1

Wermuth

Prevalence and Incidence of Parkinson's dDisease in the Faroe Islands

L.Wermuth, S. Bech2, M. Skaalum Petersen3, P. Joensen4 , P. Weihe2,3, P.Grandjean3,5

1Department of Neurology, Odense University Hospital, Odense, 2Department of Occupational and Public Health, The Faroese Hospital System, Torshavn, Faroe Islands, 3Department of Environmental Medicine, University of Southern Denmark, Odense, 4Neurologist in private practice and consultant at the National Hospital, Tórshavn, Faroe Islands, Denmark; 5Department of Environmental Health, Harvard School of Public Health, Boston, MA, USA

Address for correspondence

Lene Wermuth, MD

Department of Neurology, Odense University Hospital

Sondre Boulevard 29, 5000 Odense C, Denmark

Ph: 045 65414911; fax 045 65916003; email:

Disclosure: The authors have reported no conflicts of interest.

Short title:

Epidemiology of PD in The Faroe Islands

Key words

Epidemiology studies, Incidence, Parkinson´s disease, Prevalence, The Faroe Islands

Abstract

Objective: A study in the Faroe Islands in 1995 suggested a high prevalence of idiopathic Parkinson’s disease (IPD) and total parkinsonism of 187.6 and 233.4 per 100,000 inhabitants, respectively. Methods: Detailed case-finding methods ten years later were used and a neurologist has verified the diagnosis. Results: The crude prevalence of IPD and total parkinsonism was 206.7 per 100,000 and 227.4 per 100,000, respectively. The age-adjusted prevalence is twice as high as data from Norway and Denmark. Age at initiation of treatment and the fatality rate did not explain the increased prevalence. During 1995-2005, the average annual incidence was 21.1 per 100,000 persons for Parkinson’s disease, and 22.9 per 100,000 persons, if including atypical parkinsonism. Conclusion: The high prevalence was verified and linked to a high incidence. The cause of the high prevalence is unknown, but neurotoxic contaminants in traditional food may play a role in the pathogenesis in this population, perhaps jointly with genetic predisposition.

Introduction

Epidemiological surveys of Parkinson’s disease (PD) have shown substantial geographical variation in prevalence, even if taking into account differences in methods for case identification [1,2]. Thus, case-finding must take into account such problems as diagnostic delay and the probable existence of mild clinical cases that can cause diagnostic challenges among the elderly. Using a combination of approaches to systematic case-finding in the Faroe Islands, an age-adjusted prevalence of idiopathic PD was found to be as high as 183.3 per 100,000 persons in 1995 [3]. Using the same methods, a high prevalence was also recorded in Greenland Inuit, where the age-adjusted PD prevalence was 187.5 per 100,000 [4]. A much lower age-adjusted PD prevalence of 98.3 per 100,000 was found on the island of Als, Denmark [5], and the latter result agrees with the prevalence of 102.4 per 100,000 in Rogaland, Norway, where comparable methods for case-finding were used [6], and with results from other studies of European populations [1]. The doubled PD prevalence in the Faroe Islands is difficult to explain. Differences in fatality rates among patients in different populations could conceivably play a role. Known geographic variations in the mortality of PD patients might support this possibility [7-9], but any major impact within the Nordic communities would be unlikely. Studies on European populations include few high-quality data on prevalence and incidence, but the PD occurrence in northernmost Europe, including Iceland, appears to be increased [1,10].

The present project was therefore carried out as a case-finding study in the same population to identify all new cases during the ten years after the first study to assess the possible changes in prevalence and incidence over time.

Methods

In order to identify all potential PD cases, a combination of different case-finding methods was used. First, an invitation to participate in the study was sent to all Faroese residents receiving levodopa-containing drugs (LD) and dopa-agonists, as determined from records kept by the Chief Pharmacist in the Faroe Islands during the period between April 15, 2004 and April 15, 2005. Secondly, patients diagnosed with PD or parkinsonism by the local neurology specialist, at the National Hospital in Torshavn, or at the two regional hospitals in Suðuroy and Klaksvík were also contacted. Finally, requests were sent to all general practitioners (GPs) to contact their patients with possible or confirmed PD or parkinsonism. In addition, the study was highlighted in the news media, and self-referrals were included.

The contacts to the patients were arranged according to requirements determined by the ethical review committee serving the Faroe Islands and the institutional review board at Harvard School of Public Health. Thus, all patients known to be receiving anti-parkinson medicine received a letter of invitation and a follow-up telephone call from the Faroese Chief Pharmacist’s office. The patients diagnosed at the hospitals, by the neurologist and the GPs received a similar letter of invitation but no follow-up telephone call from these institutions. If a patient agreed to participate, they had to make contact by telephone or mail.

Patients accepting to participate in the study were invited for an examination in Tórshavn, or, when indicated, they were visited at their residence. Patients diagnosed with PD by the local neurology specialist (PJ) and with a good response to levodopa were interviewed and examined by the neurology resident who had one year of prior training in a movement disorder clinic (SB). All other patients identified as possible PD cases were also examined by a neurologist and specialist in movement disorders (LW),

who also trained and supervised the resident.

The diagnostic assessment was based on the clinical information, the course of disease development, and the response to levodopa. Subjects were diagnosed as having PD if they had at least two of the three cardinal signs of resting tremor, bradykinesia, and rigidity. The severity of the disease was rated by the Hoehn-Yahr scale, and all patients were classified into subgroups of definite, probable, and possible PD [11-132]. In doubtful cases and for classification of PD patients, we emphasized unilateral onset, any persisting asymmetry, and response to levodopa as key factors. Patients with Parkinson-like symptoms and the occurrence of additional features such as early falls, prominent autonomic symptoms, abnormalities of eye movements, dementia, cerebellar ataxia or pyramidal signs were diagnosed as having atypical parkinsonism, such as progressive supranuclear palsy (PSP), multiple system atrophy (MSA), corticobasal degeneration, and dementia with Lewy bodies (LBD) [143-176]. Patients with parkinsonism caused by cerebrovascular disease or explained by other disease or symptomatic causes were excluded. CT scan was available from most of the patients, thereby adding further reliability to the diagnosis.

The number of cases was assessed as the number alive on July 1, 2005. In estimating the crude prevalence and the age- and gender-specific prevalence, we used the 2005 Faroe population size of 48,371 inhabitants on January 1, because the age distribution was available at the Faroese Registry of Residents for the first day of each calendar year. In estimating incidence rates, account was taken that the study population increased from 43,393 inhabitants in 1995 to 48,371 inhabitants in 2005 [187].

The FaroeseEthical Review Committee and the IRB in the US approved the protocol. Written informed consent was obtained from each volunteer on the basis of verbal and written information in Faroese.

Results

We identified 153 patients with possible PD or parkinsonism. Of these, 125 patients were listed as being prescribed LD-containing drugs or dopa-agonists from Faroese pharmacies. The local neurologist’s records showed 71 patients with a diagnosis of PD. Most of these had already been identified through the Chief Pharmacist´s office, and only eight names were new. We also identified two new patients from the general practitioners, thirteen from the hospitals and an additional five from self-referrals who had become aware of the study from newspaper articles.

Of the total of 153 patients 102 were clinically examined, with the following results: 79 were confirmed to have PD, 9 had additional features consistent with a diagnosis of atypical parkinsonism, 9 were concluded to have parkinsonism due to other causes (neuroleptica usage, cerebrovascular disease), and 3 were excluded because no signs of parkinsonism were found upon examination. Two died before the prevalence date of July 1, 2005.

Of the remaining 51 patients, 19 declined clinical examination, 16 were non-responders, 12 were excluded due to other verified neurological diagnoses, and 4 had died. For the non-participating individuals, information available was limited to the diagnosis and the time of onset of symptoms obtained from clinical files within the health care system. Among the total of 35 decliners and non-responders, the diagnosis of PD was confirmed in 21, 1 had atypical parkinsonism, 5 had parkinsonism deemed to be due to cerebrovascular disease, and 3 had parkinsonism preceded by long-term neuroleptica usage. The remaining five had Alzheimer’s disease (1), epilepsy (1), no recorded history of parkinsonian symptoms (1), and two were suspected of Parkinson’s disease, but had never been evaluated by a neurologist (2).

Based on these data, a total of 127 patients were found to have parkinsonism, of whom 100 had PD according to current diagnostic criteria, and 10 had atypical parkinsonism (table 1). Specific diagnoses were sought for all atypical cases, but detailed classification was not attempted when severe dementia was present. The remaining 17 were found to have parkinsonism due to multi-infarct syndrome, as determined from clinical records including imaging results or long-term use of neuroleptica.

The relative distribution of confirmed PD, parkinsonism, other disease or exclusions does not seem to differ between participants, non-responders, and those who declined examination.

Table 2 shows that only seventeen of the 82 patients identified in the 1995 study were alive ten years later. Two of the seventeen patients were excluded from the prevalence calculation because of lack of neurological evaluation (1) and incorrect diagnosis of PD (1). During the subsequent 10-year period, 80 new PD cases had occurred, and the patients were alive on July 1, 2005 and residing in the Faroes. These subjects had been mainly identified through the pharmacy records. Two patients were identified that had started treatment before 1995 but were not a part of the study conducted in 1995. For another two patients, no information was available on the year of diagnosis or the year treatment started, and one additional patient did not receive any treatment, although diagnosed in 2002. One PD patient diagnosed in between the two studies had emigrated. Sixteen new patients had been diagnosed with PD, but died within the 10-year period.

With a total Faroese population of 48.371 inhabitants in 2005, the crude prevalence of idiopathic PD was therefore 206.7 per 100,000 persons and, if atypical parkinsonism patients were included, 227.4 per 100,000 persons. If age-adjusted to the 1995 population of Rogaland, Norway [6], the prevalence of PD was 218.0 per 100,000, i.e., about twice as high as in Norway. Based on the patients’ current residence, the prevalence was also calculated for separate island communities, taking into account the number of residents above 45 years (table 3). The prevalence in 5-year age groups and sex are shown in figure 1. Of note is that only two patients were below 50 years at the time of examination, and that a total of four men and two women were first diagnosed at an age below 50 years.

The approximate incidence rate can also be calculated. During this 10-year period, the average population size in the Faroe Islands was 45,878. When including the patient who emigrated, there were 97 and 8 incident cases of Parkinson’s disease and atypical parkinsonism, respectively, during 1995-2005. The average annual incidence rates during this period were therefore 21.1 per 100,000 persons for PD, and 22.9 per 100,000 persons, if including atypical parkinsonism.

The proportion of definite PD (61 %) in this study is higher than in the earlier study in 1995 (38 %). This difference mainly reflects the number of patients that were only considered possible PD cases at the time, rather than a change in prevalence or disease severity. The average age at the time of the examination was nearly the same, 74.3 years in 2005, as compared to 73.4 years in 1995. Likewise, average LD dosage and the mean Hoehn and Yahr score were similar in the two studies.

For the 2005 PD cases, the mean age at onset of symptoms was 66.0 years, and the mean age at onset of treatment was 67.9 years. The average dose of LD was 354 mg in 2005 (as compared to 407 mg ten years before). The LD response was good in 53 (57 %) of the cases and moderate in 29 (31.2 %).At the time of examination, 69 (75.8 %) of the cases had resting tremor. On the Hoehn-Yahr scale, 28 (32.2%) of the PD patients could be categorized as stage 1 or 1.5, an additional 27 (31.0%) as stage 2 or 2.5, 11 (12.6 %) as stage 3, and 21 (24.1 %) as stages 4 or 5. The mean score was 2.5 (as compared to 2 ten years previously). Nineteen patients (23.8%) lived in a nursing home or similar institution, two patients (2.5%) lived with relatives, and 59 resided in their own home (73.8%). Thirty-six patients (41.9%) had professional assistance to manage the activities of daily life.

Twenty-six of the 79 PD patients who were clinically examined (33 %) reported other family members (parents or siblings) with parkinsonism (Parkinson’s disease or signs of shaking palsy), and 5 of these patients (6%) reported parkinsonism in both parents and siblings.

The time since initiation of treatment may reflect the usual duration of the disease; the average was 6.7 years in this study.

Discussion

A major strength of this study is the case ascertainment and the diagnostic accuracy and was at the same high level as the previous study carried out in the Faroe Islands in 1995. While the same data sources were used for case ascertainment, the more stringent ethical requirements in 2005 only limited the extent to which the patients could be contacted by telephone; in the previous study, subjects could be contacted by telephone, if needed, to obtain a response to the invitation letter. The validity of the results obtained is supported by the distribution of cases within age groups of men and women (figure 1) being similar to expectation from other studies (19-21) and could suggest a possible environmental causative factor or more susceptible to its effects than women..

The diagnostic validity of the PD diagnosis and the differential diagnosis of parkinsonism must be carefully considered when comparing studies. In the present study, a larger percentage of the patients examined were found to have idiopathic PD than was the case ten years before. Part of this difference could possibly be due to the difference in the procedures for inviting patients for the study. In 1995, the neurologist may have examined a larger number of atypical cases, who were recruited for participation by telephone. In the present study, 16 patients did not respond to the invitation, and an additional 19 declined clinical examination. Still, clinical records allowed classification of most of these patients, and this study therefore seems to have obtained fairly complete data. Thus, in comparison with the 1995 study, the main difference was that diagnostic separation between PD and parkinsonism had improved due to the increased access to diagnostic imagining scans. Due to the higher proportion of definite PD cases in the present study, the absence of clinical examination data in non-participants is less likely to have affected this study results to any appreciable extent.

Accordingly, utilizing the best evidence available, the present study therefore verified the increased prevalence at a level similar to the one observed ten years earlier. A small decrease in the total prevalence of PD and atypical parkinsonism may be due to the more frequent application of imaging scans that likely resulted in more exclusions due to documented multi-infarct syndrome. Both studies suggest that the prevalence of idiopathic PD in the Faroe Islands is about twice as high as in two comparable Nordic populations at the island of Als, Denmark, and Rogaland, Norway, with age-adjusted prevalences of 98.3 per 100,000 [5] and 102.4 per 100,000 [6] in these two Scandinavian populations. All of these studies used very similar and meticulous methods for patient identification and diagnosis.

Methodological reasons are like to affect for the PD prevalences only to a minor extent. Because of the demographic similarity of the Faroese, Norwegian, and Danish populations, age adjustment results in only small changes of the crude prevalences. Earlier diagnosis or greater ascertainment of cases with mild PD is also unlikely reasons, given the use of similar approaches to case-finding.

In support of the validity of the higher prevalence, the incidence rate in the Faroes is also higher than in most other studies [1,198-210]. Although high-quality data for PD incidence are scarce [1], reliable results from past studies in Iceland show a similar crude incidence of 16.0 per 100,000 persons, which corresponds to an increased crude PD prevalence of 162.3 patients per 100,000 persons [221]. Another northern population with a high PD prevalence is the Greenland Inuit [4]. Thus, a possible geographic pattern may be present in these isolated populations. However, as a fairly stable and ethnically homogeneous Nordic population, the Faroese may best be compared to the residents of the island of Als, Denmark, where the PD prevalence was found to be half as high in a study using the same methodology [5].

The majority of inherited PD cases have their debut below age 50 years [232], but the present study identified only six patients first diagnosed before age 50 years. While a study from Iceland [243] demonstrated a significant familial aggregation also of patients with late-onset disease, only 33% of Faroese PD patients (22% in 1995) reported that other family members (parents or siblings) had PD or parkinsonism, and 5-6% on both occasions reported PD or parkinsonism in both parents and siblings. For comparison, a previous Danish study found that nine PD patients out of 50 (18%) reported other family members with PD or parkinsonism. Possible predisposition to PD development could be due to mutations of relevant genes, perhaps including the hematochromatosis gene [254] and one of the CYP genes [265] that are known to occur more frequently in the Faroese than in other Caucasian populations. These genes are involved in biochemical defenses against peroxidation [276] and could perhaps be of importance in degenerative diseases propagated by this mechanism.

The higher incidence of PD seen in the Faroes could also be linked to unique environmental risk factors. In this regard, the distribution of PD cases may hold some clues to possible etiologies. Although the present residence may not necessarily be long-term, the findings are similar from the two studies conducted. Thus, the highest prevalence occurred at the island of Sandoy. This island has a relatively small population, but the residents have frequent and steady access to pilot whale meat and blubber from kills that occur on the beaches of this island. Whale meat and blubber contain high concentrations of mercury and polychlorinated biphenyls (PCBs), respectively [287], and constitute the main sources of methylmercury and PCB respectively in the Faroes [298]. These neurotoxicants have been implicated as possible risk factors for the development of PD [2930-332], and past, perhaps long-term, exposure to these food contaminants could therefore be hypothesized to contribute to an increased PD risk in this island population.