Timing Patterns of Cluster Headaches and Association with Symptoms of Obstructive Sleep A p n e a

Transcription

1 Sleep Research Online 3(3): , Printed in the USA. All rights reserved X 2000 WebSciences Timing Patterns of Cluster Headaches and Association with Symptoms of Obstructive Sleep A p n e a Ronald D. Chervin, Sarah Nath Zallek, Xihong Lin*, Janette M. Hall, Namita Sharma and Kristen M. Hedger Sleep Disorders Center, Department of Neurology and *Department of Biostatistics, University of Michigan, Ann A r b o r, Michigan, USA Cluster headaches (CH) frequently recur at the same point in the circadian cycle, often during sleep. They may, in some cases, represent a susceptible individual s response to hypoxemia or other physiological changes induced by obstructive sleep apnea (OSA). If and when this mechanism exists, timing of CH close to the onset of sleep and therefore OSA might be expected. We questioned 36 subjects with CH about the times at which their CH usually occurred and about several symptoms known to be predictive of OSA, including habitual snoring, loud snoring, observed apneas and excessive daytime sleepiness. We then used logistic regression to determine whether occurrence of CH in each of six time periods was associated with OSA symptoms. The 23 subjects (64%) who reported CH in the first a typical s sleep also tended to report headaches during the midday/afternoon period. Symptoms of OSA, and in particular habitual snoring, were predictive of both first-halfof-the- and midday/afternoon CH (p<.05). Thirty-one subjects (86%) reported that their CH were sleep-related, usually occurring during any part of the or on awakening, but symptoms of OSA were not predictive of this timing pattern. In short, several OSAsymptoms showed an association with CH occurrence in the first the but not with sleep-related CH in general. These findings suggest that in some patients, physiological consequences of OSAmay trigger CH during the first few hours of sleep and thereby influence the timing of subsequent daytime headaches. CURRENT CLAIM: Occurrence of cluster headaches during the first the s sleep is associated with occurrence during the afternoon also, and both these times of headache proclivity are predicted by symptoms of obstructive sleep apnea. Cluster headaches (CH) are intense, unilateral paroxysms of head pain associated with ipsilateral cephalic autonomic changes (Headache Classification Committee of the International Headache Society, 1988). Headaches typically occur in clusters that last weeks to months, separated by remissions that last months to years, but can also occur chronically, without remission. About 0.4 to 1% of men have CH (Kudrow, 1989). The etiology of CH is not well known, but they are believed to be vascular headaches that occur during a hypothalamus-mediated period ("cluster") of increased susceptibility to triggering stimuli. One common trigger may be hypoxemia, as suggested by experimental induction of CH with hypoxia (Kudrow and Kudrow, 1990, 1993) and prompt relief from CH with supplemental oxygen (Kudrow, 1981) but conflicting data also exist (Shen et al., 1993). One cause of repeated hypoxemia is obstructive sleep apnea (OSA), which affects about 24% of men between the ages of 30 and 60 (Young et al., 1993), though most do not complain of excessive daytime sleepiness and only a small minority have been diagnosed (Young et al., 1997). Whether symptomatic or not, OSAis linked in a dose-dependent manner to hypertension and may contribute to morbidity in other disease states as well (Young et al., 1997). Some evidence suggests OSA may exacerbate chronic headaches, especially those that occur during sleep or on awakening (Paiva et al., 1997). We wondered whether OSA could play a particularly important role in CH. Several observations support this possibility. Both CH and OSA affect men more than women (Swanson et al., 1994; Young et al., 1993). Cluster headaches are particularly common at and during daytime naps, and patients may be able to prevent regular nocturnal attacks by avoiding sleep (Kunkle et al., 1952; Ekbom, 1970; Russell, 1981). The onset of nocturnal CH is usually during or just after REM sleep (Dexter and Weitzman, 1970; Manzoni et al., 1981), the stage in which OSA also tends to be most severe. One study found that among five patients with episodic CH and five with chronic CH, polysomnography demonstrated OSA in all the former and one of the latter (Kudrow et al., 1984). Two case reports have documented marked improvement in CH or elimination of CH when comorbid OSAwas treated (Buckle et al., 1993; Nath and Chervin, 1997). Larger studies of the relation between CH and OSA have not yet been published. To further explore whether OSA may trigger CH, we surveyed individuals with CH about the timing of their headaches and about symptoms of obstructive sleep apnea. Our hypothesis was that if OSA triggers CH, symptoms of OSA would be associated with distinct nocturnal patterns of CH. Although OSA cannot generally be diagnosed by symptoms alone, the latter have good predictive value for OSA and are useful in research when polysomnographic data are not available (Kump et al., 1994; Kapuniai et al., 1988; Crocker et al., 1990; Flemons et al., 1994; Haraldsson et al., 1992; Bliwise et al., 1991). C o rre s p o n d e n c e : Ronald D. Chervin, M.D., M.S., Sleep Disorders Center, University Hospital 8D8702, Box 0117, 1500 E. Medical Center Dr., Ann A r b o r, MI , Tel: , Fax: , chervin@umich.edu.

2 108 CHERVIN ETAL. METHODS Subjects Potential subjects were identified by a health system search of diagnostic codes for "migraine variant," by newspaper and radio advertisement, and by television publicity. We reviewed charts of patients with "migraine variant" to determine whether CH were likely to be present. Telephone and then personal interviews further defined headache status, and all adults who satisfied cluster headache criteria of the International Headache Society (Headache Classification Committee of the International Headache Society, 1988) were invited to participate in this study.a total of 36 subjects, 35 with episodic CH and one with the chronic form, signed informed consent and participated in this Institutional Review Board-approved study. Protocol All subjects were interviewed by one of the authors (STN), who systematically recorded responses on a sleep symptom questionnaire. Patients provided information on symptoms well-validated to have predictive value for OSA (Kump et al., 1994; Kapuniai et al., 1988; Crocker et al., 1990; Flemons et al., 1994; Haraldsson et al., 1992; Bliwise et al., 1991): whether they snored never (scored as 0), rarely (1), sometimes (2), almost always (3), or always (4); whether they snored loudly (1) or not (0); and whether apneas had been witnessed by a bed partner (1) or not (0). Each subject also completed a subjective questionnaire assessment of excessive daytime sleepiness, the Epworth Sleepiness Scale (ESS), which has been validated and reported to reflect severity of OSA; scores above 10 on a scale of 0 to 24 suggest excessive sleep propensity (Johns, 1991, 1993). Patients also completed a questionnaire about CH-related symptoms. On this questionnaire, patients specified one or more of six time periods during which CH usually began. The time period choices were labeled as follows: on waking, in the morning after waking, midday or afternoon, evening, first a typical s sleep, and second a typical s sleep. Analyses Continuous variables were summarized as means ±standard deviations. Generalized estimating equations were used to determine while accounting for correlation between multiple outcomes (timing measures of CH) within the same person (Diggle et al., 1994) whether patients were more or less likely to suffer from CH at any of several time periods than in the morning (SAS, SAS Institute Inc., Cary, NC). To assess whether patients whose CH commonly occurred in one time period also commonly reported CH in another particular period, we entered reported times of headache occurrence into a factor analysis with the principal component method and varimax rotation. However, factor analysis of non-normally distributed variables, though frequently performed, is not technically permissible; we therefore also computed odds ratios between occurrence of CH at different time periods. To test for associations between timing of CH and symptoms predictive of sleep apnea, we used logistic regression models with occurrence of CH during a particular time period as the outcome measure, and snoring frequency, loud snoring, witnessed apneas and ESS scores as explanatory variables. RESULTS Demographics Of the 36 subjects, 34 were men, and the mean age was 45.8±12.0 years. Circadian Timing of CH A question "Do your headaches wake you from sleep at?" elicited a positive response from 33 (92%) of the subjects. Sleep-related headaches CH that "usually" occurred during the nocturnal sleep period or on awakening were reported by 31 (86%) of the subjects. The proportion of subjects who reported that CH usually began in each time period is shown in Figure 1. Midday and the first the were each more often identified as periods of CH proclivity than were either awakening or morning (generalized estimating equations, p<.01 in each instance). Evening CH were also more common than awakening or morning CH (p=.0364 in each case). Headache proclivity during any part of the combined awakening and morning periods, reported by 39% of subjects, was still marginally less frequent than proclivity for CH at midday (p=.0465) or the first the (p=.0585). Factor analysis of the reported CH time periods (Table 1) indicated that patients tended to group into three categories: those whose CH usually occurred in the first the and at midday; those with CH in the second the and evenings; and those with CH mainly on awakening or in the morning. Similar patterns were observed, though results did not reach statistical significance, when odds ratios were computed between the different periods of CH occurrence Figure 1. Proportion of Subjects Who Reported that CH Usually Occurs in Each Time Period O. 7 O. 6 O. 5 O. 4 O. 3 O. 2 O. 1 O On Wa k i n g M o r n i n g M i d d a y E v e n i n g Time Periods First Half of Night Second Half of Night

3 CLUSTER HEADACHES AND OBSTRUCTIVE SLEEPAPNEA 109 Table 1. Rotated Factor Pattern Matrix From Factor Analysis for Reported Usual Time Periods of Cluster Headaches* On awakening Morning Midday Evening First Second Variance explained by each factor Factor Factor Factor * Values greater than.40 were considered to show large loading on the indicated factor and are shown in bold. On awakening Morning Midday Table 2. Odds Ratios for Occurrence of CH at Different Time Periods On Awakening Morning Midday Evening * First * 3.31 Second (Table 2). The association between CH in the first the and at midday showed a trend toward statistical significance (p=.096). Symptoms of OSA The average subject snored "sometimes" (2.1±1.3 on a scale of 0 to 4) and 39% snored almost always or always (habitual snoring). Loud snoring was reported in 60% of the subjects, and apneas had been observed by a bed partner in 20%. The mean ESS score was 7.3±4.5. Association of OSASymptoms with CH Timing Patterns A logistic regression model with presence or absence of sleep-related headaches as the outcome, and snoring frequency, loud snoring, ESS score and observed apneas as the explanatory variables, was not statistically significant (p=.32). However, models of whether CH usually began in individual time periods (outcomes) and the four sleep-related items (explanatory variables) showed statistically significant associations for CH in the first the (p=.026) and CH in the midday period (p=.025, Table 3). In each of these two models, after adjustment for other explanatory variables, snoring frequency showed the strongest associations with midday and first-half-of-the- CH, with odds ratios of 3.2 to 4.2 for a one-point increase on the snoring Likert scale. Figure 2 shows the proportions of subjects with and without each OSA symptom who reported CH in the first the. Evening DISCUSSION First Second *p< <p<.10 Table 3. Logistic Regression Models for Those CH Time Periods that were Significantly Associated with Symptoms of Sleep Apnea 0.75 This study shows that persons with CH almost always experience headaches during their nocturnal sleep periods, that nocturnal timing of headaches is associated with their diurnal 1 O. 9 O. 8 O. 7 Figure 2. Proportion of Subjects Who Reported CH in the First Half of the Night Symptom Pre s e n t Symptom A b s e n t First half of Midday Sleep Related Variables snoring loud apnea ESS snoring loud apnea ESS *LR=likelihood ratio Entire Model: LR* p-value β For Each Sleep- Related Variable: s.e p-value odds ratio O. 6 O. 5 O. 4 O. 3 O. 2 O. 1 O Habitual Snoring* Loud Snoring S l e e p i n e s s * * *chi square p-value<.05 for difference between groups **Sleepiness defined as Epworth Sleepiness Score >1 0 Observed A p n e a s

4 110 CHERVIN ETAL. timing, and that an early /midday proclivity for headaches is more common among those patients who report some symptoms of OSA, particularly habitual snoring. Our data serve to confirm previous reports, but also provide some of the first evidence of the extent to which obstructive sleep apnea a condition that affects almost one quarter of middle aged men, often without symptoms may influence the course of CH. Several previous reports on CH have emphasized the high frequency of nocturnal headaches. For example, in two series of 105 and 30 CH patients, about two-thirds reported mainly nocturnal headaches (Ekbom, 1970; Kunkle et al., 1952), and in another study of 24 patients in active cluster periods, 39 (51%) of 77 prospectively recorded headaches emerged from sleep (Russell, 1981). These data and consistent findings among our subjects suggest that morbidity in CH stems not only from pain but also from sleep disturbance. In addition to awakenings associated with headaches, insomnia can develop from the fear of sleep and associated paroxysms of pain (Sahota and Dexter, 1993). Previous reports on circadian patterns of CH have noted that attacks are particularly common within the first two hours of nocturnal sleep and also in the mid-afternoon (Ekbom, 1970; Manzoni et al., 1981). However, these studies combined data from multiple subjects rather than searching for patterns within individuals. We used factor analysis, a technique commonly used to determine which item-responses in questionnaires aggregate together, to show for the first time that those subjects who report CH in the first the tend to experience reoccurrence about 12 hours later, in the midday period, while subjects who report CH in the second the tend to experience reoccurrence the next evening. Previous authors have noted that a "refractory period" tends to follow individual CH (Ekbom, 1970), and our data raise the possibility that a physiological trigger for CH during the could influence timing of not only nocturnal headaches but diurnal headaches as well. In clinical practice, CH during sleep or wakefulness may be selectively averted by medications such as ergotamine and verapamil, respectively, sometimes without changing the timing of residual headaches in wakefulness and sleep. However, we are not aware of published data on timing of residual headaches under these circumstances. Exactly what a physiological trigger for CH might be has been debated, but the hypothesis that hypoxemia may be salient is supported by our finding that CH early in the sleep period are associated with symptoms of OSA. Oxygen desaturation in patients with OSA tends to be more frequent, prolonged and severe in REM sleep (Chervin and Aldrich, 1998), the first cycle of which generally occurs about 90 minutes after sleep onset. However, alternative explanations for a link between CH and REM sleep also exist. During REM sleep, oxygenation is often slightly lower than in non-rem sleep, even in the absence of OSA(Krieger, 1994). Increases in carbon dioxide levels, with apneas or even without apneas during REM sleep (Krieger, 1994), could also conceivably trigger CH, as could abrupt autonomic changes associated with both apneas and REM sleep (Somers et al., 1995, 1993). The similar frequencies of habitual snoring in our subjects (39%) and in population-based samples of men (44%) (Young et al., 1993) do not suggest that patients with CH are particularly prone to have OSA. Similar findings emerged from a case-control study of snoring and other characteristics among 78 CH patients (Italian Cooperative Study Group on the Epidemiology of Cluster Headache, 1995). In addition, our subjects mean ESS score was within the normal range (Johns, 1991) and did not suggest OSA was common. However, the only published polysomnographic study that included respiratory monitoring in CH patients found OSA in 6 of 10 subjects (Kudrow et al., 1984). Obstructive sleep apnea can be present, especially in milder forms, in patients who do not snore, and it is especially in milder OSAthat apneic events are likely to be linked to REM sleep (Chervin and Aldrich, 1998). Furthermore, in a recent study of the ESS among 237 patients initially suspected or confirmed to have OSA, the ESS did not reflect objective measures of sleepiness or severity of apnea (Chervin and Aldrich, 1999). Despite the absence of any indication in our data that OSA is particularly common, the association of OSAsymptoms with occurrence of headaches in the first the suggests that, at least in a subgroup of patients, OSA may trigger headaches and determine their circadian timing. This subgroup may not be small, since about one-quarter of patients with CH are likely to have OSA even if the prevalence of OSA among CH patients is no higher than that in the general male population (Young et al., 1993). Obstructive sleep apnea of mild severity can be associated with severe excessive daytime sleepiness (Chervin et al., 1995), hypertension (Young et al., 1997), or other symptoms (Engleman et al., 1997) that improve with treatment of the sleep disorder. Some evidence is now e m e rging to suggest that obstructive sleep-disordered breathing too "mild" to meet criteria for OSA may still cause excessive daytime sleepiness and hypertension (Guilleminault et al., 1993, 1996). Such data suggest that the proportion of patients with CH whose headaches are influenced by sleepdisordered breathing could substantially exceed one-quarter. F i n a l l y, sleep-disordered breathing can be treated effectively with continuous positive airway pressure or other means. Initial case reports suggest that treatment of comorbid OSA leads to marked improvement or disappearance of CH (Buckle et al., 1993; Nath and Chervin, 1997). This finding is consistent with data recently collected at a chronic headache clinic: most subjects who had sleep-related headaches were found to have sleep disorders, and in every such case, treatment of the sleep disorders led to improvement in the headaches (Paiva et al., 1997). The proportion of cluster headache patients who have sleep-related headaches is so high, at about 90%, that the recommendation to investigate all patients with sleep-related headaches for sleep disorders may not be useful for those with CH unless it is shown in the future that virtually all such patients, despite the absence of typical symptoms in some, still have OSA when tested. Our results suggest that the subgroup of patients whose sleep-related CH typically occur in the first the may be those most

5 CLUSTER HEADACHES AND OBSTRUCTIVE SLEEPAPNEA 111 likely to benefit from polysomnographic investigation. Future research should include polysomnographic studies of patients with CH to better assess the frequency of OSA, define presenting symptoms that signal the need for laboratory studies, confirm conclusions suggested at present by historical data, and explore the physiological effect of OSAon CH. ACKNOWLEDGMENTS This work was supported by the National Headache Foundation, University of Michigan GCRC grant M01- RR00042, and NINDS grant K02-NS REFERENCES 1. Bliwise DL, Nekich JC, Dement WC. Relative validity of selfreported snoring as a symptom of sleep apnea in a sleep clinic population. Chest 1991; 99: Buckle P, Kerr P, Kryger M. Nocturnal cluster headache associated with sleep apnea. A case report. Sleep 1993; 16: Chervin RD, Aldrich MS. The relation between MSLT findings and the frequency of apneic events in REM and NREM sleep. Chest 1998; 113: Chervin RD, Aldrich MS. The Epworth Sleepiness Scale may not reflect objective measures of sleepiness or sleep apnea. Neurology 1999; 52: Chervin RD, Kraemer HC, Guilleminault C. Correlates of sleep latency on the multiple sleep latency test in a clinical population. Electroencephalogr Clin Neurophysiol 1995; 95: Crocker BD, Olson LG, Saunders NA, Hensley MJ, McKeon, JL, Allen KM, Gyulay SG. Estimation of the probability of disturbed breathing during sleep before a sleep study. Am Rev Respir Dis 1990; 142: Dexter JD, Weitzman ED. The relationship of nocturnal headaches to sleep stage patterns. Neurology 1970; 20: Diggle PJ, Liang KY, Zeger SL. Analysis of Longitudinal Data. Oxford: Clarendon Press, Ekbom K. Patterns of cluster headache with a note on the relations to angina pectoris and peptic ulcer. Acta Neurol Scand 1970; 46: Engleman HM, Martin SE, Deary IJ, Douglas NJ. Effect of CPAP therapy on daytime function in patients with mild sleep apnoea/hypopnoea syndrome. Thorax 1997; 52: Flemons W W, Whitelaw WA, Brant R, Remmers JE. Likelihood ratios for a sleep apnea clinical prediction rule. Am J Resp Crit Care Med 1994; 150: Guilleminault C, Stoohs R, Clerk A, Cetel M, Maistros P. A cause of excessive daytime sleepiness: the upper airway resistance syndrome. Chest 1993; 104: Guilleminault C, Stoohs R, Shiomi T, Kushida CA, Schnittger I. Upper airway resistance syndrome, nocturnal blood pressure monitoring, and borderline hypertension. Chest 1996; 109: Haraldsson PO, Carenfelt C, Knutsson E, Persson HE, Rinder J. Preliminary report: validity of symptom analysis and daytime polysomnography in diagnosis of sleep apnea. Sleep 1992; 15: Headache Classification Committee of the International Headache Society. Classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988; 8(suppl 7): Italian Cooperative Study Group on the Epidemiology of Cluster Headache. Case-control study on the epidemiology of cluster headache I: etiological factors and associated conditions. Neuroepidemiology 1995; 14: Johns MW. A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep 1991; 14: Johns MW. Daytime sleepiness, snoring, and obstructive sleep apnea: the Epworth Sleepiness Scale. Chest 1993; 103: Kapuniai LE, Andrew DJ, Crowell DH, Pearce JW. Identifying sleep apnea from self-reports. Sleep 1988; 11: Krieger J. Breathing during sleep in normal subjects. In: Kryger MH, Roth T, Dement WC, eds. Principles and Practice of Sleep Medicine, 2 Edition. Philadelphia: W.B. Saunders, 1994: pp Kudrow L. Response of cluster headache attacks to oxygen inhalation. Headache 1981; 21: Kudrow L. Cluster headache: a review. Clin J Pain 1989; 5: Kudrow L, Kudrow DB. Association of sustained oxyhemoglobin desaturation and onset of cluster headache attacks. Headache 1990; 30: Kudrow L, Kudrow DB. The role of chemoreceptor activity and oxyhemoglobin desaturation in cluster headache. Headache 1993; 33: Kudrow L, McGinty DJ, Phillips ER, Stevenson M. Sleep apnea in cluster headache. Cephalalgia 1984; 4: Kump K, Whalen C, Tishler PV, Browner I, Ferrette V, Strohl KP, Rosenberg C, Redline S. Assessment of the validity and utility of a sleep-symptom questionnaire. Am J Resp Crit Care Med 1994; 150: Kunkle CE, Pfeiffer JB Jr, Wilhoit WM, Hamrick LW Jr. Recurrent brief headache in "cluster" pattern. Trans Am Neurol Associ 1952; 77: Manzoni GC, Terzano MG, Moretti G, Cocchi M. Clinical observations on 76 cluster headache cases. Eur Neurol 1981; 20:

6 112 CHERVIN ETAL. 29. Nath ST, Chervin RD. Improvement in cluster headache after treatment for obstructive sleep apnea. Sleep Res 1997; 26: Paiva T, Farinha A, Martins A, Batista A, Guilleminault C. Chronic headaches and sleep disorders. Arch Intern Med 1997; 157: Russell D. Cluster headache: severity and temporal profiles of attacks and patient activity prior to and during attacks. Cephalalgia 1981; 1: Sahota PK, Dexter JD. Transient recurrent situational insomnia associated with cluster headache. Sleep 1993; 16: Shen JM, Schaanning J, White L, Kruszewski P, Bjaanes E, Sjaastad O. Cluster headache: the ventilatory response to transient hypoxia with pure nitrogen. Headache 1993; 33: Somers VK, Dyken ME, Clary MP, Abboud FM. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest 1995; 96: Somers VK, Dyken ME, Mark AL, Abboud FM. Sympatheticnerve activity during sleep in normal subjects. N Engl J Med 1993; 328: Swanson JW, Yanagihara T, Stang PE, O'Fallon WM, Beard CM, Melton LJ 3rd, Guess HA. Incidence of cluster headaches: a population-based study in Olmsted County, Minnesota. Neurology 1994; 44: Young T, Evans L, Finn L, Palta M. Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middleaged men and women. Sleep 1997; 20: Young T, Palta M, Dempsey J, Skatrud J, Weber S, Safwan B. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993; 328: Young T, Peppard P, Palta M, Hla KM, Finn L, Morgan B, Skatrud J. Population-based study of sleep-disordered breathing as a risk factor for hypertension. Arch Intern Med 1997; 157: