Long-term Acceptance of Continuous Positive Airway Pressure in Obstructive Sleep Apnea 1- 3
ISOBEl ROLFE, L. G. OLSON, and N. A. SAUNDERS Introduction he use of nasal continuous positive airway pressure (CPAP) has been a major advance in the treatment of obstructive sleep apnea (OSA) (1). Many uncontrolled studies have shown benefit from long-term use of this device, particularly in patients with severe disease (2-10), and there is general agreement that it is the treatment of choice for obstructive sleep apnea. The main limitation of CPAP has been said to be initial intolerance and poor long-term compliance (5, 11). This may be of particular relevance when the treatment is used for patients with only mild daytime impairment or patients with asymptomatic hypoxemia or arrhythmia. There is, however, little information available on the long-term acceptability of nasal CPAP, especially in patients with mild disease. We reviewed 168 consecutive patients with obstructive sleep apnea of widely varying severity and report here their tolerance of nasal CPAP and the reasons for discontinuing treatment in patients who were intolerant.
T
Methods The reported data were obtained from the records of 168 consecutive patients commenced on CPAP from January 1983to June 1989at the Sleep Disorders Unit of the University of Newcastle and Royal Newcastle Hospital. The clinical features of the patients are shown in table 1. The steps that had been taken in initiating nasal CPAP treatment were as follows: 1. All patients had an initial history and examination by one of four respiratory physicians, of whom one (L.G.O.)saw about twothirds of the patients. Excessivedaytime somnolence was graded mild, unequivocal, or severe using the criteria suggested by Sullivan and Issa (12). It was the policy of the unit that all patients referred for assessment of snoring had a sleep study regardless of the presence or absence of symptoms suggesting sleep apnea. 2. Patients had an all-night diagnostic sleep study, using standard methods. Oxygenation was assessed by pulse oximeter (Biox" 3700 or 3740; Ohmeda, Louisville, CO); nasal airflow was assessed by thermistors and motion 1130
SUMMARY We studied the long-term acceptability of nasal continuous positive airway pressure (CPAP)treatment In 168 consecutive patients, 147 with obstructive sleep apnea (OSA) and 21 with snoring. Follow-up was between 1.5 and 78 months. At latest follow-up 107 of 168 (64%) were stili using CPAP. Acceptance of CPAP was least for patients with snoring alone (6 of 21 persisted) and best for patients with both excessive daytime somnolence and severe hypoxemia (minimum Sa0 2 < 75%), of whom 40 of 45 (89%) persisted with treatment. Patients with excessive daytime somnolence but without severe hypoxemia were less tolerant of CPAP (39 of 71,55%, persisted) than patients with no symptoms of excessive somnolence but with severe hypoxemia (21 of 30, 70%, persisted). The most common reasons for discontinuing CPAP were Intolerance of the mask (26 of 61), the Inconvenience of treatment (16 of 61), and the lack of symptomatic benefit from treatment (10 of 61). We concluded that long-term acceptance of CPAP was difficult to predict in advance but that it was most likely In patients with the most severe sleep apnea. Because Intolerance of the mask and inconvenience were the most common reasons for ceasing treatment, Improvements In the design of CPAP systems and careful patient training may Improve the acceptability of CPAP substantially. AM REV RESPIR DIS 1991; 144:1130-1133
of the rib cage and abdomen by inductance plethysmography (Respitrace"; Ambulatory Monitoring Inc, Ardsley, NY) Sleep was staged using one channel of EEG, two channels of EOG, and submental EMG. Signals were recorded on an ink pen recorder. All studies considered in this report were analyzed by a single technician. Blind repeat scoring of studies selected from the whole period relevant to this report did not suggest variation of scoring criteria. We recorded the apnea plus hypopnea index (AHI) and the minimum arterial oxygen saturation (Sao.}, OSA was diagnosed when the AHI wasgreater than 15. Apnea was defined as cessation of airflow for more than 10 sand hypopnea as reduction for at least 10 s of the amplitude of both oronasal airflow and ribcage and abdomen motion signals to less than half the levels prevailing before and after the event. Severe arterial desaturation was defined as Saol at any time less than 750/0. 3. Nasal CPAP was initiated during a second all-night study, at which time the airway pressure was adjusted to abolish all apneas and hypopneas associated with arterial oxygen desaturation in all sleep stages, including rapid eye movement sleep. At this time the patient's physician gave a careful explanation of the nature and purpose of CPAP treatment. The patients were told that there was no alternative therapy for OSA, but that in the absence of severe desaturation the only purpose of treatment was to relievedaytime symptoms. Patients with severe desaturation were told that they might be at risk of pulmonary hypertension, but the degree of emphasis placed
on this varied among physicians. When CPAP was initiated the fitting of the mask and the use of the machine were explained by a technician. Patients returned after a 6-wk trial of CPAP for a further sleep study and pressure check. At this first follow-up appointment problems with CPAP were discussed and the purpose of therapy was reviewed before the patient was committed to the purchase of a CPAP machine. In some cases the trial period was extended to 8 or 10wk. No home visits were made, but patients were encouraged to telephone the laboratory for help with any problem, and technical problems were reviewed without appointments. Those discontinuing CPAP werein some cases not followed thereafter and, in other cases, restudied annually. Follow-up of patients persisting with CPAP was at variable intervals, ranging from 6 to 14 months. At each follow-up visit patient compliance, changes in symptoms, and problems with treatment were noted by the attending physician. The laboratory routinely offered to purchase the CPAP machine of (Received in original form February 8, 1990 and in revised form May 14, 1991) 1 From the Faculty of Medicine, University of Newcastle, New South Wales, Australia. 2 Supported by the National Health and Medical Research Council (Australia). 3 Correspondence and requests for reprints should be addressed to Dr. L. G. Olson, Department of Medicine, University of Newcastle, David Maddison Clinical Sciences Building, Royal Newcastle Hospital, Newcastle, NSW 2300, Australia.
ACCEPTANCE OF CPAP
1131 TABLE 1
TABLE 2
CLINICAL CHARACTERISTICS OF PATIENTS (n = 168)
TOLERANCE OF CPAP IN PATIENTS WITH EDS ALONE*
BMI, * mean ± 95% CI Obeset Hypertensive:!: Observed apneic§ Complaining of frequent choking arousals Complaining of restless sleep Complaining of frequent morning fatigue Consuming ethanol> 30 g/day
% % % % % % %
31.86 ± 0.84 66 41 79 60 54 40 19.5
• BMI = weight kg/height m2 • t Obesity = 8MI > 30. t Hypertension = systolic BP > 150 mm Hg or diastolic> 95 mm Hg. § As reported by bed partner.
a patient who had bought one but later ceased to use it. Clinicians gave four major reasons for commencing patients on CPAP: 1. Snoring in the absence of GSA in patients who wished to avoid other therapy (Le., surgery) 2. Excessive daytime sleepiness (EDS) in patients who did not have severe arterial oxygen desaturation 3. Severe arterial oxygen desaturation (Sa02 < 75070) despite mild or no somnolence 4. EDS and severe arterial desaturation, both being of sufficient severity to justify therapy.
those who did and those who did not persist with treatment, although AHI tended to be higher in those who continued. All the patients who continued treatment had improvement or resolution of EDS. Of those who stopped, 9 did so because there was no improvement in their symptoms, 11 because they could not tolerate the mask, 8 because they believed that the improvement in their symptoms did not justify the inconvenience of treatment, and 3 because they could not afford to buy a CPAP device.In the remaining case the reason for stopping CPAP was uncertain. Arterial hypoxemia. There were 30 paResults tients in whom the primary indication A total of 168patients were followed for for treatment was severe arterial desatubetween 6 wk and 78 months. There were 145 (86070) males. Overall, 107(64070) pa- ration in the absence of daytime sleepitients continued to use nasal CPAP at ness thought at the initial assessment to be significant (table 3). Mean follow-up latest follow-up. was 15.5months (range 1.5to 48 months). The 21 patients (70%) who continued Outcome According to Indication CPAP had more severe arterial desatuSnoring. A group of 21patients werefol- ration at diagnosis and a tendency to a lowed for a mean of 10.3months (range higher AHI, but the difference between 2 to 52 months). Of 21patients, 6 (29%) the groups was not significant. All patients who had symptoms recontinued CPAP during follow-up. Of the patients who discontinued CPAP, 4 ported that they improved or resolved. later had uvulopalatopharyngoplasties The 9 patients who ceased therapy found with good results. The 6 patients who the device either intolerably uncomfortcontinued use of CPAP had resolution able (5 of 9) or intolerably inconvenient. of symptoms. They have continued to use EDS and Hypoxemia. A total of 45 paCPAP for 24 to 52 months. Of the 15 tients were followed for a mean of 25 patients who stopped CPAP, 8 did so be- months (range 2 to 70 months) (table 4). cause of discomfort or claustrophobia Of these, most (89%) continued therapy caused by the mask. and all reported improvement in sympExcessive daytime sleepiness. There toms. Those who continued with treatwere 71 patients treated for this indica- ment appeared to have more severe distion, followed for a mean of 22 months ease, but the differences between the (range 1.5 to 79 months) (table 2). A to- groups was not statistically significant. tal of 39 patients (55070) continued treat- Of the patients who discontinued CPAP, ment. Arterial oxygen desaturation at di- 2 did so because they could not tolerate agnosis was greater in the 39 who con- the mask, 1 because the improvement tinued with CPAP than in the 32 who resulting from treatment did not justify stopped (p < 0.025 by Wilcoxon's test), the inconvenience oftreatment, and 1 bealthough the difference was clinically in- cause of the cost of a machine. In one significant. There were no other statis- case, the reason for discontinuing treattically significant differences between ment was unclear.
Continued n Males Age, yr AHI, events/h Minimum Sa0 2, % EDS Mild Unequivocal Severe
Ceased
39 (55%)
32
35/39 58.2 ± 2.8 30.8 ± 8.1 82.1 ± 1.4
28/32 58.0 ± 4.0 22.8 ± 7.0 84.3 ± 1.5
23 12 4
20 9 3
• Means are given ± 95% CI.
TABLE 3 TOLERANCE OF CPAP IN PATIENTS WITH HYPOXEMIA ALONE* Continued n Males Age, yr AHI, events/h Minimum Sa02. %
Ceased
21 (70%)
9
19/21
8/9
57.2 ± 7.1 43.0 ± 12.3 61.5 ± 4.0
57.2 ± 8.4 35.7 ± 15.0 68.3 ± 5.9
• Means are given ± 95% CI.
TABLE 4 TOLERANCE OF CPAP IN PATIENTS WITH BOTH EDS AND HYPOXEMIA* Continued n Males Age, yr AHI, events/h Minimum Sa02, % EDS Mild Unequivocal Severe
40 (89%) 37/40 54.9 ± 3.0 51.5 ± 9.0 62.7 ± 3.4
Ceased
5 4/5
57.8 ± 5.8 39.4 ± 43.8 63.8 ± 11.9
6
23 11
4 1
• Means are given ± 95% CI.
Outcome According to Severity ofEDS When patients were grouped according to severity of EDS, there was a significant positive relationship between the severity of EDS and the probability of accepting long-term CPAP. Of those with mild EDS, 29 of 49 (59%) continued; of those with unequivocal EDS, 35 of 48 (74%) continued; and of those with severeEDS, 15of 19(79%) continued (p < 0.02 by chi-square test). This relationship was not, however, apparent in patients with milder OSA (table 2). Duration of Treatment in Patients Ceasing CPAP Of the 61 patients who discontinued CPAP, 45 (78%) did so after 2 months or less of treatment, that is, during the initial trial period. Another 10 patients
1132
ROLFE, OLSON, AND SAUNDERS
stopped using CPAP after 2 to 4 months of treatment, but thereafter only 5 patients stopped. The longest period of regular use followed by cessation was 2 yr. Discussion
Overall acceptance of long-term CPAP in these patients was 64070. Those with more severe disease at diagnosis (assessed by arterial desaturation, AHI, and EDS) were more likely to tolerate CPAP. However, even in those with minimal disease or with simple snoring a significant minority found CPAP acceptable. This is a retrospective study and may be affected by the problems inherent with this design, such as selection and observer bias and lack of a control group. Selection bias was minimized by using all patients commenced on CPAP since the establishment of a sleep center. There was probably some observer bias since clinicians' opinions on degree of EDS and indications for treatment are likely to differ. Since all patients referred for assessment of troublesome snoring had a sleep study, however, observer bias did not operate to a major degree at the diagnostic stage. It must also be noted that this is not a study of objective compliance. The study is based on the patients' reports of their usage of CPAP and is not a measure of how often or how much of each night it was used. However, there is evidence obtained from a time counter incorporated in the CPAP device that patient reports are an accurate measure of compliance (13). The finding that most patients with uncomplicated snoring did not tolerate CPAP is perhaps predictable. Several studies have shown the benefit of CPAP in abolishing snoring (14), but it is natural to expect that the absence of daytime symptoms would reduce motivation to continue therapy. It is also possible that individual physicians did not encourage continued use of CPAP by patients with mild or no OSA. Since the cost to the patient of the CPAP machine is approximately AUD 1,000, we consider that these results do not justify the routine use of CPAP for uncomplicated snoring (the cost to Australian patients of uvulopalatopharyngoplasty is substantially less). Wellmotivated patients experiencing severe social disruption who do not wish to have surgery may nevertheless respond well to CPAP. Where uvulopalatopharyngoplasty is expensive, CPAP may be the treatment of choice for many more patients than is the case in Australia.
The results we reported in relation to snoring and mild OSA should be considered in light of our arbitrary definition of OSA (apnea + hypopnea index > 15 h). In our experience and that of others (15), patients with fewer than 15 respiratory events per hour of sleep do not commonly have symptoms attributable to OSA, but those with more usually do. Our results in relation to the tolerance of CPAP may, however, have been influenced by our definition of OSA. If a higher cutoff figure had been adopted, more of those given CPAP would have been classified as being treated for snoring only, and more patients with mild disease and few symptoms might not have been offered CPAP at all. Tolerance of CPAP in those classified as having snoring and the mildest OSA might then have been better. Previous studies have suggested that it is the recurrence of daytime sleepiness on withdrawal of CPAP that is the major spur to continue treatment (7, 11).The findings of this study provide only limited support to this conclusion. Acceptance of CPAP was greater in those with more severe daytime sleepiness, but the differences were not great and do not justify a policy of restricting CPAP to severely symptomatic patients. It should be noted that EDS was assessed at presentation, and in some cases review after treatment was established, suggested that the patient's somnolence had initially been understated. Somnolence is difficult to assess by interview, and our results might have been different if we had measured somnolence objectively. On the other hand, we did not attempt to adjust our assessment of EDS to take account of the varying impact of somnolence on patients with different occupational or recreational activities. This is reflected in the significant number of patients who stopped CPAP because they found their symptoms less troublesome than the treatment (16of 61), but may also account for patients with minimal somnolence whose response to treatment was good. In our hands the best predictor of longterm acceptance of CPAP was severe hypoxemia. Hypoxemia without somnolence was associated with better tolerance of CPAP than severe somnolence without hypoxemia (tables 2 and 3). This may reflect the encouragement provided by physicians to patients whom they perceived to be at risk of serious complications, but many patients in this group also commented favorably on their improved quality of sleep and sense of well-being. These
aspects are less easily quantified than excessive sleepiness but should receive attention in studies of the symptomatic response to CPAP. The definition of severe hypoxemia used for this study (a lowest observed saturation below 75%) is arbitrary. We adopted a definition of severe hypoxemia based on the lowest observed arterial saturation because it has been suggested that in chronic lung disease this indicates those at risk of pulmonary hypertension (16), the complication of OSA most likely to be due to hypoxemia. There is no good evidence that this is the case in patients with OSA, but no other summary measure of exposure to hypoxemia has greater empiric validity, and the lowest observed arterial saturation appears in the present study to have influenced the tolerance of treatment. Of 61 patients discontinuing treatment, 10 did so because they experienced no benefit. These were predominantly patients with mild sleepiness, but 3 had unequivocal or severe EDS. It is likely that their somnolence was either not pathologic or was due to some other disorder (such as sleep myoclonus or narcolepsy) that was missed. However, of 117 patients in our sample with symptoms other than snoring, 99 reported resolution or improvement, so that we conclude that symptoms in patients with OSA can generally be assumed to be due to OSA until they fail to resolve with CPAP. The cost of CPAP was not commonly given as the reason for discontinuing treatment (6 of 61). For most of the period of study patients over 65 yr with moderate or severe disease were eligible to have CPAP provided as a Social Security benefit, but referral was at the discretion of the physician. In addition, patients with disease considered by the physician to be life threatening were provided with a machine by the unit if they could not afford to buy their own and were ineligible for the Social Security benefit. These factors may account for some of the association of both disease severity and age with continuation of treatment, particularly if the cost of treatment entered into the decision of patients discontinuing treatment because of "inconvenience." Intolerance of the mask was the major reason given for cessation of therapy (26 of 61), and this was true of all patient groups. Nasal congestion, mask discomfort, and claustrophobia have been reported in other studies as reasons for intolerance (8-10). Nasal congestion is most often transient, and if it persists usually
1133
ACCEPTANCE OF CPAP
responds to local therapy (with ipratropium bromide, vasoconstrictors, or beclomethasone). The newer CPAP machines now being developed are less noisy and have more comfortable and better fitting masks. Our results suggest that developments of this type, together with better training and education of patients, are the principal means by which the acceptability of CPAP will be improved. Acknowledgment The manuscript was typed by Mrs. Diane Kitevski. We are grateful to Dr. K. MurreeAllen and to Associate Professor M. J. Hensley for allowing us to include their patients in this report.
References 1. Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the
I
nares. Lancet 1981; 1:862-5. 2. Sullivan CE, Issa FG, Berthon-Jones M, McCauley VB, Costas LJV. Home treatment of obstructive sleep apnea with continuous positive airway pressure applied through a nose mask. Bull Eur Physiopathol Respir 1984; 2:49-54. 3. McEvoy RD, Thornton AT. Treatment of obstructive sleep apnoea syndrome with nasal continuous positive airway pressure. Sleep 1984; 7:313-25. 4. Sanders MH. Nasal CPAP effect on patterns of sleep apnoea. Chest 1984; 86:839-44. 5. Katsantonis GP, Schweitzer PK, Branham GH, Chambers G, Walsh JK. Management of obstructive sleep apnoea: comparison of various treatment modalities. Laryngoscope 1988; 98:304-9. 6. Frith RW, Cant BR. Severe obstructive sleep apnoea treated with long-term nasal continuous positive airway pressure. Thorax 1985; 40:45-50. 7. Rajagopal KR, Bennett LL, Dillard TA, Tellis CJ, Tenholder MF. Overnight nasal CPAP improves hypersomnolence in obstructive sleep apnoea. Chest 1986; 90:172-6. 8. Jiang H, Kryger MH, Zorick Fl, Conway W, Roth T. Mortality and apnoea index in obstructive sleep apnoea. Chest 1988; 94:9-14. 9. Frith RW, Cant BR. Obstructive sleep apnoea
in Auckland: diagnosis and treatment. N Z Med J 1985; 98:745-8. 10. Issa FG, Costas LV, Berthon-Jones M, McCauley VJ, Bruderer J, Sullivan CEo Nasal CPAP treatment for obstructive sleep apnea: longterm experience with 117patients. Am Rev Respir Dis 1985; 13:AI08. 11. Wagner DR, Pollack CP, Weitzman ED. Nocturnal nasal airway pressure for sleep apnoea. N Engl J Med 1983; 308:461-2. 12. Sullivan CE, Issa FG. Obstructive sleep apnoea. Clin Chest Med 1985; 6:633-48. 13. Sanders MH, Gruendle CA, Rogers RM. Patient compliance with nasal CPAP therapy for sleep apnoea. Chest 1986; 90:330-3. 14. Berry RB, Block JA. Positivenasal airway pressure eliminates snoring as well as obstructive sleep apnoea. Chest 1984; 85:15-20. 15. Gould G, Whyte K, Rhind G, et al. The sleep hypopnea syndrome. Am Rev Respir Dis 1988; 137:895-8. 16. Anthonisen NR. Home oxygen therapy. In: Isselbacher KJ, Adams RD, Braunwald E, Martin JB, Petersdorf RG, Wilson J, eds. Harrison's principles of internal medicine, update VI. New York: McGraw-Hill, 1985; 203-13.
ISOBEl ROLFE, L. G. OLSON, and N. A. SAUNDERS Introduction he use of nasal continuous positive airway pressure (CPAP) has been a major advance in the treatment of obstructive sleep apnea (OSA) (1). Many uncontrolled studies have shown benefit from long-term use of this device, particularly in patients with severe disease (2-10), and there is general agreement that it is the treatment of choice for obstructive sleep apnea. The main limitation of CPAP has been said to be initial intolerance and poor long-term compliance (5, 11). This may be of particular relevance when the treatment is used for patients with only mild daytime impairment or patients with asymptomatic hypoxemia or arrhythmia. There is, however, little information available on the long-term acceptability of nasal CPAP, especially in patients with mild disease. We reviewed 168 consecutive patients with obstructive sleep apnea of widely varying severity and report here their tolerance of nasal CPAP and the reasons for discontinuing treatment in patients who were intolerant.
T
Methods The reported data were obtained from the records of 168 consecutive patients commenced on CPAP from January 1983to June 1989at the Sleep Disorders Unit of the University of Newcastle and Royal Newcastle Hospital. The clinical features of the patients are shown in table 1. The steps that had been taken in initiating nasal CPAP treatment were as follows: 1. All patients had an initial history and examination by one of four respiratory physicians, of whom one (L.G.O.)saw about twothirds of the patients. Excessivedaytime somnolence was graded mild, unequivocal, or severe using the criteria suggested by Sullivan and Issa (12). It was the policy of the unit that all patients referred for assessment of snoring had a sleep study regardless of the presence or absence of symptoms suggesting sleep apnea. 2. Patients had an all-night diagnostic sleep study, using standard methods. Oxygenation was assessed by pulse oximeter (Biox" 3700 or 3740; Ohmeda, Louisville, CO); nasal airflow was assessed by thermistors and motion 1130
SUMMARY We studied the long-term acceptability of nasal continuous positive airway pressure (CPAP)treatment In 168 consecutive patients, 147 with obstructive sleep apnea (OSA) and 21 with snoring. Follow-up was between 1.5 and 78 months. At latest follow-up 107 of 168 (64%) were stili using CPAP. Acceptance of CPAP was least for patients with snoring alone (6 of 21 persisted) and best for patients with both excessive daytime somnolence and severe hypoxemia (minimum Sa0 2 < 75%), of whom 40 of 45 (89%) persisted with treatment. Patients with excessive daytime somnolence but without severe hypoxemia were less tolerant of CPAP (39 of 71,55%, persisted) than patients with no symptoms of excessive somnolence but with severe hypoxemia (21 of 30, 70%, persisted). The most common reasons for discontinuing CPAP were Intolerance of the mask (26 of 61), the Inconvenience of treatment (16 of 61), and the lack of symptomatic benefit from treatment (10 of 61). We concluded that long-term acceptance of CPAP was difficult to predict in advance but that it was most likely In patients with the most severe sleep apnea. Because Intolerance of the mask and inconvenience were the most common reasons for ceasing treatment, Improvements In the design of CPAP systems and careful patient training may Improve the acceptability of CPAP substantially. AM REV RESPIR DIS 1991; 144:1130-1133
of the rib cage and abdomen by inductance plethysmography (Respitrace"; Ambulatory Monitoring Inc, Ardsley, NY) Sleep was staged using one channel of EEG, two channels of EOG, and submental EMG. Signals were recorded on an ink pen recorder. All studies considered in this report were analyzed by a single technician. Blind repeat scoring of studies selected from the whole period relevant to this report did not suggest variation of scoring criteria. We recorded the apnea plus hypopnea index (AHI) and the minimum arterial oxygen saturation (Sao.}, OSA was diagnosed when the AHI wasgreater than 15. Apnea was defined as cessation of airflow for more than 10 sand hypopnea as reduction for at least 10 s of the amplitude of both oronasal airflow and ribcage and abdomen motion signals to less than half the levels prevailing before and after the event. Severe arterial desaturation was defined as Saol at any time less than 750/0. 3. Nasal CPAP was initiated during a second all-night study, at which time the airway pressure was adjusted to abolish all apneas and hypopneas associated with arterial oxygen desaturation in all sleep stages, including rapid eye movement sleep. At this time the patient's physician gave a careful explanation of the nature and purpose of CPAP treatment. The patients were told that there was no alternative therapy for OSA, but that in the absence of severe desaturation the only purpose of treatment was to relievedaytime symptoms. Patients with severe desaturation were told that they might be at risk of pulmonary hypertension, but the degree of emphasis placed
on this varied among physicians. When CPAP was initiated the fitting of the mask and the use of the machine were explained by a technician. Patients returned after a 6-wk trial of CPAP for a further sleep study and pressure check. At this first follow-up appointment problems with CPAP were discussed and the purpose of therapy was reviewed before the patient was committed to the purchase of a CPAP machine. In some cases the trial period was extended to 8 or 10wk. No home visits were made, but patients were encouraged to telephone the laboratory for help with any problem, and technical problems were reviewed without appointments. Those discontinuing CPAP werein some cases not followed thereafter and, in other cases, restudied annually. Follow-up of patients persisting with CPAP was at variable intervals, ranging from 6 to 14 months. At each follow-up visit patient compliance, changes in symptoms, and problems with treatment were noted by the attending physician. The laboratory routinely offered to purchase the CPAP machine of (Received in original form February 8, 1990 and in revised form May 14, 1991) 1 From the Faculty of Medicine, University of Newcastle, New South Wales, Australia. 2 Supported by the National Health and Medical Research Council (Australia). 3 Correspondence and requests for reprints should be addressed to Dr. L. G. Olson, Department of Medicine, University of Newcastle, David Maddison Clinical Sciences Building, Royal Newcastle Hospital, Newcastle, NSW 2300, Australia.
ACCEPTANCE OF CPAP
1131 TABLE 1
TABLE 2
CLINICAL CHARACTERISTICS OF PATIENTS (n = 168)
TOLERANCE OF CPAP IN PATIENTS WITH EDS ALONE*
BMI, * mean ± 95% CI Obeset Hypertensive:!: Observed apneic§ Complaining of frequent choking arousals Complaining of restless sleep Complaining of frequent morning fatigue Consuming ethanol> 30 g/day
% % % % % % %
31.86 ± 0.84 66 41 79 60 54 40 19.5
• BMI = weight kg/height m2 • t Obesity = 8MI > 30. t Hypertension = systolic BP > 150 mm Hg or diastolic> 95 mm Hg. § As reported by bed partner.
a patient who had bought one but later ceased to use it. Clinicians gave four major reasons for commencing patients on CPAP: 1. Snoring in the absence of GSA in patients who wished to avoid other therapy (Le., surgery) 2. Excessive daytime sleepiness (EDS) in patients who did not have severe arterial oxygen desaturation 3. Severe arterial oxygen desaturation (Sa02 < 75070) despite mild or no somnolence 4. EDS and severe arterial desaturation, both being of sufficient severity to justify therapy.
those who did and those who did not persist with treatment, although AHI tended to be higher in those who continued. All the patients who continued treatment had improvement or resolution of EDS. Of those who stopped, 9 did so because there was no improvement in their symptoms, 11 because they could not tolerate the mask, 8 because they believed that the improvement in their symptoms did not justify the inconvenience of treatment, and 3 because they could not afford to buy a CPAP device.In the remaining case the reason for stopping CPAP was uncertain. Arterial hypoxemia. There were 30 paResults tients in whom the primary indication A total of 168patients were followed for for treatment was severe arterial desatubetween 6 wk and 78 months. There were 145 (86070) males. Overall, 107(64070) pa- ration in the absence of daytime sleepitients continued to use nasal CPAP at ness thought at the initial assessment to be significant (table 3). Mean follow-up latest follow-up. was 15.5months (range 1.5to 48 months). The 21 patients (70%) who continued Outcome According to Indication CPAP had more severe arterial desatuSnoring. A group of 21patients werefol- ration at diagnosis and a tendency to a lowed for a mean of 10.3months (range higher AHI, but the difference between 2 to 52 months). Of 21patients, 6 (29%) the groups was not significant. All patients who had symptoms recontinued CPAP during follow-up. Of the patients who discontinued CPAP, 4 ported that they improved or resolved. later had uvulopalatopharyngoplasties The 9 patients who ceased therapy found with good results. The 6 patients who the device either intolerably uncomfortcontinued use of CPAP had resolution able (5 of 9) or intolerably inconvenient. of symptoms. They have continued to use EDS and Hypoxemia. A total of 45 paCPAP for 24 to 52 months. Of the 15 tients were followed for a mean of 25 patients who stopped CPAP, 8 did so be- months (range 2 to 70 months) (table 4). cause of discomfort or claustrophobia Of these, most (89%) continued therapy caused by the mask. and all reported improvement in sympExcessive daytime sleepiness. There toms. Those who continued with treatwere 71 patients treated for this indica- ment appeared to have more severe distion, followed for a mean of 22 months ease, but the differences between the (range 1.5 to 79 months) (table 2). A to- groups was not statistically significant. tal of 39 patients (55070) continued treat- Of the patients who discontinued CPAP, ment. Arterial oxygen desaturation at di- 2 did so because they could not tolerate agnosis was greater in the 39 who con- the mask, 1 because the improvement tinued with CPAP than in the 32 who resulting from treatment did not justify stopped (p < 0.025 by Wilcoxon's test), the inconvenience oftreatment, and 1 bealthough the difference was clinically in- cause of the cost of a machine. In one significant. There were no other statis- case, the reason for discontinuing treattically significant differences between ment was unclear.
Continued n Males Age, yr AHI, events/h Minimum Sa0 2, % EDS Mild Unequivocal Severe
Ceased
39 (55%)
32
35/39 58.2 ± 2.8 30.8 ± 8.1 82.1 ± 1.4
28/32 58.0 ± 4.0 22.8 ± 7.0 84.3 ± 1.5
23 12 4
20 9 3
• Means are given ± 95% CI.
TABLE 3 TOLERANCE OF CPAP IN PATIENTS WITH HYPOXEMIA ALONE* Continued n Males Age, yr AHI, events/h Minimum Sa02. %
Ceased
21 (70%)
9
19/21
8/9
57.2 ± 7.1 43.0 ± 12.3 61.5 ± 4.0
57.2 ± 8.4 35.7 ± 15.0 68.3 ± 5.9
• Means are given ± 95% CI.
TABLE 4 TOLERANCE OF CPAP IN PATIENTS WITH BOTH EDS AND HYPOXEMIA* Continued n Males Age, yr AHI, events/h Minimum Sa02, % EDS Mild Unequivocal Severe
40 (89%) 37/40 54.9 ± 3.0 51.5 ± 9.0 62.7 ± 3.4
Ceased
5 4/5
57.8 ± 5.8 39.4 ± 43.8 63.8 ± 11.9
6
23 11
4 1
• Means are given ± 95% CI.
Outcome According to Severity ofEDS When patients were grouped according to severity of EDS, there was a significant positive relationship between the severity of EDS and the probability of accepting long-term CPAP. Of those with mild EDS, 29 of 49 (59%) continued; of those with unequivocal EDS, 35 of 48 (74%) continued; and of those with severeEDS, 15of 19(79%) continued (p < 0.02 by chi-square test). This relationship was not, however, apparent in patients with milder OSA (table 2). Duration of Treatment in Patients Ceasing CPAP Of the 61 patients who discontinued CPAP, 45 (78%) did so after 2 months or less of treatment, that is, during the initial trial period. Another 10 patients
1132
ROLFE, OLSON, AND SAUNDERS
stopped using CPAP after 2 to 4 months of treatment, but thereafter only 5 patients stopped. The longest period of regular use followed by cessation was 2 yr. Discussion
Overall acceptance of long-term CPAP in these patients was 64070. Those with more severe disease at diagnosis (assessed by arterial desaturation, AHI, and EDS) were more likely to tolerate CPAP. However, even in those with minimal disease or with simple snoring a significant minority found CPAP acceptable. This is a retrospective study and may be affected by the problems inherent with this design, such as selection and observer bias and lack of a control group. Selection bias was minimized by using all patients commenced on CPAP since the establishment of a sleep center. There was probably some observer bias since clinicians' opinions on degree of EDS and indications for treatment are likely to differ. Since all patients referred for assessment of troublesome snoring had a sleep study, however, observer bias did not operate to a major degree at the diagnostic stage. It must also be noted that this is not a study of objective compliance. The study is based on the patients' reports of their usage of CPAP and is not a measure of how often or how much of each night it was used. However, there is evidence obtained from a time counter incorporated in the CPAP device that patient reports are an accurate measure of compliance (13). The finding that most patients with uncomplicated snoring did not tolerate CPAP is perhaps predictable. Several studies have shown the benefit of CPAP in abolishing snoring (14), but it is natural to expect that the absence of daytime symptoms would reduce motivation to continue therapy. It is also possible that individual physicians did not encourage continued use of CPAP by patients with mild or no OSA. Since the cost to the patient of the CPAP machine is approximately AUD 1,000, we consider that these results do not justify the routine use of CPAP for uncomplicated snoring (the cost to Australian patients of uvulopalatopharyngoplasty is substantially less). Wellmotivated patients experiencing severe social disruption who do not wish to have surgery may nevertheless respond well to CPAP. Where uvulopalatopharyngoplasty is expensive, CPAP may be the treatment of choice for many more patients than is the case in Australia.
The results we reported in relation to snoring and mild OSA should be considered in light of our arbitrary definition of OSA (apnea + hypopnea index > 15 h). In our experience and that of others (15), patients with fewer than 15 respiratory events per hour of sleep do not commonly have symptoms attributable to OSA, but those with more usually do. Our results in relation to the tolerance of CPAP may, however, have been influenced by our definition of OSA. If a higher cutoff figure had been adopted, more of those given CPAP would have been classified as being treated for snoring only, and more patients with mild disease and few symptoms might not have been offered CPAP at all. Tolerance of CPAP in those classified as having snoring and the mildest OSA might then have been better. Previous studies have suggested that it is the recurrence of daytime sleepiness on withdrawal of CPAP that is the major spur to continue treatment (7, 11).The findings of this study provide only limited support to this conclusion. Acceptance of CPAP was greater in those with more severe daytime sleepiness, but the differences were not great and do not justify a policy of restricting CPAP to severely symptomatic patients. It should be noted that EDS was assessed at presentation, and in some cases review after treatment was established, suggested that the patient's somnolence had initially been understated. Somnolence is difficult to assess by interview, and our results might have been different if we had measured somnolence objectively. On the other hand, we did not attempt to adjust our assessment of EDS to take account of the varying impact of somnolence on patients with different occupational or recreational activities. This is reflected in the significant number of patients who stopped CPAP because they found their symptoms less troublesome than the treatment (16of 61), but may also account for patients with minimal somnolence whose response to treatment was good. In our hands the best predictor of longterm acceptance of CPAP was severe hypoxemia. Hypoxemia without somnolence was associated with better tolerance of CPAP than severe somnolence without hypoxemia (tables 2 and 3). This may reflect the encouragement provided by physicians to patients whom they perceived to be at risk of serious complications, but many patients in this group also commented favorably on their improved quality of sleep and sense of well-being. These
aspects are less easily quantified than excessive sleepiness but should receive attention in studies of the symptomatic response to CPAP. The definition of severe hypoxemia used for this study (a lowest observed saturation below 75%) is arbitrary. We adopted a definition of severe hypoxemia based on the lowest observed arterial saturation because it has been suggested that in chronic lung disease this indicates those at risk of pulmonary hypertension (16), the complication of OSA most likely to be due to hypoxemia. There is no good evidence that this is the case in patients with OSA, but no other summary measure of exposure to hypoxemia has greater empiric validity, and the lowest observed arterial saturation appears in the present study to have influenced the tolerance of treatment. Of 61 patients discontinuing treatment, 10 did so because they experienced no benefit. These were predominantly patients with mild sleepiness, but 3 had unequivocal or severe EDS. It is likely that their somnolence was either not pathologic or was due to some other disorder (such as sleep myoclonus or narcolepsy) that was missed. However, of 117 patients in our sample with symptoms other than snoring, 99 reported resolution or improvement, so that we conclude that symptoms in patients with OSA can generally be assumed to be due to OSA until they fail to resolve with CPAP. The cost of CPAP was not commonly given as the reason for discontinuing treatment (6 of 61). For most of the period of study patients over 65 yr with moderate or severe disease were eligible to have CPAP provided as a Social Security benefit, but referral was at the discretion of the physician. In addition, patients with disease considered by the physician to be life threatening were provided with a machine by the unit if they could not afford to buy their own and were ineligible for the Social Security benefit. These factors may account for some of the association of both disease severity and age with continuation of treatment, particularly if the cost of treatment entered into the decision of patients discontinuing treatment because of "inconvenience." Intolerance of the mask was the major reason given for cessation of therapy (26 of 61), and this was true of all patient groups. Nasal congestion, mask discomfort, and claustrophobia have been reported in other studies as reasons for intolerance (8-10). Nasal congestion is most often transient, and if it persists usually
1133
ACCEPTANCE OF CPAP
responds to local therapy (with ipratropium bromide, vasoconstrictors, or beclomethasone). The newer CPAP machines now being developed are less noisy and have more comfortable and better fitting masks. Our results suggest that developments of this type, together with better training and education of patients, are the principal means by which the acceptability of CPAP will be improved. Acknowledgment The manuscript was typed by Mrs. Diane Kitevski. We are grateful to Dr. K. MurreeAllen and to Associate Professor M. J. Hensley for allowing us to include their patients in this report.
References 1. Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversal of obstructive sleep apnoea by continuous positive airway pressure applied through the
I
nares. Lancet 1981; 1:862-5. 2. Sullivan CE, Issa FG, Berthon-Jones M, McCauley VB, Costas LJV. Home treatment of obstructive sleep apnea with continuous positive airway pressure applied through a nose mask. Bull Eur Physiopathol Respir 1984; 2:49-54. 3. McEvoy RD, Thornton AT. Treatment of obstructive sleep apnoea syndrome with nasal continuous positive airway pressure. Sleep 1984; 7:313-25. 4. Sanders MH. Nasal CPAP effect on patterns of sleep apnoea. Chest 1984; 86:839-44. 5. Katsantonis GP, Schweitzer PK, Branham GH, Chambers G, Walsh JK. Management of obstructive sleep apnoea: comparison of various treatment modalities. Laryngoscope 1988; 98:304-9. 6. Frith RW, Cant BR. Severe obstructive sleep apnoea treated with long-term nasal continuous positive airway pressure. Thorax 1985; 40:45-50. 7. Rajagopal KR, Bennett LL, Dillard TA, Tellis CJ, Tenholder MF. Overnight nasal CPAP improves hypersomnolence in obstructive sleep apnoea. Chest 1986; 90:172-6. 8. Jiang H, Kryger MH, Zorick Fl, Conway W, Roth T. Mortality and apnoea index in obstructive sleep apnoea. Chest 1988; 94:9-14. 9. Frith RW, Cant BR. Obstructive sleep apnoea
in Auckland: diagnosis and treatment. N Z Med J 1985; 98:745-8. 10. Issa FG, Costas LV, Berthon-Jones M, McCauley VJ, Bruderer J, Sullivan CEo Nasal CPAP treatment for obstructive sleep apnea: longterm experience with 117patients. Am Rev Respir Dis 1985; 13:AI08. 11. Wagner DR, Pollack CP, Weitzman ED. Nocturnal nasal airway pressure for sleep apnoea. N Engl J Med 1983; 308:461-2. 12. Sullivan CE, Issa FG. Obstructive sleep apnoea. Clin Chest Med 1985; 6:633-48. 13. Sanders MH, Gruendle CA, Rogers RM. Patient compliance with nasal CPAP therapy for sleep apnoea. Chest 1986; 90:330-3. 14. Berry RB, Block JA. Positivenasal airway pressure eliminates snoring as well as obstructive sleep apnoea. Chest 1984; 85:15-20. 15. Gould G, Whyte K, Rhind G, et al. The sleep hypopnea syndrome. Am Rev Respir Dis 1988; 137:895-8. 16. Anthonisen NR. Home oxygen therapy. In: Isselbacher KJ, Adams RD, Braunwald E, Martin JB, Petersdorf RG, Wilson J, eds. Harrison's principles of internal medicine, update VI. New York: McGraw-Hill, 1985; 203-13.
