Abstract
The aim was to investigate the comorbidity of chronic refractory headache with obstructive sleep apnoea syndrome (OSAs). Seventy-two patients (51 women and 21 men) with chronic and refractory headaches, whose headache occurred during sleep or whose sleep was accompanied by snoring, were submitted to polysomnography. Patients diagnosed with OSAs (respiratory disturbance index > 10) began continuous positive airway pressure (C-PAP) treatment and were followed up for ≥ 6 months. Twenty-one cases of OSAs were identified (29.2% of the total investigated, 13.7% of the women and 66.6% of the men). Headaches were classified into several headache disorders, medication overuse headache and cluster headache being the most prevalent (nine and six of the 21 cases, respectively). In one case (1.4% of the total sample, 4.7% of all the men), the criteria for hypnic headache were fulfilled. Multivariate regression analysis revealed that age, male gender and body mass index were associated with OSAs. C-PAP treatment improved both sleep apnoea and headache in only a third of the cases. Patients suffering from chronic refractory headache associated with sleep or snoring, in particular those who are also middle-aged, overweight men, should be considered for polysomnography. C-PAP treatment alone does not seem to improve headache, but further investigation is needed.
Introduction
Chronic headache is the most common diagnosis among out-patients attending specialist headache centres (40%) (1). Its coexistence with other conditions may be important, as other pathological conditions could trigger, transform or even cause de novo chronic headache. Numerous sleep disorders are associated with primary headache syndromes, including hypersomnia and obstructive sleep apnoea syndrome (OSAs) (2–7). The latter seems to have a prevalence, in the general population, of 4% in adult men and 2% in adult women (8). Stereotypically, the OSAs patient is a middle-aged, overweight man, who snores loudly, but the condition is encountered in both sexes, in all age groups, in normal and underweight individuals, and even in the absence of snoring. Previous reports have explored the relationship between OSAs and specific headache syndromes, such as cluster headache and hypnic headache (2, 7–9). Clinical observations in our Headache Outpatient Clinic have shown that OSAs is often associated with chronic headache syndromes resistant to standard treatments. To investigate this association we conducted a prospective study, aiming to identify the frequency and risk factors for comorbidity of OSAs with refractory chronic headache.
Patients and methods
Headache patients attending the Headache Outpatient Clinic of the Athens Naval Hospital were selected to undergo polysomnography (PSG) when their headaches were chronic (> 15 days with headache per month, for the past 6 months), did not respond to standard treatments (failure to achieve a 50% reduction in headache frequency following at least two different prophylactic treatments, each undertaken for ≥ 2 months), and when they fulfilled at least one of the following criteria: (i) headaches more frequently occurred early in the morning, or during sleep; and (ii) the patient reported daytime sleepiness, snoring, or both, that was confirmed by their partner. The PSG was conducted in accordance with the American Academy of Neurology's description of this technique for the assessment of sleep-disordered breathing (10) and OSAs was considered when the total apnoea and hypopnoea index (respiratory disturbance index) was ≥ 10. All patients were administered a semistructured interview, specifically tailored to the needs of a headache investigation, that included the Hamilton rating scales for anxiety (HA) and depression (HD). In addition, they all underwent complete neurological and physical examination, routine blood tests and brain imaging (computed tomography or magnetic resonance imaging) to exclude structural lesions and other pathologies. Patients with OSAs were treated with continuous positive airway pressure (C-PAP) and were followed up for a period of ≥ 6 months after PSG evaluation and start of C-PAP treatment. A headache diary was used to note headache characteristics and drug consumption. Treatment with C-PAP was considered effective when, on the basis of the patients' headache diaries—the patients' conditions after 3 and 6 months of treatment were compared with those recorded at baseline, i.e. 1 month prior to PSG—a reduction of ≥ 50% in the number of days with headache per month was achieved. After C-PAP treatment was started, the previous medical treatment was gradually withdrawn over a period of 1 month. If headaches worsened after prophylactic medication withdrawal despite the C-PAP treatment introduction, the previous preventive pharmacotherapy was reintroduced. In case that combined 3-month C-PAP and prophylactic medication treatment did not improve headache, the prophylactic drugs were changed. Ten clinical parameters were analysed for possible association with OSAs: age, sex, body mass index (BMI), duration of headache attacks, severity of pain (rated on an 11-point verbal scale, from 0 to 10), timing of headache attacks, location of the pain, presence of autonomic signs, concomitant disorders, and HA and HD scores.
Results
Obstructive sleep apnoea
Of the 72 chronic headache sufferers investigated (21 men, 51 women), OSAs was diagnosed in 21 (14 men and seven women, 29.2% of the total investigated, 13.7% of the women and 66.6% of the men) (Table 1). Snoring was reported by all the OSAs patients except three (cases 2, 7 and 16). Most of them were overweight (mean BMI 31.8 ± 3.6 kg/m2) and middle-aged or older (mean age 56.6 ± 8.1 years). Nine (42.9%) of the OSAs patients recorded high HD scores (> 18), and 14 (66.6%) high HA (> 18) scores, indicating additional comorbidity with depressive or anxiety disorder. Multivariate regression analysis revealed that sex (male), age, and BMI were independently associated with OSAs. Hypertension and mood disorders were the conditions most commonly comorbid with OSAs.
Demographic and clinical characteristics of headache patients with obstructive sleep apnoea syndrome
CH, Cluster headache; HA, Hamilton rating scale for anxiety score; HD, Hamilton rating scale for depression score; MOH, medication overuse headache; NDPH, new daily persistent headache; TTH, tension-type headache; ICGD, International Classification of Headache Disorders.
Only if tired.
Headache characteristics
The headaches were moderate in severity, without a specific localization. Onset of headache attacks was reported to occur during sleep or early in the morning (by definition in our cases), and to last from 30 min to several hours (mean duration 3 ± 2 h). Medication overuse headache (MOH) was diagnosed in nine of the 21 cases (42.8%), cluster headache (CH) in six (28.5%) and tension-type headache (TTH) in three (14.28%). MOH had migrainous features in six of the nine patients affected by this form (28.57% of the 21 OSAs cases). One OSAs patient (representing 4.76% of the 21 cases and 7.14% of the affected men) met the criteria for hypnic headache (11). Autonomic signs or symptoms were reported by nine of the OSAs patients (42.8%), the most common being conjunctival injection, followed by nasal congestion, rhinorrhoea, or both. In five OSAs patients, brains scans revealed non-specific T2 signal hyperintensities.
C-PAP treatment
In five of the 21 OSAs patients (23.8%), C-PAP treatment reduced the number of days with headache per month by ≥ 50%. In 14 (66.6%), headache frequency increased and drug therapy had to be reintroduced, whereas in the other two cases (9.5%) the headache frequency did not change. No patient discontinued C-PAP treatment during the study period. To identify predictive factors for response to C-PAP treatment we looked for possible associations with age, gender, BMI, onset, duration and location of headaches, presence of autonomic signs, snoring, apnoea, hypopnoea and total index, oxygen saturation, HA and HD scores and headache classification. None of these parameters was related with headache relief after C-PAP treatment monotherapy (χ2). Nocturnal headache was reported by six patients. One case was classified as hypnic headache (case 4, Table 1). With one exception (case 17) none of these patients benefited from C-PAP treatment and headaches needed additional pharmaceutical treatment. The C-PAP mask caused discomfort in all cases at 3 months, and 12 patients (57.14%) also reported discomfort at the second assessment (6 months). In all the patients presenting MOH, pharmacotherapy was needed in addition to C-PAP treatment. In all cases, concomitant C-PAP treatment allowed a reduction of drug consumption. Hypnic headache was unresponsive to oxygen, indomethacin and C-PAP treatment when these were given singly, but did respond to the combination of C-PAP and lithium.
Discussion
Much attention has been focused on the coexistence of OSAs and headache in recent years, and several studies have investigated the frequency of this comorbidity. They have all found that CH is often comorbid with OSAs (2–4). In particular, patients with CH are 8.4 times more likely to exhibit OSAs than healthy controls (3). Surprisingly, collective investigations of headache centre (12) and sleep centre (13) patients have failed to find an association. Among 903 headache sufferers attending an out-patient headache centre, symptoms suggesting OSAs were identified only in 45 and among them PSG revealed OASs (apnoea and hypopnoea index > 5) in only 14 patients (1.5% of the total study population) (12). This lack of association perhaps suggests that only rare headache disorders, such as CH, are related to OSAs. Clinical observations in our headache centre have encouraged us to investigate patients suffering from refractory chronic headaches because they have often shown symptoms suggestive of OSAs. Therefore, those patients with both refractory chronic headache and symptoms suggestive of OSAs were grouped together. Polysomnography revealed mild to severe OSAs in approximately one-third of patients investigated. As expected, older age, male gender and high BMI were independently associated with this condition. In OASs patients, the coexistent headache was most frequently MOH (half of cases), followed by CH (one-third of cases) and migraine (one-quarter of cases). One case of hypnic headache was found. Mild autonomic signs were present in almost half of the cases (not only in CH sufferers). Headache attacks were of moderate severity, usually lasting a few hours. Hypertension and mood disorders were commonly comorbid with OSAs and chronic headache. C-PAP treatment alone was effective in less than a quarter of patients, and the majority needed additional pharmacotherapy, those suffering from MOH in particular. No specific headache characteristic was associated with response to C-PAP monotherapy.
What does this study add to current knowledge?
Although the comorbidity of OSAs with CH is known, the results of the present study suggest that chronic refractory headache generally, regardless of its specific classification, may show an association with OSAs. In particular, in addition to CH, MOH may also be comorbid with OSAs. The effectiveness of C-PAP treatment in headache was not known. These results, although not controlled, suggest that C-PAP treatment alone does not improve headache and that pharmacotherapy is also needed. The causative association between headache and OSAs, if any, is elusive. It has been speculated that OSAs may cause CH, although it is not yet known whether hypoxia, hypercapnia or hypocapnia is the decisive trigger (14, 15). The lack of effectiveness of C-PAP treatment observed in this study suggests that OSAs does not have causative associations, and is probably a cofactor in the triggering of headache attacks in MOH and CH.
Limitations
This was an open-label, observational study investigating the effectiveness of C-PAP treatment in a fairly small number of patients. It could prove useful to carry out PSG in patients with chronic refractory headache, whether or not they present symptoms of OSAs, in order to see how often these conditions are comorbid, or to study the headache incidence in all patients sent for PSG and look for headache changes in those who adopt C-PAP treatment. The American Academy of Sleep Medicine has recently published guidelines for the evaluation of C-PAP treatment (16). However, our study predated these and therefore was not conducted in accordance with them.
Clinical importance
The findings of this study underline the need for further clinical and laboratory investigation of chronic headache patients who do not respond to standard therapies. In particular, the sleep habits of overweight, middle-aged (and older) male chronic headache sufferers who snore should be carefully examined and such patients should undergo a PSG to exclude OSAs. Not only CH, but also MOH may be comorbid with OSAs. Headache patients also suffering from OSAs seem more likely to respond to C-PAP treatment in combination with prophylactic pharmacotherapy. Thus, questions designed to detect the presence of OSAs (snoring, sleep apnoea, daytime sleepiness) should be included in regular diagnostic questionnaires for headache. Although C-PAP treatment alone may not improve headaches, it may help to prevent other severe conditions related to OSAs (17).