Prescribing Practices of Anticholinergic Medications and Their Association With Cognition in an Extended Care Setting

Abstract

The current study examined prescribing patterns of anticholinergic (AC) medications and their association with cognitive function in 450 nondemented and nondelirious older adults hospitalized in a postacute extended care center. Participants completed a brief neuropsychological battery that included measures of general mental status, memory, judgment, and executive functioning as part of standard clinical care. An AC burden score was calculated for each participant based on medications taken the day of the testing using the Anticholinergic Drug Scale. Although use of AC medications was common, the majority of participants were taking medications with only minimal AC properties. AC burden and total number of AC medications were negatively correlated with age. AC burden was not associated with lower performance on any of the cognitive measures. In sum, current prescribing practices of AC medications are not associated with negative cognitive effects in a sample of older adults hospitalized in an extended care center.

Keywords

anticholinergic neuropsychology cognition burden

Introduction

Clinicians face particular challenges when prescribing medications for older adult patients. Age-related factors, such as changes in drug metabolism and elimination, multiple health comorbidities, chronic disease, and deficits in neurotransmitter transmission, result in an increased vulnerability to adverse drug effects (Feinberg, 1993). Medications with a particularly high risk of adverse drug events, often referred to as potentially inappropriate medications (PIMs), are commonly prescribed to older adults. Studies examining PIMs have shown that up to 20% of older adults in outpatient settings and up to 59% of older adults in inpatient settings were taking at least one PIM (Barton, Sklenicka, Sayegh, & Yaffe, 2008; Buck et al., 2009; Corsonello et al., 2009; Hale et al., 2008; Rajska-Neumann & Wieczorowska-Tobis, 2007). The high prevalence rate of prescriptions for PIMs can have significant negative consequences, as older adults taking PIMs endorse higher rates of drug-related problems than those who are not (Fick, Mion, Beers, & Waller, 2008). For example, Gallagher and colleagues (2008) found that over a 3-month period, 16% of emergency room admissions of acutely ill older adults could be linked to adverse effects of inappropriate prescribing.

Medications with anticholinergic (AC) properties are one class of PIMs that is commonly prescribed in older adult populations (van Eijk et al., 2000). Anticholinergic agents block the binding of the neurotransmitter acetylcholine to its receptor sites in both the central and peripheral nervous system. In the central nervous system, acetylcholine is known to be important toward cognitive functioning, particularly skills important for learning and memory. A multitude of potentially medically necessary medications have AC properties including antidepressants, antihistamines, muscle relaxants, and those used to treat urinary incontinence, vertigo, ulcers, cardiovascular problems, and gastrointestinal problems. Community-based studies have found that up to 20% of older adults sampled were taking drugs with AC properties (Cancelli et al., 2008; Lechevallier-Michel, Molimard, Dartigues, Fabrigoule, & Fourrier-Reglat, 2005). Not surprisingly, the prevalence of AC medications is particularly high in inpatient and nursing home settings (Paton et al., 2003). For example, Blazer and colleagues (1983) found that over a 1-year span, nearly 60% of nursing home residents had received AC drugs with 32% of residents taking more than one and 5% taking more than five.

AC medication in older adults has been associated with physical side-effects such as dry mouth, constipation, and falls (Ness, Hoth, Barnett, Shorr, & Kaboli, 2006; Rudolph, Salow, Angelini, & McGlinchey, 2008), greater physical and functional difficulties (Cao et al., 2008), and poorer hospitalization outcomes such as an increased length of stay and increased risk of medical complications (Agostini, Leo-Summers, & Inouye, 2001). Use of AC medications can also have detrimental effects on cognitive abilities, and studies have associated AC medication use with poor performance on measures of mental status, memory, language, and complex attention in outpatient samples (Ancelin et al., 2006; Bottiggi et al., 2006; Cancelli et al., 2008; Lechevallier-Michel et al., 2005; Low, Anstey, & Sachdev, 2009; Mulsant et al., 2003). However, it should be noted that the cognitive effects of AC medications are not clearly consistent across studies, with some studies showing impairments in some domains but not others. For example, Low, Anstey, and Sachdev (2009) showed an association between AC burden and complex attention but not verbal memory, working memory, or reaction time whereas Ancelin et al. (2006) found negative AC effects across multiple cognitive domains including reaction time, attention, and nonverbal memory but not verbal memory or reasoning.

There is less information about the association between AC medications and cognitive functioning in inpatient extended care settings, which is of particular relevance given that cognitive dysfunction may lead to poor rehabilitation outcomes in older adults (Landi et al., 2002; Nanna, Lichtenberg, BudaAbela, & Barth, 1997; Galski, Bruno, Zorowitz, & Walker, 1993). Inpatients also take more medications than their outpatient counterparts (Flaherty, Perry, Lynchard, Morley, & Chien, 1997) and are more likely to have an increased medical illness burden, both of which are risk factors for adverse drug events (Field et al., 2001; Hanlon et al., 1997; van den Bemt et al., 1999). Thus, while optimizing functioning has important implications for treatment success in inpatients, these same individuals are paradoxically at increased risk for medication-related cognitive impairment because of multiple medical comorbidities and polypharmacy.

Studies of nondemented inpatients have found an association between AC medications and delirium, decreased mental status, and decreased performance on verbal learning tests ( Agostini et al., 2001; Miller et al., 1988; Mussi, Ferrari, Ascari, & Salvioli, 1999; Nebes et al., 1997; Tune et al., 1981). One drawback to many of the previous studies is that they used relatively gross measures of cognitive functioning such as the Mini-Mental Status Exam or a delirium screen (e.g., Agostini et al., 2001; Mussi et al., 1999). These screening measures do not provide a comprehensive examination of how AC medications may affect various cognitive domains in more subtle ways. In contrast, the few studies that have assessed multiple cognitive domains have generally examined small samples of patients (i.e., less than 75; e.g., Miller et al., 1988; Nebes et al., 1997).

In addition, while physicians certainly need to be aware of potential negative effects of AC medications, it is likely not feasible to eliminate use of all AC medications in this population. Many previous studies assessed the cognitive effects of AC medications by comparing the cognitive performance of those individuals who are taking AC medications to those who are not. However, AC drug effects can arise not just from individual drugs but from an accumulation of AC burden from different medications (Boustani et al., 2008; Tune, 2001). Examining the potentially additive effects of multiple AC medications may be more relevant to inpatient populations given the increased likelihood of prescribing drugs with at least some AC properties in this population.

The current study examines prescribing practices of AC medications and their association with cognitive functioning in an inpatient extended care setting. The study assesses the possibility of additive effects of AC medications across multiple domains of cognitive functioning in a large sample of individuals whose inpatient status increases their vulnerability to adverse drug effects. It is well known that AC medications can lead to acute symptoms of delirium (Mussi et al., 1999; Tune et al., 1981) that are likely to alert clinicians to a need for a change in medications. The current study, however, examined whether AC burden is associated with more subtle cognitive difficulties that may go undetected in a sample of patients that are otherwise cognitively intact. AC burden was calculated using the Anticholinergic Drug Scale (ADS; Carnahan, Lund, Perry, & Culp, 2002), a burden scale that takes into account differences in AC properties across medications. With this scale, an AC burden score can be calculated for each patient by summing the AC effects of all prescribed medications. The ADS has been validated as a measure of AC activity in two separate studies using serum levels (Carnahan et al., 2002; Carnahan et al., 2006). In the most recent study examining 201 subjects in long-term care facilities, ADS total scores were significantly associated with serum anticholinergic activity, explaining 9.5% of the variance in serum levels (Carnahan et al., 2006).

Given that both depressive symptoms and levels of premorbid cognitive abilities can affect an individual’s current cognitive status, we controlled for these variables in our analyses. Based on previous research illustrating the effects of AC medications on cognition, we hypothesized that in a sample of older adults in an inpatient extended care unit, higher levels of AC burden would be associated with decreased cognitive functioning.

Method

Participants

Data from 450 participants (429 males) were included in the study. Participants were patients in the Extended Care Center at a Midwest VA Medical Center, an inpatient facility designed to provide rehabilitation and postacute care (average length of stay: 28-30 days). Common reasons for admission to this unit included postsurgical care, wound care, limb amputations, IV administration of medications, and acute management of complications arising from chronic illnesses. Participants ranged in age from 50 to 89 (M = 67.95, SD = 10.55) with education levels ranging from 4 to 20 years (M = 12.40, SD = 2.58). A total of 408 participants (90.7% of the sample) endorsed significant symptoms of depression with a score of 5 or above on the 15-item Geriatric Depression Scale (Sheikh & Yesavage, 1986) with an average GDS score of 6.75 (SD = 2.10). See Table 1 for details.

Table 1.

M (SD) Demographic Data for Included and Excluded Participants

	Included participants	Excluded participants
N	450	59
Males	429	59
Age**	67.95 (10.55)	73.33 (10.36)
Education*	12.40 (2.58)	11.41 (2.63)
Estimated IQ***	93.60 (10.42)	77.79 (14.29)
MMSE***	26.10 (2.63)	19.94 (4.09)
MDAS***	2.75 (1.78)	6.03 (2.59)
GDS	6.75 (2.10)	7.24 (1.90)
ADS burden	2.31 (1.61)	2.45 (1.73)

Note: MMSE = Mini-Mental Status Exam; MDAS = Memorial Delirum Assessment Scale; GDS = Geriatric Depression Scale, ADS = Anticholinergic Drug Scale.

p < .05, **p < .001, ***p < .001.

Measures

A brief neuropsychological screen was administered that assessed multiple domains including general mental status, estimated premorbid intellectual abilities, frontal functioning, memory, and depression. Specific measures are described below:

The Memorial Delirium Assessment Scale

(MDAS; Breitbart et al., 1997) is a 10-item clinician rating scale of symptoms of delirium including disorientation, disorganized thinking, delusions, perceptual disturbances, decreased psychomotor activity, and sleep-wake cycle disturbances. A score of eight or above is indicative of delirium. This measure has been shown to have good reliability and high correlations with clinician ratings of delirium severity in hospitalized individuals (Breitbart et al., 1997).

The Mini-Mental Status Exam

(MMSE; Folstein, Folstein, & Mchugh, 1975) is a brief measure of general mental status that is frequently used to screen for dementia. This 30-item exam briefly examines aspects of orientation, short-term memory, attention, language, and visuospatial abilities. Scores on this task range from 0 to 30. The MMSE has been shown to have good internal consistency and test-retest reliability and moderate to high sensitivity and specificity, particularly in hospitalized individuals (Tombaugh & McIntyre, 1992).

The Peabody Picture Vocabulary Test-III

(PPVT; Dunn & Dunn, 1997), a receptive language task, was used as an estimate of premorbid verbal intellectual abilities (Snitz, Bieliauskas, Crossland, Basso, & Roper, 2000). Standardized IQ scores were derived from raw scores on this task based on test manual age-based norms. The PPVT has been shown to have high internal consistency, test-retest reliability, and criterion validity with tests of intelligence (Dunn & Dunn, 1997).

The Frontal Assessment Battery

(FAB; Dubois, Slachevsky, Litvan, & Pillon, 2000) is a bedside cognitive exam sensitive to frontal lobe dysfunction. The FAB consists of six subtests assessing conceptualization, mental flexibility, motor programming, sensitivity to interference, inhibitory control, and environmental autonomy. Scores on this task range from 0 to 18. The FAB has been shown to have good interrater reliability, internal consistency, and discriminant validity (Dubois et al., 2000).

Neurobehavioral Cognitive Status Exam Judgment Subtest

(NCSE Judgment; Kiernan, Mueller, Langston, & Vandyke, 1987) is a measure of verbal reasoning and judgment during which patients are asked how they would respond in everyday predicaments. Scores on this task range from 0-6. Studies examining the properties of the NCSE as a screen for cognitive impairment have shown good sensitivity across patient groups with variable specificity (Malloy et al., 1997). The NCSE judgment subtest on its own has been shown to have good test-retest reliability (Mitrushina, Abara, & Blumenfeld, 1994).

The Hopkins Verbal Learning Test-Revised

(HVLT-R; Benedict, Schretlen, Groninger, & Brandt, 1998) is a 12-item word list-learning task. The list of words is read to the examinee three consecutive times. Patients are asked to recall the words immediately after each presentation and again after a 20-min delay. Variables of interest were the total number of items recalled on the first three trials (immediate recall) and the number of items recalled on the delay trial (delayed recall). Immediate recall scores range from 0-36 while delayed recall scores range from 0-12. The validity and reliability of the HVLT-R as a measure of verbal learning and memory has been demonstrated in multiple studies (Benedict, Schretlen, Groninger, & Brandt, 1998; de Jager, Hogervorst, Combrinck, & Budge, 2003; Shapiro, Benedict, Schretlen, & Brandt, 1999).

The short form of the Geriatric Depression Scale (GDS; Sheikh & Yesavage, 1986) is a 15-item self-report inventory of depressive symptoms. Clinically significant depression is indicated by scores of five and above on this scale. The short form of the GDS has been shown to have good sensitivity, specificity, and reliability (Almeida & Almeida, 1999; van Marwijk et al., 1995).

Procedure

The neuropsychological screen was administered to all patients admitted to the Extended Care Center as part of standard clinical care. After obtaining IRB approval and a waiver of informed consent, data from all screens administered from 2003 to 2006 were entered into a database and analyzed. Data were not included in the study if participants did not complete all portions of the cognitive screen or if their estimated premorbid IQ was below 70. Participants were also excluded for delirium as indicated by a score greater than 7 on the MDAS and for dementia as indicated by a score of less than 20 on the MMSE. A cut-off score of 20 was used because MMSE scores are sensitive to IQ and education levels even in healthy, nonclinical populations (Crum, Anthony, Bassett, & Folstein, 1993; Starr & Lonie, 2007; Starr, Whalley, Inch, & Shering, 1992; Whitney, Maoz, Hook, Steiner, & Bieliauskas, 2007), and a low MMSE score may not necessarily reflect dementia in individuals with low IQ. Moreover, since we hypothesized that some nondemented individuals would show declines on this task due to medication effects, we used a lower cut-off than the typical cut-off score of 24. A total of 59 individuals were excluded from the original sample due to delirium (N = 22), low performance on the PPVT (N = 11), or low score on the MMSE (N = 41); some individuals performed below criterion levels on multiple measures. Independent sample t-tests demonstrated that excluded individuals were significantly older (t[507] = 2.69, p = .007) and had lower levels of education (t[507]= −2.05, p = .04) than those individuals included in the sample. Demographic information for these excluded individuals is also presented in Table 1.

We examined five cognitive variables: score on the FAB (possible range: 0-18), score on the NCSE-Judgment (possible range: 0-6), words recalled on the HVLT-R immediate recall (possible range: 0-36), words recalled on the HVLT-R delayed recall (possible range: 0-12), and score on the MMSE (possible range: 0-30). On all measures, a higher raw score indicates better cognitive performance. Because age norms were not available for all tasks, age effects were accounted for by dividing participants into four age groups: 50-59, 60-69, 70-79, and 80 and above. Raw scores on each task were computed into age-standardized z-scores for each participant with respect to his or her age group in the sample.

To calculate AC burden, we reviewed each participant’s medical record and recorded all medications administered on the day of the neuropsychological screen. The AC properties of all medications were rated using the ADS. All drugs were rated on a scale from 0 to 3: level 0 = no known AC properties, level 1 = potentially AC as evidenced by receptor binding studies, level 2 = AC adverse effects sometimes noted, usually at excessive doses, level 3 = markedly AC. AC burden scores were calculated for each participant by summing the ratings for all medications taken on the day of testing. The list of AC medications in the ADS can be found in the appendix of Carnahan et al. (2006).

Results

Estimated premorbid verbal IQ levels ranged from 71 to 133 (M = 93.49, SD = 10.45). Means and standard deviations of raw scores for the FAB, MMSE, HVLT-R, and NCSE Judgment are presented in Table 2. Alpha levels of .05 were used for all statistical tests.

Table 2

M (SD) Scores on the Cognitive Tests, Stratified by Age Groups

Age group	MMSE	Judgment	FAB	HVLT-R immediate	HVLT-R delayed
50-59	26.87 (2.40)	5.32 (1.11)	14.67 (2.39)	18.77 (4.77)	5.76 (3.20)
60-69	26.49 (2.64)	5.46 (0.97)	13.93 (2.71)	17.57 (4.99)	5.55 (2.95)
70-79	25.66 (2.73)	4.99 (1.37)	12.69 (3.01)	14.25 (4.86)	4.08 (2.93)
80+	25.03 (2.33)	5.18 (1.21)	12.82 (2.47)	13.92 (4.37)	3.24 (2.77)
Total	26.10 (2.63)	5.2 (1.18)	13.61 (2.78)	16.35 (5.20)	4.78 (3.15)

Note: MMSE = Mini-Mental Status Exam; FAB = Frontal Assessment Battery; HVLT-R = Hopkins Verbal Learning Test-Revised.

Prescribing Patterns

On average, the total number of medications taken was 13.87 with a range of 2 to 33 (SD = 5.51). Participants were taking an average of 2.30 medications with AC properties (SD = 1.62, range: 0-8). AC burden scores ranged from 0 to 14 with a mean of 2.31 (SD = 1.61) and a mode of 2. The most common AC medications were furosemide (n = 116; a diuretic) and oxycodone (n = 174; an opioid analgesic), both drugs with level 1 AC ratings, and ranitidine (n = 81; an antihistamine with a level 2 AC rating). The most common level 3 AC medication was promethazine (n = 27), also an antihistamine. Of the 450 participants, 49 (10.89%) individuals were not taking any AC medications while 89 (19.78%) were taking five or more AC medications. Out of the 401 participants taking AC medications, 309 (77.06%) were taking only level 1 AC medications. Sixty-eight of the 401 participants (15.11%) were taking a level 3 AC medication and 12 of these 68 were taking multiple level 3 AC medications. AC burden and total number of AC medications were significantly correlated with age, even when controlling for total number of medications in a two-tailed partial correlation (r = −.17, p < .001; r = −.18, p < .001, respectively).

Anticholinergic Burden and Cognition

Premorbid verbal IQ was positively correlated with z-scores on each of the five cognitive tasks (all p’s < .001). Depression was negatively correlated with z-scores on tasks of judgment (r = −.10; p = .03), the MMSE (r = −.09; p = .04) and the FAB (r = −.11; p = .01). To control for the effects of depression and premorbid verbal IQ on cognitive performance, a two-tailed partial correlation was used with IQ and GDS scores as covariates. AC burden and z-scores on the MMSE, FAB, NCSE Judgment, HVLT-R immediate recall, and HVLT-R delayed recall were included as variables of interest in the partial correlation. AC burden was positively correlated with NCSE Judgment (r = .10, p = .04) but was not significantly correlated with any of the other cognitive variables. These partial correlations are presented in Table 3.

Table 3

Correlations Between ADS Burden Scores and Cognitive z-Scores Before and After Controlling for Premorbid IQ and Geriatric Depression Scale (GDS) Scores

	ADS burden	ADS burden (controlling for IQ and GDS)
MMSE	.04	.01
FAB	.01	−.003
Judgment	.10*	.10*
HVLT-R immediate	.02	−.02
HVLT-R delayed	.06	.03

Note: ADS = Anticholinergic Drug Scale; MMSE = Mini-Mental Status Exam; FAB = Frontal Assessment Battery; HVLT-R = Hopkins Verbal Learning Test-Revised.

p = .04.

Analyses of covariance were also performed to compare the cognitive performance of individuals with high and low levels of AC burden. Cognitive z-scores of those participants with a burden score of 0 (n = 54) were compared to those with a score of six and above (n = 62) with premorbid IQ and GDS scores included as covariates in the analyses. There were no significant differences between these two groups on any of the five cognitive tasks. Analyses of covariance were also performed on the z-scores of individuals on AC medications (n = 401) versus the scores of those who were not taking any AC medications (n = 49), with premorbid IQ and GDS scores as covariates. These analyses also did not yield any significant differences between the groups on any of the cognitive measures. We also used analyses of covariance to compare the cognitive z-scores of individuals taking at least one level 3 AC medication (n = 68) and individuals not taking any AC medications (n = 49) with premorbid IQ and GDS scores as covariates. These analyses did not yield any significant group differences on any of the cognitive measures.

To control for possible variations in physical illness among patients, post hoc partial correlations were reanalyzed with total number of medications administered also included as a covariate. The previous results were unchanged with the exception of the relation between AC burden and judgment, which was no longer significant (r = .05, p = ns). To examine post hoc a possible interaction effect between age and burden level on cognition, an age-by-burden interaction term was entered into the partial correlation with the five cognitive variables and premorbid IQ and GDS scores as covariates. The interaction term was not significantly correlated with any of the cognitive variables.

Post hoc analyses were also conducted to further explore possible explanations for our null findings. A high percentage of patients (77.06%) were taking only level 1 AC medications. While the ADS classifies level 1 AC medications as having potentially AC properties based on receptor binding studies, this may not necessarily translate into clinically meaningful adverse effects. To examine this possibility, we recalculated AC burden scores for each patient, only including AC drugs with noted adverse affects or marked AC properties (i.e., level 2 or 3 AC medications). Post hoc partial correlations between these recalculated AC burden scores and the cognitive scores, controlling for premorbid IQ and GDS, yielded no significant correlations between AC burden and any of the cognitive variables.

A post hoc partial correlation was also conducted to determine whether a different method of accounting for age variance would affect our findings. AC burden and the raw scores on the cognitive tasks were entered into the partial correlation as variables of interest and age, IQ, and GDS score were entered as control variables. The results of this post hoc analysis were consistent with the primary analysis in that none of the cognitive tasks was significantly associated with AC burden with the exception of a trend toward a positive correlation between burden and judgment (r = .08, p = .08).

We also considered that our method for measuring AC burden may have accounted for our null findings. To examine this, we recalculated AC burden using a clinician-rated AC scale developed by Han et al. (2001) and Han, Agostini, and Allore (2008). This clinician-rated AC scale is similar to the ADS in that it assigns medications a score from 0 (none) to 3 (high); however this scale is based on clinician ratings and has previously been associated with performance on the HVLT and a measure of functional daily abilities (e.g., using a telephone). The ADS burden scores were significantly correlated with the clinician-rated AC scores (r = .66, p < .001). Post hoc partial correlations between the clinician-rated AC scores and the cognitive tasks, covarying for premorbid IQ and GDS scores, were not significant.

Discussion

In the current study, we did not find an association between AC medications and poor cognitive performance in a sample of hospitalized older adults. AC burden was not significantly associated with performance on measures of mental status, memory, or executive functioning. These results suggest that even in a sample of inpatients with increased vulnerability to AC burden effects, current prescribing practices are not associated with negative cognitive side-effects.

The current results suggest that AC medication usage in our sample of inpatients is not associated with adverse cognitive effects, which is encouraging. This may reflect attempts on the part of prescribing clinicians to mitigate potential AC side-effects. We found significant negative correlations between AC burden and age and total number of AC medications and age, even when controlling for total number of medications taken. This means that older participants were less likely to be prescribed medications with AC properties and less likely to have medication regimens with high AC burden. This suggests that practitioners are more cautious in prescribing AC medications to older adults, possibly secondary to increased awareness of the potential risks of AC medications and PIMs in general. The VA Medical Center where these data were collected previously took part in a multisite intervention in 2004 aimed at increasing awareness of high-risk medications in older adults. This program targeted prescribing physicians and was found to be effective, as 50% of patients who were previously taking PIMs had these medications discontinued at follow-up (Zillich et al., 2008). These results, coupled with the current findings, suggest that efforts on the part of health care systems to increase awareness of PIMs in older patients are likely having positive effects on prescribing practices and improving standards of clinical care.

We found a positive correlation between AC burden and performance on a task of judgment. This was unexpected and should be interpreted with caution given that this has not previously been reported in the literature. Furthermore, this relationship was no longer significant when controlling for total number of medications, using a different rating scale to measure AC burden, and eliminating level 1 AC medications from the burden scores. Replications of these findings are needed before more substantial conclusions can be drawn. We did not find significant associations between AC burden and any of the other cognitive variables, which is not consistent with previous research that has found associations between AC medications and poor cognitive functioning (e.g., Bottiggi et al., 2006; Cancelli et al., 2008; Lechevallier-Michel et al., 2005; Miller et al., 1988; Mulsant et al., 2003; Nebes et al., 1997). However, differences between the current study and these previous studies may explain these discrepant findings. First, variations in samples may also account for differences in findings. For example, Nebes et al. (1997) examined AC medications and cognition in a sample of older adults with diagnosed depression. Most studies examining cognitive functioning and AC medications have studied community-based outpatient samples (e.g., Cancelli et al., 2008), which can differ greatly from inpatient samples in terms of medications and medical history, as discussed previously. Different methodologies have also been used; for example, Miller et al. (1988) used pharmacological challenge to examine the cognitive effects of AC medications.

Our null findings may also reflect the low levels of AC medications in our sample. Although a substantial number of participants (89%) in our sample were taking medications with AC effects, of these participants, over three-quarters were taking medications with only minimal AC properties (i.e., level 1 medications) and only 12 total patients in the entire sample were taking more than one medication with marked AC effects (i.e., level 3 medications). However, when we recalculated burden scores by including only medications with noted adverse effects or marked AC properties (i.e., level 2 and 3 medications), post hoc partial correlations between these recalculated burden scores and the cognitive scores were not significant. It still may be that negative cognitive effects are only found with AC burden levels that are higher than those measured in the current study. Future research examining the possible threshold at which AC burden produces significant cognitive effects would be important and directly applicable to clinical care. In addition, because our sample was relatively young (mean age = 67.95) with relatively intact mental status (mean MMSE score = 26.10), the participants in our study may not have been as vulnerable to the effects of AC medications as older individuals or individuals with compromised cognitive functioning, such as those with delirium or dementia.

Alternatively, the hospitalized participants in this sample had a number of multiple medical comorbidities and other prescribed medications (e.g., narcotics) that may have contributed to variance in cognitive performance and masked more subtle AC effects. However, we conducted the current study to examine medication effects in a sample that accurately reflects the typical cross-section of individuals in an inpatient extended care setting. While research utilizing strict exclusion criteria to identify a sample of healthy individuals could better address the direct effects of AC medications on cognitive functioning in a controlled environment, we were more interested in the state of current prescribing practices in a sample of patients that accurately reflects real-world extended care patients. Therefore, we purposefully did not exclude for medical or medication status to avoid artificially selecting a sample that was not ecologically valid and representative of the intended population. It may also be that AC medications in some individuals actually have positive effects on cognitive functioning due to the successful treatment of medical symptoms that were previously affecting cognitive abilities. For example, AC medications are commonly used to treat chronic obstructive pulmonary disease, a disease that has often been associated with negative cognitive consequences (Liesker et al., 2002; 2004).

We also considered the possibility that our choice of the ADS as a measure of AC burden led to our null findings. To our knowledge, only one previous study has used the ADS to examine AC burden and cognitive functioning. Low, Anstey, and Sachdev (2009) used the ADS to examine AC medications and effects on cognitive functioning in a large community-based sample of older adults. They found that AC medications were associated with poorer performance on a complex attention task but not measures of verbal memory, working memory, reaction time, or general mental status. To further explore whether our AC burden measure affected our findings, we recalculated AC burden using a clinician-rated AC scale developed by Han et al. (2001; 2008). This scale had previously been associated with cognitive performance in a sample of community-dwelling older men (Han, Agostini, & Allore, 2008). The ADS scores and the clinician-rated AC scores were highly correlated and post hoc partial correlations between the clinician-rated AC scores and the cognitive scores were not significant. These findings increase our confidence that the ADS is adequately measuring AC burden. In addition, our current study adds to the scant literature using the ADS to examine associations between AC burden and cognitive functioning.

There are several limitations to the current study. We did not measure serum AC activity, a method used in research to measure the degree of drug binding to muscarinic acetylcholine receptors (see Rudd, Raehl, Bond, Abbruscato, & Stenhouse, 2005). Replication of these results using a serum marker of AC activity would be informative. We also did not factor in medication dosage, although Carnahan et al. (2006) has previously shown that including dosage levels in the ADS did not increase its prediction of serum AC levels. An additional limitation is that we did not take into account differences between long-term versus acute exposure to AC medications, as only medications taken on the day of testing were factored into burden scores. It is possible that cognitive side-effects are more salient with prolonged use of AC medications. We also did not take into account the timing of medication administrations in relationship to the cognitive assessment and therefore could not account for the possibility of phasic fluctuations in AC levels. It may also be that there are idiosyncratic differences among individuals that moderate the relationship between AC burden and cognition. For example, research has found that the administration of trihexyphenidyl, an AC antiparkinsonian drug, in healthy older adults led to ratings of mental slowness only in those individuals who were APOE e4 carriers (Pomara, Belzer, Hernando, De la Pena, & Sidtis, 2008). While it may be that the current tests were not sensitive enough to detect subtle cognitive changes associated with AC burden, much of the previous research has used relatively gross measures such as the MMSE and still shown AC effects. Thus, we would expect that the tasks used in our study would show AC effects, if present.

We also recognize that by excluding individuals with delirium from our sample we may have excluded those individuals with the most significant AC side-effects. However, we excluded only 22 individuals for delirium, and these individuals had a mean burden of 2.45 (range: 0-5). This suggests that these individuals were not receiving a disproportionate level of AC medications as compared to the individuals in our final sample. In addition, we studied a sample of nondelirious, nondemented older adults who were hospitalized in an extended care unit. Therefore, our findings may not necessarily generalize to older adults seen in other settings such as ambulatory care or to older adults with symptoms of dementia or delirium.

The results of this study are highly encouraging and dissemination of these null findings is important and directly applicable to both researchers and clinicians working with older adults. We demonstrated that AC burden and total number of AC medications decreased with age in a sample of nondelirious, nondemented older adults hospitalized for extended care. For gerontological researchers, this suggests that previous research demonstrating cognitive consequences associated with AC medications has likely translated into increased awareness and caution on the part of prescribers. This optimistically illustrates the importance of clinical research and its ability to inform clinical practice. As discussed above, it is also possible that increased efforts on the part of health care systems to increase the awareness of prescribers are having a positive effect on prescribing practices, which is important and encouraging information for those in charge of such initiatives.

In addition, our findings that AC burden was not associated with poor cognitive performance suggest that medications with AC properties may not be as detrimental to cognitive functioning as previous research has suggested, at least in the population examined in the current study. This is important information for practicing gerontologists who may be faced with the difficult task of weighing the medical benefits of prescribing AC medications against the potential cognitive consequences. Our findings that AC medications may not necessarily be associated with poor cognitive functioning in an inpatient setting can increase clinicians’ confidence in their decision to prescribe AC medications. Although gerontologists certainly need to be aware of the potential for cognitive side-effects, these clinicians are often faced with a need to treat; AC medications cannot always be avoided. Even though AC medications were highly prevalent in our sample, they were not associated with poor cognitive performance and therefore clinicians treating patients similar to those in our study sample can be reassured that AC cognitive side-effects can be successfully mitigated through careful prescribing practices.

In sum, we did not find associations between low levels of AC burden and poor cognitive performance in a sample of nondemented and nondelirious older adults hospitalized in an extended care center. While AC medications were commonly prescribed in this sample, the majority of these medications did not have marked AC activity, suggesting that clinicians may be more cautious in prescribing AC medications to older adults. Consistent with this, we found that AC burden scores and total number of AC medications negatively correlated with patient age. These results are encouraging and suggest that in a population with increased vulnerability to adverse drug effects, current prescribing practices of AC medications are not associated with negative cognitive side-effects. These findings paint an optimistic picture of current prescribing practices and suggest adequate management of AC cognitive side-effects in an extended care setting.

Replications with larger sample sizes and measures of serum AC activity would further strengthen the findings of the current study.

Footnotes

Acknowledgements

The authors would like to thank Lindsey Harik, Danielle Harris, and David Kelly as well as other members of the VA Ann Arbor Extended Care Center research team for their work on this project. We thank Gus Buchtel for his assistance with manuscript edits. We thank Robert Hogikyan, M.D. for providing us with data about individuals in the VA Ann Arbor Extended Care Center.

The authors declared no potential conflicts of interests with respect to the authorship and/or publication of this article.

The authors received no financial support for the research and/or authorship of this article.

References

Almeida

(1999). Short versions of the Geriatric Depression Scale: A study of their validity for the diagnosis of a major depressive episode according to ICD-10 and DSM-IV. International Journal of Geriatric Psychiatry, 14, 858-865.

Ancelin

Artero

Portet

Dupuy

Touchon

Ritchie

(2006). Non-degenerative mild cognitive impairment in elderly people and use of anticholinergic drugs: Longitudinal cohort study. British Medical Journal, 332, 455-459.

Agostini

J. V.

Leo-Summers

L. S.

Inouye

S. K.

(2001). Cognitive and other adverse effects of diphenhydramine use in hospitalized older patients. Archives of Internal Medicine, 161, 2091-2097.

Barton

Sklenicka

Sayegh

Yaffe

(2008). Contraindicated medication use among patients in a memory disorders clinic. American Journal of Geriatric Pharmacotherapy, 6, 147-152.

Benedict

R. H. B.

Schretlen

Groninger

Brandt

(1998). Hopkins verbal learning test revised: Normative data and analysis of inter-form and test-retest reliability. Clinical Neuropsychologist, 12, 43-55.

Blazer

D. G.

Federspiel

C. F.

Ray

W. A.

Schaffner

(1983). The Risk of anticholinergic toxicity in the Elderly—A study of prescribing practices in 2 populations. Journals of Gerontology, 38, 31-35.

Bottiggi

K. A.

Salazar

J. C.

Caban-Holt

A. M.

Ryan

Mendiondo

M. S.

. (2006). Long-term cognitive impact of anticholinergic medications in older adults. American Journal of Geriatric Psychiatry, 14, 980-984.

Boustani

Campbell

Munger

Maidment

Fox

(2008). Impact of anticholinergics on the aging brain: A review and practical application. Aging Health, 4, 311-320.

Breitbart

Rosenfeld

Roth

Smith

M. J.

Cohen

Passik

(1997). The Memorial Delirium Assessment Scale. Journal of Pain and Symptom Management, 13, 128-137.

10.

Buck

M. D.

Atreja

Brunker

C. R.

Jain

Suh

T. T.

Palmer

R. M.

. (2009). Potentially inappropriate medication prescribing in outpatient practices: Prevalence and patient characteristics based on electronic health records. American Journal of Geriatric Pharmacotherapy, 7, 84-92.

11.

Cancelli

Gigli

G. L.

Piani

Zanchettin

Janes

Rinaldi

. (2008). Drugs With anticholinergic properties as a risk factor for cognitive impairment in elderly people a population-based study. Journal of Clinical Psychopharmacology, 28, 654-659.

12.

Cao

Y. J.

Mager

D. E.

Simonsick

E. M.

Hilmer

S. N.

Ling

S. M.

Windham

B. G.

. (2008). Physical and cognitive performance and burden of anticholinergics, sedatives, and ACE inhibitors in older women. Clinical Pharmacology & Therapeutics, 83, 422-429.

13.

Carnahan

Lund

Perry

Pollock

Culp

(2006). The Anticholinergic Drug Scale as a measure of drug-related anticholinergic burden: Associations with serum anticholinergic activity. Journal of Clinical Pharmacology, 46, 1481-1486.

14.

Carnahan

R. M.

Lund

B. C.

Perry

P. J.

Culp

K. R.

(2002). The relationship of an anticholinergic rating scale with serum anticholinergic activity in elderly nursing home residents. Pharmacotherapy, 22, 1341.

15.

Corsonello

Pedone

Lattanzio

Lucchetti

Garasto

Di Muzio

. (2009). Potentially inappropriate medications and functional decline in elderly hospitalized patients. Journal of the American Geriatrics Society, 57, 1007-1014.

16.

Crum

Anthony

Bassett

Folstein

(1993). Population-based norms for the Mini-Mental State Examination by age and education level. Journal of the American Medical Association, 12, 2386-2391.

17.

de Jager

Hogervorst

Combrinck

Budge

(2003). Sensitivity and specificity of neuropsychological tests for mild cognitive impairment, vascular cognitive impairment and Alzheimer’s disease. Psychological Medicine, 33, 1039-1050.

18.

Dubois

Slachevsky

Litvan

Pillon

(2000). The FAB - A frontal assessment battery at bedside. Neurology, 55, 1621-1626.

19.

Dunn

(1997). Peabody picture vocabulary test (PPVT-III, 3rd ed.). Circle Pines, MN: AGS Publishing.

20.

Feinberg

(1993). The problems of anticholinergic adverse-effects in older patients. Drugs & Aging, 3, 335-348.

21.

Fick

D. M.

Mion

L. C.

Beers

M. H.

Waller

J. L.

(2008). Health outcomes associated with potentially inappropriate medication use in older adults. Research in Nursing & Health, 31, 42-51.

22.

Field

T. S.

Gurwitz

J. H.

Avorn

McCormick

Jain

Eckler

. (2001). Risk factors for adverse drug events among nursing home residents. Archives of Internal Medicine, 161, 1629-1634.

23.

Flaherty

J. H.

Perry

Lynchard

Morley

J. E.

Chien

N. T.

(1997). The effect of polypharmacy and inappropriate medication use in older home care patients. Journal of the American Geriatrics Society, 45, 261.

24.

Folstein

M. F.

Folstein

S. E.

Mchugh

P. R.

(1975). Mini-mental state—Practical method for grading cognitive state of patients for clinician. Journal of Psychiatric Research, 12, 189-198.

25.

Gallagher

P. F.

Barry

P. J.

Ryan

Hartigan

O’Mahony

(2008). Inappropriate prescribing in an acutely ill population of elderly patients as determined by Beers Criteria. Age and Ageing, 37, 96-101.

26.

Galski

Bruno

R. L.

Zorowitz

Walker

(1993). Predicting length of stay, functional outcome, and aftercare in the rehabilitation of stroke patients—The dominant role of higher-order cognition. Stroke, 24, 1794-1800.

27.

Hale

Griffin

Cartwright

Moulin

Alford

Fleming

(2008). Potentially inappropriate medicaiton use in hospitalized older adults: A DUE using the full Beers criteria. Formulary, 43, 326-339.

28.

Han

McCusker

Cole

Abrahamowicz

Primeau

Élie

(2001). Use of medications with anticohlinergic effect predicts clinical severity of delirium symptoms in older medical inpatients. Archives of Internal Medicine, 161, 1099-1105.

29.

Han

Agostini

Allore

(2008). Cumulative anticholinergic exposure is associated with poor memory and executive function in older men. Journal of the American Geriatrics Society, 56, 2203-2210.

30.

Hanlon

J. T.

Schmader

K. E.

Koronkowski

M. J.

Weinberger

Landsman

P. B.

Samsa

G. P.

. (1997). Adverse drug events in high risk older outpatients. Journal of the American Geriatrics Society, 45, 945-948.

31.

Kiernan

R. J.

Mueller

Langston

J. W.

Vandyke

(1987). The Neurobehavioral Cognitive status examination—A brief but differentiated approach to cognitive assessment. Annals of Internal Medicine, 107, 481-485.

32.

Landi

Bernabei

Russo

Zuccala

Onder

Carosella

. (2002). Predictors of rehabilitation outcomes in frail patients treated in a geriatric hospital. Journal of the American Geriatrics Society, 50, 679-684.

33.

Lechevallier-Michel

Molimard

Dartigues

J. F.

Fabrigoule

Fourrier-Reglat

(2005). Drugs with anticholinergic properties and cognitive performance in the elderly: results from the PAQUID Study. British Journal of Clinical Pharmacology, 59, 143-151.

34.

Liesker

J. J. W.

Postma

D. S.

Beukema

R. J.

Ten Hacken

N. H. T.

van der Molen

Riemersma

R. A.

. (2004). Cognitive performance in patients with COPD. Respiratory Medicine, 98, 351-356.

35.

Liesker

J. J. W.

Wijkstra

P. J.

Ten Hacken

N. H. T.

Koeter

G. H.

Postma

D. S.

Kerstjens

H. A. M.

(2002). A systematic review of the effects of bronchodilators on exercise capacity in patients with COPD. Chest, 2121, 597-608.

36.

Low

L. F.

Anstey

K. J.

Sachdev

(2009). Use of medications with anticholinergic properties and cognitive function in a young-old community sample. International Journal of Geriatric Psychiatry, 24, 578-584.

37.

Malloy

Cummings

Coffey

Duffy

Fink

Lauterbach

E. C.

. (1997). Cognitive screening instruments in neuropsychiatry: A report of the Committee on Research of the American Neuropsychiatric Association. Journal of Neuropsychiatry and Clinical Neurosciences, 9, 189-197.

38.

Miller

P. S.

Richardson

J. S.

Jyu

C. A.

Lemay

J. S.

Hiscock

Keegan

D. L.

(1988). Association of Low serum anticholinergic levels and cognitive impairment in elderly presurgical patients. American Journal of Psychiatry, 145, 342-345.

39.

Mitrushina

Abara

Blumenfeld

(1994). Aspects of validity and reliability of the neurobehavioral cognitive status examination (NCSE) in assessment of psychiatric patients. Journal of Psychiatric Research, 28, 85-95.

40.

Mulsant

B. H.

Pollock

B. G.

Kirshner

Shen

C. Y.

Dodge

Ganguli

(2003). Serum anticholinergic activity in a community-based sample of older adults—Relationship with cognitive performance. Archives of General Psychiatry, 60, 198-203.

41.

Mussi

Ferrari

Ascari

Salvioli

(1999). Importance of serum anticholinergic activity in the assessment of elderly patients with delirium. Journal of Geriatric Psychiatry and Neurology, 12, 82-86.

42.

Nanna

M. J.

Lichtenberg

P. A.

BudaAbela

Barth

J. T.

(1997). The role of cognition and depression in predicting functional outcome in geriatric medical rehabilitation patients. Journal of Applied Gerontology, 16, 120-132.

43.

Nebes

R. D.

Pollock

B. G.

Mulsant

B. H.

Kirshner

M. A.

Halligan

Zmuda

. (1997). Low-level serum anticholinergicity as a source of baseline cognitive heterogeneity in geriatric depressed patients. Psychopharmacology Bulletin, 33, 715-719.

44.

Ness

Hoth

Barnett

Shorr

Kaboli

(2006). Anticholinergic medications in community-dwelling older veterans: Prevalence of anticholinergic symptoms, symptom burden, and adverse drug events. American Journal of Geriatric Pharmacotherapy, 4, 42-51.

45.

Paton

Lelliott

Harrington

Okocha

Sensky

Duffett

(2003). Patterns of antipsychotic and anticholinergic prescribing for hospital inpatients. Journal of Psychopharmacology, 17, 223-229.

46.

Pomara

Belzer

Hernando

De la Pena

Sidtis

J. J.

(2008). Increased mental slowing associated with the APOE epsilon 4 allele after trihexyphenidyl oral anticholinergic challenge in healthy elderly. American Journal of Geriatric Psychiatry, 16, 116-124.

47.

Rajska-Neumann

Wieczorowska-Tobis

(2007). Polypharmacy and potential inappropriateness of pharmacological treatment among community-dwelling elderly patients. Archives of Gerontology and Geriatrics, 44, 303-309.

48.

Rudd

K. M.

Raehl

C. L.

Bond

C. A.

Abbruscato

T. J.

Stenhouse

A. C.

(2005). Methods for assessing drug-related anticholinergic activity. Pharmacotherapy, 25, 1592-1601.

49.

Rudolph

J. L.

Salow

M. J.

Angelini

M. C.

McGlinchey

R. E.

(2008). The anticholinergic risk scale and anticholinergic adverse effects in older persons. Archives of Internal Medicine, 168, 508-513.

50.

Shapiro

Benedict

Schretlen

Brandt

(1999). Construct and concurrent validity of the Hopkins Verbal Learning Test-Revised. The Clinical Neuropsychologist, 13, 348-358.

51.

Sheikh

Yesavage

(1986). Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. In Brink

(Ed.), Clinical gerontology: A guide to assessment and intervention (pp. 165-173). New York: Haworth Press.

52.

Snitz

B. E.

Bieliauskas

L. A.

Crossland

Basso

M. R.

Roper

(2000). PPVT-R as an estimate of premorbid intelligence in older adults. Clinical Neuropsychologist, 14, 181-186.

53.

Starr

Lonie

(2007). The influence of premorbid IQ on minimental state examination score at time of dementia presentation. International Journal of Geriatric Psychiatry, 22, 382-384.

54.

Starr

Whalley

Inch

Shering

(1992). The quantification of the relative effects of age and NART-predicted IQ on cognitive function in healthy old people. International Journal of Geriatric Psychiatry, 7, 153-157.

55.

Tombaugh

McIntyre

(1992). The Mini-Mental State Examination: A comprehensive review. Journal of the American Geriatric Society, 40, 922-935.

56.

Tune

L. E.

(2001). Anticholinergic effects of medication in elderly patients. Journal of Clinical Psychiatry, 62, 11-14.

57.

Tune

L. E.

Holland

Folstein

M. F.

Damlouji

N. F.

Gardner

T. J.

Coyle

J. T.

(1981). Association of postoperative delirium with raised serum levels of anticholinergic drugs. Lancet, 2, 651-653.

58.

van den Bemt

P. M. L. A.

Egberts

A. C. G.

Lenderink

A. W.

Verzijl

J. M.

Simons

K. A.

van der Pol

W. S. C. J.

. (1999). Adverse drug events in hospitalized patients—A comparison of doctors, nurses and patients as sources of reports. European Journal of Clinical Pharmacology, 55, 155-158.

59.

van Eijk

M. E. C.

Bahri

Dekker

Herings

R. M. C.

Porsius

Avorn

. (2000). Use of prevalence and incidence measures to describe age-related prescribing of antidepressants with and without anticholinergic effects. Journal of Clinical Epidemiology, 53, 645-651.

60.

van Marwijk

Wallace

de Bock

Hermans

Kaptein

Mulder

(1995). Evaluation of the feasibility, reliability and diagnostic value of shortened versions of the geriatric depression scale. British Journal of General Practice, 45, 195-199.

61.

Whitney

K. A.

Maoz

Hook

J. N.

Steiner

A. R.

Bieliauskas

L. A.

(2007). IQ and scores on the minimental state examination (MMSE): Controlling for effort and education among geriatric inpatients. Aging, Neuropsychology, and Cognition, 14, 545-552.

62.

Zillich

A. J.

Shay

Hyduke

Emmendorfer

T. R.

Mellow

A. M.

Counsell

S. R.

. (2008). Quality improvement toward decreasing high-risk medications for older veteran outpatients. Journal of the American Geriatrics Society, 56, 1299-1305.