Hypercapnia in the Obstructive Sleep Apnea Syndrome

~

illlll Clinical in,ellia_iOns Hypercapnia in the Obstructive Sleep Apnea Syndrome* A Reevaluation of the "Plckwlckian Syndrome" David M. Rapoport, M.D., EC.C.R;t Stuart M. Garay, M.D.; Hadassah Epstein;§ and Roberta M. Goldring, M.D.II

*

The mechanisms of hypercapnia in eight patients with the "Piekwickian" syndrome and obstructive sleep apnea (OSAS) were evaluated pretherapy and posttherapy (tracheostomy in seven patients and chronic nocturnal use of nasal CPAP in one). Four patients (correctors) became eucapnic within two weeks of therapy. Four others (noneorrectors) remained hypercapnic. Neither residual apneas, changes in pulmonary function, change in anatomic dead space, nor changes in ventilatory chemoresponsiveness differentiated the two groups, nor did the last three factors account for return to eucapnia in the correctors. The results

indicated two separate mechanisms exist for chronic hypercapnia in OSAS:1)a critical balance between the ventilation during the time spent awake and hypoventilation due to apneas, a mechanism removed by treatment for obstructive apnea; and 2) sustained hypoventilation independent of the apnea phenomenon and therefore not correctible. The subset of patients with the second mechanism appears to represent the true "Pickwicldan" syndrome and can be identi6ed before therapy by measuring a low level of ventilation in the sustained awake state.

syndrome of obesity, hypersomnolence, and T hehypercapnia, known as the "Pickwickian" syn-

thus for acute hypercapnia), chronic waking hypercapnia remains unexplained by this mechanism." Furthermore, hypercapnia is not present in most patients with the obstructive sleep apnea syndrome while they are awake." This dilemma is highlighted by previous results from our laboratory demonstrating that patients with severe obstructive sleep apnea and chronic hypercapnia need not differ from eucapnic patients with the syndrome in number of apneas, their duration, the degree of obesity, or the clinical severity of their hypersomnolence." Thus, the mechanism of the chronic hypercapnia in the "Piekwickian" patient with obstructive sleep apnea remains to be fully elucidated. The present study was designed: (1) to define the contribution of apnea to the etiology of chronic hypercapnia in eight "Piekwickian" patients with the obstructive sleep apnea syndrome by following the response of their chronic hypercapnia to correction of the apnea phenomenon, and (2) to characterize other mechanisms of their hypercapnia before and after therapy of the apnea.

drome, has remained the focus of much physiologic and clinical interest since its first description by Burwell et ale 1 Most prior reports have shown a low minute ventilation to be characteristic of the hypercapnia. However, the reasons for this chronic hypoventilaFor editorial comment see page 621 tion and an accurate definition of the patterns of breathing associated with it remain incomplete. Initial speculation with respect to its etiology focused on diminished central responses to ventilatory stimuli! and mechanical chest wall abnormalities of the obese. 3 More recently, it has become clear that patients with the "Piekwickian" syndrome usually also have sleeprelated periodic functional obstruction of the upper airway-the obstructive sleep apnea phenomenon." While the occurrence ofobstructive apnea provides for a potential reduction of ventilation while asleep (and *From the Department of Medicine, Clinical Research Unit, and Chest Service, New York University-Bellevue Medical Center, New York. This study was supported by grants from the National Institutes of Health, Division of Research Resources RR 96 and from the New YorkLung Association. t Assistant Professor of Medicine. iAssistant Professor of Clinical Medicine. §Research Assistant. IIProfessor of Medicine. Manuscript received September 16;revision accepted November 21. Reprint requests: Dr. Rapoport, Department ofMedicine, New York University Medical Center, 550 First Avenue, New York City 10016

METHODS

Eight chronically hypercapnic patients with obstructive sleep apnea syndrome were evaluated before and after therapy for their apneas. Therapy consisted of permanent tracheostomy in seven patients; one patient was successfully treated with nightly use of a nasal mask which delivered continuous positive airway pressure (nasal CPAP). In each case, the patient was able to reduce hisarterial Pco, below 40 mm Hg by voluntary hyperventilation. All patients were obese and hypersomnolent. Six of the eight had had multiple CHEST I 89 I 5 I MAY, 1988

827

hospitalizations characterized by cardiorespiratory failure and edema. 1Wo patients had additional symptoms of chronic bronchitis. Each patient was in a clinically stable state for at least two months at the time of pretherapy study. All patients bad standard polysomnographic sleep monitoring to establish the diagnosis of obstructive sleep apnea. This included EEG, EOG, and chin EMG for sleep staging, nasal Pco, probe and oral thermistor to document air movement, detection of chest and abdominal movement by inductance plethysmography, ECG, and measurement ofarterial 0. saturation by ear oximetry. Ear oximeter saturations below 50 percent were reported as 50 percent due to the known alinearity of the instrument in this range. In two patients, an esophageal balloon was used to document persistent respiratory effort during apneas. Obstructive apnea wasdefined as a cessation of air movement lasting more than 10 seconds associated with persistent respiratory eflOrts. Apneas which began without respiratory eftOrts but during which efforts developed (mixed apneas) were also classified as obstructive. Central apnea was defined as cessation of both air movement and respiratory effort for more than 10 seconds.

Pulmonary FunctWn All patients bad standard tests to determine pulmonary function both before and after therapy for the obstructive sleep apnea. These included lung volumes by spirometry and body plethysmography, maximal expiratory flow volume loop, and multiple arterial blood gas determinations while awake. In addition, minute ventilation and pattern of respiration were determined while awake in the supine position using a three minute collection of expired gas in a TIssot spirometer. From analysis of the mixed expired gases, CO. production (YeoJ and physiologic dead space were calculated. In the patients with tracheostomy, posttreatment testing was performed with a plugged tracheostomy button which restored the upper airway anatomy to its pretracheostomy state. All tests were repeated one, six, and 12 months after therapy.

Ventilat01ll Responses to Hypercapnia and HflPoxia Ventilatory responses and pressure generated in the first 100 msec of an occluded breath (Po.J were measured during duplicate runs of progressive hypercapnia induced by rebreathing CO. on a modified circuit as described by Reed' and Whitlaw et al." Responses are reported as the slope of the best fit linear regression of ventilation and occlusion pressures to end tidal Pco, (~VEI~PC02 and ~PO.lI 6PeOt, respectively). Ventilatory response to isocapnic progressive hypoxia was measured in duplicate using the rebreathing circuit of Rebuck and Campbell" and is reported as the slope of the best fit linear regression of ventilation to 0 1 saturation, measured by ear oximetry (~VEI~O.sat). Tests were repeated before therapy on at least two occasions separated by at least one month. Because results ditTered by less than 10 percent they were averaged for each patient. Tests were then repeated one, six, and 12 months after therapy. After tracheostomy, tests were always performed through a mouthpiece, with the tracheostomy plugged using a tracheostomy button. Sleep Ventilatory Monitoring

In addition to the diagnostic sleep study, a one hour daytime sleep study was performed with a canopy ventilation monitor. U This allowed noninvasive quantitative measurement of ventilation, tidal volume, and frequency, both during wakefulness and sleep. No effort was made to sleep deprive the patients, as all were severely hypersomnolent and slept easily during testing. The canopy study was performed in association with sleep stage monitoring as above, ECG, inductance plethysmography, and ear oximetry. Sleep ventilatory studies were again performed four to six weeks after treatment. In the seven patients with tracheostomy, an all night sleep study was performed using a cuffed tracheostomy tube connected to a pneumotachograph to quantitatively monitor ventilation. The one patient treated with nightly home use of nasal CPAP

128

necessarily had a slightly modified followup study: variations in the pressure and flow through his mask were used as markers of apnea. Quantitative sleep ventilatory measurements were not possible while he was receiving therapy because of the mask. All ventilation measurements (awake and asleep) were expressed as liters/minute/meter of body surface area (BSA) in order to normalize for the differing body size of the patients. This normalization was not applied to CO. and hypoxic rebreathing ventilatory responses.

Treatment Seven of the eight patients underwent creation of a permanent tracheostomy with an H-shaped incision in the trachea and skin to create overlapping flaps of epithelial tissue lining the stoma," After two weeks of healing, a plastic cannula was inserted which did not extend past the stoma track, and which could be plugged during the daytime. The patients were instructed to remove the plug at night and during any naps to bypass the functional upper airway obstruction; they left the plug in during waking hours, which allowed them to speak and cough normally. In one case, a stricture developed above the tracheostomy site due to a wound infection; this patient was unable to use a button after an initial period of success and has continued to use a standard metal tracheostomy tube open at all times. The patient treated with nasal CPAP uses a home compressor delivering 45 L per minute of compressed heated humidified air to a mask modified from a cuffed pediatric anaesthesia mask, and attached to an overflow valve (PEEP valve) which maintains a pressure in the entire system of 12 to 14 cm H 20 . His therapy and response have been reported previously. 13 Posttreatment I nteroentWns

Posttreatment residual chronic hypercapnia was evaluated six months after treatment by temporarily returning the patients arterial Pco, and serum bicarbonate to normal values. This was accomplished either by elective mechanical ventilation with a respirator attached to a cuffed tracheostomy tube for 48 hours, or by administering an oral dose of 250 mg of acetazolamide once daily for one to two weeks. Arterial blood gas values, ventilation, and ventilatory response to rebreathing CO 2 were then retested. At a separate time, also at least six months after definitive treatment of the obstructive sleep apnea, medroxyprogesterone (20 to 40 mg sublingually three times daily for one to two months) was given to the same patients in an attempt to produce central respiratory stimulation. RESULTS

Table 1 lists anthropometric data and arterial blood gas levels of the patients before treatment, establishing the degree of chronic hypercapnia. Data from the pretreatment all-night sleep studies are also shown, documenting that each patient had severe obstructive sleep apnea. 'Iraeheostomy and nasal CPAP resulted in relief of hypersomnolence and disappearance of snoring in all cases. Peripheral edema disappeared or decreased in those in whom it was present. Sleep pattern before treatment was severely fragmented in all eight patients. Short periods of predominantly stage 2 sleep ('cmicrosleeps") were interrupted continuously every 30 to 120 seconds by arousals. These "microsleeps," defined as 10 to 60 second EEG evidence of sleep (ie, dropout of alpha waves and/or appearance of sleep spindles), usually occurred in Hypercapnia and 0bstructNe Sleep Apnea (Rapoport st 81)

Table l-Claaracterimtion ofPotien,. Lowest Sleep O 2 Saturation non-REM REM

Awake Arterial Blood Gas Values

Patient Correctors 1 2 3 4 Mean SD Noncorrectors 5 6 7 8 Mean SD P value*

Pco,

[HC03- ] (mEqlL)

Age

Weight (% Ideal)

BSA (m')

(mm Hg)

52 47 66 52 54.2 8.2

276 163 253 164 214.0 59.1

2.24 2.06 2.45 2.15 2.23 .17

45 48 52 50 48.8 3

28 28 31 29 29.0 1.4

53 67 57 45 55.5 9.2 NS

159 118 154 234 166.2 48.7 NS

2.28 1.72 2.00 2.82 2.21 .47 NS

52 55 49 49 51.2 2.9 NS

32

P0 2 (mm Hg)

Apnea Index (Eventslhr)

67 57 56 56 59.0 5.4

80 120 45 76.2 32.5

~2

57 50 75 61.0 10.6 NS

33 35

31 32.8 1.71 0.007

(%)

(%)

75 <50 50 <50

60

60 <50 <50 50

40 75 130 50 73.8 40.3 NS

55

<50 <50

<50 <50 <50

*For statistically signi6cant difference between correctors and noncorrectors.

association with an obstructive apnea; both apnea and sleep were terminated by an EEG arousal. Virtually no stage 3 to 4 non-REM or sustained REM sleep was present in any patient in the eight hours of monitoring. During the daytime nap studies, the EEG pattern was identical to the all-night studies in each patient, showing stage 1 to 2 sleep interrupted by arousals. For this reason, the quantitative measurements of ventilation during the nap were felt to be representative of events whenever "sleep" was present prior to treatment. After treatment (tracheostomy or on nasal CPAP), sleep showed prolonged uninterrupted periods of all stages including stages 3 to 4 and REM. Response of Hypercapnia to Treatment

The response of the patients' chronic hypercapnia to therapy defined two clearly separable groups (Fig 1). Four of the eight patients, including the one treated with nasal CPA~ showed a return to eucapnia docu55

--------_

..

mented within two weeks by a normalization of waking Pco, and return of the elevated serum bicarbonate to normal (correctors). This has persisted for the period of follow-up (28 to 60 months). The other four patients remained hypercapnic, despite a similar degree of improvement in the symptoms of sleep apnea (noncorrectors). This hypercapnia has persisted to the present (follow-up of 16 to 48 months). Analysis of Mechanisms of Hypercapnia Pretreatment and Posttreatment

Transient weight loss occurred in two noncorrectors and three correctors, who lost 2.3 to 13.6 kg acutely with tracheostomy, reflecting a spontaneous posttherapy diuresis and/or anorexia. The weight loss was not sustained and did not relate to correction of hypercapnia. In fact, the greatest weight loss (13.6 kg) occurred in one of the noncorrectors. Sleep-Related Apneas: Figure 2 shows the apnea 40

,..,

50

35

I

(T)

N

a u Q..

L L

.0

8~

~,

45

E ::J lCI

40 35

(J)

Pre-Treatment

Post-TrQatmQnt

cr ~

30

25

20

Pre-Treatment

Post-Treatment

FIGURE 1. Response of chronic waking hypercapnia and serum bicarbonate to treatment. Open circles represent correctors, closedcircles, noncorrectors. CHEST I 89 I 5 I MAY, 1988

821

150

5o

~

Pulmonary Function: Table 2 shows values for representative pulmonary function tests before and after treatment in both groups. Six of the patients showed abnormalities characteristic of obesity (mildly reduced vital capacity characterized by reduced expiratory reserve and mildly reduced functional residual capacity). Two of the patients had obstructive airway disease (reduced forced expiratory volume in 1 second/forced vital capacity and increased functional residual capacity). No change in pulmonary function occurred after therapy in any patient acutely or long term. Both patients with airwaysdisease were correctors. In one of these patients, the physiologic dead space was elevated. It remained elevated after treatment despite his return to eucapnia. In the other seven patients, dead space was normal, confirming the clinical impression that significant lung disease did not playa role in their hypercapnia. Thus, improvement in pulmonary function did not account for return to eucapnia in the "correctors," nor did intrinsic lung disease or worsened pulmonary function after treatment account for persistent hypercapnia in the noncorrectors. CO 2 Production and Anatomic Dead Space: CO 2 production measured in the basal resting state did not change in the patients after therapy (277 mllmin vs 290 ml/min, p = NS). This is in accord with the lack of consistent weight change and excludes a change in CO 2 production as the cause of the change in chronic arterial Pco, at the same level of awake ventilation in the "correctors." This change in arterial Pco, without change in CO 2 production implies that a change in alveolar ventilation must have occurred. The measured physiologic dead space was not changed from the preoperative value in the seven patients with tracheostomy as long as their tracheostomy button was plugged and they breathed

100

<, (I

a ~ Q. <

50

O.....- - - - - - - -.....~-~-Pre-Treatment FIGURE 2. Response of apnea phenomenon to treatment. Symbols as defined in Figure 1.

index and lowest O2 saturations during a night before and after therapy. Apnea index before treatment did not differentiate the correctors from the noncorrectors. Apnea index dropped dramatically after treatment in all patients (from 40 to 130 per hour to 0 to 15 per hour, p
Table I-Pulmonary Function Data

Patient

Vital Capacity ('I> predicted)

Funct Resid Cap (If, predicted)

82 74

137 81

65 65

58 84

90.0 33.4

.61 .85 .81 .42 .67 .20

103 112 72 87.8 23.3

64

89 91 101 76 89.2 10.3

.73 .77 .84 .72 .77 .05

NS

NS

NS

Correctors

1 2 3 4 Mean SD Noncorrectors

5 6 7 8 Mean SD P value*

71.5 8.2

FEV/FVC

Physiologic Dead Space (ml) 256

116 108 167 161.8 68.0 93

151 144 174 140.5 34.2

NS

*For statistically significant difference between correctors and noncorrectors.

130

HypercapnIa and 0bstructiYe Sleep Apnea (Rapopotf et aI)

4

N

L.

3

8 h Q.~

~

"

'" c

UJ

2

i

.> " ~

o

.2

0..-.-.---------.. . . . ----Prca-Trcaatmcant

Post-Trcaatment

Post-Treat.ent 1.6

c:

o .,.. .-.J aL ., ::J

a H• .-.J (J), N

1.2

a

C

.8

o ,i ~

'"

UJ

.4

.>

.a

O.....-..._-~--....- - Post-Treatment Pre-Treatment FICURE 3. Ventilation and occlusion pressure responses to rebreathing COl and ventilation response to ,haded bar' represent the normal range in this laboratory estimated as one standard deviation around the mean (calculated using the arctangents of the response slopesll). hypoxia pretreatment and posttreatment. Symbols as in Figure 1. The

through their upper airway(pretreatment mean 154 CC to posttreatment mean 137 ml, p = NS by paired Students t-test). Allowing these patients to breathe through their tracheostomy lowered the measured dead space by 70±10 cc (mean±SD), which was attributed to a change in anatomic dead space. To further assess the significance of this change, each of the seven patients had 70 ml of dead space tubing attached to his tracheostomy for 60 hours to increase the total dead space approximately to its preoperative value. At the end of this period, arterial blood gas levels showed no change in Pco; In particular, the correctors had not redeveloped hypercapnia. Thus, variability in the change in the dead space due to tracheostomy did not explain changes in alveolar ventilation, and the change in anatomic dead space did not account for correction ofhypercapnia in the correctors. Ventilatory Responses: Figure 3 shows the ventilatory responses to COs rebreathing (4W/4Peo s, 4PO.1I4PcoJ and to hypoxia (4W/40 I sat) of the patients before and after treatment. The ventilatory response to CO 2 was low in all but one patient and was associated with a low occlusion pressure response, indicating low central drive was present. The noncor-

rectors did not have lower responses to either COlor hypoxia than the correctors. Furthermore, the correctors showed no improvement in ventilatory responses after treatment despite their improvement in arterial blood gas levels. Repeat testing up to four years (range 16 to 48 months) after therapy has shown no late improvement in COlor hypoxic responses in these patients. Thus, differences in ventilatory responses pretreatment or changes in responses posttreatment did not account for the differingresolution ofhypercapnia after therapy in the two groups. Ventilation: Figure 4 shows representative tracings from a patient studied before treatment. The tidal volume was obtained quantitatively with a ventilation monitor: If care was taken to measure ventilation only during EEG evidence of the sustained awake state (persistent alpha wave activity for at least two minutes and no microsleeps), a distinct pattern and level of ventilation was defined for each patient, which was not periodic (see representative tracing in Fig 4B). The numerical value of this ventilation was calculated and called the sustained awake ventilation. Whenever these severely hypersomnolent patients CHEST I 89 I 5 I MAY, 1988

131

A. Spontaneous Relaxed State

TIDM.\QJ.K

I

-- - - -- ----.

B. Sustained Awake State

5

MC.

...

~.

'~

TIDAL \0..11.£

FIGURE 4. Representative tracings obtained in the canopy ventilation monitor in one patient prior to therapy. A (upper), Spontaneous relaxed state: microsleeps and periodic arousals appear on EEG; ventilation is periodic with obstructive apneas during microsleeps. B (lower), Sustained awake state by EEG: regular ventilation without apneas,

were left unstimulated, they assumed a state of consciousness characterized by continuously alternating EEG evidence of microsleep and wakefulness. This pattern is illustrated in the EEG tracing in Figure 4A and is essentially identical to the pattern of EEG seen during the all-night recording of sleep. Ventilation during this state was always periodic in all eight patients. The periodic ventilation was associated with the EEG arousals between microsleeps: unobstructed respiration occurred during the part of the EEG cycle when there was a momentary arousal; obstructive apnea occurred during the microsleep part of the EEG cycle. Figure 5 summarizes the results of these pretreatment ventilatory measurements and demonstrates the relationship between the two ventilatory patterns seen in each patient. In the spontaneous relaxed state (with periodic breathing), ventilation was low in both "correctors" (range 2.1 to 3.1 L'min/m") and noncorrectors (range 1.7 to 2.4 Uminlm2) . Although the average ventilation was lower in the noncorrectors, there was significant overlap with the correctors. In contrast, the level of ventilation in the sustained EEG-documented awake state (when periodicity was not present) clearly differed between the two groups: in the normal range in the "correctors" (range 3 to 3.8 Uminlm2) ; and lowin the noncorrectors (range 2.1 to 2.9 Umin/ml ) . Thus, the ventilation in the sustained awake state distin132

guished the correctors from the noncorrectors before treatment and anticipated their return to eucapnia after treatment. After treatment of the OSAS, the relaxed state was one of sustained wakefulness as the patients were no longer hypersomnolent, and microsleeps did not ocCUI: Furthermore, periodic ventilation did not occur 5 4

c

.......,0

N E

........ ""-c 0

...,

c >

CI

•

""

3 2

o "Correctors"

"Non-Correctors"

FIGURE 5. Levels of ventilation seen in the pretreatment state in correctors (opencircles) and noncorrectors (closedcircles). For each group, mean ventilation in the spontaneous relaxed state (shaded bar) and mean ventilation in the sustained awake state (open bar) are shown. The dashedline at 3 Uminlml represents the lower limit of normal ventilation. Hypercapnia and ObstructiveSleep Apnea (Rapoport sf 81)

in Figure 7. All three patients had returned to the control level of chronic hypercapnia within one week of discontinuation of mechanical ventilation or acetazolamide. Medroxyprogesterone (given to three patients) resulted in no change in arterial Pco, (mean 53 mm Hg vs 52 mm Hg), awake minute ventilation (mean 2.8 L'min/m" vs 2.9 L'min/m"), or response to CO 2 rebreathing (0.4 Uminlmm Hg vs 0.3 Uminlmm Hg) while the patients were receiving the drug. Thus, neither imposed eucapnia nor medroxyprogesterone succeeded in permanently correcting chronic hypercapnia in the noncorrectors.

5

4

~;;;e

:;

..

3 2

==t

.. r:

O..L--------------Post-treatllant Pra-treatlll&nt

DISCUSSION

FIGURE 6. Level of ventilation seen in the sustained awake state pretreatment and posttreatment. Symbols as in Figure 1. For each patient. there is virtual identity in the value obtained. and correctors separate from non-correctors.

In the present study, tracheostomy or chronic nocturnal use of nasal CPAP produced complete correction of the obstructive sleep apnea syndrome in all eight hypercapnic patients. However; correction of the chronic hypercapnia occurred in only four of the eight. The association of chronic hypercapnia with obstructive sleep apnea is now well recognized," and a variable response of this hypercapnia after correction of the upper airway obstruction has been observed by others. 14-16 Possible explanations for this variability in return to eucapnia include the effects of nonuniform changes in weight, pulmonary function, or level of ventilatory chemoresponsiveness after therapy for the OSAS. In our eight patients, weight loss, change in pulmonary function, and dead space effects did not account for the variability in return to eucapnia after therapy for OSAS. All patients had CO 2 responses in the low range prior to therapy, and those who reversed their chronic hypercapnia (the correctors) did so without improvement of their CO 2 response. Improved CO 2 responsiveness has been suggested as a mechanism for the improvement of hypercapnia in several other stud-

during wakefulness nor regularly during sleep as before. Figure 6 examines this posttreatment awake ventilation and compares it to the sustained awake ventilation seen before treatment. For each patient, the posttreatment awake ventilation was identical to the pretreatment sustained awake ventilation. Thus, the ventilation in the sustained awake state was characteristic for each patient. It distinguished the correctors from the noncorrectors by being higher even before therapy. Furthermore, this ventilation remained unaffected by treatment of the OSAS and was independent of the effect of treatment on hypercapnia. Posttreatment Interoentions

Eucapnia was transiently induced in three noncorrectors: in two by the use of 48 hours of elective continuous mechanical ventilation until arterial Pco, and serum bicarbonate were normalized; in one by the use of acetazolamide for one week. Results are shown

40

,.. I

35

C"J

~:::

If

40

35~----------Before A'ter • •

30 25 20

Before

During

A't.,-

Acetazol,.l. Mtlchanlcal Y..,tllatlon

FIGURE 7. EJfect of inducing eucapnia in three noncorrectors six months after therapy for obstructive sleep apnea. One patient treated with one week of acetazolamide and two with 48 hours of mechanical ventilation. All patients returned to control level of hypercapnia within one week after the intervention was discontinued. CHEST I 89 I 5 I MAY, 1988

833

ies.12.1+16 However, the published data supporting this concept remain unconvincing as responses in these studies either were not measured posttherapy," demonstrated only a shift in intercept without a change in slope," or were reported as changing only within the normal range without ever being low 12 In contrast, and in accord with our results, an earlier study demonstrated low CO 2 responses in previously hypercapnic patients long after correction of the upper airway obstruction." The implications of these observations and our results are that low CO 2 response per se did not cause hypercapnia pretreatment, and that increased CO 2 responsiveness is not the mechanism of improvement in hypercapnia posttreatment. Low CO 2 responsiveness, however, may still play a pennissive role in the development of chronic hypercapnia under the stress of untreated apnea by blunting respiratory compensations to the apnea phenomenon in the interapnea periods. 7 In the absence of intrinsic lung disease, the degree of hypercapnia must be related to the overall level of ventilation. Our data on the ventilatory patterns seen before and after therapy suggest at least two different mechanisms may exist for low ventilation and chronic hypercapnia in the "Pickwickian" patient prior to treatment for obstructive sleep apnea. The first mechanism (demonstrated by the correctors) consisted of an unfavorable balance between ventilation during time awake-which was normaland hypoventilation during periods containing apnea. This mechanism is thus clearly dependent on the presence of the apnea phenomenon itself; these patients represent a form of the "Pickwiekian" syndrome that is due to an identifiable cause other than sustained hypoventilation in the waking state. The second mechanism (demonstrated by the noncorrectors) consisted of sustained hypoventilation present even when fully awake. This hypoventilation was independent of the apnea phenomenon. The noncorrectors thus appear to represent a true "Pickwickian" syndrome. They can be identified before therapy for obstructive sleep apnea by demonstrating a low ventilation even in the fully awake state, and this low ventilation appeared to be characteristic of these patients in that it persisted after treatment of the apnea. Furthermore, their posttreatment hypercapnia could not be reversed by correction of their elevated serum bicarbonate or by temporary hyperventilation, indicating that this state represented an intrinsic abnormality of ventilatory homeostasis, and emphasizing the fundamental difference between these patients and the correctors. Chronic hypercapnia is dependent on a critical balance between the time spent over the 24 hour period in three distinct states and on the ventilation in each state: the fully awake state, the sleeping state, and the intermediate state associated with periodic breath834

ing due to microsleeps and apneas. If all of these ventilations are low, then the time spent in each is no longer critical, and hypercapnia is explained by classic alveolar hypoventilation." This appeared to be the mechanism of both pretreatment and posttreatment hypercapnia in the noncorrectors in the present study. If the ventilation during periodic breathing is low but ventilation during the sustained awake state is normal, then hypercapnia is dependent on the relative amount of time spent in the two states. This appeared to have been the mechanism of pretreatment hypercapnia in the correctors, where the pretreatment balance must have been in favor of the low periodic ventilation due to the severe hypersomnolence. It should be noted, however, that while the value of the ventilation in the awake and the apnea-eontaining states was defined for each patient in the present study, the actual percentage of time over 24 hours spent awake vs the time in periods of apnea containing ventilation has not been measured in this or other studies. In contrast to the patients of this study, most patients with obstructive sleep apnea remain chronically eucapnic." Although less apnea may explain eucapnia in less severely affected individuals, a previous study from this laboratory showed that eucapnic patients with severe obstructive apnea may have the same number and duration of apnea as hypercapnic patients." The eucapnic patients in that study did however, show a "big breath" after their apneas and had higher CO 2 rebreathing responses than the hypercapnics. A normal CO 2 response may thus playa protective role against the development of chronic hypercapnia under the stress of apnea by stimulating respiratory compensation for each apnea in the interapnea period. Figure 8 shows a proposed overview of the relationship of hypercapnia in the absence of intrinsic lung

FIGURE 8. Schema summarizing the mechanisms which determine the level of chronic arterial Pco, in patients with obstructive sleep apnea. The groups shown in the upper panels are conceptual. Whether the alveolar hypoventilation schematized on the right is congenital, acquired, or related to the longstanding obstructive apnea remains to be determined.

Hypercapnia and Obstructive Sleep Apnea (Rapoport at 81)

disease (the "Piekwickian" syndrome) and the obstructive sleep apnea syndrome. It incorporates observations from our previous study and the present data. The upper left of the schema shows a eucapnic individual at risk of developing obstructive sleep apnea. When the stress of intermittent airway obstruction is imposed, these eucapnic patients can be subdivided according to their response. The majority will tend to compensate for each apnea with an increase in ventilation just after that apnea, yielding a net ventilation which is maintained in the normal range even during the periodic breathing, independently of the number of apneas. This ventilatory compensation during periodic breathing may be related to the normal CO2 rebreathing responses seen in these patients." In the second group of eucapnic patients, compensation for the stress of apnea is inadequate: they will not raise their ventilation between each apnea sufficiently to normalize their net ventilation. A low CO2 response may be permissive for this lack of compensation. 7 These patients thus develop prolonged periods of hypoventilation during their periodic breathing. Since ventilation in this group remains normal in periods free of apnea, successful treatment of the obstructive sleep apnea results in a return to eucapnia. On the right of the schema is shown a subset of the population with sustained hypoventilation and hypercapnia independent of apnea. The hypoventilation can only be worsened by the additional mechanical stress of obstructive apnea. The patients in this group appear to have a defect in control of ventilation independent of the apnea phenomenon and a lower set-point for ventilation; they are the true "Pickwickians" in that they represent a syndrome of chronic alveolar hypoventilation of unknown etiology in the obese patient without intrinsic lung disease. The significance of delineating this subset from other hypercapnic patients with OSAS lies in that they represent a different abnormality of ventilatory control from the latter; whose hypercapnia is a direct consequence of the recurrent apneas. Whether the defect of the true "Pickwickian" is acquired (perhaps due to some aspect of obesity or to years with severe apnea in some patients), whether it results from a congenital abnormality of ventilatory control, or whether it merely represents an extreme in the spectrum of the normal population remains to be determined.

REFERENCES 1 Burwell CS, Robin ED, Whaley RD, Bickelmann AG. Extreme obesity with alveolar hypoventilation: a Pickwickian syndrome. Am J Moo 1956; 21:811-18 2 Auchincloss JH Jr, Cook E, Renzetti AD. Clinical and physiological aspects of a case of obesity, polycythemia and alveolar hypoventilation. J Clin Invest 1955; 34:1537-45 3 Rochester OF, Enson 1: Current concepts in the pathogenesis of the obesity-hypoventilation syndrome: mechanical and circulatory factors. Am J Med 1974; 57:402-20 4 Gastaut H, Tassinari CA, Duron B. Polygraphic study of the episodic diurnal and nocturnal (hypnic and respiratory) manifestations of the Pickwick syndrome. Brain Res 1966; 2:167-86 5 Sharp TI: Barrocas M, Chokroverty S. The cardiorespiratory effects of obesity. Clin Chest Med 1980; 1:103-18 6 Guilleminault C, 11lkian A, Dement WC. The sleep apnea syndromes. Ann Rev Med 1976; 27:465-84 7 Garay SM, Rapoport OM, Sorkin B, Epstein H, Feinberg I, Goldring RM. Regulation of ventilation in the obstructive sleep apnea syndrome. Am Rev Respir Dis 1981; 124:451-57 8 Read DJC. A clinical method for assessing the ventilatory response to carbon dioxide. Aust Ann Med 1966; 16:20-32 9 WhitlawWA, Derenne J~ Milic-Emili J. Occlusion pressure as a measure of respiratory center output in conscious man. Respir Physioll975; 23:181-99 10 Rebuck AS, Campbell EJM. A clinical method for assessing the ventilatory response to hypoxia. Am Rev Respir Dis 1974; 109:345-50 11 Sorkin B, Rapoport OM, Falk DB, Goldring RM. Canopy ventilation monitor for quantitative measurement of ventilation during sleep. J Appl Physiol Respirat Environ Exer Physioll980; 48:724-30 12 Guilleminault C, Simmons FB, Motta], Cummiskey J, Rosekind M, Schroeder JS, et ale Obstructive sleep apnea and tracheostomy: long-term follow-up experience. Arch Intern Med 1981; 141:985-88 13 Rapoport OM, Sorkin B, Garay SM, Goldring RM. Reversal of the "Pickwickian syndrome" by long-term use of nocturnal nasalairway pressure. N Engl J 1982, 307:931-33 14 Aubert-Thlkens G, Willems B, Veriter C, Coche E, Stanescu DC. Increase in ventilatory response to C02 following tracheostomy in obstructive sleep apnea. Bull Europ Physiopath Resp 1980; 16:587-93 15 Farmer WC, Glenn WWL, Gee JDL. Alveolar hypoventilation syndrome: studies of ventilatory control in patients selected fOr diaphragmatic pacing. Am J Moo 1978; 64:39-49 16 Sullivan CE, Berthon-Jones M, Issa FG. Remission of severe obesity-hypoventilation syndrome after short-term treatment during sleep with nasalcontinuous positive airway pressure. Am Rev Respir Dis 1983; 128:177-81 17 Ingram RH, Bishop JB. Ventilatory response to carbon dioxide after removal of chronic upper airway obstruction. Am Rev Respir Dis 1970; 102:645-47 18 Fishman ~ Goldring RM, Turino GM. General alveolar hypoventilation: a syndrome of respiratory and cardiac failure in patients with nonnallungs. Q J Med 1966; 35:261-75

Moo

CHEST I 89 I 5 I MAY, 1988

835