Effect of Uvulopalatopharyngoplasty on Upper Airway Collapsibility in Obstructive Sleep Apnea 1- 3
ALAN R. SCHWARTZ,4 NORMAN SCHUBERT, WARREN ROTHMAN, FREDERICK GODLEY, BERNARD MARSH, DAVID EISELE, JOHN NADEAU, LISA PERMUTT, lAIN GLEADHILL, PHILIP L. SMITH
Introduction SUMMARY Previous Investigators hlMl demonstrated variable responses to uvulopalatopharyn-
Obstructive sleep apnea is a common goplasty (UPP) In patients with obatructlva sleep apnea. We hypothesized that this variability Is disorder of middle-aged obese men (1), due to either (1) differences In baseline pharyngeal collapsibility preoperatlvaly or (2) differences and it is caused by recurrent obstruction In magnitude of the decrease In pharyngeal collapsibility resulting from surgery. To determine the of the upper airway during sleep (2). To relationship between changes In collapsibility and the response to UPP aurgery, wa meaaured the upper airway critical pressure (pcrlt) before and after UPP In 13 patients with obstructlva sleep apalleviate upper airway obstruction durnea. During non-REM sleep, maximal Inspiratory airflow (Vlmax) was quantitated by varying the ing sleep, various therapeutic strategies, IlMIl of nasel pressure (PH), and Perlt was determined by the leval of PH below which Vlmax cessed. including uvulopalatopharyngoplasty A posltlva response to UPP was defined by a ~ 50% fall In non-REM disordered breathing rate (DBR). (UPP), have been adopted. UPP remains In the entire group, UPP resulted In significant decreases In DBR from 71.1 ± 22.4 to 44.7 ± 38.4 an attractive therapeutic option since eplsodeslh (p = 0.025) and In Perlt from 0.2 ± 2.4 to -3.1 ± 5.4 em H20 (p = 0.016). Moreover, definitive treatment of obstructive sleep the percent change In DBR was correlated significantly with the change In Perlt (p = 0.001). Subapnea can be achieved with this procegroup analysis of responders end nonresponders demonatrated that significant differences In Perlt dure (3). Unfortunately, favorable reware confined to the responders. Specifically, responders demonatrated a significant fell In Perlt sponses to UPP have been documented from -0.8 ± 3.0 to -7.3 ± 4.9 em H 20 (p = 0.01), whereas no significant change In Perlt was detectin only approximately half 0 f patients in ed In the nonresponders (1.1 ± 1.6 varsus 0.6 ± 2.0 em H 20). No clinical, poIysomnographlc, or physiologic predictors of a fevorable response were found preoperatlvaly. We conclude that the unselected case series reported in the literresponse to UPP Is determined by the magnitude of the fall In Perlt rather than by the Initial preoperaature (4, 5). The reason for this variabilitlva leval of Perlt. SlMIral mechanisms to account for variability In the response of Perlt to surgery ty in the response to UPP, however, is are suggested. AM REV RESPIR DIS 1992; 145:527-532 unclear. Previous studies in our laboratory have demonstrated elevations in the upper airway critical pressure (Pcrit) in patients otolaryngologist if he so chose. Patients who with obstructive sleep apnea, reflecting Methods elected UPP returned to the Sleep Center increases in upper airway collapsibility Subject Selection (6-8). Moreover,wehaverecently demon- Patients with obstructive apneas and hypop- preoperatively for another sleep study to destrated that apnea resolves after weight neas referred to the Johns Hopkins Sleep Dis- termine upper airway collapsibility (Pcrit). loss when Pcrit falls to the moderately order Center were recruited for this study. UPP was then performed by staff otolaryngologists at the Johns Hopkins Medical Innegative levels that have been observed Eligible patients demonstrated a disordered stitutions who employed standard techniques of greater than 10 obbreathing rate (DBR) in asymptomatic snorers (9). These finddescribed by Fujita and coworkers (3) and ings suggest that obstructive sleep apnea structive hypopneic and apneic episodes per Simmons and colleagues (5). Patients were hour of nonrapid eyemovement (non-REM) would be expected to remit only in those allowed a minimum of 2 months recuperapatients in whom Pcrit falls below these sleep on an overnight screening sleep study. tion. Patients then repeated a postoperative Patients with concomitant medical illness on overnight screening sleep study to determine moderately negative levels after UPP. screening history, physical examination, or Therefore, favorable responses to UPP pulmonary function testing were excluded. would be most likely to occur when ei- Written informed consent for this study was (Received in original form February 25, 1991 and ther (1) a relatively low initial Pcrit pre- obtained from each patient. This study was in revised form August 26, 1991) disposes to reductions in Pcrit to moder- approved by the Johns Hopkins Medical ately or markedly negative levels after Institutions Human Investigations Review From the Department of Medicine, Pulmonary UPP or (2) the magnitude of the fall in Board. Division, and the Department of Otolaryngology, Johns Hopkins Medical Institutions, Baltimore, Pcrit with surgery is quite large. ThereStudy Design Maryland. fore, we hypothesized that patients who 2 Supported by Grant HL-37379-Q2 from the Narespond to UPP should be characterized Toexaminethe effect of UPP on apnea severi- tional Institutes of Health. ty and upper airway collapsibility, patients by either lower levels of Pcrit preopera- with newly diagnosed, predominantly obJ Correspondence and requests for reprints tively or by greater reductions in Pcrit af- structive apneas referred to the Johns Hop- should be addressed to Alan R. Schwartz, M.D., ter surgery. To test this hypothesis, weex- kins Sleep Disorder Center were counseled Division of Pulmonary and Critical Care Medicine, Johns Hopkins Asthma & Allergy Center, 301 amined how Pcrit and apnea severity by a staff physician at the Sleep Center. Each Bayview Boulevard, Baltimore, MD 21224. responded to UPP in 13patients with ob- patient was offered consultation with an Recipient of Clinical Investigator Award HLotolaryngologist or referred back to his 02031-QIAI from the National Institutes of Health. structive sleep apnea. 1
4
527
528
SCHWARTZ, SCHUBERT, ROTHMAN, E1 AL.
apnea severity and another sleep study to measure Perit.
Experimental Techniques Polysomnography, Standard polysomnographic techniques (10)were employed to define the respiratory pattern during sleep, and they are outlined briefly here. Surface C 3-A, and C 3 -Q, electroencephalographic electrodes, submental electrodes, and left and right electrooculograms were used to stage sleep. An ear oximeter (Biox Ill; Bioximetry Technology, Boulder, CO) recorded oxyhemoglobin saturation (Sao,), and heart rate and rhythm were monitored. Oronasal airflow was monitored with a thermistor, and respiratory effort and pattern were monitored with thoracic and abdominal strain gauges. A polygraph (No. 780; Grass Instruments, Quincy, MA) ran continuously at 10 mmls to simultaneously record all physiologic data throughout the night. Nocturnal sleep studies were performed between the hours of ll:oo P.M. and 8:00 A.M. Polysomnograms were scored for disordered breathing episodes (either apneas or hypopneas) and for changes in Sao" as previously described (10). In each patient, the disordered breathing rate (DBR) was calculated as the number of disordered breathing episodes (either obstructive apneas or hypopneas) per hour of nonrapid eye movement (non-REM) sleep. The mean non-REM apnea time to disordered breathing time ratio (AT/DBT) was defined by the duration of apnea during disordered breathing episodes. The baseline and nadir Sao, were measured for non-REM disordered breathing episodes to calculate the average baseline and average low Sao,. Upper airway pressure-flow relationships. In another second overnight sleep study, upper airway pressure-flow relationships were recorded, as previously described (6-8). Briefly, each subject was studied in the supine po- . sition and breathed through a tight-fitting nasal mask that was connected to a variable pressure source at the inflow port and a variable resistor at the outflow port. Pressure in the nasal mask (PN) was measured with a catheter connected to a port in the mask. Esophageal pressure was measured with a lO-cm latex balloon catheter, which was passed perinasally and positioned 10 ern above the gastroesophageal junction. Nasal inspiratory airflow (VI) was measured with a pneumotachograph, which was inserted in the breathing circuit. A thermistor was placed at the mouth, and patients were monitored visually to assure that oral breathing did not occur. All pressures, flows, and polysomnographic parameters wererecorded continuouslyon the polygraph recorder. Patients were allowed to initiate sleep with PN maintained at approximately I em H,O. After the initiation of stable non-REM sleep, PN was varied stepwise every 15 min over a range that included the PN at which maximal inspiratory airflow (Vwax) became zero,
250 200
", max
150
(mils)
100
50
o +-
ALAN R. SCHWARTZ,4 NORMAN SCHUBERT, WARREN ROTHMAN, FREDERICK GODLEY, BERNARD MARSH, DAVID EISELE, JOHN NADEAU, LISA PERMUTT, lAIN GLEADHILL, PHILIP L. SMITH
Introduction SUMMARY Previous Investigators hlMl demonstrated variable responses to uvulopalatopharyn-
Obstructive sleep apnea is a common goplasty (UPP) In patients with obatructlva sleep apnea. We hypothesized that this variability Is disorder of middle-aged obese men (1), due to either (1) differences In baseline pharyngeal collapsibility preoperatlvaly or (2) differences and it is caused by recurrent obstruction In magnitude of the decrease In pharyngeal collapsibility resulting from surgery. To determine the of the upper airway during sleep (2). To relationship between changes In collapsibility and the response to UPP aurgery, wa meaaured the upper airway critical pressure (pcrlt) before and after UPP In 13 patients with obstructlva sleep apalleviate upper airway obstruction durnea. During non-REM sleep, maximal Inspiratory airflow (Vlmax) was quantitated by varying the ing sleep, various therapeutic strategies, IlMIl of nasel pressure (PH), and Perlt was determined by the leval of PH below which Vlmax cessed. including uvulopalatopharyngoplasty A posltlva response to UPP was defined by a ~ 50% fall In non-REM disordered breathing rate (DBR). (UPP), have been adopted. UPP remains In the entire group, UPP resulted In significant decreases In DBR from 71.1 ± 22.4 to 44.7 ± 38.4 an attractive therapeutic option since eplsodeslh (p = 0.025) and In Perlt from 0.2 ± 2.4 to -3.1 ± 5.4 em H20 (p = 0.016). Moreover, definitive treatment of obstructive sleep the percent change In DBR was correlated significantly with the change In Perlt (p = 0.001). Subapnea can be achieved with this procegroup analysis of responders end nonresponders demonatrated that significant differences In Perlt dure (3). Unfortunately, favorable reware confined to the responders. Specifically, responders demonatrated a significant fell In Perlt sponses to UPP have been documented from -0.8 ± 3.0 to -7.3 ± 4.9 em H 20 (p = 0.01), whereas no significant change In Perlt was detectin only approximately half 0 f patients in ed In the nonresponders (1.1 ± 1.6 varsus 0.6 ± 2.0 em H 20). No clinical, poIysomnographlc, or physiologic predictors of a fevorable response were found preoperatlvaly. We conclude that the unselected case series reported in the literresponse to UPP Is determined by the magnitude of the fall In Perlt rather than by the Initial preoperaature (4, 5). The reason for this variabilitlva leval of Perlt. SlMIral mechanisms to account for variability In the response of Perlt to surgery ty in the response to UPP, however, is are suggested. AM REV RESPIR DIS 1992; 145:527-532 unclear. Previous studies in our laboratory have demonstrated elevations in the upper airway critical pressure (Pcrit) in patients otolaryngologist if he so chose. Patients who with obstructive sleep apnea, reflecting Methods elected UPP returned to the Sleep Center increases in upper airway collapsibility Subject Selection (6-8). Moreover,wehaverecently demon- Patients with obstructive apneas and hypop- preoperatively for another sleep study to destrated that apnea resolves after weight neas referred to the Johns Hopkins Sleep Dis- termine upper airway collapsibility (Pcrit). loss when Pcrit falls to the moderately order Center were recruited for this study. UPP was then performed by staff otolaryngologists at the Johns Hopkins Medical Innegative levels that have been observed Eligible patients demonstrated a disordered stitutions who employed standard techniques of greater than 10 obbreathing rate (DBR) in asymptomatic snorers (9). These finddescribed by Fujita and coworkers (3) and ings suggest that obstructive sleep apnea structive hypopneic and apneic episodes per Simmons and colleagues (5). Patients were hour of nonrapid eyemovement (non-REM) would be expected to remit only in those allowed a minimum of 2 months recuperapatients in whom Pcrit falls below these sleep on an overnight screening sleep study. tion. Patients then repeated a postoperative Patients with concomitant medical illness on overnight screening sleep study to determine moderately negative levels after UPP. screening history, physical examination, or Therefore, favorable responses to UPP pulmonary function testing were excluded. would be most likely to occur when ei- Written informed consent for this study was (Received in original form February 25, 1991 and ther (1) a relatively low initial Pcrit pre- obtained from each patient. This study was in revised form August 26, 1991) disposes to reductions in Pcrit to moder- approved by the Johns Hopkins Medical ately or markedly negative levels after Institutions Human Investigations Review From the Department of Medicine, Pulmonary UPP or (2) the magnitude of the fall in Board. Division, and the Department of Otolaryngology, Johns Hopkins Medical Institutions, Baltimore, Pcrit with surgery is quite large. ThereStudy Design Maryland. fore, we hypothesized that patients who 2 Supported by Grant HL-37379-Q2 from the Narespond to UPP should be characterized Toexaminethe effect of UPP on apnea severi- tional Institutes of Health. ty and upper airway collapsibility, patients by either lower levels of Pcrit preopera- with newly diagnosed, predominantly obJ Correspondence and requests for reprints tively or by greater reductions in Pcrit af- structive apneas referred to the Johns Hop- should be addressed to Alan R. Schwartz, M.D., ter surgery. To test this hypothesis, weex- kins Sleep Disorder Center were counseled Division of Pulmonary and Critical Care Medicine, Johns Hopkins Asthma & Allergy Center, 301 amined how Pcrit and apnea severity by a staff physician at the Sleep Center. Each Bayview Boulevard, Baltimore, MD 21224. responded to UPP in 13patients with ob- patient was offered consultation with an Recipient of Clinical Investigator Award HLotolaryngologist or referred back to his 02031-QIAI from the National Institutes of Health. structive sleep apnea. 1
4
527
528
SCHWARTZ, SCHUBERT, ROTHMAN, E1 AL.
apnea severity and another sleep study to measure Perit.
Experimental Techniques Polysomnography, Standard polysomnographic techniques (10)were employed to define the respiratory pattern during sleep, and they are outlined briefly here. Surface C 3-A, and C 3 -Q, electroencephalographic electrodes, submental electrodes, and left and right electrooculograms were used to stage sleep. An ear oximeter (Biox Ill; Bioximetry Technology, Boulder, CO) recorded oxyhemoglobin saturation (Sao,), and heart rate and rhythm were monitored. Oronasal airflow was monitored with a thermistor, and respiratory effort and pattern were monitored with thoracic and abdominal strain gauges. A polygraph (No. 780; Grass Instruments, Quincy, MA) ran continuously at 10 mmls to simultaneously record all physiologic data throughout the night. Nocturnal sleep studies were performed between the hours of ll:oo P.M. and 8:00 A.M. Polysomnograms were scored for disordered breathing episodes (either apneas or hypopneas) and for changes in Sao" as previously described (10). In each patient, the disordered breathing rate (DBR) was calculated as the number of disordered breathing episodes (either obstructive apneas or hypopneas) per hour of nonrapid eye movement (non-REM) sleep. The mean non-REM apnea time to disordered breathing time ratio (AT/DBT) was defined by the duration of apnea during disordered breathing episodes. The baseline and nadir Sao, were measured for non-REM disordered breathing episodes to calculate the average baseline and average low Sao,. Upper airway pressure-flow relationships. In another second overnight sleep study, upper airway pressure-flow relationships were recorded, as previously described (6-8). Briefly, each subject was studied in the supine po- . sition and breathed through a tight-fitting nasal mask that was connected to a variable pressure source at the inflow port and a variable resistor at the outflow port. Pressure in the nasal mask (PN) was measured with a catheter connected to a port in the mask. Esophageal pressure was measured with a lO-cm latex balloon catheter, which was passed perinasally and positioned 10 ern above the gastroesophageal junction. Nasal inspiratory airflow (VI) was measured with a pneumotachograph, which was inserted in the breathing circuit. A thermistor was placed at the mouth, and patients were monitored visually to assure that oral breathing did not occur. All pressures, flows, and polysomnographic parameters wererecorded continuouslyon the polygraph recorder. Patients were allowed to initiate sleep with PN maintained at approximately I em H,O. After the initiation of stable non-REM sleep, PN was varied stepwise every 15 min over a range that included the PN at which maximal inspiratory airflow (Vwax) became zero,
250 200
", max
150
(mils)
100
50
o +-
