep monitor before OPCABG. Spearman correlation was performed to investigate the relationship between hs-CRP and polygraphy test indicators; regression analysis was performed to determine whether hs-CRP is an independent inuencing factor for postoperative atrial brillation, duration of hospitalization, and hospital cost.


Introduction
Obstructive sleep apnea (OSA) is characterized by repeated or partial obstruction of the respiratory tract during sleep, with an incidence of 9-38% in adults [1]. Intermittent hypoxia induced by OSA could trigger oxidative stress and could damage the vascular endothelium; for these reasons, it is an independent risk factor for coronary heart disease (CHD) and affects its prognosis [2]. Currently, off-pump cardiac artery bypass grafting (OPCABG) is one of the primary treatments of CHD [3]. However, there are few reports on whether OSA is associated with post-OPCABG complications [4]. Traditional OSA severity markers, such as apnea hypopnea index (AHI) have limitations in predicting the prognosis of CABG. In ammatory processes are central to the pathogenesis of vascular diseases in the context of OSA. High-sensitivity C-reactive protein (hs-CRP) is a marker of acute in ammation and has strong sensitivity; a slight increase in hs-CRP level could indicate coronary plaque in ammation or coronary artery wall injury, which is closely related to CHD and its associated negative events [5].
In this study, we aimed to investigate the relationship between OSA and hs-CRP levels and to determine their effects on postoperative complications and clinical outcomes during postoperative hospitalization in patients who underwent OPCABG.

Study design
This prospective, single-center study enrolled patients who underwent OPCABG from January 2019 to December 2019 at Beijing An Zhen Hospital. This study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review board of Beijing An Zhen Hospital of Capital Medical University (approval no.: 2013025). All study participants gave informed written consent prior to participation.

Patients
The inclusion criteria were as follows: age, 40-75 years; with consent to undergo sleep monitoring tests (polygraphy; PG) using a portable sleep monitor; and provision of written informed consent for study participation. We excluded patients with valvular disease combined with other heart diseases, central sleep apnea, severe respiratory diseases (e.g., chronic obstructive pulmonary disease), severe diseases of other organs (e.g., renal failure), body temperature > 37.5℃, and preoperative use of morphine and its analogs, sedative drugs, and/or theophylline. We recorded baseline clinical data, including age, sex, body mass index (BMI), body temperature, pre-existing medical conditions (hypertension, diabetes, stroke), history of smoking, blood biochemistry ndings, and left ventricular ejection fraction (LVEF) bases on echocardiography, and PG test data.
PG tests and diagnostic criteria for OSA Eligible patients were enrolled and admitted to the hospital. PG was performed before OPCABG. Each patient in the sleep monitoring center at Beijing An Zhen Hospital wore a portable sleep monitor (ApneaLink, ResMed, Australia). We used type III PG to detect air ow by nasal catheter, respiratory movement by chest belt, heart rate by electrocardiograph, and arterial oxygen saturation (SaO 2 ) by pulse oximetry. All PG test data were analyzed by two physicians at the Sleep Center of Beijing An Zhen Hospital. In case of disagreement between the two, a third physician participated in the data analysis. Sleep apnea was de ned as the cessation of air ow through the nose and mouth for > 10 s during sleep; hypopnea, a reduction of > 50% in the air ow intensity and ≥ 4% in SaO 2 level during sleep. AHI was de ned as the total number of apnea or hypopnea episodes per hour during sleep (i.e., AHI = total number of apnea or hypopnea episodes/total sleep duration (min) × 60). Moderate-severe OSA was de ned as an AHI ≥ 15/h during a 7-h sleep. Oxygen desaturation index (ODI) ≥ 3% is the number of times that oxygen saturation decreases by > 3% per hour [6]. All moderate-severe patients were recommended for continuous positive airway pressure (CPAP) after discharge.

OPCABG
All patients underwent OPCABG after the PG test. The same cardiac surgeon performed all surgeries. The number of grafts and surgical duration were recorded. The quality of graft anastomosis met the criteria recommended by the Operation Quality Committee of Beijing An Zhen Hospital. Two ultrasound specialists performed echocardiography in all patients, and a single nurse measured the blood pressure and collected blood samples from the patients. After surgery, the patients were monitored in the intensive care unit (ICU) until ventilator removal was feasible and vital signs were stable; they were discharged from the hospital once they could move freely. The physician in the ICU and the cardiac surgeon determined the ICU stay and hospitalization duration. We recorded the following postoperative data: incidence of major adverse cardiac and cerebrovascular event (MACCE), lung infection, post-CABG atrial brillation (POAF); duration of postoperative tracheal intubation; ICU stay; duration of hospitalization, and hospital cost. POAF was de ned as the occurrence of AF within 72 h after surgery. Pulmonary infection was de ned as a postoperative increase in white blood cell count, obvious in ammation based on postoperative chest radiography and computed tomography, and meeting one of the following conditions: sputum examination reveals new characteristic changes and pathogenic bacterium could be cultivated from blood or respiratory secretions.

Blinding
The cardiac surgeon, other participating investigators, and research staff were blinded to the ndings of the PG tests. After the nal enrolled patient was discharged from the hospital in January 2020, all data were revealed to the participating investigators.

Statistical analysis
All statistical analyses were performed using SPSS, version 24.0 (SPSS Inc., Chicago, IL, USA). Continuous variables with a normal distribution were presented as means ± standard deviations and were compared using independent samples t-tests. Continuous variables without a normal distribution were presented as medians (interquartile ranges) and were compared using rank-sum tests. Categorical variables are presented as percentages and were compared using χ2 tests. If the missing quantity of measurement data was < 5%, the average value was used to replace the missing value. No missing values in the counting data were noted. Spearman correlation analysis was employed to determine the relationship between PG test data and hs-CRP; multiple linear regression was used to analyze the relationship between AHI and hs-CRP.
Univariate binary logistic regression was used to analyze the relationship between preoperative relevant indicators and POAF occurrence. Indicators with P < 0.1 on univariate analysis were included in the multivariate binary logistic regression model to determine whether hs-CRP was an independent risk factor for POAF. Using a simple linear regression, we analyzed the relationship between relevant preoperative indicators and the duration of hospitalization and hospital cost. Indicators with P < 0.1 on univariate analysis were included in the multiple linear regression model to determine whether hs-CRP was an independent in uencing factor for the duration of hospitalization and hospital cost. Two-sided P-values < 0.05 were considered statistically signi cant.

OSA prevalence and hs-CRP
The study ow chart is shown in Fig. 1. One hundred seventy-ve patients underwent PG and OPCABG from January 2019 to December 2019. A total of 123 patients were included in the nal analysis. Based on the PG ndings, 71 patients had an AHI < 15 and were assigned to the absent-mild OSA group, while the remaining 52 had an AHI ≥ 15 and were assigned to the moderate-severe OSA group. We set the threshold hs-CRP level at 2 mg/L, which is according to the results of previous large clinical trials. Seventy-eight patients had an hs-CRP level of < 2 mg/L and were assigned to the normal hs-CRP group; 45 had an hs-CRP level of ≥ 2 mg/L and were assigned to the elevated hs-CRP group.
Preoperative clinical data of the absent-mild OSA and moderate-severe OSA groups were compared (Table 1): PCO 2 (P = 0.033), hs-CRP (P = 0.001; Fig. 2), AHI, mean apnea time, maximum apnea time, and ODI ≥ 3% (all P = 0.000) were signi cantly higher in the moderate-severe OSA group than in the absentmild OSA group. LVEF (P = 0.034) and lowest and mean SaO 2 (both P = 0.000) were signi cantly lower in the moderate-severe OSA group than in the absent-mild OSA group. No signi cant difference in other preoperative indexes was found. In addition, the number of grafting performed, duration of surgery, MACCEs, lung infection, POAF, duration of ventilator use, ICU stay, duration of hospitalization, and hospital cost were not signi cantly different between the two OSA groups (Table 2).  Values are mean (± SD), median (interquartile range), or no. (%).
A comparison of postoperative clinical data between the normal hs-CRP and elevated hs-CRP groups ( Fig. 3) showed that the proportion of patients with moderate-severe OSA and that of patients with POAF were signi cantly higher in the elevated hs-CRP group than in the normal hs-CRP group (56% vs. 35%, P = 0.024; 38% vs. 19%, P = 0.024). The duration of hospitalization (P = 0.000) and hospital cost (P = 0.000) in the elevated hs-CRP group were signi cantly higher than those in the normal hs-CRP group. Mean SaO 2 (P = 0.012) in the elevated hs-CRP group was signi cantly lower than that in the normal hs-CRP group.
Correlation of hs-CRP level with POAF, duration of hospitalization, and hospital cost.
POAF, duration of hospitalization, and hospital costs were signi cantly different between the normal hs-CRP and elevated hs-CRP groups. To determine whether hs-CRP level was an independent in uencing factor for these three indicators, logistic regression analyses (for POAF) and linear regression analysis (for duration of hospitalization and hospital cost) were performed. Table 3 shows that hs-CRP level was an independent risk factor for POAF (OR = 1.17, P = 0.006). Table 4 shows that hs-CRP level (P = 0.002) and LVEF (P = 0.002) independently correlated with the duration of hospitalization; hypertension (P = 0.014), hs-CRP level (P = 0.040), and LVEF(P = 0.000) independently correlated with hospital costs.

Discussion
OSA is closely related to CHD. A study showed that the prevalence of CHD in patients with OSA was 16.2% and that of patients without OSA was 5.4% [7]; in a population with suspected CHD, the proportion of patients with moderate-severe OSA was 24%, four times higher than the prevalence in the normal population [8]. In the present study, we found that the proportion of patients with moderate-severe OSA was 42.3%, signi cantly higher than that of the aforementioned study. This nding could be because the patients with OSA in our study needed to undergo OPCABG. Moreover, the patients had severe coronary atherosclerotic lesions, which indicates that OSA is a substantial risk factor for CHD.
In addition, we found that the preoperative LVEF of patients with moderate-severe OSA was signi cantly lower than that of patients with absent-mild OSA. Previous studies also found that LVEF was independently related to moderate-severe OSA. The main mechanisms by which OSA affects cardiac function may be as follows [9,10]: rst, each respiratory obstruction event could result in an intrathoracic negative pressure of 60-70 cm H 2 O, and hypoxia could cause pulmonary vasoconstriction, resulting in preload and afterload imbalance between the left and right ventricles; subsequently, myocardial oxygen consumption increases and myocardial ischemia occurs, which in turn alters cardiac function; second, the long-term repeated uctuation of intrathoracic pressure could increase intraglomerular pressure variability, leading to impaired cardiac function; and nally, sympathetic hyperactivity affects all-day cardiopulmonary hemodynamics. In addition, hypopharyngeal edema due to decreased cardiac function could also promote the development of OSA.
CABG is one of the primary treatments of CHD. Few studies showed that OSA may affect the prognosis of patients with CABG; Uchôa et al. found that moderate-severe OSA signi cantly increased the long-term incidence of MACCEs (follow-up time of 4.5 years), revascularization rate, the proportion of angina attacks, and AF incidence in patients with CABG; there was no signi cant effect on the 30-day prognosis after CABG [11]. Another study found that AHI was an independent risk factor for increased duration of hospitalization and postoperative circulatory uctuation in patients with CABG [12]. In contrast, we found no signi cant difference in postoperative indicators, including the duration of hospitalization, between the moderatesevere OSA and absent-mild OSA groups. The possible reason for the discrepancy is that all patients included in our study underwent OPCABG, which in turn avoided the effects of extracorporeal circulation, shortening the postoperative recovery cycle, and reducing the short-term effect of OSA.
Evidence to identify the effect of OSA on the prognosis of OPCABG only using OSA classi cation is insu cient. OSA may affect the clinical outcome of CABG by in uencing other indicators. For example, our previous study found that OSA may further affect the perioperative indicators such as the cardiac function [4]. In addition, biomarkers related to OSA and CHD may also be examined to predict more accurately the effect of OSA on CABG.
CRP is an acute phase reaction protein. Repeated hypoxia and inadequate ventilation in OSA could trigger oxidative stress and systemic in ammatory response, which could in turn enhance the synthesis and release of CRP [13]. Shamsuzzaman et al. showed that CRP has a signi cant linear correlation with AHI and is an independent in uencing factor for OSA severity [14]. Moreover, in ammatory responses play key roles in the development of atherosclerosis. CRP, which is the product and mediator of in ammatory responses in atherosclerosis, is an important marker of endothelial dysfunction. Elevated CRP levels have been shown to be an independent risk factor for diseases such as myocardial infarction, peripheral vascular disease, and stroke [15]. Han et al. found that high CRP levels in uence acute renal function injury, all-cause death, duration of hospitalization, and ICU stay after CABG [16].
Compared with the CRP, hs-CRP extends the detection linear range from 3-200 mg/L to 0.005-0.10 mg/L, thereby making the determination of low-concentration CRP more accurate. hs-CRP has a long half-life, with no diurnal difference and no sex-or age-dependence and has a higher value in predicting the prognosis of cardiovascular and cerebrovascular diseases [17]. Previous studies have shown a relationship between hs-CRP levels and OSA, nevertheless, these results are controversial because of obesity and various confounding factors in previous studies [18]. In our study, we found that the hs-CRP level was signi cantly increased in the moderate-severe OSA group, and there were signi cant correlations between hs-CRP level and several indicators of PG (AHI, ODI ≥ 3%, lowest SaO 2 , mean SaO 2 ). Multiple linear regression showed that AHI was an factor affecting hs-CRP level, independent of obesity.
Hs-CRP plays an important role in predicting the prognosis of cardiovascular disease; nevertheless, only one report on the early effect of hs-CRP on OPCABG has been conducted [19]. In the present study, we found that elevated hs-CRP levels were signi cantly associated with increased AF incidence and duration of hospitalization. Our results also demonstrated that AF incidence, duration of hospitalization, and hospital costs were signi cantly higher in the elevated hs-CRP group than in the normal group after OPCABG.
Further regression analysis showed that hs-CRP level was an independent risk factor for POAF and was independently correlated with the duration of hospitalization and hospital cost.
As previously mentioned, OSA could affect hs-CRP levels, and preoperative hs-CRP level predicts postoperative indicators of OPCABG. CPAP, which could reduce the CRP level, is an important treatment for OSA [20]. For these reasons, the results of our study may suggest the utility of CPAP in patients with OSA who undergo CABG.
Stability of respiratory regulation is an important factor in determining OSA severity [21]. Compared with the normal population, patients with OSA had signi cantly reduced respiratory center responses to low PaO 2 and high PaO 2 during sleep, and the respiratory center response of some patients was also suppressed during wakefulness. Moreover, patients with OSA have long durations of apnea at night with short intervals.
While hyperventilation occurs at the end of an apneic event, it is insu cient to clear the accumulated CO 2 , thereby resulting in hypercapnia or even type II respiratory failure. In this study, we found that the periods of apnea were signi cantly longer and the PaCO 2 level was signi cantly higher in the moderate-severe OSA group than in the absent-mild OSA group when awake. Further study involving the change in PaCO 2 and internal environment is needed.
Our study has the following limitations: rst, it is a single-center study with a limited sample size; the results need to be validated using multi-center, large-sample studies. Second, we were unable to use polysomnography (PSG), which appears to offer a more accurate evaluation of OSA than does PG. Also, PG may underestimate the OSA severity. PSG requires patients to sleep at the sleep center for more than 8 h.
The sleep center in our hospital has not been provided with urgent response equipment and medical teams for patients with heart disease. All of our patients had severe coronary artery disease and were consequently at a high risk of emergent cardiac events. Therefore, we decided to use a portable PG monitor so that we could keep our patients in the cardiac surgery department for the assessment of OSA. Finally, we only analyzed hs-CRP levels at one time point; thus, it remains unclear as to whether hs-CRP level has the same predictive value at other time points.

Conclusions
Moderate-severe OSA has a higher incidence in patients who undergo OPCABG. Compared with patients with absent-mild OSA, those with moderate-severe OSA have poorer cardiac function and higher PCO 2 and hs-CRP levels. The hs-CRP level is closely related to the severity of OSA and has an important predictive value for POAF, duration of hospitalization, and hospital cost among patients with OPCABG. The effect of OSA on the short-term prognosis of OPCABG may be predicted using preoperative hs-CRP levels. Declarations