Effect of Metabolic Acidosis Upon Sleep Apnea*

Transcription

1 Effect of Metabolic Upon Sleep Apnea* john T. Sharp, M.D., F.C.C.P.; WalterS. Druz, Ph.D.; Vivian D'Souza, M.D.; and Edward Diamond, M.D. The effects of metabolic acidosis upon the pattern of apnea during sleep were assessed in ten sleep apnea patients. Four had pure obstructive apnea, two pure central apnea, and four had mixed apnea. was induced with acetazolamide. Acid-base shifts had little effect in pure obstructive and pure central apnea, but had a significant effect in mixed apnea. In two of the mixed apneic patients, metabolic acidosis converted predominantly central apnea into nearly pure obstructive apnea, prolonging apneic periods and worsening hypoxemia. A suggested explanation for this is the greater stimulating effect of acidosis upon the lower bellows muscles than upon the muscles which act to maintain patency of the upper airways. The observation that some patients with mixed sleep apnea appear to have central apnea while relatively alkalotic and obstructive apnea while acidotic emphasizes the need for more careful and detailed characterization of apneic disorders with respect to their responses to body states and therapeutic agents. The sleep apnea syndromes clearly represent disorders of ventilatory control. l-3 Because of the possibility that altering central respiratory drive might change the apnea pattern, and because metabolic acidosis induced by acetazolamide has been reported to be useful therapeutically in central sleep apnea 4 5 and certain other respiratory control disorders, 6 we undertook the present study. Mild metabolic acidosis was induced in ten patients with sleep apnea syndromes and the effects upon the sleep pattern assessed. *From the Pulmonary Section, Medical Service, Hines Veterans Hospital and the Department of Medicine, Loyola-Stritch School of Medicine, Hines, IL. This work was supported by NIH Grant Numbers HL and HL and by the Veterans Administration Research Service. Manuscript received August 9; revision accepted November 6. Reprint requests: Dr: Sharp, Hines VA Hospital, PO Box 7, Hines, IL Table!-Anthropometric and Pulmonary Function Data PATIENTS AND METHODS Anthropometric and pulmonary function data on the ten patients are summarized in Table 1. Four had purely obstructive apnea and were obese. Two of these four were hypercapnic. Two patients had purely central apnea. Both were hypercapnic, and one was obese. Four patients had mixed apnea, and of these, two were obese and two of normal weight. None was hypercapnic. In none of the ten patients was there evidence of significant lower airway obstruction as indicated by the FEV 1/FVC. Blood gas values given in Table 1 were obtained during the waking state, prior to therapy. Patients 1 and 3 had experienced right-sided heart failure related to their obesity and hypoventilation, and patient 7 had been in biventricular failure owing to coronary and hypertensive heart disease. None of the three was in failure at the time of study and all were clinically stable. Mild metabolic acidosis was induced using oral acetazolamide in doses of500 to 1,000 mg daily for four to seven days. The therapeutic aim was to reduce the plasma bicarbonate to between 16 and 20 Mild metabolic alkalosis was induced by giving furosemide, 40 mg daily, and sodium bicarbonate, 4 to 8 g daily, for four to seven days, the therapeutic goal being a plasma bicarbonate between 28 Weight Resting Arterial Blood Height Weight %of % Patient Age (em) (kg) Ideal ph Pco 2 Po 2 Hco 3 L Pred Pure obstructive apnea Pure central apnea Mixed apnea S vc CHEST I 87 I 5 I MAY,

2 Table 2-Apnea Variables in Four Patients with Obstructive Sleep Apnea during Metabolic and Relative Metabolic Alkalosis Relative Alkalosis ph 7.345± ±0.15 PaCO, 38.8± ±4.9 Hco ± ±2.3 % Apneas obstructive 100±0 96.5±3.5 % Apneas with a central period 2.3± ±7.3 Maximum 91.9± ±4.3 Nadir 75.1± ±5.3 4 so, 16.8± ± 1.8 (apneaslhr) 88.8± ±4.9 Apnea duration {sec) 23.7± ± 1.9 *Mean values± SE; 0 2 saturation in%; Pco, in mm Hg; and Hco 3 in and 32 Significant hypokalemia was not encountered in these experiments, and patients complained of no untoward symptoms related to their altered acid-base status. The diagnosis of sleep apnea and its type was established in all cases by conventional sleep studies, usually nap studies done in the morning hours after a night of sleep deprivation. In several instances, all night studies were done as well and in no patient were findings during nap studies and all night studies significantly different. In all patients, at least one sleep study had been done prior to the control observations to establish the diagnosis and to familiarize the patient with the apparatus. Most patients had had several prior studies. All patients complained of diurnal hypersomnolence, and none had any difficulty falling asleep or staying asleep during testing on any of the testing occasions. In no instance did the earlier preliminary studies show differences from the control study. Morning nap studies were done for the evaluation of the effects of altered acid-base status. During sleep studies, apnea was detected by a thermistor and/or an infrared carbon dioxide analyzer coupled to a small plastic mask covering the nose and mouth through which a constant bias How of air (12 Umin) was drawn. Motion of the abdomen and rib cage was detected by either magnetometers or an inductance plethysmograph. Oxygen saturation was monitored by an ear oximeter and sleep was monitored and staged using parietal and occipital electroencephalograms, vertical, and horizontal electroocculograms and a mental electromyogram. Sleep studies evaluating the effects of changed acid-base status were done within one week of one another, and there were no other significant clinical changes, such as weight loss or gain, between studies. In assessing apnea changes during sleep studies, the following measurements were compared: percentage of apneas which were obstructive, percentage of apneas in which there was a central apneic portion of ten seconds or greater, average maximum oxygen saturation, average nadir oxygen saturation (lowest value during apneas), average change in oxygen saturation occurring during apnea (the difference between maximum and nadir values - 4SO,), the apnea index or average number of apnea.~ per hour, and the average duration of apneic periods. In all studies, measurements were averaged over at least 90 minutes of stable uninterrupted sleep during which apneic episodes occurred regularly and continuously. Obstructive Apnea RESULTS Table 2 summarizes results observed in the four patients with purely obstructive apnea. There were no significant differences between the acidotic state and the relative alkalotic state for any of the apnea variables compared for any of the four patients as determined by Student's t-test. This is reflected in the closely similar mean values for all variables for the group as a whole. There were, however, substantial differences in the ph, Pco 2 and plasma bicarbonate for all patients between the two states. Blood gas values given in this and subsequent tables are values obtained during the waking state. Patients with alveolar hypoventilation did not differ significantly from those without in their responses to changing acid-base status. Central Apnea Th.ble 3 summarizes observations in the two patients with purely central sleep apnea. Despite substantial increases in plasma bicarbonate after furosemide and bicarbonate administration, the blood ph remained slightly acidotic because of increases in Pco 2 In both patients, the average nadir and average change (AS0 2) values for oxygen saturation were increased by induced acidosis, probably owing to an improved overall level of alveolar ventilation. However, there was no significant change in apnea index or apnea duration. Mixed Apnea Data on the four patients with mixed apnea are summarized in Tables 4 and 5. Table 3-Apnea Variables in Two Patients with Pure Central Sleep Apnea during Induced Metabolic and Relative Metabolic Alludoais. Relative Alkalosis ph Pnt Pnt PaCO, Pnt Pnt Hco 3 Pnt Pnt maximum Pnt ± ±1.2 Pnt ± ±0.1 Nadir Pnt ± ±3.0 4 so, Pnt ± ±0.3 Pnt 5 9.2± ±3.3 Pnt 6 6.7± ±0.3 Pnt Pnt Apnea duration: Pnt ± ±1.9 Pnt ± ±0.4 *Mean values± SE; Pco, in mm Hg; 0 2 saturation in%; and Hco 3 in 620 Metabolic and Sleep Apnea (ShBIP et at)

3 Table 4-Apnea Variables in Patients 7 and 8 with Mixed Sleep Apnea during Induced Metabolic and Relative Metabolic Alkalosis.* Relative Alkalosis Arterial blood ph Pnt Pnt PaC02 Pnt Pnt Hco 3 Pnt Pnt Maximum Pnt ~ ~0.8 Pnt ~ ~0.3 Nadir Pnt ~ ~3.5 Pnt ~ ~ 1.1 A so2 Pnt ~ ~3.5 Pnt 8 4.7~ ~0.9 Pnt Pnt Apnea duration: Pnt ~ ~4.1 Pnt ~ ~6.3 % Apneas with obstruction Pnt 7 93% 0% Pnt 8 89% 18% *Mean values~ SE; 0 2 saturation in%; Pco 2 in mm Hg; and Hco 3 in In contrast to the patients with pure obstructive and pure central apnea who showed little change in apnea variables with alterations in acid-base status, the patients with mixed apnea showed substantial changes. Patients 7 and 8 showed the most impressive changes (Table 4). When initially studied, both were relatively alkalotic and were thought to have central apnea on cursory examination of sleep polygraphs. More detailed examination revealed in patient 7 an occasional obstructive effort just prior to the onset of effective ventilation and in patient 8, one or more obstructed inspiratory efforts in 18 percent of apneic periods. When made mildly acidotic, 93 percent and 89 percent of apneic episodes in patients 7 and 8, respectively, showed multiple obstructed inspiratory efforts. Apneic episodes were significantly prolonged in the acidotic state in both patients by Student's t-test, and this appeared to be the main reason why nadir oxygen saturation values were lower and aso2 values greater during acidosis. Figure 1 shows a portion of the polygraph record of patient 7. In panel 1 left, tracings during relative alkalosis show the central characteristics of the apnea. Panel! right shows obstructive apnea present during acidosis as indicated by the respiratory effort undulations during the apneic period in the top three tracings. Figure 2 compares slow paper speed tracings of sleeping oxygen saturation alone in patient 7, the top tracing taken during relative alkalosis and the bottom tracing taken during acidosis. It is apparent that hypoxemia associated with apnea is more severe during acidosis than during relative alkalosis. All night sleep studies were performed in both patients 7 and 8. The structure of their nocturnal sleep was essentially the same during acidosis and relative alkalosis. In both patients, stage 2 sleep predominated with no stage 3 or stage 4 sleep. Both had short periods of REM sleep totaling less than 7 percent of the total sleeping time. In both patients, over 70 percent of apneic episodes had their terminations associated with EEG evidence of arousal (ie, well developed alpha rhythm). Findings in both patients 7 and 8 when first studied were identical to those found in relative alkalosis. Findings changed as indicated above, and in Table 4 when acidosis was induced and reverted back to the initial findings when relative alkalosis was induced, indicating the rapid reversibility of changes. In patient 7, metabolic acidosis was induced on two occasions with identical results. Maximal oxygen saturation was lower in patient 7 during acidosis despite an overall increase in alveolar ventilation. Inspection of the polygraph record suggests that this is explained by the longer duration of apneic episodes coupled with their greater number per unit of time which often did not allow the oxygen saturation to rise optimally before the next apneic episode began. Data on patients 9 and 10 are presented in Table 5. Patient 9 showed relatively few apneic periods with Table 5-Apnea Variables in Patients 9 and 10 with Mixed Sleep Apnea during Induced Metabolic and Relative Metabolic Alkalosi& * Relative Alkalosis ph Pnt PaC02 Pnt Pnt Pnt Hco 3 Pnt Pnt % Apneas with obstruction: Pnt Pnt Maximum Pnt ~ ~0.6 Pnt ~ ~0.1 Nadir Pnt ~ ::!: 1.0 A so2 Pnt ~ ~0.2 Pnt 9 2.7~ ~ 1.2 Pnt ~ ~0.3 (Apneaslhr) Pnt Pnt Apnea duration (sec) Pnt ~ ~2.9 Pnt ~ ~ 1.0 *Mean values~ SE; 0 2 saturation in%; Pco 2 in mm Hg; and Hco 3 in CHEST I 87 I 5 I MAY,

4 R.A. Alkalosis HC RA. Ac:idolla HC0,-20 ABO APNEA FIGURE I. Polygraphic tracings (redrawn) made during sleep on patient 7 while relatively alkalotic (A, left) and while acidotic (8, right). In each panel, the top three tracings are from an inductance plethysmograph showing motion of the rib cage (RC), abdomen (ABD) and their sum (RC + ABD). The fourth tracing is from the face mask-mounted thermistor which detects apnea, and the bottom tracing is from the ear oximeter. Note that the inductance plethysmograph tracings of thoracoabdominal motion show no respiratory effort during apnea during alkalosis (left), but numerous respiratory efforts throughout the apneic period during acidosis (right). obstruction in either acid-base stage (13 percent and 16 percent), and the incidence of obstructive phenomena was not affected by the change. Nor did the change appreciably affect the incidence of obstructive phe- nomena in patient 10, although most (77 percent and and a reduced.:1 S0 2 during acidosis. There was no significant change in oxygen saturation in patient 10. DISCUSSION Studies in men 7 and dogs 8 indicate that changes in 88 percent) of his apneic periods contained three or blood ph influence minute ventilation during the four obstructed inspiratory efforts. waking state and both minute ventilation and respira- The apnea index was somewhat less and apneic tory rhythm during sleep. This is explainable by periods significantly shorter (p<o.ol) during acidosis change in the degree of stimulation of central and in patient 9. In patient 10, the apnea index was peripheral chemoreceptors by hydrogen ion and conunchanged by acid-base shift, but apneic periods were sequent variation in central respiratory drive. Based significantly longer (p<o. 01) during acidosis. In pa- upon our current understanding of ventilatory control, tient 9, oxygen saturation reflected the changes in increased central respiratory drive owing to metabolic duration of apneic periods, with higher nadir values acidosis should make apneic periods shorter or fewer 100 _ R.A. Alkalosis HC Central Apnea c# _ RA. HC Obstructive Apnea 90- ~ Ear Oximeter RecordinQs FIGURE 2. Tracings of oxygen saturation from the ear oximeter (redrawn) during approximately 50 minutes of sleep in patient 7 when ht> was relatively alkalotic (top tracing) and when he was acidotic (bottom tracing). Note that he is generally more hypoxic when acidotic than when alkalotic. 822 Metabolic and Sleep Apnea (Sharp et el)

5 (or both) in patients with sleep apnea. Only one of our ten patients (patient 9) showed this expected response; however, despite the fact that acetazolamide decreased ph, PaC0 2, and plasma bicarbonate in all patients in the waking state. More striking yet was the finding in patients 7 and 8 that during metabolic acidosis, apneic episodes were significantly longer and decreases in oxygen saturation associated with apnea were greater. Indeed, it appeared in these two patients that metabolic acidosis had converted central apnea to obstructive apnea. What is implied by the appearance of upper airway obstructive apnea during acidosis in these two patients where little obstruction had been present during a more alkaline acid-base status? Certainly, it suggests that the metabolic acidosis had stimulated the inspiratory bellows muscles (diaphragm, intercostals, scaleni) to contract at times (during apnea) when they had previously been quiescent. However, it will be recalled that the duration of apneic periods was prolonged by acidosis. This suggests that the upper airway muscles responsible for opening the airways and terminating apnea were either quiescent or contracting ineffectively during acidosis at a time during apnea when they had effectively kept the upper airway open in the absence of acidosis. This, in turn, suggests that metabolic acidosis may in these two patients, have stimulated the activity of the bellows muscles to a greater extent than it stimulated the upper airway muscles. In support of this idea, studies by Weiner et al 9 indicate that respiratory stimulation by C0 2 rebreathing and by hypoxia may have proportionately lesser effects upon upper airway muscles supplied by the hypoglossal nerve than upon bellows muscles such as the diaphragm. Unfortunately, we lack EM G data on upper airway muscles such as the genioglossus which would have shed more light upon this unexpected phenomenon in our patients. If increasing central respiratory drive caused more vigorous action of the lower bellows muscles alone, leaving upper airway muscles unaffected, it is quite possible that the resulting more negative inspiratory pressures in the lower pharynx could cause more airway collapse and longer periods of obstructive apnea. Onal et al 10 point out that while animal data show different apneic thresholds and different degrees of hypoxic and hypercapnic responsiveness for diaphragm as compared to genioglossus muscle, human data show parallel and proportional changes in the activity of these two muscles during obstructive and mixed apnea. These authors assert that a greater decrease in genioglossus activity is not necessary for obstructive apnea to occur. What is needed is a decrease in upper airway muscle activity relative to pharyngeal pressure. Progressive decrease in upper airway muscle activity during the wane of the periodic cycle even with proportionate decreases in respiratory muscle activation can result in increa~ed pharyngeal resistance and transpharyngeal pressure despite decreasing negative intrathoracic pressure especially if the airway is predisposed to narrowing, as frequently occurs with obesity, tonsillar hypertrophy, or other structural abnormality of the upper airway. More recently, Shore and Millman 11 have reported a patient given acetazolamide in whom central apnea converted to mixed apnea and who concomitantly experienced worsened daytime hypersomnolence. Despite these adverse changes, apneic periods occurred less frequently, caused less hypoxemia and were unchanged in length as a result of the induced metabolic acidosis. This was in contrast to the change we observed in patients 7 and 8. These authors recommended repeating sleep studies in patients whose symptoms are not improved by therapy. It is of interest that included in the abstract reporting the usefulness of acetazolamide in central sleep apnea 4 are the following statements: The one patient who did not respond subjectively or objectively w.ts obese with loud snoring and had 31 percent obstructive apneas in the predrug study, a higher portion than most of the others studied. During acetazolamide therapy her obstmctir;e apnea became more prominent with no symptomatic improvement despite acidification of the arterial blood. This suggests that White et al 4 observed the same phenomenon in this patient, who seems to have had mixed apnea, that we observed in our patients 7 and 8. In contrast to the experience of White et al, ~our two patients with pure central apnea showed no subjective or objective improvement with induced acidosis except for slight increases in nadir and 4S0 2 values for oxygen saturation and decreased arterial Pco 2 values. It may be significant that both our patients had central alveolar hypoventilation which flattened their carbon dioxide response curve. If a further depression of ventilatory response owing to the sleeping state is added, the lack of change in sleep apnea parameters is probably understandable. The patients reported by White et al 5 all had normal waking arterial Pco 2 values, a fact which may partially explain why their patients responded favorably to acetazolamide while ours did not. Concerning the other two patients with mixed apnea, patients 9 and 10, patient 10 was similar to patients 7 and 8 in that he showed prolongation of apneic periods during acidosis. However, there was little change in the incidence of obstructive apnea, although the duration of the obstructive phase within apneas appeared somewhat prolonged. As indicated earlier, patient 9 showed the expected response to metabolic acidosis, namely shortening of apneic periods and reduction in their number. Although the major reason for lower oxygen saturations during apnea in patients 7 and 8 was the longer apneic periods resulting in lower alveolar Po 2 values, a rightward shift in the oxygen-hemoglobin dissociation owing to acidosis (the Bohr effect) might have pro- CHEST I 87 I 5 I MAY,

6 duced an additional decrement in oxygen saturation since a lower oxygen saturation would result from a given decrement in Po 2 With respect to the patients with obstructive apnea, one may postulate that patients 1 and 3 who had alveolar hypoventilation might not show a change in sleep apnea variables during metabolic acidosis for the reasons cited in the preceding paragraph dealing with the central apneic patients with alveolar hypoventilation, ie, depressed response to ph change. However, patients 2 and 4 whose waking arterial Pco 2 levels were normal also failed to show significant changes during sleep when acidotic. We have no explanation for this lack of response. We believe that the following conclusions are warranted from our observations. Differing responses were noted to the induction of metabolic acidosis in sleep apnea patients, often contrary to what one would predict a priori. This variety of responses suggests that the sleep apnea syndromes represent a heterogeneous group of patients who have varied lesions involving those portions of the brain stem concerned with respiratory control and with sleep. While this is hardly an original idea, it nevertheless bears reemphasizing. Variation in the anatomy of the upper airway may also play a role. The unexpected worsening of upper airway obstruction in two mixed apneic patients resulting from stimulating ventilation by acidifying the blood emphasizes the separation of control between the two important groups of inspiratory muscles, those controlling the inspiratory patency of the upper airway as opposed to those driving the thoracic bellows. The fact that certain patients with mixed sleep apnea may appear to have central apnea when blood ph is normal or alkalotic and obstructive apnea when acidotic indicates a need for more careful and detailed characterization of apneic disorders with respect to their responses to body states and therapeutic agents. Acetazolamide has been described as therapeutically useful in central sleep apnea. However, if such patients actually have mixed sleep apnea and are relatively alkalotic, such therapy may effect a transition to predominantly obstructive apnea, prolonging apneic episodes and worsening hypoxemia. In addition, acetazolamide may have different effects in central sleep apneic patients depending upon whether or not they have central alveolar hypoventilation. For these reasons, sleep studies should be repeated on therapy when acetazolamide use is initiated. ACKNOWLEDGMENTS: The authors thank Dr. NeilS. Cherniak for his helpful review of this manuscript. REFERENCES 1 Gastaut H, Tassinari CA, Duron B. Etude polygraphique des manifestations episodiques (hypniques et respiratoires), diurnes et nocturnes, du syndrome de Pickwick. Rev Neurol 1965; 112: Walsh RE, Michaelson MC, Harkelroad LE, Zighelboim A, Sackner MA. Upper airway obstruction in obese patients with sleep disturbance and somnolence. Ann Intern Med 1972; 76: Guilleminault C, lilkian A, Dement WC. The sleep apnea syndromes. Ann Rev Med 1976; 27: White DP, Zwillich CA, Pickett C, Hudgel OW, Weil JV. Acetazolamide therapy for central sleep apnea (abstract). Am Rev Respir Dis 1981; 123 (No.4, part 2, supplement):177 5 White DP, Zwillich CS, Pickett CK, Douglas NJ, Findley LJ, Weil JV. Central sleep apnea improvement with acetazolamide therapy. Arch Intern Med 1982; 182, Skatrud JB, Dempsey JA. Relative effectiveness of acetazolamide versus medroxyprogesterone acetate in correction of chronic carbon dioxide retention. Am Rev Respir Dis 1983; 127: Wei) JV, Kryger MH, Scoggin CH. Sleep and breathing at high altitude. In: Guilleminault C, Dement W, eds. Sleep apnea syndromes. New York: Alan R. Liss Inc, 1978: Phillipson EA, Sullivan CE. Respiratory control mechanisms during NREM and REM sleep. In: Guilleminault C, Dement W, eds. Sleep apnea syndromes. New York: Alan R. Liss Inc, 1978: Weiner D, Mitra J, Salamone J, Cherniack NS. Effect of chemical stimuli on nerves supplying upper airway muscles. J Appl Physiol Respir Environ Exer Physiol 1982; 52: Onal E, Lopata M, O'Connor T. Pathogenesis of apneas in hypersomnia-sleep apnea syndrome. Am Rev Respir Dis 1982; 125: Shore ET, Millman RP. Central sleep apnea and acetazolamide therapy. Arch Intern Med 1983; 143: Metabolic and Sleep Apnea (Sharp et al)