1007/s11940-009-0056-y
Headache
Opinion statement
Headache has been linked to a wide range of sleep disorders that may impact headache management. There are no evidence-based guidelines, but the authors believe that literature supports the following clinical recommendations:
1. Diagnose headache according to standardized criteria. Specific diagnoses are associated with increased risk for specific sleep and psychiatric disorders. 2. Collect sleep history in relation to headache patterns. Screening questionnaires and prediction equations are cost-effective. 3. Rule out sleep apnea headache in patients with awakening headache or higherrisk headache diagnoses (cluster, hypnic, chronic migraine, and chronic tensiontype headache); patients with signs and symptoms of obstructive sleep apnea warrant polysomnography and treatment according to sleep medicine practice guidelines. There is no evidence for suspending conventional headache treatment in suspected or confirmed cases of sleep apnea. Treatment of sleep apnea with CPAP may improve or resolve headache in a subset of patients. The impact on sleep apnea headache of other treatments for sleep apnea (eg, oral appliances, surgery, weight loss) is largely untested. At a minimum, sedative-hypnotic drugs should be avoided in suspected apneics until the sleep apnea is treated. 4. Among patients with migraine and tension-type headache, insomnia is the most common sleep complaint, reported by one half to two thirds of clinic patients. Patients who suffer from chronic migraine or tension-type headache may benefit from behavioral sleep modification. Pharmacologic treatment may be considered on a case-by-case basis, with hypnotics, anxiolytics, or sedating antidepressants used to manage insomnia, tailoring treatment to the symptom pattern. 5. Individuals with chronic headache are at increased risk for psychiatric disorders. Assessment for depression and anxiety may be warranted when either insomnia or hypersomnia is present. Psychiatric symptoms affect the choice of sedating versus alerting versus neutral pharmacologic agents for headache. 6. All headache patients, particularly those with episodic migraine and tension-type headaches, may benefit from inclusion of sleep variables in trigger management.
Headache
Introduction
Evidence that sleep and headache influence one another is based upon epidemiologic data, case-control studies, descriptive reports, clinical observations, and theory implicating common neuroanatomic systems [1]. Recent papers have explored brain mechanisms believed to underlie the sleepheadache relationship [2, 3]. Epidemiologic studies offer perhaps our most robust evidence of comorbidity between headache and sleep disorders, linking headache to a wide range of sleep disorders including snoring and obstructive sleep apnea (OSA), insomnia, circadian rhythm disorders, and parasomnias. Chronic daily headache not otherwise specified or a pattern of awakening frequently with any headache (termed awakening headache) is particularly suggestive of sleep disorders and occurs in 4% to 6% of the general population, 1860% of sleep apneics, 18% of insomniacs, and 21% of depressed individuals across studies [2]. OSA has been frequently studied and is particularly salient because of associated morbidity of sleep-related breathing disorders as well as its potential impact on headache management, as headache may improve after treatment. Likewise, obesity and snoring (sign and symptom of OSA) are risk factors for progression from episodic migraine to chronic daily headache and may contribute to the transformation or chronification of migraine [4]. OSA is the sleep disorder most often associated with awakening headache. Sleep apnea headache is the only formal diagnosis for headache secondary to a sleep disorder recognized by the International Classification of Headache Disorders, 2nd Edition (ICHD-II) [5]. The diagnosis is a subclassification of hypoxia-related headache under the major code heading Headache associated with metabolic disorder, which relates headache to measurable metabolic abnormalities; the headache would be expected to remit with normalization of the metabolic state [5]. The debate over pathophysiology is discussed elsewhere [6]. The diagnosis was introduced in 2004, and criteria have yet to be validated. There is some indication that a sizeable proportion of apneic patients with awakening headache would not fulfill the new ICHD-II criteria, as sleep apnea headache has been shown to present as migraine, tension, cluster, or nonspecific other headache [7]. An interesting feature of the diagnostic criteria states that sleep apnea headache can be definitively diagnosed only after successful treatment of the OSA. This is analogous to discontinuation of medication for the diagnosis of medication-overuse headache. Greater use of the sleep apnea headache diagnosis or other standardized definition would facilitate validation of the diagnosis and generalization across studies. Hypnic headache, the second sleep-related diagnosis recognized by ICHD-II, is by definition confined to sleep and is known to occur in the mid to latter portion of the night [5]. Polysomnography has not isolated hypnic headache to a specific sleep stage or sleep disorder, although anecdotal reports have suggested an association between hypnic headaches and OSA or REM-related oxygen desaturations [8]. Among patients with migraine and tension-type headache, insomnia is the most common sleep complaint. Improvements in general sleep hygiene are often clinically recommended for headache patients, but it is unclear if treatment of insomnia with either behavioral insomnia therapies or hypnotics improves headache. One randomized controlled trial demonstrated significant improvement in chronic migraine following a 12-week behavioral sleep regulation intervention, with headache change proportionate to the number of sleep behaviors changed (doseresponse relationship); patients in this study reported a high prevalence of sleep complaints but had not been diagnosed with insomnia per se [9]. Insomnia is also a symptom of depression and anxiety. Thus, clinicians should be mindful of the full painsleep mood constellation of symptoms. Sleep dysregulation (eg, sleep deprivation, oversleeping, disturbed sleep) is one of the most frequently cited acute headache triggers of migraine and tension-type headache, along with variables such as stress, menstruation, and fasting [10]. Review of this literature found lack of sleep endorsed as a trigger in 4874% of migraineurs in five studies and in 26 72% of patients with tension-type headache across studies. Oversleeping was reported as a trigger in 25% to 32% of migraineurs in three studies and 13% of tension-type headache patients in one study [11]. Prospective time-series analyses have confirmed a relationship between headache and sleep duration [12] and between headache and tiredness [13]. This paper highlights evidence for an association of sleep disorders with headache and recommends an evaluation and treatment algorithm for headache patients.
Snoring/obstructive sleep apnea Epidemiologic studies consistently find that snorers and apneics exhibit headache (especially daily headache) more frequently than controls in cross-sectional studies, with the relative risk increased at least twofold to threefold [1, 14]. Two specific subgroups appear at highest risk for OSA: patients with cluster headache or chronic headache. A study of 37 cluster headache patients who underwent polysomnography identified an 8.4fold increase in the incidence of OSA relative to controls matched for age and gender (58% vs 14%), and this risk increased more than 24-fold among patients with a body mass index (BMI) greater than 25 kg/m2 [15]. Another uncontrolled study of 31 cluster headache patients who underwent polysomnography observed OSA in 80% (25/31) of these patients [16]. Likewise, Sand et al. [17] reported that patients with OSA were seven times more likely to have chronic headache than the general population. Mitsikostas et al. [18] identified OSA in 29% (21/72) of patients with severe headache refractory to standard treatments; headache diagnoses included medication overuse headache, cluster headache, and chronic tension-type headache. Insomnia Insomnia is the most common sleep complaint among clinical headache populations, present in half to two thirds of migraineurs in neurology or specialty headache practices [10, 19]. Epidemiology has documented a relationship between sleep disturbance and headache. In a cross sectional study, Boardman et al. [20] identified a dose-response relationship between headache and sleep (eg, trouble falling or staying asleep, feeling tired or worn out): Among 2,662 respondents, headache frequency increased with the
Other sleep disorders There is some evidence linking headache to restless legs syndrome (RLS), excessive daytime sleepiness, circadian rhythm disorders, and various parasomnias. For example, dOnofrio and colleagues [22] found RLS to be greater among headache patients (n=200, of whom 114 had migraine without aura) compared with 120 controls (90 women and 30 men) matched for sex and age. Regardless of headache diagnosis, RLS was greater among headache sufferers than among controls (50% vs 33%). In both groups, RLS was associated with increased sleep complaints; depression and anxiety also were more frequent in those with RLS than in those without RLS. Daytime sleepiness appears increased in patients with migraine. Barbanti et al. [23] analyzed migraine patients in a case-control study and found questionnaire scale elevations indicative of pathologic dayt i m e s l e e p i n es s i n a b o u t a t h r e e f o l d g r e a t e r proportion of migraineurs (14%) than controls (5%). Notably, sleepiness was also related to anxiety, poor sleep quality, migraine disability, and migraine phase. Interestingly, about 10% of patients indicated that excessive sleepiness was present at some point of the migraine prodrome, headache, or recovery phase, and 7% reported sleepiness during all three migraine phases. This finding suggests an alteration in alertness intrinsic to the acute migraine process. Likewise, circadian rhythm disorders and parasomnias (eg, nightmares, bruxism, sleepwalking) are at least twice as frequent among individuals with daily headache as among controls [24].
Treatment
& We believe the literature supports the following recommendations (Fig. 1) [25].
Headache
Figure 1. Suggested algorithm for management of sleep-related headache. ICHD-IIInternational Classification of Headache Disorders, 2nd edition; OSAobstructive sleep apnea. (Adapted from Rains [25], with permission)
&
Recommendations are provided according to ICHD-II primary and secondary headache diagnoses and based upon diagnosis-specific risk for sleep and psychiatric disorders.
&
A brief 24-h sleep history includes presleep routine, sleep period (sleep latency, duration of sleep relative to time in bed, mid-cycle and early morning awakenings), nocturnal symptoms (eg, respiratory symptoms, movement, waking), daytime function (eg, napping, alertness versus sleepiness, fatigue), and measures to promote sleep or wakefulness (eg, hypnotics, caffeine, alcohol, nicotine). This his-
tory may be supplemented by a bed partner report and questionnaires to assess sleep apnea, insomnia, RLS, etc. [6].
Rule out sleep apnea headache in high-risk headache diagnoses
&
&
&
&
&
Patients diagnosed with cluster, hypnic, chronic migraine, or tensiontype headache, or any headache with usual onset during or at the termination of sleep (awakening headache) warrant screening for sleep apnea and other major sleep disorders. Sleep apnea headache is suspected in cases of new-onset headache or exacerbation of a preexisting primary headache [5]. Diagnostic criteria include headache that is present on awakening that remits and does not recur within 72 h of effective treatment for sleep apnea. Confirmation of the diagnosis of OSA includes a Respiratory Disturbance Index greater than 5 per hour on polysomnography. Empirically, sleep apnea headache has been shown to present as either bilateral (53%) or unilateral (47%); location may be frontal (33%), frontotemporal (28%), or temporal (16%); pain quality is usually pressing/tightening (79%); and pain intensity varies from mild (47%), to moderate (37%), to severe (16%) [7]. A diagnosis of cluster headache is associated with an 8.4-fold increased risk for OSA (24-fold when BMI 9 25 [15]); thus, although review of symptoms and risk factors is encouraged, the threshold for polysomnography is probably met based on diagnosis alone and the potential to improve cluster headache control with treatment for OSA [26]. Likewise, hypnic headache probably warrants polysomnography based on anecdotal evidence implicating sleep-disordered breathing. Risk for OSA may be estimated using tools such as the Berlin Sleep Questionnaire [27], which identifies patients at high risk for OSA versus low-risk patients in the primary care environment based on the patients neck circumference, habitual snoring or witnessed apnea, and hypertension (Fig. 2). Its sensitivity is 86%, specificity is 77%, positive predictive value is 89%, and likelihood ratio is 3.79. Patients with suspected OSA warrant polysomnography and treatment according to guidelines based on sleep medicine evidence [28]. Signs and symptoms, as well as treatment options, are summarized below.
Screening procedures
Headache
Figure 2. The risk of obstructive sleep apnea (apnea-hypopnea index [AHI] 9 10) can be estimated from the patients Sleep Apnea Clinical Score, based on the patients neck circumference, historical features (habitual snoring and/or partner reports of gasping, choking, or snorting), and hypertension. A clinical score 10 indicates a low likelihood ratio (1.09; 95% CI, 0.62 1.92), a score 9 10 indicates moderate probability (2.03; 95% CI, 0.944.38), and a score 9 15 indicates high probability (5.17; 95% CI, 2.5410.51). (Adapted from Flemons et al. [27], reprinted with permission of the American Thoracic Society. Copyright American Thoracic Society)
Polysomnography (portable and unattended monitoring may be available). Continuous positive airway pressure (CPAP) surgery (eg uvulopalatopharyngoplasty, tonsillectomy, tracheostomy) oral appliances (tongue retaining devices mandibular advancement) conservative measures (positional therapy to avoid supine sleep weight loss avoidance of alcohol muscle relaxants, anxiolytics, sedatives, hypnotics, narcotics, and other medications with respiratory suppression).
&
&
Chronic headache sufferers should be screened for sleep disorders, particularly in cases refractory to standard treatments such as headache prophylaxis and withdrawal from medication overuse. Patients with episodic migraine and tension-type headache may benefit from screening for sleep disorders when history suggests a significant sleep complaint or daytime sleepiness. Patients with chronic migraine and tension-type headache should be encouraged to apply behavioral sleep regulation, which has been shown to significantly improve chronic migraine [9]. More extensive behavioral insomnia treatments are described below [29]. Although there are no studies demonstrating that headache improves solely with pharmacologic treatment for insomnia, pharmacologic treatment may be warranted on a case-by-case basis. Benzodiazepines and non-benzodiazepine hypnotics are effective and there is some evidence that antidepressant medications also are effective, although the benefits must be weighed against potential adverse effects [30]. There are no data to recommend a particular
sleep aid for headache patients from the available hypnotic, anxiolytic, or sedating antidepressants; treatment should be tailored to symptom patterns.
Screen for psychiatric comorbidity
&
Individuals with chronic migraine and tension-type headache are at increased risk for psychiatric disorder [31], and the threshold for screening is met when either insomnia or hypersomnia is present. Psychiatric symptoms affect the choice of pharmacologic agents (sedating vs alerting vs neutral) [32].
&
All headache sufferers, particularly those with episodic headache, may benefit from inclusion of sleep variables in trigger management. Prospective sleep diaries track the regularity, duration, and quality of sleep as a headache trigger. Comprehensive diaries can link headache to triggers such as sleep, stress, mood, and diet [6].
&
&
&
&
Headache
severe, and avoidance of substances known to suppress respiration. Weight loss has been shown to improve the severity of sleep apnea. The degree of weight loss necessary to resolve sleep apnea is generally unknown. In clinical experience, some patients can recall a specific previous weight when they were asymptomatic and can correlate the onset of symptoms with a particular weight gain; such patients may target their weight loss to achieve a weight at which they were asymptomatic. A 10% change in body weight was shown to improve sleep apnea severity by 26% in a large epidemiologic study [39]. Because of the time required for significant weight loss and the unpredictability of conservative treatments for OSA, they are unlikely to be the treatment of choice for sleep apnea headache. Continuous positive airway pressure
Standard procedure Referral for polysomnographic evaluation is needed to confirm the diagnosis and initiate appropriate treatment for OSA. CPAP or other positive airway pressure treatments (eg, bilevel, auto) are generally titrated and evaluated in a systematic manner in a sleep laboratory or under other controlled conditions. A prescription for home use is based on laboratory response. Current home CPAP devices are often equipped with adherence-tracking devices and other indices to assess effectiveness (eg, leak, estimates of breakthrough snoring or apnea). Generally, patients initial response to treatment is assessed subjectively by interview and objectively by data retrieved from the CPAP tracking device after 1-3 months of treatment. Relevant outcomes may include impact on awakening headache, sleep quality, daytime sleepiness, and comorbid conditions or symptoms (eg, hypertension, depression, memory). Once the clinical end point is achieved, long-term follow-up by the sleep specialist for established CPAP users is an annual or semiannual office visit. Although headache may improve with treatment of sleep apnea, there is no basis for withholding headache treatment during the process of sleep evaluation or treatment. Concurrent headache interventions are usually necessary. Treatment of headache that persists after resolution of sleep-disordered breathing depends on the exact headache diagnosis, but standard headache treatments for the specific headache diagnosis will apply. Reevaluation of headache diagnosis and severity should occur after 1 month of sleep apnea treatment. It is prudent to avoid sedation with hypnotics or opiates until the breathing disorder is adequately treated. After adherence to treatment is established, narcotic and sedative-hypnotic drugs can be considered.
Special points
Complications
&
and range of motion exercises); both groups received usual medical care. Patients in the treatment and control groups reported headache frequency 24.6 days and 25.9 days per month, respectively, and more than 70% of each group was described as having a history of medication overuse. By the sixth week of treatment, 35% of the treatment group reverted to episodic headache, compared with none of the controls. After the 12-week treatment, significant improvements were observed in the treatment group versus the control group in headache frequency (39% vs 12%) and headache index (28% vs 3%). Notably, the improvement in headache was proportionate to the number of sleep behaviors changed. Though preliminary, this randomized controlled trial provides compelling evidence that a relatively brief behavioral sleep intervention delivered in the headache practice setting can improve headache outcome. The Calhoun and Ford intervention [9] includes five components: 1) Schedule a consistent bedtime that allows 8 h in bed; 2) Eliminate watching television, reading, or listening to music in bed; 3) Use visualization technique to shorten time to sleep onset; 4) Have supper at least 4 h before bedtime and limit fluids within 2 h of bedtime; 5) Discontinue naps.
&
10
Headache
&
controlled trials with either Class I or Class II evidence demonstrating effectiveness [29]. In the absence of empirical guidance, conventional treatment of insomnia with standard behavioral and pharmacologic treatments should be considered on a case-by-case basis. Patients with irregular sleep schedules or poor sleep habits and those spending less than 6.5 or more than 8.5 h in bed may benefit from sleep regulation as in the Calhoun and Ford study [9], with more extensive insomnia interventions as needed.
&
&
&
&
Behavioral and pharmacologic treatments for sleep regulation are compatible with headache care. Individuals already receiving behavioral treatment for headache may be participating in some form of relaxation training or cognitive behavioral therapy; sleep elements can be easily integrated into behavioral headache treatments. Welldeveloped behavioral and cognitive-behavioral interventions for insomnia are available and have been tailored for headache patients [43, Class IV]. A two-session behavioral insomnia intervention with demonstrated efficacy in the primary care setting is also appropriate for the headache clinic setting [42, Class I]. Participants receive these instructions: 1) eliminate sleep-incompatible activities (television watching, reading, planning, worrying) in the bed and bedroom; 2) avoid all daytime napping, and 3) follow a consistent sleep-wake schedule by adhering to an agreed-upon bedtime and rising time, which are negotiated after determining the average sleep time that the patient reported during 2 weeks of sleep-log monitoring. Each participant selects his or her standard bedtime and rising time subject to the requirement that the interval between these two times equals the pretreatment average sleep time plus 30 min. The choice of behavioral treatment is assigned based on presenting symptoms or sleep diary information. More strictly behavioral treatments (sleep restriction, stimulus control, relaxation training) are indicated for patients who spend excessive time in bed not sleeping, who have irregular sleep schedules, who engage in sleepincompatible behaviors while in bed, or who exhibit evidence of hyperarousal. Cognitive therapy is indicated for patients with unrealistic expectations concerning sleep or the consequences of not sleeping, or with anticipatory anxiety, worry, or poor coping skills. General sleep education, often termed sleep hygiene, is useful in the initial phases of treatment for nearly all patients, particularly those exhibiting inadequate sleep habits, environment, or lifestyle. Sleep restriction uses sleep deprivation to heighten sleep drive and may temporarily increase daytime sleepiness. Patients should be warned about sleepiness, especially if driving or engaged in other
11
hazardous activities. Over successive weeks, the sleepiness will remit as time in bed is gradually increased and the individuals maximum efficient sleep time is achieved. Clinicians may avoid using sleep restriction in patients with conditions that may be exacerbated by sleep deprivation, such as seizure or bipolar disorder. Some headache patients express concerns that limiting time in bed will worsen their headaches. We have not observed this to occur, but it may be preferable to initiate behavioral treatment in these patients with relaxation training or a strategy other than sleep restriction.
Pharmacologic treatment
&
&
&
&
Increasing sleep time through the use of pharmacologic agents has not been proven to improve headache. Hypnotics should be considered only after an assessment of the underlying cause of the insomnia. Sedative-hypnotic drugs can worsen comorbid conditions, especially OSA and depression. Likewise, other sleep disorders (eg, RLS, delayed sleep phase) may present as insomnia but have their own specific treatments. Pharmacologic treatment of chronic insomnia should be combined with behavioral measures. In most cases, either intermittent use of the sedative-hypnotic drug or nightly use for a short time is desired, although long-term treatment may be appropriate for select patients. The best initial choice of a sedative-hypnotic depends on the nature of the insomnia, comorbid medical and psychiatric diagnoses, prior experience with hypnotics, substance abuse history, and pharmacokinetic properties of the drug (Tmax, half-life, side effect profile, dependency risk). Drug choice also depends on whether the insomnia pattern involves sleep onset, maintenance (frequent awakenings during the night), or early morning awakening. Medications approved by the FDA for insomnia include benzodiazepines (triazolam, temazepam, estazolam, flurazepam, quazepam), selective benzodiazepine receptor agonists (zolpidem, zolpidem CR, zaleplon, eszopiclone), and a melatonin receptor agonist, ramelteon. All of the these medications have demonstrated efficacy in multiple Class I trials for short-term treatment of insomnia, with improvement in sleep latency, total sleep time, and/or wakefulness after sleep onset [41]. Antidepressants (eg, trazodone, mirtazapine, doxepin, amitriptyline) are often used off-label.
Detailed pharmacokinetic data and risk profiles are available elsewhere [41]. For sleep onset insomnia, zaleplon and ramelteon are indicated, as they have very short half-lives and principal impact on reducing the latency to sleep onset, but they do not increase the tendency to maintain sleep. Alternatively, short-acting to intermediate-acting agents are indicated for sleep onset and sleep maintenance insomnia (eg, eszopiclone,
12
Headache zolpidem, zolpidem CR, estazolam, temazepam). Use of agents with long half-lives (eg, flurazepam, quazepam) is generally discouraged because of morning residual effect, but sometimes these agents can be helpful for patients with anxiety because they exert an anxiolytic effect well into the next day.
13
3.
4.
&
Practicing self-management skills. Patients may modify their sleep environments, schedule, or behaviors (eg, sleep hygiene, relaxation training). They also may eliminate avoidable triggers (eg, dietary triggers) and predict and prepare for uncontrollable triggers (eg, weather, menstrual cycle). Trigger management can be carried out with minimal cost to staff and patients. Patients often respond positively to trigger management even when they may resist a more psychological or cognitive-behavioral treatment.
14
Headache
Disclosure
No potential conflicts of interest relevant to this article were reported.
12.
13.
14.
15.
16.
17. 18.
19.
20.
21.
22.
23.
15
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Davis JA, Fine ED, Maniglia AJ: Uvulopalatopharyngoplasty for obstructive sleep apnea in adults: clinical correlation with polysomnographic results. Ear Nose Throat J 1993, 72(1):6366. 38. Abad VC, Guilleminault C: Treatment options for obstructive sleep apnea. Curr Treat Options Neurol 2009, 11:358367. 39. Peppard PE, Young T, Dempsey J, Skatrud J: Longitudinal study of moderate weight change and sleepdisordered breathing. JAMA 2000, 248:30153021. 40. Smith MT, Perlis ML, Park A, et al.: Comparative meta-analysis of pharmacotherapy and behavior therapy for persistent insomnia. Am J Psychiatry 2002, 159(1):511. 41. Schutte-Rodin S, Broch L, Buysse D, et al.: Clinical guideline for the evaluation and management of chronic insomnia in adults. J Clin Sleep Med 2008, 4 (5):487504. This paper provides a thorough review of chronic insomnia and evidence-based guidelines for its treatment. 42. Edinger JD, Sampson WS: A primary care friendly cognitive behavioral insomnia therapy. Sleep 2003, 2:177182. 43. Rains JC: Optimizing circadian cycles and behavioral insomnia treatment in migraine. Curr Pain Headache Rep 2008, 12:213219. 44. Cohen GL: Protriptyline, chronic tension-type headaches, and weight loss in women. Headache 1997, 37(7):433436. 45. Feren S, Katyal A, Walsh JK: Efficacy of hypnotic medications and other medications used to treat insomnia. Sleep Med Clinics 2006, 1:387397. 46. James SP, Mendelson WB: The use of trazodone as a hypnotic: a critical review. J Clin Psychiatry 2004, 65:752755. 47. Rains JC, Penzien DB, McCrory DC, et al.: Behavioral headache treatment: history, review of the empirical literature, and methodological critique. Headache 2005, 45(Suppl 2):S92S109. 48. Lipchik GL, Smitherman TA, Penzien DB, et al.: Basic principles and techniques of cognitive-behavioral therapies for comorbid psychiatric symptoms among headache patients. Headache 2006, 46 (Suppl 3):S119S132. 49. Nash JM, Park ER, Walker BB, et al.: Cognitive-behavioral group treatment for disabling headache. Curr Headache Rep 2003, 2:475481. 50. Martin VT, Behbehani MM: Toward a rational understanding of migraine trigger factors. Med Clin North Am 2001, 85(4):911941.