Effect of upper airway surgery on heart rate variability in patients with obstructive sleep apnoea syndrome
Yoo-Sam Chung, MD, PhD, Department of Otolaryngology, Asan Medical Center, Ulsan University College of Medicine, Asanbyeongwon-gil 86, Songpagu, Seoul 138-736, Korea. Tel.: 82 2 3010 3710; fax: 82 2 489 2773; e-mail: firstname.lastname@example.org
To determine whether surgery influences cardiovascular autonomic modulation in obstructive sleep apnoea syndrome (OSAS), the present study was performed to evaluate the effect of upper airway (UA) surgery on heart rate variability (HRV) using frequency domain analysis for patient groups who have had either successful or unsuccessful surgery. We compared body mass index (BMI), polysomnographic [apnoea index (AI), apnoea-hypopnoea index (AHI), minimum SaO2] and HRV [very low frequency (VLF) power, low frequency (LF) power, high frequency (HF) power, HF/LF ratio, LFnu = LF/(LF + HF), HFnu = HF/(LF + HF)] parameters between the unsuccessful (n = 14) and successful (n = 22) surgical groups before and after UA surgery. Significant changes were observed for the successful patient group with respect to mean AI (from 29.1 ± 21.3 to 2.0 ± 3.2 events h−1, P < 0.001), AHI (from 38.6 ± 20.0 to 5.6 ± 5.1 events h−1, P < 0.001), minimum SaO2 (from 73.3 ± 12.7 to 86.3 ± 6.5%, P < 0.001), VLF power (from 25599 ± 12906 to 20014 ± 9839 ms2, P = 0.013), LF power (from 17293 ± 7278 to 14155 ± 4980 ms2, P = 0.016), LFnu (from 0.700 ± 0.104 to 0.646 ± 0.128, P = 0.031) and HFnu (from 0.300 ± 0.104 to 0.354 ± 0.128, P = 0.031); however, mean BMI, HF power and LF/HF ratio did not change significantly after UA surgery. No significant changes were observed in the unsuccessful surgical group. Successful UA surgery may improve cardiac sympathetic and parasympathetic modulation in patients with OSAS.
Obstructive sleep apnoea syndrome (OSAS) is a chronic multi-factorial disease characterized by recurrent interruptions of breathing during sleep such as apnoea and hypopnoea. Repeated sleep-related respiratory disturbances are associated with various harmful conditions including hypoxaemia, hypercapnia, arousals, noticeable fluctuations in intrathoracic pressure and elevated sympathetic activity (Patil et al., 2007; Strollo and Rogers, 1996). As a result of these conditions, individuals may experience numerous symptoms, including excessive daytime sleepiness, non-refreshing sleep, morning headache, reduced libido, memory loss and decreased concentration (Guilleminault, 1987; Patil et al., 2007; Strollo and Rogers, 1996). Furthermore, if left untreated OSAS can cause serious complications, especially cardiovascular morbidities such as hypertension, coronary artery disease, heart failure and stroke (Kales et al., 1984; Strollo and Rogers, 1996; Yaggi et al., 2005).
While OSAS is associated with numerous detrimental effects, the exact pathophysiological mechanisms underlying the relationship between OSAS and cardiovascular disorder are not fully understood. However, previous studies suggest that an imbalance of cardiovascular autonomic nervous system (ANS) function or increased nocturnal sympathetic activity in OSAS may play a significant role in the development of cardiovascular disease (Hedner et al., 1988, 1995).
Heart rate variability (HRV) refers to variations in the beat-to-beat time interval of consecutive sinus rhythm and reflects cardiovascular autonomic regulation (Rajendra Acharya et al., 2006). HRV analysis is a useful, sensitive and non-invasive method for investigating the balance between sympathetic and parasympathetic activity responsible for modulating cardiovascular autonomic activity (Khoo et al., 1999; Rajendra Acharya et al., 2006).
There are three primary therapies for OSAS, namely, continuous positive airway pressure (CPAP), oral appliance (OA) and upper airway (UA) surgery (Kushida et al., 2006a,b; Powell, 2009). Recent studies suggest that efficient treatment for OSAS using compliable long-term CPAP or OA may improve parameters associated with HRV or cardiac autonomic modulation in patients with OSAS (Coruzzi et al., 2006; Khoo et al., 2001; Roche et al., 1999). However, there is little research examining the relationship between successful surgical treatment and the improvement of cardiovascular autonomic nervous function in patients with OSAS (Jiang et al., 2004). Therefore, this study was designed to evaluate the effect of UA surgery on HRV using frequency domain analysis for patient groups who have had either successful or unsuccessful surgery.
Materials and Methods
This study was conducted at two centres and was approved by the Institutional Review Board (IRB) at Asan Medical Center (AMC) and Korea University Ansan Hospital (KUAsH). The IRBs of both AMC and KUAsH were exempted this study from having to provide verification for informed consent forms because the present study was based on a retrospective investigation which used previously collected data from past clinical examinations/practices. Inclusion criteria were: (1) adults (18 ≤ age < 61 years old) who were diagnosed with OSAS; (2) had refused to accept CPAP treatment for OSAS; and (3) who were treated by UA surgery and then followed-up with polysomnography at 3 months after surgery. Exclusion criteria were: (1) patients with underlying diseases [e.g. diabetes mellitus, cardiovascular (hypertension, coronary artery disease, stroke, etc.), respiratory, chronic hepatic and renal disorders, etc.]; (2) taking drugs (e.g. beta-blockers, etc.) that affect the cardiovascular ANS; (3) having other sleep disorders (e.g. periodic limb movements during sleep, etc.); and (4) having inadequate or missing HRV recordings. The body mass index (BMI) was calculated by dividing the weight (in kilograms) by the height squared (in meters). A total of 36 consecutive eligible subjects consisting of 14 surgical non-success patients (14 males) and 22 surgical success patients (21 males, one female) were included in the final statistical analysis. The mean age of the patients included in the study was 37.6 ± 9.1 years and the mean BMI (kg m−2) was 26.4 ± 2.3.
Full-night attended polysomnography was performed using a computerized system at AMC (Grass-Telefactor, Astro-Med, West Warwick, RI, USA) and KUAsH (Alice 4; Respironics, Atlanta, GA, USA). The recorded physiological signals included electroencephalogram (EEG), electro-oculogram (EOG), submental and leg electromyogram (EMG), electrocardiogram (ECG), airflow at the nose and mouth, chest and abdominal respiratory movement, arterial oxygen saturation with pulse oximetry, snoring microphone and body position sensor. All the sleep studies were interpreted manually by a sleep technician according to standard sleep stage and arousal criteria (Atlas Task Force of the American Sleep Disorders Association, 1992; Rechtschaffen and Kales, 1968). Apnoea was defined as absence of airflow for a period lasting at least 10 s and hypopnoea was defined as at least a 30% reduction in airflow associated with a 4% or greater decrease in oxygen saturation. The apnoea index (AI) was defined as the number of apnoeic episodes per hour of total sleep time (TST), and the apnoea–hypopnoea index (AHI) was defined as the number of episodes of apnoea and hypopnoea per hour of TST. The arousal index was defined as the number of arousals per hour of TST. A diagnosis of OSAS was established by an AHI ≥ 5 and possessing symptoms related to OSAS such as excessive daytime sleepiness (American Academy of Sleep Medicine, 2005).
Upper airway surgery
In this study, we defined UA surgery as surgical treatment for OSAS, such as pharyngeal surgery [uvulopalatopharyngoplasty (UPPP), uvulopalatal flap] with or without nasal surgery (endoscopic sinus surgery, septoplasty, turbinoplasty). Surgical success for OSAS was defined as an AHI < 20 with a > 50% decrease in postoperative AHI and surgical cure was defined as an AHI < 5 with a > 50% decrease in postoperative AHI, as determined by comparison with baseline data (pre-operative AHI).
Heart rate variability analysis
Full-night ECG signals obtained from the polysomnographic data were analysed according to standard guidelines (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996). Frequency domain parameters including absolute values of power (ms2), ratio and normalized units (nu) were calculated (Burr, 2007) and were as follows: (1) very low frequency (VLF, 0.003–0.04 Hz) power, which is the power in the VLF band of the HRV spectrum and is not well defined; (2) low frequency (LF, 0.04–0.15 Hz) power, which is the power in the LF band of the HRV spectrum and may reflect both sympathetic and parasympathetic tone; (3) high frequency (HF, 0.15–0.4 Hz) power, which is the power in the HF band of the HRV spectrum and may reflect parasympathetic (vagal) tone; (4) LF/HF (LF : HF ratio), which is used as an indicator of the balance between sympathetic and parasympathetic tone; (5) LFnu [LF/(LF + HF)], which is used as a marker of sympathetic modulation; and (6) HFnu [HF/(LF + HF)], which is used as a marker of parasympathetic modulation.
All the data obtained in this study are presented as the means ± standard deviation (SD) for continuous variables, and as frequencies (percentage) for categorical variables. Statistical analysis of baseline data (age, BMI, polysomnographic and HRV parameters) between the unsuccessful (uncured) and successful (cured) groups was performed using the Mann–Whitney U-test. The Wilcoxon’s signed rank test was used to compare BMI, polysomnographic and HRV parameters in the unsuccessful (uncured) and successful (cured) groups before and after UA surgery. Statistical analysis was performed using spss version 12.0 statistical software (SPSS Inc., Chicago, IL, USA), and P < 0.05 was considered statistically significant.
The baseline data, including demographic, polysomnographic and HRV parameters, are given in Table 1. There were no significant differences in mean age, BMI, AI, AHI, minimum SaO2, VLF power, LF power, HF power, LF/HF ratio, LFnu and HFnu between the unsuccessful and successful groups. Significant differences were not found in all parameters except LF power between the uncured and cured groups.
Table 1. Baseline data
|Sex (male/female)||14/0||21/1|| ||21/1||14/0|| |
|Age, years||39.3 ± 8.0||36.5 ± 9.7||0.490||37.7 ± 8.5||37.4 ± 10.3||0.962|
|BMI, kg m−2||26.0 ± 2.3||26.7 ± 2.3||0.597||26.4 ± 2.7||26.4 ± 1.7||0.962|
|AI, events h−1 of TST||29.4 ± 22.1||29.1 ± 21.3||0.911||29.2 ± 20.9||29.3 ± 22.7||0.987|
|AHI, events h−1 of TST||37.9 ± 22.6||38.6 ± 20.0||0.936||39.3 ± 20.7||36.8 ± 21.5||0.665|
|Minimum SaO2, %||76.4 ± 9.0||73.3 ± 12.7||0.597||75.1 ± 9.3||73.6 ± 14.4||0.962|
|VLF power, ms2||20 828 ± 8772||25 599 ± 12 906||0.253||22 074 ± 12 009||26 369 ± 10 767||0.203|
|LF power, ms2||14 009 ± 4712||17 293 ± 7278||0.160||14 208 ± 6360||18 858 ± 5943||0.027*|
|HF power, ms2||6446 ± 3043||6914 ± 2547||0.343||6235 ± 2867||7514 ± 2353||0.116|
|LF/HF ratio||2.523 ± 1.339||2.796 ± 1.454||0.553||2.673 ± 1.6||2.716 ± 1.057||0.432|
|LFnu||0.688 ± 0.086||0.700 ± 0.104||0.619||0.688 ± 0.100||0.708 ± 0.093||0.432|
|HFnu||0.311 ± 0.086||0.300 ± 0.104||0.642||0.312 ± 0.100||0.292 ± 0.093||0.432|
The pre-operative and postoperative data, including BMI, polysomnographic and HRV parameters for the unsuccessful and successful groups, are shown in Table 2. In the unsuccessful group, no significant changes were observed after UA surgery for BMI, polysomnographic and HRV parameters. However, in the success group there were significant changes in mean AI (P < 0.001), AHI (P < 0.001), minimum SaO2 (P < 0.001), VLF power (P = 0.013), LF power (P = 0.016), LFnu (P = 0.031) and HFnu (P = 0.031) after UA surgery. Other parameters did not change before and after UA surgery.
Table 2. Pre-operative and postoperative data for the unsuccessful and successful groups
|BMI, kg m−2||26.0 ± 2.3||26.6 ± 3.2||0.650||26.7 ± 2.3||27.0 ± 2.6||0.286|
|AI, events h−1 of TST||29.4 ± 22.1||27.5 ± 26.9||0.551||29.1 ± 21.3||2.0 ± 3.2||< 0.001*|
|AHI, events h−1 of TST||37.9 ± 22.6||44.7 ± 24.4||0.551||38.6 ± 20.0||5.6 ± 5.1||< 0.001*|
|Minimum SaO2, %||76.4 ± 9.0||77.6 ± 7.6||0.572||73.3 ± 12.7||86.3 ± 6.5||< 0.001*|
|VLF power, ms2||20 828 ± 8772||21 489 ± 15253||0.600||25 599 ± 12 906||20 014 ± 9839||0.013*|
|LF power, ms2||14 009 ± 4712||14 435 ± 6504||0.972||17 293 ± 7278||14 155 ± 4980||0.016*|
|HF power, ms2||6446 ± 3043||5788 ± 2734||0.152||6914 ± 2547||7751 ± 3729||0.485|
|LF/HF ratio||2.523 ± 1.339||2.699 ± 1.342||0.650||2.796 ± 1.454||2.256 ± 1.354||0.072|
|LFnu||0.688 ± 0.086||0.705 ± 0.077||0.507||0.700 ± 0.104||0.646 ± 0.128||0.031*|
|HFnu||0.311 ± 0.086||0.295 ± 0.077||0.650||0.300 ± 0.104||0.354 ± 0.128||0.031*|
The pre-operative and postoperative data, including BMI, polysomnographic and HRV parameters for the uncured and cured groups, are presented in Table 3. In the uncured group, there were no significant changes in all parameters except mean AI after UA surgery. However, in the cured group, significant changes were found in mean AI (P = 0.001), AHI (P = 0.001), minimum SaO2 (P = 0.005), VLF power (P = 0.003) and LF power (P = 0.006) after UA surgery. There were no significant changes in other parameters before and after UA surgery.
Table 3. Pre-operative and postoperative data for the uncured and cured groups
|BMI, kg m−2||26.4 ± 2.7||26.9 ± 3.2||0.445||26.4 ± 1.7||26.7 ± 2.0||0.433|
|AI, events h−1 of TST||29.2 ± 20.9||19.2 ± 24.1||0.042*||29.3 ± 22.7||0.5 ± 0.5||0.001*|
|AHI, events h−1 of TST||39.3 ± 20.7||32.5 ± 25.4||0.140||36.8 ± 21.5||2.4 ± 1.9||0.001*|
|Minimum SaO2, %||75.1 ± 9.3||79.5 ± 6.9||0.051||73.6 ± 14.4||88.3 ± 6.9||0.005*|
|VLF power, ms2||22 074 ± 12009||20 675 ± 13640||0.455||26 369 ± 10767||20 451 ± 9484||0.003*|
|LF power, ms2||14 208 ± 6360||13 374 ± 5533||0.614||18 858 ± 5943||15 664 ± 5434||0.006*|
|HF power, ms2||6235 ± 2867||6622 ± 3842||0.931||7514 ± 2353||7564 ± 2839||0.730|
|LF/HF ratio||2.673 ± 1.6||2.474 ± 1.412||0.590||2.716 ± 1.057||2.356 ± 1.289||0.198|
|LFnu||0.688 ± 0.100||0.671 ± 0.118||0.555||0.708 ± 0.093||0.666 ± 0.110||0.140|
|HFnu||0.312 ± 0.100||0.329 ± 0.118||0.476||0.292 ± 0.093||0.334 ± 0.110||0.140|
The results of this study, which examined the effect of UA surgery on HRV in patients with OSAS, suggest that UA surgery may improve cardiovascular autonomic regulation in patients who have undergone successful surgery relative to patients who did not experience surgical success. To the best of our knowledge, this is the first study to compare polysomnographic and HRV parameters in the unsuccessful and successful groups before and after UA surgery. Further, this study is the largest to date to document the relationship between surgical therapy for OSAS and cardiovascular autonomic nervous activity using HRV analysis.
It is well recognized that patients with OSAS have increased cardiac sympathetic tone during sleep (Somers et al., 1995). In this study, LF power and LFnu were decreased significantly after successful surgical treatment for OSAS, indicating that increased cardiac sympathetic tone and altered cardiac sympathetic modulation caused by OSAS may be normalized with the improvement of respiratory disturbances and hypoxic condition after successful UA surgery (Burr, 2007; Khoo et al., 1999). In addition, HFnu increased significantly after UA surgery in the successful patient group, indicating that successful UA surgery may improve cardiac parasympathetic modulation of the sino-atrial node in patients with OSAS (Burr, 2007; Khoo et al., 1999).
There are several treatment methods for OSAS, including CPAP, OA, UA surgery, positional therapy and weight loss (Kushida et al., 2006a; Randerath et al., 2011). Each therapy for OSAS has advantages and disadvantages, respectively. Of these therapeutic options, effective therapy using CPAP or OA may be related to the normalization of deranged cardiovascular autonomic variability in patients with OSAS (Coruzzi et al., 2006; Khoo et al., 2001; Roche et al., 1999). Conversely, the connection between other treatments and cardiovascular autonomic modulation has not yet been demonstrated in patients with OSAS.
Continuous positive airway pressure works as a pneumatic splint of the UA and is a very efficient treatment in compliant patients. CPAP should be considered in the treatment of patients with moderate to severe OSAS and is optionally indicated for the management of patients with mild OSAS (Epstein et al., 2009). However, because there are some disadvantages, such as discomfort and side effects, CPAP adherence rates have been reported to range from 40 to 80% (Kribbs et al., 1993; Pepin et al., 1999). Several studies have established that long-term CPAP therapy leads to a substantial improvement of altered cardiac autonomic function in patients with OSAS (Khoo et al., 2001; Roche et al., 1999). Roche et al. (1999) assessed the effect of long-term CPAP treatment on the parameters of cardiac autonomic activity using HRV analysis in patients with OSAS and found a significant decrease in VLF power, LF power, LF/HF ratio and LFnu, as well as a significant increase in HFnu during the night after CPAP therapy for 3 months. These results are in close agreement with the results of the present study, which together suggest that successful surgical treatment for OSAS significantly improved the indicators of altered cardiac autonomic modulation, especially deranged cardiac sympathetic activation in patients with OSAS.
Oral appliance acts by enlarging the UA and/or by increasing UA muscle tone. OA is usually indicated for the treatment of patients with mild to moderate OSAS; however, it does have some disadvantages, such as inconvenience and complications (Epstein et al., 2009; Randerath et al., 2011). Itzhaki et al. (2007) reported that long-term OA treatment may improve OSAS with cardiovascular risk factors such as endothelial function and oxidative stress markers. In a recent study evaluating the effect of OA treatment for OSAS on cardiovascular autonomic variability, Coruzzi et al. (2006) concluded that 3 months of therapy using OA may improve indices related to cardiac autonomic modulation in patients with OSAS of mild degree. In their study, the exact mechanisms between the improvement of cardiac autonomic dysregulation and OA treatment for OSAS were not found; however, they proposed that reduced chemoreflex sensitivity, especially a decline in sympathetic tone with the alleviation of respiratory disturbances, may play an important role.
Surgical treatment is an efficacious method in a minority of patients with OSAS who have a surgically correctable obstruction related to OSAS (Epstein et al., 2009; Randerath et al., 2011). However, it is difficult to predict the success of surgical results in the majority of patients who do not have a severe anatomical abnormality (Sher et al., 1996). Therefore, this study was designed to investigate what effect the result of surgery (successful or unsuccessful) has on HRV. Through this study, we confirm that only successful surgery may influence improvement of cardiac autonomic parameters. Many studies have reported an association between UA surgery for OSAS and improvement of clinical outcomes, including symptoms, quality of life, automobile accidents, cardiovascular risk and mortality (Haraldsson et al., 1995; Peker et al., 2002; Riley et al., 2000; Weaver et al., 2004). However, to determine the role of HRV underlying the link between the positive effect of surgery for OSAS and the alleviation of clinical outcomes, especially a reduction of cardiovascular risk or mortality rate, there have been no previous studies that compared polysomnographic and HRV parameters systemically in unsuccessful and successful surgical groups before and after UA surgery. The results of this study indicate that improved sympathetic and parasympathetic modulation after successful surgery may have a marked effect on decreasing incidence of cardiovascular disease in patients with OSAS. Jiang et al. (2004) found that LF and HF components of HRV and LF/HF ratio were reduced significantly in 16 patients with OSAS 3 months after UPPP. These results are partially consistent with our findings that successful UA surgery may improve cardiac sympathetic modulation in patients with OSAS. However, one of the limitations of their study is that patients who did not have successful surgery were included in the study analysis. Thus, it may be difficult to evaluate the pure effect of successful surgical treatment on the improvement of altered cardiac autonomic variability in patients with OSAS. Peker et al. (2002) investigated the causal relationship between OSAS and cardiovascular disease by performing a 7-year follow-up study in a consecutive sleep clinic cohort of 182 middle-aged men with or without OSAS. They showed a significantly increased incidence of a cardiovascular disease in patients with OSAS during a follow-up period and that effective therapy for OSAS is associated with a significant decrease of cardiovascular risk. In their study, UPPP had a critical role in diminishing the incidence of cardiovascular disease due to relatively low compliance of CPAP. Weaver et al. (2004) compared the survival rate between CPAP-treated and UPPP-treated patients with OSAS, and found that UA surgical treatment provides greater long-term survival than CPAP therapy in patients with OSAS. Indeed, it is thought that these results may be related to a large proportion of CPAP-intolerant patients.
One of the strengths of the present study was that the surgical non-success group acted as a control group, which we believe may be helpful to understand the results of the current study. This study had several limitations. First, the study is not a randomized controlled trial (RCT). We expect that this study might trigger the performance of a larger RCT to compare the surgical therapy with CPAP or OA. Secondly, we analysed the data according to cure criteria (50% reduction in AHI and AHI < 5) and LFnu (from 0.708 ± 0.093 to 0.666 ± 0.110) and HFnu (from 0.292 ± 0.092 to 0.334 ± 0.110) improved after UA surgery. However, there were no statistical significances in these parameters. It is thought that these outcomes may be caused primarily by a small sample size (n = 14). Further studies using a large sample size will be required. Thirdly, in this study, HRV parameters did not change significantly in all patients, including successful and unsuccessful groups, after surgery. CPAP or OA studies related with HRV showing benefits did not focus only on responders or adherent patients, but demonstrated improvements in general. This study does not allow for any conclusion on the efficacy of surgical approaches and improvement of cardiovascular parameters, as we analysed selectively those who responded to treatment. This study suggested that only successful surgery may improve cardiovascular autonomic variability in OSAS patients. Fourthly, demographic characteristics of our patients were generally different from those of standard OSAS population (e.g. age, sex, BMI, etc.). Therefore, the outcomes might not be representative of the entire OSAS population.
In conclusion, the results of the present study provide evidence that increased cardiac sympathetic tone and altered cardiac sympathetic and parasympathetic modulation caused by OSAS might be improved with the alleviation of respiratory disturbances and hypoxic condition after successful surgical treatment. Successful UA surgery may improve cardiac sympathetic and parasympathetic modulation in patients with OSAS.
Declarations of Interest
This study was supported by a grant of the Korea Healthcare technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A090084). This study was supported by the KIST Institutional Program.