eTable. Anticholinergic Drugs and their Anticholinergic Drug Scale (ADS) Score Considered in This Study
Customize your JAMA Network experience by selecting one or more topics from the list below.
Aldebert G, Faillie J, Hillaire-Buys D, et al. Association of Anticholinergic Drug Use With Risk for Late Age-Related Macular Degeneration. JAMA Ophthalmol. 2018;136(7):770–778. doi:10.1001/jamaophthalmol.2018.1719
Is anticholinergic drug use associated with late age-related macular degeneration?
In this case-control study, at least 3 months’ use of anticholinergic drugs was associated with increased risk of AMD. This association was greater with prolonged use and high Anticholinergic Burden Score.
While confounders that were not addressed could have contributed or accounted entirely for the results, these findings suggest the risk of late AMD may be increased with at least 3 months’ use of anticholinergic drugs, with a dose-effect association.
Amyloid-β is a major component of retinal drusen, the primary lesions of age-related macular degeneration (AMD), and autopsy and animal models suggested that anticholinergic drug (ACD) use increased brain amyloid-β deposition.
To investigate the association between exposure to ACDs and late AMD (features of neovascular AMD or geographic atrophy of the retinal pigment epithelium in at least 1 eye).
Design, Setting and Participants
A multicenter case-control study in 4 French ophthalmologic centers comprising 200 cases with late AMD and 200 controls enrolled from July 2016 to June 2017.
Exposure to at least 3 months of ACDs started before AMD diagnosis was recorded during a specific interview. A dose-effect association with cumulative exposure duration and Anticholinergic Burden Score was explored. The association between ACD exposure and AMD was assessed by multivariate logistic regression analysis adjusted for age, sex, smoking status, family history of AMD, alcohol consumption, and use of anticoagulant and anti-inflammatory drugs. Odds ratios (ORs) and 95% confidence intervals were estimated.
Main Outcomes and Measures
Association between exposure to ACDs and late AMD.
Among case participants, the mean (SD) age was 74.8 (9.2) years, 129 (64.5%) were women, 192 (96%) were white, 65 (32.5%) had geographic atrophy, 135 (67.5%) had neovascular AMD, 116 (58%) had unilateral AMD, and 84 (42%) had bilateral AMD. Among control participants, the mean (SD) age was 75.5 (7.2) years, with 116 (58%) women and 187 (93.5%) white participants. Twenty-six cases (13%) and 10 controls (5%) were exposed to ACDs throughout life for at least 3 months before AMD onset. Risk of AMD was increased with ever exposure to ACDs (adjusted OR [aOR], 2.84; 95% CI, 1.33-6.06; P = .007), high Anticholinergic Burden Score (≥3) (aOR, 6.42; 95% CI, 1.38-29.92; P = .02), and longest cumulative exposure to ACD (≥15 years) (aOR, 5.88; 95% CI, 1.22-28.31; P = .03).
Conclusions and Relevance
Risk of late AMD may be increased with at least 3 months’ use of ACDs. A dose-effect association was suggested by a greater association with prolonged use and high Anticholinergic Burden Score. Further studies, in particular those with longitudinal design, are needed to confirm this association.
Age-related macular degeneration (AMD) is the leading cause of vision loss in people 50 years and older in the developed world.1-3 It is characterized by macular extracellular deposits (drusen) and degeneration of retinal pigmented epithelium and photoreceptors (with or without choroidal neovascularization). Pathogenic pathways mediating this development are discussed. Inflammation, hypoxia, and oxidative stress seem to have a central role.3,4
Several risk factors of AMD have been identified, aging being the strongest. Other major risk factors are smoking (with a reversibility of the risk after cessation) and family history of AMD (some genetic markers have been identified). Further risk factors show a moderate strength of association: previous cataract surgery, high body mass index (BMI), other cardiovascular risk factors, history of cardiovascular disease, nutritional factors (protective action of antioxidants, lutein, and ω3), sunlight exposure, and iris color.3,5
No drug use has been clearly associated with the occurrence of AMD. The protective effect of statins6 and anti-inflammatory drugs use is uncertain.7,8 Medications suspected to increase AMD risk are aspirin,9 thyroxin,10 and β-blockers.11,12 Bias relating this drug use to patients’ medical history could not be excluded.
Older adults frequently use medications with anticholinergic activity (prevalence from 7.5% to 27%),13-16 but panels of experts have suggested that the benefits of these agents in older populations may be surpassed by the risks.17,18 Some anticholinergic drugs (ACDs) achieve the intended therapeutic effect by blocking the effect of acetylcholine at the muscarinic receptor within specific organ systems (eg, bladder antimuscarinics, bronchodilators, and anti-Parkinson agents). However, other medications have unintended anticholinergic effects that are not the primary therapeutic activity (eg, first-generation antihistamines, tricyclic antidepressants, and some antipsychotic agents).
Anticholinergic drug use has been associated with cognitive decline and dementia,13-15,19-22 and reduced cholinergic transmission is suspected to increase brain amyloid-β deposition.23-26 Additionally, cognitive impairment has been associated with AMD,27-30 and characteristic lesions of AMD (drusen) and Alzheimer disease (senile plaques) both contain amyloid-β deposits,31-33 suspected to contribute to local inflammation and oxidative stress leading to retinal degenerative events in AMD.34-39 The aim of this study was to explore the association between ACD use and late AMD occurrence.
This multicenter case-control study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist items for reporting case-control studies.40 This study was conducted in 4 French ophthalmologic centers in Montpellier and Nîmes, France. The research was approved by the ethics committee of Marseille University Hospital, France, and informed oral consent was obtained from participants, thus conforming to the provisions of the 1995 Declaration of Helsinki (revised in Edinburgh 2000).
Participants were enrolled from July 2016 to June 2017. Eligible case participants were defined as any person consulting for late AMD, already diagnosed or not, defined by feature of neovascular AMD (nAMD) with active or inactive choroidal neovacularisation, or geographic atrophy (GA) or the retinal pigment epithelium, in at least 1 eye confirmed by an ophthalmologist. Control participants were selected among people older than 60 years who were consulting in the same centers during the same period for refraction assessment but without ophthalmologic disease as confirmed by a detailed ophthalmic examination. Patients with early AMD (large drusen or pigment abnormality) were not included in this study.
Age at late AMD diagnosis and at the beginning of neovascular complications for each eye were from medical files and patient questioning. The index date was defined as the late AMD diagnosis date in cases and as the inclusion date in controls.
All drug exposures throughout life, for at least 3 months, and starting before the index date were recorded, with specific recording of ACDs; benzodiazepines; antihypertensive drugs; statins; fibrates; anti-aggregate, anticoagulate, and anti-inflammatory drugs; levothyroxine; and amiodarone. We specifically questioned patients and relatives accompanying them during an interview, relying on a questionnaire listing all drugs (classified by indication) and on medical prescriptions from a general practitioner.
We established the ACD list according to Moulis et al41 using the Anticholinergic Drug Scale,42 each drug having a score ranging from 0 to 3 depending on the intensity of its antimuscarinic effects. Because we wanted to consider only ACDs with a clinical antimuscarinic effect,43 we excluded drugs with score 1 on the Anticholinergic Drug Scale and not listed by Durán et al44 because of potential antimuscarinic effects in vitro but without clinical antimuscarinic effects consensually accepted. Finally, because some ACDs marketed in France are not listed in the Anticholinergic Drug Scale, we added the drugs with a clinical anticholinergic effect listed by Laroche et al.45 The included drugs with their corresponding Anticholinergic Drug Scale score are listed in the eTable in the Supplement. Ever exposure to ACDs was defined as use of at least 1 of these drugs throughout the lifetime, for at least 3 months, and starting before the index date. Previous exposure was defined as use discontinued before AMD diagnosis, and current exposure was defined as use ongoing at AMD diagnosis.
To assess a dose-effect association, we calculated the Anticholinergic Burden Score for each patient as described by Moulis et al41 by summing the Anticholinergic Drug Scale score for all ACDs ever used.42 For drugs marketed in France not listed in the Anticholinergic Drug Scale, we attributed the same anticholinergic weight as that assigned by a college of pharmacologists in previous French studies.41,43 The Anticholinergic Burden Score was classified as 0, 1 to 2, and at least 3, with cut points based on previous studies41 and on the distribution observed in our sample. This indicates low, medium, and high burden.
Age at onset and exposure duration to each ACD before AMD diagnosis were recorded. We calculated the cumulative exposure duration to ACDs by summing the exposure duration of all ACDs ever used before AMD diagnosis. Cumulative exposure duration was classified as no use and use for less than 5, 5 to 15, and more than 15 years, with cut points based on clinical interpretability and on the distribution observed in our sample.
From a literature review, we selected covariates that may confound the relation between ACD use and late AMD.3,5 Age, sex, college education level, race/ethnicity, eye color, and first-degree family history of AMD were identified. Body mass index was calculated as weight in kilograms divided by height in meters squared, with overweight defined as between 25 and 30 and obesity defined as more than 30. Smoking at the index date was classified as nonsmoker, ex-smoker, or current smoker. Alcohol intake was classified as nondrinker, rare drinker (<1 units/d), moderate drinker (1-2 units/d), and excessive drinker (≥3 units/d). Data on fruits, vegetables, fish, red meat, and poultry consumption by week were collected. We also recorded the presence of the following potential confounding factors: high blood pressure, diabetes (types 1 and 2), hyperlipidemia, coronary artery disease, cerebrovascular disease, peripheral vascular disease, cataract surgery, and known memory disorders with significant effect on everyday life. We identified medical indications associated with ACD use such as major depressive episode, Parkinson disease, and epilepsy. Only diagnoses made before the index date were included in the analyses.
Descriptive statistics were used to summarize the baseline characteristics of cases and controls. In the primary analysis, to assess the association between AMD and ever exposure to ACDs, we calculated odds ratios (ORs) with 95% CIs by using a logistic regression model adjusted for sex, age, and other risk factors for AMD that were known (eg, smoking status and AMD family history) or identified in our analyses (alcohol consumption and use of anticoagulant and anti-inflammatory drugs), estimating adjusted ORs (aORs) and 95% confidence intervals. In a secondary analysis, we used the same model to assess the association between AMD and current exposure to ACDs, previous exposure to ACDs, cumulative exposure duration to ACDs, and Anticholinergic Burden Score. Third, we performed subgroup analyses by unilateral or bilateral AMD, GA, or nAMD. Assuming a prevalence of ACD use of 7.5% in the older population15 and a clinically significant OR of 2.5, with a statistical power of 80% and a risk of error of 5%, we estimated a sample size of at least 382 people (191 cases and 191 controls). Analyses involved use of SAS, version 9.2 (SAS Institute Inc). P values were 2-sided and were considered significant at a level of .05.
A total of 200 cases and 200 controls were included. Characteristics of cases and controls are listed in Table 1. The number of women was 129 (64.5%) and 116 (58%), respectively. At the index date, the mean (SD) age was 74.8 (9.2) years and 75.5 (7.2) years, respectively, with a slightly higher prevalence of age 70 to 80 years for controls. At the index date, 65 cases (32.5%) had GA, 135 (67.5%) had nAMD, 116 (58%) had unilateral AMD, and 84 (42%) had bilateral AMD.
Compared with controls, cases were more likely than controls to be current smokers (OR, 2.09; 95% CI, 1.13-3.87; P = .02), be overweight (OR, 1.61; 95% CI, 1.04-2.49; P = .03), and have a first-degree family history of AMD (OR, 2.88; 95% CI, 1.55-5.33; P = .001); they were less likely to be moderate drinkers (OR, 0.42; 95% CI, 0.24-0.74; P = .003). Other characteristics did not differ between the groups.
Overall, 26 cases (13%) vs 10 controls (5%) were ever exposed to ACDs (Table 2). Risk of late AMD was increased with ever use of ACDs (OR, 2.84; 95% CI, 1.33-6.06; P = .007), which remained after adjustment for age, sex, smoking status, AMD family history, alcohol consumption, and use of anticoagulant and anti-inflammatory drugs (aOR, 2.39; 95% CI, 1.07-5.33; P = .03). The most common ACD classes used were antidepressants (15 cases [7.5%] and 7 controls [3.5%]). Compared with never use of ACDs, current and previous ACD use were not associated with AMD (aOR, 2.22; 95% CI, 0.86-5.77; P = .10 and aOR, 2.79; 95% CI, 0.69-11.25; P = .15). Risk of AMD was increased with high Anticholinergic Burden Score (≥3) (aOR, 6.42; 95% CI, 1.38-29.92; P = .02) but not moderate Anticholinergic Burden Score (1 or 2) (aOR, 1.38; 95% CI, 0.52-3.67; P = .52) and with longest cumulative exposure to ACDs (>15 years) (aOR, 5.88; 95% CI, 1.22-28.31; P = .03) but not intermediate (5-15 years) and shortest cumulative exposure (<5 years) (aOR, 1.48; 95% CI, 0.41-5.33; P = .55 and aOR, 1.69; 95% CI, 0.45-6.29; P = .44).
On stratifying analyses by AMD subclasses (Table 3), ever exposure to ACDs was associated with increased risk of GA (OR, 3.46; 95% CI, 1.37-8.73; P = .009), nAMD (OR, 2.56; 95% CI, 1.12-5.82; P = .03), and bilateral and unilateral AMD (OR, 2.86; 95% CI, 1.17-7.03; P = .02 and OR, 2.82; 95% CI, 1.22-6.51; P = .02).
Risk of AMD was reduced with use of anticoagulant and steroidal anti-inflammatory drugs (OR, 0.32; 95% CI, 0.11-0.89; P = .03 and OR, 0.26; 95% CI, 0.07-0.95; P = .04). The groups did not differ in other drug classes used (Table 4).
This case-control study revealed an association of late AMD with ever exposure to ACDs for at least 3 months before AMD diagnosis. The association was observed after adjustment for the main AMD risk factors known or identified in our analyses. A dose-effect relation was observed, with a greater association of AMD with long cumulative ACD exposure and high Anticholinergic Burden Score.
Among the different ACD lists and scales available, we chose the methods of Moulis et al41: we restricted the Anticholinergic Drug Scale42 to drugs with clinical antimuscarinic effects consensually accepted (using the list from Durán et al44) and added the drugs with clinical antimuscarinic effects marketed in France (using the list from Laroche et al).45 The Anticholinergic Risk Scale46 is debatable in geriatrics studies because it does not include some ACDs widely used in older patients.47 Finally, the Anticholinergic Cognitive Burden scale48 overestimates the anticholinergic clinical effect of drugs such as paroxetine and olanzapine.47 It does not include some drugs with clinical anticholinergic effects such as fluoxetine47; therefore, it was not appropriate for our purpose. We chose to collect exposure for at least 3 months because of the difficulty for patients to remember medications used for a short time and because of the inability of previous studies to show adverse effects for cognitive function of ACD intake for less than 3 months.21
Several mechanisms may be suggested to explain the association of ACD use and AMD. One is the increase in macular inflammation owing to retinal amyloid-β deposition secondary to ACD use. Anticholinergic drug use is suspected to increase brain amyloid-β deposition. An autopsy study of patients with Parkinson disease found increased levels of Alzheimer disease neuropathologic features in participants who used ACDs for 2 years or longer.23 Moreover, in animal models, reduced cholinergic transmission (via atropin or cortical cholinergic denervation) increased brain amyloid-β concentration.24-26 Several studies founded amyloid-β in drusen of eyes with AMD.31-33 Amyloid-β is suspected, as in Alzheimer disease, to promote mitochondrial dysfunction leading to oxidative stress of retinal cells and induce local inflammation, with complement pathways and microglial cell activation, contributing to retinal degenerative events.34-39
Additionally, exposure of retinal pigment epithelium to ultraviolet radiation may be increased via mydriasis induced by ACDs. The association of ultraviolet radiation exposure and late AMD has been studied,49-51 but to our knowledge, not that of pupil size and AMD.
We observed a lower rate of AMD among patients using anticoagulant drugs, which was not previously described in the literature. The finding may be explained by the reduced activity of factor Xa and thrombin, which are suspected to stimulate proinflammatory and profibrotic mediator production by retinal pigment epithelial cells.52 We also observed a lower rate of AMD among patients using anti-inflammatory drugs, which was previously described7 but with conflicting results.8
The data concerning covariates and exposure to ACDs were self-reported, which suggests potential recall bias in this study and possible underestimation of ACD use. However, these data were assessed similarly in cases and controls. Anticholinergic drug use in controls was lower than in previous studies13-16 probably because these studies considered an older population and shorter use as well as topical use (eg, bronchodilatators) and drugs with in vitro anticholinergic effects only.
Age at GA onset is sometimes difficult to determine accurately, which raises some concern in the temporal association of ACD and GA. We observed a higher association between ACD and nAMD with a lower risk of temporal bias than in GA.
Cases were representative of the AMD population, with similar diagnostic age and proportion of GA and nAMD53 and the same exposure to the main other risk factors (OR between 2 and 3 for AMD family history and current smoking). Concerning other covariates, such as medical indications associated with ACD use, including major depressive syndrome, controls were similar to cases. Moreover, ACD use was the only medication recorded in this study associated with increased risk of AMD. Therefore, a selection bias of controls is unlikely.
The case-control design of this study did not allow for assessing a causative relation. We were not able to assess the risk reduction or persistence after discontinuation of ACDs. There were no other investigations than anticholinergic association with AMD, which was the primary objective of this prospective study. The case-control design of this study did not allow for assessing a temporal association. Moreover, consistency of this association cannot be affirmed by other studies, and because AMD is permanent, we cannot look for reversibility. Therefore, beyond the strength of the association and the dose-effect relation suggested, we cannot affirm a causative relation.
This study suggests an association of late AMD with ACD use for at least 3 months. A dose-effect relation was suggested by a greater association with prolonged intake and high Anticholinergic Burden Score. Further studies are needed to confirm this association and assess potential causative pathways between ACD use and late AMD.
Corresponding Author: Vincent Daien, MD, PhD, Department of Ophthalmology, Guide Chauliac Hospital, CHU Montpellier, 80 Ave Augustin Fliche, 34295 Montpellier, France (email@example.com).
Accepted for Publication: February 25, 2018.
Published Online: May 24, 2018. doi:10.1001/jamaophthalmol.2018.1719
Author Contributions: Dr Aldebert had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Aldebert, Faillie, Hillaire-Buys, Mura, Daien.
Acquisition, analysis, or interpretation of data: Aldebert, Faillie, Hillaire-Buys, Carrière, Delcourt, Creuzot-Garcher, Daien.
Drafting of the manuscript: Aldebert, Faillie.
Critical revision of the manuscript for important intellectual content: Aldebert, Faillie, Hillaire-Buys, Mura, Carrière, Delcourt, Creuzot-Garcher, Daien.
Statistical analysis: Faillie, Mura, Carriere.
Study Supervision: Daien.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Daien reports personal fees from Bayer, Novartis, Thea, and Horus outside the submitted work. Dr Creuzot-Garcher reports grants and personal fees from Baush and Lomb, Horus, and Thea and personal fees from Alcon, Allergan, Bayer, and Novartis outside the submitted work. Dr Delcourt reports personal fees from Allergan Bauschand Lomb, Novartis, and Roche and grants and personal fees from Laboratoires Théa outside the submitted work. Dr Villain reports personal fees from Allergan. No other disclosures were reported.
Funding/Support: This study was funded by the University Hospital of Montpellier, France.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Sandrine Allieu, MD, Sophie Navarre, MD, and Charlotte Rohart, MD, Department of Ophthalmology, Beausoleil Clinic, Montpellier, France, contributed in enrolling patients and data collection. Olivier Rebollo, MD, Pierre-André Duval, MD, and Timothé Busetto, MD, Saint-Jean Clinic, Montpellier, France, contributed in enrolling patients and data collection. Didier Audemard, MD, Department of Ophthalmology, Caremeau Hospital, Nîmes, France, contributed in enrolling patients and data collection. Christelle Schneider, MD, Anne-Lise Montcriol, MD, Florent Verhaeghe, MD, Julien Plat, MD, and Arnaud Payerols, MD, Department of Ophthalmology, Gui De Chauliac Hospital, Montpellier, France, contributed in enrolling patients and data collection. Thibault Archard, orthoptist student in the Department of Ophthalmology, Gui De Chauliac Hospital, Montpellier, France, contributed in enrolling patients. Caroline Diaz and Matthieu Trocellier, pharmacy students in in the Department of Ophthalmology, Gui De Chauliac Hospital, Montpellier, France, contributed in data collection. No compensation was received for these contributions.
Create a personal account or sign in to: