Figures
Abstract
Background
Impaired balance is a major problem in patients with idiopathic Parkinson’s disease (PD) resulting in an increased risk of falls and fall-related fractures. Most studies which analyzed the risk of femoral fractures in patients with idiopathic PD were performed either in specialized centers or excluded very frail patients. The current study used a large population-based dataset in order to analyze the risk of femoral fractures in patients with idiopathic PD.
Methods
Data from more than 880.000 individuals aged 65 years or older and insured between 2004 and 2009 at a large German health insurance company were used for the analyses. Persons with idiopathic PD were identified by the dispensing of Parkinson-specific medication and by hospital diagnoses, if available. People without PD served as the reference group. Incident femoral fractures were obtained from hospital diagnoses. Analyses were stratified by gender and information on severe functional impairment (care need) as provided by reimbursement claims.
Results
Compared with the reference group, persons with idiopathic PD had a more than doubled risk to sustain a femoral fracture. The risk was higher in men (HR = 2.61; 95%-CI: 2.28–2.98) than in women (HR = 1.79; 95%-CI: 1.66–1.94). The increased risk was only observed in people without severe functional impairment. The sensitivity analysis using a refined definition of idiopathic PD patients yielded similar results.
Citation: Benzinger P, Rapp K, Maetzler W, König H-H, Jaensch A, Klenk J, et al. (2014) Risk for Femoral Fractures in Parkinson’s Disease Patients with and without Severe Functional Impairment. PLoS ONE 9(5): e97073. https://doi.org/10.1371/journal.pone.0097073
Editor: Mathias Toft, Oslo University Hospital, Norway
Received: November 22, 2013; Accepted: April 15, 2014; Published: May 22, 2014
Copyright: © 2014 Benzinger et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The analysis was supported by the Bundesministerium für Bildung und Forschung (Förderkennzeichen: 01EC1007A). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Parkinson’s disease (PD) is the second most common neurodegenerative disorder after Alzheimer’s disease [1]. Postural instability and impaired balance is one of the cardinal symptoms of PD posing these patients at an increased risk of falling [2]–[7]. Fall-related injuries represent the most frequent reason for hospital admission in PD [8]. The occurrence of falls in patients with PD has been shown to be associated with increasing age, severity of symptoms, and use of dopaminergic drugs [3], [4], [7], [9].
The increased risk of falling in combination with a low bone mineral density is putting PD patients at high risk for osteoporotic fractures [10]. In several studies, PD patients were found to have a high rate of fractures [2], [11]–[16]. Among these, femoral fractures are the most common type of non-vertebral fractures [7], [12]–[14], [16]. Femoral fractures are associated with high morbidity, mortality, and costs in the general older population [17]–[19]. In PD patients, femoral fractures are associated with a particularly high risk of unfavorable outcomes such as admission to nursing homes [20]. However, estimates of the magnitude of these patients’ increased risk are mainly based on observational studies with a limited number of PD patients and only a low number of femoral fractures [13]–[16].
PD is characterized by loss of functional abilities as the disease progresses [21]–[23]. However, the influence of functional limitations on the excess risk for femoral fractures in PD patients has not been explored so far. A large German population-based dataset containing routine data allowed us to identify older individuals with severe functional limitations in activities of daily living. The aim of this study was (1) to estimate the risk of idiopathic PD patients to sustain a femoral fracture and (2) to stratify this estimate for the presence of severe functional impairment.
Materials and Methods
Data Source
The routine data collection systems of the largest health insurance company in Bavaria, the Allgemeine Ortskrankenkasse Bayern (AOK Bavaria), was used to select data on gender, age, long term care need, admission to hospital, admission and discharge diagnosis as well as dispensed medication. Health insurance and long term care insurance is statutory in Germany. The AOK Bavaria covers nearly 50% of the population aged 65 years and over in Bavaria, a federal state with 12.5 million inhabitants.
Data on Medication to Treat PD
Within the stored data of the AOK Bavaria, all prescriptions for medication for the treatment of PD were identified (Tab. 1). Substances were categorized according the Anatomical Therapeutic Chemical Classification System (ATC code) to substance classes. In Germany, outside hospitals Parkinson medication is available only at pharmacies with a written prescription by a physician and is reimbursed by the person’s health insurance. For reimbursement, a person’s insurance number along with type, dose and amount of drug prescribed, as well as prescription date is transferred to the health insurance. No information on the dosage regime or diagnosis is transferred. Data on reimbursement held by health insurances give complete information on all prescriptions filled in.
Study Population
The dataset consisted of 932,197 people aged 65 years and over who were member at AOK Bavaria between January 1st 2004 and June 30th 2009. Data from the year 2004 (January 1st to December 31st) regarding admission or discharge diagnoses as well as PD medications were used to identify patients with idiopathic PD. Individuals dying in 2004 were not considered (N = 44,333). Identification of the study population and sub-groups is displayed in figure 1.
Identification was based on characteristics in 2004 as provided by the health insurance company (AOK: Allgemeine Ortskrankenkasse Bavaria) PD: Parkinson’s disease; N: number of participants.
Identification of Individuals with Idiopathic PD
Identification of patients with PD using routine data is not straightforward. In our dataset we could use two types of information to identify PD patients: (1) primary hospital admission and discharge diagnoses of hospitalized patients, and (2) information about antiparkinson medication dispensed by pharmacies outside the hospital. Hospital admission and discharge diagnoses covered only a small percentage of PD patients. Therefore, the main information to identify patients with idiopathic PD was the medication to treat PD.
We defined individuals as patients with an idiopathic PD if a) antiparkinson medication was dispensed at least twice in 2004, or once in 2004 and at least once in 2005 or b) idiopathic PD was the primary hospital admission or discharge diagnosis in 2004 (ICD-10: G20) and antiparkinson medication was dispensed at least once in 2004.
Most patients with idiopathic PD are treated with antiparkinson drugs and are covered by the chosen method. However, some other disease entities are also treated by antiparkinson drugs like atypical Parkinsonism, secondary Parkinsonism or restless legs syndrome (RLS). Therefore, these individuals were excluded if they could be identified by hospital admission or discharge diagnoses (ICD-10) (G21/G22/G23/G25.81). The final dataset comprised 323, 404 men and 560, 453 women.
This dataset still included patients with atypical Parkinsonism, secondary Parkinsonism or RLS. Atypical Parkinsonism and secondary Parkinsonism are very rare diseases. RLS, however, has a higher prevalence and might account for a relevant percentage of the dispensed antiparkinson medication. Medically treated RLS is usually treated by monotherapy with dopaminergic drugs. Therefore, we applied two different models to perform a sensitivity analysis. Model I (‘possible’ idiopathic PD patients) used data of all patients with antiparkinson medication (see above). Most patients with a RLS are still included in this model. Model II (‘probable’ idiopathic PD patients) excluded all individuals (n = 11,078) from model I who received levodopa only or dopamine agonists only (change of substance within and between substance class possible). In this model most patients with a RLS but also patients with an idiopathic PD treated by monotherapy with dopaminerigic drugs are excluded. Therefore, model II may over-represent patients with an advanced idiopathic PD.
Please note that these terms are not synonymous with the terms for clinical definition of PD [24] and are therefore put in brackets throughout this article.
Reference group: The reference group comprised of individuals without intake of antiparkinson medication and no hospital admission or discharge diagnosis of atypical Parkinsonism, secondary Parkinsonism or RLS in 2004.
Definition of Severe Functional Impairment
The need for long-term nursing care was used as a marker for severe functional impairment and was assessed at the beginning of the observation period (January 1st 2005). In Germany, most persons with a minimum of six months of need for nursing care are eligible to receive reimbursement for long-term care by the long-term care insurance. The long-term care insurance was introduced in the German social insurance system in 1995. All employed citizens are members by law [25]. Experts confirm a person’s eligibility to receive long-term care benefit. Long-term care benefits are granted for professional, family, and/or informal help. In order to claim long-term care benefits, people must have a daily minimum of 90 minutes of assistance with basic activities of daily living (ADL) such as washing, eating, or dressing, and instrumental activities of daily living (IADL) such as cleaning or shopping. This information thus defines relatively well functional impairment, independent of underlying diseases.
Identification of Femoral Fractures
Hospital discharge diagnoses were used to identify femoral fractures (ICD-10: S72). The observation period started on January 1st 2005 and ended on June 30th 2009. Hospital admissions coding a femoral fracture that occurred within less than 30 days to a previous fracture were excluded in order to avoid double coding.
Statistics
The accumulation of the individual person-years of individuals with PD and without (reference group) started at beginning of observation period (January 1st 2005) and ended with censoring due to death or at the end of the observation period. Fracture rates and 95% confidence intervals were estimated per 1000 person-years and adjusted for age using negative binomial regression models. To quantify the relative risk for the first femoral fracture during observation time in people with ‘possible’ or ‘probable’ idiopathic PD compared to people without PD hazard ratios with 95% confidence intervals were calculated. In the applied proportional hazards models age of participants was used as ‘survival time’-variable to adjust for age and to account for the fact that the participants were of different ages at study begin. All models were stratified for sex and considered severe functional impairment by stratification. All statistical calculations were carried out using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA).
Results
A total of 883,857 individuals aged 65 years and older were included in our study population. The median follow-up was 4.5 years. As ‘possible’ PD patients (model I) 23,469 individuals (2.7%) were identified and 12,391 (1.4%) as ‘probable’ idiopathic PD patients (model II). Table 2 shows baseline characteristics of the reference group as well as of individuals with ‘possible’ PD (model I) and ‘probable’ PD (model II). In both models, PD-patients were older than the reference group (77.9, 76.9 years vs. 74.4 years in individuals with ‘possible’ PD, ‘probable’ PD, and in individuals of the reference group, respectively). A higher proportion of these individuals had severe functional impairment (47.7, 50.3 and 12.7% of the population with ‘possible’ PD, ‘probable’ PD, and of individuals of the reference group, respectively). However, the median age of individuals with severe functional impairment was younger in both models compared to the reference group (81.2/79.6 years vs. 83.5 years, as for individuals with ‘possible’ PD, ‘probable’ PD, and for individuals of the reference group, respectively).
As displayed in Table 3, the most commonly prescribed medication of individuals with ‘possible’ PD was levodopa either alone or in combination (79.0%). Only 1.8% of individuals received dopamine agonists alone and 9.4% in combination.
Estimates of fracture rates were based on 34,861 femoral fractures overall. Table 4 shows age adjusted fracture rates stratified by severe functional impairment and gender. Rates for femoral fractures are ranging from 3/1000 to 33/1000 person-years. In all strata and models, women had a higher fracture rate than men. The relative difference was less pronounced in individuals with ‘possible’ or ‘probable’ idiopathic PD with severe functional impairment. Figure 2 illustrates hazard ratios (HR) using the same strata. Narrowing the criteria from model I to model II increased the HR from 2.27 to 2.61 in men, and from 1.51 to 1.79 in women. The HR was higher in men than in women in both models (HR 2.27 vs. 1.51/2.61 vs. 1.79 in model I/II, respectively). In individuals without severe functional impairment, the relative fracture risk was higher than in those with functional impairment. In fact, women with ‘probable’ idiopathic PD and severe functional impairment did not have an increased risk to sustain a femoral fracture compared to women of the reference group. Men with ‘probable’ idiopathic PD and functional impairment had a low risk increase for femoral fractures of 24%. Both models showed similar results (Fig. 2).
Individuals with ‘possible’ (N = 23,469) or ‘probable’ (N = 12,391) idiopathic Parkinson’s disease (PD, for definition see text) were compared to 860,388 people without PD, stratified by gender and care need.
Discussion
In this large-scale analysis of population-based data of a health insurance company we compared femoral fracture rates in older individuals with ‘possible’ or ‘probable’ PD to a reference group without PD. The rates of femoral fractures found in our analysis are comparable to rates of femoral fractures previously reported [26]–[28].Our data demonstrate that individuals aged 65 years or older with medically treated idiopathic PD have a more than doubled (men) or almost doubled (women) risk to sustain a femoral fracture compared to the reference group. To test the robustness of our results and the susceptibility to the inclusion and exclusion criteria, two models with refined definitions were used. ‘Possible’ PD (model I) showed lower hazard ratios compared to ‘probable’ PD (model II). Both models showed a similar picture with a higher hazard ratio in those PD patients without severe functional impairment compared to PD patients with severe functional impairment and higher hazard ratio in men compared to women.
In PD patients, one out of four falls results in an injury [6]. Falls and fractures represent the leading diagnosis for admission to hospital in PD patients [8]. Our finding of an increased risk for femoral fractures in PD is in line with previous studies that demonstrated an association of PD with an increased risk of falling and a decreased bone mineral density [10]. A meta-analysis pooling data from six prospective studies found 46% of PD patients to fall over a three month period [6]. In addition to known risk factors such as age, female gender and previous falls, various disease-specific risk factors are under discussion such as pathologic gait characteristics [5], [29], [30], balance problems [5], [29], impaired cognition [29], [31], and the use of dopaminergic drugs [5], [9].
However, the excess risk we found is lower than the values reported previously ranging from a 2.2- to a 4.6-fold risk [12], [14]–[16], [32]. Our study is based on an unselected population-based sample of individuals including all stages of PD and including community-dwelling older individuals as well as nursing home residents (in both, PD patients and the reference group). This approach is in clear contrast to a number of studies recruiting from either specialist clinics or cohort studies that excluded the frailest of these patients. Such an example of a non-representative sample would be the Osteoporotic Fractures in Men study (MrOS) that explicitly excluded men unable to walk without personal assistance [11], [33]. A similar approach was chosen for the recruitment of another cohort [14]. Considering the fact that the risk of falling increases with the disease stages [4], [6], the advantage of a population-based approach is apparent.
This study introduces an additional approach which may be useful for a better interpretation of the observed risk. We were able to stratify the cohorts according to the need for long-term nursing care as a functional parameter. About every other PD patient in our dataset required long-term nursing care. Stratifying our analyses by this indicator for severe functional impairment, we retrieved a more complex picture. PD patients without severe functional impairment had a more than two fold risk to sustain a femoral fracture, compared to the reference group. In contrast, PD patients with severe functional impairment did not (women) or only slightly (men) differ from the reference group with severe functional impairment. Although we could confirm that PD is a risk factor for femoral fractures, the above data indicate that not the diagnosis per se, but functional impairment as a consequence of PD contributes relevantly to the risk of femoral fractures. Severe functional impairment expressed as care need has been shown to be a risk factor for factures [34], [35] The analysis of PD patients versus a reference group presented in this study implies that not the underlying disease is attributable for this risk but the need for help with ADLs.
Our approach has several limitations that need to be considered when interpreting the findings. The dataset used does not allow diagnosis of PD with very high security. However, by introducing a ‘possible’ and ‘probable’ PD model (which showed comparable results) and by choosing definitions that make the inclusion of severe atypical Parkinsonism (with an even higher risk of falls) in the PD cohort very probable, we argue that our approach might rather over- than under-estimate the fracture risk in the PD group. Moreover, data from the Rotterdam study indicates that our model approach indeed has a high accuracy to detect the persons of interest: They found 78% of users of Parkinson medication to have clinically confirmed idiopathic PD; intake of more than one substance class yield 100% specificity [36]. The most prevalent other condition treated with dopaminergic drugs is the restless legs syndrome (RLS) [37]. Further arguments for the usefulness of the PD models introduced here are provided by results of epidemiological studies investigating treatment behavior of the most relevant “confounding” disease, i.e. RLS. These studies found that, contrary to what is recommended in the guidelines, only a minority of patients are treated with dopaminergic drugs [38]–[40].
Another shortcoming of our data is the lack of information on dosage. Data from Denmark demonstrated a dose-dependent association between fracture risk and Parkinson’s medication [15]. Reimbursement data in the dataset used in our study did not allow reliable calculation of dosage in the absence of regimes.
We identified persons as ‘possible’ PD (model I) and ‘probable’ PD based on medication records in 2004. We could not trace back the duration of dispensing for a longer time period. Hence, this study does not allow any conclusions with respect to the duration of PD.
Our data were derived only from one health insurance company and may not be representative for the whole German population. However, the AOK is the by far largest statutory health care insurance and is open to all people. In Bavaria, the AOK covers almost 50% of the population aged 65 years and older, thus we argue that this is indeed a representative sample.
Conclusions
Our findings confirm the increased risk of femoral fractures in PD patients. However, according to our large population-based dataset it is lower than previously reported. Interestingly, diagnosis-specific differences in fracture risk was observable only in PD patients without severe functional impairment, whereas PD patients and the reference group with severe functional impairment did not show relevant differences in fracture risk. This may have relevant implications for disease-specific falls prevention programs in PD patients. Such programs may be most effective in early rather than in advanced stages of the disease.
Acknowledgments
We thank Regina Merk-Bäuml, Ralf Brum, Markus Gindl and Stefanie Büttner from the Allgemeine Ortskrankenkasse (AOK) for their support of our analyses.
Author Contributions
Conceived and designed the experiments: PB KR JK AJ GB WM HHK. Performed the experiments: PB GB AJ. Analyzed the data: PB KR JK AJ GB WM. Wrote the paper: PB KR JK AJ GB WM HHK.
References
- 1. De Lau LML, Breteler MMB (2006) Epidemiology of Parkinson’s disease. Lancet Neurol 5: 525–535
- 2. Abou-Raya S, Helmii M, Abou-Raya A (2009) Bone and mineral metabolism in older adults with Parkinson’s disease. Age Ageing 38: 675–680
- 3. Balash Y, Peretz C, Leibovich G, Herman T, Hausdorff JM, et al. (2005) Falls in outpatients with Parkinson’s disease: frequency, impact and identifying factors. J Neurol 252: 1310–1315
- 4. Bloem BR, Grimbergen YA, Cramer M, Willemsen M, Zwinderman AH (2001) Prospective assessment of falls in Parkinson’s disease. J Neurol 248: 950–958.
- 5. Matinolli M, Korpelainen JT, Sotaniemi KA, Myllylä VV, Korpelainen R (2011) Recurrent falls and mortality in Parkinson’s disease: a prospective two-year follow-up study. Acta Neurol Scand 123: 193–200
- 6. Pickering RM, Grimbergen YAM, Rigney U, Ashburn A, Mazibrada G, et al. (2007) A meta-analysis of six prospective studies of falling in Parkinson’s disease. Mov Disord 22: 1892–1900
- 7. Wielinski CL, Erickson-Davis C, Wichmann R, Walde-Douglas M, Parashos SA (2005) Falls and injuries resulting from falls among patients with Parkinson’s disease and other parkinsonian syndromes. Mov Disord 20: 410–415
- 8. Temlett JA, Thompson PD (2006) Reasons for admission to hospital for Parkinson’s disease. Intern Med J 36: 524–526
- 9. Arbouw MEL, Movig KLL, van Staa TP, Egberts ACG, Souverein PC, et al. (2011) Dopaminergic drugs and the risk of hip or femur fracture: a population-based case-control study. Osteoporos Int 22: 2197–2204
- 10. Van den Bos F, Speelman AD, Samson M, Munneke M, Bloem BR, et al. (2013) Parkinson’s disease and osteoporosis. Age Ageing 42: 156–162
- 11. Fink HA, Kuskowski MA, Taylor BC, Schousboe JT, Orwoll ES, et al. (2008) Association of Parkinson’s disease with accelerated bone loss, fractures and mortality in older men: the Osteoporotic Fractures in Men (MrOS) study. Osteoporos Int 19: 1277–1282
- 12. Genever RW, Downes TW, Medcalf P (2005) Fracture rates in Parkinson’s disease compared with age- and gender-matched controls: a retrospective cohort study. Age Ageing 34: 21–24
- 13. Johnell O, Melton LJ 3rd, Atkinson EJ, O’Fallon WM, Kurland LT (1992) Fracture risk in patients with parkinsonism: a population-based study in Olmsted County, Minnesota. Age Ageing 21: 32–38.
- 14. Schneider JL, Fink HA, Ewing SK, Ensrud KE, Cummings SR (2008) The association of Parkinson’s disease with bone mineral density and fracture in older women. Osteoporos Int 19: 1093–1097
- 15. Vestergaard P, Rejnmark L, Mosekilde L (2007) Fracture risk associated with parkinsonism and anti-Parkinson drugs. Calcif Tissue Int 81: 153–161
- 16. Melton LJ 3rd, Leibson CL, Achenbach SJ, Bower JH, Maraganore DM, et al. (2006) Fracture risk after the diagnosis of Parkinson’s disease: Influence of concomitant dementia. Mov Disord 21: 1361–1367
- 17. Bleibler F, Konnopka A, Benzinger P, Rapp K, König H-H (2013) The health burden and costs of incident fractures attributable to osteoporosis from 2010 to 2050 in Germany-a demographic simulation model. Osteoporos Int 24: 835–847
- 18. Ioannidis G, Papaioannou A, Hopman WM, Akhtar-Danesh N, Anastassiades T, et al. (2009) Relation between fractures and mortality: results from the Canadian Multicentre Osteoporosis Study. CMAJ 181: 265–271
- 19. Johnell O, Kanis JA (2006) An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int 17: 1726–1733
- 20. Clubb VJ, Clubb SE, Buckley S (2006) Parkinson’s disease patients who fracture their neck of femur: a review of outcome data. Injury 37: 929–934
- 21. Healy GN, Owen N (2010) Sedentary behaviour and biomarkers of cardiometabolic health risk in adolescents: an emerging scientific and public health issue. Rev Esp Cardiol 63: 261–264.
- 22. Jankovic J, Kapadia AS (2001) Functional decline in Parkinson disease. Arch Neurol 58: 1611–1615.
- 23. Harrison MB, Wylie SA, Frysinger RC, Patrie JT, Huss DS, et al. (2009) UPDRS activity of daily living score as a marker of Parkinson’s disease progression. Mov Disord 24: 224–230
- 24. Hughes AJ, Daniel SE, Kilford L, Lees AJ (1992) Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatr 55: 181–184.
- 25.
Becker C, Leistner K, Nikolaus T (1998) Introducing a statutory insurance system for long-term care (Pflegeversicherung) in Germany. Michel JP, Rubenstein LZ, Vellas BJ, Albarede JL: Geriatric Programs and departments around the world Serdi-Springer, Paris-New York: 55–64.
- 26. Icks A, Haastert B, Wildner M, Becker C, Meyer G (2008) Trend of hip fracture incidence in Germany 1995–2004: a population-based study. Osteoporosis International 19: 1139–1145.
- 27. Icks AIA, Arend WAW, Becker CBC, Rapp KRK, Haastert BHB (2013) Incidence of hip fractures in Germany, 1995–2010. Archives of osteoporosis 8: 1–7.
- 28. Kanis JA, Odén A, McCloskey EV, Johansson H, Wahl DA, et al. (2012) A systematic review of hip fracture incidence and probability of fracture worldwide. Osteoporosis International 23: 2239–2256.
- 29. Latt MD, Lord SR, Morris JGL, Fung VSC (2009) Clinical and physiological assessments for elucidating falls risk in Parkinson’s disease. Mov Disord 24: 1280–1289
- 30. Plotnik M, Giladi N, Dagan Y, Hausdorff JM (2011) Postural instability and fall risk in Parkinson’s disease: impaired dual tasking, pacing, and bilateral coordination of gait during the “ON” medication state. Exp Brain Res 210: 529–538
- 31. Wood BH, Bilclough JA, Bowron A, Walker RW (2002) Incidence and prediction of falls in Parkinson’s disease: a prospective multidisciplinary study. J Neurol Neurosurg Psychiatr 72: 721–725.
- 32. Bhattacharya RK, Dubinsky RM, Lai SM, Dubinsky H (2012) Is there an increased risk of hip fracture in Parkinson’s disease? A nationwide inpatient sample. Mov Disord 27: 1440–1443
- 33. Fink HA, Kuskowski MA, Orwoll ES, Cauley JA, Ensrud KE (2005) Association between Parkinson’s disease and low bone density and falls in older men: the osteoporotic fractures in men study. J Am Geriatr Soc 53: 1559–1564
- 34.
Benzinger P, Becker C, Kerse N, Bleibler F, Büchele G, et al.. (2013) Pelvic Fracture Rates in Community-Living People With and Without Disability and in Residents of Nursing Homes. J Am Med Dir Assoc. doi:10.1016/j.jamda.2013.03.012.
- 35. Rapp K, Becker C, Cameron ID, Klenk J, Kleiner A, et al. (2012) Femoral fracture rates in people with and without disability. Age Ageing 41: 653–658
- 36. Van de Vijver DA, Stricker BH, Breteler MM, Roos RA, Porsius AJ, et al. (2001) Evaluation of antiparkinsonian drugs in pharmacy records as a marker for Parkinson’s disease. Pharm World Sci 23: 148–152.
- 37. Trenkwalder C, Hening WA, Montagna P, Oertel WH, Allen RP, et al. (2008) Treatment of restless legs syndrome: an evidence-based review and implications for clinical practice. Mov Disord 23: 2267–2302
- 38. Högl B, Kiechl S, Willeit J, Saletu M, Frauscher B, et al. (2005) Restless legs syndrome: a community-based study of prevalence, severity, and risk factors. Neurology 64: 1920–1924
- 39. Hening W, Walters AS, Allen RP, Montplaisir J, Myers A, et al. (2004) Impact, diagnosis and treatment of restless legs syndrome (RLS) in a primary care population: the REST (RLS epidemiology, symptoms, and treatment) primary care study. Sleep Med 5: 237–246
- 40. Cirillo DJ, Wallace RB (2012) Restless legs syndrome and functional limitations among American elders in the Health and Retirement Study. BMC Geriatr 12: 39