BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 26, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from

An analysis of anticholinergic and sedative medicine effects on medicine-induced deterioration and frailty: a cross- sectional study ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-029221

Article Type: Research

Date Submitted by the 17-Jan-2019 Author:

Complete List of Authors: Lim, Renly; University of South Australia Division of Health Sciences, Quality Use of Medicines and Pharmacy Research Centre; University of South Australia Kalisch Ellett, Lisa; University of South Australia, Sansom Institute for Health Resarch, School of Pharmacy and Medical Sciences, Quality Use of Medicines and Pharmacy Research Centre Widadgo, Imaina; University of South Australia, Sansom Institute for Health Resarch, School of Pharmacy and Medical Sciences, Quality Use of Medicines and Pharmacy Research Centre Pratt, Nicole; University of South Australia, Sansom Institute for Health Resarch Roughead, Elizabeth; University of South Australia, Sansom Institute for

Health Research http://bmjopen.bmj.com/

Keywords: anticholinergics, frail older adults, sedatives

on September 26, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 26 BMJ Open

1 2

3 Title: An analysis of anticholinergic and sedative medicine effects on medicine-induced BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from 4 5 6 deterioration and frailty: a cross-sectional study 7 8 9 10 Author names and institution: 11 12 Renly Lim1, PhD, Lisa M. Kalisch Ellett1, PhD, Imaina S. Widagdo1, PhD, Nicole L. Pratt1, 13 14 15 PhD, Elizabeth E. Roughead1, PhD 16 17 18 For peer review only 19 1 Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical 20 21 22 Sciences, University of South Australia, Adelaide, Australia. 23 24 25 26 Corresponding author 27 28 29 Renly Lim 30 31 Quality Use of Medicines and Pharmacy Research Centre, 32 33 School of Pharmacy and Medical Sciences, 34 35 University of South Australia, 36

37 http://bmjopen.bmj.com/ 38 GPO Box 2471, Adelaide, SA 5001, Australia. 39 40 E-mail: [email protected] 41 42 Tel: +(61) 883 022 307 43 44 Fax: +(61) 8 8302 2466

45 on September 26, 2021 by guest. Protected copyright. 46 47 ORCID ID: https://orcid.org/0000-0003-4135-2523 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 26

2 1 2

3 Abstract BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from 4 5 6 7 Objective 8 To test the association between use of medicines with anticholinergic or sedative properties 9 10 and medicine-induced deterioration (physical function, cognitive function, appetite), and 11 12 frailty. 13 14 15 Design, setting and participants 16 17 This cross-sectional study analysed baseline data collected as part of the Australian 18 For peer review only 19 Longitudinal Study of Ageing (ALSA), a population based cohort of 2087 participants aged 20 21 65 years or over living in South Australia. 22 23 24 Main outcome measures 25 26 Physical function was measured at baseline using measures including hand grip strength, 27 walking speed, chair stands, activities of daily living (ADL), instrumental activities of daily 28 29 living (IADL). Cognitive function was measured using Mini Mental State Examination 30 31 (MMSE). Appetite was measured using Center for Epidemiologic Studies Depression (CES- 32 33 D) question 2. Frailty was measured using Frailty Index. The association between use of 34 anticholinergics or sedatives and physical or cognitive function, appetite, or frailty, was 35 36 assessed using analysis of covariance and ordinal or binary logistic regression.

37 http://bmjopen.bmj.com/ 38 39 40 Results 41 Almost half of the population were using anticholinergics or sedatives (n=954, 45.7%). Use 42 43 of anticholinergics was significantly associated with poorer grip strength (women only), 44 slower walking speed, poorer instrumental activities of daily living (IADL) and poorer

45 on September 26, 2021 by guest. Protected copyright. 46 appetite. Use of sedatives was significantly associated with poorer grip strength (men only), 47 48 slower walking speed and poorer IADL. We found no significant association between 49 50 medicine use and cognitive function. Users of anticholinergics or sedatives were significantly 51 52 more likely to be frail compared with non-users. 53 54 55 Conclusion 56 57 Use of medicines with anticholinergic or sedative properties is significantly associated with 58 medicine-induced deterioration (physical function and appetite) and frailty. Early 59 60 identification of signs and symptoms of medicine-induced deterioration is particularly

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 26 BMJ Open

3 1 2

3 important in older people so that worsening frailty and subsequent adverse events are BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from 4 5 prevented. 6 7 8 Keywords: anticholinergics; frail older adults; sedatives 9 10 11 12 Word count: 2905 13 14 15 16 17 Article Summary 18 For peer review only 19 Strengths and limitations of this study 20 21  Our findings on the association between use of anticholinergics and poorer appetite 22 contribute uniquely to the literature. 23 24  We analysed the association between medicines with anticholinergic or sedative 25 26 properties and grip strength based on gender, which previous studies have not done. 27 28  The study is limited by its cross-sectional design. It is not possible to identify a cause- 29 effect relationship between use of medicines with anticholinergic or sedative 30 31 properties and reduced physical function or poorer appetite using this study design. 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 26

4 1 2

3 Introduction BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from 4 5 Medicines with anticholinergic or sedative properties are commonly prescribed to older 6 7 people (1), with prevalence estimates ranging from 13 to 42% for medicines with sedative 8 properties (2, 3) and from 8 to 36% for medicines with anticholinergic properties (3, 4). 9 10 Adverse effects of these medicines, like falls and confusion are detrimental to older people. 11 12 This risk is increased due to age-related changes in pharmacokinetics and 13 14 pharmacodynamics, the presence of multiple co-morbidities and the use of multiple 15 medicines with subsequent increased probability of drug interactions (5, 6). The cumulative 16 17 medication burden of anticholinergic and sedative medicines is often unintentional and 18 For peer review only 19 compounded by the fact that many medicines with anticholinergic properties also have 20 21 sedative properties (5). 22 23 24 Use of medicines with anticholinergic or sedative properties can directly lead to adverse 25 26 events (5) which are easily recognised and potentially rectifiable. Use of medicines with 27 anticholinergic or sedative properties has also been associated with frailty (7), and frailty may 28 29 contribute to an increased risk of adverse events (8). However, medicines with 30 31 anticholinergic or sedative properties often have what might be considered “minor side 32 33 effects” which are difficult to detect and frequently unrecognised. These side effects, which 34 we describe as medicine-induced deterioration, particularly if the cumulative effect builds 35 36 slowly over time, is often misattributed as geriatric syndromes, frailty or simply changes due

37 http://bmjopen.bmj.com/ 38 to aging. Medicine-induced deterioration may include decline in physical function, decline in 39 40 cognitive function and loss of appetite. While these same symptoms may occur independently 41 due to aging, medicines with anticholinergic or sedative properties have side effect profiles 42 43 that also contribute to these declines. The lack of recognition of these signs and symptoms as 44 medicine-induced may subsequently contribute to increased risk of frailty and subsequent

45 on September 26, 2021 by guest. Protected copyright. 46 increased risk of adverse events such as falls and fractures. 47 48 49 50 There are several pharmacological pathways by which medicines with anticholinergic or 51 52 sedative properties may contribute to medicine-induced deterioration. Medicines with 53 sedative properties enhance the effects of the neurotransmitter gamma-aminobutyric acid 54 55 (GABA) at the GABAA receptor in the central nervous system, causing sedative and muscle 56 57 relaxant effects (9). The side effects related to the sedative and muscle-relaxing actions may 58 cause a decline in both cognitive and physical function. Additionally, some medicines with 59 60 sedative properties may cause a decline in cognitive function by blocking the dopamine-D2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 26 BMJ Open

5 1 2

3 receptors in the central nervous system. Similarly, use of medicines with anticholinergic BMJ Open: first published as 10.1136/bmjopen-2019-029221 on 4 September 2019. Downloaded from 4 5 properties may lead to medicine-induced deterioration via several pathways. As some 6 7 medicines with anticholinergic properties have sedative properties, these medicines may 8 directly lead to a decline in physical function due to the sedative side effects. Medicines with 9 10 anticholinergic properties may affect cognitive function by blocking the muscarinic receptors 11 12 in the central nervous system which regulates learning and memory (10). Medicines with 13 14 anticholinergic properties block the effects of acetylcholine at the M3-muscarinic receptors of 15 the salivary glands causing dry mouth, which may lead to loss of appetite. 16 17 18 For peer review only 19 We have created a model of the relationship between medicine use, medicine-induced 20 21 deterioration, and frailty (Figure 1). Using data from the Australian Longitudinal Study of 22 Ageing (ALSA) (11), the study aims to assess (i) the association between use of medicines 23 24 with anticholinergic or sedative properties and medicine-induced deterioration (physical 25 26 function, cognitive function, appetite) and (ii) the association between use of medicines with 27 anticholinergic or sedative properties and frailty. We hypothesized that use of medicines with 28 29 anticholinergic or sedative properties is associated with medicine-induced deterioration 30 31 (decline in physical function and decline in cognitive function for anticholinergics or 32 33 sedatives, and poorer appetite for anticholinergics), and increased frailty. If anticholinergic or 34 sedative medicine-induced deterioration was monitored for and detected in practice, this 35 36 would provide an opportunity for healthcare professionals to prevent deterioration by

37 http://bmjopen.bmj.com/ 38 moderating medicine use, mitigating progression to frailty and potentially avoiding 39 40 subsequent adverse events (Figure 1). 41 42 43 44

45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 26

6 1 2 3 Method 4 5 Study population and data collection 6 7 This study analysed baseline data collected as part of the Australian Longitudinal Study of 8 Ageing (ALSA), a population based cohort of 2087 participants aged 65 years or over at 9 10 baseline (1992) living in South Australia (11). Both community-dwelling and people living in 11 12 residential care were eligible and were recruited by stratified random sampling from the 13 14 South Australian electoral roll. Data collection was undertaken by comprehensive personal 15 interviews, self-completed questionnaires and clinical assessments on physical function. 16 17 18 For peer review only 19 Medicines with anticholinergic or sedative properties 20 21 We identified medicines with anticholinergic properties using Duran’s scale (12) with the 22 addition of other anticholinergic medicines used in Australia, as identified in the Australian 23 24 Medicines Handbook, Martindale and expert opinion (13). Medicines with sedative properties 25 26 were identified from the Australian Pharmaceutical Formulary and Handbook. Medicines 27 with sedative properties were defined as those which cause sedation, and which are required 28 29 by legislation to be dispensed with a specific sedation warning on the pharmacy label. 30 31 Participants were stratified into one of four groups: i) non-users, ii) users of medicines with 32 33 anticholinergic properties only, iii) users of medicines with sedative properties only or iv) 34 users of medicines with sedative and anticholinergic properties. 35 36 37 38 Measures of physical function, cognitive function, appetite and frailty 39 40 As part of baseline assessments in ALSA (11), participants had their physical function 41 objectively assessed using hand grip strength, a timed 8-feet (2.4 metre) walk at a normal 42 43 pace and repeated chair stands. Hand grip strength was assessed using a handheld 44 45 dynamometer. The maximum grip strength (kg) was recorded based on the best of two trials. 46 Walking speed was calculated using distance in meters and time in seconds. Participants were 47 48 instructed to walk at normal pace from a standing start over a distance of 8 feet. Repeated 49 50 chair stands were conducted by asking participants to rise unassisted from a chair five times. 51 52 Functional status was subjectively assessed by interviewers using activities of daily living 53 (ADL) and instrumental activities of daily living (IADL) scores. The ADL includes walking, 54 55 bathing, personal grooming, dressing, eating, using the toilet and getting from bed to 56 57 chair(14). The IADL encompasses the following tasks: housekeeping, meal preparation, 58 telephone use, handling finances, use of transportation and shopping(15). Cognitive function 59 60 was measured using the Mini Mental State Examination (MMSE) questionnaire (range 0-30)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 26 BMJ Open

7 1 2 3 (16). Appetite was measured using the Center for Epidemiologic Studies Depression (CES-D) 4 5 question 2 “I did not feel like eating; my appetite was poor”, with scores of 0=rarely or none 6 7 of the time, 1=some or a little of the time, 2=occasionally or a moderate amount of time, and 8 3=most or all of the time. The frailty index (8), a multidimensional assessment which 9 10 included physical, medical, psychological and social factors, was used to measure frailty. The 11 12 frailty index, which has been validated in the Australian setting (17, 18), included 39 13 14 variables with a total score range of between 0 and 1. A higher frailty index indicates worse 15 frailty status. The subjective functional status (ADL and IADL) formed part of the frailty 16 17 index while the objective physical function (grip strength, walking speed, chair stands), 18 For peer review only 19 cognitive function (MMSE) and appetite (CES-D question 2) were not directly measured in 20 21 the frailty index. 22 23 24 Covariates 25 26 Univariate analysis was performed to determine which measures were significantly 27 associated with use of anticholinergics or sedatives. The final analyses were adjusted for age, 28 29 body mass index, residential status, smoking status, cognitive impairment, depressive 30 31 symptoms and number of co-morbidities using the Functional Comorbidity Index. The 32 33 Functional Comorbidity Index contains a list of diseases reported to be associated with 34 function in older populations (19). We excluded depressive symptoms from the Functional 35 36 Comorbidity Index when we used it as a covariate for adjustment as depression was assessed 37 38 independently. We did not adjust for covariates for the frailty analysis because it is unclear 39 40 how the covariates such as age and comorbidities affect the relationship between medicines 41 use and frailty (20). 42 43 44 45 Statistical analysis 46 Descriptive statistics were used to report the baseline characteristics of the study population. 47 48 The one-way analysis of variance was used to compare continuous variables including age, 49 50 body mass index, Functional Comorbidity Index, MMSE, CES-D, while the chi-square test 51 52 was used to compare categorical variables including gender, smoking, presence of cognitive 53 impairment and presence of depression. The prevalence of use of medicines was grouped by 54 55 Anatomical Therapeutic Chemical (ATC) class (21). 56 57 58 Unadjusted analysis of variance and adjusted analysis of covariance were performed to 59 60 compare differences in physical or cognitive function between users and non-users. Given

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 26

8 1 2 3 that hand grip strength has been shown to be significantly different between men and women 4 5 with distinct gender specific cut-off points (22), this performance measure was analysed 6 7 separately based on gender. All other analyses were not stratified by gender. Ordinal logistic 8 regression was used to compare differences in appetite between users and non-users. Binary 9 10 logistic regression was used to compare differences in frailty between users and non-users. 11 12 Analysis was performed using SAS version 9.4 for Window (SAS Institute Inc., Cary, NC, 13 14 USA). A priori, a p-value of <0.05 was considered statistically significant. 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 26 BMJ Open

9 1 2 3 Results 4 5 Characteristics of study population 6 7 Table 1 shows the baseline characteristics of study participants. The mean age ± standard 8 deviation of participants was 78 ± 7 years old; 94% were community-dwelling participants 9 10 and 6% lived in residential care. Nearly half of the population were using medicines with 11 12 anticholinergic or sedative properties (n=954, 45.7%); 18.3% were using medicines with 13 14 anticholinergic properties only, 11.3% were using medicines with sedative properties only 15 and 16.1% were using medicines with sedative and anticholinergic properties. The most 16 17 commonly used medicines with sedative properties were benzodiazepines, antidepressants 18 For peer review only 19 and antihistamines (Table 2, Additional file 1). The most frequently used medicines with 20 21 anticholinergic properties were frusemide, H2-receptor antagonists and digoxin (Table 2). 22 23 24 Use of medicines with anticholinergic or sedative properties and physical function 25 26 In the unadjusted analyses, users of medicines with anticholinergic properties only, medicines 27 with sedative properties only, or medicines with sedative and anticholinergic properties had 28 29 significantly poorer function in all physical function measures (p<0.05). 30 31 32 33 After adjusting for covariates, use of medicines with anticholinergic properties only was 34 significantly associated with poorer grip strength (women only, mean difference M -1.43 35 diff 36 95% confidence interval CI -2.56, -0.31), poorer performance on chair stands (Mdiff 0.90, 37 38 95% CI 0.13, 1.68) slower walking speed (Mdiff -0.05, 95% CI -0.09, -0.02), and poorer IADL 39 40 score (Mdiff 0.22, 95% CI 0.04, 0.41) (Table 3). After adjusting for covariates, use of 41 medicines with sedative properties only was significantly associated with poorer grip strength 42 43 (men only, Mdiff -1.92, 95% CI -3.54, -0.30), slower walking speed (Mdiff -0.05, 95% CI -0.10, 44 45 -0.01) and poorer IADL score (Mdiff 0.23, 95% CI 0.01, 0.45) (Table 3). After adjusting for 46 covariates, use of medicines with anticholinergic and sedative properties was significantly 47 48 associated with slower walking speed (Mdiff -0.08, 95% CI -0.11, -0.04), and poorer IADL 49 50 score (Mdiff 0.42, 95% CI 0.22, 0.61) (Table 3). 51 52 53 Use of medicines with anticholinergic or sedative properties and cognitive function 54 55 Use of medicines with anticholinergic or sedative properties was significantly associated with 56 57 poorer cognitive function in the unadjusted analysis (p<0.05), but the associations were not 58 significant after adjusting for covariates (p>0.05) (Table 3). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 26

10 1 2 3 4 5 Use of medicines with anticholinergic properties and appetite 6 7 Use of medicines with anticholinergic properties was significantly associated with poorer 8 appetite in both the unadjusted (OR 2.25, 95% CI 1.79, 2.82) and adjusted analysis (OR 1.77, 9 10 95% CI 1.27, 2.45). 11 12 13 14 Use of medicines with anticholinergic or sedative properties and frailty 15 Participants who used anticholinergics or sedatives had three times or more the odds of being 16 17 frail compared with non-users (anticholinergics only: OR 3.9, 95% CI 2.9, 5.3; sedatives 18 For peer review only 19 only: OR 3.3, 95% CI 2.3, 4.8; sedatives and anticholinergics: OR 6.2, 95% CI 4.6, 8.5). 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 26 BMJ Open

11 1 2 3 Discussion 4 5 Our results showed that use of medicines with anticholinergic or sedative properties was 6 7 significantly associated with medicine-induced deterioration. In addition, we found that 8 participants who used medicines with anticholinergic or sedative properties were significantly 9 10 more likely to be frail. We have previously demonstrated, using a longitudinal design in the 11 12 same study population that frail older people (as identified by the frailty index), were more 13 14 likely to have adverse events at follow-up including an increased risk of mortality (OR 3.2, 15 95% CI 2.4, 4.1), hospitalisation (OR 2.3, 95% CI 1.7, 3.0), nursing home admission (OR 16 17 3.3, 95% CI 1.6, 7.0) and fall (OR 3.4, 95% CI 2.7, 4.3) (8). Collectively, our results support 18 For peer review only 19 the hypothesis that use of medicines with anticholinergic or sedative properties may 20 21 contribute to frailty via the intermediary pathways of medicine-induced deterioration or may 22 directly contribute to frailty, leading to an increased risk of adverse events (Figure 1). 23 24 25 26 Our findings on the association of anticholinergics or sedatives with physical function are 27 similar to another Australian cross-sectional study (3). Similar associations have also been 28 29 observed in studies involving community-dwelling older adults in United States (23, 24), 30 31 Italy (25) and Finland (26). A growing body of evidence has shown that limitations in 32 33 physical function in older people (a component of frailty), predict risk of fracture, 34 complications or length of hospital stay, disability and mortality (27-30). An adequate degree 35 36 of mobility is crucial for independent living and maintenance of quality of life in older people 37 38 (31). With the number of adults aged over 65 years old in Australia projected to rise from 39 40 14% in 2012 to 27% in 2101 (32), an important goal of geriatric medicine is to reduce a 41 patient’s medication burden to prevent medicine-induced functional impairment. 42 43 44 45 The mean difference of 1.9kg in adjusted grip strength between users versus non-users of 46 medicines with sedative properties in men was statistically significant. Similarly, the mean 47 48 difference in adjusted grip strength between users versus non-users of medicines with 49 50 anticholinergic properties (1.4kg) in women was statistically significant and clinically 51 52 relevant (33). Use of anticholinergics or sedatives was associated with a difference in walking 53 speed of 0.05-0.08m/s between users and non-users, which was considered to be a clinically 54 55 meaningful change (34). Grip strength and walking speed are two commonly included 56 57 components of frailty, and have been shown to relate to clinically significant outcomes such 58 as functional disability (27, 28) and mortality (35) in older people. 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 26

12 1 2 3 Our results indicate that use of anticholinergics or sedatives was not significantly associated 4 5 with poorer cognitive function. While studies in other countries have reported contrasting 6 7 results (36), our findings are in keeping with another Australian cross-sectional study 8 involving 1705 community-dwelling man which reported that anticholinergic and sedative 9 10 medicine use as measured by the drug burden index was not associated with limitations in 11 12 cognitive function (37). It may be that use of anticholinergics or sedatives is not associated 13 14 with poorer cognitive function in our study population, however, it may also be that the 15 MMSE is not sensitive enough to detect mild cognitive impairment (38). Other measures, 16 17 such as the Montreal Cognitive Assessment tool have been shown to be better for detecting 18 For peer review only 19 mild cognitive impairment. The lack of association between medicines with anticholinergic 20 21 or sedative properties and poorer cognitive function should be interpreted with caution due to 22 the high mean MMSE score (27 points) and the small percentage of participants with 23 24 cognitive impairment (16%) in our population. In addition, the dose and duration of use of 25 26 medicines with anticholinergic or sedative properties could have been below the threshold 27 required to decrease cognitive function. Most patients in our study population used only one 28 29 medicine with anticholinergic or sedative properties. 30 31 32 33 Our findings on the association between use of anticholinergics and poorer appetite 34 contribute uniquely to the literature. While a cause-effect relationship cannot be established, 35 36 it is widely known that anticholinergics can cause dry mouth (39) and it is plausible that this 37 38 contributed to our finding. In older people, loss of appetite due to dry mouth is difficult to 39 40 detect and loss of appetite may be misattributed to be part of the natural ageing process. It is 41 likely that loss of appetite due to medicines-induced dry mouth is an under-detected problem; 42 43 however this topic has not been well researched. 44 45 46 There are several strengths to our study. We proposed an intermediary pathway by which 47 48 medicines use contributes to frailty in older people. In analysing the intermediary pathways 49 50 of medicine-induced deterioration, we assessed grip strength, chair stands and walking speed 51 52 all of which were not variables included in our calculation of frailty. We analysed the 53 association between medicines with anticholinergic or sedative properties and grip strength 54 55 based on gender, which previous studies have not done. The outcome measures chosen were 56 57 both objective and clinically relevant. The study is limited by its cross-sectional design. It is 58 not possible to identify a cause-effect relationship between use of medicines with 59 60 anticholinergic or sedative properties and reduced physical function or poorer appetite using

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 26 BMJ Open

13 1 2 3 this study design. Since the majority of participants used only one medicine with 4 5 anticholinergic or sedative properties, we did not examine the association between the 6 7 different number of medicines used and medicine-induced deterioration. Another limitation is 8 that we did not examine the effects of dose or the duration of treatment. Although we 9 10 adjusted our analysis for various potential confounding factors, there is a possibility of 11 12 residual confounding. 13 14 15 In conclusion, use of medicines with anticholinergic or sedative properties is significantly 16 17 associated with medicine-induced deterioration (physical function and appetite) and frailty. 18 For peer review only 19 Early identification of signs and symptoms of medicine-induced deterioration is particularly 20 21 important in older people so that worsening frailty and subsequent adverse events are 22 prevented. Grip strength, chair stands and walking speed are simple assessment tools that 23 24 could be used routinely in general practice or pharmacy practice to monitor for medicine 25 26 induced deterioration in persons considered at risk of frailty. Use of these tools would 27 provide an objective measure by which clinicians could assess medicine-induced 28 29 deterioration. Clinicians and pharmacists should review use of medicines with anticholinergic 30 31 or sedative properties in the older population who are at the highest risk of medicine-induced 32 33 deterioration. 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 26

14 1 2 3 Funding 4 5 This study used data collected as part of the Australian Longitudinal Study of Ageing 6 7 (ALSA). ALSA was initially funded by a grant from the US National Institute of Health (AG 8 08523-02), with additional funding from the South Australian government, Flinders 9 10 University and other NGOs. Subsequent funding has been provided by the Australian 11 12 Research Council (ARC-LP 0669272, ARC-LP 100200413, ARC-DP 0879152 and ARC-DP 13 14 130100428) and the National Health and Medical Research Council (NHMRC 179839 and 15 229922). 16 17 18 For peer review only 19 Competing interests 20 21 The authors declare that they have no conflicts of interest. 22 23 24 Author contributions 25 26 RL conceived and designed the study, participated in data analysis and drafted the 27 manuscript. LKE, ISW and NP participated in data analysis and critically revised the 28 29 manuscript for important intellectual content. EER conceived and designed the study, and 30 31 critically revised the manuscript for important intellectual content. All authors read and 32 33 approved the final manuscript. 34 35 36 Acknowledgement 37 38 We thank the participants in the Australian Longitudinal Study of Ageing (ALSA), who have 39 40 given their time over many years, and gratefully acknowledge the work of the project team at 41 the Flinders Centre for Ageing Studies, Flinders University, Adelaide, who carried out the 42 43 ALSA and provided data for this paper. 44 45 46 Data sharing statement 47 48 The data that support the findings of this study are available from the Flinders Centre for 49 50 Ageing Studies, Flinders University but restrictions apply to the availability of these data, 51 52 which were used under license for the current study, and so are not publicly available. Data 53 are however available from the authors upon reasonable request and with permission from 54 55 Flinders Centre for Ageing Studies, Flinders University. 56 57 58 Ethics approval and consent to participate: This study used data collected as part of the 59 60 Australian Longitudinal Study of Ageing (ALSA). Ethical approval for the study was

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 26 BMJ Open

15 1 2 3 obtained from the Clinical Investigation Committee of Flinders Medical. Written informed 4 5 consent was given by each participant in the study. 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 26

16 1 2 3 References 4 5 1. Gadzhanova S, Roughead E, Robinson M. Use of Medicines with Anticholinergic and 6 Sedative Effect Before and After Initiation of Anti-Dementia Medications. Drugs - real world 7 outcomes. 2015; 2: 53-60. 8 2. Taipale HT, Bell JS, Gnjidic D, et al. Sedative load among community-dwelling 9 10 people aged 75 years or older: association with balance and mobility. Journal of clinical 11 psychopharmacology. 2012; 32: 218-24. 12 3. Gnjidic D, Cumming RG, Le Couteur DG, et al. Drug Burden Index and physical 13 function in older Australian men. British journal of clinical pharmacology. 2009; 68: 97-105. 14 4. Carriere I, Fourrier-Reglat A, Dartigues JF, et al. Drugs with anticholinergic 15 properties, cognitive decline, and dementia in an elderly general population: the 3-city study. 16 17 Arch Intern Med. 2009; 169: 1317-24. 18 5. Mintzer J, BurnsFor A. Anticholinergicpeer review side-effects of onlydrugs in elderly people. Journal of 19 the Royal Society of Medicine. 2000; 93: 457-62. 20 6. McLean AJ, Le Couteur DG. Aging biology and geriatric clinical pharmacology. 21 Pharmacological reviews. 2004; 56: 163-84. 22 7. Gnjidic D, Hilmer SN, Blyth FM, et al. High-risk prescribing and incidence of frailty 23 among older community-dwelling men. Clinical pharmacology and therapeutics. 2012; 91: 24 25 521-8. 26 8. Widagdo IS, Pratt N, Russell M, et al. Predictive performance of four frailty measures 27 in an older Australian population. Age and ageing. 2015; 44: 967-72. 28 9. Rudolph U, Crestani F, Benke D, et al. Benzodiazepine actions mediated by specific 29 gamma-aminobutyric acid(A) receptor subtypes. Nature. 1999; 401: 796-800. 30 10. Volpicelli LA, Levey AI. Muscarinic acetylcholine receptor subtypes in cerebral 31 cortex and hippocampus. Progress in brain research. 2004; 145: 59-66. 32 33 11. Flinders Centre for Ageing Studies. ALSA: The Australian Longitudinal Study of 34 Ageing. 2016. 35 12. Duran CE, Azermai M, Vander Stichele RH. Systematic review of anticholinergic risk 36 scales in older adults. European journal of clinical pharmacology. 2013; 69: 1485-96. 37 13. Australian Government Department of Veterans' Affairs. Topic 39: Thinking clearly 38 about the anticholinergic burden., 2014. 39 40 14. Katz S, Downs TD, Cash HR, et al. Progress in development of the index of ADL. 41 The Gerontologist. 1970; 10: 20-30. 42 15. Lawton MP, Brody EM. Assessment of older people: self-maintaining and 43 instrumental activities of daily living. The Gerontologist. 1969; 9: 179-86. 44 16. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for 45 grading the cognitive state of patients for the clinician. Journal of psychiatric research. 1975; 46 12: 189-98. 47 48 17. Widagdo I, Pratt N, Russell M, et al. How common is frailty in older Australians? 49 Australasian journal on ageing. 2015; 34: 247-51. 50 18. Widagdo IS, Pratt N, Russell M, et al. Construct Validity of Four Frailty Measures in 51 an Older Australian Population: A Rasch Analysis. The Journal of frailty & aging. 2016; 5: 52 78-81. 53 19. Groll DL, To T, Bombardier C, et al. The development of a comorbidity index with 54 physical function as the outcome. Journal of clinical epidemiology. 2005; 58: 595-602. 55 56 20. Rodríguez-Mañas L, Féart C, Mann G, et al. Searching for an Operational Definition 57 of Frailty: A Delphi Method Based Consensus Statement. The Frailty Operative Definition- 58 Consensus Conference Project. The Journals of Gerontology Series A: Biological Sciences 59 and Medical Sciences. 2013; 68: 62-67. 60 21. WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index 2017.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 26 BMJ Open

17 1 2 3 22. Bahat G, Tufan A, Tufan F, et al. Cut-off points to identify sarcopenia according to 4 European Working Group on Sarcopenia in Older People (EWGSOP) definition. Clinical 5 6 nutrition. 2016. 7 23. Cao YJ, Mager DE, Simonsick EM, et al. Physical and cognitive performance and 8 burden of anticholinergics, sedatives, and ACE inhibitors in older women. Clinical 9 pharmacology and therapeutics. 2008; 83: 422-9. 10 24. Hilmer SN, Mager DE, Simonsick EM, et al. Drug burden index score and functional 11 decline in older people. Am J Med. 2009; 122: 1142-49 e1-2. 12 25. Landi F, Russo A, Liperoti R, et al. Anticholinergic drugs and physical function 13 14 among frail elderly population. Clinical pharmacology and therapeutics. 2007; 81: 235-41. 15 26. Lampela P, Lavikainen P, Garcia-Horsman JA, et al. Anticholinergic drug use, serum 16 anticholinergic activity, and adverse drug events among older people: a population-based 17 study. Drugs & aging. 2013; 30: 321-30. 18 27. Studenski S,For Perera S,peer Patel K, et reviewal. Gait speed and only survival in older adults. Jama. 19 2011; 305: 50-58. 20 21 28. Guralnik JM, Ferrucci L, Simonsick EM, et al. Lower-Extremity Function in 22 Persons over the Age of 70 Years as a Predictor of Subsequent Disability. New England 23 Journal of Medicine. 1995; 332: 556-62. 24 29. Cooper R, Kuh D, Cooper C, et al. Objective measures of physical capability and 25 subsequent health: a systematic review. Age and ageing. 2011; 40: 14-23. 26 30. Bohannon RW. Hand-grip dynamometry predicts future outcomes in aging adults. 27 Journal of geriatric physical therapy. 2008; 31: 3-10. 28 29 31. Guralnik JM, LaCroix AZ, Abbott RD, et al. Maintaining mobility in late life. I. 30 Demographic characteristics and chronic conditions. American journal of epidemiology. 31 1993; 137: 845-57. 32 32. Australian Bureau of Statistics. Population Projections, Australia, 2012 (Base) to 33 2101, catalogue no. 3222.0. ACT: Australian Bureau of Statistics, 2013. 34 33. Villafane JH, Valdes K, Bertozzi L, et al. Minimal Clinically Important Difference of 35 Grip and Pinch Strength in Women With Thumb Carpometacarpal Osteoarthritis When 36 37 Compared to Healthy Subjects. Rehabilitation nursing : the official journal of the Association 38 of Rehabilitation Nurses. 2014. 39 34. Perera S, Mody SH, Woodman RC, et al. Meaningful change and responsiveness in 40 common physical performance measures in older adults. Journal of the American Geriatrics 41 Society. 2006; 54: 743-9. 42 35. Sasaki H, Kasagi F, Yamada M, et al. Grip strength predicts cause-specific mortality 43 44 in middle-aged and elderly persons. Am J Med. 2007; 120: 337-42. 45 36. Fox C, Smith T, Maidment I, et al. Effect of medications with anti-cholinergic 46 properties on cognitive function, delirium, physical function and mortality: a systematic 47 review. Age and ageing. 2014; 43: 604-15. 48 37. Gnjidic D, Le Couteur DG, Naganathan V, et al. Effects of drug burden index on 49 cognitive function in older men. Journal of clinical psychopharmacology. 2012; 32: 273-7. 50 38. Trzepacz PT, Hochstetler H, Wang S, et al. Relationship between the Montreal 51 52 Cognitive Assessment and Mini-mental State Examination for assessment of mild cognitive 53 impairment in older adults. BMC geriatrics. 2015; 15: 107. 54 39. Feinberg M. The Problems of Anticholinergic Adverse Effects in Older Patients. 55 Drugs & aging. 1993; 3: 335-48. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 26

18 1 2 3 4 5 6 Figure 1: The hypothesised relationship between medicines, medicine-induced deterioration, 7 8 9 frailty and adverse events. 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 26 BMJ Open

19 1 2 3 Table 1. Baseline characteristics of users and non-users of medicines with anticholinergic or 4 5 sedative properties 6 7 Baseline Total Use of Use of Use of Non- p-value 8 characteristics (n=2087) anticho- sedatives anticholinergics users 9 linergics (n=235, and sedatives (n=1133, 10 (n=383, 11.3%) (n=336, 16.1%) 54.3%) 11 18.3%) 12 13 Age, 78.2 (6.7) 80.3 (6.6) 79.1 78.5 (6.9) 77.2 <0.001 14 mean (SD) years (6.1) (6.6) 15 16 Sex (male), 1056 235 109 128 (38.1) 584 <0.001 17 n (%) (50.5) (61.36) (46.4) (51.5) 18 For peer review only 19 20 Body mass 26.1 (4.1) 26.2 (4.3) 26.2 26.1 (4.4) 26.0 0.852 21 index, mean (3.7) (4) 22 (SD) kg/m2 23 24 Community 1961 356 (93) 222 293 (87.2) 1090 <0.001 25 living, n (%) (93.9) (94.5) (96.2) 26 27 28 Current smoker, 176 (8.4) 24 (6.3) 23 34 (10.2) 95 0.441 29 n (%) (9.8) (8.4) 30 31 MMSE score, 26.9 (4.2) 26.4 (4.5) 26.5 26.4 (4.8) 27.2 <0.001 32 mean (SD) (4.7) (3.7) 33 34 Presence of 325 70 (18.3) 39 63 (18.8) 153 0.035 35 36 cognitive (15.6) (16.6) (13.5) 37 impairment, 38 n (%)* 39 40 CES-D score, 8.2 (7.4) 9.1 (7.5) 9.7 11.7 (8.7) 6.6 <0.001 41 mean (SD) (7.4) (6.5) 42 43 44 Presence of 295 63 (16.5) 36 87 (25.9) 109 <0.001 45 depressive (14.1) (15.3) (9.6) 46 symptoms, 47 n (%)# 48 49 Functional 2.2 (1.6) 2.9 (1.7) 2.2 2.9 (1.7) 1.7 <0.001 50 51 comorbidity (1.5) (1.4) 52 index, mean 53 (SD) 54 CES-D: Center for Epidemiologic Studies Depression; MMSE: Mini Mental State 55 Examination; SD: standard deviation 56 * Presence of cognitive impairment was defined as MMSE score of less than 24. 57 # Presence of depressive symptoms was defined as CES-D score of 16 or more. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 26

20 1 2 3 Table 2. Prevalence of use of drugs with anticholinergic or sedative effects* 4 5 Drug group Sedative or anticholinergic Frequency 6 7 (%) 8 Benzodiazepines 312 (14.9) 9 10 Temazepam (N05CD07) Sedative 88 (4.2) 11 Oxazepam (N05BA04) Sedative 84 (4.0) 12 Nitrazepam (N05CD02) Sedative 76 (3.6) 13 Diazepam (N05BA01) Sedative 51 (2.4) 14 Diuretics: Frusemide (C03CA01) Anticholinergic 287 (13.8) 15 H2-receptor anatagonists Anticholinergic 201 (9.7) 16 Ranitidine (A02BA02) Anticholinergic 114 (5.5) 17 Cimetidine (A02BA01) Anticholinergic 87 (4.2) 18 Cardiac glycosides:For Digoxin peer (C01AA05) reviewAnticholinergic only 179 (8.6) 19 Antidepressants 138 (6.6) 20 Doxepin (N06AA12) Sedative, Anticholinergic 48 (2.3) 21 22 Drugs for obstructive airway diseases Anticholinergic 65 (3.1) 23 Antihistamines Sedative, Anticholinergic 64 (3.1) 24 Antipsychotics Sedative, Anticholinergic 54 (2.6) 25 Antihypertensives Sedative 47 (2.3) 26 Opioids Sedative 53 (2.5) 27 Antiepileptics Sedative, Anticholinergic 36 (1.7) 28 29 Drugs for functional gastrointestinal Anticholinergic 24 (1.2) 30 disorders 31 Antigout preparations Anticholinergic 15 (0.7) 32 Anti-parkinson drugs Sedative, Anticholinergic 10 (0.5) 33 Other nervous system drugs Sedative, Anticholinergic 7 (0.3) 34 Antidiarrheals Anticholinergic 4 (0.2) 35 Drugs for bipolar disorder Anticholinergic 4 (0.2) 36 37 Antimigraine preparations Sedative 3 (0.1) 38 Cough suppresants Sedative 3 (0.1) 39 Antiarrhythmics, class I and III Anticholinergic 2 (0.1) 40 Muscle relaxant Sedative, Anticholinergic 1 (0.1) 41 42 *For some ATC groups, not all medicines within the group have anticholinergic or sedative 43 properties. Full list of medicines are included in Additional file 1. 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 26 BMJ Open

21 1 2 3 Table 3. Adjusted means (95% confidence interval) of physical function and cognitive function according to use of medicines with 4 anticholinergic or sedative properties 5 6 Non-users Anticholinergics Mean p-value Sedatives Mean difference p-value Sedatives and Mean p-value 7 (n=1133) (n=383) difference (n=235) (95% CI) anticholinergics difference 8 (95% CI) (n=336) (95% CI) 9 Physical 10 function 11 Grip strength, 29.4 29.2 -0.27 0.67 27.5 -1.92 0.020 28.3 -1.18 0.12 12 malea (27.1, 31.8) (26.8, 31.6)For(-1.47, peer 0.94) review(24.8, 30.2) (-3.54, -0.30)only (25.7, 30.8) (-2.68, 0.32) 13 Grip strength, 18.6 17.2 -1.43 0.013 18.6 -0.01 0.99 17.8 -0.80 0.100 14 femalea (17.1, 20.1) (15.5, 18.8) (-2.56, -0.31) (16.9, 20.3) (-1.14, 1.12) (16.3, 19.3) (-1.76, 0.15) 15 Walking 0.7 0.7 -0.05 0.0038 0.7 -0.05 0.0098 0.6 -0.08 <0.001 16 speedb (0.7, 0.8) (0.6, 0.7) (-0.09, -0.02) (0.6, 0.7) (-0.10, -0.01) (0.6, 0.7) (-0.11, -0.04) 17 Chair standsb 14.1 15.0 0.90 0.022 14.6 0.49 0.29 14.6 0.48 0.24 18 (12.5, 15.6) (13.4, 16.5) (0.13, 1.68) (12.9, 16.3) (-0.41, 1.4) (12.9, 16.2) (-0.32, 1.28) 19 ADLb 0.6 0.7 0.07 0.25 0.7 0.05 0.43 0.7 0.04 0.48 20 (0.5, 0.8) (0.5, 0.9) (-0.05, 0.18) (0.5, 0.9) (-0.08, 0.19) (0.5, 0.9) (-0.08, 0.16) b 21 IADL 1.1 1.4 0.22 0.020 1.4 0.23 0.043 1.6 0.42 <0.001 22 (0.9, 1.4) (1.1, 1.7) (0.04, 0.41) (1.0, 1.7) (0.01, 0.45) (1.3, 1.9) (0.22, 0.61) 23 Cognitive 24 function MMSEc 26.4 26.5 0.05 0.84 26.7 0.34 0.20 26.6 0.24 0.31 25 (25.7, 27.1) (25.7, 27.2) (-0.40, 0.49) (26.0, 27.5) (-0.18 , 0.86) (25.9, 27.3) (-0.22, 0.70) 26 27 Grip strength (kg) – a higher value indicates stronger muscle strength, Walking speed (m/s) – a higher value indicates faster walking speed, chair stands (s) – 28 29 a higher value indicates slower time required to complete chair stands, ADL (activities of daily living) and IADL (instrumental activities of daily living)– a 30 higher value indicates higher impairment. a 31 Adjusted for age, body mass index, residential status, smoking status, cognitive impairment, depressive symptoms and functional comorbidity index b 32 Adjusted for age, gender, body mass index, residential status, smoking status, cognitive impairment, depressive symptoms and functional comorbidity index 33 a Adjusted for age, gender, body mass index, residential status, smoking status, depressive symptoms and functional comorbidity index 34 35 36 37 38 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 22 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 26 BMJ Open

1 2 3 Additional file 1: List of medicines with anticholinergic or sedative properties included in the 4 5 study. 6 7 Drug group (ATC code) Anticholinergic or Frequency 8 sedative (%) 9 Benzodiazepines (N05BA, N05CD) 10 Temazepam (N05CD07) Sedative 88 (4.2) 11 Oxazepam (N05BA04) Sedative 84 (4.0) 12 Nitrazepam (N05CD02) Sedative 76 (3.6) 13 Diazepam (N05BA01) Sedative 51 (2.4) 14 Flunitrazepam (N05CD03) Sedative 11 (0.5) 15 Bromazepam (N05BA08) Sedative 1 (0.05) 16 Flurazepam (N05CD01) Sedative 1 (0.05) 17 Diuretics (C03) 18 For peer review only 19 Frusemide (C03CA01) Anticholinergic 287 (13.8) 20 H2-receptor anatagonists (A02BA) 21 Ranitidine (A02BA02) Anticholinergic 114 (5.5) 22 Cimetidine (A02BA01) Anticholinergic 87 (4.2) 23 Antidepressants (N06A) 24 Doxepin (N06AA12) Sedative, Anticholinergic 48 (2.3) 25 Amitriptyline (N06AA09) Sedative, Anticholinergic 28 (1.3) 26 Dothiepin (N06AA16) Sedative, Anticholinergic 24 (1.1) 27 Imipramine (N06AA02) Sedative, Anticholinergic 22 (1.1) 28 Mianserin (N06AX03) Sedative 6 (0.3) 29 Fluoxetine (N06AB03) Sedative, Anticholinergic 3 (0.1) 30 Nortriptyline (N06AA10) Sedative, Anticholinergic 3 (0.1) 31 Moclobemide (N06AG02) Sedative 2 (0.1) 32 Clomipramine (N06AA04) Sedative, Anticholinergic 1 (0.05) 33 Tranylcypromine (N06AF04) Sedative 1 (0.05) 34 Cardiac glycosides (C01A) 35 Digoxin (C01AA05) Anticholinergic 179 (8.6) 36 Drugs for obstructive airway diseases (R03) 37 38 Theophylline (R03DA04) Anticholinergic 52 (2.5) 39 Ipratropium (R03BB01) Anticholinergic 13 (0.6) 40 Antihistamines (R06A) 41 Promethazine (R06AD02) Sedative, Anticholinergic 17 (0.8) 42 Astemizole (R06AX11) Sedative 10 (0.5) 43 Dexchlorpheniramine (R06AB02) Sedative, Anticholinergic 9 (0.4) 44 Terfenadine (R06AX12) Sedative 7 (0.3) 45 Diphenhydramine (R06AA02) Sedative, Anticholinergic 5 (0.2) 46 Loratadine (R06AX13) Anticholinergic 4 (0.2) 47 Methdilazine (R06AD04) Sedative 4 (0.2) 48 Azatadine (R06AX09) Sedative 4 (0.2) 49 Chlorphenamine, combinations Sedative 2 (0.1) 50 (R06AB54) 51 Cyproheptadine (R06AX02) Sedative, Anticholinergic 1 (0.05) 52 Diphenylpyraline (R06AA07) Sedative 1 (0.05) 53 Dexchlorpheniramine, Sedative 1 (0.05) 54 combinations (R06AB52) 55 Thiethylperazine (R06AD03) Sedative 1 (0.05) 56 Promethazine, combinations Sedative 1 (0.05) 57 58 (R06AD52) 59 Hydroxyzine (R06AE) Sedative 1 (0.05) 60 Antipsychotics (N05A)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 26

1 2 3 Prochlorperazine (N05AB04) Sedative, Anticholinergic 37 (1.8) 4 Thioridazine (N05AC02) Sedative, Anticholinergic 7 (0.3) 5 Trifluoperazine (N05AB06) Sedative, Anticholinergic 4 (0.2) 6 Pericyazine (N05AC01) Sedative, Anticholinergic 4 (0.2) 7 Fluphenazine (N05AB02) Sedative, Anticholinergic 1 (0.05) 8 9 Haloperidol (N05AD01) Sedative, Anticholinergic 1 (0.05) 10 Antihypertensives (C02) 11 Methyldopa (C02AB01) Sedative 47 (2.3) 12 Clonidine (C02AC01) Sedative 6 (0.3) 13 Opioids (N02A) 14 Codeine/paracetamol (N02AA59) Sedative 30 (1.4) 15 Dextropropoxyphene (N02AC04) Sedative 19 (0.9) 16 Morphine (N02AA01) Sedative 4 (0.2) 17 Antiepileptics (N03A) 18 Phenytoin (N03AB02)For peer reviewSedative only 22 (1.1) 19 Carbamazepine (N03AF01) Sedative, Anticholinergic 5 (0.2) 20 Clonazepam (N03AE01) Sedative 4 (0.2) 21 Phenobarbital (N03AA02) Sedative 3 (0.1) 22 Valproate (N03AG01) Sedative 2 (0.1) 23 Drugs for functional gastrointestinal disorders (A03) 24 Belladona alkaloids (A03BA04) Anticholinergic 14 (0.7) 25 Domperidone (A03FA03) Anticholinergic 6 (0.3) 26 Metoclopramide (A03FA01) Anticholinergic 4 (0.2) 27 Antigout preparations (M04A) 28 29 Colchicine (M04AC01) Anticholinergic 15 (0.7) 30 Anti-parkinson drugs (N04) 31 Bromocriptine (N04BC01) Anticholinergic 3 (0.1) 32 Biperiden (N04AA02) Sedative, Anticholinergic 3 (0.1) 33 Orphenadrine (N04AB02) Sedative, Anticholinergic 2 (0.1) 34 Benzatropine (N04AC01) Sedative, Anticholinergic 1 (0.05) 35 Amantadine (N04BB01) Sedative, Anticholinergic 1 (0.05) 36 Other nervous system drugs (N07) 37 Betahistine (N07CA01) Sedative 5 (0.2) 38 Methadone (N07BC02) Sedative, Anticholinergic 2 (0.1) 39 Antidiarrheals (A07) 40 Loperamide (A07DA03) Anticholinergic 4 (0.2) 41 Drugs for bipolar disorder 42 Lithium (N06AX) Anticholinergic 4 (0.2) 43 Antimigraine preparations (N02C) 44 Pizotifen (N02CX01) Sedative 3 (0.1) 45 Cough suppresants (R05D) 46 Pholcodine (R05DA08 ) Sedative 3 (0.1) 47 48 Antiarrhythmics, class I and III (C01B) 49 Disopyramide (C01BA03) Anticholinergic 2 (0.1) 50 Muscle relaxants (M03) 51 Baclofen (M03BX01) Sedative, Anticholinergic 1 (0.05) 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 26 BMJ Open

1 2 STROBE 2007 (v4) Statement—Checklist of items that should be included in reports of cross-sectional studies 3 4 5 Item Section/Topic Recommendation Reported on page # 6 # 7 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract 1 8 9 (b) Provide in the abstract an informative and balanced summary of what was done and what was found 2 10 Introduction 11 12 Background/rationale 2 Explain the scientificFor background peer and rationale forreview the investigation being reported only 4 13 Objectives 3 State specific objectives, including any prespecified hypotheses 5 14 15 Methods 16 Study design 4 Present key elements of study design early in the paper 6 17 18 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data 6 19 collection 20 Participants 6 (a) Give the eligibility criteria, and the sources and methods of selection of participants 6 21 22 23 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if 7 24 25 applicable 26 Data sources/ 8* For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe 7 27 measurement comparability of assessment methods if there is more than one group 28 Bias 9 Describe any efforts to address potential sources of bias 7 29 30 Study size 10 Explain how the study size was arrived at N/A 31 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and 7 32 why 33 34 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 7 35 (b) Describe any methods used to examine subgroups and interactions N/A 36 37 (c) Explain how missing data were addressed N/A 38 (d) If applicable, describe analytical methods taking account of sampling strategy N/A 39 (e) Describe any sensitivity analyses N/A 40 41 Results 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 26 of 26

1 2 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, 9 3 confirmed eligible, included in the study, completing follow-up, and analysed 4 5 (b) Give reasons for non-participation at each stage N/A 6 (c) Consider use of a flow diagram N/A 7 Descriptive data 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential 9, Table 1 8 9 confounders 10 (b) Indicate number of participants with missing data for each variable of interest N/A 11 Outcome data 15* Report numbers of outcome events or summary measures 9-10 12 Main results 16 (a) Give unadjustedFor estimates peer and, if applicable, confounder-adjustedreview estimates only and their precision (eg, 95% confidence 9-10, Table 2 13 14 interval). Make clear which confounders were adjusted for and why they were included 15 (b) Report category boundaries when continuous variables were categorized 9-10, Table 2 16 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period N/A 17 18 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses N/A 19 Discussion 20 Key results 18 Summarise key results with reference to study objectives 11 21 22 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and 12 23 magnitude of any potential bias 24 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from 12 25 similar studies, and other relevant evidence 26 27 Generalisability 21 Discuss the generalisability (external validity) of the study results 12 28 Other information 29 30 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on 14 31 which the present article is based 32 33 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 34 35 36 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE 37 checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 38 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is available at www.strobe-statement.org. 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open

An analysis of anticholinergic and sedative medicine effects on physical function, cognitive function, appetite and frailty: a cross-sectional study in Australia ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-029221.R1

Article Type: Research

Date Submitted by the 21-May-2019 Author:

Complete List of Authors: Lim, Renly; University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical Sciences Kalisch Ellett, Lisa; University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical Sciences Widagdo, Imaina; University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical Sciences Pratt, Nicole; University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical Sciences Roughead, Elizabeth; University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical Sciences

Primary Subject Epidemiology Heading:

Secondary Subject Heading: Geriatric medicine

Keywords: anticholinergics, frail older adults, sedatives

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 26 BMJ Open

1 2 3 Title: An analysis of anticholinergic and sedative medicine effects on physical function, 4 5 6 cognitive function, appetite and frailty: a cross-sectional study in Australia 7 8 9 10 Author names and institution: 11 12 Renly Lim1, PhD, Lisa M. Kalisch Ellett1, PhD, Imaina S. Widagdo1, PhD, Nicole L. Pratt1, 13 14 15 PhD, Elizabeth E. Roughead1, PhD 16 17 18 For peer review only 19 1 Quality Use of Medicines and Pharmacy Research Centre, School of Pharmacy and Medical 20 21 22 Sciences, University of South Australia, Adelaide, Australia. 23 24 25 26 Corresponding author 27 28 29 Renly Lim 30 31 Quality Use of Medicines and Pharmacy Research Centre, 32 33 School of Pharmacy and Medical Sciences, 34 35 University of South Australia, 36 37 38 GPO Box 2471, Adelaide, SA 5001, Australia. 39 40 E-mail: [email protected] 41 42 Tel: +(61) 883 022 307 43 44 45 Fax: +(61) 8 8302 2466 46 47 ORCID ID: https://orcid.org/0000-0003-4135-2523 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 26

2 1 2 3 Abstract 4 5 6 7 Objective 8 To test the association between use of medicines with anticholinergic or sedative properties 9 10 and physical function, cognitive function, appetite, and frailty. 11 12 13 14 Design, setting and participants 15 This cross-sectional study analysed baseline data collected as part of the Australian 16 17 Longitudinal Study of Ageing (ALSA), a population based cohort of 2087 participants aged 18 For peer review only 19 65 years or over living in South Australia. 20 21 22 Main outcome measures 23 24 Physical function was measured at baseline using measures including hand grip strength, 25 26 walking speed, chair stands, activities of daily living (ADL), instrumental activities of daily 27 living (IADL). Cognitive function was measured using Mini Mental State Examination 28 29 (MMSE). Appetite was measured using Center for Epidemiologic Studies Depression (CES- 30 31 D) question 2. Frailty was measured using Frailty Index. The association between use of 32 33 anticholinergics or sedatives and physical or cognitive function, appetite, or frailty, was 34 assessed using analysis of covariance and ordinal or binary logistic regression. 35 36 37 38 Results 39 40 Almost half of the population were using anticholinergics or sedatives (n=954, 45.7%). Use 41 of anticholinergics was significantly associated with poorer grip strength, slower walking 42 43 speed, poorer instrumental activities of daily living (IADL) and poorer appetite. Use of 44 45 sedatives was significantly associated with poorer grip strength, slower walking speed and 46 poorer IADL. We found no significant association between medicine use and cognitive 47 48 function. Users of anticholinergics or sedatives were significantly more likely to be frail 49 50 compared with non-users. 51 52 53 Conclusion 54 55 Use of medicines with anticholinergic or sedative properties is significantly associated with 56 57 poorer physical function, poorer appetite and increased frailty. Early identification of signs 58 and symptoms of deterioration associated with medicine use is particularly important in older 59 60 people so that worsening frailty and subsequent adverse events are prevented.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 26 BMJ Open

3 1 2 3 4 5 Keywords: anticholinergics; frail older adults; sedatives 6 7 8 Word count: 2983 9 10 11 12 13 Article Summary 14 15 Strengths and limitations of this study 16 17  Our findings on the association between use of anticholinergics and poorer appetite 18 For peer review only 19 contribute uniquely to the literature. 20 21  We analysed the association between medicines with anticholinergic or sedative 22 properties and grip strength based on sex, which previous studies have not done. 23 24  The outcome measures chosen were both objective and clinically relevant. 25 26  The study is limited by its cross-sectional design. 27 28  It is not possible to identify a cause-effect relationship between use of medicines with 29 anticholinergic or sedative properties and reduced physical function or poorer appetite 30 31 using this study design. 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 26

4 1 2 3 Introduction 4 5 Medicines with anticholinergic or sedative properties are commonly prescribed to older 6 7 people (1), with prevalence estimates ranging from 13 to 42% for medicines with sedative 8 properties (2, 3) and from 8 to 36% for medicines with anticholinergic properties (3, 4). 9 10 Adverse effects of these medicines, like falls and confusion are detrimental to older people. 11 12 This risk is increased due to age-related changes in pharmacokinetics and 13 14 pharmacodynamics, the presence of multiple co-morbidities and the use of multiple 15 medicines with subsequent increased probability of drug interactions (5, 6). The cumulative 16 17 medication burden of anticholinergic and sedative medicines is often unintentional and 18 For peer review only 19 compounded by the fact that many medicines with anticholinergic properties also have 20 21 sedative properties (5). 22 23 24 Use of medicines can directly lead to adverse events (5) which are easily recognised and 25 26 potentially rectifiable. Use of medicines with anticholinergic or sedative properties has also 27 been associated with frailty (7), and frailty may contribute to an increased risk of adverse 28 29 events (8). However, medicines with anticholinergic or sedative properties often have what 30 31 might be considered “minor side effects” which are difficult to detect and frequently 32 33 unrecognised. These side effects, which we describe as medicine-induced deterioration, 34 particularly if the cumulative effect builds slowly over time, is often misattributed as geriatric 35 36 syndromes, frailty or simply changes due to aging. Medicine-induced deterioration may 37 38 include decline in physical function, decline in cognitive function and loss of appetite. While 39 40 these same symptoms may occur independently due to aging, medicines with anticholinergic 41 or sedative properties have side effect profiles that also contribute to these declines. The lack 42 43 of recognition of these signs and symptoms as medicine-induced may subsequently 44 45 contribute to increased risk of frailty and subsequent increased risk of adverse events such as 46 falls and fractures. 47 48 49 50 There are several pharmacological pathways by which medicines with anticholinergic or 51 52 sedative properties may contribute to medicine-induced deterioration. Medicines with 53 sedative properties enhance the effects of the neurotransmitter gamma-aminobutyric acid 54 55 (GABA) at the GABAA receptor in the central nervous system, causing sedative and muscle 56 57 relaxant effects (9). The side effects related to the sedative and muscle-relaxing actions may 58 cause a decline in both cognitive and physical function. Additionally, some medicines with 59 60 sedative properties may cause a decline in cognitive function by blocking the dopamine-D2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 26 BMJ Open

5 1 2 3 receptors in the central nervous system. Similarly, use of medicines with anticholinergic 4 5 properties may lead to medicine-induced deterioration via several pathways. As some 6 7 medicines with anticholinergic properties have sedative properties, these medicines may 8 directly lead to a decline in physical function due to the sedative side effects. Medicines with 9 10 anticholinergic properties may affect cognitive function by blocking the muscarinic receptors 11 12 in the central nervous system which regulates learning and memory (10). Medicines with 13 14 anticholinergic properties block the effects of acetylcholine at the M3-muscarinic receptors of 15 the salivary glands causing dry mouth, which may lead to loss of appetite. 16 17 18 For peer review only 19 We have created a model of the relationship between medicine use, medicine-induced 20 21 deterioration, and frailty (Figure 1). If anticholinergic or sedative medicine-induced 22 deterioration was monitored for and detected in practice, this would provide an opportunity 23 24 for healthcare professionals to prevent deterioration by moderating medicine use, mitigating 25 26 progression to frailty and potentially avoiding subsequent adverse events (Figure 1). Using 27 data from the Australian Longitudinal Study of Ageing (ALSA) (11), the study aims to assess 28 29 (i) the association between use of medicines with anticholinergic or sedative properties and 30 31 medicine-induced deterioration (physical function, cognitive function, appetite) and (ii) the 32 33 association between use of medicines with anticholinergic or sedative properties and frailty. 34 We hypothesized that use of medicines with anticholinergic or sedative properties is 35 36 associated with decline in physical function and decline in cognitive function for 37 38 anticholinergics or sedatives, poorer appetite for anticholinergics, and increased frailty. 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 26

6 1 2 3 Method 4 5 Study population and data collection 6 7 This study analysed baseline data collected as part of the Australian Longitudinal Study of 8 Ageing (ALSA), a population based cohort of 2087 participants aged 65 years or over at 9 10 baseline (1992) living in South Australia (11). Both community-dwelling and people living in 11 12 residential care were eligible and were recruited by stratified random sampling from the 13 14 South Australian electoral roll. At baseline, 93 percent of participants lived in the community 15 and 6 percent lived in residential care (11). Data collection was undertaken by comprehensive 16 17 personal interviews, self-completed questionnaires and clinical assessments on physical 18 For peer review only 19 function. 20 21 22 Medicines with anticholinergic or sedative properties 23 24 We identified medicines with anticholinergic properties using Duran’s scale (12) with the 25 26 addition of other anticholinergic medicines used in Australia based on the lists used in 27 previous Australian studies (1, 13). Medicines with sedative properties were identified from 28 29 the Australian Pharmaceutical Formulary and Handbook. Medicines with sedative properties 30 31 were defined as those which cause sedation, and which are required by legislation to be 32 33 dispensed with a specific sedation warning on the pharmacy label. Participants were stratified 34 into one of four groups: i) non-users, ii) users of medicines with anticholinergic properties 35 36 only, iii) users of medicines with sedative properties only or iv) users of medicines with 37 38 sedative and anticholinergic properties. 39 40 41 Measures of physical function, cognitive function, appetite and frailty 42 43 As part of baseline assessments in ALSA (11), participants had their physical function 44 45 objectively assessed using hand grip strength, a timed 8-feet (2.4 metre) walk at a normal 46 pace and repeated chair stands. Hand grip strength was assessed using a handheld 47 48 dynamometer in the dominant hand (14). The maximum grip strength (kg) was recorded 49 50 based on the best of two trials. Walking speed was calculated using distance in meters and 51 52 time in seconds. Participants were instructed to walk at normal pace from a standing start 53 over a distance of 8 feet. Repeated chair stands were conducted by asking participants to rise 54 55 unassisted from a chair five times. Functional status was subjectively assessed by 56 57 interviewers using activities of daily living (ADL) and instrumental activities of daily living 58 (IADL) scores. The ADL includes walking, bathing, personal grooming, dressing, eating, 59 60 using the toilet and getting from bed to chair(15). The IADL encompasses the following

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 26 BMJ Open

7 1 2 3 tasks: housekeeping, meal preparation, telephone use, handling finances, use of transportation 4 5 and shopping(16). Cognitive function was measured using the Mini Mental State 6 7 Examination (MMSE) questionnaire (range 0-30) (17). Appetite was measured using the 8 Center for Epidemiologic Studies Depression (CES-D) question 2 “I did not feel like eating; 9 10 my appetite was poor”, with scores of 0=rarely or none of the time, 1=some or a little of the 11 12 time, 2=occasionally or a moderate amount of time, and 3=most or all of the time. The frailty 13 14 index (8), a multidimensional assessment which included physical, medical, psychological 15 and social factors, was used to measure frailty. The frailty index, which has been validated in 16 17 the Australian setting (18, 19), included 39 variables with a total score range of between 0 18 For peer review only 19 and 1. A higher frailty index indicates worse frailty status. The subjective functional status 20 21 (ADL and IADL) formed part of the frailty index while the objective physical function (grip 22 strength, walking speed, chair stands), cognitive function (MMSE) and appetite (CES-D 23 24 question 2) were not directly measured in the frailty index. 25 26 27 Covariates 28 29 Univariate analysis was performed to determine which measures were significantly 30 31 associated with use of anticholinergics or sedatives. The final analyses were adjusted for age, 32 33 body mass index, residential status, smoking status, cognitive impairment, depressive 34 symptoms and number of co-morbidities using the Functional Comorbidity Index. The 35 36 Functional Comorbidity Index contains a list of diseases reported to be associated with 37 38 function in older populations (20). We excluded depressive symptoms from the Functional 39 40 Comorbidity Index when we used it as a covariate for adjustment as depression was assessed 41 independently. We did not adjust for covariates for the frailty analysis because it is unclear 42 43 how the covariates such as age and comorbidities affect the relationship between medicines 44 45 use and frailty (21). 46 47 48 Statistical analysis 49 50 Descriptive statistics were used to report the baseline characteristics of the study population. 51 52 The one-way analysis of variance was used to compare continuous variables including age, 53 body mass index, Functional Comorbidity Index, MMSE, CES-D, while the chi-square test 54 55 was used to compare categorical variables including sex, smoking, presence of cognitive 56 57 impairment and presence of depression. The prevalence of use of medicines was grouped by 58 Anatomical Therapeutic Chemical (ATC) class (22). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 26

8 1 2 3 Unadjusted analysis of variance and adjusted analysis of covariance were performed to 4 5 compare differences in physical or cognitive function between users and non-users. Given 6 7 that hand grip strength has been shown to be significantly different between men and women 8 with distinct sex specific cut-off points (23), this performance measure was analysed 9 10 separately based on sex. All other analyses were not stratified by sex. Ordinal logistic 11 12 regression was used to compare differences in appetite between users and non-users. Binary 13 14 logistic regression was used to compare differences in frailty between users and non-users. 15 Analysis was performed using SAS version 9.4 for Window (SAS Institute Inc., Cary, NC, 16 17 USA). A priori, a p-value of <0.05 was considered statistically significant. 18 For peer review only 19 20 21 Patient and public involvement 22 There was no direct participant involvement in the present study. We use data collected as 23 24 part of the ALSA study (11). 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 26 BMJ Open

9 1 2 3 Results 4 5 Characteristics of study population 6 7 Table 1 shows the baseline characteristics of study participants. The mean age ± standard 8 deviation of participants was 78 ± 7 years old; 94% were community-dwelling participants 9 10 and 6% lived in residential care. Nearly half of the population were using medicines with 11 12 anticholinergic or sedative properties (n=954, 45.7%); 18.3% were using medicines with 13 14 anticholinergic properties only, 11.3% were using medicines with sedative properties only 15 and 16.1% were using medicines with sedative and anticholinergic properties. The most 16 17 commonly used medicines with sedative properties were benzodiazepines, antidepressants 18 For peer review only 19 and antihistamines (Table 2). The most frequently used medicines with anticholinergic 20 21 properties were frusemide, H2-receptor antagonists and digoxin (Table 2). 22 23 24 Use of medicines with anticholinergic or sedative properties and physical function 25 26 In the unadjusted analyses, users of medicines with anticholinergic properties only, medicines 27 with sedative properties only, or medicines with sedative and anticholinergic properties had 28 29 significantly poorer function in all physical function measures (p<0.05). 30 31 32 33 After adjusting for covariates, use of medicines with anticholinergic properties only was 34 significantly associated with poorer grip strength (women only, mean difference M -1.43 35 diff 36 95% confidence interval CI -2.56, -0.31), poorer performance on chair stands (Mdiff 0.90, 37 38 95% CI 0.13, 1.68) slower walking speed (Mdiff -0.05, 95% CI -0.09, -0.02), and poorer IADL 39 40 score (Mdiff 0.22, 95% CI 0.04, 0.41) (Table 3). After adjusting for covariates, use of 41 medicines with sedative properties only was significantly associated with poorer grip strength 42 43 (men only, Mdiff -1.92, 95% CI -3.54, -0.30), slower walking speed (Mdiff -0.05, 95% CI -0.10, 44 45 -0.01) and poorer IADL score (Mdiff 0.23, 95% CI 0.01, 0.45) (Table 3). After adjusting for 46 covariates, use of medicines with anticholinergic and sedative properties was significantly 47 48 associated with slower walking speed (Mdiff -0.08, 95% CI -0.11, -0.04), and poorer IADL 49 50 score (Mdiff 0.42, 95% CI 0.22, 0.61) (Table 3). 51 52 53 Use of medicines with anticholinergic or sedative properties and cognitive function 54 55 Use of medicines with anticholinergic or sedative properties was significantly associated with 56 57 poorer cognitive function in the unadjusted analysis (p<0.05), but the associations were not 58 significant after adjusting for covariates (p>0.05) (Table 3). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 26

10 1 2 3 4 5 Use of medicines with anticholinergic properties and appetite 6 7 Use of medicines with anticholinergic properties was significantly associated with poorer 8 appetite in both the unadjusted (OR 2.25, 95% CI 1.79, 2.82) and adjusted analysis (OR 1.77, 9 10 95% CI 1.27, 2.45). 11 12 13 14 Use of medicines with anticholinergic or sedative properties and frailty 15 Participants who used anticholinergics or sedatives had three times or more the odds of being 16 17 frail compared with non-users (anticholinergics only: OR 3.9, 95% CI 2.9, 5.3; sedatives 18 For peer review only 19 only: OR 3.3, 95% CI 2.3, 4.8; sedatives and anticholinergics: OR 6.2, 95% CI 4.6, 8.5). 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 26 BMJ Open

11 1 2 3 Discussion 4 5 Our results showed that use of medicines with anticholinergic or sedative properties was 6 7 significantly associated with poorer physical function and poorer appetite. In addition, we 8 found that participants who used medicines with anticholinergic or sedative properties were 9 10 significantly more likely to be frail. We have previously demonstrated, using a longitudinal 11 12 design in the same study population that frail older people (as identified by the frailty index), 13 14 were more likely to have adverse events at follow-up including an increased risk of mortality 15 (OR 3.2, 95% CI 2.4, 4.1), hospitalisation (OR 2.3, 95% CI 1.7, 3.0), nursing home 16 17 admission (OR 3.3, 95% CI 1.6, 7.0) and fall (OR 3.4, 95% CI 2.7, 4.3) (8). Collectively, our 18 For peer review only 19 results support the hypothesis that use of medicines with anticholinergic or sedative 20 21 properties may contribute to frailty via the intermediary pathways of deterioration associated 22 with medicine use or may directly contribute to frailty, leading to an increased risk of adverse 23 24 events (Figure 1). 25 26 27 Our findings on the association of anticholinergics or sedatives with physical function are 28 29 similar to another Australian cross-sectional study (3). Similar associations have also been 30 31 observed in studies involving community-dwelling older adults in United States (24, 25), 32 33 Italy (26) and Finland (27). A growing body of evidence has shown that limitations in 34 physical function in older people (a component of frailty), predict risk of fracture, 35 36 complications or length of hospital stay, disability and mortality (28-31). An adequate degree 37 38 of mobility is crucial for independent living and maintenance of quality of life in older people 39 40 (32). With the number of adults aged over 65 years old in Australia projected to rise from 41 14% in 2012 to 27% in 2101 (33), an important goal of geriatric medicine is to reduce a 42 43 patient’s medication burden to prevent functional impairment associated with medicine use. 44 45 46 The mean difference of 1.9kg in adjusted grip strength between users versus non-users of 47 48 medicines with sedative properties in men was statistically significant. Similarly, the mean 49 50 difference in adjusted grip strength between users versus non-users of medicines with 51 52 anticholinergic properties (1.4kg) in women was statistically significant and clinically 53 relevant (34). Use of anticholinergics or sedatives was associated with a difference in walking 54 55 speed of 0.05-0.08m/s between users and non-users, which was considered to be a clinically 56 57 meaningful change (35). Grip strength and walking speed are two commonly included 58 components of frailty, and have been shown to relate to clinically significant outcomes such 59 60 as functional disability (28, 29) and mortality (36) in older people.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 26

12 1 2 3 4 5 Our results indicate that use of anticholinergics or sedatives was not significantly associated 6 7 with poorer cognitive function. While studies in other countries have reported contrasting 8 results (37), our findings are in keeping with another Australian cross-sectional study 9 10 involving 1705 community-dwelling man which reported that anticholinergic and sedative 11 12 medicine use as measured by the drug burden index was not associated with limitations in 13 14 cognitive function (38). It may be that use of anticholinergics or sedatives is not associated 15 with poorer cognitive function in our study population, however, it may also be that the 16 17 MMSE is not sensitive enough to detect mild cognitive impairment (39). Other measures, 18 For peer review only 19 such as the Montreal Cognitive Assessment tool have been shown to be better for detecting 20 21 mild cognitive impairment. The lack of association between medicines with anticholinergic 22 or sedative properties and poorer cognitive function should be interpreted with caution due to 23 24 the high mean MMSE score (27 points) and the small percentage of participants with 25 26 cognitive impairment (16%) in our population. In addition, the dose and duration of use of 27 medicines with anticholinergic or sedative properties could have been below the threshold 28 29 required to decrease cognitive function. Most patients in our study population used only one 30 31 medicine with anticholinergic or sedative properties. 32 33 34 Our findings on the association between use of anticholinergics and poorer appetite 35 36 contribute uniquely to the literature. While a cause-effect relationship cannot be established, 37 38 it is widely known that anticholinergics can cause dry mouth (40) and it is plausible that this 39 40 contributed to our finding. In older people, loss of appetite due to dry mouth is difficult to 41 detect and loss of appetite may be misattributed to be part of the natural ageing process. It is 42 43 likely that loss of appetite due to medicines-induced dry mouth is an under-detected problem; 44 45 however this topic has not been well researched. 46 47 48 There are several strengths to our study. We proposed an intermediary pathway by which 49 50 medicines use contributes to frailty in older people. In analysing the intermediary pathways 51 52 of deterioration associated with medicine use, we assessed grip strength, chair stands and 53 walking speed all of which were not variables included in our calculation of frailty. We 54 55 analysed the association between medicines with anticholinergic or sedative properties and 56 57 grip strength based on sex, which previous studies have not done. The outcome measures 58 chosen were both objective and clinically relevant. 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 26 BMJ Open

13 1 2 3 The study is limited by its cross-sectional design. It is not possible to identify a cause-effect 4 5 relationship between use of medicines with anticholinergic or sedative properties and reduced 6 7 physical function or poorer appetite using this study design. We did not differentiate between 8 potent and probable anticholinergic medicines, or between peripherally and centrally acting 9 10 medicines. Another limitation is that we did not examine medication adherence or the effects 11 12 of dose or the duration of treatment. Although we adjusted our analysis for various potential 13 14 confounding factors, there is a possibility of residual confounding and confounding by 15 indication. Although the CES-D has been validated, the question on appetite has not been 16 17 validated by itself. The method of assessing appetite may be subject to recall bias. 18 For peer review only 19 20 21 In conclusion, use of medicines with anticholinergic or sedative properties is significantly 22 associated with poorer physical function, poorer appetite and increased frailty. Early 23 24 identification of signs and symptoms of deterioration associated with medicine use is 25 26 particularly important in older people so that worsening frailty and subsequent adverse events 27 are prevented. Grip strength, chair stands and walking speed are simple assessment tools that 28 29 could be used routinely in general practice or pharmacy practice to monitor for deterioration 30 31 in persons considered at risk of frailty. Use of these tools would provide an objective measure 32 33 by which clinicians could assess deterioration associated with medicine use. Clinicians and 34 pharmacists should review use of medicines with anticholinergic or sedative properties in the 35 36 older population who are at the highest risk of deterioration associated with medicine use. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 26

14 1 2 3 Funding 4 5 This study used data collected as part of the Australian Longitudinal Study of Ageing 6 7 (ALSA). ALSA was initially funded by a grant from the US National Institute of Health (AG 8 08523-02), with additional funding from the South Australian government, Flinders 9 10 University and other NGOs. Subsequent funding has been provided by the Australian 11 12 Research Council (ARC-LP 0669272, ARC-LP 100200413, ARC-DP 0879152 and ARC-DP 13 14 130100428) and the National Health and Medical Research Council (NHMRC 179839 and 15 229922). 16 17 18 For peer review only 19 Competing interests 20 21 The authors declare that they have no conflicts of interest. 22 23 24 Author contributions 25 26 RL conceived and designed the study, participated in data analysis and drafted the 27 manuscript. LKE, ISW and NP participated in data analysis and critically revised the 28 29 manuscript for important intellectual content. EER conceived and designed the study, and 30 31 critically revised the manuscript for important intellectual content. All authors read and 32 33 approved the final manuscript. 34 35 36 Acknowledgement 37 38 We thank the participants in the Australian Longitudinal Study of Ageing (ALSA), who have 39 40 given their time over many years, and gratefully acknowledge the work of the project team at 41 the Flinders Centre for Ageing Studies, Flinders University, Adelaide, who carried out the 42 43 ALSA and provided data for this paper. 44 45 46 Data sharing statement 47 48 The data that support the findings of this study are available from the Flinders Centre for 49 50 Ageing Studies, Flinders University but restrictions apply to the availability of these data, 51 52 which were used under license for the current study, and so are not publicly available. Data 53 are however available from the authors upon reasonable request and with permission from 54 55 Flinders Centre for Ageing Studies, Flinders University. 56 57 58 Ethics approval and consent to participate: This study used data collected as part of the 59 60 Australian Longitudinal Study of Ageing (ALSA). Ethical approval for the study was

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 26 BMJ Open

15 1 2 3 obtained from the Committee on Clinical Investigation at Flinders Medical Centre (Minute 4 5 1462: Research Application 9/88). Written informed consent was given by each participant in 6 7 the study. 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 26

16 1 2 3 References 4 5 1. Gadzhanova S, Roughead E, Robinson M. Use of Medicines with Anticholinergic and 6 7 Sedative Effect Before and After Initiation of Anti-Dementia Medications. Drugs - real world 8 outcomes. 2015; 2: 53-60. 9 10 2. Taipale HT, Bell JS, Gnjidic D, et al. Sedative load among community-dwelling 11 12 people aged 75 years or older: association with balance and mobility. Journal of clinical 13 14 psychopharmacology. 2012; 32: 218-24. 15 3. Gnjidic D, Cumming RG, Le Couteur DG, et al. Drug Burden Index and physical 16 17 function in older Australian men. British journal of clinical pharmacology. 2009; 68: 97-105. 18 For peer review only 19 4. Carriere I, Fourrier-Reglat A, Dartigues JF, et al. Drugs with anticholinergic 20 21 properties, cognitive decline, and dementia in an elderly general population: the 3-city study. 22 Arch Intern Med. 2009; 169: 1317-24. 23 24 5. Mintzer J, Burns A. Anticholinergic side-effects of drugs in elderly people. Journal of 25 26 the Royal Society of Medicine. 2000; 93: 457-62. 27 6. McLean AJ, Le Couteur DG. Aging biology and geriatric clinical pharmacology. 28 29 Pharmacological reviews. 2004; 56: 163-84. 30 31 7. Gnjidic D, Hilmer SN, Blyth FM, et al. High-risk prescribing and incidence of frailty 32 33 among older community-dwelling men. Clinical pharmacology and therapeutics. 2012; 91: 34 521-8. 35 36 8. Widagdo IS, Pratt N, Russell M, et al. Predictive performance of four frailty measures 37 38 in an older Australian population. Age and ageing. 2015; 44: 967-72. 39 40 9. Rudolph U, Crestani F, Benke D, et al. Benzodiazepine actions mediated by specific 41 gamma-aminobutyric acid(A) receptor subtypes. Nature. 1999; 401: 796-800. 42 43 10. Volpicelli LA, Levey AI. Muscarinic acetylcholine receptor subtypes in cerebral 44 45 cortex and hippocampus. Progress in brain research. 2004; 145: 59-66. 46 11. Flinders Centre for Ageing Studies. ALSA: The Australian Longitudinal Study of 47 48 Ageing. 2016. 49 50 12. Duran CE, Azermai M, Vander Stichele RH. Systematic review of anticholinergic risk 51 52 scales in older adults. European journal of clinical pharmacology. 2013; 69: 1485-96. 53 13. Australian Government Department of Veterans' Affairs. Topic 39: Thinking clearly 54 55 about the anticholinergic burden., 2014. 56 57 14. Miller MD GL, Crotty M, Harrison, JE, Andrews GR. A clinically relevant criterion 58 for grip strength: relationship with falling in a sample of older adults. Nutr Diet. 2003; 61: 59 60 248-52.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 26 BMJ Open

17 1 2 3 15. Katz S, Downs TD, Cash HR, et al. Progress in development of the index of ADL. 4 5 The Gerontologist. 1970; 10: 20-30. 6 7 16. Lawton MP, Brody EM. Assessment of older people: self-maintaining and 8 instrumental activities of daily living. The Gerontologist. 1969; 9: 179-86. 9 10 17. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for 11 12 grading the cognitive state of patients for the clinician. Journal of psychiatric research. 1975; 13 14 12: 189-98. 15 18. Widagdo I, Pratt N, Russell M, et al. How common is frailty in older Australians? 16 17 Australasian journal on ageing. 2015; 34: 247-51. 18 For peer review only 19 19. Widagdo IS, Pratt N, Russell M, et al. Construct Validity of Four Frailty Measures in 20 21 an Older Australian Population: A Rasch Analysis. The Journal of frailty & aging. 2016; 5: 22 78-81. 23 24 20. Groll DL, To T, Bombardier C, et al. The development of a comorbidity index with 25 26 physical function as the outcome. Journal of clinical epidemiology. 2005; 58: 595-602. 27 21. Rodríguez-Mañas L, Féart C, Mann G, et al. Searching for an Operational Definition 28 29 of Frailty: A Delphi Method Based Consensus Statement. The Frailty Operative Definition- 30 31 Consensus Conference Project. The Journals of Gerontology Series A: Biological Sciences 32 33 and Medical Sciences. 2013; 68: 62-67. 34 22. WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index 2017. 35 36 23. Bahat G, Tufan A, Tufan F, et al. Cut-off points to identify sarcopenia according to 37 38 European Working Group on Sarcopenia in Older People (EWGSOP) definition. Clinical 39 40 nutrition. 2016. 41 24. Cao YJ, Mager DE, Simonsick EM, et al. Physical and cognitive performance and 42 43 burden of anticholinergics, sedatives, and ACE inhibitors in older women. Clinical 44 45 pharmacology and therapeutics. 2008; 83: 422-9. 46 25. Hilmer SN, Mager DE, Simonsick EM, et al. Drug burden index score and functional 47 48 decline in older people. Am J Med. 2009; 122: 1142-49 e1-2. 49 50 26. Landi F, Russo A, Liperoti R, et al. Anticholinergic drugs and physical function 51 52 among frail elderly population. Clinical pharmacology and therapeutics. 2007; 81: 235-41. 53 27. Lampela P, Lavikainen P, Garcia-Horsman JA, et al. Anticholinergic drug use, serum 54 55 anticholinergic activity, and adverse drug events among older people: a population-based 56 57 study. Drugs & aging. 2013; 30: 321-30. 58 28. Studenski S, Perera S, Patel K, et al. Gait speed and survival in older adults. Jama. 59 60 2011; 305: 50-58.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 26

18 1 2 3 29. Guralnik JM, Ferrucci L, Simonsick EM, et al. Lower-Extremity Function in 4 5 Persons over the Age of 70 Years as a Predictor of Subsequent Disability. New England 6 7 Journal of Medicine. 1995; 332: 556-62. 8 30. Cooper R, Kuh D, Cooper C, et al. Objective measures of physical capability and 9 10 subsequent health: a systematic review. Age and ageing. 2011; 40: 14-23. 11 12 31. Bohannon RW. Hand-grip dynamometry predicts future outcomes in aging adults. 13 14 Journal of geriatric physical therapy. 2008; 31: 3-10. 15 32. Guralnik JM, LaCroix AZ, Abbott RD, et al. Maintaining mobility in late life. I. 16 17 Demographic characteristics and chronic conditions. American journal of epidemiology. 18 For peer review only 19 1993; 137: 845-57. 20 21 33. Australian Bureau of Statistics. Population Projections, Australia, 2012 (Base) to 22 2101, catalogue no. 3222.0. ACT: Australian Bureau of Statistics, 2013. 23 24 34. Villafane JH, Valdes K, Bertozzi L, et al. Minimal Clinically Important Difference of 25 26 Grip and Pinch Strength in Women With Thumb Carpometacarpal Osteoarthritis When 27 Compared to Healthy Subjects. Rehabilitation nursing : the official journal of the Association 28 29 of Rehabilitation Nurses. 2014. 30 31 35. Perera S, Mody SH, Woodman RC, et al. Meaningful change and responsiveness in 32 33 common physical performance measures in older adults. Journal of the American Geriatrics 34 Society. 2006; 54: 743-9. 35 36 36. Sasaki H, Kasagi F, Yamada M, et al. Grip strength predicts cause-specific mortality 37 38 in middle-aged and elderly persons. Am J Med. 2007; 120: 337-42. 39 40 37. Fox C, Smith T, Maidment I, et al. Effect of medications with anti-cholinergic 41 properties on cognitive function, delirium, physical function and mortality: a systematic 42 43 review. Age and ageing. 2014; 43: 604-15. 44 45 38. Gnjidic D, Le Couteur DG, Naganathan V, et al. Effects of drug burden index on 46 cognitive function in older men. Journal of clinical psychopharmacology. 2012; 32: 273-7. 47 48 39. Trzepacz PT, Hochstetler H, Wang S, et al. Relationship between the Montreal 49 50 Cognitive Assessment and Mini-mental State Examination for assessment of mild cognitive 51 52 impairment in older adults. BMC geriatrics. 2015; 15: 107. 53 40. Feinberg M. The Problems of Anticholinergic Adverse Effects in Older Patients. 54 55 Drugs & aging. 1993; 3: 335-48. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 26 BMJ Open

19 1 2 3 Figure 1: The hypothesised relationship between medicines, medicine-induced deterioration, 4 5 6 frailty and adverse events. 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 26

20 1 2 3 Table 1. Baseline characteristics of users and non-users of medicines with anticholinergic or 4 5 sedative properties 6 7 Baseline Total Use of Use of Use of Non- p-value 8 characteristics (n=2087) anticho- sedatives anticholinergics users 9 linergics (n=235, and sedatives (n=1133, 10 (n=383, 11.3%) (n=336, 16.1%) 54.3%) 11 18.3%) 12 13 Age, 78.2 (6.7) 80.3 (6.6) 79.1 78.5 (6.9) 77.2 <0.001 14 mean (SD) years (6.1) (6.6) 15 16 Sex (male), 1056 235 109 128 (38.1) 584 <0.001 17 n (%) (50.5) (61.36) (46.4) (51.5) 18 For peer review only 19 20 Body mass 26.1 (4.1) 26.2 (4.3) 26.2 26.1 (4.4) 26.0 0.852 21 index, mean (3.7) (4) 22 (SD) kg/m2 23 24 Community 1961 356 (93) 222 293 (87.2) 1090 <0.001 25 living, n (%) (93.9) (94.5) (96.2) 26 27 28 Current smoker, 176 (8.4) 24 (6.3) 23 34 (10.2) 95 0.441 29 n (%) (9.8) (8.4) 30 31 MMSE score, 26.9 (4.2) 26.4 (4.5) 26.5 26.4 (4.8) 27.2 <0.001 32 mean (SD) (4.7) (3.7) 33 34 Presence of 325 70 (18.3) 39 63 (18.8) 153 0.035 35 36 cognitive (15.6) (16.6) (13.5) 37 impairment, 38 n (%)* 39 40 CES-D score, 8.2 (7.4) 9.1 (7.5) 9.7 11.7 (8.7) 6.6 <0.001 41 mean (SD) (7.4) (6.5) 42 43 44 Presence of 295 63 (16.5) 36 87 (25.9) 109 <0.001 45 depressive (14.1) (15.3) (9.6) 46 symptoms, 47 n (%)# 48 49 Functional 2.2 (1.6) 2.9 (1.7) 2.2 2.9 (1.7) 1.7 <0.001 50 51 comorbidity (1.5) (1.4) 52 index, mean 53 (SD) 54 CES-D: Center for Epidemiologic Studies Depression; MMSE: Mini Mental State 55 Examination; SD: standard deviation 56 * Presence of cognitive impairment was defined as MMSE score of less than 24. 57 # Presence of depressive symptoms was defined as CES-D score of 16 or more. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 26 BMJ Open

21 1 2 3 4 5 Table 2: List of medicines with anticholinergic or sedative properties included in the study. 6 7 Drug group (ATC code) Anticholinergic or Frequency 8 9 sedative (%) 10 Benzodiazepines (N05BA, N05CD) 11 Temazepam (N05CD07) Sedative 88 (4.2) 12 Oxazepam (N05BA04) Sedative 84 (4.0) 13 Nitrazepam (N05CD02) Sedative 76 (3.6) 14 Diazepam (N05BA01) Sedative 51 (2.4) 15 Flunitrazepam (N05CD03) Sedative 11 (0.5) 16 Bromazepam (N05BA08) Sedative 1 (0.05) 17 Flurazepam (N05CD01) Sedative 1 (0.05) 18 Diuretics (C03) For peer review only 19 Frusemide (C03CA01) Anticholinergic 287 (13.8) 20 H2-receptor anatagonists (A02BA) 21 Ranitidine (A02BA02) Anticholinergic 114 (5.5) 22 Cimetidine (A02BA01) Anticholinergic 87 (4.2) 23 Antidepressants (N06A) 24 Doxepin (N06AA12) Sedative, Anticholinergic 48 (2.3) 25 26 Amitriptyline (N06AA09) Sedative, Anticholinergic 28 (1.3) 27 Dothiepin (N06AA16) Sedative, Anticholinergic 24 (1.1) 28 Imipramine (N06AA02) Sedative, Anticholinergic 22 (1.1) 29 Mianserin (N06AX03) Sedative 6 (0.3) 30 Fluoxetine (N06AB03) Sedative, Anticholinergic 3 (0.1) 31 Nortriptyline (N06AA10) Sedative, Anticholinergic 3 (0.1) 32 Moclobemide (N06AG02) Sedative 2 (0.1) 33 Clomipramine (N06AA04) Sedative, Anticholinergic 1 (0.05) 34 Tranylcypromine (N06AF04) Sedative 1 (0.05) 35 Cardiac glycosides (C01A) 36 Digoxin (C01AA05) Anticholinergic 179 (8.6) 37 Drugs for obstructive airway diseases (R03) 38 Theophylline (R03DA04) Anticholinergic 52 (2.5) 39 Ipratropium (R03BB01) Anticholinergic 13 (0.6) 40 Antihistamines (R06A) 41 Promethazine (R06AD02) Sedative, Anticholinergic 17 (0.8) 42 Astemizole (R06AX11) Sedative 10 (0.5) 43 Dexchlorpheniramine (R06AB02) Sedative, Anticholinergic 9 (0.4) 44 45 Terfenadine (R06AX12) Sedative 7 (0.3) 46 Diphenhydramine (R06AA02) Sedative, Anticholinergic 5 (0.2) 47 Loratadine (R06AX13) Anticholinergic 4 (0.2) 48 Methdilazine (R06AD04) Sedative 4 (0.2) 49 Azatadine (R06AX09) Sedative 4 (0.2) 50 Chlorphenamine, combinations Sedative 2 (0.1) 51 (R06AB54) 52 Cyproheptadine (R06AX02) Sedative, Anticholinergic 1 (0.05) 53 Diphenylpyraline (R06AA07) Sedative 1 (0.05) 54 Dexchlorpheniramine, Sedative 1 (0.05) 55 combinations (R06AB52) 56 Thiethylperazine (R06AD03) Sedative 1 (0.05) 57 Promethazine, combinations Sedative 1 (0.05) 58 (R06AD52) 59 Hydroxyzine (R06AE) Sedative 1 (0.05) 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 26

22 1 2 3 Antipsychotics (N05A) 4 Prochlorperazine (N05AB04) Sedative, Anticholinergic 37 (1.8) 5 Thioridazine (N05AC02) Sedative, Anticholinergic 7 (0.3) 6 Trifluoperazine (N05AB06) Sedative, Anticholinergic 4 (0.2) 7 Pericyazine (N05AC01) Sedative, Anticholinergic 4 (0.2) 8 9 Fluphenazine (N05AB02) Sedative, Anticholinergic 1 (0.05) 10 Haloperidol (N05AD01) Sedative, Anticholinergic 1 (0.05) 11 Antihypertensives (C02) 12 Methyldopa (C02AB01) Sedative 47 (2.3) 13 Clonidine (C02AC01) Sedative 6 (0.3) 14 Opioids (N02A) 15 Codeine/paracetamol (N02AA59) Sedative 30 (1.4) 16 Dextropropoxyphene (N02AC04) Sedative 19 (0.9) 17 Morphine (N02AA01) Sedative 4 (0.2) 18 Antiepileptics (N03A)For peer review only 19 Phenytoin (N03AB02) Sedative 22 (1.1) 20 Carbamazepine (N03AF01) Sedative, Anticholinergic 5 (0.2) 21 Clonazepam (N03AE01) Sedative 4 (0.2) 22 Phenobarbital (N03AA02) Sedative 3 (0.1) 23 Valproate (N03AG01) Sedative 2 (0.1) 24 Drugs for functional gastrointestinal disorders 25 (A03) 26 Belladona alkaloids (A03BA04) Anticholinergic 14 (0.7) 27 Domperidone (A03FA03) Anticholinergic 6 (0.3) 28 29 Metoclopramide (A03FA01) Anticholinergic 4 (0.2) 30 Antigout preparations (M04A) 31 Colchicine (M04AC01) Anticholinergic 15 (0.7) 32 Anti-parkinson drugs (N04) 33 Bromocriptine (N04BC01) Anticholinergic 3 (0.1) 34 Biperiden (N04AA02) Sedative, Anticholinergic 3 (0.1) 35 Orphenadrine (N04AB02) Sedative, Anticholinergic 2 (0.1) 36 Benzatropine (N04AC01) Sedative, Anticholinergic 1 (0.05) 37 Amantadine (N04BB01) Sedative, Anticholinergic 1 (0.05) 38 Other nervous system drugs (N07) 39 Betahistine (N07CA01) Sedative 5 (0.2) 40 Methadone (N07BC02) Sedative, Anticholinergic 2 (0.1) 41 Antidiarrheals (A07) 42 Loperamide (A07DA03) Anticholinergic 4 (0.2) 43 Drugs for bipolar disorder 44 Lithium (N06AX) Anticholinergic 4 (0.2) 45 Antimigraine preparations (N02C) 46 Pizotifen (N02CX01) Sedative 3 (0.1) 47 48 Cough suppresants (R05D) 49 Pholcodine (R05DA08 ) Sedative 3 (0.1) 50 Antiarrhythmics, class I and III (C01B) 51 Disopyramide (C01BA03) Anticholinergic 2 (0.1) 52 Muscle relaxants (M03) 53 Baclofen (M03BX01) Sedative, Anticholinergic 1 (0.05) 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 26 BMJ Open

23 1 2 3 Table 3. Adjusted means (95% confidence interval) of physical function and cognitive function according to use of medicines with 4 anticholinergic or sedative properties 5 6 Non-users Anticholinergics Mean p-value Sedatives Mean difference p-value Sedatives and Mean p-value 7 (n=1133) (n=383) difference (n=235) (95% CI) anticholinergics difference 8 (95% CI) (n=336) (95% CI) 9 Physical 10 function 11 Grip strength, 29.4 29.2 -0.27 0.67 27.5 -1.92 0.020 28.3 -1.18 0.12 12 malea (27.1, 31.8) (26.8, 31.6)For(-1.47, peer 0.94) review(24.8, 30.2) (-3.54, -0.30)only (25.7, 30.8) (-2.68, 0.32) 13 Grip strength, 18.6 17.2 -1.43 0.013 18.6 -0.01 0.99 17.8 -0.80 0.100 14 femalea (17.1, 20.1) (15.5, 18.8) (-2.56, -0.31) (16.9, 20.3) (-1.14, 1.12) (16.3, 19.3) (-1.76, 0.15) 15 Walking 0.7 0.7 -0.05 0.0038 0.7 -0.05 0.0098 0.6 -0.08 <0.001 16 speedb (0.7, 0.8) (0.6, 0.7) (-0.09, -0.02) (0.6, 0.7) (-0.10, -0.01) (0.6, 0.7) (-0.11, -0.04) 17 Chair standsb 14.1 15.0 0.90 0.022 14.6 0.49 0.29 14.6 0.48 0.24 18 (12.5, 15.6) (13.4, 16.5) (0.13, 1.68) (12.9, 16.3) (-0.41, 1.4) (12.9, 16.2) (-0.32, 1.28) 19 ADLb 0.6 0.7 0.07 0.25 0.7 0.05 0.43 0.7 0.04 0.48 20 (0.5, 0.8) (0.5, 0.9) (-0.05, 0.18) (0.5, 0.9) (-0.08, 0.19) (0.5, 0.9) (-0.08, 0.16) b 21 IADL 1.1 1.4 0.22 0.020 1.4 0.23 0.043 1.6 0.42 <0.001 22 (0.9, 1.4) (1.1, 1.7) (0.04, 0.41) (1.0, 1.7) (0.01, 0.45) (1.3, 1.9) (0.22, 0.61) 23 Cognitive 24 function MMSEc 26.4 26.5 0.05 0.84 26.7 0.34 0.20 26.6 0.24 0.31 25 (25.7, 27.1) (25.7, 27.2) (-0.40, 0.49) (26.0, 27.5) (-0.18 , 0.86) (25.9, 27.3) (-0.22, 0.70) 26 27 Grip strength (kg) – a higher value indicates stronger muscle strength, Walking speed (m/s) – a higher value indicates faster walking speed, chair stands (s) – 28 29 a higher value indicates slower time required to complete chair stands, ADL (activities of daily living) and IADL (instrumental activities of daily living)– a 30 higher value indicates higher impairment. a 31 Adjusted for age, body mass index, residential status, smoking status, cognitive impairment, depressive symptoms and functional comorbidity index b 32 Adjusted for age, sex, body mass index, residential status, smoking status, cognitive impairment, depressive symptoms and functional comorbidity index 33 a Adjusted for age, sex, body mass index, residential status, smoking status, depressive symptoms and functional comorbidity index 34 35 36 37 38 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 24 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 The hypothesised relationship between medicines, medicine-induced deterioration, frailty and adverse 33 events. 34 35 271x214mm (300 x 300 DPI) 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 26 BMJ Open

1 2 STROBE 2007 (v4) Statement—Checklist of items that should be included in reports of cross-sectional studies 3 4 5 Item Section/Topic Recommendation Reported on page # 6 # 7 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract 1 8 9 (b) Provide in the abstract an informative and balanced summary of what was done and what was found 2 10 Introduction 11 12 Background/rationale 2 Explain the scientificFor background peer and rationale forreview the investigation being reported only 4 13 Objectives 3 State specific objectives, including any prespecified hypotheses 5 14 15 Methods 16 Study design 4 Present key elements of study design early in the paper 6 17 18 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data 6 19 collection 20 Participants 6 (a) Give the eligibility criteria, and the sources and methods of selection of participants 6 21 22 23 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if 7 24 25 applicable 26 Data sources/ 8* For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe 7 27 measurement comparability of assessment methods if there is more than one group 28 Bias 9 Describe any efforts to address potential sources of bias 7 29 30 Study size 10 Explain how the study size was arrived at N/A 31 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and 7 32 why 33 34 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 7 35 (b) Describe any methods used to examine subgroups and interactions N/A 36 37 (c) Explain how missing data were addressed N/A 38 (d) If applicable, describe analytical methods taking account of sampling strategy N/A 39 (e) Describe any sensitivity analyses N/A 40 41 Results 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 26 of 26

1 2 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, 9 3 confirmed eligible, included in the study, completing follow-up, and analysed 4 5 (b) Give reasons for non-participation at each stage N/A 6 (c) Consider use of a flow diagram N/A 7 Descriptive data 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential 9, Table 1 8 9 confounders 10 (b) Indicate number of participants with missing data for each variable of interest N/A 11 Outcome data 15* Report numbers of outcome events or summary measures 9-10 12 Main results 16 (a) Give unadjustedFor estimates peer and, if applicable, confounder-adjustedreview estimates only and their precision (eg, 95% confidence 9-10, Table 2 13 14 interval). Make clear which confounders were adjusted for and why they were included 15 (b) Report category boundaries when continuous variables were categorized 9-10, Table 2 16 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period N/A 17 18 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses N/A 19 Discussion 20 Key results 18 Summarise key results with reference to study objectives 11 21 22 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and 12 23 magnitude of any potential bias 24 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from 12 25 similar studies, and other relevant evidence 26 27 Generalisability 21 Discuss the generalisability (external validity) of the study results 12 28 Other information 29 30 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on 14 31 which the present article is based 32 33 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 34 35 36 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE 37 checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 38 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is available at www.strobe-statement.org. 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60