Abstract
Objective:
Continuous positive airway pressure (CPAP) has been established as an effective treatment for obstructive sleep apnea-hypopnea syndrome (OSAHS). Recently, several auto-CPAP devices that can detect upper airway obstructive events and provide information about residual events while patients are on CPAP have come into clinical use. The purpose of this study was to compare the apnea-hypopnea index (AHI) determined by the S8 auto-CPAP device with the AHI derived by polysomnography in patients with OSAHS.
Method:
Consecutive patients with OSAHS titrated on S8 auto-CPAP were included. The correlation between AHI determined by manual scoring (AHI-PSG) and by S8 (AHI-S8) during an overnight in-hospital polysomnogram with the patient on CPAP was assessed. Furthermore, the apnea index (AI) and the hypopnea index (HI) were evaluated separately.
Results:
Seventy patients with OSAHS (94% men) were enrolled. The mean AHI on the diagnostic study was 51.9 ± 2.4. During the titration, this device markedly suppressed the respiratory events (AHI-PSG, 4.2 ± 0.4; AI, 1.9 ± 0.3; HI, 2.3 ± 0.3). On the other hand, the AHI-S8 was 9.9 ± 0.6 (AI-S8, 2.4 ± 0.3; HI-S8, 7.5 ± 0.4). There was a strong correlation between the overall AHI-PSG and the AHI-S8 (r = 0.85, p < 0.001), with a stronger correlation in the apnea component AI-PSG and the AI-S8 (r = 0.93, p < 0.001), whereas there was a weaker correlation between the HI-PSG and the HI-S8 (r = 0.67, p < 0.001).
Conclusions:
Using the same airflow signals as those of the CPAP device, a strong correlation between the AHI-PSG and the AHI-S8 was observed. However, the correlation was weakened when the analysis was limited to the HI.
Citation:
Ueno K; Kasai T; Brewer G; Takaya H; Maeno K; Kasagi S; Kawana F; Ishiwata S; Narui K. Evaluation of the apnea-hypopnea index determined by the S8 auto-CPAP, a continuous positive airway pressure device, in patients with obstructive sleep apnea-hypopnea syndrome. J Clin Sleep Med 2010;6(2):146-151.
Keywords: Apnea, hypopnea, polysomnography, titration
Obstructive sleep apnea-hypopnea syndrome (OSAHS) has been recognized as a highly prevalent disorder and a significant predictor of all-cause and cardiovascular mortality.1–3 On the other hand, continuous positive airway pressure (CPAP) has been established as an effective treatment for the attenuation of apneas and hypopneas and, therefore, the symptoms associated with OSAHS,4–7 along with a reduction in overall mortality.8 In fact, a large number of subjects with OSAHS have been identified,1 and CPAP has become widely used as the therapy for OSAHS in recent years. In most patients with OSAHS, CPAP titration is performed to confirm the suppression of apneas and hypopneas, and, then, CPAP with the appropriate pressure setting is started. However, CPAP is often not retitrated after the initial titration study unless patients complain about CPAP use.
BRIEF SUMMARY
Current Knowledge/Study Rationale: Currently, several auto-CPAP devices designed to detect upper airway obstruction and provide information about the residual apnea-hypopnea index (AHI) while patients are on CPAP have been developed and have come into clinical use for patients with obstructive sleep apnea-hypopnea syndrome (OSAHS). However, data evaluating the AHI determined by the S8 auto-CPAP device are sparse.
Study Impact: This study has value in the clinical setting in helping to assess whether respiratory events are optimally treated with the prescribed pressure setting after initiation of CPAP therapy. Furthermore, this study suggests that residual AHI should be interpreted with caution in cases with severe OSAHS that had more leakage.
Currently, several auto-CPAP devices designed to detect upper airway obstructive events and provide information about the residual apnea-hypopnea index (AHI) while patients are on CPAP have been developed and have come into clinical use. However, although some data have shown the accuracy of the AHI determined by such devices during diagnostic studies,9 comparison data about residual AHI on CPAP do not exist. Furthermore, the factors contributing to differences between the AHI scored on polysomnography and that determined by the auto-CPAP devices have not been clarified. The aim of the this study was to compare the residual AHI on CPAP determined by a current auto-CPAP device (S8 Respond, ResMed Inc, Sydney, NSW, Australia) with that obtained by manually scored polysomnography and to identify associated factors contributing to these differences.
METHOD
Subjects
Consecutive patients diagnosed with moderate to severe OSAHS on diagnostic polysomnography followed by overnight titration using the S8 auto-CPAP device were enrolled from our sleep center (Toranomon Hospital, Tokyo, Japan) between May and November 2006. The exclusion criteria were (1) patients with heart failure, who had had a stroke, or who were on dialysis; (2) patients with central AHI of more than 5 events per hour and at least 10% of the total AHI was related to central events during diagnostic polysomnography; and (3) patients with any central events on CPAP titration.
The Toranomon Hospital Ethics Committee approved the study, and informed consent to participate was obtained from all patients.
Measurements
All polysomnography was performed using a digital polygraph (SomnoStar α Sleep System; SensorMedics Corp., Yorba Linda, CA) monitored by a sleep technician. The accepted definitions and scoring methods for the diagnosis of OSAHS were used for both diagnostic polysomnography and CPAP titration,10,11 i.e., apnea was defined as a cessation of air flow for at least 10 seconds; hypopnea was defined as a reduction of airflow by more 50% for at least 10 seconds or discernible reduction of airflow with desaturation (using a baseline of 120 s) or an arousal. During CPAP titration, polysomnography was performed while patients were wearing the S8 auto-CPAP device, and, therefore, the air flow signals of the polysomnogram were obtained from the nasal mask of the CPAP device. The pressure level was titrated manually using standard titration procedures.12,13 The S8 auto-CPAP device automatically recorded apneas and hypopneas according to the following definition: apnea events were scored when a decrease in air flow was detected that was greater than 75%, compared with the recent average (time constant of 100 s), for more than 10 seconds; apnea-hypopnea events were scored when a decrease in air flow was observed that was greater than 50%, compared with the recent average (time constant of 100 s), for more than 10 seconds. All recordings were made simultaneously during CPAP titration.
In each case, the respiratory events were manually scored, and the AHI (AHI-PSG) was calculated per hour of sleep; the apnea index (AI-PSG) and the hypopnea index (HI-PSG) were also determined. Furthermore, the AHI determined by the S8 auto-CPAP device (AHI-S8), as well as the apnea index (AI-S8) and the hypopnea index (HI-S8), were calculated per hour of recording. The S8 auto-CPAP device does not specifically report the HI, but the HI can be calculated as the difference between the AHI and the AI. Simultaneously, the S8 auto-CPAP device reported the overnight summary data for leakage and provided the pressure level (maximum, 95th percentile, and the median level of leakage or provided pressure).
After each CPAP titration, polysomnography scoring was performed before generating the S8 report.
Statistical Analysis
The data are expressed as the mean ± SEM for continuous variables or numbers and percentages for categorical variables. The data derived from polysomnography were compared with those derived from the S8 using the Wilcoxon rank-sum test, correlation analysis, and Bland and Altman plots.14 The associations between several characteristics and polysomnographic findings of patients, including age, sex, body mass index, Epworth Sleepiness Scale scores, findings on diagnostic polysomnography (AHI, AI, percentage of slow wave sleep, percentage of rapid eye movement (REM) sleep, periodic limb movement index), findings on CPAP titration (AHI, AI, percentage of slow wave sleep, percentage of REM sleep, periodic limb movement index), data from the S8 (AHI-S8, 95th percentile pressure, 95th percentile leakage), and the difference between the AHI-S8 and the AHI-PSG, were analyzed using simple and stepwise multiple regression analyses. A p value of less than 0.05 was taken as the level of statistical significance.
RESULTS
Overall, 70 patients were assessed. The characteristics and findings of diagnostic polysomnography are shown in Table 1. The mean total recording times were 7.2 ± 0.1 hours for polysomnography and 7.1 ± 0.1 hours for the S8. The total sleep time during polysomnography was 6.5 ± 0.1 hours. The polysomnographic findings and data from the S8 on CPAP titration are summarized in Tables 2 and 3, respectively.
Table 1.
Patient characteristics and findings from diagnostic polysomnography.
| Age, y | 52.2 ± 1.4 |
| Men | 66 (94) |
| BMI, kg/m2 | 28.7 ± 0.6 |
| ESS score | 10.3 ± 0.6 |
| AHI, events/h | 51.9 ± 2.4 |
| AI, events/h | 41.6 ± 2.8 |
| HI, events/h | 10.5 ± 1.1 |
| Arousal index, events/h | 48.8 ± 2.4 |
| Sleep stage, % of TST | |
| SWS | 3.6 ± 4.2 |
| REM | 7.3 ± 1.0 |
Data are presented as mean ± SEM, except men, which is shown as number (%). BMI refers to body mass index; ESS, Epworth Sleepiness Scale; AHI, apnea-hypopnea index; AI, apnea index; HI, hypopnea index; TST, total sleep time; SWS, slow-wave sleep; REM, rapid eye movement.
Table 2.
Polysomnographic findings on CPAP titration.
| TST, h | 6.5 ± 0.1 |
| AHI-PSG, events/h | 4.2 ± 0.4 |
| AI-PSG, events/h | 1.9 ± 0.3 |
| HI-PSG, events/h | 2.3 ± 0.3 |
| Arousal Index, events/h | 12.6 ± 0.8 |
| Sleep stage, % of TST | |
| SWS | 18.2 ± 1.0 |
| REM | 21.7 ± 0.8 |
Data are presented as mean ± SEM. CPAP refers to continuous positive airway pressure; TST, total sleep time; AHI-PSG, apnea-hypopnea index assessed by polysomnography; AI-PSG, apnea index assessed by polysomnography; HI-PSG, hypopnea index assessed by polysomnography; SWS, slow-wave sleep; REM, rapid eye movement.
Table 3.
Data from the S8 during CPAP titration.
| AHI-S8, events/h | 9.9 ± 0.6 |
| AI-S8, events/h | 2.4 ± 0.3 |
| HI-S8, events/h | 7.5 ± 0.4 |
| Pressure, cm H2O | |
| Maximum | 13.5 ± 0.3 |
| 95th percentile | 12.0 ± 0.3 |
| Median | 9.7 ± 0.3 |
| Leak, L/s | |
| Maximum | 0.58 ± 0.03 |
| 95th percentile | 0.26 ± 0.02 |
| Median | 0.09 ± 0.01 |
Data are presented as mean ± SEM. CPAP refers to continuous positive airway pressure; AHI-S8, apnea-hypopnea index determined by the S8 device; AI-PSG, apnea index assessed on polysomnography; HI-S8, hypopnea index determined by the S8 device.
The mean AHI-PSG was 4.2 ± 0.4, whereas the mean AHI-S8 was 9.9 ± 0.6 (p < 0.001). Figure 1A shows the correlation between the AHI-PSG and the AHI-S8 (r = 0.85; p < 0.001). Figure 1B shows a Bland and Altman plot, which showed a mean AHI difference of −5.7, and the limits of agreement for the AHI were from +0.1 to −11.5.
Figure 1A.
Scatter plots of the apnea-hypopnea index (AHI) derived from the polysomnogram (AHI-PSG) and the AHI from the S8 auto-continuous positive airway pressure (CPAP) device (AHI-S8).
A–There is a strong correlation between the AHI-PSG and the AHI-S8.
Figure 1B.
Bland-Altman plots according to the AHI.
B–The Y axis indicates the difference between the AHI-PSG and the AHI-S8 ([AHI-PSG]-[AHI-S8]). The X axis indicates the mean values. The solid line represents the mean difference; the dashed lines represent the limit of agreement (± 2 SD).
Similarly, the difference between the AI-PSG (4.2 ± 0.4) and the AI-S8 (2.4 ± 0.3) was significant (p < 0.001). There was also a strong correlation between the AI-PSG and the AI-S8 (r = 0.93; p < 0.001, Figure 2A). Figure 2B shows a Bland and Altman plot, which showed a mean AI difference of −0.5; the limits of agreement for the AI were from +1.3 to −2.3.
Figure 2A.
Scatter plots of the apnea index (AI) derived from the polysomnogram (AI-PSG) and the AI from the S8 auto-continuous positive airway pressure device (AI-S8).
A–There is a strong correlation between the AI-PSG and the AI-S8.
Figure 2B.
Bland-Altman plots according to the AI.
B–The Y axis indicates the difference between the AI-PSG and the AI-S8 ([AI-PSG]-[AI-S8]). The X axis indicates the mean values. The solid line represents the mean difference; the dashed lines represent the limit of agreement (± 2 SD).
Furthermore, the difference between the HI-PSG (2.3 ± 0.3) and the HI-S8 (7.5 ± 0.4) was also significant (p < 0.001). There was a modest correlation between the HI-PSG and the HI-S8 (r = 0.67; p < 0.001, Figure 3A). Figure 3B shows a Bland and Altman plot, which showed a mean HI difference of −5.3; the limits of agreement for the HI were from +0.1 to −10.7.
Figure 3A.
Scatter plots of the hypopnea index (HI) derived from the polysomnogram (HI-PSG) and the HI from the S8 auto-continuous positive airway pressure (CPAP) device (HI-S8).
A–There is a strong correlation between the HI-PSG and the HI-S8.
Figure 3B.
Bland-Altman plots according to the HI. The Y axis indicates the difference between the HI-PSG and the HI-S8 ([HI-PSG]-[HI-S8]).
B–The X axis indicates the mean values. The solid line represents the mean difference; the dashed lines represent the limit of agreement (± 2 SD).
The pressure profile and the data about leakage during CPAP titration are also shown in Table 3. The pressure level was approximately 10 cm H2O, and there was no significant leakage during CPAP titration.
Factors that were significantly associated with greater differences between the AHI-S8 and the AHI-PSG in the simple and stepwise multiple regression analyses are summarized in Table 4. Higher 95th percentile leakage and higher AHI on diagnostic polysomnography were associated with greater differences between the AHI-S8 and the AHI-PSG in the final stepwise regression model.
Table 4.
Simple and stepwise multiple regression analyses.
| Simple |
Stepwise multiple |
|||
|---|---|---|---|---|
| β | p | β | p | |
| AHI from the diagnostic study | 0.302 | 0.012 | 0.325 | 0.048 |
| SWS from the diagnostic study | 0.276 | 0.023 | - | - |
| AHI-PSG | 0.298 | 0.012 | - | - |
| 95th percentile pressure | 0.287 | 0.016 | - | - |
| 95th percentile leakage | 0.292 | 0.014 | 0.329 | 0.044 |
AHI refers to apnea-hypopnea index; SWS, slow-wave sleep; AHI-PSG, apnea-hypopnea index assessed by polysomnography during continuous positive airway pressure (CPAP) titration.
DISCUSSION
Our data suggest a strong correlation between the AHI determined by the S8 auto-CPAP device and the AHI derived from simultaneous polysomnography. However, the S8 device tended to report higher indexes of respiratory events (approximately 5 for the AHI and the HI, but only 0.5 for the AI) than those derived from polysomnography. In addition, greater severity of sleep apnea and degree of mask leakage were associated with a greater difference between the AHI-S8 and the AHI-PSG. These results have significant value in the clinical setting to assess whether respiratory events are optimally treated with the prescribed pressure setting after initiation of CPAP therapy.
Similar results have been reported in other studies in which the possibility of replacing the diagnostic evaluations using polysomnography by those using a CPAP device was investigated. Gugger et al have shown comparison data between AHI determined by a CPAP device and that obtained by manual scoring using an older version of a similar device (AutoSet software version 3.03, which was manufactured by ResMed, Inc., and was specialized for diagnostic evaluation).9 They found that AHI determined by the CPAP device correlated strongly with manually scoring, albeit the AHI scored by the former was systematically greater (the mean difference: 4.2). In the present study, we compared the residual AHI reported by the S8 device in the on-CPAP situation with the AHI derived from simultaneous polysomnography; the results were compatible with those in Gugger et al's report and were acceptable, though the results of the present study differed from those of Gugger et al in that respiratory events were suppressed by the S8 per se, which is more significant information for clinicians.
Considering the result of the specific analysis of the HI, hypopnea events, compared with the apnea events, by the S8 device were more often overestimated, and this caused the difference between the AHI-PSG and the AHI-S8. This finding is also compatible with the results of previous studies that have shown the comparison data of the indexes of respiratory events determined by a similar CPAP device and that from diagnostic polysomnography.9,15,16
Such overestimation of HI-S8 events is not an unexpected if we take into account how the auto-CPAP works. To test whether the CPAP level is too high, whenever the auto-CPAP detects that airflow is maintained as normal breathing by the provided pressure level, it begins to lower the pressure level until an appearance of a hypopnea or a change in the shape of the airflow suggests airway collapse or flow limitation. When this is detected, the pressure immediately rises in principle before there is an arousal. This might be perceived as a subtle hypopnea if the analysis includes information derived only from the airflow signal. In contrast, on polysomnography, the same reduction in airflow will fail to qualify as a hypopnea because there is no arousal and often no desaturation. Therefore, HI determined by auto-CPAP might be overestimated. However, such overestimation of hypopneas by the auto-CPAP, which is derived from testing whether the CPAP level is too high, is the key of the auto-CPAP concept. From this point of view, the AHI determined by auto-CPAP should be evaluated by a cut-off value of AHI different from that derived by manual scoring of polysomnography, i.e., based on our analysis, an AHI cut-off of less than 10 improves accuracy, compared with an AHI less than 5.
However, in other studies that have compared the AHI determined by auto-CPAP and the AHI derived from diagnostic polysomnography, the outliers were observed in patients who had periodic limb movement disorder, although no obvious reason for the discrepancy in the AHI obtained by the 2 methods was found.9,15,16 In the present study, the regression analysis showed no significant relationship between the AHI discrepancy and the periodic limb movement index either on the diagnostic study or during CPAP titration. Factors associated with the AHI discrepancy on the simple regression analysis were a greater severity of sleep apnea and less slow wave sleep on the diagnostic study, greater AHI-PSG, higher 95th percentile pressure level, and higher 95th percentile leakage. This implies that the AHI-S8 was overscored in cases with more severe OSAHS and a greater percentage of slow wave sleep on the diagnostic study, in those with a higher residual AHI on polysomnography during CPAP titration, in those who required a higher pressure to suppress respiratory events, and in those with greater leakage. Furthermore, considering the results of multiple regression analysis, the AHI-S8, compared with the AHI-PSG, tended to be overestimated in patients with severe OSAHS who had more leakage. To the best of our knowledge, this is the first report to identify the factors associated with AHI-S8 overestimation. From this perspective, the present study has significant value.
There were some limitations associated with our study. First, we used airflow signal given by the S8 device for polysomnography and did not use an independent way to measure the airflow during CPAP; this might have favorably influenced correlation between indexes of respiratory events determined by S8 and those derived from polysomnography. Second, if the patient is predominantly mouth breathing, there may be many reductions of airflow without desaturation and arousal, and such episodes are also associated with overestimation of the AHI by the S8. Third, we excluded patients with central events on CPAP titration because the algorithm of the S8 device is basically used for detecting upper airway obstructive events and cannot detect central events with confidence. This may also have favorably influenced correlation between indexes of respiratory events determined by S8 and those derived from polysomnography. Fourth, because this study was conducted in an in-laboratory attended situation, the results must be interpreted with caution in the follow-up clinic, where data are obtained in the unattended situation (i.e., at patients' home).
In conclusion, a strong correlation between the AHI-PSG and the AHI-S8 was observed. However, the correlation was weakened when the analysis was limited to the HI. Furthermore, independent factors associated with a discrepancy in the AHI scored by the 2 methods were AHI during the diagnostic study and leakage during CPAP titration. These results provide an impetus for further investigation in the arena of auto-CPAP devices and very important information for clinicians.
DISCLOSURE STATEMENT
This was not an industry supported study. The authors have indicated no financial conflicts of interest.
ABBREVIATIONS
- OSAHS
obstructive sleep apnea-hypopnea syndrome
- CPAP
continuous positive airway pressure
- AHI
apnea-hypopnea index
- AHI-PSG
apnea-hypopnea index determined by manually scored polysomnography
- AI-PSG
apnea index determined by manually scored polysomnography
- HI-PSG
hypopnea index determined by manually scored polysomnography
- AHI-S8
apnea-hypopnea index determined by the S8 auto-CPAP device
- AI-S8
apnea index determined by the S8 auto-CPAP device
- HI-S8
hypopnea index determined by the S8 auto-CPAP device
REFERENCES
- 1.Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med. 1993;328:1230–1235. doi: 10.1056/NEJM199304293281704. [DOI] [PubMed] [Google Scholar]
- 2.He J, Kryger MH, Zorick FJ, Conway W, Roth T. Mortality and apnea index in obstructive sleep apnea: experience in 385 male patients. Chest. 1988;94:9–14. [PubMed] [Google Scholar]
- 3.Shahar E, Whitney CW, Redline S, et al. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med. 2001;163:19–25. doi: 10.1164/ajrccm.163.1.2001008. [DOI] [PubMed] [Google Scholar]
- 4.Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversal of obstructive sleep apnea by continuous positive airway pressure applied through the nares. Lancet. 1981;i:862–865. doi: 10.1016/s0140-6736(81)92140-1. [DOI] [PubMed] [Google Scholar]
- 5.Engleman HM, Martin SE, Deary IJ, Douglas NJ. Effect of continuous positive airway treatment on daytime function in sleep apnoea/hypopnoea syndrome. Lancet. 1994;343:572–575. doi: 10.1016/s0140-6736(94)91522-9. [DOI] [PubMed] [Google Scholar]
- 6.Meurice JC, Paquereau J, Neau JP, et al. Long-term evolution of daytime somnolence in patients with sleep apnea/hypopnea syndrome treated by continuous positive airway pressure. Sleep. 1997;20:1162–1166. [PubMed] [Google Scholar]
- 7.Kawahara S, Akashiba T, Akahoshi T Horie T. Nasal CPAP improves the quality of life and lessens the depressive symptoms in patients with obstructive sleep apnea syndrome. Intern Med. 2005;44:422–427. doi: 10.2169/internalmedicine.44.422. [DOI] [PubMed] [Google Scholar]
- 8.Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365:1046–1053. doi: 10.1016/S0140-6736(05)71141-7. [DOI] [PubMed] [Google Scholar]
- 9.Gugger M. Comparison of ResMed AutoSet (version 3.03) with polysomnography in the diagnosis of the sleep apnoea/hypopnoea syndrome. Eur Respir J. 1997;10:587–591. [PubMed] [Google Scholar]
- 10.American Academy of Sleep Medicine Task Force. Sleep related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. The Report of an American Academy of Sleep Medicine Task Force. Sleep. 1999;22:667–689. [PubMed] [Google Scholar]
- 11.Rechtshaffen A, Kales A, editors. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Los Angeles, CA: UCLA Brain Information Service/Brain Research Institute; 1968. [Google Scholar]
- 12.Standards of Practice Committee Task Force. Practice parameters for the indications for polysomnography and related procedures. Sleep. 1997;20:406–422. [PubMed] [Google Scholar]
- 13.Littner M, Hirshkowitz M, Davila D, et al. Practice parameters for the use of auto-titrating continuous positive airway pressure devices for titrating pressures and treating adult patients with obstructive sleep apnea syndrome. A report of the American Academy of Sleep Medicine. Sleep. 2002;25:143–147. doi: 10.1093/sleep/25.2.143. [DOI] [PubMed] [Google Scholar]
- 14.Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;i:307–310. [PubMed] [Google Scholar]
- 15.Gugger M, Mathis J, Bassetti C. Accuracy of an intelligent CPAP machine with in-built diagnostic abilities in detecting apnoeas: a comparison with polysomnography. Thorax. 1995;50:1199–1201. doi: 10.1136/thx.50.11.1199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bradley P, Mortimore IL, Douglas NJ. Comparison of polysomnography with ResCare Autoset in the diagnosis of the sleep apnoea/hypopnoea syndrome. Thorax. 1995;50:1201–1203. doi: 10.1136/thx.50.11.1201. [DOI] [PMC free article] [PubMed] [Google Scholar]