Making a Bad Diagnosis Worse? Suspect Drug Management of Urinary Incontinence in Persons with Dementia

The Beers Criteria expert panel recommends avoidance of bladder medications with antimuscarinic effects, commonly referred to as anticholinergic medications, in people with dementia for several reasons.¹ First, people with dementia are highly sensitive to the adverse cognitive and psychiatric effects of anticholinergics. Individuals with Alzheimer’s disease given scopolamine experienced psychiatric side effects and greater dose-dependent worsening of cognition than age-matched controls.² Substantial evidence illustrates that anticholinergic medications can lead to adverse cognitive outcomes ranging from memory deficits to delirium.³ Second, bladder anticholinergic medications may be given to individuals with dementia because of side effects from cholinesterase inhibitors offered for preserving cognition.^4,5 This use of an additional drug to treat the side effects of the first (as opposed to stopping the drug causing the side effect) often leads to polypharmacy—called the prescribing cascade. Simultaneous use of procholinergic and anticholinergic drugs not only has theoretical disadvantages, but also has been associated with functional decline.⁶ Third, emerging evidence suggests that anticholinergic medication use is associated with incident dementia and brain atrophy, suggesting that cognitive effects are not necessarily reversible with discontinuation.^7,8 Establishing causality is difficult because urinary incontinence (UI) is a common indication for anticholinergics and incontinence can be an early sign of ischemic white matter disease, which also poses risk of cognitive decline,⁹ but this evidence remains troubling.

In this issue of the Journal of the American Geriatrics Society, Green and colleagues present data on the use of bladder antimuscarinics in older adults with impaired cognition.¹⁰ They studied individuals who underwent examinations at Alzheimer’s Disease Centers from 2005 to 2015. Six percent of participants with dementia, 5.6% of those with mild cognitive impairment (MCI), and 4.0% of those with normal cognition used anticholinergic bladder medications. Although UI is common in people with dementia—particularly as the disease progresses^11,12—using anticholinergic medications for UI or lower urinary tract symptoms (LUTS) in individuals with cognitive impairment is troubling. Even worse for the individuals studied, of those with dementia taking bladder antimuscarinics, 16% were simultaneously taking another drug classified as having clinically significant anticholinergic properties; 27% were taking a cholinesterase inhibitor at the time of the initiation of a bladder anticholinergic medication, and 41% of those taking bladder anticholinergics with dementia were simultaneously taking a procholinergic.¹⁰ Green and colleagues’ article is important, yet work remains to be done in terms of gaps in this research and its clinical implications.

This study has several unique strengths. First, past investigations of cognition and bladder dysfunction have been limited because standardized diagnostic criteria for dementia were not used.¹¹ Criterion standard diagnostic procedures were used in this study. Second, research-trained clinicians collected standardized medication inventories, which included over-the-counter medications. The authors discuss prescribing that might be suspect in terms of frequency and specific drug choice; oxybutynin and tolterodine were the most commonly used bladder antimuscarinics.¹⁰ Oxybutynin immediate release (IR) is frequently acknowledged as a poor drug choice in older adults because it is relatively lipophilic, with limited selectivity for bladder muscarinic receptors. It appears to cause more adverse effects than other bladder antimuscarinics, although extended-release and transdermal preparations may be safer.¹³ Oxybutynin and tolterodine are relatively nonselective antimuscarinic agents, although oxybutynin is more lipophilic. Both can cause at least some level of cognitive impairment.^14,15 If an individual has MCI or Alzheimer’s disease, and it is thought that bladder antimuscarinics are necessary, it would seem a better choice to use one with a better theoretical safety profile. Although drug selection in this study may have been a function of the study period (2005–15), it also could be from policies driven by cost considerations. Providers prescribing within the context of formularies or working with insurers are often given fail-first instructions: “Make sure that your patient does not benefit from the cheaper drug first.” It’s not just efficacy that matters, but also safety, and information on this in patients with dementia is limited. Inferences about comparative effectiveness and safety can be made based on preclinical studies and studies in relatively healthy older adults, but until strong evidence is available in this population, these are only theories, and people will continue to receive medications that may worsen cognitive outcomes. More study is needed, particularly to improve understanding of the severity and type of UI or LUTS, specific details about activity of daily living impairments, and effectiveness or harms of UI drugs in this population.

The article by Green and colleagues lacks meaningful data on participants’ LUTS or likely UI type. The severity and effect of LUTS are unknown at the individual level, and the prevalence of LUTS is unknown at the group level. A prescriber should always consider the UI type (stress UI, urgency UI, UI with detrusor hypocontractility), physical signs, and the context of the LUTS in the overall life of the individual and caregiver. An anticholinergic bladder medication would not be useful for diagnoses such as postprostatectomy UI, with which there is continuous leakage, or pure stress UI, with which there is a finding of a substantially high postvoid residual because of bladder outlet obstruction, or in situations in which an individual makes no attempt to toilet.

Although there is some aggregate information about functional impairment, more information would be welcome at the level of the individual. Comorbidities were not very different across diagnostic categories, but those with normal cognition had many fewer functional impairments (0.5, range 0–30; higher is worse) than those with MCI (3.9) or dementia (19.2). Not all individuals with dementia have UI (preserved transfer and dressing ability might facilitate continence), but cognitive impairment with loss of dressing or transfer ability usually results in complete UI.¹⁶

In theory, it is known that antimuscarinic drugs are a poor choice for individuals with dementia, but little is learned from this study about the outcomes (safety and efficacy) of this prescribing, for which data are limited. A placebo-controlled trial of extended-release oxybutynin in 50 female nursing home residents with dementia focused on safety and did not evaluate efficacy.¹⁷ Surprisingly, no one in either group experienced delirium, and cognitive and behavioral outcomes were similar, which could relate to lower peak blood levels with the extended-release formulation, but another study of the safety and effectiveness of bladder antimuscarinics in long-term care residents of Veterans Affairs community living centers found that the number needed to treat to observe a difference in incontinence measures on the Minimum Data Set (n=32) was similar to the number needed to harm for hip fracture (n=36). Immediate-release oxybutynin was used by nearly all patients.¹⁸ The unfavorable risk–benefit profile in this study raises the questions of how often bladder antimuscarinics are effective in individuals with dementia and to what extent safety and effectiveness are monitored. One author (RC) recalled asking a caregiver whether the bladder antimuscarinic her husband took reduced his UI: the answer, “He never makes it to the bathroom.” There are potentially better, safer LUTS treatments in this population. Scheduled toileting or prompted voiding with assistance for those who retain the physical ability to toilet and have an able caregiver may be the best first-line treatment for UI in many people with dementia. There are more data for the effectiveness of this intervention in the LTC setting than in the home, including evidence that oxybutynin does not improve outcomes beyond prompted voiding in nursing home residents.^12,19 In men who experience urinary frequency and urgency, alpha-adrenergic antagonists (α-blockers), phosphodiesterase type-5 inhibitors, or 5-α reductase inhibitors (when prostate enlargement is present) might be useful medications and would not have anticholinergic effects. For urge incontinence, mirabegron, a beta-3 agonist, does not bind muscarinic receptors. It may have a good cognitive safety profile, but this needs investigation. It can cause high heart rate and blood pressure, so it is not completely benign.²⁰ Of the anticholinergic medications for urge incontinence, some have properties suggesting better cognitive safety profiles. Trospium is a quaternary amine that appears not to penetrate the blood–brain barrier, at least in healthy older adults, and so seems a good choice,²¹ although the blood–brain barrier may be dysfunctional in individuals with dementia and other illnesses, making cognitive safety uncertain.²² Darifenacin is more selective for M3 receptors found in the bladder than for the M1 receptors, which are more important for cognition and so may be safer. It is more lipophilic than tolterodine, but it and trospium had poor central nervous system penetration in rats. An active metabolite of fesoterodine and tolterodine, 5-hydroxymethyl tolterodine, also had limited brain penetration. Solifenacin is not especially selective for M3 muscarinic receptors and penetrated the brain much more than trospium or darifenacin in rats.²³ Studies of darifenacin and solifenacin showed no adverse cognitive effects, but these were conducted in healthy older adults, not in people with dementia.¹³ The effect of these drugs in people with dementia is unclear.

This article by Green and colleagues contributes meaningfully to the literature on prescribing patterns for UI in individuals known to have dementia. Prescribers should be aware of the potential negative effect of anticholinergic bladder relaxants in this population.