Abstract
INTRODUCTION
Hypertension contributes to both early cerebrovascular brain aging and cognitive decline [1]. It is a modifiable risk factor for both Alzheimer’s disease (AD) and vascular dementia [2]. The prevalence of hypertension among people with dementia ranges from 35% to 84% [3–5]. While antihypertensive prescribing rates have increased significantly over recent years in the general older population with hypertension [6, 7], such trends have not specifically been investigated in those with dementia.
Previous research highlights discrepancies in the detection and pharmacological treatment of hypertension in people with dementia. European studies have found that community-dwelling older people with dementia are prescribed fewer evidence-based cardiovascular disease (CVD) medications such as beta-blockers and angiotensin-converting enzyme (ACE) inhibitors than people without dementia [8, 9]. Conversely, other studies have found no differences in hypertension treatment [10] or greater use of antihypertensives in people with dementia [11]. Disparities in antihypertensive treatment may also relate to socioeconomic, racial, and ethnic differences within people with dementia [12].
Appropriately managing hypertension in people with dementia is important because clinical and observational studies have reported that antihypertensive therapy can slow the progression of cognitive and functional decline [13–15]. Conversely, strict blood pressure control and overtreatment with antihypertensives may worsen cognition in people with dementia and cognitive impairment [16]. Among people with cognitive impairment aged 65 years and older treated with antihypertensive medications, low daytime systolic blood pressure (SBP) (≤128 mm Hg) was independently associated with a greater progression of cognitive decline over a median 9-month follow-up [16].
No previous national population studies have investigated trends in antihypertensive use in people with dementia. This is an important gap because people with dementia represent a special population in which the benefit to risk ratio of antihypertensive treatment is likely to differ to that of the general population [17]. Additionally, this population are at an increased risk of functional decline and institutionalization [18].
The objectives of this study were to (i) investigate national trends in antihypertensive medication prescribing in people with dementia between 2006 and 2012 in the United States, and (ii) to investigate clinical and demographic factors associated with different antihypertensive prescribing patterns in people with dementia over this period.
METHODS
Data source
Data from the 2006–2012 National Ambulatory Medical Care Survey (NAMCS) and the outpatient department component of the National Hospital Ambulatory Medical Care Survey (NHAMCS) were combined to analyze physician visits by people with dementia aged ≥65 years over a seven-year period. Data were analyzed from 2006 after the introduction of Medicare Part D so that any changes in medication use following Part D would not influence our results. At the time of this study, the 2012 NHAMCS dataset was unavailable and therefore excluded from analysis.
The NAMCS and NHAMCS are annual national probability sample surveys conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention. Detailed descriptions of study methodologies have been presented previously [19]. The NAMCS captures visits to nonfederal office-based physicians, and the NHAMCS samples from hospital-based outpatient departments and emergency departments. The physician visit is the unit of observation, and data are weighted to produce unbiased national estimates. Survey data are collected annually using a multistage probability design. Data collected on the patient encounter forms include patient characteristics, physician practice information, diagnoses (up to 3) and medications (up to 8). Data collection was performed by physicians, office staff or census field representatives. All medical and medication coding and keying operations were performed centrally by SRA International, Inc. Durham, North Carolina and were subject to quality control procedures. These validated data have been used previously to investigate antihypertensive prescribing [7, 21] and medication use in people with dementia [22].
Study sample
People with dementia were identified as those with a documented diagnosis of dementia on the encounter form or those receiving a prescription for an anti-dementia medication. Dementia was defined using the following International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes: 290.XX, 291.2, 292.82, 294.1, 294.10, 294.11, 331.19, 331.82, 331.0, and 331.1X [22]. Anti-dementia medications included memantine, donepezil, rivastigmine, or galantamine.
Antihypertensives
Medications included those that were ordered, supplied, administered, or continued during the visit. This included medications intended for regular or as-needed use. Medications were coded according to the Multum Lexicon Drug Database (Denver, Colorado). Antihypertensives were identified according to generic or brand-name medications documented on the encounter form. Antihypertensives were defined as diuretics (049), beta-blockers (047), calcium-channel blockers (048), ACE inhibitors (042), angiotensin II receptor blockers (ARBs) (056), and others (antiadrenergics [043, 044], vasodilators [053], renin inhibitors [342]). Antihypertensives prescribed for any indication were included. For fixed-dose combination antihypertensive products, individual ingredients were counted separately.
Explanatory variables: Physician visit characteristics
Demographic, socioeconomic, clinical, and prescriber factors were explored. These were patient age, sex, race (white, black, other), source of payment (private insurance, public insurance [Medicare, Medicaid, workers’ compensation], and other [other insurance, selfpay]), vascular comorbidities documented on the encounter form (hypertension, cerebrovascular disease, congestive heart failure, diabetes, ischemic heart disease, chronic renal failure), geographical region, and metropolitan statistical area.
Statistical analyses
Descriptive analyses were used to evaluate sample characteristics over the study period. For temporal trends in diagnoses, two separate complex samples multivariate logistic regression models were used to assess linear trends across time by estimating the p-value of the coefficient for year of visit as a continuous explanatory variable for (i) dementia diagnosis and (ii) hypertension in those with dementia. For temporal trends in antihypertensive prescribing, three separate complex samples multivariate logistic regression models using year as a categorical explanatory variable were used to explore trends for the prescription of (i) any antihypertensive, (ii) multiple (≥2) antihypertensives, and (iii) each antihypertensive class. Separate complex samples multivariate logistic regression models were then used to estimate unadjusted and adjusted odds ratios (AORs) and 95% confidence intervals (CIs) for factors associated with the prescribing of (i) any antihypertensive, (ii) ≥2 antihypertensives, and (iii) each antihypertensive class in those with dementia. According to the NAMCS/NHAMCS analytical guidelines, estimates based on less than 30 sample cases or with more than a 30% relative standard error may be unreliable. Hence, ACE inhibitors and ARBs were combined into one category to increase the reliability of estimates. All multivariate models were adjusted for demographic and vascular comorbidity variables. All analyses were performed using IBM SPSS Statistics for Windows, Version 22.0. (Armonk, NY: IBM Corp).
Ethics approval
Institutional Review Board exemption was granted by the Harvard Pilgrim Health Care Human Studies Committee, Boston, MA, USA and the Monash University Human Research Ethics Committee, Melbourne, Australia, for this study.
RESULTS
Demographics
The weighted number of physician visits by people 65 years and older for the period from 2006 to 2012 was 1.89 billion. Of these, 42 million (2.2%, 95% CI 2.0–2.5%) were by people with a documented diagnosis of dementia. The characteristics of these office visits are summarized in Table 1. The majority of physician visits made by people with dementia were among those who were female (62.6%, 95% CI 59.7–65.4%), white (88.4%, 95% CI 85.5–0.8%), resided in the Southern regions of the United States (40.3%, 95% CI 34.2–46.6%) and in a metropolitan statistical area (87.3%, 95% CI 80.1–92.1%).
Hypertension was the most common vascular comorbidity and was documented in 58.0% (95% CI 55.1–60.8%) of physician visits for people with dementia. Overall, 60.3% (95% CI 57.1–63.4%) of physician visits by people with dementia involved the prescription of at least one antihypertensive medication. The most commonly prescribed antihypertensives were ACE inhibitors/ARBs (31.4%, 95% CI 28.1–34.9%), beta-blockers (23.3%, 95% CI 20.9–25.9%), and diuretics (20.3%, 95% CI 17.7–23.1%). Overall, 29.7% (95% CI 27.3% –32.4%) of physician visits involved the prescription of two or more antihypertensive medication classes.
Temporal trends in dementia and hypertension
The overall number of physician visits among people with a diagnosis of dementia increased from 4.9 million (95% CI 3.7–6.3 million; 2.0%) in 2006 to 5.7 million (95% CI 4.9–6.6 million; 2.3%) in 2012 (p = 0.08 for linear trend). In people with dementia, the number of physician visits with a recorded hypertension diagnosis increased from 49.3% (95% CI 41.3% –57.4%) in 2006 to 70.5% (95% CI 63.4% –76.7%) in 2011 before declining to 55.7% (95% CI 50.2% –61.2%) in 2012 (p < 0.01 for linear trend).
Temporal trends in antihypertensive prescribing in people with dementia
Figure 1 and Supplementary Table 1 present the temporal trends in antihypertensive prescribing. The proportion of physician visits involving prescription of antihypertensives increased non-significantly from 57.3% (95% CI 48.7% –65.9%) in 2006 to 58.2% (95% CI 52.1% –63.9%) in 2012 (p = 0.80). There were no statistically significant differences in trends of use of major antihypertensive classes over the study period, except for calcium channel blockers where there was a statistically significant decrease from 20.5% (95% CI 14.4% –28.4%) to 12.5% (95% CI 9.6% –16.2%) in 2007 (p = 0.02) and 10.9% (95% CI 7.2–16.1%) in 2009 (p = 0.01), followed by a slight increase from 2010 to 2012 (17.9% [95% CI 14.0% –22.6%] in 2012). The use of other antihypertensive classes increased significantly from 6.2% (95% CI 3.1% –11.0%) in 2006 to 13.3% (95% CI 9.4% –18.6%) in 2012 (p = 0.03).
Physician visits involving the prescription of multiple antihypertensive medication classes increased non-significantly from 2006 to 2012 in people with dementia, with the proportion of those using two or more antihypertensive medication classes increasing from 11.3% (95% CI 7.8% –16.3%) in 2006 to 15.1% (95% CI 12.1% –18.6%) in 2012 (p = 0.29). The proportion of physician visits involving prescription of three or more classes increased from 8.2% (95% CI 4.4% –14.5%) in 2006 to 12.5% (9.0% –17.1%) (p = 0.37) in 2012 (Fig. 2 and Supplementary Table 2). Conversely, the use of fixed-dose combination products decreased from 11.5% (95% CI 6.5% –19.5%) in 2008 to 5.3% (95% CI 3.3% –8.2%) in 2012 (p = 0.52).
Factors associated with prescribing antihypertensives in people with dementia
Tables 2–4 present the results of the multivariate analyses for factors associated with the prescription of (i) any antihypertensive, (ii) ≥2 antihypertensives, and (iii) each antihypertensive class, respectively.
Any antihypertensive
Males were more likely to use antihypertensives than females (AOR 1.37, 95% CI 1.02–1.84). Hypertension diagnosis was significantly associated with antihypertensive use (AOR 4.96, 95% CI 3.58–6.88). Congestive heart failure was significantly associated with the use of antihypertensives (AOR 4.10, 95% CI 2.2–7.61).
Multiple antihypertensives
Male sex was associated with multiple antihypertensive class use (AOR 1.52, 95% CI 1.14–2.04), as was black race (AOR 2.04, 95% CI 1.12–3.71, p = 0.02) and Midwest residence (AOR 2.03, 95% CI 1.46–2.82). Clinical diagnoses associated with multiple antihypertensive use included hypertension (AOR 3.62, 95% CI 2.69–4.87), congestive heart failure (AOR 2.42, 95% CI 1.26–4.63), and chronic kidney disease (AOR 2.53, 95% CI 1.60–3.98). Diabetes was inversely associated with multiple antihypertensive use (AOR 0.65, 95% CI 0.47–0.90).
Antihypertensive classes
Hypertension diagnosis was significantly associated with antihypertensive use across all classes and regimens. Diuretic use was more likely to occur in blacks (AOR 1.89, 95% CI 1.17–3.06, p < 0.01), individuals residing in the Midwest (AOR 1.98, 95% CI 1.16–3.38) and those with congestive heart failure (AOR 2.49, 95% CI 1.57–3.93). Self-pay or use of other forms of payment were inversely associated with diuretic use (AOR 0.27, 95% CI 0.10–0.70). Black race (AOR 2.52, 95% CI 1.40–4.55) and chronic renal failure (AOR 2.53, 95% CI 1.60–3.98) were associated with calcium channel blocker use. Public payment (AOR 1.81, 95% CI 1.08–3.01), cerebrovascular disease (AOR 1.54, 95% CI 1.06–2.23), ischemic heart disease (AOR 2.14, 95% CI 1.43–3.22), and congestive heart failure (AOR 3.19, 95% CI 1.79–5.67) were associated with beta-blocker use. Males were more likely to use ACE-inhibitors/ARBs (AOR 1.46, 95% CI 1.04–2.05) compared to females. Males (AOR 3.62, 95% CI 2.31–5.67), blacks (AOR 2.20, 95% CI 1.14–4.24), and those with chronic renal failure (AOR 2.76, 95% CI 1.56–4.91) were more likely to use other antihypertensive classes. There was no significant association between age and any of the antihypertensive outcomes.
DISCUSSION
Our study found a statistically significant increase in documented hypertension in physician visits by older people with dementia from 2006 to 2012. There were non-significant increases in the proportion of visits that involved prescribing of antihypertensives and multiple antihypertensive medication classes over the study period. Male sex, black race, private health insurance, and Midwest region were associated with prescription of certain antihypertensives and regimens in people with dementia after controlling for hypertension and vascular comorbidities.
The increase in the proportion of visits with a recorded diagnosis of hypertension among people with dementia may indicate increasing recognition and documentation of hypertension in this vulnerable group of individuals. Our study found a non-significant increase in overall antihypertensive use and the use of multiple antihypertensive classes over the study period in people with dementia. Trends remained non-significant across classes, except for some changes in the use of calcium channel-blockers and “other” antihypertensives. Conversely, other studies have found significant upward trends in use of antihypertensives in the general older population with hypertension [6, 7]. Gu et al. [7] report that the proportion of hypertension visits with antihypertensive prescriptions increased significantly from 2003 to 2010, with significant increases in beta-blocker and ARB use seen in those patients aged 60 years and older. These findings highlight that the trends in antihypertensive use in people with dementia may differ to the general older population. It may indicate that the overall prescribing trends of antihypertensives in those with dementia do not necessarily coincide with the increased diagnosis of hypertension seen over the same study period. However, given that Gu et al. analyzed data from a slightly early time period and focused on hypertension visits only, these studies may not be directly comparable.
Our study found a significant increase in the use of “other” antihypertensive classes, often used as last line therapy when other first line agents are ineffective. Increasing use of combination therapy and therapies reserved as last line for refractory hypertension may indicate appropriate management in the presence of severe disease. However, older people with dementia may be susceptible to adverse drug events associated with higher treatment intensity. People with dementia are more likely to be frail, and have greater susceptibility to orthostatic hypotension, falls, cerebral hypoperfusion, especially with increasing age [5]. Despite this, we did not find a difference in the intensity of antihypertensives prescribed between younger (65–74 years) and the oldest (≥85 years old) people with dementia. The increasing use of more intense forms of antihypertensive therapy needs to be carefully assessed for potential risks among people with dementia.
The most commonly prescribed antihypertensive class in physician visits were ACE inhibitors/ARBs, beta-blockers and diuretics. The prescription of these medications is consistent with the previous 7th Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) antihypertensive prescribing guidelines in the general older population, first published in 2003 and in place during the study period [23]. However, conflicting results exist in relation to whether or not use of these medications in older people with dementia is optimal. A large, prospective, cohort study reported beta-blocker use was associated with a lower risk of cognitive impairment compared to non-use of antihypertensive medications in older Japanese American men who were free of cognitive impairment at baseline [24]. However, the protective effects of beta-blockers have not been confirmed in other studies [25]. Investigations into the cognitive effects of beta-blocker use in people with established dementia is limited. Preliminary evidence from a single prospective cohort study found beta-blockers were associated with a 40% decrease in the rate of functional decline in people with AD, while diuretics were associated with a 59% increase in the rate of functional decline and other antihypertensives were found to have no significant effect [14]. The potential for CNS active beta-blockers to impair memory in people with pre-existing cognitive impairment has also been demonstrated in a cross-sectional pilot study [26]. However, these studies are limited by small sample sizes. In contrast, there is considerable evidence to support the use of beta-blocker therapy in reducing morbidity and mortality in people with certain cardiovascular diseases, especially heart failure [27]. Clinicians should carefully weigh the cardiovascular benefits of beta-blocker use with the potential for worsening cognitive impairment, and periodically screen patients taking these medications for signs of cognitive effects. The greater use of beta-blockers and diuretics may relate to their lower cost and physicians having greater clinical familiarity with them compared to other antihypertensive classes. This is reflected in the finding that beta-blockers were more likely to be used by patients with public health insurance compared to privately insured patients.
ACE inhibitors and ARBs have been shown to delay cognitive decline in people with established dementia, including AD. A four-year observational study reported older adults with AD using ACE inhibitors had a significantly slower rate of cognitive decline than those who did not use any antihypertensive medications or other antihypertensives after adjusting for hypertension (Mini-Mental State Exam [MMSE]: –7.5±0.9 versus –9.7±0.4; p = 0.03) [13]. Clinical trials have also suggested that ARBs can slow cognitive decline in people with AD compared to other classes [15, 28]. Proposed mechanisms include increased regional cerebral blood flow and modulation of serum adipocytokines and glucose homeostasis [15, 28]. The relatively high, collective use of ACE-inhibitors/ARBs in this cohort was thus promising to see.
The greater use of diuretics and calcium channel blockers in black people is consistent with JNC 8 guidelines for antihypertensive prescribing in this population [29]. Blacks were more likely to use multiple antihypertensives and other antihypertensive classes reserved for last line use or refractory hypertension. This is reflective of current evidence that blacks not only have a higher prevalence of hypertension, but also more severe or poorly controlled hypertension than white individuals [30, 31]. However, as this study population was older people with dementia and at a greater likelihood of being frail, further exploration into whether these intense medication regimens represent appropriate blood pressure control or potential overtreatment is needed.
Females were less likely to receive antihypertensives, including ACE inhibitors/ARBs and more intense forms of antihypertensive therapy (multiple and other antihypertensive classes) compared with males after adjusting for hypertension and comorbid vascular disease. While this may indicate that females in the study sample had less severe disease, it may also highlight under-recognition and under-treatment of hypertension and CVD in this group [31]. This is further highlighted by the fact that in people aged 65 years and older in the US, the prevalence of CVD is similar between men and women, with the prevalence of hypertension being higher in women than men [31]. Studies have also shown this is the case in those with dementia [3]. It was unclear as to why there was an inverse association between diabetes and multiple antihypertensive use. However, the use of less intense antihypertensive therapy in people with diabetes compared to those without has been reported in previous studies [32].
This study has investigated trends in antihypertensive prescribing in people with dementia, and the demographic and clinical factors associated with their use. Generating evidence regarding the management of cardiovascular comorbidity in dementia is important. Optimal management of these comorbidities may offer potential improvement to cognition and wellbeing in those with dementia [33]. Early management of comorbidity and prevention of associated complications have the potential to reduce the incidence and disease-progression of dementia, as well as morbidity and mortality. Future work should further explore the therapeutic effects of different antihypertensive classes on people with established dementia. Additionally, research into the optimal clinical targets for people with dementia to ensure appropriate pharmacological management of hypertension should be undertaken.
Our study has several strengths and limitations. Firstly, NAMCS and NHAMCS produce national estimates with minimal biases. However, these estimates were based on physician visits rather than individual patients. This may oversample patients who have more comorbidities or require more frequent follow-ups visits. Additionally, these surveys do not include individuals with dementia who are institutionalized. Underestimation of dementia may have occurred as dementia diagnosis was derived from the encounter form that allowed only a maximum of three recorded diagnoses. As patients were likely to have multiple comorbid conditions, dementia diagnosis may have been underreported on the form. However, we attempted to overcome this limitation by supplementing this information with prescription for anti-dementia medications. Underestimation of antihypertensive medication may also have occurred as information was derived from medications recorded on the encounter form with an upper limit of eight medications recorded. It was not possible to investigate medication adherence, non-pharmacological treatment or lifestyle factors, or clinical appropriateness of medication in this study. As this was an analysis of physician visits, an individual patient could theoretically be included in the analysis more than once, however this is unlikely given the sampling design.
CONCLUSION
In older people with dementia attending outpatient physician visits in the United States, hypertension diagnoses increased significantly from 2006 to 2012, with non-significant increases in overall antihypertensive use and the use of multiple classes. Gender, racial, regional, and socioeconomic factors influence antihypertensive prescribing in this population. As people with dementia are prone to adverse drug events, the risk to benefit ratio of antihypertensive regimens in people with dementia needs to be explored further.
