Anticholinergic Exposure and Risk of Pneumonia in Persons with Alzheimer’s Disease: A Nested Case-Control Study

Abstract

Background: Risk of pneumonia is increased in persons with Alzheimer’s disease (AD). In some studies, anticholinergic drugs (AC) have been associated with an increased pneumonia risk.

Objective: We analyzed the risk of pneumonia associated with ACs in persons with AD.

Methods: We performed a nested case-control study using register-based data from a Finnish nationwide MEDALZ cohort including all community-dwelling persons diagnosed with AD during 2005–2011. Cases were identified based on pneumonia diagnoses (n = 12,442) from hospital discharge and causes of death registers. Up to two controls without pneumonia were matched based on time since AD diagnoses, age, and gender for each case; AC use was measured using Anticholinergic Drug Scale.

Results: Use of AC was associated with an increased risk of pneumonia (adjusted odds ratio (OR) 1.36, 95% confidence interval (CI) 1.29–1.43). However, there was no increased pneumonia risk in persons using level 3 ACs. Incident use was associated with higher risk of pneumonia (OR 2.68, 95% CI 2.15–3.34) than prevalent use (OR 1.48, 95% CI 1.40–1.57). Among persons using cholinesterase inhibitors (AChEIs), risk of pneumonia was increased in persons using also ACs (OR 1.53, 95% CI 1.41–1.66).

Conclusion: ACs were associated with an increased risk of pneumonia in persons with AD, especially at the time of initiation of these drugs. AC use was associated with increased pneumonia risk also in persons using AChEIs. This risk should be carefully considered when treating AD patients.

Keywords

Aged Alzheimer’s disease anticholinergics pneumonia

INTRODUCTION

Pneumonia is common among the older population and is responsible for a large proportion of hospitalizations among the elderly [1]. Its incidence in adults is slightly over 1 per 1000 person-years but increases with age up to 9.9 in persons aged 65–74 years and 29.4 in persons ≥85 years [2, 3]. Aspiration is a prominent mechanism for pneumonia resulting from dysphagia [4]. There are several conditions increasing the risk of pneumonia, including poor nutrition, tobacco use, dysphagia, chronic cardiovascular disease, epilepsy, respiratory conditions such as asthma and chronic obstructive pulmonary disease (COPD), and mental health diagnoses such as depression and bipolar disorder [3].

Alzheimer’s disease (AD) is the most common neurodegenerative disorder among older persons [5], and it causes progressive impairment of cognitive functions. Dementia doubles the risk of pneumonia [3], and pneumonia is one of the most common causes for hospitalization among persons with AD [6]. Dementia increases the risk of pneumonia-related mortality more than two-fold [7].

Pharmacotherapy of AD is based on two different groups of medicines. Cholinesterase inhibitors (AChEI) increase acetylcholine concentration in cholinergic synapses, resulting in small improvements in cognitive function [8]. Memantine has a different mechanism of action, as it blocks NMDA receptors on the glutaminergic system [8].

Anticholinergic drugs (ACs) are commonly used among older people. However, these drugs have several adverse effects including central (e.g., drowsiness, confusion, delirium, and cognitive decline) and peripheral (e.g., dry mouth, constipation, urinary retention, vision changes) adverse effects [9]. In addition, cholinergic activity is lower in older persons with dementia compared with those with normal cognition, making them more susceptible to AC adverse effects [10]. ACs are actually associated with decreased health-related quality of life in older persons with dementia [11]. Furthermore, although the mechanism of action of ACs is opposite to cholinesterase inhibitors and therefore they should not be used together, their concomitant use is common (9–13%) [12 –14]. It has been suggested that AC use may increase pneumonia risk through sedation and altered mental status in addition to peripheral effects such as dryness of mouth (leading to impaired oropharyngeal bolus transport), low levels of thick mucosal secretion and depression of mucociliary transport (increased risk of bacterial growth), and low esophageal sphincter pressure (acid reflux and aspiration) [15, 16].

Recent studies have indicated an increased risk of pneumonia among older persons using ACs [15, 16], but no studies have been conducted among persons with dementia or AD (although use of ACs is frequent in persons with dementia [17]). The objective of our study was to investigate whether the use of ACs increases the risk of pneumonia in older persons with AD.

MATERIALS AND METHODS

Study population

We used data from the MEDALZ-study [18], which includes all 70,718 community-dwelling persons who were diagnosed with AD during 2005–2011 in Finland. These persons were identified from the Special Reimbursement register which includes persons entitled for higher reimbursement of drugs due to certain chronic diseases [19]. Finnish current care guideline on cognitive disorders recommends that all persons with clinically verified AD having no contraindications should be prescribed anti-dementia drugs [20]. The diagnostic process should be conducted according to a predefined protocol and approved by the Social Insurance Institution of Finland in order to receive special reimbursement. The diagnostic process includes computed tomography or MRI scan and confirmation of diagnoses by a neurologist or geriatrician, according to the NINCDS-ADRDA [21] and DSM-IV criteria.

Data sources

Data have been collected from the Finnish nationwide registers. These include the Special Reimbursement register (comorbidities, 1972–2012), the Prescription register (drug use, years 1995–2012), and the Hospital Discharge register (hospitalizations with diagnoses, 1972–2012). Statistics Finland provided causes of death 2005–2012 and socioeconomic data since 1970.

Drug use in the Prescription register is maintained by the Social Insurance Institution. It includes purchases of reimbursed drugs and is classified according to Anatomical Therapeutic Chemical–classification system (ATC) [22]. Drug purchases by community-dwelling persons are recorded in the register whereas drugs used during stays in hospitals and public nursing homes are provided by the caring unit. Previously utilized method PRE2DUP [18, 23] was used to model purchases to drug use periods (when drug use started and ended). The method takes into account regularity of purchases, hospitalizations, and stockpiling of drugs, and is based on sliding averages of daily dose for each ATC code for each person. The PRE2DUP is based on evaluation of personal drug purchasing patterns to calculate for how long each drug purchase will last based on local dose and regularity of purchases. If the purchased amount is not enough to reach next purchase, the drug use period is ended and use is re-started at the next purchase. “Any AC use” periods were derived by combining all overlapping use periods of AC drug substances together.

Cases

We conducted a nested case-control study among persons diagnosed with AD included in the MEDALZ study. Pneumonia diagnoses (N = 14,888) were based on data from the Hospital Discharge and the Causes of Death registers. Pneumonia was defined according to ICD-10 (J10.0, J11.0, J12, J13, J14, J15, J16, J18, and J69.0). Only the first event (hospitalization or death) was considered for each person. Causes of death register records direct (final disease or condition resulting in death) and/or underlying (the disease or injury which initiated the cascade of morbid events leading directly to death) causes of deaths, and those cases where pneumonia was recorded as direct or underlying cause of death were extracted. At the time of pneumonia diagnoses, only community-dwelling persons were included as cases in this study (2,446 cases were excluded due to long hospitalization), resulting in 12,442 cases included in the analyses. National guidelines [24] present basis of diagnosis and treatment of pneumonia in Finland.

Controls

For each pneumonia case for the date of pneumonia diagnoses, up to two control persons without pneumonia were selected from the cohort of AD cases. Cases were available for being controls before their event date, otherwise controls were chosen without replacement. At the matching date, a control person needed to be community-dwelling and without pneumonia since AD diagnoses. A control person was matched based on time since AD diagnoses (≤90 days), age (±2 years), and gender. For 407 cases, only one control person could be identified.

Exposure

Exposure to ACs was defined according to the Anticholinergic Drug Scale [25], which has been widely used (e.g., [16]). It ranks drugs based on AC potential from 0 (no AC activity) to 3 (strong AC). Any use of ACs and the highest level of AC potential was considered based on drugs used during a time period of 14 days before the index date (the date of pneumonia for cases and the corresponding matching date for controls). The highest AC level in use means that a person may also use weaker AC levels concomitantly. Use of AC levels was compared with non-use of any ACs. To examine recent initiations, any AC use was categorized as incident use (use 1–30 days before the index date but no use during 31–365 days), past use (use 31–365 days before but no use 1–30 days before), prevalent use (use 1–30 and 31–365 days before), and no use (no use 1–30 and 31–365 before the index date). Duration of any AC use during the previous year before the index date was categorized as chronic use (≥274 days, i.e., 9 months), non-chronic use (<274 days), and no use (365 days). All drug use periods during the previous year were summed together for each person. As we aimed to better describe the actual burden of ACs, the above-mentioned duration of use was multiplied by the corresponding sum of AC scores (cumulative AC burden, Fig. 1). The cumulative burden was divided into quartiles and compared with no use.

Comorbid conditions

Data on asthma/chronic obstructive pulmonary disease (COPD), cardiovascular disease, diabetes, and epilepsy were collected from the Special Reimbursement register. All diagnoses recorded before the AD diagnosis were considered. Cardiovascular diseases were defined as having one or several of the following: chronic heart failure, arterial hypertension, coronary artery disease, and chronic arrhythmia. Other diagnoses were extracted from Hospital Discharge register according to ICD-10 codes (with corresponding ICD-8 and -9 codes). Schizophrenia (schizophrenia, schizotypal or delusional disorders F20-29), and depression/bipolar disorder (F30-34, F38-39) since 1972 until 5 years before the diagnosis of AD were considered. History of substance abuse was defined as hospitalization based on the diagnoses of alcohol or narcotic use (F10-19) or alcoholic pancreatitis (K860) at any time point before the index date. Similarly we obtained history of hip fracture (S72.0–72.2), stroke (I60–I64), and previous ischemic heart event (I20–I25 or revascularization procedures). In addition, previous pneumonia during one year before the AD diagnoses were collected from Hospital Discharge register. Data on use of proton pump inhibitors (PPIs, A02BC), AChEIs (N06DA), memantine (N06DX01), drugs for obstructive airway diseases (short as asthma medication, R03), oral glucocorticoids (H02AB), and antibiotics (J01) at the index date were collected from the Prescription register.

Socioeconomic position (categorized as high, medium, low, or unknown) was defined as the highest position recorded for study participants in their middle age (45–55 years old), according to classification by the Statistics Finland. The highest class included entrepreneurs and higher clerical workers, medium class included lower clerical workers and employees, and the lowest class included unemployed, retired, and students. Class “unknown” includes persons with unknown socioeconomic class and those with missing data at Statistics Finland (about 5% of the cohort).

Statistical analyses

Categorical variables were reported as percentages, 95% confidence intervals (CIs) and compared with chi squared tests. Continuous variables were presented as means with standard deviations (SDs). The analyses were conducted with conditional logistic regression models due to matched design. The reference exposure category was nonuse of any ACs. The models were adjusted for the following covariates: gender, age group, history of hip fracture, stroke and ischemic heart event, schizophrenia, depression or bipolar disorder, cardiovascular disease, asthma or COPD, epilepsy, cancer, substance abuse, as well as use of asthma medication, oral corticosteroids, PPIs, cholinesterase inhibitors and memantine. Subgroup analyses were conducted among persons using AChEIs if they have higher risk of pneumonia during AC use. For this subgroup analyses, those who were AChEI nonusers at the index date were excluded and the matching retained so that each case has at least one control person (5,799 cases and 8,886 matched controls included).

RESULTS

Mean age between cases and controls was similar (83.6, standard deviation (SD) 6.7 for cases and 83.3, SD 6.5 for controls) years. Those persons who had pneumonia more often had epilepsy, diabetes, and cardiovascular disease, and had had a hip fracture, stroke, or ischemic heart event more often than control persons. In addition, they more often used PPIs, oral corticosteroids and memantine, but a smaller proportion of them used AChEIs. Substance abuse as well as use of antibiotics and asthma medication were more common among persons with pneumonia (Table 1).

Persons using ACs were more often female, aged at least 80 years, and had poorer health status than persons not using ACs (Table 2). In addition, substance abuse was more common in persons using ACs. Use of AChEIs was less common, but memantine was used more often in persons using ACs. ACs were used by 61.2% of the population. The majority (93.6%) of persons using ACs used only level 1ACs whereas use of level 2 (3.1%) and level 3 (3.3%) were uncommon. Most frequently used level 3 ACs were tolterodine (1.0%), darifenacin (0.5%), and oxybutynin (0.7%). Carbamazepine (1.6%), oxcarbazepine (0.8%), and ranitidine (0.6%) were most often used level 2 ACs. Most often used level 1 ACs were cardiovascular drugs: furosemide(+potassium-sparing diuretic) (47.6%), isosorbide mono/dinitrate (33.5%), and warfarin (29.1%). For the full list, see Supplementary Table 1.

The majority (63.5%) of persons using carbamazepine/oxcarbazepine were diagnosed with epilepsy.

When compared to persons without any ACs, those on ACs had an increased risk of pneumonia (unadjusted OR 1.72, 95% CI 1.64–1.80, adjusted OR 1.36 (1.29–1.43). The risk among those persons who only used level 1 ACs was similar. The highest risk of pneumonia was on persons using level 2 ACs (unadjusted OR 2.18, 95% CI 1.86–2.55, adjusted OR 1.40, 95% CI 1.17–1.68), whereas level 3 AC use was not associated with pneumonia (unadjusted OR 1.33, 95% CI 1.13–1.56, adjusted OR 1.03, 95% CI 0.87–1.23) (Table 3).

Incident users of ACs (those who used ACs 0–30 days but not 31–365 days before the index date) had the highest risk of pneumonia (unadjusted OR 3.00, 95% CI 2.43–3.71, adjusted OR 2.68 (95% CI 2.15–3.34), compared with prevalent users (AC use in both 0–30 and 31–365 days before the index date) (unadjusted OR 1.90, 95% CI 1.80–2.00, adjusted OR 1.48 (95% CI 1.40–1.57) and those persons with past use (use 31–365 days but no use 1–30 days before the index date) (unadjusted OR 1.64, 95% CI 1.50–1.78, adjusted OR 1.51, 95% CI 1.38–1.65). Persons using ACs within previous year before the index date in either a chronic (≥274 days) or non-chronic (<274 days) manner had increased risk of pneumonia compared with nonusers (unadjusted OR 1.73, 95% CI 1.64–1.82 and 1.83, 95% CI 1.71–1.96, adjusted OR 1.35, 95% CI 1.28–1.43 and 1.63, 95% CI 1.51–1.75, respectively).

When the effect of cumulative AC burden on pneumonia risk was measured, all quartiles were associated with and increased risk of pneumonia (Table 3). The lowest quartile was associated with the highest risk and a trend of higher risk for higher cumulative AC burden was observed within the following quartiles.

In addition, we performed a subgroup analysis on persons using AChEIs. Among those persons, use of ACs increased the risk of pneumonia (unadjusted OR 1.93, 95% CI 1.79–2.07, adjusted OR 1.53, 95% CI 1.41–1.66) (Table 4).

DISCUSSION

The main finding of our study was that use of ACs was associated with increased risk of pneumonia in persons with AD. To our knowledge, this is the first study conducted on persons with AD. There are two previous publications on the association between AC use and pneumonia in the general aged population whereas our study concerned the AD population which is in the increased risk of pneumonia compared with cognitively intact older persons. In addition, our study population was more than ten times larger than in previous studies, thus increasing the statistical power and reliability of the results.

In general, persons using ACs were on poorer health condition, and they had more conditions recognized as risk factors for pneumonia. However, after adjusting with these factors the risk associated with ACs persisted. Incident use was associated with higher risk than prevalent use.

In our study, overall use of any AC was associated with pneumonia with the same odds ratio as in the study by Chatterjee et al. [16] (1.65, 95% CI 1.20–2.28). In our study, level 2 ACs had similar risk of pneumonia than level 1 ACs, which is in contrast to Chatterjee et al. [16], who reported no association with pneumonia and level 2 or 3 AC drugs. However, they stated that this may be due to lack of statistical power in their study. In our study, majority of level 2 ACs use consisted of antiepileptic drugs carbamazepine and oxcarbazepine. This drug use pattern differs from Chatterjee et al., as in their study the most used level 2 drugs were cyclobenzaprine (not marketed in Finland) and paroxetine (rarely used in our study population). Thus, the difference may be explained with differences in study populations, prescribing practices as well as in drugs available in different countries. Epilepsy itself is a risk factor for pneumonia [3], however our results were adjusted for epilepsy. In our study, one third of persons using carbamazepine or oxcarbamazepine had no diagnosis of epilepsy. Carbamazepine and oxcarbazepine have also been used to treat behavioral symptoms in AD [26], however this is uncommon compared to use of antipsychotics for that purpose. Persons with AD have also lowered threshold for convulsions [27] and it is possible, that these drugs have been used for that indication. Paul et al. [15] found no difference between low- and high-potency ACs.

Pneumonia risk was increased regardless of duration of AC medication among persons using level 1ACs, although the risk was highest in the lowest quartile (the shortest duration of use). For level 2 AC (mainly carbamazepine and oxcarbazepine) use, only the shortest duration of use was associated with pneumonia, which may be due to indications of these drugs (epilepsy and behavioral symptoms that may also increase pneumonia risk). The majority of level 1 AC drugs used in this group were cardiovascular drugs and this may be partially explained by comorbidity.

We found some dose-response effect between pneumonia and use of ACs but the effect was not constant as the lowest cumulative AC burden was associated with similar risk as the highest burden. The reason for this is unknown. However, with an exception of some strong ACs, definition of an AC has been challenging and there is no consensus for the anticholinergicity of several drugs. There are several different lists that rank drugs based on their pharmacological properties, and they differ more or less from each other [28]. ADS is one of the widest lists published, and especially the number of level 1 ACs is high, including a wide variety of drugs used in different purposes [25]. ADS was chosen for our study to allow comparison with previously reportedresults [16].

On the other hand, five known types of muscarinic receptors (M₁-M₅) are widely distributed throughout the body with different effects (e.g., M₁, but also M₂ and M₄ have been linked to cognitive processes in the central nervous system [29]). Therefore, the specificity of a drug to a specific muscarinic receptor determines its effects in the body. In addition, characteristics of a drug molecule (such as lipid solubility) and condition of the patient (e.g., diseases increasing the blood-brain barrier permeability) also alter a person’s susceptibility to suffer from AC side-effects.

Level 3 ACs were not associated with risk of pneumonia in our study, which is in agreement with Chatterjee et al. [16]. They speculated that other, yet unknown mechanisms may contribute to the observed relationship in addition to anticholinergic mechanisms [16]. In our study, level 3 ACs were mainly urinary ACs whose primary therapeutic effect is mediated by M₃ receptors. They potentially cause cognitive impairment although the risk differs between different molecules [29]. In addition, AC penetration to the central nervous system is increased, while cholinergic activity is decreased in AD [10, 29], and this would further increase the potential of central adverse effects. Furthermore, the M₃ subtype of muscarinic receptors is responsible of mucus secretion in the lung [30]. Therefore, an increased pneumonia risk in persons using level 3 ACs would have been expected in our AD population if the increased pneumonia risk would be solely due to AC effects. More pharmacological research is required to further investigate this matter.

It has been hypothesized that central adverse effects of ACs would increase risk of pneumonia through aspiration caused by sedation and altered mental status [15]. We found the highest risk of pneumonia in incident users of ACs (OR 2.68 (95% CI 2.15–3.34), which is in same scale with [15]. This may support the hypothesis as sedation often is more pronounced in the initiation phase of drugs.

Since the majority of persons with AD use antidementia drugs, we wanted to see if AC use was associated with higher risk of pneumonia in subgroup of persons using AChEIs. Concomitant use of ACs and AChEIs was associated with an increased risk compared to those using AChEIs but not ACs. This may be due to opposite mechanisms of action of these two drug groups leading to decreased therapeutic effect but increased risk of adverse effects. Finnish guidelines state that antidementia medication should be initiated at an early stage of AD, and its use should be ceased when the disease has progressed to a level where antidementia drugs have no effect. In addition, nonusers include persons who did not tolerate these drugs due to adverse effects and thus, it is possible that nonusers of antidementia drugs are a highly selected group of persons with AD. As it is possible that persons not using antidementia drugs may have more advanced stage of AD, we selected the controls for pneumonia cases based on time since AD diagnoses as a crude measure of duration of the disease. Difference in time since AD diagnoses was minimized in the sampling of controls. However, as we lacked measures of disease severity for AD, it is possible that nonusers of antidementia drugs had worse health conditions making them more susceptible to pneumonia than those who still were on antidementia drugs. Therefore these results should be interpreted with great caution.

The major strength of this study is that our nationwide cohort included all persons newly diagnosed with AD regardless of their socioeconomic status. Nationwide registers enabled the follow-up of their health status and drug use. Drug use in our study represents actual drug purchases instead of prescribed drugs, which avoids the primary nonadherence problem, i.e., not all prescribed drugs are purchased from the pharmacy. In addition, we modeled drug use with a novel method by taking into account on personal drug use behavior and hospitalizations.

Pneumonia cases in our study represent the most severe infections, whereas pneumonia cases treated in home care were excluded. However, in this patient group the majority of pneumonias are severe and therefore, this should not create bias in our results. Furthermore, the analyses were adjusted for use of antibiotics, which should capture infections treated at home care. An important limitation of our study, based on limitations of the registers, was the lack of data on severity of AD or behavioral and psychological symptoms of dementia and lifestyle factors such as smoking.

In conclusion, the risk of pneumonia is increased in AD patients using ACs, and the risk was highest shortly after initiation of ACs. The lowest and the highest cumulative AC burden categories were associated with a similar risk and higher than the middle categories. However, it is possible that the risk increase does not result from direct antagonism to muscarinic receptors, and further studies are required to clarify the mechanism leading to increased pneumonia risk. However, the risk of pneumonia should be carefully considered when treating this vulnerable group of older persons.

DISCLOSURE STATEMENT

Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/16-0956r1).

Footnotes

Appendix

References

Simonetti

, Viasus

, Garcia-Vidal

, Carratalá

(2014) Management of community-acquired pneumonia in older adults. Ther Adv Infect Dis 2, 3–16.

Ochoa-Gondar

, Vila-Córcoles

, de Diego

, Arija

, Maxenchs

, Grive

, Martin

, Pinyol

; EVAN-65 Study Group (2008) The burden of community-acquired pneumonia in the elderly: The Spanish EVAN-65 study. BMC Public Health 8, 222.

Torres

, Peetermans

, Viegi

, Blasi

(2013) Risk factors for community-acquired pneumonia in adults in Europe: A literature review. Thorax 68, 1057–1065.

Marik

, Daplan

(2003) Aspiration pneumonia and dysphagia in the elderly. Chest 124, 328–336.

Pasqualetti

, Tognini

, Calsolaro

, Polini

, Monzani

(2015) Potential drug-drug interactions in Alzheimer patients with behavioral symptoms. Clin Interv Aging 10, 1457–1466.

Rudolph

, Zanin

, Jones

, Marcantonio

, Fong

, Yang

, Yap

, Inouve

(2010) Hospitalization in community-dwelling persons with Alzheimer’s disease: Frequency and causes. J Am Geriatr Soc 58, 1542–1548.

Foley

, Affoo

, Martin

(2015) A systematic review and meta-analysis examining pneumonia-associated mortality in dementia. Dement Geriatr Cogn Disord 39, 52–67.

Buckley

, Salpeter

(2015) A risk-benefit assessment of dementia medications: Systematic review of the evidence. Drugs Aging 32, 453–467.

Wawruch

, Macuqova

, Kostkova

, Luha

, Dukat

, Murin

, Drobna

, Wilton

, Kuzelova

(2012) The use of medications with anticholinergic properties and risk factors for their use in hospitalised elderly patients. Pharmacoepidemiol Drug Saf 32, 170–176.

10.

Konishi

, Hori

, Uchida

, Watanabe

, Tominaga

, Kimura

, Hosoyamada

, Shibasaki

, Kataoka

, Hachisu

(2010) Adverse effects of anticholinergic activity on cognitive functions in Alzheimer’s disease. Psychogeriatrics 10, 34–38.

11.

Sura

, Carnahan

, Chen

, Aparasu

(2015) Anticholinergic drugs and health-rated quality of life in older adults with dementia. J Am Pharm Assoc 55, 282–287.

12.

Johnell

, Fastbom

(2008) Concurrent use of anticholinergic drugs and cholinesterase inhibitors: Register-based study of over 700,000 elderly patients. Drugs Aging 25, 871–877.

13.

Sink

, Thomas

3rd , Xu

, Craig

, Kritchevsky

, Sands

(2008) Dual use of bladder anticholinergics and cholinesterase inhibitors: Long-term functional and cognitive outcomes. J Am Geriatr Soc 56, 847–853.

14.

Palmer

, Albrecht

, Park

, Dutcher

, Rattinger

, Simoni-Wastila

, Walker

, Zuckerman

(2015) Use of drugs with anticholinergic properties among nursing home residents with dementia: A national analysis of Medicare beneficiaries from 2007 to 2008. Drugs Aging 32, 79–86.

15.

Paul

, Walker

, Dublin

(2015) Anticholinergic medications and risk of community-acquired pneumonia in elderly adults: A population-based case-control study. J Am Geriatr Soc 63, 476–485.

16.

Chatterjee

, Carnahan

, Chen

, Holmes

, Johnson

, Aparasu

(2016) Anticholinergic medication use and risk of pneumonia in elderly adults: A nested case-control study. J Am Geriatr Soc 64, 394–400.

17.

Barry

, Cooper

, Ryan

, Passmore

, Robinson

, Molloy

, Darcy

, Buchanan

, Hughes

(2016) Potentially inappropriate prescribing among people with dementia in primary care: A retrospective cross-sectional study using the enhanced prescribing database. J Alzheimers Dis 52, 1503–1513.

18.

Taipale

, Koponen

, Tanskanen

, Tolppanen

, Tiihonen

, Hartikainen

(2014) High prevalence of psychotropic drug use among persons with and without Alzheimer’s disease in Finnish nationwide cohort. Eur Neuropsychopharmacol 24, 1729–1737.

19.

Tolppanen

, Taipale

, Koponen

, Lavikainen

, Tanskanen

, Tiihonen

, Hartikainen

(2016) Cohort profile: The Finnish Medication and Alzheimer’s disease (MEDALZ) study. BMJ Open 6, e012100.

20.

Finnish Medical Society Duodecim, Current care guideline: Memory disorders [in Finnish with English summary]. http://www.kaypahoito.fi. Last updated August 13, 2010, Accessed on March 10, 2016.

21.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA work group under the auspices of department of health and human services task force on Alzheimer’s disease. Neurology 34, 939–944.

22.

WHO Collaborating Centre for Drug Statistics Methodology. The Anatomical Therapeutic Chemical Classification System. Structure and principles. http://www.whocc.no/atc/structure_and_principles/. Accessed on March 10, 2016.

23.

Tanskanen

, Taipale

, Koponen

, Tolppanen

, Hartikainen

, Ahonen

, Tiihonen

(2015) From prescription drug purchases to drug use periods - a second generation method (PRE2DUP). BMC Med Inform Decis Mak 15, 21-015-0140-z.

24.

Finnish Medical Society Duodecim, Current care guideline: Lower respiratory infections [in Finnish with English summary]. http://www.kaypahoito.fi. Last updated: January 2, 2015, Accessed on April 11, 2016.

25.

Carnahan

, Lund

, Perry

, Pollock

, Culp

(2006) The anticholinergic drug scale as a measure of drug-related anticholinergic burden: Associations with serum anticholinergic activity. J Clin Pharmacol 46, 1481–1486.

26.

Ballard

, Corbett

, Chitramohan

, Aarsland

(2009) Management of agitation and aggression associated with Alzheimer’s disease: Controversies and possible solutions. Curr Opin Psychiatry 22, 532–540.

27.

Horváth

, Szucs

, Barcs

, Noebels

, Kamondi

(2016) Epileptic seizures in Alzheimer disease: A review. Alzheimer Dis Assoc Disord 30, 186–192.

28.

Salahudeen

, Duffull

, Nishtala

(2015) Anticholinergic burden quantified by anticholinergic risk scales and adverse outcomes in older people: A systematic review. BMC Geriatr 15, 31.

29.

Kay

, Ebinger

(2008) Preserving cognitive function for patients with overactive bladder: Evidence for a differential effect with darifenacin. Int J Clin Pract 62, 1792–1800.

30.

Meurs

, Dekkers

BGJ

, Maarsingh

, Halayko

, Zaagsma

, Gosens

(2013) Muscarinic receptors on airway mesenchymal cells: Novel findings for an ancient target. Pulm Pharmacol Ther 26, 145–155.