Voluntary Hyperventilation in Obesity Hypoventilation*

Transcription

1 Voluntary Hyperventilation in Obesity Hypoventilation* Judith l ech, M.D., F.G.G.R;t Ergiin Onal, M.D., F.G.G.R; Robert Aronson, M.D., F.G.G.R; and Melvin lajpata, M.D., F.G.G.R Arterial blood gas analysis was performed before and after 6 to 9 s of voluntary hyperventilation in 27 consecutive patients with occlusive sleep apnea syndrome (OSA) and daytime hypercapnia. The percentage offall in PaCO I from baseline was examined in relationship to age, body mass index, sleep-disordered breathing indices, and pulmonary function variables. In 14 subjects without airftow obstruction, only one individual could not voluntarily hyperventilate into the normal range, whereas 6 of 13 subjects with airftow obstruction could not hyperventilate to eucapnia. The average percentage of fall in PaCO I was 16 mm Hg (SEM = 1.3 mm Hg). The percentage of fall in PaCO z correlated significantly with FEV.IFVC ratio (r=.47, p=o.oi) and with FEV. (r=.5, p=o.oo8). Although the baseline PaCO I did not correlate with FEV., the posthy- perventilation PaCO z did (r =.54, p=.3). Voluntary hyperventilation studies herein suggest a predominant role for impairment ofventilatory control in the maintenance of hypercapnia in OSA since a fall of PaCO. into the normal range can usually be obtained. The correlation between the percentage of fall in PaCO I and spirometric measures of respiratory mechanics, as well as the inability of some subjects to normalize the PaCO z voluntarily suggests an added role for respiratory mechanical impairment in obesity hypoventilation. (Cheat 1991; 1: ) AHI= apnealhypopnea index; 8= obesity hypoventilation; OSA=occlusive sleep apnea syndrome; SaO.= arterial oxygen saturation T he determinants of the presence and severity of hypercapnia in the subgroup of patients with obesity hypoventilation (OU) within the occlusive sleep apnea syndrome (OSA) are probably multifactorial. Mechanical impairment (both obstructive and restrictive), the severity of the sleep disorder, and gender have all been associated with elevated levels ofarterial carbon dioxide tension (PaC 2 ).... Previous analysis in a large group ofour patients with OSA has indicated that, although the presence of hypercapnia was not associated with mechanical impairment, the degree of hypercapnia, once established, was significantly related to the degree ofmechanical impairment as measured by the forced vital capacity (FVC). In the present study, voluntary hyperventilation, a clinical tool recommended to assess Ucan't breathe" vs Uwon't breathe" in patients with hypercapnia,s.6 was examined in subsequent subjects with OSA and hypercapnia. On voluntary hyperventilation, subjects with primary alveolar hypoventilation are classically able to reduce the arterial PaC 2 by at least 11 mm Ug and usually into the normal range. Subjects with hypercapnia due solely to airflow obstruction are usually unable to reduce the arterial PaC 2 to this degree by voluntary effort. 5 *From the Department of Medicine, Division of Respirology, Ottawa Civic Hospital, Ottawa, Ontario, Canada, the Department of Medicine, Section of Respiratory and Critical Care Medicine, University of Illinois College of Medicine at Chicago, Illinois and the West Side Veterans Hospital, Chicago. Supported in part a grant from the Ontario Thoracic Society. trecipient ofa Career Scientist Award, Ontario Ministry ofhealth. Manuscript received November 26; revision accepted March 19. METHODS Consecutive patients with resting hypercapnia defined as a PaC 2 ofgreater than 45 mm Hg and OSA defined as a history ofdaytime hypersomnolence plus an apnealhypopnea index (AHI) of greater than 1 events per hour 7 were studied in a prospective manner. Patients were excluded ifa neuromuscular disease was known or if thyroid function studies revealed hypothyroidism. Informed consent was obtained from each subject. The index of obesity used was the body mass index (BMI) reported in kilograms per square meter.'l Subjects underwent spirometry using a pneumotachygraph linear to 15 Us and applying the American Thoracic Society (ATS) criteria for selection of curves. 9 The maximum inspiratory and expiratory pressures obtained were the best of three efforts using the methodology described by Black and Hyatt. 1 During nocturnal polysomnography, sleep was defined by electroencephalogram (C4,AI), electro-oculogram, and submental electromyogram by the criteria of RechtschafJen and Kales. II AirOow at the nose and mouth was recorded by thermocouples fitted into a loose-fitting face mask. Thoracoabdominal motion was monitored by inductive plethysmography (Respitrace, Ardsley, NY). Arterial oxygen saturation (SaOJ was continuously measured by ear oximetry (472IA, Hewlett-Packard, Waltham, MA) and all these variables were recorded on a 16-channel strip chart recorder. An apnea was defined as cessation of airoow at the nose and mouth lasting for at least 1 S7 and hypopnea was defined as a decrease in airoo~ rib cage, or abdominal excursion by greater than 5 percent associated with an oxygen desaturation ofat least 4 percent below the precedin~ baseline The AHI was calculated as the number of apneas and hypopneas per hour of total sleep time. 7. 1! The mean SaO! was defined as 1 percent minus the mean ofmeasured Sa 2 at the desaturation troughs in each patient. Initial arterial blood gas values were obtained with the patient awake, seated, and breathing R)()m air. The patient was then instructed to voluntarily hyperventilate (as rapidly and deeply as possible) over at least a 6O-s interval while the blood drawer obtained a second arterial puncture at between 6 and 9 s as the patient continued to voluntarily hyperventilate. Continuous coach Voluntary Hyperventilation in Obesity Hypoventilation (Leech et 81)

2 ing and encouragement by a second technician to maintain hyperventilation were usually necessary as hypersomnolent patients were often unable to stay awake voluntarily between punctures. In the analysis, baseline PaCO. in this study group was fint correlated to the PaCO. predicted in subjects with OH from our previous study to see ifthe same acton again influenced the level of PaCO. The prediction equation previously derived from 4 subjects with OH in OSA was as follows: PaCO.=-.19 (PaOJ+O.5 (AHI)-O.11 (FVC) +66. Next, the change in PaCO. on voluntary hyperventilation was expressed as percentage of fall in PaCO. from baseline. This percentage offall was examined in relationship to age, BMI, sleepdisordered breathing indices, and pulmonary function variables by Pearson correlation. RESULTS Twenty-seven patients with OH were studied. Ten ofthese were women. Ages ranged from 34 to 71 years with a mean age of51 years. Body mass index ranged from 21 kg/m 2 to 68 kg/m 2 with an average of 44 kg/ sqm. All subjects were hypersomnolent, frequently severely so. Many had to be aggressively coached during hyperventilation to avoid sleep onset. Nocturnal polysomnograms demonstrated an average AHI of 73 events perhour (SEM = 8.2 events perhour, minimum 19 events per hour, maximum 176 events per hour). The average oxyhemoglobin saturation trough recorded following sleep-disordered breathing events was 74 percent (SEM=2 percent, minimum 53 percent, maximum 89 percent). Awake seated arterial blood gas values demonstrated varying degrees ofhypoxemia (mean Pa 2 = 6 mm Hg, SEM = 2.4 mm Hg). Resting PaC 2 ranged from 46 to 67 mm Hg (mean PaC 2 = 55 mm Hg, SEM=1.3 mm Hg). The FEV.IFVC percentwas greater than 75 percent in 14 subjects or 52 percent ofcases. For the purposes of this discussion, airhow obstruction was defined as an FEV.IFVC ofless than 75 percent. A mild degree ofairflow obstruction (FEV.IFVC percent of6 to 75 percent) was observed in 11 subjects or 41 percent of cases and a moderate obstructive defect was observed in two subjects (FEV.IFVC percent between 5 and 6 percent). The absolute value ofthe FEV. was less than 1.2 L in four subjects (actual values 1., 1.1,.9, and 1. L), all ofwhom had mild to moderate airhow obstruction. The mean percent predicted FVC was reduced at 6 percent predicted (SEM = 3.4 percent, minimum value 32 percent, maximum value 15 percent). In summ~ pulmonary function tests frequently revealed a restrictive pattern, with a mild to moderate element ofairhow obstruction in about half the patients. In only four patients was the absolute value ofthe FEV. decreased to less than 1.2 L, a level below which hypercapnia may be seen solely due to chronic airhow obstruction. 5 These heterogeneous pulmonary function results are similar to previous values reported in 34 men with OSA from our laooratory. 3 Subjects' individual and mean results for age, BMI, FEV.IFVC percent, FEV., AHI, and before and after hyperventilation PaC 2 are reported grouped by the absence (n = 14) ('Iable 1) or presence (n = 13) (Table 2) ofairhow obstruction. The majority ofsubjects were massively obese and many had severe obstructive sleep apnea as manifested by the AHI. Maximum inspiratory mouth pressure mean values were -57 cm H 2 (SEM= -7.6 cm H 2, minimum - 1 cm H 2, maximum -157 cm H 2 ). Similarly mean maximal expiratory pressures were within the normal range at 14 cm H 2 (SEM = 8.6 cm H 2, minimum 3 cm H 2, maximum ISO cm H 2 ). The baseline PaC 2 correlated with a predicted Table I-H~ OSA ltjtienta tdithout ObaructioeAinDt DiIeae Hyperventilation PaCO., mm Hg Age, BMI, FEV/FVC FEV. (% pred), AHI, Sex yr kwm If, L Eventslh Before After F (57) F (6) M (35) M (64) M (71) M (53) M (35) M (62) M (42) F (71) M (73) F (53) F (81) F (76) X (6) ±ISEM (4) CHEST I 1 I 5 I NOVEMBER,

3 Hyperventilation PaCO., mm Hg Age, 8MI, FEV/FVC FEV1 (If, pred), AHI, Sex yr ~m If, L Eventslh Before After F (5) M (71) M (27) M (27) M (3) M (19) M (42) M (42) F (46) M (36) F (45) F (66) M (54) X (5) ±ISEM (6) PaCO! from the equation we had previously reported with a Pearson correlation coefficient of r=o.6 (P =.9). As a basis for comparison ofthe strength of this relationship, the maximal inspiratory pressure correlated with the maximal expiratory pressure within the present study group at a similar level (r=o.65, p=o.ooo5). The average fall in arterial PaCO! on voluntary hyperventilation was 16 mm Hg (SEM = 1.3 mm Hg, minimum decrease5 mm Hg, add maximum decrease 26 mm Ug). Seven subjects with airflow obstruction could not decrease the PaCO! by greater than 11 mm Ug, although all subjects without airflow obstruction could. These seven subjects did not differ in age, BMI, AUI, or mean nocturnal saturation from their six counterparts with airflow obstruction who could successfully voluntarily hyperventilate. When expressed as a percentage decrease from baseline, the mean decrease was 29 percent (SEM =2.6 percent, minimum 9 percent, maximum 52 percent). Seven patients could not lower the PaCO! into the normal range; and only one of these (posthyperventilation PaCO!=46) had a normal FEV.IFVC ratio. Examination ofthe correlates ofthe percentage fall in PaCO! within the total study group did not demonstrate a significant relationship (P>O.5) with age, BMI, the severity of sleep-disordered breathing as measured by AHI or mean nocturnal desaturation, or effort or muscular force as represented in the maximal inspiratory/expiratory pressures. The percentage fall in PaCO! did relate to FEV.IFVC percent (r=o.47, p=o.oi), to FEV. percent predicted (r=o.41, p=o.3), and to FEV. in absolute terms (r=o.56, p=o.o(2). Figure 1 depicts the relationship between percentage of fall in PaCO! and FEVl' expressed as percent predicted in those with and without airflow obstruction. Even in the subjects without a decrease in the FEV.IFVC percent ratio, there remains a ~ : ~ FIGURE 1. Relationship of fall in Peo. to FEV. percent predicted. Open circles=fev/fvc ratio ~75"; closed circles=fev/fvc ratio <75 percent FEY1 '" precid8d

4 (. ~~ L.- I~2 l' o _52 o ~. ~: I: ' ' FEY 1 (UtNI) FEV 1 (Utres) FIGURE 2. Relationship to FEV. ofbefore and after hyperventilation Poo. Open circles=fev/fyc ratio ~75 percent; closed circles = FEV/FYC ratio <75 percent. significant relationship between FEV. percent predicted and percentage offall in PaCO t In a stepwise hierarchical regression analysis with the fall of PaCO I as the dependent variable afterthe FEVIIFVC percent entered the model (r'=.21, p ==.1), no other variables added significantly to the model. Although baseline PaCO I did not correlate with FEV I (r=o.34, p=o.9), the posthyperventilation PaC 2 correlated with FEVI both in absolute terms (r=o.54, p==o.oo3) and in terms ofpercent predicted FEV I (r==o.66, p=o.oo(2). These relationships are demonstrated in Figure 2. DISCUSSION The mechanisms of development of hypercapnia are four: alterations in carbon dioxide production, disturbances in gas exchange, abnormalities in the mechanical system, and impairment ofthe ventilatory control system. In anyone patient or diagnostic category, these mechanisms may interact in a complex fashion} In OH, a combination of the latter two mechanisms led to the development of hypercapnia. This report uses voluntary hyperventilation, a simple test that overrides the inhuence of the resting ventilatory control system to tease out the relative weights ofthese two components. Rhoads and Brady6 first described voluntary hyperventilation in seven patients with hypercapnia. Prior to this, itwas believedonecould notmakethediagnosis of idiopathic alveolar hypoventilation in the presence of respiratory mechanical impairment, either restrictive orobstructive. Theirseven patientswith moderate degrees of obstructive or restrictive defects had a mean PaCO t of 64 mm Hg and, as expected, diminished ventilatory responses to carbon dioxide. Absolute values for FEV1 were greater than levels usually associated with COl retention and subjects were able to voluntarily hyperventilate to normal, or near normal PaCO Sh with a decrease of greater than 11 mm Hg. This was in contrast to 2 patients with bronchitis whose resting PaCO I averaged 58 mm Hg but who were, on the average, only able to voluntarily decrease the PaCO I by 5 mm Hg. These findings were published before the syndrome of obstructive sleep apnea was understood to be associated with alveolar hypoventilation, but it is of note that several of their subjects were massively obese. Investigations in mass loading and OH have documented that in OH, as compared with obese or mass loaded eucapnia, the respiratory system is less compliant,15 respiratory muscles do not respond as well to equivalent neural stimuli,16 lung volumes are lower,i and load compensation as measured by EMG di and inspiratory effort (PO. IS) is impaired.i These findings suggest both mechanical impairment and decreased central neural drive in OB. Bradley et al 3 suggested that a mild to moderate degree ofdiffuse airflow obstruction was necessary for the development of hypercapnia in obese subjects with OSA and hypercapnia. In a larger group, we also found that the sev~rityofhypercapnia, once present, related significantly to mechanical impairment represented by the percent predicted FVC, although there were certainly obese individuals with OSA and hypercapnia who had completely normal results ofspiromet~ CHEST I 1 I 5 I NOVEMBER,

5 The present study demonstrates in 27 patients with OSA and OH that, like the seven patients of Rhoads and Brady,6 and unlike their patients with severe airhow obstruction, most subjects were abje tq,voluntarily hyperventilate to normal or near no~al PaC 2 This suggests a predominant role for nonmechanical causes ofhypercapnia in OH, ie, impairment ofneuromuscular respiratory control systems. On the other hand, of significant correlation between the percentage offall in PaC 2 and respiratory mechanics (FEV.IFVC ratio and percent predicted FEV.) suggests that airflow obstruction and/or restriction, perhaps due to the mass loading of obesi~ have a measurable effect on the ability to voluntarily hyperventilate and may playa role in the development or maintenance ofhypercapnia. Apart from airflow obstruction per se, a difference in minute ventilation, dead space partitioning, or the development of inspiratory muscle fatigue might also account for an inability to lower the PaC 2 in some individuals. The present study's experimental design does not enable examination ofthese possibilities and they are problematic to study adequately in the hypersomnolent morbidly obese subject during voluntary hyperventilation. Voluntary hyperventilation is an artificial stress of the resting properties of the respiratory system that overrides the resting inhuence of the central controller. Just as Rhoads and Brady6 used it to discern Ucan't breath" from uwon't breathe;' we suggest that it can be used to demonstrate that a combination of both impaired central drive and respiratory mechanical impairment interact in the etiology ofhypercapnia in individuals with the OH syndrome. REFERENCES 1 Leech ja, na1 E, Baer ~ Lopata M. Determinants of hypercapnia in occlusive sleep apnea syndrome. Chest 1987; 92: Lopata M, Onal E. Mass loading, sleep apnea and the pathogenesis of obesity hypoventilation. Am Rev Respir Dis 1982; 126: Bradley}D, Rutherford R, Lue F, Moldofsky F, Grossman RF, Zamel N, et ale Role of diffuse airway obstruction in the hypercapnia of obstructive sleep apnea. Am Rev Respir Dis 1986; 134: Suratt PM, Wilhoit SC, Atkinson RL. Elevated pulse 8w resistance in awake obese subjects with obstructive sleep apnea. Am Rev Respir Dis 1983; 127: Lopata H, LoureJl9> ~ Evaluation of respiratory control. In: Williams MH et ale Disturbance of respiratory control: Clinics in chest medicine. Philadelphia: WB Saunders; Rhoads GG, Brody }S. Idiopathic alveolar hypoventilation: clinical spectrum. Ann Intern Med 1969; 71: Guilleminault C, Cummiskey}, Dement WC. Sleep apnea syndromes. Annu Rev Moo 1976; 27: Arnold CB, ed. Measurement of overweight. Statistical bull, Metropolitan Co 1984; 65:2-3 9 Ferris BG. Principal investigator: epidemiology standardization project. Am Rev Respir Dis 1978; 118:6 1 Black LF, Hyatt RE. Maximal respiratory pressures: normal values and relationship to age and sex. Am Rev Respir Dis 1969; 99: Rechtscbaffen A, Kales A. A manual ofstandardized terminol~ techniques and scoring system for sleep stages of human subjects. Los Angeles: BISIBR}, UCLA; Guilleminault C, Tilldan A, Dement WC. The sleep apnea syndromes. Annu Rev Med 1976; 27: na1 E, Leech ja, Lopata M. Relationship between pulmonary function and sleep-induced respiratory abnormalities. Chest 1985; 87: Weinberger SE, Schwartzstein RM, Weiss JW Hypercapnia. N Eng} J Med 1989; 321: SharpjT, Henry J~ Sweany SIC, etale The total work ofbreathing in normal and obese men. J Clio Invest 1964; 43: Louren~ ~ Diaphragm activity in obesity. J Clio Invest 1969; 48: Voluntary HyperventiIaIioIn ObesIty Hypcw8ntIIatIon (Leech et ai)