Prevalence and risk factors of RLS in an elderly population

Methods.

The MEMO Study is a follow-up project of the 1989–1990 World Health Organization Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) Survey–Augsburg, Germany.9 MEMO examines cognitive function and cardiovascular risk factors for neurodegenerative diseases in the elderly. The study was restricted to participants of the second MONICA survey, who were 65 years or older on October 1, 1997, and lived in the city of Augsburg or one of two neighboring suburbs. Sampling of participants in MEMO was done by age groups (65 to 69, 70 to 74, 75+) to represent a broad range between 65 and 83 years. The overall response rate among those eligible was 60.6%, yielding a total of 385 participants. Sixteen individuals had to be excluded because of incomplete information on RLS, leaving a study population of 369 for this analysis. The data were collected between October 1997 and April 1998. All participants were interviewed by two RLS-trained physicians in face-to-face interviews. The following standardized questions were used to assess RLS according to the International Restless Legs Syndrome Study Group8:

1) Do you have unpleasant sensations like crawling, paraesthesias, or pain in the legs combined with a motor restlessness and an urge to move?

2) Do these symptoms occur only at rest and does moving improve them?

3) Are these symptoms worse in the evening or at night compared with the morning?

The three answer categories included “Yes,” “No,” or “I don’t know.” Participants were only classified as RLS positive if they answered all three questions with Yes.

General medical history and history of neurologic diseases including screening questions for parkinsonism10 and psychiatric diseases were assessed in interview form. Current medication, taken within the last 7 days, and lifetime exposure to neuroleptic drugs were noted. Participants who did not know their current drug therapy received a medication form and a stamped envelope to fill in and send back to the study center. The Center of Epidemiologic Studies Depression Scale (CES-D)11-12 was used to assess symptoms of depression, and the Short Form (SF)-3613-15 to evaluate quality of life. All participants were examined using standardized neurologic examination that included items from the motor subscore of the Unified Parkinson’s Disease Rating Scale.16 They also completed a neuropsychological test battery, including the standardized version of the Mini-Mental State Examination (MMSE).17

Results.

Table 1 summarizes sociodemographic characteristics of the entire study population. The mean age was 72.7 years with a range from 65 to 83. Almost half of the participants were women. Diabetes was reported by 10% of the population and about 1 in 4 subjects was found to be considerably overweight (body mass index [BMI] ≥ 30).

Thirty-six of 369 participants were classified as RLS positive, resulting in an overall prevalence of 9.76% for RLS. Table 2 shows the prevalence of RLS in the entire population, as well as the gender and age distribution. There was a significant gender difference (p = 0.012), with RLS prevalence of 13.9% in women and 6.1% in men. Men showed a trend toward decrease of RLS with age, whereas prevalences in women were similar in all age groups.

Clinical characteristics are given in table 3. Male RLS cases were on average taller but had the same weight as male non-RLS cases, who were more often overweight (not shown). Female cases had the same BMI as noncases. RLS-positive subjects reported less diabetes and significantly lower rates of hypertension. A high percentage were never smokers. No differences were observed on performance tests. In self-perceived health status, the RLS cases showed significantly lower scores in depression and mental health; the cases also had lower self-perceived physical health scores, but the differences were not significant. No differences in cognitive function as measured by the MMSE were seen between cases and noncases.

Forty-seven participants were excluded from the drug analysis because they did not send in the medication form, even after the reminder. RLS-positive individuals were prescribed benzodiazepines, neuroleptics, and estrogens more often than RLS negatives, but the differences were not statistically significant. The use of thyroid medication was similar in both groups.

The number of general practitioner visits and utilization of medical care in the previous month were not different between RLS positives and negatives (not shown). RLS positives sought advice from patient support groups more often (p = 0.02) than did RLS negatives.

In multivariate analysis we identified female gender as the only significant risk factor (OR 2.46) among all. Additionally, in men, but not in women, we observed a significant decreasing trend in risk of RLS with age and a high depression score to be significantly associated with RLS (table 4).

Discussion.

We found an overall prevalence of 9.8% of RLS-positive participants in a general population 65 years of age and older. The prevalence of RLS was higher in women compared with men. Elderly women showed stable prevalence rates in all age groups, whereas men had a decreasing prevalence between 65 and 83 years. RLS-positive subjects took slightly more benzodiazepines, neuroleptics, and estrogen, although the absolute number of participants taking these drugs was too small to be statistically significant. Individuals with RLS symptoms had higher depression scores and lower quality of mental health compared with RLS-negative participants.

This is the first population-based study that applies the minimal diagnostic criteria for RLS8 to assess the prevalence of RLS. In earlier studies, prevalence estimations of RLS varied between 2.51 and 29%.2 A number of reasons might be responsible for this wide range of prevalences: 1) Whereas one study group1 examined healthy participants, another2 assessed the frequency of RLS in outpatient veterans. 2) Because RLS seems to be influenced by genetic susceptibility, regional variations of genetically different populations such as in Atlantic provinces and Quebec may occur.6 3) Previous studies applied different diagnostic criteria. In some studies RLS was diagnosed clinically by a physician1; in others, the participants filled out a questionnaire.2,6-7

Our result of 9.8% RLS-positive participants is similar to the results of the two other population-based studies.6,7 Both studies reported RLS-typical symptoms occurring in 10 to 15% of the population older than 18 years. In the current study, we found a clear and significant gender difference, with RLS affecting women twice as often as men. Among all participants we identified female gender as the only significant risk factor (OR 2.46). In the first of the two other population-based studies,6 women showed “bedtime leg restlessness” significantly more often, and in the second,7 women more often had RLS-typical symptoms (19% versus 14.4%). This gender difference has not been explained. One hypothesis is that the postmenopausal intake of estrogen may play a role in the clinical manifestation of RLS in elderly women. However, the overall number of women taking estrogen in our study is too small to assume any relationship between estrogen intake and the occurrence of RLS. This association could be of particular interest, because an increased occurrence of RLS during the last trimester of pregnancy, a condition associated with high estrogen levels, has been described.1,18

Previous studies discussed the relationship between age and RLS. Some studies found either no age dependency1 or an increasing prevalence with age.6 RLS patients with an early manifestation of the syndrome, i.e., in childhood or adolescence, exist mostly within families of hereditary RLS.19-22 Population-based studies in young adults have not been performed. In our study, the prevalence of RLS seemed to decrease with age in elderly men and remained nearly stable in elderly women. Other study groups described associations between RLS and diabetes5 and RLS and parkinsonian syndromes.23 We were unable to evaluate these associations because only one RLS-positive person had diabetes and none had a parkinsonian syndrome. There was no clear relationship between RLS and dopaminergic or dopaminantagonistic treatment in our study population. The higher percentage of benzodiazepine intake among RLS positives (6.3% compared with 1.7% in RLS negatives) is probably due to attempts to treat the sleep disorders by the patients and their general practitioners.

RLS-positive subjects had significantly higher depression scores than did negatives, even if the question about “sleep disorders” in the CES-D was excluded. Only among men, but not among women, a high depression score was significantly associated with RLS. Also only among men, the risk associated with RLS decreased with age (significant trend). However, a cross-sectional study such as MEMO cannot differentiate if depression is a consequence of the syndrome or if it existed before RLS symptoms occurred. Prospective cohort studies are needed to answer this question. However, symptoms such as anxiety and concentration that are assessed in the CES-D could also be RLS related. A Swiss study reported a higher score for depression using the SCL-90-R-items (SCL-90-R-scale24) in RLS patients compared with patients with varicosis and compared with normal controls,25 but applied different clinical criteria for diagnosing RLS. The significantly lower mental and general health scores and nonsignificant lower physical health scores in the RLS group demonstrate the considerable impact of RLS on well-being. Another study group7 reported a similar association between RLS and physical health. Utilization of medical care was not different among RLS cases and noncases, but cases sought significantly more help from patient support groups. We interpret this as a consequence of not having been diagnosed with RLS by general practitioners, as RLS is still an underrecognized disease. We did not directly ask the participants if RLS was diagnosed or treated before the study. None of the RLS-positive participants received dopaminergic medication, which is considered as therapy of first choice in RLS.26 Therefore, we assume that the 36 participants with RLS symptoms have never been diagnosed with RLS.