Article Text
Abstract
Background: Previous prevalence studies of Parkinson’s disease (PD) in the UK have spanned a 40 year period and have predominantly been in the North of the country. These have presented rates by current age but have not examined this by age at disease onset.
Methods: A community based prevalence study was undertaken which attempted to identify all clinically diagnosed cases of PD from primary and secondary care for the city of Cardiff, Wales, UK. A meta-analysis of all past studies in the UK, including our own, was also undertaken.
Results: Overall, 380 cases of PD were identified from a population of 292 637 residents, giving a crude prevalence rate of 130 per 100 000 (95% CI 117 to 144) and an age standardised rate of 142 per 100 000 (95% CI 128 156), standardised to the 1997 England and Wales population. Our prevalence rates were very similar to the weighted average of previous UK studies although there was evidence of between study heterogeneity (p = 0.0006). 5.4% and 31.2% of prevalent PD patients had their disease onset below the age of 50 or 65 years, respectively.
Conclusions: The data suggest that there are no major geographical variations in the prevalence of PD in the UK and that the age adjusted prevalence rate has remained relatively stable over the past 40 years. Although PD risk is far greater in older subjects, patients with young onset are not that uncommon in the community, and health and social care provision should reflect their needs.
Statistics from Altmetric.com
Previous epidemiological studies in the UK reported the crude prevalence of Parkinson’s disease (PD) as 113–164 per 100 000 persons.1–8 PD prevalence studies usually present data by current age and ignore age at disease onset. While young onset PD (YOPD) is rare, with an incidence of approximately 7 per million per year under the age of 50 years,9 YOPD cases may be more common among all prevalent cases because of increased survival. We have completed a community based prevalence study of PD in Cardiff, South Wales, UK. The aim of this study was to specifically examine what proportion of prevalent cases had young onset disease, to study how age at onset influences source of health care and undertake a meta-analysis of all published UK prevalence studies to examine if there are geographical and temporal variations.
METHODS
We identified potential PD cases from primary care by undertaking two standardised searches using diagnostic Read Code F12 and sub-codes and any of the dopaminergic group of drugs used in PD (see appendix 1 online), excluding bromocriptine and amantadine because of their frequent use in conditions other than PD. After removing duplicates, a unique list of general practitioner case records were screened (MMW). Cases were defined as prevalent if they (a) were alive and symptomatic during the period January to December 2006, (b) were registered with a participating practice, (c) had been diagnosed as having PD and (d) met the Queen Square Brain Bank criteria10 at clinic assessment or if the information held in the clinical notes were consistent with a diagnosis of PD and without recognised exclusion criteria. Patients were invited by the general practitioners after excluding any miscoded patients and those felt unsuitable for contact (eg, patients with a terminal illness). Participating patients completed questionnaires and attended a clinical assessment. Anonymised patient information was used to monitor recruitment and calculate prevalence rates. Notes were cross checked for each patient registered with the participating Cardiff general practices and attending the main secondary care PD Clinic in Cardiff over a 6 month period and where available from consultant neurologists in Cardiff using pseudo-identifiers. In addition, neurology consultants in Cardiff were asked to recruit patients who resided in Cardiff and whom had not yet been approached.
Crude prevalence rates and 95% confidence intervals (CI) were calculated using the Poisson distribution and standardised to the 1997 England and Wales population. Previous studies were reanalysed using the same standard population. We calculated the between study variation from a mixed effects Poisson model.
RESULTS
Forty-five out of 54 (88%) of the eligible primary care practices participated, providing a population denominator of 292 637 (96% of the Cardiff population). The diagnostic search yielded 731 unique potential cases (see web fig 1 online). A total of 378 cases were excluded (210 were prescribed dopaminergic agents for an alternative diagnosis, 96 cases had secondary parkinsonism, 72 patients did not have PD as either miscoded (26), essential tremor (35) or did not have their diagnosis revised after specialist review (11)).
We identified an additional 49 cases (20 from general neurology clinics, 27 from PD clinic, one self-referral, one other), resulting in 402 potential cases. Fifteen cases were excluded before examination as not being PD. A total of 132 of the 387 potential prevalent cases were examined (35%) and seven additional cases (5%) were excluded as not PD. Only one diagnosis revised case had not seen a secondary care specialist. This resulted in 380 prevalent cases (182 females, 198 male).
Prevalence rates increased with age (see table 1) although males rates were greater at all ages. The age adjusted rates were 142 per 100 000 (95% CI 128 to 156 per 100 000) overall, and male and female rates were 171 (95% CI 147 to 195) and 120 (95% CI 103 to 138), respectively. The male to female prevalence rate ratio was 1.43 (95% CI 1.17 to 1.76; p = 0.001).
Age at symptom onset was obtained for 336 of the 380 (88%) prevalent patients (see web table 1 online). For men, mean age at onset was 67.7 years (95% CI 66.0 to 69.3 years) and for women mean age at onset was 69.7 years (95% CI 68.1 to 71.4 years). There was a bimodal distribution of age at onset with peaks at 40–44 years and 75–79 years. PD had begun in 3.6% of patients before the age of 45 years. The average disease duration for the prevalent group was 6.2 years. The YOPD group (onset <45 years) had longer disease duration (11.1 years, 95% CI 7.9 to 14.4 years) than the later onset (onset ⩾45 years) PD group (6.0 years, 95% CI 5.4 to 6.5) with a significant difference of 5.1 years (95% CI 2.2 to 8.0; p = 0.0006). Cases who did not attend the clinic had an older age at onset but had the same duration of disease as those who did (see web table 2 online). In total, 85% of PD patients identified in primary care had been seen by a neurologist or PD specialist. Neurologists were most likely to have made the diagnosis of PD (44%). For YOPD, a neurologist made the diagnosis in 90% of cases (p<0.001) although in a substantial number of cases (31%) there was no information available. Overall follow-up care was mainly provided by the geriatrician led PD clinic (63%) and neurologists (17%) although, again, details were not available on follow-up care in 31%. Of those patients assessed in the research clinic, 45.5% were under follow-up by the movement disorder geriatrician, 22.7% by neurologists, 4.5% by general practitioners and 2.3% by other consultants.
Our prevalence rate (142 per 100 000) was very close to the global average for all previous UK prevalence studies (fig 1) but there was evidence of statistical heterogeneity (p value = 0.0006). Prevalence rates showed no systematic temporal change over the past 42 years.
DISCUSSION
We found that the prevalence of PD in Cardiff is similar to other studies in the UK (at approximately 140 per 100 000). Although the prevalence of young PD cases is much lower than that for older cases, 5.4% and 31.2% of the total prevalent PD population have their disease onset before 50 years or 65 years of age, respectively. Approximately 1 in 20 PD patients develop disease before the age of 45 years, normally considered to be a cut-off at which genetic autosomal recessive PD should be considered. While the crude rate of PD in those under 50 years in Cardiff is 2.4 per 100 000, the prevalence of those whose disease started before the age of 50 years is greater at 6.2 per 100 000. These figures equate to approximately 3700 people in the UK currently alive with YOPD and means that YOPD, defined by an age of onset below 50 years, is at least as common as motor neuron disease, Huntington disease and progressive supranuclear palsy.
Our meta-analysis of all UK studies suggests that the prevalence of PD in the UK is fairly uniform and has been relatively stable over the past 40 years. This is somewhat surprising given the increase in overall life expectancy, increased awareness and better management of PD, which would if anything increase prevalence. It is possible that incidence has declined but the incidence trends from 1976 to 1990 from Rochester, Minnesota, were relatively stable.11 More recent studies are likely to have excluded other parkinsonian conditions, such as multiple system atrophy (approximately 10% of cases),12 which may have been misdiagnosed as PD in the past.
We believe that the heterogeneity between studies can be explained by methodological differences. The study from Aberdeen actively searched for cases in nursing homes, which would have an over-representation of PD cases.6 In contrast, the study from rural Wales5 did not use diagnostic databases but relied on prescription records and referrals. In addition, it is not known whether PD survival is worse in rural compared with urban areas.
Almost all of the YOPD patients were diagnosed by a neurologist but only one-third of the later onset patients. Recent national guidelines13 suggest that the diagnosis of PD should always be made by a specialist but this was not always the case at the time of our study.
Neurologists only looked after about half of the YOPD patients, and few (9%) of those whose disease started after the age of 65 years. A substantial minority (15%) of PD patients were followed-up solely by their general practitioners.
Limitations
Although we obtained a very high rate of participation from our primary care practices in Cardiff, we only clinically examined about one-third of all the cases, and older cases were less likely to be seen. We also did not invite cases with a diagnosis of essential tremor although the previous London study3 identified patients with unrecognised tremor dominant PD. In that study, 11 of the 56 patients with non-parkinsonian tremor had probable or possible PD and their inclusion increased the number of PD cases by 7.9%.14
In conclusion, our prevalence estimate suggests that rates have remained relatively constant over the past 40 years. Although PD is a disease of the elderly, YOPD is far more common than appreciated, and the provision of health and social care for these patients should be an important consideration.
Acknowledgments
We would like to thank all the patients, practice managers and general practitioners who helped with this study. Ross Harris provided statistical and programming support for the forest plot. We would like to thank John Gallacher who provided some corrected data for the Hobson et al (2005)2 paper. We are very grateful to Pete Shiarly for database design and management.
REFERENCES
Supplementary materials
Web Only Data JNNP.2008.162222
Files in this Data Supplement:
Footnotes
Competing interests: None.
Funding: This work was supported by the Parkinson’s Disease Society of the United Kingdom (grant No 8047) and The Ipsen Research Fund.
Ethics approval: Approval for this study was obtained from the Multicentre Research Ethics Committee for Wales (MREC), the Local Health Board and the appropriate Research and Development Departments.
▸ Appendices 1 and 2 and an additional figure and table are published online only at http://jnnp.bmj.com/content/vol80/issue7