Upsala J Med Sci 97: 141-148 Airway Obstruction, Obesity and C 0 2 Ventilatory Responsiveness in the Sleep Apnea Syndrome Thorarinn Gislason' and Ritva Tammivaara* From the Departments of 'Pulmonary Medicine and 2Clinical Physiology, Akademiska Sjukhuset, University of Uppsala, Uppsala, Sweden ABSTRAn In 32 patients with sleep apnea syndrome (SAS), pulmonary function, blood gases and the ventilatory response to CO, (CO, VR) were studied before and 6 months after uvulopalato- pharyngoplasty. Nine of the SAS patients had airway obstruction (AO-SAS), defined as FEV1.o I 72 % of the predicted value. They had a significantly higher PaCO,, lower PaO, and a lower CO, VR than the remaining SAS patients. Preoperatively 4 SAS patients were hypercapnic (PaCO, >5.8 Wa) and compared with the normocapnic ones they were more obese; in 3 of them FEV,.o was I72%. The hypercapnic SAS patients had a significantly lower CO, VR. The CO, VR was significantly correlated to A 0 and the degree of oxygen desaturation during sleep, but not to the number of episodes of apnea and hypopnea nor their length. The VR to CO, did not predict the postoperative outcome. Postoperatively 2 hypercapnic obese AO-SAS patients showed a large decrease in episodes of apnea and hypopnea and an increase in CO, VR, and became normocapnic. Other patients showed no consistent changes in CO, VR postoperatively. INTRODUCTION The sleep apnea syndrome (SAS) is a frequently diagnosed disorder which is characterized by loud snoring and repeated upper airway obstructions during sleep (1,2). Many aspects of the inter- relationship between SAS, pulmonary function, obesity, hypercapnea and the ventilatory response to CO, (CO, VR) are still unclear. Aubert-Tulken and coworkers found a shift to the left in the curve of the CO, VR after tracheostomy in a 41-year-old male patient who was hypercapnic and had pulmonary airway obstruction (3). Guilleminault and Cumminskey reported that among 5 eucapnic, nonobese men with SAS the CO, VR was doubled after tracheostomy (4). Among 8 hypercapnic SAS patients 4 became eucapnic after treatment (5). In 19 closely followed SAS patients there was no change in the slope of the CO, VR during successful treatment with continuous positive airway pressure (CPAP) (6). However, in 10 of the patients, who were initially hypercapnic, there was an increase in the CO, VR measured as the position on the CO, VR 141 curve (6). We have so far found no reports on the possible results of uvulopalatopharyngoplasty (UPPP) on hypercapnea and the CO, VR. The aim of this prospective study was to evaluate the CO, VR among SAS patients in relation to airway obstruction, obesity, the severity of the SAS, and arterial blood gases, before and after UPPP. METHODS Patients; This study comprised all SAS patients who underwent UPPP for SAS between September 1984 and April 1986 at Akademiska Sjukhuset, Uppsala, Sweden. In all there were 31 men and 3 women, with a mean age of 49 years (range 30-68). Fourteen were current smokers and one had stopped smoking less than one year previously. The study protocol had been approved by the Ethics Committee of the Medical Faculty of Uppsala University. The clinical characteristics of the SAS patients and the operative procedure have been described in detail elsewhere (7). Five had a history of chronic obstructive pulmonary disease and 1 had bronchial asthma. Two were treated with beta-2 agonists, and 2 with theophylline preparations. All but 1 were habitual snorers, and all-night polysomnographic studies had confirmed the SAS diagnosis. An apnea was defined as a 10 second or longer complete cessation of both nasal and oral air flow and a hypopnea was defined as a marked decrease in oro-nasal air flow for at least 10 seconds, followed by a fall in baseline oxygen saturation by at least 4% or arousal. An index of apneas (A) and hypopneas (H), was calculated as the total number of such events per hour of sleep (1). Preoperatively these 34 SAS patients had a mean (A+H) index of 44 (median 30, range 9- 101). Six months postoperatively all 34 patients came for follow-up studies and 22 patients (65%) showed a decrease in the (A+H) index by more than 50% and were therefore classed as responders to UPPP (7). Sixteen patients had an (A+H) index below 10 postoperatively. All but 4 SAS patients were overweight, with a body mass index (BMI) of 2 29 kg/m2 and the mean BMI for all patients was 32.7 (SD 5.7) kg/m2 (Table 1). Lunp function tests and blood gases: Lung volumes and maximal flow volume curves were determined by standard spirometry and body plethysmography. Arterial pH, PO,, and PCO, were measured with a Coming 168 pH, PO, and PCO, blood gas analyzer (Coming Medical, Halstead, England). Oxygen saturation was measured continuously during sleep with a BIOX I11 Pulse oximeter. The number of apneas and hypopneas per hour of sleep that caused a fall in SaO, below 85% were especially recorded. Testing of ventilatory regulation: All tests of the VR to hyperoxic hypercapnea were performed at the same time of day by a modification of the rebreathing method described by Read (8), while the patients were awake and at rest. We have previously described the CO, rebreathing test in detail (9). The VR slope was calculated as the change in ventilation (AV l/min) caused by a rise in end- tidal PCO, by 1 kPa (=7.5 mmHg). The position of the VR line (l/min) was expressed as the 142 minute ventilation at a PCO, of 8 kPa (6). Table 1. SAS patients with airway obstruction (AO-SAS) compared with the other SAS patients. BMI: body mass index; FEVl.o: forced expiratory volume in one second; V C vital capacity; RV: residual volume; TLC: total lung capacity; FRC functional residual capacity. A: apneas; H: hypopneas; E V F erythrocyte volume fraction. AO-SAS Other SAS patients (n=9) (n=23) Mean (S.D.) Mean (S.D.) p-value Age 54 BMI (kg/m2) 37.0 EVF (%) 46 Pulmonary function tests FEV1,o (% pred.) 65 RV (% pred) 129 TLC (% pred) 84 FEv,.pc 71 Blood gases PaO, ( H a ) 9.1 PaCO, (kPa) 5.8 AV/APC02 (I/min/kPa) 13.1 Sleep data (A+H) index (No. per hour) Mean length of A+H ( s e c ) A+H with SaO, < 85% Postoperative reduction 58 21 (No. per hour) 32 in (A+H) index (%) 58 47 (10) 31.4 (4.9) 44 (3) 97 (12) 102 (30) 100 (10) 77 (6) 10.7 (1.2) 5.0 (0.5) 24.8 (10.1) 39 (26) 24 (7) 9 (15) 58 (38) =o. 1 <0.01 =0.2 <0.05 <0.001 <0.05 <0.002 <0.01 <0.01 =0.09 =0.4 <0.005 =0.97 Statistics: Statistical probability was tested by a Student’s t test on unpaired values, except in comparisons of changes in VR, when paired values were used. The correlations between different parameters were assessed by least square linear regression. RESULTS Among the 34 patients treated by UPPP, all but 2 underwent representative lung function tests. Flow parameters revealed airway obstruction (AO), defined as an FEV,, of 272% of the 143 predicted value, among 9 SAS. These patients will be referred to as AO-SAS. Compared with the remainder they were more overweight ( ~ ~ 0 . 0 1 ) and had lower PaO, and higher PaCO, at rest (p5.8 kPa (Cases No 1,2,17 and 31 in ref. 7). Compared with the normocapnic patients they were more obese and showed a tendency to a higher (A+H) index, but the length of A+H was the same in these 2 groups (Table 2). Three of the hypercapnic patients also had an FEV,.o value I 7 2 % , but the fourth had a high FEV, o, so the difference was not significant for this small sample (p=O. 1). Table 2. Hypercapnic (n=4) and normocapnic (n=28) SAS patients. (Abbreviations see Table 1). Hypercapnic Normocapnic p-value BMI (kg/mz) EVF (%) PaO, &Pa) FEVl.0 (% P W FEV,.flC(%) RV (% pred) AV/AE'CO, (l/min/kF'a) (A+H) index (No. per hour) A+H with SaO, < 85% (No. per hour) 41.2 (7.6) 48 (4) 9.1 (1.6) 76 (13) 77(11) 156 (43) 11 (7) 65 (28) 43 (15) 31.9 (4.8) 44 (3) 10.4 (1.2) 91 (21) 7x71 100 (27) 23 (10) 40 (26) <0.003 =0.06 =0.06 =0.2 =0.6 <0.001 <0.02 =0.09 <0.002 Ventilatory response to CO2; Altogether 32 SAS patients underwent a test of the CO, VR before UPPP and all but two were tested again 6 months postoperatively (Table 3). The mean slope of the CO, VR curve was 21 (10) l/min/kF'a and the mean ventilation at 8 kPa was 47 (23) - 144 l/min (Table 3). There was a significant negative correlation between the number of A+H causing an Sa02 desaturation < 85% and the CO, VR (VR = 35.3 +(-0.95 x No. of desaturations), p<0.02). The CO, VR was also significantly correlated to obesity (VR = 37.6 +(0.21 x BMI) p<0.05) and also to the degree of A 0 (VR = 71.6 + 0.81 x FEV1, p<0.05). The 9 AO-SAS patients displayed a significantly lower CO, VR than the others (p<0.005), estimated both as the slope and as minute ventilation at 8 kPa (Table 3). The 4 hypercapnic SAS patients also exhibited a very low CO, VR (1.8, 18.0, 13.0 and 9.6 V m i m a ) , and this was significantly lower than in the rest of the patients with SAS (Table 2). The CO, VR was not significantly correlated to (A+H) index or the mean length of apneas and hypopneas. Table 3. The ventilatory response to CO, in Vmin, pre- and postoperatively. No statistical difference was observed. Preoperat ivel y Postoperatively Mean (SD) Mean(SD) Mean (SD) Mean (SD) Slope (AV/kPa) Vent. at 8 kPa Slope (AVikPa) Vent. at 8 kPa All patients (n=30) 21 (10) 47 (23) 24 (14) 45 (22) AO-SAS patients (n=8) 13 ( 6 ) 31 (18) 13 (7) 33 (20) Other SAS patients (n=22) 24 (10) 53 (22) 28 (15) 49 (22) Responders to UPPP (n=20) 23 (1 1) 53 (22) 26 (16) 49 (23) Nonresponders to UPPP (n=10)18 (8) 37 (22) 19 (8) 36 (16) PostoDerative findings: There was no consistent change postoperatively in the CO, VR, either among responders or nonresponders to UPPP, or among AO-SAS patients (Table 3). Postoperatively one hypercapnic patient had no apneas or hypopneas and the slope of his CO, VR increased from 1.8 to 7.3 VminlkPa, and PaCO, decreased from 7.5 to 5.8 kPa. Another patient showed a 35% reduction in (A+H) index postoperatively and his PaCO, fell from 5.9 to 5.4 kPa; the slope of the CO, VR curve increased from 18 to 25 VminMa and ventilation at 8 kPa from 36 to 45 Vmin. The remaining 2 patients showed only a minor decrease in the (A+H) index postoperatively and were still hypercapnic. Smokers vs. nonsmokers: Among the SAS patients the smokers were more hypoxic (PaO, = 9.7 (1.4) kPa vs.10.6 (1.1) kPa, p< 0.05) than the nonsmokers and they also had higher PaCO, values (5.5 (0.9) kPa vs. 5.0 (0.4 ) kPa, p< 0.05). Their FEV,.o was also somewhat lower (FEVl.o = 83 (18) vs. 95 (19) % pred, p=0.08). There was no difference between the smokers and nonsmokers as to sleep data, the CO, VR or the erythrocyte volume fraction. 145 DISCUSSION The findings in our study group of SAS patients clearly illustrate the great subject variability in the severity of this disorder and their heterogeneous ventilatory and pulmonary status. The C 0 2 VR varied considerably among SAS patients, but was significantly related to A 0 and obesity. We used ventilation as a parameter of chemosensitivity, but in patients with obstructive lung disease ventil-ation may be limited by expiratory airflow and not actual drive to breathe (10). However, as the airway obstruction in our patients was not so great (7), we consider that the low CO, VR among the AO-SAS patients, can only partly be explained by mechanical limitation imposed by lower airway dysfunction. This connection between airway obstruction, a low CO, VR and hypercapnea has been pointed out in a previous case report (4) and also by Bradley and coworkers (1 1). They found that among 50 consecutive SAS patients the presence of diffuse airway obstruction was an important predisposing factor to the development of CO, retention and all of their seven hypercapnic patients had A 0 (I 1). One of our hypercapnic patients had no signs of A 0 (FEVl.o=95 % (of pred), but he suffered from extreme obesity (BMI=40.9 kg/m2> (7). In a group of 28 SAS patients with airway obstruction hypercapnea was much more common among SAS patients with high lifetime alcohol comsumption (12). Our data on alcohol consumption are only based on information from our standard interview and we know that among our 4 hypercapnic SAS patients, 2 reported heavy alcohol intake, 1 never drinker, and no information was available from the fourth. Patients with low CO, VR more frequently desaturated below 85%, but other data from sleep studies were not related to the CO, VR. These results are similar to the results from Kunitomo et al, but they reported that the awake hypercapnic ventilatory drive was significantly inversely correlated with maximal desaturation during REM sleep but showed no correlation with any other sleep desaturation parameters (13). In our study this is at least partly because these patients had greater pulmonary obstruction and their baseline awake SaO, position is closer to the steep portion of the oxyhemoglobin dissociation curve. There was no consistent change in the CO, VR after UPPP, either in the group of SAS patients as a whole or in the different subgroups (Table 3). This is contrary to a previous finding of a shift in the VR curve to the left and an increase in the slope of the VR curve after tracheostomy (3) and also to a report on 5 initially normocapnic men, for whom an increase in the slope was observed 3 months after tracheostomy (4). In the largest study in this field hitterto it was found that among 19 SAS patients treated with CPAP, only those 10 patients with elevated daytime C 0 2 showed a progressive increase in ventilation at 8 kPa, but there was no change in the slope of the CO, VR (6). One possible explanation for the lack of increase in the mean CO, VR in our study might be that so few of our patients were hypercapnic. Another reason may be that compared to CPAP and tracheostomy, UPPP does not eliminate all apneas and hypopneas as effectively. The patients in references 3 and 4 were all tracheostomized and this might possibly have influenced the VR 146 independently of the relief of apneas and hypopneas. After tracheostomy there is a decrease in dead space ventilation, which increases the alveolar ventilation and might lead to lower PaCO,, at least during sleep. Although 16 of our SAS patients had an (A+H) index below 10 postoperatively (7) they might still be suffering from partial upper airway obstruction (14). The mechanisms that cause a decrease in the CO, VR might therefore still be operative. A small subgroup among our SAS patients with airways obstruction and extreme obesity also had hypercapnea and a low CO, VR. Two of them who showed postoperatively the greatest improvement in the (A+H) index, became normocapnic and in their cases the low CO, VR was partly reversible. A possible explanation for this is that increased endogenous opioid activity might occur as an adaptive reaction (15). ACKNOWLEDGEMENTS Supported by grants from the Swedish National Association against Heart and Chest Diseases, Stockholm, The King Oscar I1 Jubilee Foundation, Stockholm and the Thuring Foundation, Stockholm. 1. 2. 3. 4 . 5 . 6. 7 . 8 . 9. 10. REFERENCES Guilleminault, C., Cumminsky, J., & Dement, W.C.: Sleep Apnea Syndromes: Recent Advances. Adv Intern Med 26:347-74, 1980. Krieger, J.: Sleep apnea syndromes in adults. Bull Eur Physiopath Respir 22:147-89, 1986. Aubert-Tulkens, G., Willems, B., Veriter, Cl., Coche, E. & Stanescu, D.C.: Increase in ventilatory response to C 0 2 following tracheostomy in obstructive sleep apnea. Bull Eur Physiopath Respir 16587-93, 1980. Guilleminault, C., & Cumminsky, J.: Progressive improvement of apnea index and ventilatory response to C 0 2 after tracheostomy in obstructive sleep apnea syndrome. Am Rev Respir Dis 126:14-20, 1982. Rapoport, D.M., Stuart, M.G., Epstein, H., & Goldring, R.M.: Hypercapnea in obstructive sleep apnea syndrome. Chest 89:627-35, 1986. Berthon-Jones, M. & Sullivan, C.E.: Time course of change in the ventilatory response to C 0 2 with long-term CPAP therapy for obstructive sleep apnea. Am Rev Respir Dis 135:144-7, 1987. Gislason, T., Lindholm, C-E., Almqvist, M., Bimng, E., Boman, G., Eriksson, G . , Larsson, S.G., Lidell, C., & Svanholm, H.: Uvulopalatopharyngoplasty in the sleep apnea syndrome - predictors of results. Arch Otolaryngol 144:45-51, 1988. Read, D.J.C.: A clinical method for assessing the ventilatory response to carbon dioxide. Aust Ann Med 16:20-32, 1967. Gislason, T., Sandhagen, B., & Boman, G . : Transcutaneous C 0 2 monitoring in adults with sleep-related breathing disorders. Upsala J Med Sci 94:171-81, 1989. Plotkowski, L.M., Hannhart, B, Elfassi, R., Sautegeau, A., Peslin, R., & Sadoul, P.: Role 147 of the mechanical impairment on the ventilatory response to C02 in chronic airway obstruction. Clin Respir Physiol 2351-6, 1987. 1 1 . Bradley, T.D., Rutherford, R., Lue, F., Moldofsky, H., Grossman, R.F., Zamel, N., & Phillipson, E.A.: Role of diffuse airway obstruction in the hypercapnia of obstructive sleep apnea. Am Rev Respir Dis 134:920-4, 1986. 12. Chan, S.C., Gurnstein, R.R., Bye, P.T.P., Woolcock, A.J., & Sullivan, C.E.: Obstructive sleep apnea with severe chronic airflow limitation. Am Rev Respir Dis 140:1274-8, 1989. 13. Kuniotomo, F., Kimura, H., Tatsumi, K., Kuriyama, T., Watanabe, S . , & Honda, Y.: Abnormal breathing during sleep and chemical control of breathing during wakefulness in patients with sleep apnea syndrome. Am Rev Respir Dis 139:164-9, 1989. 14. Polo, O., Brissaud, L., Fraga, J., Dejeaa, Y., & Billiard, M.: Partial upper airway obstruction in sleep after uvulopalatopharyngoplasty. Arch Otolaryngol Head Neck Surg 115~1350-54, 1989. 15. Gislason, T., Almqvist, M., Boman, G., Lindholm, C-E., & Terenius, L.: Increased CSF odioid activity in sleep apnea syndrome-regress after successful treatment. Chest 96:250-54, 1989. Correspondence and request for reprints: Thorarinn Gislason, M.D Dept. of Pulmonary Medicine Vifilsstadir, 210 Gardabaer, Iceland 148