Eucapnia and Hypercapnia in Patients with Chronic Airflow Limitation The Role of the Upper Airway1.2
C. SHU CHAN, PETER T. P. BYE, ANN J. WOOLCOCK, and COLIN E. SULLIVAN
Introduction
The pathogenesis of hypercapnia in patients with chronic airflow limitation (CAL) has remained elusive. Historically, clinicians have recognized two extreme forms of this chronic disease: the "pink puffer" and the "blue bloater" (1-3). Clinically, the "pink puffer" is recognized by severe dyspnea, hyperinflation, mild hypoxemia, and normocapnia. In contrast, the "blue bloater" is characterized by chronic productive cough, mild dyspnea, cor pulmonale, excess weight, polycythemia, moderate hypoxemia, and hypercapnia. However, this clinical classification is imprecise, and many patients cannot be classified clearly into either category. Furthermore, the coexistence of the obstructive sleep apnea (OSA) syndrome, a common condition in patients with CAL, is likely to have been present in many patients in previous studies. The recent advances in the understanding of obstructive sleep apnea provide important clues concerning the mechanism of hypercapnia in patients with CAL. In particular, the importance of snoring (4), alcohol consumption (5), and upper airway size (6, 7) in patients with obstructive sleep apnea is well recognized. However, the interaction of these factors with CAL is unknown. In this study, we avoided the classification of "pink puffer" and "blue bloater" and instead identified two groups of patients with CAL according to their stable awake Pac0 2 : eucapnic (Pac0 2 < 45 mm Hg) and hypercapnic (Pac0 2 > 45 mm Hg), Patients with polysomnographic evidence of obstructive sleep apnea were excluded. Using these strict criteria, we compared the eucapnic and hypercapnic patients. We examined a range of potentially important factors that may be involved in the development of chronic hypercapnia.
SUMMARY In this study, we examined two groups of patients with chronic airflow limitation (CAL) separated according to their awake, stable arterial CO2 level. The aim was to Identify factors that may contribute to the development of chronic hypercapnic respiratory failure. Patients with obstructive sleep apnea were excluded from the study. Detailed lifetime histories of smoking, alcohol, and snoring were obtained from all patients together with measurements of lung function and of upper airway size. Thirty-three patients with FEV, < 1.5 L were studied, of whom 19 were eucapnic and 14 were hypercapnic. Both groups had a similar degree of chronic airflow limitation and similar lung volumes and OLeo. The hypercapnic group had more hypopneas and desaturated more severely during sleep. The greatest differences between the groups were in their alcohol consumptions, snoring histories, and upper airway dimensions. The eucapnic patients were characterized by lower lifetime alcohol intake, minimal snoring, and large upper airway size. In contrast, the hypercapnic patients were characterized by excessive lifetime alcohol consumption, habitual snoring over many years, and a small upper airway size. Our findings suggest that chronic, heavy alcohol use and upper airway dysfunction are important factors in the development of hypercapnic respiraAM REV RESPIR DIS 1990; 141:861-865 tory failure.
Methods Patient Selection We studied 33 male patients with CAL and FEY 1 less than 1.5 L. All patients had one night of acclimatization in the laboratory before undergoing full all-night sleep studies, and those with evidence of obstructive sleep apnea were excluded. All patients were clinically stable and had not received long-term oxygentherapy at the time of the study. Week1y awake arterial blood gas tensions weremeasured in these patients over a 4-wk period to confirm their clinical stability. There was no overlap in the Pac0 2 of the two groups.
Snoring, Smoking, and Alcohol History Detailed lifetime smoking and alcohol history were obtained from all patients using a structured interview schedule that had previously been used in the study of alcoholic liver disease (8). This technique involved helping the patients to remember their past smoking and alcohol consumptions by using key events in their lives, e.g., first and subsequent jobs, service in the armed forces, arrival of first child, year of emigration, etc. Each patient was interviewed on two separate occasions at least a month apart to check the reproducibility of the history. Snoring history was obtained from both the patients and their sleepingpartners. UtiIiz-
ing the same interview technique as outlined in the last paragraph, we were able to obtain a detailed history of snoring pattern and frequency over the years. Patients were classified as nonsnorers if they snored infrequently, occasional snorers if they snored three or less nights a week, and as habitual snorers if they snored four or more nights a week.
Awake Measurements On the day of or immediately after the sleep study, the patients had measurements of spirometry, lung volumes, diffusing capacity, and arterial blood gas tensions. The lung volumes were measured by the single-breath helium dilution technique, and the diffusing capacity was measured by the carbon monoxide single-breath technique. Arterial blood gas tensions were measured using a Corning
(Received in original form October 6, 1988 and in revised form August 21, 1989) 1 From the Department of Thoracic Medicine and the Sleep Unit, Royal Prince Alfred Hospital and the University of Sydney, Sydney, New South Wales. 2 Correspondence and requests for reprints should be addressed to Associate Professor C. E. Sullivan, Department of Medicine, University of Sydney, Sydney, N.S.w. 2006, Australia.
861
862
175blood gas analyzer (Corning Glass, Medfield, MA). We calculated the body mass index (BMI) from each patient's height and weight. BMI is defined as the weight in kilograms divided by the square of the height in meters. This measure provides an index of the degree of obesity. The normal range is between 20 to 24, and above 25 is considered overweight.Obesity is defined as a valuegreater than 30. The upper airway was assessed both clinically and with CT scan. The patients were scanned in the awake supine position with the neck placed in a neutral position and scanned at 8-mm intervals during quiet tidal breathing. Images were obtained from the superior aspect of the pharynx down to the glottic area. Each cross-sectionalarea was then digitized from the scanner's computer. The space posterior to the hard palate down to the tip of the uvula was defined as the nasopharynx. The narrowest segment in the upper airway of these patients was in the nasopharyngeal space, and the area of this segment was used as a measure of the upper airway dimension.
Sleep Studies The sleep state was assessed with two EEG records (C3/A2, C4/Al), a postural (submental) EMG record and two ocular movement (EOG) records. ECG and heart rate were recorded continuously. The respiratory variables recorded were oxyhemoglobin saturation (HP-47201A; Hewlett-Parkard, Waltham, MA), chest wall and abdominal movement were measured by respiratory inductive plethysmograph (Respitrace'"; Ambulatory Monitoring Inc., Ardsley, NY). Oronasal airflow was measured with a Venturi mask attached to a pressure transducer (Statham PM 131 TC transducer; Gould-Statham Instruments, Oxnard, CAl. All variables were recorded with a Grass Model 78-16 channel EEG polygraph system (Grass Instruments, Quincy, MA). Snoring was recorded qualitatively with a sound recorder. Apnea was defined as cessation of airflow for at least 10 s. Hypopnea was defined as a reduction in airflow amplitude by greater than 50070 for at least 10 s. Obstructive sleep apnea was defined as an apnea index greater than 5 per hour in REM or in NREM sleep.
CHAN, BYE, WOOLCOCK, AND SULLIVAN
TABLE 1 ANTHROPOMETRIC AND LUNG FUNCTION DATA Eucapnic Number Age, yr Height, m Weight, kg BMI Smoking, pack-years pH Pao" mm Hg Paco" mm Hg AaPo" mm Hg Hemoglobin, gldl FEV" L FEV,IFVC, % TLC, L TLC, % pred FRC, L FRC, % pred RV, L RV, % pred OLco, % pred
Hypercapnic
19 63 171 68 23 42 7.39 78 38 25 14.8 0.92 33 7.18 120 5.01 143 3.96 177 42
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
p Value
14 3 2 3 1 4 0.01 2 1 2 0.5 0.05 2 0.29 3 0.25 5 0.19 8 5
59 173 85 28 67 7.37 63 50 25 16.4 0.90 36 7.20 116 5.04 141 4.47 197 42
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
2 1 4 1 11 0.01 3 1 2 0.6 0.08 3 0.37 5 0.32 9 0.32 15 4
NS NS
< 0.01 < 0.001 NS NS
< 0.001 < 0.001 NS
< 0.05 NS NS NS NS NS NS NS NS NS
Definition of abbreviations: BMI = body mass index; NS = not signiticant.
TABLE 2 SLEEP DATA Eucapnic Hypopnea index (NREM), nih Hypopnea index (REM), nih Mean awake Sao" % Mean lowest NREM Sao" % Mean lowest REM Sao,. %
20 16 94 91 87
any evidence of obstructive sleep apnea, hypopneas werecommon in both groups, but they were significantly higher in the hypercapnic group than in the eucapnic group during NREM sleep, as shown in table 2. The mean hypopnea index in the eucapnic patients during NREM sleep was similar to that of the "pink puffers" reported by Catterall and coworkers (10). During REM sleep, the hypopnea index was similar in both the eucapnic and hypercapnic groups of patients. During the sleep study night, snoring was obStatistics served in more than half of the hyperAll measured variables were expressed as capnic patients, whereas only four of the mean ± SEM. Comparison between the two eucapnic patients snored (lightly). Both groups was performed by statistical analysis groups had similar degree of airflow limiusing an unpaired t test. For nonparametric tation, hyperinflation, gas trapping, difdata (alcohol and cigarette consumption), the Mann-Whitney U test was used. The reliabil- fusion limitation, and widened alveolarity coefficient for the alcohol and cigarette arterial oxygen difference. However, the consumptions were estimated by Pearson's resting awake Pao.Ievel revealed that the product-moment coefficient of correlation (9). hypercapnic patients were more hypoxemic than were the eucapnic patients, Results thus confirming the importance of The anthropometric and lung function chronic alveolar hypoventilation as a data are shown in table 1. Of 33 patients, cause of the chronic hypoxemia. 19 were eucapnic and 14 were hypercapBoth the weight and the BMI of the nic. Although none of these patients had hypercapnic group were significantly
± ± ± ± ±
9 5 0 1 1
Hypercapnic
66 34 91 82 76
± ± ± ± ±
13 15 1 3 3
p Value
< 0.01 NS
< 0.05 < 0.01
< 0.001
higher than those of the eucapnic group. However, within the hypercapnic group, only four patients werein the obese range. The degree of oxyhemoglobin desaturation during sleep was similar to previous studies (10, 11); table 2 shows that the hypercapnic group desaturated much more than the eucapnic group, and that within each group the greatest desaturation was observed during REM sleep. The reliability coefficients for the alcohol and cigarette history are shown in table 3. The test-retest reliability coefficient for the alcohol history in this group of patients ranged from 0.95 to 1.00. These results compare very favorably
TABLE 3 RELIABILITY OF SMOKING AND ALCOHOL HISTORY Reliability Coefficient Lifetime alcohol consumption Daily alcohol intake Duration of alcohol consumption Lifetime cigarette consumption
0.98 1 .00 0.98 0.95
863
COMPARISON OF EUCAPNIC AND HYPERCAPNIC CHRONIC AIRFLOW LIMITATION
eo 50
0..
~I : O~ ~
!!;i"
100
ecnou_
CJ EUCAPNIA ~
ISI_Dolly~,,*tho
L..aM.F"Iwy.,.
0..
It:
50
~
Gms/day
:
~~
20 10
HYPERCAPNIA
75
,a ft..:J o '--J.....l'""--'--:-'.:>o.L.....L,,-J~-----_. ,-eo 51-100 eo
~~ l:5 ~
CJ Avo. Dally U!W1lmo
:::> 0 Cl u, 0
50
25
0'--~>.J.....1..:-'::-'--'::-''':::'--'='=~~--'-:=
Gms/day
•
O'--....L.._......._ --'--_.L---'--Non-snarer
Occasslonal
HGbltual
Snorer
Snorer
Fig. 1. Average daily alcohol consumption of eucapnic (upper panel) and hypercapnic (lower panel) patients. The asterisk represents no patient in these categories.
Fig. 4. Snoring pattern obtained from the history of eucapnic and hypercapnic patients. The asterisk indicates that there were no patients in these categories.
with figures obtained from Saunders and associates (8), and Skinner and Sheu (12), which ranged from 0.81 to 0.92 and 0.85 to 0.95, respectively. The reliability coefficient for the cigarette consumption was also high when the same interview technique was used. The alcohol consumption of the two groups is shown in figures 1 to 3. About half the eucapnic patients were nondrinkers and the rest were mild drinkers. There was no difference in the drinking pattern during the last five years of their
lives. In contrast, many of the hypercapnic patients were moderate to very heavy drinkers. Interestingly, heavy drinkers had either reduced their drinking to the mild-to-moderate range or had ceased completely in the last 5 yr of their lives. In terms of lifetime consumption, figure 2 shows a clear separation between these two groups of patients, with almost no overlap. The mean duration of alcohol consumption in the eucapnic group was 34 yr, and in the hypercapnic group it was 31 yr. The cigarette consumption was higher in the hypercapnic group than in the eucapnic group (table 1), but this did not reach statistical significance. The mean duration of consumption was similar in both groups; 38 ± 3 yr in the eucapnic group and 36 ± 3 yr in the hypercapnic
60
50 0..
:::>
~
...
30
~
20
Cl
CJ EUCAPNIA fZ2J HYPERCAPNIA
40
group. At the time of their sleep study, 18 of the 19 eucapnic patients had given up smoking, whereas in the hypercapnic group six of the 14 patients were still smoking. It can be seen in figure 4 that the snoring history was strikingly different between the eucapnic and hypercapnic patients. All hypercapnic patients had a history of habitual snoring, whereas in the eucapnic patients 53070 were nonsnorers, 37% were occasional snorers, and 10% were habitual snorers. . Clinical assessment of the upper airway revealed a marked difference between the eucapnic and the hypercapnic patients. The eucapnic patients had wide oropharyngeal dimensions and no redness or edema of the tonsillar folds and soft palate. In contrast, the hypercapnic patients showed changes typical of chronic snoring. These changes were narrow oropharynx and redness and edema of the tonsillar folds and soft palate. Examples of the awake measurement of upper airway dimensions as assessed by CT scanning are given in figures 5 and 6. A typical eucapnic patient with wide upper airway is shown in figure 5, and, in contrast, a typical hypercapnic patient with small upper airways is shown in figure 6. The site and degree of upper airway narrowing in these hypercapnic patients with CAL were very similar to those initially observed in patients with obstructive sleep apnea (6, 7). Furthermore, the upper airway dimensions of the eucapnic patients with CAL were larger
0
10 0
Kgm Fig. 2. Cumulative lifetime alcohol consumption of eucapnic and hypercapnic patients. The asterisk indicates that there were no patients in these categories.
E
o
~ c
600
E
500
~ ~
"
400
0 s: 0 c
300
e 0
o
..
~
'0 0
"'"
E
"
~
200 100
a
o 000 ·0l:ll:l8880000 Eucapnia
Hypercapnia
Fig. 3. Average daily alcohol consumption in 34 patients with CAL (open circles). The closed circles indicate mean ± SEM. Eucapnia = 17 ± 6g/day, hypercapnia = 177 ± 46 g/day.
Fig. 5. CT scan of a eucapnic patient. Cross-sectional view through the retropalatal space (indicated by the marker). The upper airway is wide with a cross-sectional area of 6.60 em'.
864
CHAN, BYE, WOOLCOCK, AND SULLIVAN
Fig. 6. CT scan of a hypercapnic patient. Cross-sectional view through the retropalatal space (indicated by the marker). The upper airway is smaller than in figure 4 (cross-sectional area = 0.77 em').
and comparable in size to the nonapneic group reported by Haponik and coworkers (6). The individual and mean upper airway dimensions of the eucapnic and hypercapnic group are shown in figure 7, and figure 8 shows the relationship between the BMI and the upper airway dimension. The correlation coefficient was 0.46 at p < 0.01. Discussion
In this study of 33 patients with severe CAL without obstructive sleep apnea, we have identified important differences between the eucapnic and the hypercapnic patients. The sleep studies revealed patterns of abnormality very similar to those described by Catterall and coworkers (10). Thus, it is unlikely that there is any
c:E
5
fundamental difference between our patients with CAL and those reported previously. The hypercapnic patients were predominantly heavy snorers in their detailed life history. They had a long history of excessivealcohol consumption and had anatomically narrow upper airways on CT scan. In contrast, the eucapnic patients were characterized by an absence of snoring in their detailed life history and were essentially lifelong very light or nondrinkers. Their awake upper airway dimensions werelarger, and in some, the dimensions were "super-normal." There wereessentially no differences in the lung function of these two groups. These findings suggest that upper airway dysfunction and heavy alcohol consumption are important factors in the pathogenesis of chronic hypercapnia. Snoring is a sign of partial upper air-
0
~
C. SHU CHAN, PETER T. P. BYE, ANN J. WOOLCOCK, and COLIN E. SULLIVAN
Introduction
The pathogenesis of hypercapnia in patients with chronic airflow limitation (CAL) has remained elusive. Historically, clinicians have recognized two extreme forms of this chronic disease: the "pink puffer" and the "blue bloater" (1-3). Clinically, the "pink puffer" is recognized by severe dyspnea, hyperinflation, mild hypoxemia, and normocapnia. In contrast, the "blue bloater" is characterized by chronic productive cough, mild dyspnea, cor pulmonale, excess weight, polycythemia, moderate hypoxemia, and hypercapnia. However, this clinical classification is imprecise, and many patients cannot be classified clearly into either category. Furthermore, the coexistence of the obstructive sleep apnea (OSA) syndrome, a common condition in patients with CAL, is likely to have been present in many patients in previous studies. The recent advances in the understanding of obstructive sleep apnea provide important clues concerning the mechanism of hypercapnia in patients with CAL. In particular, the importance of snoring (4), alcohol consumption (5), and upper airway size (6, 7) in patients with obstructive sleep apnea is well recognized. However, the interaction of these factors with CAL is unknown. In this study, we avoided the classification of "pink puffer" and "blue bloater" and instead identified two groups of patients with CAL according to their stable awake Pac0 2 : eucapnic (Pac0 2 < 45 mm Hg) and hypercapnic (Pac0 2 > 45 mm Hg), Patients with polysomnographic evidence of obstructive sleep apnea were excluded. Using these strict criteria, we compared the eucapnic and hypercapnic patients. We examined a range of potentially important factors that may be involved in the development of chronic hypercapnia.
SUMMARY In this study, we examined two groups of patients with chronic airflow limitation (CAL) separated according to their awake, stable arterial CO2 level. The aim was to Identify factors that may contribute to the development of chronic hypercapnic respiratory failure. Patients with obstructive sleep apnea were excluded from the study. Detailed lifetime histories of smoking, alcohol, and snoring were obtained from all patients together with measurements of lung function and of upper airway size. Thirty-three patients with FEV, < 1.5 L were studied, of whom 19 were eucapnic and 14 were hypercapnic. Both groups had a similar degree of chronic airflow limitation and similar lung volumes and OLeo. The hypercapnic group had more hypopneas and desaturated more severely during sleep. The greatest differences between the groups were in their alcohol consumptions, snoring histories, and upper airway dimensions. The eucapnic patients were characterized by lower lifetime alcohol intake, minimal snoring, and large upper airway size. In contrast, the hypercapnic patients were characterized by excessive lifetime alcohol consumption, habitual snoring over many years, and a small upper airway size. Our findings suggest that chronic, heavy alcohol use and upper airway dysfunction are important factors in the development of hypercapnic respiraAM REV RESPIR DIS 1990; 141:861-865 tory failure.
Methods Patient Selection We studied 33 male patients with CAL and FEY 1 less than 1.5 L. All patients had one night of acclimatization in the laboratory before undergoing full all-night sleep studies, and those with evidence of obstructive sleep apnea were excluded. All patients were clinically stable and had not received long-term oxygentherapy at the time of the study. Week1y awake arterial blood gas tensions weremeasured in these patients over a 4-wk period to confirm their clinical stability. There was no overlap in the Pac0 2 of the two groups.
Snoring, Smoking, and Alcohol History Detailed lifetime smoking and alcohol history were obtained from all patients using a structured interview schedule that had previously been used in the study of alcoholic liver disease (8). This technique involved helping the patients to remember their past smoking and alcohol consumptions by using key events in their lives, e.g., first and subsequent jobs, service in the armed forces, arrival of first child, year of emigration, etc. Each patient was interviewed on two separate occasions at least a month apart to check the reproducibility of the history. Snoring history was obtained from both the patients and their sleepingpartners. UtiIiz-
ing the same interview technique as outlined in the last paragraph, we were able to obtain a detailed history of snoring pattern and frequency over the years. Patients were classified as nonsnorers if they snored infrequently, occasional snorers if they snored three or less nights a week, and as habitual snorers if they snored four or more nights a week.
Awake Measurements On the day of or immediately after the sleep study, the patients had measurements of spirometry, lung volumes, diffusing capacity, and arterial blood gas tensions. The lung volumes were measured by the single-breath helium dilution technique, and the diffusing capacity was measured by the carbon monoxide single-breath technique. Arterial blood gas tensions were measured using a Corning
(Received in original form October 6, 1988 and in revised form August 21, 1989) 1 From the Department of Thoracic Medicine and the Sleep Unit, Royal Prince Alfred Hospital and the University of Sydney, Sydney, New South Wales. 2 Correspondence and requests for reprints should be addressed to Associate Professor C. E. Sullivan, Department of Medicine, University of Sydney, Sydney, N.S.w. 2006, Australia.
861
862
175blood gas analyzer (Corning Glass, Medfield, MA). We calculated the body mass index (BMI) from each patient's height and weight. BMI is defined as the weight in kilograms divided by the square of the height in meters. This measure provides an index of the degree of obesity. The normal range is between 20 to 24, and above 25 is considered overweight.Obesity is defined as a valuegreater than 30. The upper airway was assessed both clinically and with CT scan. The patients were scanned in the awake supine position with the neck placed in a neutral position and scanned at 8-mm intervals during quiet tidal breathing. Images were obtained from the superior aspect of the pharynx down to the glottic area. Each cross-sectionalarea was then digitized from the scanner's computer. The space posterior to the hard palate down to the tip of the uvula was defined as the nasopharynx. The narrowest segment in the upper airway of these patients was in the nasopharyngeal space, and the area of this segment was used as a measure of the upper airway dimension.
Sleep Studies The sleep state was assessed with two EEG records (C3/A2, C4/Al), a postural (submental) EMG record and two ocular movement (EOG) records. ECG and heart rate were recorded continuously. The respiratory variables recorded were oxyhemoglobin saturation (HP-47201A; Hewlett-Parkard, Waltham, MA), chest wall and abdominal movement were measured by respiratory inductive plethysmograph (Respitrace'"; Ambulatory Monitoring Inc., Ardsley, NY). Oronasal airflow was measured with a Venturi mask attached to a pressure transducer (Statham PM 131 TC transducer; Gould-Statham Instruments, Oxnard, CAl. All variables were recorded with a Grass Model 78-16 channel EEG polygraph system (Grass Instruments, Quincy, MA). Snoring was recorded qualitatively with a sound recorder. Apnea was defined as cessation of airflow for at least 10 s. Hypopnea was defined as a reduction in airflow amplitude by greater than 50070 for at least 10 s. Obstructive sleep apnea was defined as an apnea index greater than 5 per hour in REM or in NREM sleep.
CHAN, BYE, WOOLCOCK, AND SULLIVAN
TABLE 1 ANTHROPOMETRIC AND LUNG FUNCTION DATA Eucapnic Number Age, yr Height, m Weight, kg BMI Smoking, pack-years pH Pao" mm Hg Paco" mm Hg AaPo" mm Hg Hemoglobin, gldl FEV" L FEV,IFVC, % TLC, L TLC, % pred FRC, L FRC, % pred RV, L RV, % pred OLco, % pred
Hypercapnic
19 63 171 68 23 42 7.39 78 38 25 14.8 0.92 33 7.18 120 5.01 143 3.96 177 42
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
p Value
14 3 2 3 1 4 0.01 2 1 2 0.5 0.05 2 0.29 3 0.25 5 0.19 8 5
59 173 85 28 67 7.37 63 50 25 16.4 0.90 36 7.20 116 5.04 141 4.47 197 42
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
2 1 4 1 11 0.01 3 1 2 0.6 0.08 3 0.37 5 0.32 9 0.32 15 4
NS NS
< 0.01 < 0.001 NS NS
< 0.001 < 0.001 NS
< 0.05 NS NS NS NS NS NS NS NS NS
Definition of abbreviations: BMI = body mass index; NS = not signiticant.
TABLE 2 SLEEP DATA Eucapnic Hypopnea index (NREM), nih Hypopnea index (REM), nih Mean awake Sao" % Mean lowest NREM Sao" % Mean lowest REM Sao,. %
20 16 94 91 87
any evidence of obstructive sleep apnea, hypopneas werecommon in both groups, but they were significantly higher in the hypercapnic group than in the eucapnic group during NREM sleep, as shown in table 2. The mean hypopnea index in the eucapnic patients during NREM sleep was similar to that of the "pink puffers" reported by Catterall and coworkers (10). During REM sleep, the hypopnea index was similar in both the eucapnic and hypercapnic groups of patients. During the sleep study night, snoring was obStatistics served in more than half of the hyperAll measured variables were expressed as capnic patients, whereas only four of the mean ± SEM. Comparison between the two eucapnic patients snored (lightly). Both groups was performed by statistical analysis groups had similar degree of airflow limiusing an unpaired t test. For nonparametric tation, hyperinflation, gas trapping, difdata (alcohol and cigarette consumption), the Mann-Whitney U test was used. The reliabil- fusion limitation, and widened alveolarity coefficient for the alcohol and cigarette arterial oxygen difference. However, the consumptions were estimated by Pearson's resting awake Pao.Ievel revealed that the product-moment coefficient of correlation (9). hypercapnic patients were more hypoxemic than were the eucapnic patients, Results thus confirming the importance of The anthropometric and lung function chronic alveolar hypoventilation as a data are shown in table 1. Of 33 patients, cause of the chronic hypoxemia. 19 were eucapnic and 14 were hypercapBoth the weight and the BMI of the nic. Although none of these patients had hypercapnic group were significantly
± ± ± ± ±
9 5 0 1 1
Hypercapnic
66 34 91 82 76
± ± ± ± ±
13 15 1 3 3
p Value
< 0.01 NS
< 0.05 < 0.01
< 0.001
higher than those of the eucapnic group. However, within the hypercapnic group, only four patients werein the obese range. The degree of oxyhemoglobin desaturation during sleep was similar to previous studies (10, 11); table 2 shows that the hypercapnic group desaturated much more than the eucapnic group, and that within each group the greatest desaturation was observed during REM sleep. The reliability coefficients for the alcohol and cigarette history are shown in table 3. The test-retest reliability coefficient for the alcohol history in this group of patients ranged from 0.95 to 1.00. These results compare very favorably
TABLE 3 RELIABILITY OF SMOKING AND ALCOHOL HISTORY Reliability Coefficient Lifetime alcohol consumption Daily alcohol intake Duration of alcohol consumption Lifetime cigarette consumption
0.98 1 .00 0.98 0.95
863
COMPARISON OF EUCAPNIC AND HYPERCAPNIC CHRONIC AIRFLOW LIMITATION
eo 50
0..
~I : O~ ~
!!;i"
100
ecnou_
CJ EUCAPNIA ~
ISI_Dolly~,,*tho
L..aM.F"Iwy.,.
0..
It:
50
~
Gms/day
:
~~
20 10
HYPERCAPNIA
75
,a ft..:J o '--J.....l'""--'--:-'.:>o.L.....L,,-J~-----_. ,-eo 51-100 eo
~~ l:5 ~
CJ Avo. Dally U!W1lmo
:::> 0 Cl u, 0
50
25
0'--~>.J.....1..:-'::-'--'::-''':::'--'='=~~--'-:=
Gms/day
•
O'--....L.._......._ --'--_.L---'--Non-snarer
Occasslonal
HGbltual
Snorer
Snorer
Fig. 1. Average daily alcohol consumption of eucapnic (upper panel) and hypercapnic (lower panel) patients. The asterisk represents no patient in these categories.
Fig. 4. Snoring pattern obtained from the history of eucapnic and hypercapnic patients. The asterisk indicates that there were no patients in these categories.
with figures obtained from Saunders and associates (8), and Skinner and Sheu (12), which ranged from 0.81 to 0.92 and 0.85 to 0.95, respectively. The reliability coefficient for the cigarette consumption was also high when the same interview technique was used. The alcohol consumption of the two groups is shown in figures 1 to 3. About half the eucapnic patients were nondrinkers and the rest were mild drinkers. There was no difference in the drinking pattern during the last five years of their
lives. In contrast, many of the hypercapnic patients were moderate to very heavy drinkers. Interestingly, heavy drinkers had either reduced their drinking to the mild-to-moderate range or had ceased completely in the last 5 yr of their lives. In terms of lifetime consumption, figure 2 shows a clear separation between these two groups of patients, with almost no overlap. The mean duration of alcohol consumption in the eucapnic group was 34 yr, and in the hypercapnic group it was 31 yr. The cigarette consumption was higher in the hypercapnic group than in the eucapnic group (table 1), but this did not reach statistical significance. The mean duration of consumption was similar in both groups; 38 ± 3 yr in the eucapnic group and 36 ± 3 yr in the hypercapnic
60
50 0..
:::>
~
...
30
~
20
Cl
CJ EUCAPNIA fZ2J HYPERCAPNIA
40
group. At the time of their sleep study, 18 of the 19 eucapnic patients had given up smoking, whereas in the hypercapnic group six of the 14 patients were still smoking. It can be seen in figure 4 that the snoring history was strikingly different between the eucapnic and hypercapnic patients. All hypercapnic patients had a history of habitual snoring, whereas in the eucapnic patients 53070 were nonsnorers, 37% were occasional snorers, and 10% were habitual snorers. . Clinical assessment of the upper airway revealed a marked difference between the eucapnic and the hypercapnic patients. The eucapnic patients had wide oropharyngeal dimensions and no redness or edema of the tonsillar folds and soft palate. In contrast, the hypercapnic patients showed changes typical of chronic snoring. These changes were narrow oropharynx and redness and edema of the tonsillar folds and soft palate. Examples of the awake measurement of upper airway dimensions as assessed by CT scanning are given in figures 5 and 6. A typical eucapnic patient with wide upper airway is shown in figure 5, and, in contrast, a typical hypercapnic patient with small upper airways is shown in figure 6. The site and degree of upper airway narrowing in these hypercapnic patients with CAL were very similar to those initially observed in patients with obstructive sleep apnea (6, 7). Furthermore, the upper airway dimensions of the eucapnic patients with CAL were larger
0
10 0
Kgm Fig. 2. Cumulative lifetime alcohol consumption of eucapnic and hypercapnic patients. The asterisk indicates that there were no patients in these categories.
E
o
~ c
600
E
500
~ ~
"
400
0 s: 0 c
300
e 0
o
..
~
'0 0
"'"
E
"
~
200 100
a
o 000 ·0l:ll:l8880000 Eucapnia
Hypercapnia
Fig. 3. Average daily alcohol consumption in 34 patients with CAL (open circles). The closed circles indicate mean ± SEM. Eucapnia = 17 ± 6g/day, hypercapnia = 177 ± 46 g/day.
Fig. 5. CT scan of a eucapnic patient. Cross-sectional view through the retropalatal space (indicated by the marker). The upper airway is wide with a cross-sectional area of 6.60 em'.
864
CHAN, BYE, WOOLCOCK, AND SULLIVAN
Fig. 6. CT scan of a hypercapnic patient. Cross-sectional view through the retropalatal space (indicated by the marker). The upper airway is smaller than in figure 4 (cross-sectional area = 0.77 em').
and comparable in size to the nonapneic group reported by Haponik and coworkers (6). The individual and mean upper airway dimensions of the eucapnic and hypercapnic group are shown in figure 7, and figure 8 shows the relationship between the BMI and the upper airway dimension. The correlation coefficient was 0.46 at p < 0.01. Discussion
In this study of 33 patients with severe CAL without obstructive sleep apnea, we have identified important differences between the eucapnic and the hypercapnic patients. The sleep studies revealed patterns of abnormality very similar to those described by Catterall and coworkers (10). Thus, it is unlikely that there is any
c:E
5
fundamental difference between our patients with CAL and those reported previously. The hypercapnic patients were predominantly heavy snorers in their detailed life history. They had a long history of excessivealcohol consumption and had anatomically narrow upper airways on CT scan. In contrast, the eucapnic patients were characterized by an absence of snoring in their detailed life history and were essentially lifelong very light or nondrinkers. Their awake upper airway dimensions werelarger, and in some, the dimensions were "super-normal." There wereessentially no differences in the lung function of these two groups. These findings suggest that upper airway dysfunction and heavy alcohol consumption are important factors in the pathogenesis of chronic hypercapnia. Snoring is a sign of partial upper air-
0
~
