March 2011
Reviewer: Jeff C. Huffman, MD
Association of anticholinergic drugs with hospitalization and mortality among older cardiovascular patients: A prospective study
Uusvaara J, Pitkala KH, Kautiainen H, Tilvis RS, Strandberg TE
Drugs Aging 2011; 28(2):131-138
Background: Anticholinergic medications are linked with increased rates of delirium and falls, and they can have other adverse effects, such as tachycardia. In general, it is recommended that medications with anticholinergic properties be avoided in older adults whenever possible, especially in patients with dementia or other cognitive disorders. However, there has been relatively limited prospective study of the impact of anticholinergic medications on broader medical outcomes such as admissions and mortality.
Methods: This was a secondary analysis from a randomized trial of 400 Finnish older adults (aged 75-90) with CVD (prior MI, CAD, previous CVA/TIA, or peripheral arterial disease). The primary trial studied the impact of an intervention to tailor subjects’ medication regimen to fit evidence-based treatment guidelines. This intervention was compared to usual care and the results of this trial were negative. The goal of the secondary analysis was to prospectively compare the number of days hospitalized and mortality rates among these participants who were and were not taking anticholinergic medications at study outset. Anticholinergic medications were defined using a list of 32 medications determined to have these properties. Data on readmissions were obtained via national hospital discharge register and mortality data via national registry. Regression analyses were used to assess an independent association between anticholinergic use and adverse outcomes.
Results: Overall, 295 subjects (74%) were prescribed one of the 32 anticholinergic agents at study outset. Over an average follow-up period of 3.3 years, subjects getting anticholinergic agents were more likely to suffer mortality (20.7% vs. 9.5%), and had more hospital days per year (14.9 vs. 5.2). When multivariate analysis was performed that accounted for age, gender, and medical comorbidity (using a comorbidity index), anticholinergic medication use at baseline was not associated with mortality, but continued to be significantly associated with the number of days spent in the hospital.
Discussion: This is an important line of work—many elderly patients get anticholinergic medications, and we are all well aware of their adverse effects, especially on cognition. It is important to have prospective studies that assess the impact of these medications on objective medical outcomes to better understand the risks of these medications in older adults. Therefore, this project was a good first step in exploring this area of research by prospectively measuring such outcomes over years.
Unfortunately, this specific study had substantial limitations. The list of anticholinergic medications used in the study is problematic in several ways. First, and perhaps most importantly, it does not include several medications (e.g., diphenhydramine) that have substantial anticholinergic effects, while including medications with rather modest (e.g., warfarin) effects.
Furthermore, though classifications (e.g., 1-3 rating scale) exist that detail the degree to which an agent acts at muscarinic cholinergic receptors, there was no severity gradient used in this study, so a patient on oxybutinin, doxepin, or hydroxyzine would be rated the same as a patient on furosemide. There was also no account for being on single vs. multiple anticholinergic medications. Therefore, overall anticholinergic burden was not really accounted for in any way aside from an on/off system.
In addition, other agents that may impact rates of cognition and falls by other mechanisms—e.g., benzodiazepines—are on the list, and it may be that patients on benzos may have done worse (and that this was unrelated to anticholinergic effects). Finally, measuring only anticholinergic medication use at study outset likely does not accurately reflect such medication use over a >3 year period.
In short, this is an important topic, and at the same time the lack of apparent completeness of the anticholinergic medication list and the lack of any in-depth assessment of anticholinergic use in terms of degree of anticholinergic effects, overall anticholinergic burden via multiple medications, or repeated assessments of use greatly limits this specific investigation.