EMMA DANIELS, MARTIN A. KING, IAN E. SMITH and JOHN M. SHNEERSON

 Onset is usually in the second or third decades of life, although around 6% of narcoleptics have symptoms before they are lO-years-old (Broughton and Broughton 1994). Narcolepsy can affect performance at school, obtaining and maintaining employment, formation of relationships, family life and participation in social activities.

There is no cure for narcolepsy, but the severity of daytime sleepiness can be reduced by stimulant medication, and antidepressants are used to suppress cataplexy. It can be difficult to find the optimum dose of stimulant medication with which sleepiness is improved without the induction of side effects.

There have been previous reports of the psychosocial effects of narcolepsy from a small number of authors (Broughton and Ghanem 1976; Broughton et al. 1981; Kales et al. 1982; Broughton et al. 1983; Broughton et al. 1984; Rogers 1984; Alaia 1992; Broughton 1992a, b), but few concerning subjective reports of how narcoleptics perceive their own health status in terms of functional and emotional status and general wellbeing (Ferrans et al. 1992; Beusterien et al. 1999). Broughton et al. (1981) compared the life effects of narcolepsy in three different countries, but there has been no large United Kingdom (UK) study and many published reports are more than 15-years-old.

The aim of this study is to describe the health status and psychosocial aspects of those with narcolepsy in the UK, compared to normative data, and to assess the relationship of drug treatment for the symptoms of narcolepsy.

METHODS
Subjects

A questionnaire was sent to 500 members of the United Kingdom Association of Narcolepsy (UKAN). These were randomly selected by the UKAN membership secretary from the UKAN confidential database. The total membership of UKAN was unknown at the time of the study.

Instruments
The questionnaire was divided into three sections. The first section comprised: questions regarding subjects' age and sex; their diagnosis and treatment of narcolepsy; other concurrent conditions; the Ullanlinna narcolepsy scale (UNS) (Hublin et al. 1994a); and psychosocial aspects of narcolepsy. The second section comprised the UK Short Form 36 health survey questionnaire (SF-36) (Brazier et al. 1992; Garratt et al. 1993) and the third section comprised the Beck depression inventory (BDI) (Beck et al. 1961).

Narcolepsy scale

The UNS is an epidemiological tool designed to diagnose narcolepsy (Hublin et al. 1994a). It comprises 11 questions about sleepiness and cataplexy, each of which has five possible answers. Each answer is allocated a score between 0 and 4, giving a total possible score of 0-44. A higher score indicates more severe and frequent symptoms. The authors suggest that a score greater than or equal to 14 indicates a diagnosis of narcolepsy.

Psychosocial aspects

The questions regarding the psychosocial aspects of narcolepsy were designed specifically for this study. They enquired whether having narcolepsy had caused avoidance of, or problems with, different aspects of life such as school, work, relationships, home and recreation. In addition to answering specific questions, subjects were invited to add further comments about particular problems they had experienced due to narcolepsy.

The health status of the subjects was assessed using the SF-36 (Garratt et al. 1993), a widely used and validated questionnaire (Brazier et al. 1992; Jenkinson et al. 1993), which quantifies subjective reports of health in terms of functional status, emotional status and general well being. The SF-36 comprises 36 questions divided into the following eight domains representing different aspects of health status:

1 Physical functioning, which is the subject's ability to walk up stairs, carry groceries and so forth.
2 Role limitations due to physical problems, i.e. whether physical health has caused problems with everyday life.
3 Role limitations due to emotional problems, i.e. whether emotional health has caused problems with everyday life.
4 Social functioning, which is how emotional or physical
health has affected normal social activities.
5 Mental health, which is a reflection of mood. 6 Energy Ivitality.
7 Bodily pain.
8 General health perceptions.
Thus the dimensions effectively cover functional status, wellbeing and overall evaluation of health (Brazier et al. 1992). A lower score indicates poorer health status.

Beck depression inventory

The prevalence and severity of depression was measured by the BDI (Beck et al. 1961). This comprises 21 questions each with four possible answers which are allocated a higher score the more severe the symptom. The total possible score range is 0-63. The higher the total score, the more severely depressed the subject is.

Data handling and statistical analysis

All questionnaires were checked for completeness of the age and sex categories, and for the medication that was being taken for narcolepsy. Subjects who had not specified their age or gender, or who were under 18 years of age, were removed from the dataset. Subjects taking modafinil were also excluded from the dataset as they formed too small a group for separate analysis.

The SF-36 was scored according to the instructions outlined in the UK SF-36 analysis and interpretation manual (Jenkinson et al. 1996). The UNS was scored according to the method outlined in the paper by Hublin et al. (1 994a). The BDI was scored in the conventional way and the scores divided into the following severity categories as suggested by normative data (Bowling 1991): normal (no depression) 0-9; mild depression 10-15; mild-moderate depression 16-19; moderate-severe depression 20-29; severe depression> 29.

The SPSS computer package was used for statistical analysis. The relationships between the UNS score and each of the eight SF-36 domain scores, and the UNS score with the BDI were tested using the Kendall Correlation Coefficients. Each subject was allocated the mean normative score for each SF-36 dimension according to their age and sex. These values were taken from Jenkinson et al. (1996) for those under the age of 65 years and from data used in the paper by Sharples et al. (2000) for those aged 65 years and over. There is no UK published data that covers the whole age range required for this study. The two studies cited above are similar in that they both used general practitioner lists to randomly select people from the general population. They differ in terms of the area of the UK that the people were selected from, and how the survey was administered (postal vs. face-to-face interview). There has been no analysis to show if and how these differences may affect the results. A Wilcoxon Matched-Pairs Signed

Ranks

Test was used to compare the UKAN SF-36 data with the normative data, as none of the UKAN SF-36 scale scores followed a normal distribution.
The subjects were divided into four groups according to the combination of medication they were taking for narcolepsy (Table 1). Cross-tabulation using a chi-square test was used to ensure there were no differences in sex distribution between the four groups. Kruskal-Wallis one-way analysis of variance was used to test for differences in age distribution, SF-36 domain scores and BDI scores. The Mann-Whitney U-test was used where appropriate to assess whether anyone particular medication group had significantly different SF-36 or BDI scores to the other groups.
 

RESULTS

A total of 313 questionnaires were returned anonymously to the Sleep Centre, Papworth Hospital (return rate = 62.6%). One subject was excluded for not specifying their gender, three were excluded for being under the age of 18 years, and four were excluded for taking modafinil. Thus the total number of subjects included in the analysis was 305.

Of the 305 subjects, 60.7% were female. The age range was 18 years to 89 years, with a median of 56 years. Almost half (48.9%) were diagnosed with narcolepsy by a neurologist, 28.5% were diagnosed by hospital consultants with different specialities and 18.4% by their general practitioner. The earliest diagnosis of narcolepsy recorded in the UKAN population was in 1927. An increasing number of people received a diagnosis over each decade and 32.1 % of the population were diagnosed from 1990 to the time of the survey in 1997. Only 26.6% reported that they had regular appointments with a consultant or general practitioner to review their narcolepsy. The UNS scores of the UKAN population ranged from 4 to 43 with a median of 22. Two hundred and seventy people had a UNS score greater or equal to 14, which has been suggested to indicate a diagnosis of narcolepsy (Hublin et al. 1994a).
 

Psychosocial aspects

The percentages of the UKAN population who answered positively to each of the psychosocial questions are listed in Table 2. Additional comments included being unable to bath when alone in the house because of the risk of falling asleep and problems using electrical and gas appliances. Subjects also reported being punished at school by teachers who thought they were lazy.

Beck depression inventory

The BDI scores of the UKAN population ranged from 0 (no depression) to 48 (severe depression) with a median score of 11, which falls into the mild depression category. The BDI indicated that 56.9% of the UKAN population were depressed (BDI<10) with moderate or severe depression (BDI < 20) in 15.1 % of responders. There was no correlation between the UNS score and the BDI.

SF-36
The SF-36 scores showed great variability in all dimensions. There was no correlation between the UNS score and any of the SF-36 scale scores. The UKAN group had significantly lower SF-36 scores on all scales than age and sex matched normative data. The scale that showed the greatest difference between medians was role limitation due to physical problems, followed by (in descending order): energy/vitality, social functioning, role limitation due to emotional problems, general health perceptions, mental health, physical functioning and bodily pain.

Medication subgroups

There was no significant difference in the sex distribution or age distribution between the four medication groups. The only SF-36 domains in which a significant difference was seen between medication groups 1-4 (Table 1), were physical function (P = 0.003) and social function (P = 0.048). Those who took both stimulant and anticataplexy medication (group 1) had significantly lower scores, indicating poorer health status than the rest of the sample population for both of these scales. Group 1 also had the lowest median score for the other six scales but the differences were not significant.

There was a significant difference in the BDI scores between the four medication groups (P = 0.02) with the mean rank for group 1 being higher than the mean ranks for the other three groups. A comparison of the BDI scores for group 1 with the other three groups combined together showed that the scores of group 1 were significantly higher, indicating more severe depression than the rest of the UKAN population (P = 0.0002).

DISCUSSION

The results of this study show that members of UKAN have a poorer health status, as measured by all eight scales of the SF36, than the general UK population. It has also shown that narcolepsy has a detrimental effect on the psychosocial aspects of life and that over half of the sample are depressed to some extent. These results are consistent with studies regarding people with narcolepsy living in other countries (Roth and Nevsimalova 1975; Broughton and Ghanem 1976; Broughton et al. 1981; Broughton et al. 1983; Broughton et al. 1984; Ferrans et al. 1992; Broughton 1992b; Merritt et al. 1992; Yang and Clerk 1998), except for one study which showed even greater effects (Alaia 1992).
The greatest effect of narcolepsy in this study, as reflected by the SF-36, is on role limitation due to physical problems (role physical). Physical health limited the subjects by restricting the type of, and the amount of time spent on, work or other activities. They had not accomplished as much as they had hoped and/or had difficulties performing work or other activities. Narcoleptics commonly avoid situations which could be embarrassing or harmful if they were to fall asleep inappropriately or have cataplexy.

The biggest reported impact of narcolepsy was on leisure activities such as visiting the cinema or theatre (76.7%), playing sports (39.7%) and taking holidays (37.4%).

Narcolepsy also affected schooling, with over 50% having difficulty concentrating in class and feeling that they had achieved less than they were capable. Presumably in the majority of these cases, narcolepsy had remained untreated at this important time as only 8% of the sample had received a diagnosis by 16 years of age with a further 14.5% being diagnosed by the age of 21 years. The earliest age of diagnosis was 10 years (n = 3). Almost 20% of the sample were bullied, teased or had difficulty making friends. After leaving school, 27% have been unable to use the qualifications they did gain and 36.7% have lost or left a job because of narcolepsy. It has been reported that many narcoleptics perceive that their job performance has deteriorated due to memory and concentration problems, which they attribute to narcolepsy (Broughton and Ghanem 1976,; Broughton et al. 1981; Alaia 1992).

The subjects reported difficulty in performing everyday tasks such as cooking (38.7%), ironing (27.9%), or supervising children (31.5%). Other studies have shown that narcoleptics have more accidents when driving (Alaia 1992), at home, at work, and attributable to smoking than the general population (Broughton et al. 1981; Broughton et al. 1984; Broughton 1992a, b; Cohen et al. 1992).

It has been suggested that the accumulation of all these effects of narcolepsy can cause deterioration in emotional health (Broughton and Ghanem 1976; Kales et al. 1982; Alaia 1992). It has also been proposed that depression may be endogenous to narcolepsy (Roth and Nevsimalova 1975; Broughton et al. 1981), as abnormalities in rapid eye movement (REM) sleep, such as reduced REM sleep latency, are common to depression and narcolepsy (Broughton et al. 1981). In the current study, 56.9% of participants were depressed to some extent, as reflected by the BDI scores, and as a group the narcolepsy sample had significantly poorer scores on the SF-36 domain role limitation due to emotional health, than a sample from the UK general population. The prevalence of depression in this sample is similar to that found in other studies (Broughton and Ghanem 1976; Broughton et al. 1981; Broughton et al. 1983; Broughton et al. 1984; Mosko et al. 1989; Merritt et al. 1992).

The energy/vitality scale was the second most affected aspect of health status in the UKAN sample compared to normative data. People with narcolepsy have excessive daytime sleepiness, with unwelcome sleep attacks rather than feelings of physical tiredness, although they may have difficulty distinguishing between the meanings of the two when faced with a questionnaire with no accompanying explanation. Sleepiness may have been better measured in our group by the Epworth Sleepiness Score (ESS) (Johns 1991). There is only a moder-

ately close correlation between the ESS and the SF-36 energy/ vitality scale (Briones et al. 1996).
A similar pattern of impairment of health status has been shown using the SF-36 in a North American population of narcoleptics who were not taking any stimulant medication (Beusterien et al. 1999). The majority of our subjects were taking stimulant and/or anticataplexy medication. However, whatever the medication combination being taken, health status was not returned to normal. In our study, those who were taking both stimulant and anti-cataplexy treatment had significantly lower scores for two of the eight SF-36 domains, and significantly higher scores for the BDI than the other three medication groups.

The results show that the treatments being used are not sufficiently effective to restore health status to normal. Mosko et al. (1989) reported that depression associated with narcolepsy did not improve with stimulant medication and Merrit et al. (1992) reported a greater prevalence of depression in narcolepsy than in the general population, which was independent of pharmacological treatment. It has also been found that some narcoleptics choose to put up with some of their symptoms because the drug-induced side-effects are worse (Alaia 1992).
Only a quarter of the UKAN population reported having regular follow-up appointments with their general practitioner or consultant to review their narcolepsy symptoms, and presumably the medication of the others is not being checked regularly to see if it is optimal. Almost a quarter of respondents were taking no medication at all, and nearly three-quarters were not taking specific anti-cataplexy medication, although two-thirds of these may be obtaining some relief from cataplexy from their stimulant medication. These results suggest that the drug management of most people with narcolepsy is inadequate.

The new non-amphetamine wake-promoting drug modafinil has been shown to produce higher scores than a placebo for the role physical, energy/vitality, social function and role emotional on the SF-36 in narcoleptics. But even with this and careful patient management, the majority of scores did not return to normal (Beusterien et al. 1999). The number of subjects in our survey taking modafinil was too small to enable any statistical comparisons to be made with this study.

Even though our study shows that narcolepsy has a detrimental effect on health status, it did not find any correlation between the severity of the two main symptoms, as reflected by the UNS, and the BDT or SF-36 scale scores. One possible reason is that worsening narcolepsy symptoms do not progressively impair health status. Some other authors have shown that depression is not related to the severity of symptoms in narcolepsy (Alaia 1992), although other research has shown the opposite.

It is possible that the UNS does not truly reflect the severity of narcolepsy. Using specific scoring criteria duplicated in the current study, the UNS was reported to be 98.8% specific and 100% sensitive in determining the narcoleptic syndrome (Hublin et al. 1994a). However, in a subsequent survey of the Finnish Twin Cohort, 75 people met the UNS scoring criteria for narcolepsy; but after further examination, a diagnosis of narcolepsy was only verified in three people, i.e. 4% of the original 75 (Hublin et al. 1994b). This suggests that the UNS is not as reliable for diagnosing narcolepsy or determining its severity as was first claimed.

Since the survey used in the current study was designed and completed, two more questionnaires have been developed and validated in a bid to diagnose narcolepsy without laboratory investigations or clinical interview by a sleep disorder specialist (Parkes et al. 1998; Anic- Labat et al. 1999). These could be considered for use in addition to, or as an alternative to the UNS.

The sample used for the current study was taken from members of UKAN, an association for people with narcolepsy. It may not be representative of the total UK narcolepsy population, and some members may not have a definite diagnosis of narcolepsy. It could be that only those with particularly severe symptoms of narcolepsy join a support group, but equally the support of the association may improve their health status. The questionnaire was returned by 62.6% of those to whom it was sent, who may not be representative of the original 500. The majority of people who took part in the survey were female, when no difference in sex ratio was expected. These factors should be taken into account in the interpretation of the results, but it is unlikely that any other methodology could provide information about such a large sample of those with narcolepsy living in the UK where there are few sleep centres carrying out full diagnostic procedures.

Despite its limitations, the current study has the advantage of being based on a larger UK narcolepsy population than any other health-related quality of life study. It is apparent that the symptoms of narcolepsy affect many aspects of working, home and social life. Although patients receive treatment for their symptoms, mainly with conventional amphetamine and/or anticataplexy medication, their health-related quality of life remains much poorer than that of the general UK population.