Archives of Clinical Neuropsychology 30 (2015) 105– 113
The Interaction Between Medical Burden and Anticholinergic Cognitive Burden on Neuropsychological Function in a Geriatric Primary Care Sample Cady K. Block1, Erin Logue1, Nicholas S. Thaler2, David M. Scarisbrick1, James J. Mahoney, III 1, James Scott1,*, Kevin Duff3 1
*Corresponding author at: Tel.: +1-405-271-5253; Fax: +1-405-271-8802; E-mail address: [email protected]. Accepted 3 December 2014
Abstract Poorer neuropsychological function is associated with increased medical burden (MB) and the use of more anticholinergic medications. However, the interaction between MB and anticholinergic cognitive burden (AB) on neuropsychological performance is unknown. In a sample of 290 elderly primary care patients, those with a greater level of AB demonstrated poorer Total Index performance on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Further, an interaction effect was noted such that there was a greater anticholinergic effect on RBANS Total, Attention, and Delayed Memory Index scores for participants with fewer MB. Participants with more MB demonstrated poorer performance irrespective of their level of AB. These results indicate that MB effects may be overshadowed by anticholinergic effects in older patients. Keywords: Geriatrics; Attention; Assessment
The number of chronic health conditions per person is rapidly rising in the United States. By 2030, the number of Americans with one or more chronic conditions is projected to increase by 37% (Anderson, 2010). Some of the most common chronic health conditions facing our society include high blood pressure, dementia, diabetes, heart disease, arthritis, and COPD (Anderson, 2010). Presently, approximately one in four Americans is diagnosed as having multiple chronic health conditions, which is defined as more than two chronic conditions (Anderson, 2010). Moreover, a growing number of Americans have five or more chronic medical conditions, including 6% of females and 4% of males (Anderson, 2010). In our nation’s Medicare system, the statistics are even greater; 67.3% of beneficiaries have two or more chronic medical conditions with an additional 14% having six or more conditions (Lochner, Goodman, Posner, & Parekh, 2013). It appears that older Americans are disproportionately affected, with as many as 75% of adults aged 65 and older have multiple chronic medical conditions. That number will likely only increase over time as our older population ages (Anderson, 2010; U.S. Census Bureau, 2008). This increased number of chronic medical conditions, or medical burden (MB), has been shown to be associated with a concomitant increase in the risk of hospitalizations, functional limitations and disability, and mortality (Anderson, 2010). As noted by Duff, Mold, Roberts, and McKay (2007), MCCs have been shown to have a deleterious and cumulative effect on neuropsychological functioning across a number of geriatric health populations. Duff and colleagues (2007) examined neuropsychological performance in a group of 690 community-dwelling geriatric primary care patients and, after adjusting for age and gender, observed a strong inverse relationship between MB and neuropsychological functioning. In particular, attention was noted to be selectively worse for older adults diagnosed with five or more MCCs. Other studies of geriatric inpatients and outpatients examined the relationship between MB and performance on selected measures of cognitive function, and discovered an inverse relationship (e.g., D-KEFS Trails motor speed, Karp et al., 2006; Mattis Dementia Rating Scale-2 total score, MacNeill, Lichtenberg, & LaBuda, 2000). More broadly, a 2005 population-based study indicated that older adults diagnosed with cognitive # The Author 2014. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. doi:10.1093/arclin/acu073 Advance Access publication on 30 December 2014
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Department of Psychiatry and Behavioral Sciences, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73126-0901, USA 2 UCLA Semel Institute for Neuroscience and Human Behavior, Los Angeles, CA 90095-1759, USA 3 The University of Utah, Salt Lake City, UT 84108, USA
106
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Fig. 1. A sample listing of medications with anticholinergic effects.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
impairment-not dementia (CIND) or dementia were significantly more likely to experience significant medical comorbidity, which in turn was associated with worse daily functioning and lower mini-mental status examination scores (MMSE; Lyketsos et al., 2005). Other studies have also reported that individuals with a greater number of comorbid medical conditions demonstrated larger declines in cognition (Kutlay, 2001). In individuals whom underwent coronary artery bypass grafting surgery, greater cognitive dysfunction was noted at 7 days and 6 months post-surgery in those patients diagnosed with both diabetes and coronary artery disease (CAD), when compared with individuals diagnosed with CAD alone (Kadoi, 2005). Older adults with MCCs in particular are also prone to polypharmacy. This presents its own risks including the prescription of potentially inappropriate medications that can result in poorer health status, increased risk for harmful medication interactions, and increased exposure to acute and chronic adverse medication effects (Fu, Liu, & Christensen, 2004; Yeh, Liu, Peng, Lin, & Chen, 2013). For example, medications with anticholinergic effects (see Fig. 1) are especially known to be associated with a number of adverse cognitive effects including reductions in attention/concentration, learning, and memory (Kersten & Wyller, 2014; Ray et al., 1992). That being said, Kersten and Wyller (2014) also reported that, after anticholinergic drug burden was reduced, no significant cognitive improvements have been demonstrated in randomized controlled trials, thus indicating that cognitive impairment may be prolonged well past the period of drug discontinuation and/or other factors may be involved in contributing to these deficits. Indeed, studies evaluating the use of these medications in the geriatric population have identified objective neuropsychological deficits that appear to increase with the number of medications, including general cognitive decline, increased proneness to errors, reductions in psychomotor speed, and executive dysfunction (e.g., Boustani, Campbell, Munger, Maidment, & Fox, 2008; Campbell et al., 2009; Pasina et al., 2013). These changes may be more substantial for older adults given the age-related changes in central cholinergic receptor function, blood – brain barrier integrity, and slowed medication metabolism/elimination (Koyoma, Steinman, Ensrud, Hillier, & Yaffe, 2014; Peters, 1989). Of note, older adults who used medications with anticholinergic effects for as little as 90 days were 2.73 times more likely to have a diagnosis of mild cognitive impairment (MCI), and have an increased likelihood of amyloid and neurofibrillary neuropathology with a 2-year duration of use (Cai, Campbell, Khan, Callahan, & Boustani, 2013; Perry, Kilford, Lees, Burn, & Perry, 2003). The present study sought to examine the interaction between the potential neuropsychological burden imposed by multiple chronic health conditions (MB) and by medications with anticholinergic effects (anticholinergic cognitive burden, AB) in a sample of community-dwelling geriatric primary care patients. We hypothesized that increases in both MB and AB would interact to produce adverse effects on neuropsychological function above and beyond the effects of age, education, gender, and ethnicity. Given that older adults with neuropsychological dysfunction are at risk for poorer medical compliance and other potentially adverse outcomes in trying to manage complex medical regimens and activities of daily living, it is hoped that research considering
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
107
these factors will result in positive local (e.g., improved patient care, health, and quality of living) and systemic (e.g., reduced healthcare costs) effects.
Methods Study Sample
Table 1. Characteristics of the sample (N ¼ 290) M (SD) Age Gender Female Male Ethnicity Caucasian African American Native American descent Hispanic/Latino Level of education College degree Some college High school graduate Some high school ,8th grade
% (n)
72.76 + 5.47 43.8% (n ¼ 127) 56.2% (n ¼ 163) 88.6% (n ¼ 257) 8.6% (n ¼ 25) 1.7% (n ¼ 5) 1.0% (n ¼ 3) 13.1% (n ¼ 38) 34.5% (n ¼ 100) 22.4% (n ¼ 65) 5.5% (n ¼ 16) 2.4% (n ¼ 7)
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Data in the present study represent a subsample of the Oklahoma Longitudinal Assessment of Health Outcomes in Mature Adults (OKLAHOMA) studies cohort (Duff et al., 2003, 2007), whereby participants were recruited from the practices of primary care family physicians throughout central Oklahoma with the goal of assessing the effect of various symptoms, medical conditions, and health care delivery qualities on health outcomes across time. In this cohort, eligible individuals were identified from billing records and required to meet the following criteria: first, 65 years or older; second, a clinical visit in the past 18 months; and third, sufficient awareness and orientation as determined by a physician to complete informed consent. General demographic data are presented in Table 1. As discussed in the recruitment procedures and outlined in Fig. 2, we excluded from statistical analyses individuals with over neurologic conditions (i.e., concussion, head injury, seizures, stroke/transient ischemic attack (TIA)/RIND, Parkinson disease, and/or brain hemorrhage), and demographic data do not include these individuals. In an attempt to maintain study generalizability and feasibility, we elected to keep participants diagnosed with conditions with potential neurologic sequelae (e.g., depression and diabetes). Our resultant actual sample included the following conditions: anemia (n ¼ 40, 13.8%), arthritis (n ¼ 177, 61%), asthma (n ¼ 16, 5.5%), cancer (not encephalitic; n ¼ 51, 17.6%), hepatitis (n ¼ 5, 1.7%), depression (n ¼ 49, 16.9%), diabetes (n ¼ 46, 15.9%), diverticulitis (n ¼ 44, 15.2%), emphysema (n ¼ 11, 3.8%), glaucoma (n ¼ 25, 8.6%), hearing loss (n ¼ 85, 29.3%), heart disease (n ¼ 81, 27.9%), hereditary neuropathy (n ¼ 2, 0.7%), hiatal hernia (n ¼ 41, 14.1%), high blood cholesterol (n ¼ 105, 36.6%), hypertension (n ¼ 133, 45.9%), kidney/bladder stones (n ¼ 25, 8.6%), liver problems (n ¼ 7, 2.4%), lupus (n ¼ 5, 1.7%), macular degeneration (n ¼ 20, 6.9%), osteoarthritis (n ¼ 63, 21.7%), osteoporosis (n ¼ 24, 8.3%), peripheral neuropathy (n ¼ 6, 2.1%), rheumatoid arthritis (n ¼ 21, 7.2%), sarcoidosis (n ¼ 2, 0.7%), scleroderma (n ¼ 1, 0.3%), sinus problems (n ¼ 62, 21.4%), stomach or duodenal ulcers (n ¼ 39, 13.4%), surgical removal of large intestine (n ¼ 9, 3.1%), surgical removal of part of stomach or small intestine (n ¼ 4, 1.4%), thyroid problems (n ¼ 52, 17.9%), frequent UTI (n ¼ 21, 7.2%), and vitamin B-12 deficiency (n ¼ 15, 5.2%). In addition to MB, medication data were collected for all participants in the study. The average number of medications people were taking, including those that were anticholinergic was 5.27, with a range of 1– 14 medications per person. Table 2 shows data on the percentage of participants prescribed the top 10 most prescribed medications in the database not included on the ACB scale. No participants were taking anticholinesterase inhibitors. In our sample, 45.2% were taking anticholinergic medications at the time of the study. The average ACB score for the sample was 1.1 (SD ¼ 1.6).
108
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Table 2. Percentage of patients prescribed the top 10 most prescribed medications not on ACB scale Medication
Participants (%)
Levothyroxine Atorvastatin Conjugated estrogens Lisinopril Simvastatin Alendronate Hydrochlorothiazide Amlodipine Celecoxib Rofecoxib
3.8 3.1 3.1 2.2 2.2 1.8 1.6 1.4 1.4 1.4
Materials and Procedures All procedures underwent review and approval by the Institutional Review Boards at all institutions involved in the study. Recruitment is detailed in Fig. 2, with a total sample size of 290 participants. Individuals who agreed to participate were asked to complete a questionnaire mailed to them 2 weeks prior to the study enrollment visit. Although many more individuals completed the background and neuropsychological measures as part of a larger study (i.e., Duff et al., 2003), only 290 possessed completed medication data and were included for analysis in the present study. Participants completed a demographics questionnaire and medical history questionnaire that assessed past and/or current diagnosis of 48 possible medical conditions including: acromegaly; AIDS; amyloidosis; anemia; arthritis; asthma; brain hemorrhage; cancer; hepatitis B or C; concussion; Crohn’s disease; depression; diabetes; diverticulosis/itis; emphysema; glaucoma; head injury; hearing loss; heart disease; hereditary neuropathy; hiatal hernia; hypercholesteremia; hypertension; kidney/bladder stones; leprosy; liver problems; lupus; macular degeneration; multiple myeloma; neurofibromatosis; osteoarthritis; osteoporosis;
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Fig. 2. Recruitment and enrollment procedures for the study sample (n ¼ 290).
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
109
Parkinson disease; peripheral neuropathy; porphyria; rheumatoid arthritis; sarcoidosis; scleroderma; seizures; sinus problems; stomach/duodenal ulcers; stroke, TIA, or reversible ischemic neurologic deficit; surgical removal of large intestines/colon or small intestines/stomach; thyroid problems; frequent urinary tract infections; vitamin B12 deficiency; and/or Whipple’s disease. Participants then completed Form A of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS; Randolph, 1998). Statistical Analyses
Results Index Analysis The correlation between MB and AB is 0.27 which indicates that these two predictors do not substantially overlap in variance. The first regression model examined the RBANS Total Scale score. When the two predictors and their interaction were entered into the equation, the overall equation was significant (adjusted R2 ¼ .05, p , .01). Significant predictors included the AB score (b ¼ 20.23, p , .01) and the MB × AB interaction score (b ¼ 0.16, p ¼ .01). As demonstrated in the two-way interaction in Fig. 3, individuals possessing a lower MB score but higher AB score demonstrated the poorest overall RBANS performance, suggesting increased sensitivity to the potential cognitive effects of their prescribed medication regimen. Low MB/low AB participants exhibited the best RBANS Total Index performance than those prescribed a greater number of medications. Individuals with a high MB appeared less affected by increased AB. A series of regressions were next performed on each of the RBANS indexes to better identify which indexes were most affected by the observed relationship. See Table 3 for results. The AB score predicted RBANS scores across all indexes. Further, analyses of the Attention and Delayed Memory indexes revealed significant MB × AB interactions. For both indexes, a similar pattern emerged as identified with the RBANS Total Scale score such that individuals with a lower MB were significantly sensitive to the burden imposed by medications with anticholinergic cognitive effects; however, those with a higher MB demonstrated poorer performance irrespective of their level of AB. Subtest Analysis Post hoc analyses were performed for each of the following 12 RBANS subtests: List Learning, Story Memory, Figure Copy, Line Orientation, Picture Naming, Semantic Fluency, Digit Span, Coding, List Recall, List Recognition, Story Recall, and Figure Recall. Overall models were significant for the List Learning, Story Memory, Line Orientation, Coding, and Story Recall subtests. In each case, greater AB predicted poorer performance on the respective RBANS subtest score (p , .01 in all cases). In addition, a significant AB main effect and MB × AB interaction effect was revealed for the Digit Span subtest. Both
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Two predictor variables were examined in this study. The first variable was the number of self-reported medical conditions at baseline (i.e., MB). As noted above, MB variable has been previously demonstrated to have an inverse relationship with neuropsychology function (Duff et al., 2007). The number of self-reported medical conditions was summed to yield a total MB score for all statistical analyses. The second predictor variable was the AB of participants’ individual medication regimens. This variable was measured using the Anticholinergic Cognitive Burden (ACB) Scale developed by the Indiana University Center for Aging Research (Campbell et al., 2009; Boustani et al., 2008). The ACB scale identifies the severity of anticholinergic effects of prescription and over-the-counter medications divided into “possible” (score of 1) and “definite” categories (score of 2 or 3, depending upon the strength of the effect). Each medication is scored, and all medication scores are subsequently summed to result in a total AB score for that particular individual. For the purposes of this paper, we were interested in analyzing the effect of anticholinergic medications only. As both the MB and AB variables are continuous, we employed a series of moderated multiple linear regressions to examine their main and interactive effects on cognition. Following convention, predictors were first centered around their means to eliminate multicollinearity (Tabachnick & Fiddell, 2007). The predictor variables were then entered into the first block of the regression model, while the interaction of the two predictor variables was entered into the second block. The interaction variable was obtained by multiplying the value of the two predictor variables together. The initial criterion variable was the RBANS Total Scale score. Significant two-way interactions were plotted using R-code syntax. To better identify which domains were most affected by the interaction between MB and AB, secondary exploratory analyses were conducted on the five RBANS indexes. Significant findings were further evaluated at the subtest level.
110
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
the List Recall and List Recognition subtests had significant MB × AB interaction effects. These interactions followed the pattern of the RBANS total score and indexes, such that individuals with low MB had more sensitivity to AB, while those with high MB had poorer scores regardless of their AB. No significant overall or main effects were identified for the other RBANS subtests.
Discussion The purpose of the present study was to examine the interaction between the MB and AB posed by participants’ self-reported medication regimen on neuropsychological function in a sample of community-dwelling geriatric outpatients. Our findings revealed an interaction between MB and AB for selected cognitive variables. Surprisingly, however, only participants possessing a lower MB appeared to be more sensitive to increased AB given the observed reductions in RBANS Total Scale score, as well as index scores in the domains of Attention and Delayed Memory. This extends Duff and colleagues (2007) finding that decreased attention was the primary area of concern in populations with MCCs, and is consistent with previous findings that use of anticholinergic medications in older adults can decrease cognition (Boustani et al., 2008; Campbell et al., 2009). When post hoc analyses were performed, the interactions on these indexes were found to be driven primarily by Digit Span, List Recall, and List Recognition subtests. In summary, although MB and AB each independently contributed to poorer attention and delayed memory, increased AB overshadowed the effects of increased MB as indicated by total RBANS index scores (Fig. 3). Although the reasons for this surprising finding are unclear, one potential explanation may be that an increasing number of medical conditions possess a greater synergistic effect on neuropsychological function than does consuming an increased number of medications possessing anticholinergic cognitive effects. However, our results may also have been influenced by the duration of medication use. Indeed, research has demonstrated that long-term consumers of anticholinergics potentially
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Fig. 3. Multiple linear regression two-way interaction of medical burden (MB) and anticholinergic cognitive burden on RBANS Total Index Scores. ACB ¼ anticholinergic cognitive burden; Med ¼ medical burden. Ranges for both variables are represented in z-score values. For MB: “High Med” z-score ¼ +1.00, “Medium Med” z-score ¼ 0.00, and “Low Med” z-score ¼ 21.00. The anticholinergic cognitive burden score range is also represented by a range of z-score values.
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
111
Table 3. Moderated multiple regression analyses for the RBANS total and index scores b
t
p
.05
— 20.02 20.23 20.16 — 20.00 20.21 20.11
— 20.39 23.71 22.61 — 20.06 23.41 21.79
20.07 20.20 20.08 — 20.02 20.01 20.11 — 20.07 20.27 20.17 — 20.02 20.14 20.15
21.20 23.17 21.23 — 20.32 20.20 21.74 — 21.23 24.47 22.74 — 20.26 22.19 22.40
,.01 .70 ,.01 .01 ,.01 .95 ,.01 .07 ,.01 .23 ,.01 .22 .33 .75 .84 .08 ,.01 .22 ,.01 ,.01 ,.05 .79 ,.05 ,.05
.04
.04
.00
.08
.02
Note: RBANS ¼ Repeatable Battery for the Assessment of Neuropsychological Status; ACB ¼ anticholinergic cognitive burden; MB ¼ medical burden.
show less of a decline in neuropsychological functioning over time than do novel or incident users of these medications (Shah et al., 2013). It stands to reason that participants with greater MB may have been on these medications longer. However, since data regarding the duration of medication use was not available in this particular sample, we feel that this is a critical area for further investigation. In addition, data were not available on medication dosing in order to evaluate whether a dose –response relationship exists. An important implication of our findings is how they apply to the concept of diminishing returns in medicine. Mold, Hamm, and McCarthy (2010) posit that the additive influence of an increasing number of medical interventions is inversely related to the maximal intervention benefit. Furthermore, there is a concomitant increase in the cost– benefit ratio of a given medical interventions. These data suggest that increasing the number of medications with anticholinergic cognitive effects has the subsequent potential to do more harm than good in older adults with fewer medical conditions. Fortunately, efforts to increase provider awareness about appropriate medication prescription practices for older adults already exist (e.g., the Beers List; Fick et al., 2003). Programs to evaluate anticholinergic burden and help to decrease its effects in the geriatric population have also been suggested (see He and Ball, 2013; Karimi, Dharia, Floria, & Slattum, 2012). Thus, it would be beneficial to include neuropsychologists in these efforts as well as in future endeavors to evaluate neuropsychological functioning as it relates to anticholinergic burden. We demonstrated a relatively simple way that neuropsychologists could accomplish this in both research and daily clinical practice (i.e., by measuring the number of medical conditions and rating medications with the ACB scale). Projects similar to the Pasina et al. (2013) study could benefit from the use of more detailed neuropsychological measures. Our results are limited by several issues. First, to be able to examine medication effects, we were required to exclude two thirds of the original OKLAHOMA dataset. It is unfortunate that so many participants did not have this data available, and certainly future comparisons should include some record of medications as well as data on dosage, duration of use, and stability of use. The nature of the findings gleaned from correlational research is problematic as well. Given the cross-sectional nature of this study, it is unknown which came first: lower cognitive functioning or increasing MB and use of anticholinergic drugs. In fact, a 2005 population-based study indicated that older adults diagnosed with CIND or dementia were significantly more likely to experience significant medical comorbidity, which in turn was associated with worse daily functioning and lower MMSE scores (Lyketsos et al., 2005). We recommend that more research of a longitudinal nature be conducted in this area. Furthermore, the sample employed was fairly homogenous; i.e., they were primarily white, well-educated, and relatively young older adults who lived independently within the community and were without major cognitive impairment. Thus, the applicability of these findings to samples with more diverse characteristics is not yet known and should be examined.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
RBANS Total Index MB ACB MB × ACB RBANS Immediate Memory Index MB ACB MB × ACB RBANS Visuospatial/Constructional Index MB ACB MB × ACB RBANS Language Index MB ACB MB × ACB RBANS Attention Index MB ACB MB × ACB RBANS Delayed Memory Index MB ACB MB × ACB
Adjusted R 2
112
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Conflict of Interest None declared.
References Anderson, G. (2010). Chronic care: Making the case for ongoing care. Princeton, NJ: Robert Woods Johnson Foundation. Retrieved February 24, 2014, from http ://www.rwjf.org/en/research-publications/find-rwjf-research/2010/01/chronic-care.html. Boustani, M. A., Campbell, N. L., Munger, S., Maidment, I., & Fox, G. C. (2008). Impact of anticholinergics on the aging brain: A review and practical application. Aging Health, 4, 311–320. Cai, X., Campbell, N., Khan, B., Callahan, C., & Boustani, M. (2013). Long-term anticholinergic use and the aging brain. Alzheimer‘s & Dementia, 9, 377– 385. Campbell, N., Boustani, M., Limbil, T., Ott, C., Fox, C., Maidment, I., et al. (2009). The cognitive impact of anticholinergics: A clinical review. Clinical Interventions in Aging, 4, 225–233. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40, 373 –383. Duff, K., Mold, J. W., Roberts, M. M., & McKay, S. (2007). Medical burden and cognition in older patients in primary care: Selective deficits in attention. Archives of Clinical Neuropsychology, 22, 569–575. Duff, K., Patton, D., Schoenberg, M. R., Mold, J., Scott, J. G., & Adams, R. L. (2003). Age- and education-corrected independent normative data for the RBANS in a community dwelling elderly sample. Clinical Neuropsychology, 17, 351– 366. Fick, D. M., Cooper, J. W., Wade, W. E., Waller, J. L., Maclean, J. R., & Beers, M. H. (2003). Updating the beers criteria for potentially inappropriate medication use in older adults. Archives of Internal Medicine, 163, 2716–2724. Fu, A. Z., Liu, G. G., & Christensen, D. B. (2004). Inappropriate medication use and health outcomes in the elderly. Journal of the American Geriatric Society, 52, 1934– 1939. He, Z., & Ball, P. A. (2013). Can medication management review reduce anticholinergic burden (AB) in the elderly? Encouraging results from a theoretical model. International Psychogeriatrics, 25, 1425–1431.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Some of the measures used constitute another limitation. The OKLAHOMA health questionnaire lacks the inclusion of some conditions such as dementia and MCI. However, dementia was not a primary concern as the study explicitly excluded these individuals from participation. More specifically, the study was designed to study long-term health outcomes in a sample of independently living community-dwelling geriatric individuals. Participants, along with being able to provide informed consent, were required to be functionally independent to be in the study. Although formal neuropsychological testing beyond the RBANS is not available, these patients were well above the level of impairment that would characterize dementia during their initial evaluation. It is certainly possible that some participants were on the cusp of MCI that did not necessarily interfere yet with ADL or IADL performance. However, we elected not to outright exclude these individuals because MCI itself is a heterogeneous condition that is often prevalent in this population, and removing all cases that met some criteria (e.g., score ,85 on at least one index) would limit the generalizability of findings. Without signs of actual functional impairment, it is also difficult to ascertain whether MCI for a specific individual actually represents the beginning of cognitive decline or is simply a transient phenomenon that resolves on later testing. We elected to remove neurological conditions with known pathophysiology and a co-occurring prototypical neurological profile (e.g., Parkinson’s disease) because these disorders impair cognition in a predictable manner. However, in an attempt to maintain study generalizability and feasibility, we opted to let participants diagnosed with conditions with potential neurologic sequelae (e.g., depression and diabetes) remain. The RBANS is a brief measure of cognition that is particularly limited in its ability to assess executive functioning, and MB was determined by self-report via a questionnaire that was neither standardized nor exhaustive. Future research in this area should assess cognition and MB more comprehensively by using more in-depth measures of neuropsychological functioning as well as a validated, weighted questionnaire of MB (e.g., the Index of Comorbidity; Charlson, Pompei, Ales, & MacKenzie, 1987). Similarly, the potential association between cognition and other medications participants were prescribed was not addressed. Given the potential for total medication cognitive burden, we recommend this as an area for future study. Finally, only a small proportion of the variance (2% to 8%) in performance on the RBANS was explained by the main effects of an interaction between MB and AB. This leaves a significant amount of the variance in cognition as measured by the RBANS unexplained. Even when taking these limitations into account, we consider the findings of the present study to be important. To our knowledge, this is the first study to date to examine the complex relationship between MB and AB and their potential influence on neuropsychological performance. Statistics for the financial impact of MCCs on our healthcare system are striking (Lochner et al., 2013); 85% of U.S. healthcare spending dollars are funneled towards individuals with MCCs (Anderson, 2010; Lochner et al., 2013). It will become more important than ever to delineate the burden imposed by MCCs and the medications designed to treat these conditions due to the advent of the ACA and its allowing an increased catchment of older Americans into the U.S. healthcare system.
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
113
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Kadoi, Y., Saito, S., & Fujita, N. (2005). Risk factors for cognitive dysfunction after coronary artery bypass graft surgery in patients with type 2 diabetes. Journal of Thoracic and Cardiovascular Surgery, 3, 576– 593. Karimi, S., Dharia, S. P., Floria, D. S., & Slattum, P. W. (2012). Anticholinergic burden: Clinical implications for seniors and strategies for clinicians. Consult Pharmacy, 27, 564 –82. Karp, J. F., Reynolds, C. F., Butters, M., Dew, M. A., Mazumdar, S., & Begley, A. E., et al. (2006). Correlates of neuropsychological impairment in older adult pain clinic patients. Pain Medicine, 7, 444– 452. Kersten, H., & Wyller, T. B. (2014). Anticholinergic drug burden in older people’s brain—How well is it measured? Basic & Clinical Pharmacology & Toxicology, 114, 151– 159. Koyoma, A., Steinman, M., Ensrud, K., Hillier, T. A., & Yaffe, K. (2014). Long-term cognitive and functional effects of potentially inappropriate medications in older women. The Journals of Gerontology Series A: Biological Sciences & Medical Sciences, 69, 423 –9. Kutlay, S., Nergizoglu, G., Duman, N., Atli, T., Keven, K., Eturk, S., et al. (2001). Recognition of Neurocognitive Dysfunction in Chronic Hemodialysis Patients. Renal Failure, 23, 781–787. Lochner, K. A., Goodman, R. A., Posner, S., & Parekh, A. (2013). Multiple chronic conditions among Medicare beneficiaries: State-level variations in prevalence, utilization, and cost, 2011. Medicare & Medicaid Research Review, 3, E1– E19. Lyketsos, C. G., Toone, L., Schanz, J., Rabins, P. V., Steinberg, M., Onyike, C. U., et al. (2005). Population-based study of medical comorbidity in early dementia and “cognitive impairment, no dementia (CIND):” Association with functional and cognitive impairment: The Cache County Study. American Journal of Geriatric Psychiatry, 13, 656 –664. MacNeill, S. E., Lichtenberg, P. A., & LaBuda, J. (2000). Factors affecting return to living alone after medical rehabilitation: A cross-validation study. Rehabilitation Psychology, 45, 356–364. Mold, J. W., Hamm, R. M., & McCarthy, L. H. (2010). The law of diminishing returns in clinical medicine: How much risk reduction is enough? Journal of the American Board of Family Medicine, 23, 371–375. Pasina, L., Djade, C. D., Lucca, U., Nobili, A., Tettamanti, M., Franchi, C., et al. (2013). Association of anticholinergic burden with cognitive and functional status in a cohort of hospitalized elderly: Comparison of the Anticholinergic Burden Scale and Anticholinergic Risk Scale. Drugs and Aging, 30, 103– 112. Perry, E. K., Kilford, L., Lees, A. J., Burn, D. J., & Perry, R. H. (2003). Increased Alzheimer pathology in Parkinson’s disease related to antimuscarinic drugs. Annals of Neurology, 54, 235–238. Peters, N. L. (1989). Snipping the thread of life: Antimuscarinic side effects of medications in the elderly. Archives of Internal Medicine, 149, 2414–2420. Randolph, C. (1998). Repeatable battery for the assessment of neuropsychological status (RBANS). San Antonio, Harcourt, TX: The Psychological Corporation. Ray, P. G., Meador, K. J., Loring, D. W., Zamrini, E. W., Yang, X. H., & Buccasfusco, J. J. (1992). Central anticholinergic hypersensitivity in aging. Journal of Geriatric Psychiatry & Neurology, 5, 72– 77. Shah, R. C., Janos, A. L., Kline, J. E., Yu, L., Leurgeans, S. E., & Wilson, R. S., et al. (2013). Cognitive decline in older persons initiating anticholinergic medications. PLoS ONE, 8, doi: 10.1371/journal.pone.0064111. Tabachnick, B. G., & Fiddell, L. S. (2007). Using Multivariate Statistics (5th ed.). Boston: Allyn and Bacon. U.S. Bureau of the Census. (2008). Projected Population by Single Year of Age, Sex, Race, and Hispanic Origin for the United States: July 1, 2000 to July 1, 2050. Washington. Retrieved February 24, 2014, from http://www.census.gov/population/projections/data/national/2008/downloadablefiles.html. Yeh, Y. C., Liu, C. L., Peng, L. N., Lin, M. H., & Chen, L. K. (2013). Potential benefits of reducing medication-related anticholinergic burden for demented older adults: A prospective cohort study. Geriatrics and Gerontology International, 13, 694–700.
The Interaction Between Medical Burden and Anticholinergic Cognitive Burden on Neuropsychological Function in a Geriatric Primary Care Sample Cady K. Block1, Erin Logue1, Nicholas S. Thaler2, David M. Scarisbrick1, James J. Mahoney, III 1, James Scott1,*, Kevin Duff3 1
*Corresponding author at: Tel.: +1-405-271-5253; Fax: +1-405-271-8802; E-mail address: [email protected]. Accepted 3 December 2014
Abstract Poorer neuropsychological function is associated with increased medical burden (MB) and the use of more anticholinergic medications. However, the interaction between MB and anticholinergic cognitive burden (AB) on neuropsychological performance is unknown. In a sample of 290 elderly primary care patients, those with a greater level of AB demonstrated poorer Total Index performance on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Further, an interaction effect was noted such that there was a greater anticholinergic effect on RBANS Total, Attention, and Delayed Memory Index scores for participants with fewer MB. Participants with more MB demonstrated poorer performance irrespective of their level of AB. These results indicate that MB effects may be overshadowed by anticholinergic effects in older patients. Keywords: Geriatrics; Attention; Assessment
The number of chronic health conditions per person is rapidly rising in the United States. By 2030, the number of Americans with one or more chronic conditions is projected to increase by 37% (Anderson, 2010). Some of the most common chronic health conditions facing our society include high blood pressure, dementia, diabetes, heart disease, arthritis, and COPD (Anderson, 2010). Presently, approximately one in four Americans is diagnosed as having multiple chronic health conditions, which is defined as more than two chronic conditions (Anderson, 2010). Moreover, a growing number of Americans have five or more chronic medical conditions, including 6% of females and 4% of males (Anderson, 2010). In our nation’s Medicare system, the statistics are even greater; 67.3% of beneficiaries have two or more chronic medical conditions with an additional 14% having six or more conditions (Lochner, Goodman, Posner, & Parekh, 2013). It appears that older Americans are disproportionately affected, with as many as 75% of adults aged 65 and older have multiple chronic medical conditions. That number will likely only increase over time as our older population ages (Anderson, 2010; U.S. Census Bureau, 2008). This increased number of chronic medical conditions, or medical burden (MB), has been shown to be associated with a concomitant increase in the risk of hospitalizations, functional limitations and disability, and mortality (Anderson, 2010). As noted by Duff, Mold, Roberts, and McKay (2007), MCCs have been shown to have a deleterious and cumulative effect on neuropsychological functioning across a number of geriatric health populations. Duff and colleagues (2007) examined neuropsychological performance in a group of 690 community-dwelling geriatric primary care patients and, after adjusting for age and gender, observed a strong inverse relationship between MB and neuropsychological functioning. In particular, attention was noted to be selectively worse for older adults diagnosed with five or more MCCs. Other studies of geriatric inpatients and outpatients examined the relationship between MB and performance on selected measures of cognitive function, and discovered an inverse relationship (e.g., D-KEFS Trails motor speed, Karp et al., 2006; Mattis Dementia Rating Scale-2 total score, MacNeill, Lichtenberg, & LaBuda, 2000). More broadly, a 2005 population-based study indicated that older adults diagnosed with cognitive # The Author 2014. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. doi:10.1093/arclin/acu073 Advance Access publication on 30 December 2014
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Department of Psychiatry and Behavioral Sciences, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73126-0901, USA 2 UCLA Semel Institute for Neuroscience and Human Behavior, Los Angeles, CA 90095-1759, USA 3 The University of Utah, Salt Lake City, UT 84108, USA
106
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Fig. 1. A sample listing of medications with anticholinergic effects.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
impairment-not dementia (CIND) or dementia were significantly more likely to experience significant medical comorbidity, which in turn was associated with worse daily functioning and lower mini-mental status examination scores (MMSE; Lyketsos et al., 2005). Other studies have also reported that individuals with a greater number of comorbid medical conditions demonstrated larger declines in cognition (Kutlay, 2001). In individuals whom underwent coronary artery bypass grafting surgery, greater cognitive dysfunction was noted at 7 days and 6 months post-surgery in those patients diagnosed with both diabetes and coronary artery disease (CAD), when compared with individuals diagnosed with CAD alone (Kadoi, 2005). Older adults with MCCs in particular are also prone to polypharmacy. This presents its own risks including the prescription of potentially inappropriate medications that can result in poorer health status, increased risk for harmful medication interactions, and increased exposure to acute and chronic adverse medication effects (Fu, Liu, & Christensen, 2004; Yeh, Liu, Peng, Lin, & Chen, 2013). For example, medications with anticholinergic effects (see Fig. 1) are especially known to be associated with a number of adverse cognitive effects including reductions in attention/concentration, learning, and memory (Kersten & Wyller, 2014; Ray et al., 1992). That being said, Kersten and Wyller (2014) also reported that, after anticholinergic drug burden was reduced, no significant cognitive improvements have been demonstrated in randomized controlled trials, thus indicating that cognitive impairment may be prolonged well past the period of drug discontinuation and/or other factors may be involved in contributing to these deficits. Indeed, studies evaluating the use of these medications in the geriatric population have identified objective neuropsychological deficits that appear to increase with the number of medications, including general cognitive decline, increased proneness to errors, reductions in psychomotor speed, and executive dysfunction (e.g., Boustani, Campbell, Munger, Maidment, & Fox, 2008; Campbell et al., 2009; Pasina et al., 2013). These changes may be more substantial for older adults given the age-related changes in central cholinergic receptor function, blood – brain barrier integrity, and slowed medication metabolism/elimination (Koyoma, Steinman, Ensrud, Hillier, & Yaffe, 2014; Peters, 1989). Of note, older adults who used medications with anticholinergic effects for as little as 90 days were 2.73 times more likely to have a diagnosis of mild cognitive impairment (MCI), and have an increased likelihood of amyloid and neurofibrillary neuropathology with a 2-year duration of use (Cai, Campbell, Khan, Callahan, & Boustani, 2013; Perry, Kilford, Lees, Burn, & Perry, 2003). The present study sought to examine the interaction between the potential neuropsychological burden imposed by multiple chronic health conditions (MB) and by medications with anticholinergic effects (anticholinergic cognitive burden, AB) in a sample of community-dwelling geriatric primary care patients. We hypothesized that increases in both MB and AB would interact to produce adverse effects on neuropsychological function above and beyond the effects of age, education, gender, and ethnicity. Given that older adults with neuropsychological dysfunction are at risk for poorer medical compliance and other potentially adverse outcomes in trying to manage complex medical regimens and activities of daily living, it is hoped that research considering
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
107
these factors will result in positive local (e.g., improved patient care, health, and quality of living) and systemic (e.g., reduced healthcare costs) effects.
Methods Study Sample
Table 1. Characteristics of the sample (N ¼ 290) M (SD) Age Gender Female Male Ethnicity Caucasian African American Native American descent Hispanic/Latino Level of education College degree Some college High school graduate Some high school ,8th grade
% (n)
72.76 + 5.47 43.8% (n ¼ 127) 56.2% (n ¼ 163) 88.6% (n ¼ 257) 8.6% (n ¼ 25) 1.7% (n ¼ 5) 1.0% (n ¼ 3) 13.1% (n ¼ 38) 34.5% (n ¼ 100) 22.4% (n ¼ 65) 5.5% (n ¼ 16) 2.4% (n ¼ 7)
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Data in the present study represent a subsample of the Oklahoma Longitudinal Assessment of Health Outcomes in Mature Adults (OKLAHOMA) studies cohort (Duff et al., 2003, 2007), whereby participants were recruited from the practices of primary care family physicians throughout central Oklahoma with the goal of assessing the effect of various symptoms, medical conditions, and health care delivery qualities on health outcomes across time. In this cohort, eligible individuals were identified from billing records and required to meet the following criteria: first, 65 years or older; second, a clinical visit in the past 18 months; and third, sufficient awareness and orientation as determined by a physician to complete informed consent. General demographic data are presented in Table 1. As discussed in the recruitment procedures and outlined in Fig. 2, we excluded from statistical analyses individuals with over neurologic conditions (i.e., concussion, head injury, seizures, stroke/transient ischemic attack (TIA)/RIND, Parkinson disease, and/or brain hemorrhage), and demographic data do not include these individuals. In an attempt to maintain study generalizability and feasibility, we elected to keep participants diagnosed with conditions with potential neurologic sequelae (e.g., depression and diabetes). Our resultant actual sample included the following conditions: anemia (n ¼ 40, 13.8%), arthritis (n ¼ 177, 61%), asthma (n ¼ 16, 5.5%), cancer (not encephalitic; n ¼ 51, 17.6%), hepatitis (n ¼ 5, 1.7%), depression (n ¼ 49, 16.9%), diabetes (n ¼ 46, 15.9%), diverticulitis (n ¼ 44, 15.2%), emphysema (n ¼ 11, 3.8%), glaucoma (n ¼ 25, 8.6%), hearing loss (n ¼ 85, 29.3%), heart disease (n ¼ 81, 27.9%), hereditary neuropathy (n ¼ 2, 0.7%), hiatal hernia (n ¼ 41, 14.1%), high blood cholesterol (n ¼ 105, 36.6%), hypertension (n ¼ 133, 45.9%), kidney/bladder stones (n ¼ 25, 8.6%), liver problems (n ¼ 7, 2.4%), lupus (n ¼ 5, 1.7%), macular degeneration (n ¼ 20, 6.9%), osteoarthritis (n ¼ 63, 21.7%), osteoporosis (n ¼ 24, 8.3%), peripheral neuropathy (n ¼ 6, 2.1%), rheumatoid arthritis (n ¼ 21, 7.2%), sarcoidosis (n ¼ 2, 0.7%), scleroderma (n ¼ 1, 0.3%), sinus problems (n ¼ 62, 21.4%), stomach or duodenal ulcers (n ¼ 39, 13.4%), surgical removal of large intestine (n ¼ 9, 3.1%), surgical removal of part of stomach or small intestine (n ¼ 4, 1.4%), thyroid problems (n ¼ 52, 17.9%), frequent UTI (n ¼ 21, 7.2%), and vitamin B-12 deficiency (n ¼ 15, 5.2%). In addition to MB, medication data were collected for all participants in the study. The average number of medications people were taking, including those that were anticholinergic was 5.27, with a range of 1– 14 medications per person. Table 2 shows data on the percentage of participants prescribed the top 10 most prescribed medications in the database not included on the ACB scale. No participants were taking anticholinesterase inhibitors. In our sample, 45.2% were taking anticholinergic medications at the time of the study. The average ACB score for the sample was 1.1 (SD ¼ 1.6).
108
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Table 2. Percentage of patients prescribed the top 10 most prescribed medications not on ACB scale Medication
Participants (%)
Levothyroxine Atorvastatin Conjugated estrogens Lisinopril Simvastatin Alendronate Hydrochlorothiazide Amlodipine Celecoxib Rofecoxib
3.8 3.1 3.1 2.2 2.2 1.8 1.6 1.4 1.4 1.4
Materials and Procedures All procedures underwent review and approval by the Institutional Review Boards at all institutions involved in the study. Recruitment is detailed in Fig. 2, with a total sample size of 290 participants. Individuals who agreed to participate were asked to complete a questionnaire mailed to them 2 weeks prior to the study enrollment visit. Although many more individuals completed the background and neuropsychological measures as part of a larger study (i.e., Duff et al., 2003), only 290 possessed completed medication data and were included for analysis in the present study. Participants completed a demographics questionnaire and medical history questionnaire that assessed past and/or current diagnosis of 48 possible medical conditions including: acromegaly; AIDS; amyloidosis; anemia; arthritis; asthma; brain hemorrhage; cancer; hepatitis B or C; concussion; Crohn’s disease; depression; diabetes; diverticulosis/itis; emphysema; glaucoma; head injury; hearing loss; heart disease; hereditary neuropathy; hiatal hernia; hypercholesteremia; hypertension; kidney/bladder stones; leprosy; liver problems; lupus; macular degeneration; multiple myeloma; neurofibromatosis; osteoarthritis; osteoporosis;
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Fig. 2. Recruitment and enrollment procedures for the study sample (n ¼ 290).
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
109
Parkinson disease; peripheral neuropathy; porphyria; rheumatoid arthritis; sarcoidosis; scleroderma; seizures; sinus problems; stomach/duodenal ulcers; stroke, TIA, or reversible ischemic neurologic deficit; surgical removal of large intestines/colon or small intestines/stomach; thyroid problems; frequent urinary tract infections; vitamin B12 deficiency; and/or Whipple’s disease. Participants then completed Form A of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS; Randolph, 1998). Statistical Analyses
Results Index Analysis The correlation between MB and AB is 0.27 which indicates that these two predictors do not substantially overlap in variance. The first regression model examined the RBANS Total Scale score. When the two predictors and their interaction were entered into the equation, the overall equation was significant (adjusted R2 ¼ .05, p , .01). Significant predictors included the AB score (b ¼ 20.23, p , .01) and the MB × AB interaction score (b ¼ 0.16, p ¼ .01). As demonstrated in the two-way interaction in Fig. 3, individuals possessing a lower MB score but higher AB score demonstrated the poorest overall RBANS performance, suggesting increased sensitivity to the potential cognitive effects of their prescribed medication regimen. Low MB/low AB participants exhibited the best RBANS Total Index performance than those prescribed a greater number of medications. Individuals with a high MB appeared less affected by increased AB. A series of regressions were next performed on each of the RBANS indexes to better identify which indexes were most affected by the observed relationship. See Table 3 for results. The AB score predicted RBANS scores across all indexes. Further, analyses of the Attention and Delayed Memory indexes revealed significant MB × AB interactions. For both indexes, a similar pattern emerged as identified with the RBANS Total Scale score such that individuals with a lower MB were significantly sensitive to the burden imposed by medications with anticholinergic cognitive effects; however, those with a higher MB demonstrated poorer performance irrespective of their level of AB. Subtest Analysis Post hoc analyses were performed for each of the following 12 RBANS subtests: List Learning, Story Memory, Figure Copy, Line Orientation, Picture Naming, Semantic Fluency, Digit Span, Coding, List Recall, List Recognition, Story Recall, and Figure Recall. Overall models were significant for the List Learning, Story Memory, Line Orientation, Coding, and Story Recall subtests. In each case, greater AB predicted poorer performance on the respective RBANS subtest score (p , .01 in all cases). In addition, a significant AB main effect and MB × AB interaction effect was revealed for the Digit Span subtest. Both
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Two predictor variables were examined in this study. The first variable was the number of self-reported medical conditions at baseline (i.e., MB). As noted above, MB variable has been previously demonstrated to have an inverse relationship with neuropsychology function (Duff et al., 2007). The number of self-reported medical conditions was summed to yield a total MB score for all statistical analyses. The second predictor variable was the AB of participants’ individual medication regimens. This variable was measured using the Anticholinergic Cognitive Burden (ACB) Scale developed by the Indiana University Center for Aging Research (Campbell et al., 2009; Boustani et al., 2008). The ACB scale identifies the severity of anticholinergic effects of prescription and over-the-counter medications divided into “possible” (score of 1) and “definite” categories (score of 2 or 3, depending upon the strength of the effect). Each medication is scored, and all medication scores are subsequently summed to result in a total AB score for that particular individual. For the purposes of this paper, we were interested in analyzing the effect of anticholinergic medications only. As both the MB and AB variables are continuous, we employed a series of moderated multiple linear regressions to examine their main and interactive effects on cognition. Following convention, predictors were first centered around their means to eliminate multicollinearity (Tabachnick & Fiddell, 2007). The predictor variables were then entered into the first block of the regression model, while the interaction of the two predictor variables was entered into the second block. The interaction variable was obtained by multiplying the value of the two predictor variables together. The initial criterion variable was the RBANS Total Scale score. Significant two-way interactions were plotted using R-code syntax. To better identify which domains were most affected by the interaction between MB and AB, secondary exploratory analyses were conducted on the five RBANS indexes. Significant findings were further evaluated at the subtest level.
110
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
the List Recall and List Recognition subtests had significant MB × AB interaction effects. These interactions followed the pattern of the RBANS total score and indexes, such that individuals with low MB had more sensitivity to AB, while those with high MB had poorer scores regardless of their AB. No significant overall or main effects were identified for the other RBANS subtests.
Discussion The purpose of the present study was to examine the interaction between the MB and AB posed by participants’ self-reported medication regimen on neuropsychological function in a sample of community-dwelling geriatric outpatients. Our findings revealed an interaction between MB and AB for selected cognitive variables. Surprisingly, however, only participants possessing a lower MB appeared to be more sensitive to increased AB given the observed reductions in RBANS Total Scale score, as well as index scores in the domains of Attention and Delayed Memory. This extends Duff and colleagues (2007) finding that decreased attention was the primary area of concern in populations with MCCs, and is consistent with previous findings that use of anticholinergic medications in older adults can decrease cognition (Boustani et al., 2008; Campbell et al., 2009). When post hoc analyses were performed, the interactions on these indexes were found to be driven primarily by Digit Span, List Recall, and List Recognition subtests. In summary, although MB and AB each independently contributed to poorer attention and delayed memory, increased AB overshadowed the effects of increased MB as indicated by total RBANS index scores (Fig. 3). Although the reasons for this surprising finding are unclear, one potential explanation may be that an increasing number of medical conditions possess a greater synergistic effect on neuropsychological function than does consuming an increased number of medications possessing anticholinergic cognitive effects. However, our results may also have been influenced by the duration of medication use. Indeed, research has demonstrated that long-term consumers of anticholinergics potentially
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Fig. 3. Multiple linear regression two-way interaction of medical burden (MB) and anticholinergic cognitive burden on RBANS Total Index Scores. ACB ¼ anticholinergic cognitive burden; Med ¼ medical burden. Ranges for both variables are represented in z-score values. For MB: “High Med” z-score ¼ +1.00, “Medium Med” z-score ¼ 0.00, and “Low Med” z-score ¼ 21.00. The anticholinergic cognitive burden score range is also represented by a range of z-score values.
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
111
Table 3. Moderated multiple regression analyses for the RBANS total and index scores b
t
p
.05
— 20.02 20.23 20.16 — 20.00 20.21 20.11
— 20.39 23.71 22.61 — 20.06 23.41 21.79
20.07 20.20 20.08 — 20.02 20.01 20.11 — 20.07 20.27 20.17 — 20.02 20.14 20.15
21.20 23.17 21.23 — 20.32 20.20 21.74 — 21.23 24.47 22.74 — 20.26 22.19 22.40
,.01 .70 ,.01 .01 ,.01 .95 ,.01 .07 ,.01 .23 ,.01 .22 .33 .75 .84 .08 ,.01 .22 ,.01 ,.01 ,.05 .79 ,.05 ,.05
.04
.04
.00
.08
.02
Note: RBANS ¼ Repeatable Battery for the Assessment of Neuropsychological Status; ACB ¼ anticholinergic cognitive burden; MB ¼ medical burden.
show less of a decline in neuropsychological functioning over time than do novel or incident users of these medications (Shah et al., 2013). It stands to reason that participants with greater MB may have been on these medications longer. However, since data regarding the duration of medication use was not available in this particular sample, we feel that this is a critical area for further investigation. In addition, data were not available on medication dosing in order to evaluate whether a dose –response relationship exists. An important implication of our findings is how they apply to the concept of diminishing returns in medicine. Mold, Hamm, and McCarthy (2010) posit that the additive influence of an increasing number of medical interventions is inversely related to the maximal intervention benefit. Furthermore, there is a concomitant increase in the cost– benefit ratio of a given medical interventions. These data suggest that increasing the number of medications with anticholinergic cognitive effects has the subsequent potential to do more harm than good in older adults with fewer medical conditions. Fortunately, efforts to increase provider awareness about appropriate medication prescription practices for older adults already exist (e.g., the Beers List; Fick et al., 2003). Programs to evaluate anticholinergic burden and help to decrease its effects in the geriatric population have also been suggested (see He and Ball, 2013; Karimi, Dharia, Floria, & Slattum, 2012). Thus, it would be beneficial to include neuropsychologists in these efforts as well as in future endeavors to evaluate neuropsychological functioning as it relates to anticholinergic burden. We demonstrated a relatively simple way that neuropsychologists could accomplish this in both research and daily clinical practice (i.e., by measuring the number of medical conditions and rating medications with the ACB scale). Projects similar to the Pasina et al. (2013) study could benefit from the use of more detailed neuropsychological measures. Our results are limited by several issues. First, to be able to examine medication effects, we were required to exclude two thirds of the original OKLAHOMA dataset. It is unfortunate that so many participants did not have this data available, and certainly future comparisons should include some record of medications as well as data on dosage, duration of use, and stability of use. The nature of the findings gleaned from correlational research is problematic as well. Given the cross-sectional nature of this study, it is unknown which came first: lower cognitive functioning or increasing MB and use of anticholinergic drugs. In fact, a 2005 population-based study indicated that older adults diagnosed with CIND or dementia were significantly more likely to experience significant medical comorbidity, which in turn was associated with worse daily functioning and lower MMSE scores (Lyketsos et al., 2005). We recommend that more research of a longitudinal nature be conducted in this area. Furthermore, the sample employed was fairly homogenous; i.e., they were primarily white, well-educated, and relatively young older adults who lived independently within the community and were without major cognitive impairment. Thus, the applicability of these findings to samples with more diverse characteristics is not yet known and should be examined.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
RBANS Total Index MB ACB MB × ACB RBANS Immediate Memory Index MB ACB MB × ACB RBANS Visuospatial/Constructional Index MB ACB MB × ACB RBANS Language Index MB ACB MB × ACB RBANS Attention Index MB ACB MB × ACB RBANS Delayed Memory Index MB ACB MB × ACB
Adjusted R 2
112
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
Conflict of Interest None declared.
References Anderson, G. (2010). Chronic care: Making the case for ongoing care. Princeton, NJ: Robert Woods Johnson Foundation. Retrieved February 24, 2014, from http ://www.rwjf.org/en/research-publications/find-rwjf-research/2010/01/chronic-care.html. Boustani, M. A., Campbell, N. L., Munger, S., Maidment, I., & Fox, G. C. (2008). Impact of anticholinergics on the aging brain: A review and practical application. Aging Health, 4, 311–320. Cai, X., Campbell, N., Khan, B., Callahan, C., & Boustani, M. (2013). Long-term anticholinergic use and the aging brain. Alzheimer‘s & Dementia, 9, 377– 385. Campbell, N., Boustani, M., Limbil, T., Ott, C., Fox, C., Maidment, I., et al. (2009). The cognitive impact of anticholinergics: A clinical review. Clinical Interventions in Aging, 4, 225–233. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40, 373 –383. Duff, K., Mold, J. W., Roberts, M. M., & McKay, S. (2007). Medical burden and cognition in older patients in primary care: Selective deficits in attention. Archives of Clinical Neuropsychology, 22, 569–575. Duff, K., Patton, D., Schoenberg, M. R., Mold, J., Scott, J. G., & Adams, R. L. (2003). Age- and education-corrected independent normative data for the RBANS in a community dwelling elderly sample. Clinical Neuropsychology, 17, 351– 366. Fick, D. M., Cooper, J. W., Wade, W. E., Waller, J. L., Maclean, J. R., & Beers, M. H. (2003). Updating the beers criteria for potentially inappropriate medication use in older adults. Archives of Internal Medicine, 163, 2716–2724. Fu, A. Z., Liu, G. G., & Christensen, D. B. (2004). Inappropriate medication use and health outcomes in the elderly. Journal of the American Geriatric Society, 52, 1934– 1939. He, Z., & Ball, P. A. (2013). Can medication management review reduce anticholinergic burden (AB) in the elderly? Encouraging results from a theoretical model. International Psychogeriatrics, 25, 1425–1431.
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Some of the measures used constitute another limitation. The OKLAHOMA health questionnaire lacks the inclusion of some conditions such as dementia and MCI. However, dementia was not a primary concern as the study explicitly excluded these individuals from participation. More specifically, the study was designed to study long-term health outcomes in a sample of independently living community-dwelling geriatric individuals. Participants, along with being able to provide informed consent, were required to be functionally independent to be in the study. Although formal neuropsychological testing beyond the RBANS is not available, these patients were well above the level of impairment that would characterize dementia during their initial evaluation. It is certainly possible that some participants were on the cusp of MCI that did not necessarily interfere yet with ADL or IADL performance. However, we elected not to outright exclude these individuals because MCI itself is a heterogeneous condition that is often prevalent in this population, and removing all cases that met some criteria (e.g., score ,85 on at least one index) would limit the generalizability of findings. Without signs of actual functional impairment, it is also difficult to ascertain whether MCI for a specific individual actually represents the beginning of cognitive decline or is simply a transient phenomenon that resolves on later testing. We elected to remove neurological conditions with known pathophysiology and a co-occurring prototypical neurological profile (e.g., Parkinson’s disease) because these disorders impair cognition in a predictable manner. However, in an attempt to maintain study generalizability and feasibility, we opted to let participants diagnosed with conditions with potential neurologic sequelae (e.g., depression and diabetes) remain. The RBANS is a brief measure of cognition that is particularly limited in its ability to assess executive functioning, and MB was determined by self-report via a questionnaire that was neither standardized nor exhaustive. Future research in this area should assess cognition and MB more comprehensively by using more in-depth measures of neuropsychological functioning as well as a validated, weighted questionnaire of MB (e.g., the Index of Comorbidity; Charlson, Pompei, Ales, & MacKenzie, 1987). Similarly, the potential association between cognition and other medications participants were prescribed was not addressed. Given the potential for total medication cognitive burden, we recommend this as an area for future study. Finally, only a small proportion of the variance (2% to 8%) in performance on the RBANS was explained by the main effects of an interaction between MB and AB. This leaves a significant amount of the variance in cognition as measured by the RBANS unexplained. Even when taking these limitations into account, we consider the findings of the present study to be important. To our knowledge, this is the first study to date to examine the complex relationship between MB and AB and their potential influence on neuropsychological performance. Statistics for the financial impact of MCCs on our healthcare system are striking (Lochner et al., 2013); 85% of U.S. healthcare spending dollars are funneled towards individuals with MCCs (Anderson, 2010; Lochner et al., 2013). It will become more important than ever to delineate the burden imposed by MCCs and the medications designed to treat these conditions due to the advent of the ACA and its allowing an increased catchment of older Americans into the U.S. healthcare system.
C. K. Block et al. / Archives of Clinical Neuropsychology 30 (2015); 105–113
113
Downloaded from http://acn.oxfordjournals.org/ at Heriot-Watt University on September 27, 2016
Kadoi, Y., Saito, S., & Fujita, N. (2005). Risk factors for cognitive dysfunction after coronary artery bypass graft surgery in patients with type 2 diabetes. Journal of Thoracic and Cardiovascular Surgery, 3, 576– 593. Karimi, S., Dharia, S. P., Floria, D. S., & Slattum, P. W. (2012). Anticholinergic burden: Clinical implications for seniors and strategies for clinicians. Consult Pharmacy, 27, 564 –82. Karp, J. F., Reynolds, C. F., Butters, M., Dew, M. A., Mazumdar, S., & Begley, A. E., et al. (2006). Correlates of neuropsychological impairment in older adult pain clinic patients. Pain Medicine, 7, 444– 452. Kersten, H., & Wyller, T. B. (2014). Anticholinergic drug burden in older people’s brain—How well is it measured? Basic & Clinical Pharmacology & Toxicology, 114, 151– 159. Koyoma, A., Steinman, M., Ensrud, K., Hillier, T. A., & Yaffe, K. (2014). Long-term cognitive and functional effects of potentially inappropriate medications in older women. The Journals of Gerontology Series A: Biological Sciences & Medical Sciences, 69, 423 –9. Kutlay, S., Nergizoglu, G., Duman, N., Atli, T., Keven, K., Eturk, S., et al. (2001). Recognition of Neurocognitive Dysfunction in Chronic Hemodialysis Patients. Renal Failure, 23, 781–787. Lochner, K. A., Goodman, R. A., Posner, S., & Parekh, A. (2013). Multiple chronic conditions among Medicare beneficiaries: State-level variations in prevalence, utilization, and cost, 2011. Medicare & Medicaid Research Review, 3, E1– E19. Lyketsos, C. G., Toone, L., Schanz, J., Rabins, P. V., Steinberg, M., Onyike, C. U., et al. (2005). Population-based study of medical comorbidity in early dementia and “cognitive impairment, no dementia (CIND):” Association with functional and cognitive impairment: The Cache County Study. American Journal of Geriatric Psychiatry, 13, 656 –664. MacNeill, S. E., Lichtenberg, P. A., & LaBuda, J. (2000). Factors affecting return to living alone after medical rehabilitation: A cross-validation study. Rehabilitation Psychology, 45, 356–364. Mold, J. W., Hamm, R. M., & McCarthy, L. H. (2010). The law of diminishing returns in clinical medicine: How much risk reduction is enough? Journal of the American Board of Family Medicine, 23, 371–375. Pasina, L., Djade, C. D., Lucca, U., Nobili, A., Tettamanti, M., Franchi, C., et al. (2013). Association of anticholinergic burden with cognitive and functional status in a cohort of hospitalized elderly: Comparison of the Anticholinergic Burden Scale and Anticholinergic Risk Scale. Drugs and Aging, 30, 103– 112. Perry, E. K., Kilford, L., Lees, A. J., Burn, D. J., & Perry, R. H. (2003). Increased Alzheimer pathology in Parkinson’s disease related to antimuscarinic drugs. Annals of Neurology, 54, 235–238. Peters, N. L. (1989). Snipping the thread of life: Antimuscarinic side effects of medications in the elderly. Archives of Internal Medicine, 149, 2414–2420. Randolph, C. (1998). Repeatable battery for the assessment of neuropsychological status (RBANS). San Antonio, Harcourt, TX: The Psychological Corporation. Ray, P. G., Meador, K. J., Loring, D. W., Zamrini, E. W., Yang, X. H., & Buccasfusco, J. J. (1992). Central anticholinergic hypersensitivity in aging. Journal of Geriatric Psychiatry & Neurology, 5, 72– 77. Shah, R. C., Janos, A. L., Kline, J. E., Yu, L., Leurgeans, S. E., & Wilson, R. S., et al. (2013). Cognitive decline in older persons initiating anticholinergic medications. PLoS ONE, 8, doi: 10.1371/journal.pone.0064111. Tabachnick, B. G., & Fiddell, L. S. (2007). Using Multivariate Statistics (5th ed.). Boston: Allyn and Bacon. U.S. Bureau of the Census. (2008). Projected Population by Single Year of Age, Sex, Race, and Hispanic Origin for the United States: July 1, 2000 to July 1, 2050. Washington. Retrieved February 24, 2014, from http://www.census.gov/population/projections/data/national/2008/downloadablefiles.html. Yeh, Y. C., Liu, C. L., Peng, L. N., Lin, M. H., & Chen, L. K. (2013). Potential benefits of reducing medication-related anticholinergic burden for demented older adults: A prospective cohort study. Geriatrics and Gerontology International, 13, 694–700.
