- Email: [email protected]
Transtracheal Oxygen, Nasal CPAP and Nasal Oxygen in Five Patients with Obstructive Sleep Apnea
Transtracheal Oxygen, Nasal CPAP and Nasal Oxygen in Five Patients with Obstructive Sleep Apnea*
Robert J Farney, M.D., F.C.C.R; James M. Walker; Ph.D.; Jeffrey C. Elmer; M.D., F.C.C.R; Vincent A. VIScomi, M.D., F.C.C.R; and R. ]on Ord, M.D. The effect of transtracheal oxygen administration by means of a 9-French (2.7 mm) percutaneous catheter was assessed in 6ve patients with severe obstructive sleep apnea. We hypothesized that the delivery of oxygen below the site of airway obstruction should reduce the arterial oxygen desaturation during apneas and hypopneas, thereby increasing respiratory stability. Standard sleep and respiratory measurements were recorded in these subjects with a11night polysomoography on nonconsecutive nights during four experimental conditions: room air (BL), nasal continuous positive airway pressure (CPAP), nasal 0. (NC 0.), and transtracheal O. (TT Os). In three of these subjects, room air was infused (TT RA) at Bow rates comparable to TT 0 •. Compared with baseline room air measurements, TT 0. DOt only signi6cantly increased the SaO. nadir from 70.4 percent to 89.7 percent (p
sleep apnea/hypopnea (13.81h sleep) but did not abolish desaturations when apneas occurred (mean SaO. nadir, 80.0 percent). Compared with oxygen, transtracheal infusion of room air appeared to be somewhat effective; however, the small number of studies with TT RA precluded statistical analysis. We believe that TT 0. is superior to NC 0. for some patients with obstructive sleep apnea because continuous oxygen Bow below the site of airway obstruction more reliably prevents alveolar hypoxia and respiration is stabilized. Infusion of air or oxygen through the tracheal catheter Bow may also increase mean airway pressure and reduce obstructive apnea similar to nasal CPAP. We CODelude that TT 0. may be an effective alternative mode of therapy for some patients with severe sleep apnealhypopDea when nasal CPAP is not tolerated or when combined oxygen and nasal CPAP are required. (Cheat 1992; 101:1228-35)
purpose of this study was to evaluate the efficacy T ofhe transtracheal oxygen (TT OJ in patients with
clear that oxygen plus CPAP is not only more complex but also more expensive. We were interested in assessing the short-term effects of administration of oxygen below the site of airway obstruction by means of a 9-French (2.7 mm) transtracheal catheter (TT O 2) that was recently developed for patients with chronic lung disease.v" We hypothesized that alveolar hypoxia would be prevented during apnea if the catheter Howrate exceeded oxygen consumption. Although ventilation/perfusion mismatching and increased venous admixture may accompany obstructive apneas, increasing the alveolar P02 should ameliorate arterial hypoxemia, stabilize respiratory control, and reduce the frequency of apnea. 9-11 Consequently, both the time spent apneic and hypoxic would be reduced. U sing all-night polysomnography we compared TT O2 with nasal CPAP and NC O 2 in its ability to reduce oxygen desaturations and obstructive apnea. In some of these patients, we also measured the effects of room air administered through the catheter (TT RA) since infusion of gas beneath the site of airway obstruction could increase the airway pressure and exert a similar effect as CPAE We report herein the first study in which these three modalities and room air have been
severe obstructive sleep apnea. Although therapy with nasal continuous positive airway pressure (CPAP) is highly effective in the majority of these patients, the combined immediate and long-term failure rate measures 25 to 40 percent.l-" Accordingly, some patients may receive a tracheostomy or simply oxygen via nasal cannula. While previous studies have demonstrated that supplemental oxygen given by either face mask or nasal cannula (NC OJ attenuates the magnitude of oxygen desaturation, there is generally only a modest reduction in apnea frequencys" In addition, some patients require continuous oxygen therapy as well as CPAP during sleep. To our knowledge, the frequency of combined therapy has not been reported, but it is
*From the Intermountain Sleep Disorders Center, LDS Hospital, and Department of Medicine, University of Utah and Salt Lake Clinic, Salt Lake City. Supported in part by a grant from the Deseret Foundation, LDS Hospital, Salt Lake City. These data were presented in part at the Annual Meeting, American Thoracic Society, May 10, 1988. Manuscript received June 3; revision accepted September 3. Reprint requests: Dr. mmey, LDS Hospital, Sleep Lab, Salt Lake City 84143
1228
BL=baseline on room air; etCO.=end tidal CO,; NC 0.= oxygen given via nasal cannula; NREM = nonrapid eye movement; TST = total sleep time; TT O. = transtracheal oxygen; TT RA = room air administered through catheter
TTO, NasalCPAP and NasalO2 in OSA(Fameyet 8/)
Table I-Study Population· PatientlSexlAge, yr
Height, em
lIFl60
kg
157 183 183 167 188
2IMI57 ~48
41M147 51M1J7
PaO., mmHg
~ight,
106 135 147 116
_
..., ..........
,.....
5fJ 62
71 58 49
206
SaO., % 90
93 93 88 82
UB, gldl
mm Hg
pH
SAHI
14 17 15 16 15
37 31 40 43 45
7.47 7.50 7.41 7.43 7.43
35 50 74 78
PaCO.,
Additional Diagnoses CB CB
86
*SAHI = sleep apnea plus hypopnea index computed as total apneas and hypopneasltotal sleep time in hours; CB = chronic bronchitis.
compared in sleep apnea patients with polysomnography using one therapeutic condition across the night. METHODS
Five patients with severe obstructive sleep apnea were selected to participate in this study because they were either noncompliant with nasal CPAP or also required supplemental oxygen 24 hlday. All bad been previously tested by means of all-night polysomnography with and without nasal CPAE The anthropometric and baseline arterial blood gas data for the patient population are presented in 'Dlble 1. Male subjects predominated (415) and two subjects had chronic bronchitis secondary to cigarette smoking.
Study Protocol The research protocol was approved by the Institutional Review Board and informed consent was obtained. Following selection from
patients with previously documented sleep apnea, subjects were restudied with polysomnography on nonconsecutive nights using the following experimental conditions: baseline on room air (BL), nasal CP~ NC 0., 11 0., and 11 RA (three patients only). The sequence of tests, intervals between studies, and ultimate therapeutic levels used in each case are shown in 1llble 2. An unusually long interval occurred in one subject (No.3) who bad been tested on CPAP three months before his repeated baseline polysomnogram. This person's testing was delayed because he temporarily moved out of state. Not only was he completely intolerant of nasal CPA~ but he also had severe sleep apnea on baseline testing subsequent to the CPAP trial (sleep apnea plus hypopnea index 741b sleep). A transtracheal catheter (SCOOP) was placed using the technique described by Christopher et al.· If possible, one week was allowed for the subjects to become accommodated to the catheter before restudying with 1T 0.. Since experimental data were also being used for clinical decision making, maximal levels of nasal CPAP and oxygen were arbitrarily set that would be practical to continue after the stud~ Nasal CPAP was adjusted in approximate increments of
2.5 cm 8.0 pressure up to a maximum of 15 cm 8.0. Oxygen Bow rates were adjusted in approximate increments of 1 Umin using a pediatric Bowmeter up to a total Bow of 6 Umin. Oxygen or nasal CPAP was titrated as quickly as possible, generally within the first hour following the initiation of sleep to a target SaO. of 90 percent, after which therapeutic levels remained constant. If desaturations developed after the initial titration period, for example, with rapid eye movement (REM) sleep, no changes in therapy were made so that the experimental condition would remain constant throughout the night. In our laboratory, independent simultaneous recording of the SaO. at slow paper speed (20 em/h) has been a convenient method of titrating CPAE When respiratory disturbances occur at our elevation, desaturations or oscillations of the SaO. almost always result. 11 The SaO. pattern thereby provides a reproducible endpoint that may be less subject to observer error and artifactual changes compared with the detection of hypopneas, for example. Furthermore, since a major therapeutic objective was to maintain the SaO. at a physiologic level, the SaO. of90 percent was chosen as the end point (for both CPAP and O. trials). Consequently therapy was not speci6caUy adjusted to eradicate apneas and hypopneas. In three subjects, room air (Tr RA) was administered through the catheter and compared with TT O. at identical flow rates (3 Umin). Patient 5, who was restudied seven months after his original TT O. test, underwent a second night of testing with IT O. with a resultant decrease in his O. requirements to 3 Umin.
Sleep and Respiratory Meaaurements Polysomnographic recordings included standard placements for continuous monitoring of central and occipital electroencephalogram (~4 and 01J(2), horizontal electro-oculogram, and submental and anterior tibialis electromyogram. Airftow at the nose and mouth was sensed by measurement of end tidal CO. (etCOJ (Biochem MicroSpan Capnograph). Qualitative tidal volume (VT), thoracic excursion, and abdominal excursion were recorded by inductive plethysmography (Respigraph). Qualitative diagnostic
Table I-Sequence O!Polyaomnograplay and Treatment Conditiont· Case
Study 1
Interval
Study 2
Interval
Study 3
Interval
Study 4
1
6 days
BL
2 days
2mo
5 days
CPAP 10cm BL
1 day
NCO. lUmin BL
4
CPAP 10 em NCO. 5Umin CPAP 15cm BL
5
BL
ITO. lUmin 170. 3Umin ITO. 3Umin ITO. 2Umin 110. 6Umin
2
3
3mo 2 days 7 days
NCO. 2.5Umin NCO. 6Umin
7 days 3 days 7 days
NCO. 6Umin CPAP 15cm CPAP 15cm
34 days
1 day 26 days 6 days
Interval
Study 5
6 days
TIM 3Umin ITO. 3Umin TIO. 3Umin
9mo 7mo
Interval
Study 6
37 days
ITRA 3Umin l1RA 3Umin
8 days
*BL = baseline polysomnography; CPAP = nasal continuous positive airway pressure; numbers underneath de6ne 6nallevel of pressure in em B,O; NC O.=nasal cannula oxygen; numbers underneath refer to 6nal oxygen Bow; IT 0.= transtracheal oxygen; numbers underneath refer to final oxygen flow; 11 RA = transtracheal room air; numbers underneath refer to Bow of room air through transtracheal catheter. CHEST I 101 151 MA'f, 1992
1229
Table 3-Meana and Standard Deviationsfor Apnea and HfI1JOPfIeIJ Indicea ocrou AU Conditions· Baseline NREM Central apnea Obstructive apnea Hypopnea Apneas and hypopneas REM Central apnea Obstructive apnea Hypopnea Apneas and hypopneas Total sleep time Central apnea Obstructive apnea Hypopnea Apneas and hypopneas
NC02
CPAP
TI02
2.1 ±2.9 1.3 ± 1.9 2.5±3.8BL 5.9±5.6"
O.O±O.O 5.88±8.1 2O.6± 18.8 26.6±22.8
1.3±2.8 36.0 ± 19. 7~·:!! 27.9±20.7 65.3±21.€·:!!
1.4±2.0 40.7 ± 15. 7~·:!! 19.5± 19.0 61.6± 10.7~·:!!
7.7± 12.4 35.4 ± 20.9"' 21.2± 18.7 64.4 ± 29.6"·cp
8.9± 13.9 33.7 ± 26.()1T 6.2±5.5 48.8±20.3
10.0± 17.6 11.3± 10.5 9.1 ± 18.6 30.3± 17.6
2.5±3.3 4.4±4.0 16.6± 14.3 23.5± 13.2
2.1 ±4.0 35.7 ± 18.te·:!! 26.7 ± 20.5
2.3±2.7 39.0± 15.~·:!! 17.7± 17.3 59.0± 11.~·!!:
3.8±5.8 4.1±3.9 6.0± 11.0 13.8± 10.0
0.4±0.4 5.7±7.5 20.1 ± 16.9 26.2 ± 20.7
64.5±21.~·:!!
$BL·!!:=differs from baseline p
Data Analysis Using standard criteria,a3.14 records were analyzed manually for sleep stages and respiratory events without knowledge of treatment condition. Apneas were defined by an 80 to 100 percent reduction in the airflow signal (etC OJ compared with baseline, whereas hypopneas were defined as a 50 to 80 percent reduction. The occurrence and duration of respiratory events were corroborated by concurrent changes in VT. However, arterial oxygen desaturations were not required to define an apnea or hypopnea since desaturations were attenuated by oxygen administration. Almost all apneas and hypopneas were terminated by briefelectrophysiologic arousals. By de6nition, all apneas and hypopneas lasted at least 10 s. Obstructive apneas were indicated by the presence of respiratory effort or paradoxic thoracic-abdominal motion while central events were defined by the absence of any apparent respiratory effort. If there were both central and obstructive components to a respiratory event, an event was defined as mixed. Since there were few pure obstructive apneas for meaningful analysis and since there is no evidence that mixed and obstructive apneas have unique patho-
cP·~=differs
physiologic mechanisms or clinical consequences, mixed and obstructive apneas were grouped together as obstructive. Hypopneas were not differentiated as obstructive or central. The frequencies for apneas and hypopneas were defined for nonrapid eye movement (NREM), rapid eye movement (REM), and total sleep time (I'ST). The sleep apnea index was computed as the total of all apneas divided by TST in hours. The sleep hypopnea index and the sleep apnea/hypopnea index were computed similarly. Data obtained during the initial oxygen or CPAP titration period were excluded from the analysis. Means and standard deviations were determined on the following parameters: apnea index, hypopnea index, apnea/hypopnea index, Sa02 nadir during respiratory events, duration of apneas and hypopneas, TST, sleep efficiency (TST/total recording time x 1(0), electrophysiologic arousals per hour of sleep, number of awakenings, percentage of stage REM, and percentage of stage 1 to 4 NREM sleep. One-way analysis of variance with repeated measurements was performed on the data. When a significant F test was obtained, a Duncan's multiple range test was used to compare specific means. Statistical comparisons were not performed on TT RA data because of the small number of subjects. RESULTS
There were no procedure-related complications in these subjects such as hemorrhage, subcutaneous emphysema, pneumothorax, or infection. When the
Table 4-Meana and Standard Deviations ofUngth (a) and Arterial Oxygen De8tJturation (Percent)for Apneaa
and HypopnetJB GCroB8 All Conditions· Baseline
Length, s Central apnea Obstructive apnea Hypopnea Apneas and hypopneas Mean nadir desaturation, % Central apnea Obstructive apnea Hypopnea Apneas and hypopneas
NCOI
CPAP
TI02 23.0±23.0 28.9± 13.3 26.2± 13.8 26.3± 13.5 89.4±5.7 89.2±5.0 9O.2±4.0 89.7±4.6
16.9± 10.1 27.6± 10.6 20.8± 11.9 26.1± 10.6
19.6± 13.3 28.6± 10.9 28.9± 10.8
12.1 ±3.1 24.9± 14.0 12.8±3.9:!! 21.9± 13.8
69.7±21.8" 69.4 ± 12.~·:!! 72.0± 13.~·~·!!: 70.4 ± 12.~·cP.!!:
85.0± 14.5 85.5±5.9 87.9±4.4 86.2±5.6
81.4± 12.5 77.3± 16.9 86.7±9.1 8O.0± 12.9"'
28.2±13.~
from baseline p
$BL·~=differs
1230
cP·~=differs
TTO. Nasal CPAP and Nasal O2 in OSA (Farney et aJ)
Table 5-1ndividutJl Value. for Apneaa, Hypopneaa, and SaO. Nadir acroa. All Conditiona* Baseline Case No. 1 2 3 4 S Mean SD
CPAP
~COI
AI
HI
Sa02 Nadir, %
AI
34.S 13.S S1.1 23.2 66.9 37.8 ±21.4
0.0 36.4 23.3 55.0 19.0 26.7 ±20.S
86.2 73.2 73.7 66.4 S2.6 70.4 ±12.3
37.3 36.S 54.2 22.9 55.7 41.3 ±13.7
HI "
SaO. Nadir, % 92.S OO.S 84.5 85.1 78.3 86.2 ±S.6
9.3 11.6 7.0 48.S 12.2 17.7 ±17.3
AI 1.9 8.6 2.2 11.2 lS.3 7.8 ±S.9
TrO.
HI
SaO. Nadir, %
AI
HI
SaO. Nadir, %
0.0 0.0 25.3 0.0 4.S 6.0 ±11.0
86.8 87.1 92.7 71.9 61.6 80.0 ±12.9
0.7 4.0 19.1 2.6 4.3 6.1 ±7.4
4.1 .7.9 30.8 43.1 4.6 20.1 ±16.9
94.3 89.1 93.0 89.S 82.6 89.7 ±4.6
• AI = apnea index computed as total apneasltotal sleep time in hours; HI = hypopnea index computed as total hypopneasltotal sleep time in hours.
study was completed, all patients elected to continue using TT O2 rather than use alternative therapies. Objective studies to assess the state of daytime sleepiness were not performed in this study. However, subsequent clinical interviews revealed that these subjects felt more alert and snoring was reportedly diminished with TT 02. Respiratory Parameters
The results of standard respiratory parameters are shown in Tables 3 through 5. Arterial Oxygen Saturation: Compared with room air breathing, all treatment conditions improved Sa02. With TT 02' the Sa02 was almost always maintained at about 90 percent with oximetric patterns suggesting respiratory stabihty " Oxygen therapy via nasal cannula also increased the Sa02 level, but respiratory instability was evident by persistent fluctuations. The Sa02de saturation during NC O2was greater than with TT O2 but did not reach statistical significance. Although the Sa02 was usually maintained near 90 percent with CPA~ respiratory disturbances, particularly during REM sleep, resulted in substantial desaturations (mean Sa02 nadir during REM sleep, 78.5 percent, and during NREM sleep, 83.7 percent). Representative recordings of the Sa02 from patient 5 are shown in Figure 1. This subject presented with typical symptoms of extremely loud snoring and excessive sleepiness associated with marked peripheral
edema, obesity, and alveolar hypoventilation. While breathing room air, the apnealhypopnea index measured 85.9Ihsleep and the mean oxygen desaturation was 52.6 percent. Treatment with nasal CPAP reduced the frequency of respiratory disturbances effectively but due to baseline hypoxia, the oxygen saturation was not adequately corrected (mean nadir, 61.6 percent). Oxygen via nasal cannula resulted in a substantial increase in Sa02 but there were still frequent apneas, hypopneas, and arterial oxygen desaturations. Transtracheal oxygen not only reduced the frequency of all respiratory events but also provided a satisfactory Sa02. Paired transtracheal oxygen and room air studies were performed seven months following the initial studies (Fig 1 through 3). ApnealHypopnea Frequency: Table 5 shows the individual data for each patient. The frequency of apneas and hypopneas was most effectively reduced by means ofCPAP (apnealhypopnea index 13.8Ih sleep vs baseline of64.6Ih sleep) while the effects ofTT O 2 were more variable. Despite the stable appearance by oximetry, moderately frequent hypopneas were present with TT O 2 in some patients (hypopnea index 20.1Ih sleep). However, the effect ofTT O 2 on apnea index (6.1Ih sleep) was equivalent to CPA}! Although there was a trend, the apnealhypopnea index was not significantly reduced with NC O2 compared with room air. We have considered the possibility that the treat-
Table 6-Mearu and Standard DeoitJtiona for Sleep VtJritJbla tJCrOII All Conditiona. Baseline Total sleep time, h Stage 1 NREM, % Stage 2 NREM, % Stage 3 and 4 NREM, % REM,% Sleep efficiency, % Arousal index, events/h
6.7±0.8 2O.4± 12.3 64.6±7.2 0.S±0.6 lS.1±6.0 88.6±6.2 74.0±46.1
NCO. 6.8±0.6 10.9 ± 6.()8L 74.6±8.1 0.3±0.S 14.S±4.8 9O.7±S.8 53.S ± 29.9
.BL·!!:=dilfers from baseline p
CPAP
TrO.
S.6± 1.4 13.6±S.l B1• 48.8 ± 19.f)8L.~.!! 8.9 ± 10.SBL. Ne •n 29.2 ± 10.8BL. Ne •n
6.8±O.9 11.1 ±3.Q8L 72.6±3.2 1.8±2.6 14.9±4.7 9O.4±7.S 41.S ± 28.1 BL
84.9± 12.0 21.9 ± 9.e!:·NC
differs from NC 0. p
CHEST I 101 151 MA~
1992
1231
100j ROOM AIR
PATIENT 5
~'OO80
80
60 40
SAl 66.9
100j N-CPAP (15cm) 80
SAl 15.5
60 40
60 40
SAHl 85.9
[100
~:~ 20.0
40
100 j NC 02 (6LIm) z
0
.....
0:
.....
['00 80
80
60
SAHl 67.9
60 40
40
::)
(f)
z
L.LJ ~
>-
x 0
100j TT 02~~~ 80
SAl 4.3
60 40
SAHl 8.9
[100 80 60 40
roo
100j TT RA (3L1m) 80
80
60 40
I~~ j TT O2(3L1m l 60 40 I
SAl 14.2
SAHI 49.5
SAl 3.9
SAHl 10.7
roo
80 60
40 i
I
2
60 40
4
3 TIM E ( hours)
FIGURE 1. Oximetric recordings from patient 5. The Sa02 recorded from the second to fourth hour of each experimental condition are shown. The more severe desaturations in each panel are related to REM sleep.
ment conditions may have led to artifactual changes in respiratory scoring. However, the concordance between measurements of etC02 , VT, chest wall excursion, and abdominal excursion in association with electrophysiologic arousals at the termination of respiratory events indicates that their frequency and duration were not underestimated. Note that the changes in the arousal index paralleled changes in apnea plus hypopnea index (Tables 3 and 6). Duration of Apnea/Hypopnea: None of the treatments resulted in a significant increase in duration of apnea compared with the baseline room air condition. However, there was a significant difference in the hypopnea duration (p
shown in Figures 2 and 3. Since apneas and hypopneas did not result in consistent differences in desaturation, mean Sa02 values for all respiratory events are plotted. Compared with baseline studies, TT RA resulted in a slight increase in the Sa02 nadir and a mild reduction in frequency of apneas and hypopneas (apnea plus hypopnea index: BL, 79.5; TT RA, 66.7; TT O2,39.9). In one case (No.5), the apnea index was markedly decreased with TT RA but there were frequent hypopneas associated with substantial oxygen desaturations (Figs 1 through 3). In two subjects, frequent hypopneas were also observed with TT 02' but the overall Sa02 was about 90 percent (mean nadir, 88.7 percent). Sleep Variables: Table 6 shows means and SDs for standard sleep parameters across experimental conditions. In general, sleep architecture was more normal during CPAE The percentage of REM and stage 314 sleep increased, whereas the percentage of stage 2 no, Nasal CPAPand NasalO2 in OSA (Farney et aI)
100 90 e,
80
---APNEA + HYPOPNEA INDEX -----APNEA INDEX
5 4
UJ UJ -J
3
:I:
5~
en 70 ...... en
60
~
z LtJ > 50 LtJ
\
,,
\
\
\
3 .. ---~~\----------\
a: 40 0 ~
<[
a: Q..
en
20
\
\
\\
" ,,
,
\',
"
"
"
\
4 .-----------__'-\-,.
......... -
'
10
o
"
\
30
UJ
a::
" \
~
-~,
--I------.........,.-------~
BASE LINE
"
-.
,
-... _- ...... -..... -
~'-I
...
..... -
I
TT RA STUDY CONDITIONS
2. Apnea and hypopnea indices of three subjects while receiving either room air or oxygen via transtracheal catheter at comparable Bow rates. Subjects 4 and 5 were tested nine and seven months, respectively; after their original evaluations. FIGURE
NREM sleep decreased in the CPAP condition as compared with other conditions. There was a correspondence between the effectiveness of respiratory therapy and changes in sleep parameters. The number of arousals decreased significantly with both CPAP and TT O 2 administratio~ as compared with the baseline study Treatment with TT O 2 resulted in the next greatest reduction in apnea plus hypopnea index to 26.2 (± 20.1) with a corresponding reduction in the arousal index to 41.5(±28.1). Treatment with NC O 2 had the least impact on both respiratory parameters and frequency of arousals. Stage 1 NREM sleep was reduced (p
The most important observation from this study is that transtracheal oxygen maintained a physiologically adequate 8a02 and reduced the frequency of sleepdisordered breathing in these patients with severe sleep apnea. In this study, a pragmatic end point (Sa02 of 90 percent) was selected because of previous experience." Greater reductions in apnea and hypopnea frequency may have been demonstrated if the measurement of airflow tidal volume signal, and/or electrophysiologic arousals had been used. Our results are consistent with other more limited studies showing
marked reductions in the apnea/hypopnea index with TT O2 administration. 15.16 Because TT O 2 was effective and well tolerated, the need for tracheostomy was eliminated. Oxygen via nasal cannula was found to be the least effective form of therapy in the present study As expected, the frequency of periodic breathing was most effectively reduced by CPAE Except when there was significant hypoxia during wakefulness, CPAP also resulted in desirable Sa02 levels. One of the most intriguing questions raised by this study concerns the mechanism of TT O 2 on reducing respiratory instability, and in particular the sleep apnea index. In some cases, the reduction in both apneas and hypopneas was dramatic (Fig 1). In other cases, there were residual respiratory events that were predominantly hypopneas but without significant desaturation. Thus, even when respiratory disturbances were not completely eliminated by TT 02' the more severe grades of obstruction appear to have been reduced which ameliorated oxygen desaturation. Information concerning the effect of either nasal or transtracheal oxygen therapy in patients with sleep apnea is limited.3-5·15-17 Various mechanisms have been proposed for the salutary effect of oxygen on eliminating or reducing periodic breathingv'? that include the following: (I) a direct stimulation of the central nervous system; (2) reduction of upper airway resistance; 804 (3) stabilization of chemical feedback by reduction of peripheral chemoreceptor activity According to models of the respiratory control system that incorporate negative feedback, ventilatory instability is directly 100 90 N
o ~
z
0 ~
80
70
<[
a::
:::>
~
<[
3
4
60
en
z 50
5
L\J
~
X
0
40
-J
<[
a:: 30
L\J
~
a: <[
20 10
o
- I- - - - - - - - I --------BASE LINE TT RA STUDY CONDITIONS
FIGURE 3. Mean SaOI nadir during apneas and hypopneas of three subjects while receiving either room air or oxygen via transtracheal catheter at comparable Row rates.
CHEST 1101 1 5 1 MA'f, 1992
1~
correlated with hypoxia, hypercapnia, reduced lung O 2 and CO2 storage volumes, and increased gain of the peripheral chemoreceptors." An additional mechanism may be operative with TT 02. As a result of continuous oxygen How through the catheter, an increase in the mean airway pressure might develop sufficient to maintain airway patency similar to the effect of nasal CPAE Increased airway pressure may also increase functional residual capacity, which has been shown to enlarge pharyngeal crosssectional area in patients with obstructive sleep apnea.!" Transtracheal room air at comparable How rates to TT O 2 resulted in slight improvement in both apnea! hypopnea index and Sa02 nadir compared with baseline data. The transtracheal room air studies were interesting but unfortunately inconclusive because of the small number and because the infusion of room air still provides oxygen, albeit at a lower concentration than TT 02. Thus, the effect of oxygen could not be completely separated from increased airway pressure. Based on the present study and consistent findings by other investigators, we believe that TT O 2 exerts its major effect by stabilizing chemoreceptor activity and by reducing the central component of apnea, although increased airway pressure may playa role. The superiority of TT O 2 over NC O 2 most likely stems from the more reliable delivery of oxygen to the alveolar gas compartment when upper airway occlusion is present. Previous investigators have demonstrated that oxygen via nasal prongs or face mask consistently reduces both the severity of oxygen desaturation and the frequency of apnea in patients with obstructive sleep apnea syndrome.3-5 Because of different study designs, it is difficult to compare our results. We were concerned that oxygen therapy would prolong the duration of respiratory events and increase accumulated apnea time.3.5.15.19.20 However, our data and those of Chauncey and Aldrich" did not indicate that oxygen therapy by transtracheal catheter or nasal cannula significantly increased the duration of apneic events. In contrast to some studies.P-" our subjects were examined using one steady-state treatment modality throughout the entire sleep period, which we believe may provide more realistic data. It is conceivable that the order of the studies or intervals between tests may have biased these results. Long-term use of CPAP has been shown to change the ventilatory response to CO 2,21 which could possibly also modify the extent of obstructive sleep apnea. It is unlikely that the prior use of CPAP was a confounding factor in this study because these patients were selected on the basis of being poorly compliant with CPAP and none was using this therapy regularly. Transtracheal oxygen studies were delayed in several patients because of scheduling conflicts and one pa1234
tient (No.1) received oxygen via nasal cannula. Gold et al5 have demonstrated that long-term oxygen therapy has no effect on apnea frequency beyond the period of administration. Therefore, the use of oxygen before TT O 2 studies would not likely skew the results. Finally, the order of tests preceding TT O 2 studies was randomized so that there would be no consistent effect from any therapy. Nasal CPAP is the optimal therapy for immediately eliminating obstructive apneas and normalizing sleep architecture in the majority of patients. However, some will be intolerant of nasal CPAP and others may require concomitant and continuous use of oxygen. This study suggests that TT O 2 may be a viable therapeutic alternative. When supplemental oxygen was required in addition to CPAE transtracheal oxygen was successfully used as a single modality, thus reducing complexity and treatment expense. We did not address potential long-term complications of transtracheal catheter therapy nor did we evaluate other catheter systems. Chronic therapy with TT O 2 could be complicated by mucosal ulcerations and infection. Transtracheal oxygen therapy requires frequent cleaning and the catheter can be easily occluded by mucus in patients with heavy secretions. Some of the subjects in this study may not be representative of the majority of patients with severe sleep apnea. All were more difficult to treat than usual which, in fact, was the specific reason for considering them for an experimental treatment. Two had chronic bronchitis and all were studied at moderately high elevation. Nevertheless, we do not believe that these considerations seriously detract from the main thrust of this article, which is that TT O 2 may be useful when other modalities of therapy are not successful or are impractical. Given the success of TT O 2 in these patients, transtracheal oxygen therapy may also be beneficial in less severe cases, although additional subjects should be studied before such conclusions can be drawn. In some subjects, hypopneas may still be present but perhaps the single most significant physiologic consequence, hypoxia, can be prevented. The frequency of arousals was also reduced with transtracheal oxygen but further studies are necessary to determine the long-term effects on sleep architecture and daytime symptoms as well as to better define the role of airway pressure. ACKNOWLEDGMENTS: The authors wish to thank Kathy Bradley for her invaluable assistance and patience in the preparation of this manuscript, William Clark and Jan Kramer for their technical expertise with polysomnography, Julian Maack for medical illustrations, Alan Abdulla, M. D., for his contributions in the initial studies, and Alan H. Morris, M.D., Robert O. Crapo, M.D., and Arthur S. Slutsky, M. D., for their critical reviews and suggestions.
REFERENCES 1 Sanders MH, Gmendl CA, Rogers RM. Patient compliance with nasal CPAP therapy for sleep apnea. Chest 1986; 90:330-37 ITO. NasalCPAPand NasalO2 in OSA (Farney et 8/)
2 Waldorn RE, Herrick rw Nguyen MC, O'Donnell AE, Sodero J, Potolicchio SJ. Long-term compliance with nasal continuous positive airway pressure therapy of obstructive sleep apnea. Chest 1990; 97:33-8 3 Martin RJ, Sander MH, Gray BA, Pennock BE~~ute and.longterm ventilatory effects of hyperoxia in the adult sleep apnea syndrome. Am Rev Respir Dis 1982; 125:175-80 4 Smith PL, Haponik EF, Bleecker ER. The effects of oxygen in patients with sleep apnea. Am Rev Respir Dis 1984; 130:958-63 5 Gold AR, Schwartz AR, Bleecker ER, Smith PL. The effects of chronic nocturnal oxygen administration upon sleep apnea. Am Rev Respir Dis 1986; 134:925-29 6 Heimlich H, Carr G. Transtracheal catheter technique for pulmonary rehabilitation. Ann Otol Rhinol Laryngol 1985; 94:502-04 7 Heimlich H. Respiratory rehabilitation with transtracheal oxygen system. Ann Otol Rhinol Laryngoll982; 91:643-47 8 Christopher KL, Spofford BT, Petrun MD, McCarty DC, Goodman JR, Petty TL. A program for transtracheal oxygen delivery: assessment of safety and efficacy. Ann Intern Med 1987; 107:802-oB 9 Berssenbrugge A, Dempsey J, Iber C, Skatrud J, Wilson ~ Mechanisms of hypoxia-induced periodic breathing during sleep in humans. J Physioll983; 343:507-24 10 Phillipson EA. Control of breathing during sleep. Am Rev Respir Dis 1978; 118:909-39 11 Khoo MCK, Kronauer RE, Strohl K~ Slutsky AS. Factors inducing periodic breathing in humans: a general model. J Appl Physiol: Respir Environ Exercise Physiol 1982; 53:644-59 12 Farney RJ, Walker LE, Jensen RL, Walker JM. Ear oximetry to detect apnea and differentiate rapid eye movement (REM) and
non-REM (NREM) sleep. Chest 1986; 89:533-39 13 Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. In: Brain information service/brain research institute. LosMgeles: University of California, 1968 14 Bornstein SK. Respiratory monitoring during sleep: polysomnography. In: Guilleminault C, 00. Sleeping and waking disorders: indications and techniques. Menlo Park, Calif: AddisonWesley Publishing Co, 1982:183-212 15 Spofford BT, Christopher KL, Hoddes ES. Transtracheal oxygen therapy for obstructive sleep apnea [abstract]. Chest 1986; 89:484(S) 16 Chauncey }B, Aldrich MS. Preliminary findings in the treatment of obstructive sleep apnea with transtracheal oxygen. Sleep 1990; 13:167-74 17 Elmer JC, Farney RJ, Walker JM, 000 RJ, Viscomi VA. The comparison of transtracheal oxygen with other therapies for obstructive sleep apnea. Am Rev Respir Dis 1988; 137:311A 18 Hoffstein ~ Zamel N, Phillipson E. Lung volume dependence of pharyngeal cross-sectional area in patients with obstructive sleep apnea. Am Rev Respir Dis 1984; 130:175-78 19 Hudgel D~ Hendricks C, Dooley A. Alteration in obstructive apnea pattern induced by changes in oxygen- and carbondioxide-inspired concentrations. Am Rev Respir Dis 1988; 138:16-9 20 Motta J, Guilleminault C. Effects of oxygen administration in sleep-induced apneas. In: Guilleminault C, Dement ~ eds. Sleep apnea syndrome. New York: Alan R Liss Inc, 1978:137-44 21 Berthon-Jones M, Sullivan CEo lime course of change in ventilatory response to CO 2 with long-term CPAP therapy for obstructive sleep apnea. Am Rev Respir Dis 1987; 135:144-47
CHEST I 101 I 5 I MA'Y, 1992
1235
Robert J Farney, M.D., F.C.C.R; James M. Walker; Ph.D.; Jeffrey C. Elmer; M.D., F.C.C.R; Vincent A. VIScomi, M.D., F.C.C.R; and R. ]on Ord, M.D. The effect of transtracheal oxygen administration by means of a 9-French (2.7 mm) percutaneous catheter was assessed in 6ve patients with severe obstructive sleep apnea. We hypothesized that the delivery of oxygen below the site of airway obstruction should reduce the arterial oxygen desaturation during apneas and hypopneas, thereby increasing respiratory stability. Standard sleep and respiratory measurements were recorded in these subjects with a11night polysomoography on nonconsecutive nights during four experimental conditions: room air (BL), nasal continuous positive airway pressure (CPAP), nasal 0. (NC 0.), and transtracheal O. (TT Os). In three of these subjects, room air was infused (TT RA) at Bow rates comparable to TT 0 •. Compared with baseline room air measurements, TT 0. DOt only signi6cantly increased the SaO. nadir from 70.4 percent to 89.7 percent (p
sleep apnea/hypopnea (13.81h sleep) but did not abolish desaturations when apneas occurred (mean SaO. nadir, 80.0 percent). Compared with oxygen, transtracheal infusion of room air appeared to be somewhat effective; however, the small number of studies with TT RA precluded statistical analysis. We believe that TT 0. is superior to NC 0. for some patients with obstructive sleep apnea because continuous oxygen Bow below the site of airway obstruction more reliably prevents alveolar hypoxia and respiration is stabilized. Infusion of air or oxygen through the tracheal catheter Bow may also increase mean airway pressure and reduce obstructive apnea similar to nasal CPAP. We CODelude that TT 0. may be an effective alternative mode of therapy for some patients with severe sleep apnealhypopDea when nasal CPAP is not tolerated or when combined oxygen and nasal CPAP are required. (Cheat 1992; 101:1228-35)
purpose of this study was to evaluate the efficacy T ofhe transtracheal oxygen (TT OJ in patients with
clear that oxygen plus CPAP is not only more complex but also more expensive. We were interested in assessing the short-term effects of administration of oxygen below the site of airway obstruction by means of a 9-French (2.7 mm) transtracheal catheter (TT O 2) that was recently developed for patients with chronic lung disease.v" We hypothesized that alveolar hypoxia would be prevented during apnea if the catheter Howrate exceeded oxygen consumption. Although ventilation/perfusion mismatching and increased venous admixture may accompany obstructive apneas, increasing the alveolar P02 should ameliorate arterial hypoxemia, stabilize respiratory control, and reduce the frequency of apnea. 9-11 Consequently, both the time spent apneic and hypoxic would be reduced. U sing all-night polysomnography we compared TT O2 with nasal CPAP and NC O 2 in its ability to reduce oxygen desaturations and obstructive apnea. In some of these patients, we also measured the effects of room air administered through the catheter (TT RA) since infusion of gas beneath the site of airway obstruction could increase the airway pressure and exert a similar effect as CPAE We report herein the first study in which these three modalities and room air have been
severe obstructive sleep apnea. Although therapy with nasal continuous positive airway pressure (CPAP) is highly effective in the majority of these patients, the combined immediate and long-term failure rate measures 25 to 40 percent.l-" Accordingly, some patients may receive a tracheostomy or simply oxygen via nasal cannula. While previous studies have demonstrated that supplemental oxygen given by either face mask or nasal cannula (NC OJ attenuates the magnitude of oxygen desaturation, there is generally only a modest reduction in apnea frequencys" In addition, some patients require continuous oxygen therapy as well as CPAP during sleep. To our knowledge, the frequency of combined therapy has not been reported, but it is
*From the Intermountain Sleep Disorders Center, LDS Hospital, and Department of Medicine, University of Utah and Salt Lake Clinic, Salt Lake City. Supported in part by a grant from the Deseret Foundation, LDS Hospital, Salt Lake City. These data were presented in part at the Annual Meeting, American Thoracic Society, May 10, 1988. Manuscript received June 3; revision accepted September 3. Reprint requests: Dr. mmey, LDS Hospital, Sleep Lab, Salt Lake City 84143
1228
BL=baseline on room air; etCO.=end tidal CO,; NC 0.= oxygen given via nasal cannula; NREM = nonrapid eye movement; TST = total sleep time; TT O. = transtracheal oxygen; TT RA = room air administered through catheter
TTO, NasalCPAP and NasalO2 in OSA(Fameyet 8/)
Table I-Study Population· PatientlSexlAge, yr
Height, em
lIFl60
kg
157 183 183 167 188
2IMI57 ~48
41M147 51M1J7
PaO., mmHg
~ight,
106 135 147 116
_
..., ..........
,.....
5fJ 62
71 58 49
206
SaO., % 90
93 93 88 82
UB, gldl
mm Hg
pH
SAHI
14 17 15 16 15
37 31 40 43 45
7.47 7.50 7.41 7.43 7.43
35 50 74 78
PaCO.,
Additional Diagnoses CB CB
86
*SAHI = sleep apnea plus hypopnea index computed as total apneas and hypopneasltotal sleep time in hours; CB = chronic bronchitis.
compared in sleep apnea patients with polysomnography using one therapeutic condition across the night. METHODS
Five patients with severe obstructive sleep apnea were selected to participate in this study because they were either noncompliant with nasal CPAP or also required supplemental oxygen 24 hlday. All bad been previously tested by means of all-night polysomnography with and without nasal CPAE The anthropometric and baseline arterial blood gas data for the patient population are presented in 'Dlble 1. Male subjects predominated (415) and two subjects had chronic bronchitis secondary to cigarette smoking.
Study Protocol The research protocol was approved by the Institutional Review Board and informed consent was obtained. Following selection from
patients with previously documented sleep apnea, subjects were restudied with polysomnography on nonconsecutive nights using the following experimental conditions: baseline on room air (BL), nasal CP~ NC 0., 11 0., and 11 RA (three patients only). The sequence of tests, intervals between studies, and ultimate therapeutic levels used in each case are shown in 1llble 2. An unusually long interval occurred in one subject (No.3) who bad been tested on CPAP three months before his repeated baseline polysomnogram. This person's testing was delayed because he temporarily moved out of state. Not only was he completely intolerant of nasal CPA~ but he also had severe sleep apnea on baseline testing subsequent to the CPAP trial (sleep apnea plus hypopnea index 741b sleep). A transtracheal catheter (SCOOP) was placed using the technique described by Christopher et al.· If possible, one week was allowed for the subjects to become accommodated to the catheter before restudying with 1T 0.. Since experimental data were also being used for clinical decision making, maximal levels of nasal CPAP and oxygen were arbitrarily set that would be practical to continue after the stud~ Nasal CPAP was adjusted in approximate increments of
2.5 cm 8.0 pressure up to a maximum of 15 cm 8.0. Oxygen Bow rates were adjusted in approximate increments of 1 Umin using a pediatric Bowmeter up to a total Bow of 6 Umin. Oxygen or nasal CPAP was titrated as quickly as possible, generally within the first hour following the initiation of sleep to a target SaO. of 90 percent, after which therapeutic levels remained constant. If desaturations developed after the initial titration period, for example, with rapid eye movement (REM) sleep, no changes in therapy were made so that the experimental condition would remain constant throughout the night. In our laboratory, independent simultaneous recording of the SaO. at slow paper speed (20 em/h) has been a convenient method of titrating CPAE When respiratory disturbances occur at our elevation, desaturations or oscillations of the SaO. almost always result. 11 The SaO. pattern thereby provides a reproducible endpoint that may be less subject to observer error and artifactual changes compared with the detection of hypopneas, for example. Furthermore, since a major therapeutic objective was to maintain the SaO. at a physiologic level, the SaO. of90 percent was chosen as the end point (for both CPAP and O. trials). Consequently therapy was not speci6caUy adjusted to eradicate apneas and hypopneas. In three subjects, room air (Tr RA) was administered through the catheter and compared with TT O. at identical flow rates (3 Umin). Patient 5, who was restudied seven months after his original TT O. test, underwent a second night of testing with IT O. with a resultant decrease in his O. requirements to 3 Umin.
Sleep and Respiratory Meaaurements Polysomnographic recordings included standard placements for continuous monitoring of central and occipital electroencephalogram (~4 and 01J(2), horizontal electro-oculogram, and submental and anterior tibialis electromyogram. Airftow at the nose and mouth was sensed by measurement of end tidal CO. (etCOJ (Biochem MicroSpan Capnograph). Qualitative tidal volume (VT), thoracic excursion, and abdominal excursion were recorded by inductive plethysmography (Respigraph). Qualitative diagnostic
Table I-Sequence O!Polyaomnograplay and Treatment Conditiont· Case
Study 1
Interval
Study 2
Interval
Study 3
Interval
Study 4
1
6 days
BL
2 days
2mo
5 days
CPAP 10cm BL
1 day
NCO. lUmin BL
4
CPAP 10 em NCO. 5Umin CPAP 15cm BL
5
BL
ITO. lUmin 170. 3Umin ITO. 3Umin ITO. 2Umin 110. 6Umin
2
3
3mo 2 days 7 days
NCO. 2.5Umin NCO. 6Umin
7 days 3 days 7 days
NCO. 6Umin CPAP 15cm CPAP 15cm
34 days
1 day 26 days 6 days
Interval
Study 5
6 days
TIM 3Umin ITO. 3Umin TIO. 3Umin
9mo 7mo
Interval
Study 6
37 days
ITRA 3Umin l1RA 3Umin
8 days
*BL = baseline polysomnography; CPAP = nasal continuous positive airway pressure; numbers underneath de6ne 6nallevel of pressure in em B,O; NC O.=nasal cannula oxygen; numbers underneath refer to 6nal oxygen Bow; IT 0.= transtracheal oxygen; numbers underneath refer to final oxygen flow; 11 RA = transtracheal room air; numbers underneath refer to Bow of room air through transtracheal catheter. CHEST I 101 151 MA'f, 1992
1229
Table 3-Meana and Standard Deviationsfor Apnea and HfI1JOPfIeIJ Indicea ocrou AU Conditions· Baseline NREM Central apnea Obstructive apnea Hypopnea Apneas and hypopneas REM Central apnea Obstructive apnea Hypopnea Apneas and hypopneas Total sleep time Central apnea Obstructive apnea Hypopnea Apneas and hypopneas
NC02
CPAP
TI02
2.1 ±2.9 1.3 ± 1.9 2.5±3.8BL 5.9±5.6"
O.O±O.O 5.88±8.1 2O.6± 18.8 26.6±22.8
1.3±2.8 36.0 ± 19. 7~·:!! 27.9±20.7 65.3±21.€·:!!
1.4±2.0 40.7 ± 15. 7~·:!! 19.5± 19.0 61.6± 10.7~·:!!
7.7± 12.4 35.4 ± 20.9"' 21.2± 18.7 64.4 ± 29.6"·cp
8.9± 13.9 33.7 ± 26.()1T 6.2±5.5 48.8±20.3
10.0± 17.6 11.3± 10.5 9.1 ± 18.6 30.3± 17.6
2.5±3.3 4.4±4.0 16.6± 14.3 23.5± 13.2
2.1 ±4.0 35.7 ± 18.te·:!! 26.7 ± 20.5
2.3±2.7 39.0± 15.~·:!! 17.7± 17.3 59.0± 11.~·!!:
3.8±5.8 4.1±3.9 6.0± 11.0 13.8± 10.0
0.4±0.4 5.7±7.5 20.1 ± 16.9 26.2 ± 20.7
64.5±21.~·:!!
$BL·!!:=differs from baseline p
Data Analysis Using standard criteria,a3.14 records were analyzed manually for sleep stages and respiratory events without knowledge of treatment condition. Apneas were defined by an 80 to 100 percent reduction in the airflow signal (etC OJ compared with baseline, whereas hypopneas were defined as a 50 to 80 percent reduction. The occurrence and duration of respiratory events were corroborated by concurrent changes in VT. However, arterial oxygen desaturations were not required to define an apnea or hypopnea since desaturations were attenuated by oxygen administration. Almost all apneas and hypopneas were terminated by briefelectrophysiologic arousals. By de6nition, all apneas and hypopneas lasted at least 10 s. Obstructive apneas were indicated by the presence of respiratory effort or paradoxic thoracic-abdominal motion while central events were defined by the absence of any apparent respiratory effort. If there were both central and obstructive components to a respiratory event, an event was defined as mixed. Since there were few pure obstructive apneas for meaningful analysis and since there is no evidence that mixed and obstructive apneas have unique patho-
cP·~=differs
physiologic mechanisms or clinical consequences, mixed and obstructive apneas were grouped together as obstructive. Hypopneas were not differentiated as obstructive or central. The frequencies for apneas and hypopneas were defined for nonrapid eye movement (NREM), rapid eye movement (REM), and total sleep time (I'ST). The sleep apnea index was computed as the total of all apneas divided by TST in hours. The sleep hypopnea index and the sleep apnea/hypopnea index were computed similarly. Data obtained during the initial oxygen or CPAP titration period were excluded from the analysis. Means and standard deviations were determined on the following parameters: apnea index, hypopnea index, apnea/hypopnea index, Sa02 nadir during respiratory events, duration of apneas and hypopneas, TST, sleep efficiency (TST/total recording time x 1(0), electrophysiologic arousals per hour of sleep, number of awakenings, percentage of stage REM, and percentage of stage 1 to 4 NREM sleep. One-way analysis of variance with repeated measurements was performed on the data. When a significant F test was obtained, a Duncan's multiple range test was used to compare specific means. Statistical comparisons were not performed on TT RA data because of the small number of subjects. RESULTS
There were no procedure-related complications in these subjects such as hemorrhage, subcutaneous emphysema, pneumothorax, or infection. When the
Table 4-Meana and Standard Deviations ofUngth (a) and Arterial Oxygen De8tJturation (Percent)for Apneaa
and HypopnetJB GCroB8 All Conditions· Baseline
Length, s Central apnea Obstructive apnea Hypopnea Apneas and hypopneas Mean nadir desaturation, % Central apnea Obstructive apnea Hypopnea Apneas and hypopneas
NCOI
CPAP
TI02 23.0±23.0 28.9± 13.3 26.2± 13.8 26.3± 13.5 89.4±5.7 89.2±5.0 9O.2±4.0 89.7±4.6
16.9± 10.1 27.6± 10.6 20.8± 11.9 26.1± 10.6
19.6± 13.3 28.6± 10.9 28.9± 10.8
12.1 ±3.1 24.9± 14.0 12.8±3.9:!! 21.9± 13.8
69.7±21.8" 69.4 ± 12.~·:!! 72.0± 13.~·~·!!: 70.4 ± 12.~·cP.!!:
85.0± 14.5 85.5±5.9 87.9±4.4 86.2±5.6
81.4± 12.5 77.3± 16.9 86.7±9.1 8O.0± 12.9"'
28.2±13.~
from baseline p
$BL·~=differs
1230
cP·~=differs
TTO. Nasal CPAP and Nasal O2 in OSA (Farney et aJ)
Table 5-1ndividutJl Value. for Apneaa, Hypopneaa, and SaO. Nadir acroa. All Conditiona* Baseline Case No. 1 2 3 4 S Mean SD
CPAP
~COI
AI
HI
Sa02 Nadir, %
AI
34.S 13.S S1.1 23.2 66.9 37.8 ±21.4
0.0 36.4 23.3 55.0 19.0 26.7 ±20.S
86.2 73.2 73.7 66.4 S2.6 70.4 ±12.3
37.3 36.S 54.2 22.9 55.7 41.3 ±13.7
HI "
SaO. Nadir, % 92.S OO.S 84.5 85.1 78.3 86.2 ±S.6
9.3 11.6 7.0 48.S 12.2 17.7 ±17.3
AI 1.9 8.6 2.2 11.2 lS.3 7.8 ±S.9
TrO.
HI
SaO. Nadir, %
AI
HI
SaO. Nadir, %
0.0 0.0 25.3 0.0 4.S 6.0 ±11.0
86.8 87.1 92.7 71.9 61.6 80.0 ±12.9
0.7 4.0 19.1 2.6 4.3 6.1 ±7.4
4.1 .7.9 30.8 43.1 4.6 20.1 ±16.9
94.3 89.1 93.0 89.S 82.6 89.7 ±4.6
• AI = apnea index computed as total apneasltotal sleep time in hours; HI = hypopnea index computed as total hypopneasltotal sleep time in hours.
study was completed, all patients elected to continue using TT O2 rather than use alternative therapies. Objective studies to assess the state of daytime sleepiness were not performed in this study. However, subsequent clinical interviews revealed that these subjects felt more alert and snoring was reportedly diminished with TT 02. Respiratory Parameters
The results of standard respiratory parameters are shown in Tables 3 through 5. Arterial Oxygen Saturation: Compared with room air breathing, all treatment conditions improved Sa02. With TT 02' the Sa02 was almost always maintained at about 90 percent with oximetric patterns suggesting respiratory stabihty " Oxygen therapy via nasal cannula also increased the Sa02 level, but respiratory instability was evident by persistent fluctuations. The Sa02de saturation during NC O2was greater than with TT O2 but did not reach statistical significance. Although the Sa02 was usually maintained near 90 percent with CPA~ respiratory disturbances, particularly during REM sleep, resulted in substantial desaturations (mean Sa02 nadir during REM sleep, 78.5 percent, and during NREM sleep, 83.7 percent). Representative recordings of the Sa02 from patient 5 are shown in Figure 1. This subject presented with typical symptoms of extremely loud snoring and excessive sleepiness associated with marked peripheral
edema, obesity, and alveolar hypoventilation. While breathing room air, the apnealhypopnea index measured 85.9Ihsleep and the mean oxygen desaturation was 52.6 percent. Treatment with nasal CPAP reduced the frequency of respiratory disturbances effectively but due to baseline hypoxia, the oxygen saturation was not adequately corrected (mean nadir, 61.6 percent). Oxygen via nasal cannula resulted in a substantial increase in Sa02 but there were still frequent apneas, hypopneas, and arterial oxygen desaturations. Transtracheal oxygen not only reduced the frequency of all respiratory events but also provided a satisfactory Sa02. Paired transtracheal oxygen and room air studies were performed seven months following the initial studies (Fig 1 through 3). ApnealHypopnea Frequency: Table 5 shows the individual data for each patient. The frequency of apneas and hypopneas was most effectively reduced by means ofCPAP (apnealhypopnea index 13.8Ih sleep vs baseline of64.6Ih sleep) while the effects ofTT O 2 were more variable. Despite the stable appearance by oximetry, moderately frequent hypopneas were present with TT O 2 in some patients (hypopnea index 20.1Ih sleep). However, the effect ofTT O 2 on apnea index (6.1Ih sleep) was equivalent to CPA}! Although there was a trend, the apnealhypopnea index was not significantly reduced with NC O2 compared with room air. We have considered the possibility that the treat-
Table 6-Mearu and Standard DeoitJtiona for Sleep VtJritJbla tJCrOII All Conditiona. Baseline Total sleep time, h Stage 1 NREM, % Stage 2 NREM, % Stage 3 and 4 NREM, % REM,% Sleep efficiency, % Arousal index, events/h
6.7±0.8 2O.4± 12.3 64.6±7.2 0.S±0.6 lS.1±6.0 88.6±6.2 74.0±46.1
NCO. 6.8±0.6 10.9 ± 6.()8L 74.6±8.1 0.3±0.S 14.S±4.8 9O.7±S.8 53.S ± 29.9
.BL·!!:=dilfers from baseline p
CPAP
TrO.
S.6± 1.4 13.6±S.l B1• 48.8 ± 19.f)8L.~.!! 8.9 ± 10.SBL. Ne •n 29.2 ± 10.8BL. Ne •n
6.8±O.9 11.1 ±3.Q8L 72.6±3.2 1.8±2.6 14.9±4.7 9O.4±7.S 41.S ± 28.1 BL
84.9± 12.0 21.9 ± 9.e!:·NC
differs from NC 0. p
CHEST I 101 151 MA~
1992
1231
100j ROOM AIR
PATIENT 5
~'OO80
80
60 40
SAl 66.9
100j N-CPAP (15cm) 80
SAl 15.5
60 40
60 40
SAHl 85.9
[100
~:~ 20.0
40
100 j NC 02 (6LIm) z
0
.....
0:
.....
['00 80
80
60
SAHl 67.9
60 40
40
::)
(f)
z
L.LJ ~
>-
x 0
100j TT 02~~~ 80
SAl 4.3
60 40
SAHl 8.9
[100 80 60 40
roo
100j TT RA (3L1m) 80
80
60 40
I~~ j TT O2(3L1m l 60 40 I
SAl 14.2
SAHI 49.5
SAl 3.9
SAHl 10.7
roo
80 60
40 i
I
2
60 40
4
3 TIM E ( hours)
FIGURE 1. Oximetric recordings from patient 5. The Sa02 recorded from the second to fourth hour of each experimental condition are shown. The more severe desaturations in each panel are related to REM sleep.
ment conditions may have led to artifactual changes in respiratory scoring. However, the concordance between measurements of etC02 , VT, chest wall excursion, and abdominal excursion in association with electrophysiologic arousals at the termination of respiratory events indicates that their frequency and duration were not underestimated. Note that the changes in the arousal index paralleled changes in apnea plus hypopnea index (Tables 3 and 6). Duration of Apnea/Hypopnea: None of the treatments resulted in a significant increase in duration of apnea compared with the baseline room air condition. However, there was a significant difference in the hypopnea duration (p
shown in Figures 2 and 3. Since apneas and hypopneas did not result in consistent differences in desaturation, mean Sa02 values for all respiratory events are plotted. Compared with baseline studies, TT RA resulted in a slight increase in the Sa02 nadir and a mild reduction in frequency of apneas and hypopneas (apnea plus hypopnea index: BL, 79.5; TT RA, 66.7; TT O2,39.9). In one case (No.5), the apnea index was markedly decreased with TT RA but there were frequent hypopneas associated with substantial oxygen desaturations (Figs 1 through 3). In two subjects, frequent hypopneas were also observed with TT 02' but the overall Sa02 was about 90 percent (mean nadir, 88.7 percent). Sleep Variables: Table 6 shows means and SDs for standard sleep parameters across experimental conditions. In general, sleep architecture was more normal during CPAE The percentage of REM and stage 314 sleep increased, whereas the percentage of stage 2 no, Nasal CPAPand NasalO2 in OSA (Farney et aI)
100 90 e,
80
---APNEA + HYPOPNEA INDEX -----APNEA INDEX
5 4
UJ UJ -J
3
:I:
5~
en 70 ...... en
60
~
z LtJ > 50 LtJ
\
,,
\
\
\
3 .. ---~~\----------\
a: 40 0 ~
<[
a: Q..
en
20
\
\
\\
" ,,
,
\',
"
"
"
\
4 .-----------__'-\-,.
......... -
'
10
o
"
\
30
UJ
a::
" \
~
-~,
--I------.........,.-------~
BASE LINE
"
-.
,
-... _- ...... -..... -
~'-I
...
..... -
I
TT RA STUDY CONDITIONS
2. Apnea and hypopnea indices of three subjects while receiving either room air or oxygen via transtracheal catheter at comparable Bow rates. Subjects 4 and 5 were tested nine and seven months, respectively; after their original evaluations. FIGURE
NREM sleep decreased in the CPAP condition as compared with other conditions. There was a correspondence between the effectiveness of respiratory therapy and changes in sleep parameters. The number of arousals decreased significantly with both CPAP and TT O 2 administratio~ as compared with the baseline study Treatment with TT O 2 resulted in the next greatest reduction in apnea plus hypopnea index to 26.2 (± 20.1) with a corresponding reduction in the arousal index to 41.5(±28.1). Treatment with NC O 2 had the least impact on both respiratory parameters and frequency of arousals. Stage 1 NREM sleep was reduced (p
The most important observation from this study is that transtracheal oxygen maintained a physiologically adequate 8a02 and reduced the frequency of sleepdisordered breathing in these patients with severe sleep apnea. In this study, a pragmatic end point (Sa02 of 90 percent) was selected because of previous experience." Greater reductions in apnea and hypopnea frequency may have been demonstrated if the measurement of airflow tidal volume signal, and/or electrophysiologic arousals had been used. Our results are consistent with other more limited studies showing
marked reductions in the apnea/hypopnea index with TT O2 administration. 15.16 Because TT O 2 was effective and well tolerated, the need for tracheostomy was eliminated. Oxygen via nasal cannula was found to be the least effective form of therapy in the present study As expected, the frequency of periodic breathing was most effectively reduced by CPAE Except when there was significant hypoxia during wakefulness, CPAP also resulted in desirable Sa02 levels. One of the most intriguing questions raised by this study concerns the mechanism of TT O 2 on reducing respiratory instability, and in particular the sleep apnea index. In some cases, the reduction in both apneas and hypopneas was dramatic (Fig 1). In other cases, there were residual respiratory events that were predominantly hypopneas but without significant desaturation. Thus, even when respiratory disturbances were not completely eliminated by TT 02' the more severe grades of obstruction appear to have been reduced which ameliorated oxygen desaturation. Information concerning the effect of either nasal or transtracheal oxygen therapy in patients with sleep apnea is limited.3-5·15-17 Various mechanisms have been proposed for the salutary effect of oxygen on eliminating or reducing periodic breathingv'? that include the following: (I) a direct stimulation of the central nervous system; (2) reduction of upper airway resistance; 804 (3) stabilization of chemical feedback by reduction of peripheral chemoreceptor activity According to models of the respiratory control system that incorporate negative feedback, ventilatory instability is directly 100 90 N
o ~
z
0 ~
80
70
<[
a::
:::>
~
<[
3
4
60
en
z 50
5
L\J
~
X
0
40
-J
<[
a:: 30
L\J
~
a: <[
20 10
o
- I- - - - - - - - I --------BASE LINE TT RA STUDY CONDITIONS
FIGURE 3. Mean SaOI nadir during apneas and hypopneas of three subjects while receiving either room air or oxygen via transtracheal catheter at comparable Row rates.
CHEST 1101 1 5 1 MA'f, 1992
1~
correlated with hypoxia, hypercapnia, reduced lung O 2 and CO2 storage volumes, and increased gain of the peripheral chemoreceptors." An additional mechanism may be operative with TT 02. As a result of continuous oxygen How through the catheter, an increase in the mean airway pressure might develop sufficient to maintain airway patency similar to the effect of nasal CPAE Increased airway pressure may also increase functional residual capacity, which has been shown to enlarge pharyngeal crosssectional area in patients with obstructive sleep apnea.!" Transtracheal room air at comparable How rates to TT O 2 resulted in slight improvement in both apnea! hypopnea index and Sa02 nadir compared with baseline data. The transtracheal room air studies were interesting but unfortunately inconclusive because of the small number and because the infusion of room air still provides oxygen, albeit at a lower concentration than TT 02. Thus, the effect of oxygen could not be completely separated from increased airway pressure. Based on the present study and consistent findings by other investigators, we believe that TT O 2 exerts its major effect by stabilizing chemoreceptor activity and by reducing the central component of apnea, although increased airway pressure may playa role. The superiority of TT O 2 over NC O 2 most likely stems from the more reliable delivery of oxygen to the alveolar gas compartment when upper airway occlusion is present. Previous investigators have demonstrated that oxygen via nasal prongs or face mask consistently reduces both the severity of oxygen desaturation and the frequency of apnea in patients with obstructive sleep apnea syndrome.3-5 Because of different study designs, it is difficult to compare our results. We were concerned that oxygen therapy would prolong the duration of respiratory events and increase accumulated apnea time.3.5.15.19.20 However, our data and those of Chauncey and Aldrich" did not indicate that oxygen therapy by transtracheal catheter or nasal cannula significantly increased the duration of apneic events. In contrast to some studies.P-" our subjects were examined using one steady-state treatment modality throughout the entire sleep period, which we believe may provide more realistic data. It is conceivable that the order of the studies or intervals between tests may have biased these results. Long-term use of CPAP has been shown to change the ventilatory response to CO 2,21 which could possibly also modify the extent of obstructive sleep apnea. It is unlikely that the prior use of CPAP was a confounding factor in this study because these patients were selected on the basis of being poorly compliant with CPAP and none was using this therapy regularly. Transtracheal oxygen studies were delayed in several patients because of scheduling conflicts and one pa1234
tient (No.1) received oxygen via nasal cannula. Gold et al5 have demonstrated that long-term oxygen therapy has no effect on apnea frequency beyond the period of administration. Therefore, the use of oxygen before TT O 2 studies would not likely skew the results. Finally, the order of tests preceding TT O 2 studies was randomized so that there would be no consistent effect from any therapy. Nasal CPAP is the optimal therapy for immediately eliminating obstructive apneas and normalizing sleep architecture in the majority of patients. However, some will be intolerant of nasal CPAP and others may require concomitant and continuous use of oxygen. This study suggests that TT O 2 may be a viable therapeutic alternative. When supplemental oxygen was required in addition to CPAE transtracheal oxygen was successfully used as a single modality, thus reducing complexity and treatment expense. We did not address potential long-term complications of transtracheal catheter therapy nor did we evaluate other catheter systems. Chronic therapy with TT O 2 could be complicated by mucosal ulcerations and infection. Transtracheal oxygen therapy requires frequent cleaning and the catheter can be easily occluded by mucus in patients with heavy secretions. Some of the subjects in this study may not be representative of the majority of patients with severe sleep apnea. All were more difficult to treat than usual which, in fact, was the specific reason for considering them for an experimental treatment. Two had chronic bronchitis and all were studied at moderately high elevation. Nevertheless, we do not believe that these considerations seriously detract from the main thrust of this article, which is that TT O 2 may be useful when other modalities of therapy are not successful or are impractical. Given the success of TT O 2 in these patients, transtracheal oxygen therapy may also be beneficial in less severe cases, although additional subjects should be studied before such conclusions can be drawn. In some subjects, hypopneas may still be present but perhaps the single most significant physiologic consequence, hypoxia, can be prevented. The frequency of arousals was also reduced with transtracheal oxygen but further studies are necessary to determine the long-term effects on sleep architecture and daytime symptoms as well as to better define the role of airway pressure. ACKNOWLEDGMENTS: The authors wish to thank Kathy Bradley for her invaluable assistance and patience in the preparation of this manuscript, William Clark and Jan Kramer for their technical expertise with polysomnography, Julian Maack for medical illustrations, Alan Abdulla, M. D., for his contributions in the initial studies, and Alan H. Morris, M.D., Robert O. Crapo, M.D., and Arthur S. Slutsky, M. D., for their critical reviews and suggestions.
REFERENCES 1 Sanders MH, Gmendl CA, Rogers RM. Patient compliance with nasal CPAP therapy for sleep apnea. Chest 1986; 90:330-37 ITO. NasalCPAPand NasalO2 in OSA (Farney et 8/)
2 Waldorn RE, Herrick rw Nguyen MC, O'Donnell AE, Sodero J, Potolicchio SJ. Long-term compliance with nasal continuous positive airway pressure therapy of obstructive sleep apnea. Chest 1990; 97:33-8 3 Martin RJ, Sander MH, Gray BA, Pennock BE~~ute and.longterm ventilatory effects of hyperoxia in the adult sleep apnea syndrome. Am Rev Respir Dis 1982; 125:175-80 4 Smith PL, Haponik EF, Bleecker ER. The effects of oxygen in patients with sleep apnea. Am Rev Respir Dis 1984; 130:958-63 5 Gold AR, Schwartz AR, Bleecker ER, Smith PL. The effects of chronic nocturnal oxygen administration upon sleep apnea. Am Rev Respir Dis 1986; 134:925-29 6 Heimlich H, Carr G. Transtracheal catheter technique for pulmonary rehabilitation. Ann Otol Rhinol Laryngol 1985; 94:502-04 7 Heimlich H. Respiratory rehabilitation with transtracheal oxygen system. Ann Otol Rhinol Laryngoll982; 91:643-47 8 Christopher KL, Spofford BT, Petrun MD, McCarty DC, Goodman JR, Petty TL. A program for transtracheal oxygen delivery: assessment of safety and efficacy. Ann Intern Med 1987; 107:802-oB 9 Berssenbrugge A, Dempsey J, Iber C, Skatrud J, Wilson ~ Mechanisms of hypoxia-induced periodic breathing during sleep in humans. J Physioll983; 343:507-24 10 Phillipson EA. Control of breathing during sleep. Am Rev Respir Dis 1978; 118:909-39 11 Khoo MCK, Kronauer RE, Strohl K~ Slutsky AS. Factors inducing periodic breathing in humans: a general model. J Appl Physiol: Respir Environ Exercise Physiol 1982; 53:644-59 12 Farney RJ, Walker LE, Jensen RL, Walker JM. Ear oximetry to detect apnea and differentiate rapid eye movement (REM) and
non-REM (NREM) sleep. Chest 1986; 89:533-39 13 Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. In: Brain information service/brain research institute. LosMgeles: University of California, 1968 14 Bornstein SK. Respiratory monitoring during sleep: polysomnography. In: Guilleminault C, 00. Sleeping and waking disorders: indications and techniques. Menlo Park, Calif: AddisonWesley Publishing Co, 1982:183-212 15 Spofford BT, Christopher KL, Hoddes ES. Transtracheal oxygen therapy for obstructive sleep apnea [abstract]. Chest 1986; 89:484(S) 16 Chauncey }B, Aldrich MS. Preliminary findings in the treatment of obstructive sleep apnea with transtracheal oxygen. Sleep 1990; 13:167-74 17 Elmer JC, Farney RJ, Walker JM, 000 RJ, Viscomi VA. The comparison of transtracheal oxygen with other therapies for obstructive sleep apnea. Am Rev Respir Dis 1988; 137:311A 18 Hoffstein ~ Zamel N, Phillipson E. Lung volume dependence of pharyngeal cross-sectional area in patients with obstructive sleep apnea. Am Rev Respir Dis 1984; 130:175-78 19 Hudgel D~ Hendricks C, Dooley A. Alteration in obstructive apnea pattern induced by changes in oxygen- and carbondioxide-inspired concentrations. Am Rev Respir Dis 1988; 138:16-9 20 Motta J, Guilleminault C. Effects of oxygen administration in sleep-induced apneas. In: Guilleminault C, Dement ~ eds. Sleep apnea syndrome. New York: Alan R Liss Inc, 1978:137-44 21 Berthon-Jones M, Sullivan CEo lime course of change in ventilatory response to CO 2 with long-term CPAP therapy for obstructive sleep apnea. Am Rev Respir Dis 1987; 135:144-47
CHEST I 101 I 5 I MA'Y, 1992
1235