Social Functioning, Psychological Functioning, and Quality of Life in Epilepsy

Authors


Address correspondence and reprint requests to Dr. Th.P.B.M. Suurmeijer at Department of Sociology/ICS, University of Groningen, Grote Rozenstraat 31, 9712 TG Groningen, The Netherlands. E-mail: thsuurmeijer@med.rug.nl or th.p.b.m.suurmeijer@ppsw.rug.nl

Abstract

Summary:  Purpose: Part of our research intended to explain “Quality of Life” (QoL) differences between people with epilepsy. To this end, a series of already existing generic and disease-specific health status measures were used. In this study, they were considered as determinants of people's QoL, whereas QoL itself was conceived as a general “value judgment” about one's life.

Methods: From the records of four outpatient clinics, 210 persons with epilepsy were randomly selected. During their visit to the outpatient clinic, they completed a questionnaire assessing, among other things, health perceptions and social and psychological functioning. Additional information about their medical and psychosocial status was gathered from the patient files. Data were analysed by using a hierarchical regression analysis.

Results: In decreasing order of importance, “psychological distress,”“loneliness,”“adjustment and coping,” and “stigma perception” appeared to contribute most significantly to the outcome QoL as judged by the patients themselves, regardless of their physical status. In the final model, none of the clinical variables (onset, seizure frequency, side effects of antiepileptic drugs) contributed significantly anymore to the patients' “quality-of-life judgement.” Apparently the effect of other variables such as seizure frequency and health perceptions, medication and side effects, life fulfilment, self-esteem, and mastery is mediated by these variables.

Conclusions: Because all of the variance in QoL of the patients was explained by the psychosocial variables included in this study, health professionals should be aware of the significance of the psychosocial functioning of the patients and the role it plays in the achievement of a good QoL. Both informal and professional support may be an adjunct to conventional treatment. In future research, this issue should be given high priority.

Epilepsy and psychosocial status

Once people have been diagnosed as “epilepsy patients,” that is, persons who have had at least two epileptic seizures, they are most probably being treated by a medical specialist, a pediatrician or a neurologist, and antiepileptic drugs (AEDs) will be prescribed and expected to be properly used.

The medical specialist's efforts are directed primarily at the achievement of seizure control. It is often believed that doctors in charge also are providing care for psychosocial problems. One may wonder whether that is done or is done adequately. From a pilot study (1), it appeared that medical files contained little information regarding this domain. Knowledge and adequate support in this area are important because these psychosocial sequelae largely determine what it means for the person concerned “to have epilepsy,” even after complete seizure control has been achieved. For the sequelae of a condition such as epilepsy are determined not only by the epilepsy itself, but also, and perhaps even more, by the meaning ascribed to the condition by the patient's social environment (2–6). In addition to the direct effects epilepsy can have on the daily functioning of people with epilepsy, (prejudiced) attitudes, stereotypes, and behaviors encountered in society both from lay people and professionals (for example from close relatives, friends, neighbors, but also from colleagues, teachers, employers, or physicians) can be more disabling than the seizures themselves and will strongly burden their lives (3–12). To improve their opportunities to participate fully in society, and consequently, to maintain or improve their daily functioning and position in society, it is necessary to change the way society views epilepsy and treats the people who have it.

Differently put, the general assumption underlying the medical model is that the disease process itself primarily determines psychosocial distress. A worse clinical course is assumed to result in greater psychosocial upset and stress over time. Conversely, a benign course of epilepsy should be associated with levels of psychosocial functioning that do not differ significantly from those of the period before illness. The social science model posits that a set of causal influence factors on adjustment processes in chronic disease (i.e., epilepsy) is located in the social environment of the patient and determines the outcome of short- and long-term adjustment processes. Therefore, in addition to the disease process itself and in accordance with the so-called biopsychosocial model, a supportive environment is considered a second major determinant of the psychosocial status of the patient.

Because the consequences are more far-reaching than only the physical aspect, it is obvious that not only aspects in the “physical domain” should be assessed and evaluated but also those of other life domains such as emotional functioning, role activities and social functioning, health perceptions, and general satisfaction with life (3,13–17). Basically, this argues for a multidimensional approach to the consequences of epilepsy. The “quality of life” (QoL) construct was introduced to assess and evaluate more comprehensively the outcomes or effects of (chronic) illness and medical interventions.

Quality of life

A broad approach to the conceptualization and assessment of QoL in epilepsy (as well as in other chronic disorders) is often strongly advocated (15,18–20). QoL is now often defined as the “individual's overall satisfaction with life and their general sense of well-being”(21) and comprises mostly three to six dimensions or life domains (15,18,21,22). For example, both Spilker's (15) and Baker's (18) comprehensive models specify three broad QoL domains (physical, social, and psychological), and within each one, various specific, interrelated aspects and dimensions are identified. According to Spilker, QoL can be seen on three levels (“pyramid model”) with the overall assessment of well-being at the top, the broad QoL domains in the middle, and the lower level includes aspects of these broad domains.

A more elaborate definition of QoL was recently presented by the so-called World Health Organization Quality of Life group (WHOQOL-group). According to this group, QoL can be defined as the: “individuals' perception of their position in life in the context of the culture and the value system in which they live and in relation to their goals, expectations, standards, and concerns”(23). The patient's perspective on QoL is at the core of this definition and encompasses the personal assessment of the patient of his physical health status (e.g., daily physical functioning) and his psychological and social status (e.g., mood, companionship, and recreational activities) (24).

In our opinion, this definition or conceptualization rightly excludes the use of so-called “objective social indicators” (e.g., employment, socioeconomic status, recreational facilities) or “objective” health-related indicators of QoL (e.g., comorbidity, health care use), which can be considered as “necessary conditions for satisfaction and happiness”(25). They are assumed to be conducive to QoL on the individual level. As such, they are very important resources and should of course be included in studies of QoL. Consequently, QoL is “subjective” or, more adequately phrased, “subject-bound.” Or as Devinsky and Cramer (24) put it:

“The patient is the only one who knows how he or she feels, how the disorder affects that person's vigor, self-confidence, ability to socialize, obtain work, and function at home and on the job. Surely the patient's reports may be biased, as may be the doctor's. However, the patient is the person who must define his own quality of life (…). Only the patient knows if an imbalance exists between expectations and reality (…).”

Nevertheless, despite these widely accepted definitions, QoL is still conceptualized and measured in a variety of ways. There is disagreement regarding what constitutes a good life versus what influences it (21,25–27). Some experts in the field such as Leventhal and Colman (27) and Hunt (28) have recently raised fundamental criticisms against the usual way of conceptualizing and measuring of QoL. They emphasize that a generic value judgement about life in general or life domains is at the core of the QoL concept and not the simple assessment of physical, social, and/or psychological functioning or status of the patient. Measures assessing physical, social, and psychological functioning or status are considered as determinants of QoL, not as QoL itself. These measures are often denoted “health status measures”(14,29–32), and there is no need to reconceptualize them as QoL measures instead of determinants of QoL. This latter approach of QoL is central in the present study on QoL in Dutch patients with epilepsy.

The purpose of the present study was to explore the relative contribution of the physical, social, and psychological status or functioning of the patient to the patient's value judgment about his or her QoL.

Research model

The assumption is made that someone's overall QoL judgement is not only directly affected by a chronic physical disorder but also (perhaps even mainly) indirectly affected and reflects the stress and strains induced by a decrease in or loss of personal, social, and economic resources (for example, by the aggravation of domestic tasks, occupational changes, loss of income, lowered feelings of self-esteem, mastery, or control) (33).

This is in line with the [Devins et al. (34)] concept of “illness intrusiveness,” which refers to the interference of the illness with continued involvement (5) in valued activities and interests (e.g., sociable activities, leisure activities, or fulfilling occupational roles). It parallels the Pearlin et al. (35,36) process of “stress proliferation.” Stress proliferation refers to “the expansion or emergence of stressors within and beyond a situation whose stressfulness was initially more circumscribed. As the initial stressor or set of stressors gives rise to an accumulation of stressors, the lives of individuals may become enmeshed in a configuration of stressors quite different from those they faced earlier”(36).

The first group of stressors is called “primary stressors,” and the second one, “secondary stressors.” Applied to people with a chronic condition such as epilepsy, hardships and problems anchored in the illness and the course itself (block 2 in Table 1) can be considered primary stressors, while secondary stressors are surfacing in the social, leisure, and occupational realm (block 3 in Table 1). These two groups of stressors will affect the psychological status of the patient (block 4 in Table 1) and ultimately his or her value judgment about the “overall quality of life.”

Table 1. Research model
Main concepts and components
Block 1:Sociodemographic status
Block 2:Physical status
 Subblock a.Clinical(ly related) aspects
 Subblock b.Health status perceptions
Block 3.Social status
 Subblock a.Social functioning
 Subblock b.Self-efficacy
Block 4.Psychological status
General mental health evaluation
Psychological distress
OutcomeOverall quality of life

These assumptions are reflected in the hierarchical biopsychosocial research model, presented in Table 1, and can be seen as an extension and specification of Spilker's “pyramid model of QoL” (described before) and supplemented by the Pearlin et al. “stress proliferation model.” So,

  • 1after adjusting statistically for the influence of sociodemographic variables age and sex (block 1), the other variables will be brought into the analysis. As specified by the hierarchical model, presented in the Table 1,
  • 2first, variables representing the medical model were entered into the analysis. They refer to the physical status of the patient and comprise both clinical/clinically related aspects and health perceptions (block 2). First, the clinical(ly related) aspects of subblock 2a were entered: onset, seizure frequency, side effects of AEDs, followed by the health-status perceptions of subblock 2b: perception of epilepsy seizures and general health perceptions.
  • 3Next, variables representing the social science model (comprising both blocks 3 and 4), were entered. The variables of block 3 represent the social status of the patient. First, variables referring to the social functioning of the patient in subblock 3a were entered into the analysis: stigma, life fulfilment, loneliness, followed by self-efficacy variables of subblock 3b: self-esteem, mastery or control, and general adjustment.
  • 4This was followed by a set of variables considered to represent the psychological status or functioning (block 4) of the patient, comprising the patient's general mental health evaluation and psychological distress.
  • 5The outcome variable is QoL as assessed by a visual analogue scale with “delighted-terrible faces” (VAS-DT) and with scores running from 0 (“as bad as possible”) to 10 (“as good as possible“). This is in accordance with our theoretical position as just formulated.

MATERIALS AND METHODS

Sampling procedures and response

Patients were selected from outpatients' clinics of the Department of Neurology of the university hospital in Groningen and of a special epilepsy centre, in total, four outpatients' clinics, all in the northern region of The Netherlands. Patients who visited the outpatient clinic were randomly included. To be included/excluded from our study, the following criteria were used:

  • inclusion criteria: age between 18 and 65 years, and a definite diagnosis of epilepsy;

  • exclusion criteria: malign epilepsy syndromes accompanied by a mental handicap or the presence of severe psychiatric disturbances, as reflected in an earlier psychiatric admission.

Of the 275 patients with epilepsy who all fulfilled these inclusion/exclusion criteria for enrolment in the study, 225 (82%) participated. Because of missing data or unclear answers, 210 persons were used in the final analyses.

The patients were asked to fill out a questionnaire on the day they visited their neurologist at the outpatient clinic. In addition, demographic and clinical data were collected by means of both a short interview with the patients and information from their medical files. A letter of informed consent, signed by the patients, was obtained from all the subjects participating in the study.

Composition of the research group: sociodemographic and clinical characteristics of the respondents

An overview of some of the patients' sociodemographic, social, and clinical(ly related) data are presented in Table 2.

Table 2. Sample composition: sociodemographics, educational level, employment status, and clinical status
Parameter 
  1. AED, antiepileptic drug.

Sociodemographic status 
 Mean age (yr)38 (SD, 11.4)
 Sex (% males)51
 Married (% yes)66
Educational status (%) 
 Primary level3
 Junior vocational56
 Junior secondary22
 Senior secondary/vocational18
 Higher vocational/university1
Employment status (%) 
 Full-time job21
 Part-time job14
 Sheltered workshop12
 Disability pension24
 Unemployed4
 Retired2
 Housewife16
 Full-time education7
Clinical status (%) 
 Onset before 21 yr78
 Unknown etiology87
 Type of epilepsy 
  Generalized12
  Partial81
  Unclassifiable7
Seizure type 
 Tonic–clonic23
 Complex partial (CP)61
 CP + Secondarily generalized31
 Simple partial17
 Absence6
 Other11
Polypharmacy (two or more AEDs)71
Side effects of AEDs (no side effects)23

Of the 210 respondents, 51% were male and 49% female patients. The mean age was 38 years (SD, 11.4 years). Sixty-six percent were married, a high percentage as compared with the results of some recent other studies (37,38). In the general Dutch population older than 20, 77% are married (39), which is not that much higher than the figures found in this study, considering that younger people also were included in our sample.

As compared with the Dutch population (39), the educational level of our group was generally below average: 81% had only lower types of education as opposed to 19% with a higher level of education. Consequently, the employment status was rather low: only 35% had paid jobs. About 30% were members of the EVN (Epilepsie Vereniging Nederland), which is the Dutch patient organization for patients with epilepsy, their parents, or partners. This percentage is about 10 times as high as the corresponding percentage for the total group of people with epilepsy.

Seventy-eight percent of our respondents had an onset before age 21 years (mean age at onset, 15 years; SD, 11.0 years). Furthermore, 12% had a generalized type of epilepsy: 10% idiopathic and 2% symptomatic. As many as 81% of our respondents had partial epilepsy. Consequently, complex partial seizures were most frequently encountered.

The rather high percentage of partial epilepsy may be a consequence of the composition of the population from which we drew our sample (40) and the enrolment procedures used in our research. The possibility of a sample bias was discussed in a former publication based on these data (20). Twenty-four percent of the patients were seizure free for geqslant R: gt-or-equal, slanted1 year, and 6% for 6 months. The remainder of the patients (70%) had at least one seizure during the last 6 months. In 6%, a status epilepticus occurred. Seventy-one percent used two or more AEDs, with carbamazepine (CBZ) and sodium valproate (VPA) most frequently used.

Quality-of-life domains

The questionnaire contained a number of generic and disease-specific health status measures (see Table 3), covering several parts of the components of Spilker's pyramid QoL model, discussed before, and have been reported elsewhere (20). In our study, no health status measures were developed de novo. Instead, already existing, (mostly) generic health status measures were used. The main reasons to propose such a procedure were of a theoretic, comparative, and resource-optimizing nature (20).

Table 3. Measures used in the analyses
DomainMeasure (response categories, and
Cronbach's alpha between brackets)
Items/Range of scoresReference
  1. AED, antiepileptic drug.

OnsetAge at first seizure0–59 
Seizure frequencyNumber of seizures in the past year (per type of seizure; maximum, four types) (none, 1–6, 7–12, 12–24, >24 seizures/half year)0–24 
Side effects of AEDsIncludes headache, dizziness, sleepiness, flurry, slowness, memory problems, etc., no, yes140–14 
Perception of epilepsy seizures (PES)Includes worries about having a seizure, getting hurt, social problems, feeling bothersome (very worried, afraid, always, definite to not at all, never) (alpha, 0.71)55–2241
Health perceptions (HP)Includes feeling ill, excellent health feeling badly, etc. (absolutely true to absolutely not true) (alpha, 0.80)66–3142, 43
Perception of stigma of epilepsy (PSE)Includes having to prove yourself, treated fairly by employers, putting unreasonable limits on me, etc. (strongly agree to strongly disagree) (alpha, 0.62)66–249
Life-fulfillment questionnaire (LFQ)Fulfillment relevant aspects of social life including having a good family life, friends, getting help with problems, able to spend leisure time, save money, have a job, secure and stable future, etc. (extremely important to no importance and yes/no) (alpha, 0.68)40−60–+6044–46
Loneliness scale (LS)Includes feeling rejected, missing people around, pleasure of company, people to lean on, etc. (agree, disagree) (alpha, 0.89)110–1147–49
General adjustment to epilepsy (GATE)Feeling adjusted, able to cope (strongly agree to strongly disagree)11–520
Self-esteem (RSE)Rosenberg Self-Esteem (RSE) scale including being satisfied, not good at all, feeling useless, proud, etc. (strongly agree to strongly disagree) (alpha, 0.86)1010–4050
Mastery (MAS)Includes feeling being pushed around, having little control, self-reliant, feeling helpless, etc. (strongly agree to strongly disagree) (alpha, 0.75)77–2851
Mental health (MH)Includes being happy, nervous, downhearted, etc. (always to never) (alpha, 0.78)55–3042, 43
Psychological distress (GHQ)General Health Questionnaire (GHQ); includes items about depressive feelings, anxiety, somatic symptoms, and social dysfunctioning (not at all to much more than usual) (alpha, 0.93)280–8452, 53
Visual Analogue Scale (VAS-DT)Delighted–Terrible (DT) faces10–1029, 41

Statistical methods

Data were analyzed by using bivariate (Pearson correlation) and multivariate (multiple regression ) methods. As to the latter, in the present study, a hierarchical regression model was used. In most other studies, a stepwise procedure is used. It is controversial because it affords priority to statistical criteria of inclusion rather than to theoretical ones (36,54–56). Therefore we preferred a hierarchical regression model, as was discussed before at Table 1. The significance level was set on 0.05 (one-sided). Data were analyzed with SPSS (57).

RESULTS

To decide which variables to include in the regression analysis, all independent variables have been correlated with the dependent variable: overall judgment of QoL as assessed with the VAS-DT (Table 4).

Table 4. Correlations between independent and dependent variables
Independent variablesDependent variable
quality of life (VAS-DT)
r
  1. AED, antiepileptic drug.

  2. a  p = 0.02, b0.01 < p < 0.001, cp < 0.001 (all one-sided).

Sex0.01
Age0.07
Onset−0.06
Seizure frequency−0.14a
Side effects of AEDs−0.21b
Perception of epilepsy seizures0.36c
Health perceptions0.41c
Perception of stigma of epilepsy−0.17b
Life fulfillment0.36c
Loneliness0.43c
General adjustment to epilepsy0.46c
Self-esteem0.44c
Mastery0.46c
Mental health0.52c
Psychological distress−0.56c

All variables but sex, age, and onset correlated significantly with the dependent variable “overall judgement of QoL.” Therefore we decided to omit sex, age, and onset from the regression analysis. Most of the correlations between the independent variables (not presented) were statistically significant and varied from |0.12| to |0. 58|. Three correlation coefficients (all with seizure frequency) were not statistically significant.

Figure 1 (right) shows the increase of the explained variance in the dependent variable after each step or block is successively entered into the equation according to the model presented in Table 1. The variance explained by each block separately (i.e., irrespective of other variables) has been mentioned for each block.

Figure 1.

Research model: results of hierarchical regression analysis*

From this it appears that each separate block or group of variables has a substantive contribution to the QoL of our respondents. However, the independent variables within and between blocks are related to each other as well. Consequently, when entered hierarchically into the analysis according to our theoretical model, it appears that the contribution of each block or group of variables to the QoL of our respondents decreases, because part of the variance explained by a later block is already included in the variance explained by a former block or former blocks.

In the first column of Table 5, the incremental R2 (cf. Fig. 1, right) and corresponding p values are presented. This table also shows the variables (left in the final model after all variables have been entered into the analysis according to the specified order) that contributed most and significantly to the outcome “overall QoL judgment.” In decreasing order of importance, these variables are psychological distress, loneliness, adjustment and coping, and stigma perception. The β coefficients in the second column refer to the relative contribution of the variables in the final model. Only significant β coefficients are shown.

Table 5. Results of the hierarchical multiple regression analysis
Dependent variable:
QoL [VAS-DT]
Increment
R2 (%)
pchangeβp
  • a

     Measures in bolded type refer to significant β coefficients.

  • AED, antiepileptic drug.

Independent variables:    
Physical status    
1. Seizure frequency, side-effects AEDs5.01  
2. Health status perceptions [Perception epilepsy seizures and health perceptions]15<.0001  
Social status    
3. Social functioning [Perception stigma epilepsy,a9].0002.16/.17.03/.04
3. Life fulfillment and lonelinessa]15   
4. Self-efficacy [Self-esteem, Mastery and General adjustment toward epilepsya]6.003.19.01
Psychological status    
5. Mental health and Psychological distressa9<.0001−.33.0004
Total R.67   
Total R2 (%)44   

None of the clinical variables contributed significantly to the outcome “QoL judgment.” All of the variance was explained by the psychosocial variables in the model. Apparently the influence of the clinical (and other psychosocial) variables is mediated by the other variables in the model.

DISCUSSION

In the introduction, the WHOQOL-group definition of QoL was referred to as someone's perception of his or her position in life within a social and cultural context (23). More specifically, it was assumed that QoL is a generic value judgment reflecting “an individual's evaluation of the level of his or her functioning within a number of life domains, and the value or importance assigned to these domains”(27). Expectations, aspirations, behavioral repertoires, coping strategies, former and actual personal and symbolic experiences will result “in a unique combination of weightings (importance and value; TS) for a set of life domains”(26) and will produce a generic value judgment of life. From this conceptualization of QoL, we developed a research model on which our analyses were based.

The results can be summarized as follows. Most of the variables used in this study had a moderately strong bivariate relation with the QoL of our respondents. After all variables were entered into the equation according to the specified order in our model, only a small number of variables were left in the final model that maintained a direct relation with the overall QoL judgment. If people with epilepsy feel or actually are less stigmatized, are less socially isolated, are more able to cope with the disorder and its sequelae, and feel less distressed, their QoL appeared to be better, as judged by the patients themselves, regardless their physical status. This does not imply that these other variables are unimportant, certainly not. The variables left in the final model, however, (partly) mediate the effect of the other variables. Daily concerns such as, for example, employment, driving, medication and side effects, seizure unpredictability, and so on (10,14,19,58–61), may run through or manifest themselves in stigma perception, feeling lonely, self-efficacy, and mood.

Furthermore, we may conclude that the variables assigned to the “medical model” contributed only 5% to the patients' QoL (Fig. 1, block 1). Together with the clinically related variables (Fig. 1, block 2), the “physical status” of the patient explained 15% of the variance in the QoL of our respondents. The sociological and psychological variables belonging to the “social science model” (Fig. 1, blocks 3–5) contributed twice as much to the patients' QoL as the clinical(ly related) variables.

Health professionals should be aware of the significance of the psychosocial functioning of patients and the role it plays for their QoL. Both informal and professional support (the latter referring to doctor–patient relationships in the first place) may be an important adjunct to conventional treatment. As far as professional support is concerned, one may think, among other things, of creating an atmosphere of understanding, which seems to be very important, together with a keen interest in aspects of daily life such as relations in the nuclear family, child socialization, sexuality, pregnancy, social relations at work and school, educational and occupational achievements, and problems. Only in this way can one gain insight into the specific (psycho-)social situation of a specific patient, which is indispensable to give adequate guidance to the patient, including an eventual referral to another professional.

More generally, social support refers to exchanges of social, emotional, and instrumental resources between the individual and his or her social environment, intended to enhance the well-being of the individual concerned. In this sense, it also can be conceived of as a resource that contributes to the attainment of QoL outcomes. In nonepilepsy studies, social support is reported to have a beneficial effect on health and well-being, and recently it was conceptualized as “coping assistance”(64–66). Apart from the studies mentioned later, the role of this important social domain variable (67,68) in promoting well-being has not been studied systematically in any detail among people with epilepsy. Collings (45), in a recent study of life fulfilment and epilepsy, found that aspects of social support (such as family, friends, access to help) were the most highly rated sources of life fulfilment and, when present, provided immunity from some of the negative consequences of an epilepsy diagnosis. In two other studies on “epilepsy and social support”(69,70), it was found that perceived social support from both friends and family were factors in emotional adjustment, even when sociodemographic and seizure-related variables were considered. It was emphasized that social support should be considered when further investigating this area (70). Therefore, issues related to both informal and formal social support should be given high priority in future research. Finally, in the “chain of causality,” mood and distress lie closest to QoL and, consequently, explain most (but certainly not all) of the QoL variation among our patients with epilepsy. Therefore, decreasing QoL and increasing psychological distress, in our opinion, not only and perhaps not even primarily, calls for some kind of psychological treatment but for an improvement of the daily living situation of people with epilepsy, which may be conceived as “disadvantaged”(71–73). To improve their opportunities to participate fully in society, and consequently, to maintain or improve their daily functioning, it is necessary to change the way society views epilepsy and treats the people who have it. That is where the initiative of the WHO and ILAE, “Out of the Shadow,” is directed. A firm knowledge base is required for an effective development of this initiative and can be provided—among other things—by social science research such as ours.

Acknowledgment: We thank all the persons with epilepsy who were willing to participate in our study. Without their kind cooperation, this study would not have been possible. We also thank A.M. Beun, O.G. Mulder, J. Overweg, and J. Vermeulen from the outpatient clinics of the special centre for epilepsy “Meer en Bosch” in Heemstede and O.G. Sie from the Clinic for Neurology of the Academic Hospital Groningen. Finally, we are very grateful to Dr. W. van Schuur for his very valuable comments. The study was financially supported by a grant of NovoNordisk Netherlands.

Ancillary