Abstract
INTRODUCTION
Parkinson’s disease (PD) in institutionalized patients has a profound impact on quality of life and is characterized by a wide range of motor and non-motor signs [1, 2]. Orthostatic hypotension (OH) is a common non-motor trait, defined as a blood pressure drop within three minutes of standing of at least 20 mmHg systolic or at least 10 mmHg diastolic [3]. OH has a prevalence in PD ranging from 30% to 64.9% [4], and can be symptomatic, causing symptoms like generalized weakness, lightheadedness, head and neck pain, vertigo, falls and syncope [5]. However, it can also be asymptomatic, making it clinically less relevant. Patients can describe symptoms reminiscent of OH without measurements showing a blood pressure drop, either because the symptoms are not due to OH or because the blood pressure was measured at a time when OH was not pronounced. The prevalence of symptomatic orthostatic hypotension has been estimated to be about 19% [6, 7], with a large proportion of PD-patients (30%) having asymptomatic OH. Treatment of OH can be pharmacological and non-pharmacological. Suggested pharmacological treatments include midodrine and fludrocortisone [8, 9]. In Dutch guidelines, domperidone is suggested as a treatment for OH [10].
No studies have specifically addressed the prevalence of OH in institutionalized PD-patients, even though the prevalence in these severely affected patients might be particularly high. It is currently unclear how often OH is recognized and treated in long-term care settings. Prior work suggests that misdiagnosis and undertreatment are common among institutionalized PD-patients [2, 11], and this may be especially the case for an elusive symptom such as OH. The aim of this study was to explore the prevalence of OH in PD patients living in Dutch nursing homes, and to determine whether it is recognized and treated adequately.
MATERIALS AND METHODS
Population
The study, described in detail previously [1, 2], was performed in 12 large nursing home organizations in the South of the Netherlands between January and November 2010. Only residents receiving long-term care were included. Eligible residents were identified through medical files of elderly care physicians and registries of pharmacists identifying all nursing home residents who used antiparkinsonian medication (defined according to the Anatomical Therapeutic Chemical (ATC) classification system: N04). All nursing home medical charts were reviewed, and all primary and secondary diagnoses were recorded. All prescribed medication was recorded, and fludrocortison, midodrine and domperidone were specifically registered for analysis. For most residents, the hospital outpatient file was present in the nursing home and was also reviewed. A single researcher (NW) with experience in the field of movement disorders confirmed the diagnosis of PD according to the UK Parkinson’s Disease Brain Bank criteria [12]. Participants scoring less than 18 on the Mini-Mental State Examination (MMSE)—representing moderate to severe cognitive decline, which could hamper obtaining reliable clinimetrics [13]—were excluded. Participants currently or recently treated with (typical) antipsychotic drugs or other offensive drugs, defined by an extensive review [14], were also excluded. An exception was made for patients in whom a diagnosis had been made prior to the start of the offending drug [11].
Measurement
OH was measured according to a standard protocol consisting of three measurements with a routine sphygmomanometer [15]. A first measurement was carried out after at least 10 minutes of supine rest. The second and third measurements were performed following standing up, after 1 and 3 minutes respectively [3]. A nurse or nurse assistant carried out the measurement after instruction by a movement disorder specialist (NW). Patients able to stand, with or without help of caregiver, were included. Patients who could not stand for 3 minutes due to motor problems were excluded. Patients who could not stand for 3 minutes due to OH were included. Symptoms of orthostatic intolerance were determined using two questions of the first domain of the Non-motor Symptom Scale (NMSS): “Lightheadedness, dizziness, weakness on standing”, and “Falling because of fainting /blacking out”. Answers comprise frequency (range 0–4) and severity (range 0–3) assessing the symptom, if present, over the last month. Participants presenting with frequency ≥1 on either of these were determined as having orthostatic symptoms. The severity is scored as mild (causing little distress or disturbance), moderate (causing some distress or disturbance) or severe (causing major distress or disturbance) [16]. We defined thefollowing combinations of OH and symptoms: probably symptomatic OH denotes patients with orthostatic symptoms and OH, possibly symptomatic OH concerns those with orthostatic symptoms but no OH. Probably asymptomatic OH concerned patients with OH but without orthostatic symptoms; finally, no OH describes patients without OH and without symptoms. All current medication was retrieved from the nursing home medical chart. Antihypertensive medications considered were diuretics, B-adrenergic antagonist, calcium antagonist and ACE-inhibitors or angiotensin receptor antagonist. Co-morbid disorders were retrieved from medical files and classified according to the ICD-10 coding system. Cardiovascular disease was considered present if medical files showed cerebrovascular accident (ICD-10: I60–I69, G45), myocardial infarction (ICD-10: I21, I22, I252), heart failure (ICD-10 I50), peripheral vascular disease (ICD-10: I71, I79.0, I73.9, R02, Z95.8, Z95.9) or other cardiovascular diseases (I0–I99). Also hypertension (ICD-10: I10–I15) and diabetes mellitus (ICD-10: E10-E14) were extracted frommedical files.
Statistics
For descriptive statistics SPSS version 20.0.0.1 (2011) was used. Taking antihypertensives, history of hypertension, history of diabetes mellitus and history of cardiovascular disease were analyzed as potential confounding variables in an explorative analysis, using the chi-square test. A critical p-value of 0.05 was applied.
RESULTS
Of 258 patients considered, 152 patients were diagnosed with PD; 73 met the inclusion criteria of MMSE >18. Blood pressure measurements were available for 64 patients, with no measurement (N = 7) and failed third measurement (N = 2) explaining the remainder of the group (Fig. 1). One patient completed the blood pressure measurement but not the NMSS questionnaire. This patient was left out of the further analyses. The mean age of patients was 78.8 years; disease duration was 9.9 years. The majority of patients had advanced disease (85.9% Hoehn & Yahr stage ≥4). The prevalence of OH was 51.6% (Table 1).
OH was recognized in 52.4% of the population; 25.4% had probably symptomatic OH and 27.0% had probably asymptomatic OH. 46.0% of all patients reported orthostatic symptoms. 20.6% had possible symptomatic OH and 27.0% had no OH (Table 2). 6 reported orthostatic symptoms as a major source of distress, 8 as causing some distress and 15 as causing little distress.
Three diagnosis alluding to OH were found in medical files: “collapse”, “prone to collapse” and “autonomic dysfunction”. A diagnosis was present in the medical files of one patient with probably symptomatic OH, one patient with probably asymptomatic OH and one patient with possibly symptomatic OH. Six patients were treated with either domperidone or fludrocortison. None received midodrine. Domperidone was prescribed for two patients with probable symptomatic OH and three patients with possible symptomatic OH. Fludrocortison was prescribed to one patient with possibly symptomatic OH and one patient with no OH (Table 3).
Comorbidity was frequently present: 35 patients had a history of cardiovascular disease, 17 patients hypertension and 11 patients had diabetes mellitus. Another 32 patients were using antihypertensives (Table 3). No statistically significant differences between OH-groups were found on comorbidity and use of antihypertensive medication.
DISCUSSION
In this relatively small sized cohort study the prevalence of OH in institutionalized PD patients was 51.6%. In a systematic review in 2011 prevalence in PD-patients was estimated lower at 30% [4]. As no comparison was made in this study with unmedicated PD-patients, less disabled PD-patients or non-PD nursing home residents no conclusions can be drawn on why prevalence seems higher in our cohort. Several hypothesis can be made based on this study. A potential determinant of OH is PD progression. An earlier study showed that OH is equally prevalent in mild, moderate and severe PD [17]. However, the average HY stage of 4.3 in our cohort is much higher than the HY stage in any cohort published earlier [4, 17]. Other determinants of OH are cardiovascular disease, diabetes mellitus and antihypertensive medication [18]. These factors were shown to be present in >50% of patients and could potentially explain the increased prevalence in this patient group. No differences in frequency of determinants were found between PD-patients with and without OH in an explorative analysis. This could be because our study was underpowered to detect a difference due to the small sample size. More research is warranted to explain the finding of high OH prevalence in institutionalized PD-patients.
About half of patients had probably symptomatic OH. In this paper we defined three degrees of probability linking patient’s symptoms with the outcome of an orthostatic blood pressure measurement. The Consensus statement [3] defines OH as a blood pressure fall, but does not provide guidelines how to link symptoms with this sign. There are several reasons why studying this relation is complex. First, a fall of systolic blood pressure of 20 mm Hg is abnormal, but whether it causes clinically manifest cerebral hypoperfusion probably depends more on the level of the nadir than on the magnitude of the drop. In other words, a systolic drop from 60 to 40 mmHg can be clinically more relevant than a drop of 180 to 80 mm Hg [19]. Second, the manifestation of OH can differ during the course of the day, with symptoms typically presenting most frequently in the morning. Also, OH can worsen due to the variable presence of exacerbating factors like heat, food, alcohol, exercise, activities which increased intrathoracic pressure (e.g. coughing, defecating) and certain drugs [20, 21]. Third, the reliability of self-reported symptoms can be biased as patients have difficulty distinguishing symptoms of OH from other symptoms, e.g. balance problems or benign paroxysmal positional vertigo [22, 23].
Several other methodological issues could hamper the interpretation of our findings. First, an inherent risk of selection bias is associated with cohorts with a small sample size. We tried to minimize the risk of selection with a double search recruitment strategy identifying virtually all eligible residents in the approached nursing homes. Second, PD-patients with an MMSE <18 were excluded from this study. The prevalence of OH is known to be high (50%) in patients with PD and dementia [24], so the prevalence observed here is likely an underestimate. Third, we followed the definition of classical OH, so initial OH (occurring very rapidly after rising, hence missed by the traditional blood pressure measurements) and delayed OH (occurring after prolonged standing, well beyond the second measurement taken after 3 min of standing) may have been missed also resulting in a probable underestimation [25, 26]. Finally, we did not use an Ewing test battery as our measurement routine of autonomic dysfunction. We focused on OH as a clinically most relevant expression of autonomic failure. To do so, we diagnosed OH using a validated and clinically widely used measurement technique, tailored to this particular group of patients who are not capable of visiting an outpatient clinic equipped with a tilt table [6, 27] In addition, we went further than merely measuring a blood pressure drop and wanted to gain insight into the symptoms of OH. The Non-Motor Symptom Scale is suggested for evaluating the presence and severity of OH-related symptoms in a Movement Disorder Task Force-review of dysautonomia rating scales [28]. There is however an inherent risk of informational bias associated with our measurement routine.
Probably symptomatic OH was present in 25.4% of our institutionalized patient population. An additional 20.6% had possibly symptomatic OH. Only three patients had a diagnosis alluding to OH in the medical record, suggesting a marked under-diagnosis of OH. A minority of symptomatic patients received fludrocortison or domperidone, despite sometimes debilitating symptoms upon rising or standing, suggesting undertreatment. As the clinical manifestations of OH can be very burdensome, a more robust approach to address these complaints is needed.
CONFLICTS OF INTEREST
This study was made possible by a research grant of Dutch Parkinson Patient Federation and the Prinses Beatrix Fonds. Drs ALAJ Hommel, Dr MJ Faber and Prof BR Bloem were supported by a JPND grant. The authors have no conflict of interest to report.
Footnotes
ACKNOWLEDGMENTS
We would like to thank Bart van Lange, MD, for helping with the data collection.