Abstract
INTRODUCTION
Parkinson’s disease (PD) is the second most common neurodegenerative disease affecting an estimated one million individuals in the United States (US) and five million individuals globally [1]. With increasing age, the prevalence of PD also increases. The prevalence of PD is estimated to be 1% among individuals 60 years or older [1] and it increases to 4% amongindividuals 80 or older [2] and it is projected that there will be a two-fold increase in the number of individuals with PD by 2030 [3]. Because PD prevalence is projected to increase, healthcare management of elderly individuals should be studied in an effort to achievebetter health outcomes and to decrease the steeply rising cost curve.
Individuals with PD suffer from several comorbid conditions; psychoses co-occurs in 20%–60% of this with PD [4–6]. Psychoses in those with PD leads to a wide array of negative health outcomes such as increased caregiver burden, difficulty managing PD symptoms, functional impairment and possible nursing home placement [6, 7]. Despite the chances of PD symptom deterioration and increased risk of mortality, antipsychotic agents are widely used to manage psychoses among PD patients [8]. Although efficacy data are limited, second generation antipsychotics (SGAs) are preferred to first generation antipsychotics for treatment of psychoses in patients with PD due to their better adverse effect profiles [9]. According to the American Academy of Neurology (AAN), clozapine and quetiapine are recommended SGAs for treating psychoses among PD patients as these two agents have a lower potential of causing or exacerbating existing movement disorders [10]. Despite the potential risks associated with antipsychotic use in PD patients, a study by Zarowitz et al. (2011), found that a large segment of skilled nursing facility residents with PD who did not have documented psychoses were taking antipsychotic drugs [11].
In addition to psychoses, elderly individuals with PD often suffer from other conditions such as anxiety, depression, and difficulty sleeping. Hence, anti-anxiety, antidepressants and sedative/hypnotics medications are commonly used to manage these comorbid conditions [12]. Recent epidemiological studies have demonstrated that psychotropic polypharmacy is associated with increased mortality risk among elderly individuals with PD [12]. Hence, it is critical to understand concurrent psychotropic polypharmacy use among elderly individuals with PD so that these risks can be avoided. However, to the best of our knowledge, no study has been conducted in the US, that focuses on the psychotropic polypharmacy among elderly individuals with PD. Therefore, the primary objectives of this study were to examine the prevalence, patterns and predictors of psychotropic polypharmacy among elderly individuals with PD.
METHODS
Conceptual framework
The Andersen behavioral model (ABM) was used as the basic conceptual framework for this study [13]. The ABM suggests that an individual’s medication use is a function of predisposing, enabling and need factors.
Data sources
Two nationally representative long-term care surveys [2004 National Nursing Home Survey (NNHS) and 2007 National Home and Hospice Care Survey (NHHCS)] were used for this study. The focus included nursing home and home health settings since patients with PD predominantly reside in these long-term care settings. It should be kept in mind that neither the 2004 NNHS nor the 2007 NHHCS captures elderly individuals with PD at home who do not receive home-based services
National nursing home survey 2004
The National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC) conducted the 2004 NNHS, which is the latest publicly available data set. National estimates of nursing home services provided, as well as staff and resident information can be obtained from the NNHS. A two-stage stratified probability design was used for the 2004 NNHS. The first stage involved facility selection while the second stage involved selection of residents. The 2004 NNHS consisted of 13,507 residents who resided in 1,174 facilities, which translated into a 78% response rate. The 2004 NNHS consists of three file types: (1) facility data; (2) resident data; and (3) prescription data. For this study, the resident and prescription data files were used. Information on demographic characteristics, diagnoses, payment source, functional status and services were gathered from the resident file. The resident file contains up to 34 diagnoses per resident, including: two primary diagnosis codes from nursing home admission; two current primary diagnosis codes; and up to 30 secondary diagnosis codes. Information on up to 40 medications (up to 25 medications taken within the previous 24 hours of the interview and up to 15 medications taken on a regular basis during the month preceding the interview, but not in the past 24 hours) can be obtained from the prescription data file. Psychotropic drugs were identified by the generic drug codes [14].
National home and hospice care survey 2007
The 2007 NHHCS is a nationally representative survey of the US home health and hospice agencies that includes: a patient file; an agency file; and a medication use file. For the purposes of this study, the patient and the medication use files were used. The patient file contains information on: demographic characteristics; functional status; diagnoses; and services provided to patients. The patient file consists of two primary diagnosis codes (one primary admission code and one current primary diagnosis code), and up to 16 secondary diagnosis codes. The medication use file provides information on use of up to 25 medications per patient. Psychotropic medications were identified using Multum drug codes. A stratified two-stage probability sample design was used for the 2007 NHHCS. The first stage consisted of selection of a representative sample US home health and hospice agencies. In the second stage, patients were selected during agency interviews. For this study, the home health component of the NHHCS, which contained information for 4,683 home health patients, wasused [15].
Study sample and analysis
This study used a cross-sectional design. The sample included from each dataset consisted of elderly (age ≥65 years) individuals with PD [identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code of 332.xx]. Psychotropic drugs included four drug categories: antipsychotics, antidepressants, anxiolytics, and sedatives/hypnotics [16]. Inter-class psychotropic polypharmacy was defined as the concurrent use of two or more medications from different psychotropic classes, whereas concurrent use of two or more medications within the same psychotropic medication class was defined as intra-class psychotropic polypharmacy. A list of drugs included in each category is provided in Supplementary Table 1. Comorbidities such as depression, delirium, dementia, psychoses etc. were identified by ICD-9-CM codes (list provided in Supplementary Table 2). Descriptive statistics of selected characteristics and medication use are presented as frequencies. Multivariable logistic regression was used to assess predictors of psychotropic polypharmacy among elderly individuals with PD. National estimates were obtained by using the sampling weight provided for each dataset. All analyses were adjusted for the complex survey design of NNHS and NHHCS. The Institutional Review Board of the University of Arizona approved this study as Non-Human Subject Research(Exempt status).
RESULTS
According to the 2004 NNHS and 2007 NHHCS, there were 93,648 and 37,439 elderly individuals with PD residing in nursing home and home healthcare settings respectively in the US. In both groups, the majority were white, non-Hispanic, older than 75, and had public (Medicare/Medicaid/Veterans Affairs) insurance. However, the majority of nursing home residents with PD were females (63.05%) while those residing in home healthcare settings were predominantly males (60.07%). Most of the elderly individuals with PD who resided in the US nursing homes had impaired decision making ability (60.48%) and approximately one in four (24.37%) had some form of behavioral symptoms, while only 10.62% had a psychoses. Additional detail is presented in Table 1.
Among elderly nursing home residents with PD, the nationally representative prevalence of psychotropic polypharmacy was 26.68% (95% CI, 22.81%–29.75%), whereas, it was 21.36% (95% CI, 8.12%–34.6%) in home health. Use of antidepressant medications constituted the majority of the psychotropic medication use among both nursing home (48.91%, 95% CI, 44.9%–52.8%) and home health (40.98%, 95% CI, 25.37%–56.61%) residents with PD. Nearly one-third (31.27%, 95% CI, 27.6%–34.9%) nursing home residents with PD were prescribed antipsychotic medications among whom less than one-fourth had a documented diagnosis of psychotic symptoms. Selective serotonin reuptake inhibitors (SSRIs) were the most prescribed antidepressant drugs in both nursing home (66.07% of antidepressant use) and home healthcare settings (67.85% of antidepressant use). Serotonin-norepinephrine reuptake inhibitors (SNRIs) constituted 7.68% and 2.83% of the antidepressant use among nursing home and home healthcare residents respectively. Mirtazapine, which is a noradrenergic and specific serotonergic antidepressant (belonging to the miscellaneous category) was the most widely used antidepressant in nursing home settings (10.08% of elderly PD nursing home residents). The overwhelming majority of elderly individuals with PD who were on antipsychotics in both nursing home and home healthcare settings were prescribed SGAs (95.93% and 96.36% of antipsychotic use respectively). The majority of anti-anxiety drugs prescribed for elderly individuals with PD in nursing homes were short acting benzodiazepines (51.92% of anti-anxiety drug use), whereas in home healthcare settings it was long acting benzodiazepines (79.98% of anti-anxiety drug use). The prevalence of use of sedative/hypnotics among elderly nursing home residents with PD was 2.84% (95% CI, 1.60%–4.35%), while in the home healthcare setting was 22.1% (95% CI, 8.23%–35.97%). There was a marked intra-class antidepressant polypharmacy among elderly individuals with PD in nursing home (4.74%, 95% CI, 3.18%–6.33%) and home healthcare settings (8.69%, 95% CI, 0.00%–18.69%), no other psychotropic drug class showed intra-class polypharmacy. Table 2 provides additional detail.
Among elderly nursing home residents with PD, 88.15% were prescribed one or more antiparkinsonian medications, with carbidopa-levodopa (85.62% of elderly nursing home residents with PD) being the most commonly used antiparkinsonian medication followed by carbidopa (4.5% of elderly nursing home residents with PD) and levodopa. Among elderly individuals with PD residing in the home healthcare settings, only 56.55% received one or more antiparkinsonian medications [majority (46.37% of elderly home healthcare residents with PD) being prescribed carbidopa-levodopa] (This data is not presented in tabular form.).
Multiple logistic regression analyses that adjusted for the complex survey design revealed that only need characteristics, such as specific comorbidities were significantly associated with psychotropic polypharmacy among elderly nursing home residents with PD. The likelihood of psychotropic polypharmacy was approximately four times higher [adjusted odds ratio (AOR): 3.705; 95% CI, 1.781 –7.705] among elderly nursing home residents with psychoses compared to those without a psychoses diagnosis. Similarly, presence of schizophrenia (AOR: 3.166; 95% CI, 1.092 –9.184), depression (AOR: 2.551; 95% CI, 1.526 –4.263) and anxiety (AOR: 3.474; 95% CI, 1.967 –6.135) were positively associated with the chances of having psychotropic polypharmacy. Multivariate analyses for identifying predictors of psychotropic polypharmacy was not conducted due to small sample size in home healthcare settings. Additional data are provided in Table 3.
DISCUSSION
The most common comorbidities observed in both the nursing home and home health care settings were depression, delirium, and dementia with the vast majority of patients being over the age of 75 years. This is consistent with other studies which estimate the prevalence of dementia and depression in patients with PD to be 24–31% and 7–35% respectively [17, 18]. The overall prevalence of serious mental illnesses such as schizophrenia and bipolar disorder is less than that found for depression. Additionally life expectancy for patients with serious mental illness, such as schizophrenia, is lower than for the general population, resulting in fewer patients found with these psychiatric disorders in older age [19, 20]. Therefore these findings were not surprising.
Psychotropic polypharmacy was positively associated with specific comorbidities, notably schizophrenia, depression, and anxiety. This may reflect a symptom presentation that requires multiple agents for control of the disease. Monotherapy is generally recommended in patients with new-onset psychiatric disorders but most guidelines recommend augmentation and combination therapy in patients who have failed adequate trials of monotherapy or meet the criteria as having a treatment-resistant psychiatric disorder. Polypharmacy with either antidepressants or benzodiazepines, both commonly used medication classes for the treatment of depressive and anxiety disorders, has been associated with an increased number of falls in older adults [21]. As anxiety disorders are common in patients with PD with an estimated 49% lifetime prevalence, it is important to minimize polypharmacy with agents from multiple classes (e.g. benzodiazepines) if either a SSRI or SNRI, both first-line treatments, can be used as monotherapy instead [22]. The risk versus the benefit must always be weighed when considering the use of multiple psychotropic agents in patients with PD even when the combination is supported by evidence-based guidelines.
A diagnosis of psychoses for patients in nursing homes resulted in a greater likelihood of those patients being a recipient of psychotropic polypharmacy. As an episode of psychosis is not limited to persons with a diagnosis of a schizophrenia spectrum or other psychotic disorder but can occur with other diagnoses such as depression, two medications may be needed. For example, an antidepressant to treat the mood disorder and an antipsychotic to target the psychosis may be necessary. Treatment for neurocognitive disorders (also known as dementias) with behavioral disturbance includes consideration of citalopram based upon the findings of The Citalopram for Agitation in Alzheimer Disease study [23]. An antipsychotic may later be added to the patient’s medication regimen if his/her behavioral disturbances become an imminent danger to the patients and/or others, resulting in the use of two concomitant psychotropic medications.
A number of medications used to treat PD including carbidopa-levodopa, COMT inhibitors, and dopamine agonists can cause neuropsychiatric symptoms such as agitation and hallucinations. Anticholinergic medications may also raise the risk of developing psychosis [24]. Of the subjects in this study, 56%–88% were prescribed one or more antiparkinsonian medications with 46%–86% prescribed carbidopa-levodopa, a medication with a known adverse effect profile of hallucinations, psychosis, and agitation. It is possible that some of the subjects in this study were prescribed an antipsychotic medication due to antiparkinsonian medication-induced psychosis. If these patients also had another psychiatric comorbidity such as depression, this could account for the use of multiple psychotropic agents.
Nearly one-third of nursing home residents with PD were prescribed antipsychotic medications among whom less than one-fourth had a documented diagnosis of psychotic symptoms. This finding is consistent with an earlier study in which veterans with dementia but without documentation of any psychotic symptoms were just as likely to receive antipsychotics as those patients who did have a diagnosis of psychotic symptoms [25]. Patients with PD, especially those with dementia or medication-induced psychosis, may display aggressive or agitated behavior, a common reason for antipsychotic use. Although the Food and Drug Administration issued a warning regarding the risk of increased mortality with the use of antipsychotics in patients with dementia, short-term use of antipsychotics may be considered in cases when the benefit outweighs the risk (e.g. danger to self and others) after non-pharmacological techniques have been exhausted [26]. Antipsychotics are also used to treat bipolar disorder and some SGAs are indicated as adjunctive therapy for patients with major depressive disorder regardless of the presence or absence of psychosis. This may partially explain antipsychotics prescribed to PD patients without a documented diagnosis of psychotic symptoms. However, use of antipsychotics in nursing homes for non-approved indications for which evidence is lacking continues to be of concern and needs to be further addressed [27].
Antidepressants were the most commonly prescribed psychotropic medication class for patients with PD in this study. This finding is not surprising as antidepressants are used for a multitude of FDA approved and off-label uses including depressive disorders, anxiety disorders, post-traumatic stress disorder, obsessive-compulsive disorder, bipolar disorder (in conjunction with a SGA), insomnia, fibromyalgia, neuropathy, chronic musculoskeletal pain, stress incontinence, migraine prophylaxis, and vasomotor symptoms related to menopause. SSRIs, the most prescribed antidepressant class for this population, are considered first-line treatment for older patients due to tolerability and overall safety compared with some of the other antidepressant classes such as tricyclic antidepressants. Mirtazapine was the most widely used antidepressant in nursing home settings. Older individuals with PD may experience symptoms of depression such as insomnia and loss of appetite. Due to its side effect profile that includes both drowsiness and weight gain, mirtazapine is often chosen to treat depression in patients who experience these particular symptoms.
There was marked intra-class antidepressant polypharmacy found in both of the study populations. Polypharmacy with trazodone specifically accounted for 4.74% of all intra-class antidepressant pharmacy. Trazodone is most commonly used as an augmenting agent to enhance the actions of other antidepressant medications rather than used as monotherapy. As trazodone has high rates of sedation but without the accompanying weight gain observed with mirtazapine, it is considered a treatment of choice for insomnia in patients with depression [28].
There was a difference found in the type of benzodiazepine prescribed for elderly patients with PD depending upon their setting. In the nursing home setting, short-acting benzodiazepines were most prescribed whereas in home healthcare settings, long-acting benzodiazepines accounted for the vast majority of anxiolytics. This may reflect either a variation in practice or an environment which differs between the two settings. For example, it may be more likely that patients in home healthcare settings still share a bed with a partner. As sleep disorders such as rapid eye movement behavior disorder occur in 37.5% of patients with PD, clonazepam may be used to treat the sleep disorder which otherwise may be disruptive for the partner [29]. Likewise, patients requiring a nursing home setting may experience more agitation which is usually treated with a shorter acting agent such as lorazepam. However, this may still be an opportunity for intervention if the patient is experiencing adverse effects from benzodiazepine use. In general, short-acting benzodiazepines are preferred if a benzodiazepine is indicated, when possible, due to changes in metabolism that result in a longer half-life for these medications in elderly patients. Based upon the results found in this study, it is possible that less than ideal benzodiazepine prescribing may be occurring in settings in which the patient is still living at home. Cognitive impairment and/or falls resulting from these agents may result in the patient prematurely requiring a supervised living setting [30].
Sedative/hypnotic use was also found to be markedly different between the two settings with much higher prevalence of use of sedative/hypnotics in the home health setting. Two of the most commonly used medications found in this class include temazepam and zolpidem. This is especially concerning considering that 20% of all emergency room visits resulting from a psychotropic medication adverse event in patients ages 65 and older involve zolpidem [31]. Although concomitant anti-anxiety and sedative/hypnotic use was very low, the use of certain medications from either of these classes, even as monotherapy, may still pose a risk to this patient population.
There are several limitations that should be noted. First, the majority of patients included in this sample were white, non-Hispanic, ages 75 years and older, and had public insurance. Therefore, disparities in healthcare for minority populations may not have been reflected in study. As this study just looked at patients with PD in nursing homes or who were receiving home health services, the extent of psychotropic polypharmacy in other populations with PD (e.g. family caregiver) remains unknown. Moreover, there has been some conflicting evidence in the existing literature regarding the use of ICD-9-CM code of 332.0 to identify PD, as it is also used to capture several different conditions which includes atypical parkinsonism, drug-induced parkinsonism and idiopathic PD. A study by Swarztrauber et al. (2005) concluded that the ICD-9-CM code of 332.0, which is usually used identify PD, was not able to differentiate between parkinsonism and PD [32]. Hence, another limitation of this study is that the use of the 332.xx code may not be able to distinguish between PD and other forms of parkinsonism. Multivariate analyses for identifying predictors of psychotropic polypharmacy was not conducted due to the small sample size in home healthcare settings. Finally, due to the retrospective nature of this study, only data which was included in the original database could be analyzed.
CONCLUSIONS
More than one-in-four nursing home and one-in-five home healthcare elderly residents with PD in the US had a documented use of more than one type of concurrent psychotropic medication, with antidepressants being the most used psychotropic medication in both the settings. These findings underscore the importance of evidence-based prescribing when psychotropic medications are used in elderly individuals with PD to reduce unnecessary psychotropic polypharmacy.
CONFLICTS OF INTEREST
None.
Footnotes
ACKNOWLEDGMENTS
Funded by a grant from the University of Arizona Health Sciences (UAHS) Center for Biomedical Informatics and Biostatistics (CB2) at The University of Arizona. The content is solely the responsibility of the authors and does not necessarily represent the official views of Arizona CB2 core.