The Epidemiology of Narcolepsy in Olmsted County, Minnesota: A Population-Based Study

Transcription

1 NARCOLEPSY The Epidemiology of Narcolepsy in Olmsted County, Minnesota: A Population-Based Study Michael H. Silber MB, ChB, 1 Lois E. Krahn MD, 2 Eric J. Olson MD, 3 V. Shane Pankratz PhD Sleep Disorders Center, 1 Departments of Neurology; 2 Psychiatry; 3 Internal Medicine; and 4 Health Sciences Research, Mayo Clinic and Mayo Medical School, Rochester, Minnesota Study Objectives: To determine the age- and sex-specific incidence rates and prevalence of narcolepsy in a United States community. Design: The records-linkage system of the Rochester Epidemiology Project was utilized to ascertain all patients with narcolepsy seen in Olmsted County, Minnesota between 1960 and Age- and sex-specific incidence rates were calculated, using census data. Prevalence of narcolepsy on January 1, 1985 was calculated. Setting: N/A Patients or Participants: Community patients diagnosed with narcolepsy by a validated set of diagnostic criteria. INTRODUCTION DESPITE INTENSE RECENT INTEREST IN THE PATHO- GENESIS OF NARCOLEPSY, 1 THE INCIDENCE AND PREVALENCE OF THE DISORDER REMAIN UNCERTAIN. A number of prevalence studies have been reported using a variety of methodologies but, to the best of our knowledge, there have been no incidence studies. In 1957, Roth estimated the prevalence in Czechoslovakia to be between 20 and 30 per 100, Two prevalence studies, published only in abstract form, were conducted in the United States. The first, based on responses to a newspaper advertisement, estimated the minimum prevalence of narcolepsy with cataplexy as 50 per 100, The second, based on responses to a television broadcast, estimated a minimum prevalence of 67 per 100, The prevalence of narcolepsy with cataplexy in Japan was estimated at 160 per 100,000, based on a questionnaire and short personal interview with 478 teenagers. 5 Based on sleep laboratory data and questionnaires circulated to all neurologists in the country, the prevalence of narcolepsy in Israel was estimated at between 1:500,000 and 1:660, The Ullanlinna Narcolepsy Scale, a validated questionnaire for the diagnosis of narcolepsy, was used to identify possible narcoleptics in a large cohort of Finnish twins. 7 These patients were interviewed and narcolepsy eventually confirmed by sleep studies in three patients. A prevalence for narcolepsy Disclosure Statement The authors have no financial involvement in any organization with a direct commercial interest in the subject matter discussed. No off-label or investigational use is discussed. Submitted for publication October 2001 Accepted for publication December 2001 Address correspondence to: Michael H. Silber, M.B.Ch.B., Department of Neurology, Mayo Clinic, 200 1st Street SW, Rochester, MN, 55905; Tel: ; Fax: ; msilber@mayo.edu SLEEP, Vol. 25, No. 2, Interventions: N/A Measurements and Results: The incidence rate per 100,000 persons per year was 1.37 (1.72 for men and 1.05 for women). The incidence rate was highest in the 2nd decade, followed in descending order by the 3rd, 4th and 1st decades. The prevalence on January 1, 1985 was 56.3 per 100,000 persons. Approximately 36% of prevalence cases did not have cataplexy. Conclusions: Narcolepsy is not a rare disorder. It appears to be commoner in men. It originates most commonly in the 2nd decade. Narcolepsy without cataplexy is an important subgroup, warranting further study. with cataplexy of 26 per 100,000 population was calculated. A study in Hong Kong, based on sleep laboratory data, estimated the overall prevalence as between one and 40 per 100,000 with 60% of the prevalence cases having cataplexy. 8 Following a telephone questionnaire of a nationwide sample of 4972 subjects, a prevalence of narcolepsy with cataplexy in the United Kingdom of 40 per 100,000 was calculated. 9 The aims of the present study were to determine the age- and gender-specific incidence rates and prevalence for narcolepsy, with and without cataplexy, in Olmsted County in Minnesota. The study was approved by the Mayo Institutional Review Board. METHODS Study Population Olmsted County is located in southeastern Minnesota and has a population primarily of northern, western, and central European descent. Over 70% of the population lives in Rochester, the centrally located county seat, which lies approximately 90 miles southeast of Minneapolis. The local economy is based on farming, health care, and light industry. To be included in the incidence study, subjects had to reside within the county limits at the time of onset of their narcolepsy (either sleepiness or cataplexy, whichever came first); patients who may have moved to the county for management of previously diagnosed narcolepsy were thus excluded. Incidence data was based on patients whose narcolepsy commenced between 1960 and We included in the prevalence study only living patients with already present narcolepsy who resided in the county on the prevalence date, January 1, Case Ascertainment We ascertained patients through the records-linkage system of the Rochester Epidemiology Project. 10 Medical care in Olmsted

2 Table 1 Mayo classification of narcolepsy Category A Definite Narcolepsy History of excessive daytime sleepiness. History of cataplexy, defined as definite bilateral weakness of brief duration brought on by emotion. Mean initial sleep latency of <8 minutes on MSLT *. Two or more SOREMP on MSLT *, OR 1 SOREMP on MSLT * and 1 SOREMP on the preceding nocturnal PSG. Apnea-hypopnea index (AHI) < 10 per hour on nocturnal PSG preceding the MSLT*. (The last 3 criteria can be replaced by cataplexy witnessed by a physician with documented recoverable areflexia, or cataplexy recorded by polysomnography (PSG) and video recording.) Category B Subgroup B1 Probable Narcolepsy (Laboratory Confirmation) History of excessive daytime sleepiness. A history of cataplexy, defined as definite bilateral weakness of brief duration brought on by emotion. Mean initial sleep latency of < 8 minutes on MSLT *. One or fewer SOREMP on MSLT * or on the preceding nocturnal PSG. AHI <10 per hour on the nocturnal PSG preceding the MSLT *. Subgroup B2 History of excessive daytime sleepiness. No history of cataplexy. Mean initial sleep latency of < 8 minutes on MSLT *. Two or more SOREMP on MSLT or 1 SOREMP on the MSLT * and 1 SOREMP on the preceding nocturnal PSG. AHI <10 per hour on the nocturnal PSG preceding the MSLT *. Category C Probable Narcolepsy (Clinical) History of excessive daytime sleepiness History of cataplexy, defined as definite bilateral weakness of brief duration brought on by emotion. No or inadequate sleep studies performed * MSLT performed under standard conditions, including a total sleep time of 6 hours on the preceding night PSG [17] County is predominantly provided by the Mayo Clinic at primary, secondary, and tertiary levels, but other providers include a smaller independent clinic and its associated hospital, and several independent practitioners. All these providers participate in the Rochester Epidemiology Project, which supplies the infrastructure for indexing and linking essentially all medical information in the county. Medical diagnoses and other key clinical information from medical records are routinely entered into computerized indices using the hospital version of the International Classification of Diseases, Adapted (H-ICDA). 11 These indices can be searched for all medically diagnosed cases of a given condition. We searched the indices for the following seven H-ICDA diagnostic codes: narcolepsy or cataplexy; cataplexy; sleep epilepsy; sleeping disease; Gelineau s syndrome; narcolepsy; brain disease not otherwise specified. All medical records of each potential case were reviewed by a clinical neurophysiology technologist, with special training as a data abstractor for the project. All physician authors were diplomates of the American Board of Sleep Medicine. A physician (MHS) reviewed all charts of patients who fulfilled residency requirements within the specified time periods. He determined whether the patients fulfilled our criteria for narcolepsy, assigned them to subtypes (see below), and verified the date of onset of symptoms. When any of these were uncertain, another physician independently reviewed the chart (LEK) and a consensus was reached. We also obtained the records of a neurologist who, after his retirement from the staff of the Mayo Clinic, maintained in the region a private practice devoted to narcolepsy from 1982 to 1995, but whose records were not included in the records-linkage system. These records were reviewed in the same way. We also searched the recordslinkage index for the following thirteen H-ICDA diagnoses, as a check that we had not inadvertently missed cases of narcolepsy: sleepiness, not otherwise specified; sleepiness, cause specified; hypnogenic episode; disorder of excessive somnolence; hypersomnia, not otherwise specified; hypersomnolence, not otherwise specified; somnolence; psychophysiological hypersomnia; psychophysiological hypersomnolence; sleep drunkenness; Kleine- Levin syndrome; Prader-Willi syndrome. Four hundred of the 781 charts identified were reviewed by one of the physician authors (EJO). Because it may take several years for patients with narcolepsy to seek medical attention, we reviewed all medical SLEEP, Vol. 25, No. 2,

3 Table 2 Incidence Rates of Narcolepsy (Categories A, B, and C) Incidence Cases Incidence Population Incidence Rates ( X 100,000) , , , , , , , , , , , , , , , , , , ,950 70, , ,411 40, , ,511 14,967 47, ,111 1,928 8, Total ,339,133 1,219,036 2,558, records for a period of six years following termination of the incidence study ( ) and 11 years following the prevalence date ( ). Diagnostic Criteria Incidence Rates of Narcolepsy with Cataplexy (Categories A, B1, and C) Incidence Incidence Population Incidence Rates (X 100,000) , , , , , , , , , , , , , , , , , , ,950 70, , ,411 40, , ,511 14,976 47, ,111 1,928 8, Total ,339,133 1,219,036 2,558, Existing criteria for the diagnosis of narcolepsy include the criteria of the International Classification of Sleep Disorders (ICSD) 12 and the Diagnostic and Statistical Manual of Mental Disorders-IV criteria. 13 These are complex, somewhat unclear, or have not been validated. Therefore we developed new research criteria for the diagnosis of narcolepsy, and tested their validity and inter-rater reliability. 14 The criteria are summarized in Table 1. In summary, cases were divided into Category A (Definite Narcolepsy), Category B1 (Probable Narcolepsy Laboratory Confirmation), and Category C (Probable Narcolepsy Clinical). Category B was subdivided into category B1 (objective sleepiness, cataplexy but no sleep onset REM periods [SOREMS]) and category B2 (objective sleepiness, SOREMS, but no cataplexy). Categories A, B1, and C comprised cases of narcolepsy with cataplexy, while category B2 comprised cases of narcolepsy without cataplexy. All cases were classified according to this system. It was also noted whether or not cases fulfilled the diagnostic criteria of the ICSD. 12 Data Analysis We obtained population counts for Olmsted County, disaggregated by gender and age, from U.S. Census data for the years 1960, 1970, 1980, and We obtained counts for the intercensual years by assuming a linear growth of the population between censuses. We calculated incidence rates using incident cases between January 1, 1960 and December 31, 1989 as the numerator and population counts as the denominator (personyears). Prevalence was calculated using cases resident in Olmsted County on the prevalence date of January 1, 1985 as the numerator and the interpolated census population count on that date as the denominator. 15 We computed separate incidence rates and prevalence for the different diagnostic categories we used. Age- and gender-specific incidence rates were calculated. Since the study involved the target population in its entirety, no sampling was involved, and we did not use statistical tests to interpret our findings. 16 RESULTS Medical records of 325 patients with the seven most likely H- ICDA diagnostic codes were reviewed. A total of 72 patients fulfilling our diagnostic, residency, and time criteria were identified, 71 from the record-linkage system and one from the records of the private practice neurologist specializing in narcolepsy. These SLEEP, Vol. 25, No. 2,

4 4 All Narcolepsy Cases (Categories A, B & C Narcolepsy with Cataplexy (Categories A, B1 & C) Incidence per 100,000 per year Figure 1 Incidence of narcolepsy in Olmsted County, MN included 55 patients fulfilling criteria for the prevalence study and 35 fulfilling criteria for the incidence study. No cases fulfilling our criteria for narcolepsy were identified among the 400 records linked to the 13 additional H-ICDA codes. All patients were Caucasian. The mean time from onset of sleepiness to onset of cataplexy in the Group A, B1, and C patients (narcolepsy with cataplexy) was 5.3+/-9.2 years. The age- and sex-specific average annual incidence rates for narcolepsy as a whole and narcolepsy with cataplexy are given in Table 2. The average incidence rate per 100,000 persons per year for narcolepsy (all categories) between 1960 and 1989 was 1.37 (men 1.72 and women 1.05). Calculated per decade, the incidence rate was 1.10 between 1960 and 1969, 1.62 between 1970 and 1979, and 1.34 between 1980 and The incidence rate for narcolepsy with cataplexy (categories A, B1, and C) was 0.74 (men 0.82 and women 0.67). Using ICSD criteria, the incidence rate was The incidence rates for narcolepsy as a whole (categories A, B, and C), narcolepsy with cataplexy (categories A, B1, and C) and narcolepsy by ICSD criteria were all highest in the second decade, followed by the third, fourth and first decades in descending order. (See Figure 1). The age- and sex specific prevalence of narcolepsy as a whole and narcolepsy with cataplexy in Olmsted County, Minnesota on January 1, 1985 is given in Table 3. The overall prevalence of narcolepsy was 56.3/100,000 persons (men 72.5, women 41.3). The prevalence for narcolepsy with cataplexy (categories A, B1, and C) was 35.8 (men 42.7, women 29.5). Using ICSD criteria, the prevalence was In similar fashion to the incidence rates, the prevalence was highest in the second decade. Years of Age DISCUSSION Our results indicate that narcolepsy is far from a rare disorder in the United States. The prevalence of narcolepsy with or without cataplexy is one in 1, 786 people and the prevalence of narcolepsy with cataplexy is one in 2,793 people. Thus, one would expect there to be 560 narcoleptics in an area of 1 million population. The incidence rate of 1.37 per 100,000 population per year (0.74 for narcolepsy with cataplexy) implies that one would expect 14 new narcoleptics to be diagnosed every year in an area with 1 million population. Our prevalence of 36 narcoleptics with cataplexy per 100,000 population is very close to the per 100,000 found in previous U.S. and European studies using different methodologies. 2-4,7,9 This prevalence is similar to the prevalence of multiple sclerosis in southern Europe. 17 Primary care physicians need to have a high index of suspicion for the disorder in sleepy young patients. The diagnostic criteria we developed have been validated and tested for specificity and interrater reliability. 14 Incidence rates and prevalences using the nonvalidated ICSD criteria fell inbetween those for narcolepsy as a whole and those for narcolepsy with cataplexy using the Mayo criteria. Apart from the Finnish twin cohort study, 7 no previous research used validated criteria. It should be noted that all our patients were Caucasian, and thus we cannot reach conclusions regarding the epidemiology of narcolepsy in the African American, U.S. Hispanic, or Asian populations. We showed a male predominance in the incidence rates (1.6:1) and prevalence (1.8:1) of narcolepsy as a whole. In the subgroup of narcolepsy with cataplexy, the gender effect was still present but less marked with incidence rate ratios SLEEP, Vol. 25, No. 2,

5 Table 3 Prevalence of Narcolepsy (Categories A, B, and C) January 1, 1985 Prevalence Prevalence Population Incidence Rates (X 100,000) ,859 8,215 16, ,483 7,629 15, ,215 7,999 17, ,509 8,154 16, ,828 5,670 11, ,185 4,124 8, ,265 2,780 6, ,583 1,592 4, , , Total ,799 46,868 97, Prevalence of Narcolepsy with Cataplexy (Categories A, B1, and C) January 1, 1985 Prevalence Prevalence Population Incidence Rates (X 100,000) ,859 8,215 16, ,483 7,629 15, ,215 7,999 17, ,509 8,154 16, ,828 5,670 11, ,185 4,124 8, ,265 2,780 6, ,583 1,592 4, , , Total ,799 46,868 97, of 1.2:1 and prevalence ratios of 1.4:1. Previous epidemiological studies have not examined gender, but case-based series have at times suggested a male predominance. 2,18,19 The explanation for this gender effect is unknown. The suggestion that it is an artifact related to men seeking more medical attention due to their alleged predominance in the work force is unlikely to be correct, as the gender effect was present in all four decades of our study. We have shown that narcolepsy commences most frequently in the second decade with incidence rates more than twice that seen in the third decade and more than three times that seen in the first and fourth decades. Although this incidence pattern has previously been noted in case series, 2,18 this is the first confirmation in a population-based study. A clinic based study of narcoleptics in Quebec and France suggested a bimodal pattern of age of onset, with a larger peak in the second decade and a smaller peak in the fourth decade. 19 Possible explanations for this discrepancy are the different ethnic populations studied, the slightly different diagnostic criteria used, and the fact that the French and Canadian study was not population based. In large clinical series, patients with narcolepsy without cataplexy comprise approximately 25% of the total group. 20,21 We found the prevalence of narcolepsy without cataplexy to be 36% of the prevalence of narcolepsy as a whole, suggesting that this subgroup may be somewhat more common than generally perceived. A number of issues need to be considered in assessing the reliability of our results. First, were all patients with narcolepsy seen at Mayo Clinic and other Olmsted County health providers identified? The records-linkage methodology of the Rochester Epidemiology Project has been used in many previous studies and accurately identifies patients coded with a specific diagnosis. In a Rochester study of the epidemiology of Parkinson s disease, 321 charts of patients with H-ICDA codes unrelated to Parkinsonism were reviewed; no patients with Parkinsonism were identified. 22 We reviewed 400 charts of patients diagnosed with disorders of sleepiness other than narcolepsy, and found that no cases of narcolepsy had been missed. We reviewed the records of the only private practice neurologist in the county not linked to the system and identified one additional patient. Second, did patients with narcolepsy consult other health providers outside of Olmsted County? Rochester is in the center of the county and the Mayo Clinic serves as the tertiary care center for southeastern Minnesota. None of the surrounding counties have tertiary care medical facilities, and specifically none of them have sleep disorders centers. Even if some patients were receiving care for narcolepsy elsewhere, it is highly likely that they would have received some of their general health care within the county, resulting in the diagnosis of narcolepsy being captured. Thus, while we cannot completely rule out the possibility that some patients sought care elsewhere, this is unlikely to have had a major impact on our findings. Third, many patients with nar- SLEEP, Vol. 25, No. 2,

6 colepsy delay seeking medical attention for their complaint. Could some patients have been missed as a result? We reviewed all medical records for a period of six years following termination of the incidence study ( ). If we had missed large numbers of patients with more recent onset of narcolepsy, we would have expected a lower incidence rate in the decade , compared to previous decades. However, the incidence per 100,000 persons per year was very similar between 1960 and 1989 (1.10 between 1960 and 1969, 1.62 between 1970 and 1979, and 1.34 between 1980 and 1989), suggesting that we had included most recent patients. We also selected a prevalence date that allowed us 11 years of subsequent data review ( ) in order to minimize the possibility of missed patients. Fourth, could some patients with narcolepsy have never consulted a physician for their symptoms? The population of southeastern Minnesota is medically sophisticated and public information regarding sleep disorders and their consequences has been widely disseminated in recent years. Nevertheless, we acknowledge that this possibility cannot be excluded by our methodology. Fifth, could group B2 (narcolepsy without cataplexy) include patients whose cataplexy had not yet manifested? The mean difference between the onset of sleepiness and that of cataplexy in Groups A, B1, and C (narcolepsy with cataplexy) was 5.3 years. Thus the additional six years that records were reviewed beyond termination of the incidence study and the 11 years beyond termination of the prevalence study would have likely resulted in the ascertainment of most, but not necessarily all, patients whose cataplexy developed after sleepiness. However, it is well recognized that some narcoleptic patients only develop cataplexy decades later, and no intermediate term epidemiological study can fully account for this. Sixth, can we be certain that patients with narcolepsy did not move to Olmsted County to obtain treatment for their disorder at Mayo Clinic? This should not have affected our incidence study, as we required that the onset of disease occurred while patients were already resident in the county. While theoretically, this might have inflated the prevalence figures, review of the patients charts identified no one documented to have moved for health reasons. We conclude that narcolepsy is not a rare disorder. It appears to be more common in men, a phenomenon needing further study. It originates most frequently in the second decade of life and commences only rarely after the fourth decade. Narcolepsy without cataplexy is a significant subgroup. Primary care physicians should increase their awareness of this disease. ACKNOWLEDGMENT We would like to thank Dr. Walter Rocca of the Departments of Health Sciences Research and Neurology of Mayo Clinic for his advice and assistance with this project and Ms. Nancy Slocumb for her hard work in abstracting the data. Supported by Mayo Foundation (Piscopo Funds). REFERENCES 1. Krahn LE, Black JL, Silber MH. Narcolepsy: new understanding of irresistible sleep. Mayo Clin Proc 2001;76: Roth B. Narcolepsy and Hypersomnia, Chapter 10: S. Karger, Dement WC, Zarcone V, Varner V, et al. The prevalence of narcolepsy (abstract). Sleep Res 1972;1: Dement WC, Carskadon MA, Ley R. The prevalence of narcolepsy SLEEP, Vol. 25, No. 2, II (abstract). Sleep Res 1973;2: Honda Y. Census of narcolepsy, cataplexy and sleep life among teen-agers in Fujisawa City (abstract). Sleep Res 1979;8: Peled R, Lavie P. Narcolepsy-cataplexy: an extremely rare disorder in Israel (abstract). Sleep Res 1987;16: Hublin C, Kaprio J, Partinen M, et al. The prevalence of narcolepsy: an epidemiologic study of the Finnish twin cohort. Ann Neurol 1994;35: Wing YK, Chiu CKW, Chiu HFK, Chen CN. Narcolepsy in Hong Kong Chinese a preliminary experience. Aust NZ J Med 1994;24: Ohayon MM, Priest RG, Caulet M, Guilleminault C. Hypnagogic and hypnopompic hallucinations: pathological phenomena? Br J Psychiatr 1996;169: Melton LJI. History of the Rochester epidemiology project. Mayo Clin Proc 1996;71: Commission on Professional and Hospital Activities HCFA. Hospital adaptation of the international classification of diseases, adapted (H-ICDA), 2nd ed. Ann Arbor, MI: Commission on Professional and Hospital Activities, American Sleep Disorders Association. The International classification of sleep disorders, revised. diagnostic and coding manual. Rochester, MN: ASDA, American Psychiatric Association. Diagnostic and Statistical manual of mental disorders (DSM-IV). Washington, DC: American Psychiatric Association, Silber MH, Krahn LE, Olson EJ. Diagnosing narcolepsy: validity and reliability of new diagnostic criteria. Sleep Medicine 2002;In Press. 15. Bergstralh EJ, Offord KP, Chu CP, Beard CM, O Fallon WM, Melton LJI. Calculating incidence, prevalence and mortality rates in Olmsted County, Minnesota: an update. Technical Report Series No ; Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota. 16. Anderson DW, Mantel N. On epidemiologic surveys. Am J Epidemiol 1983;118: Weinshenker BW. Epidemiology of multiple sclerosis. Neurol Clin 1996;14: Daniels LE. Narcolepsy. Medicine 1934;13: Dauvilliers Y, Montplaisir J, Molinari N et al. Age at onset of narcolepsy in two large populations of patients in France and Quebec. Neurology 2001;57: Mignot E, Hayduk R, Black J, Grumet FC, Guilleminault C. HLA DQB1*0602 is associated with cataplexy in 509 narcoleptic patients. Sleep 1997;20: Guilleminault C, Mignot E, Partinen M. Controversies in the diagnosis of narcolepsy. Sleep 1994;17:S1-S Bower JH, Maraganore DM, McDonnell SK, Rocca WA. Incidence and distribution of parkinsonism in Olmsted County, Minnesota, Neurology 1990;49: ABBREVIATIONS USED H-ICDA Hospital version of the International Classification of Diseases, Adapted; ICSD International Classification of Sleep Disorders; SOREM Sleep onset REM periods; MSLT Multiple sleep latency test.