Recent large epidemiological observational studies have identified an association between the use of anticholinergic drugs and cognitive impairment [1].  Although limited to patients with high levels of cumulative exposure, anticholinergic drugs to treat overactive bladder (OAB) have been identified as potential culprit agents [2].  Despite conflicting data, no causal link and the reported low prevalence of cognitive adverse events, patients are concerned about their potential occurrence following OAB treatment [3].  Such heightened concern may lead to a lack of effective treatment for OAB in older people should there be no alternative. 
Although short term specific cognitive testing and clinical trial data suggest no impairment of cognition in cognitively intact older people and the incidence of “cognitive “ adverse events in clinical trials is low, there are few data concerning exposure to OAB antimuscarinics in those who may be at greater risk of cognitive impairment.  This study therefore aimed to examine the changes in Mini-Mental State Examination score (MMSE), in an attempt to define any changes in cognitive performance in patients with both normal and low baseline scores at trial entry. We hypothesized that patients with low MMSE scores at baseline would show no consistent pattern of change in response to treatment with fesoterodine.

Data from 2 randomized, placebo controlled phase IV clinical trials of flexibly dosed fesoterodine over 12 weeks duration (A0221045, A0221049) which prospectively collected data on men and women with OAB ≥65y formed the dataset for this study.  Each study also collected data on MMSE at trial baseline and end of study.  Analyses were based on the Full Analysis Set (FAS) which included all subjects who took at least one dose of assigned study drug and contributed data to at least one baseline and post-baseline efficacy assessment with baseline value of the outcome variable > 0. MMSE scores were collected at baseline and end of study.  The MMSE comprises 11 items across 6 domains and includes such items as recalling the year, correctly identifying a familiar object like a pencil, counting backward, and writing a sentence. The measure has a standardized delivery and scoring system, resulting in scores ranging from 0 to 30 with higher scores indicating less impairment. Score change in MMSE was defined as the score at Week 12 – score at baseline and scores were categorized based on raw score change as: Improved – a positive score change of 1 or more points; No change- difference of scores is 0, or Declined - a reduction in score change of 1 or more points.

There were 673 patients treated with fesoterodine and 674 patients treated with placebo, 93.9% patients were white. Demographic details are shown in Table 1. 

In the A0221045 study, the MMSE score remained unchanged in 115/356 (32.3%) placebo treated and 124/341 (36.4) fesoterodine treated patients.  For the A0221049 study, these figures were 79/220 (35.9%) placebo treated and 75/223 (33.6%) fesoterodine treated patients. Figure 1 shows the proportions of patients in each trial whose MMSE score either improved, remained unchanged or worsened. There was no statistically significant difference between placebo and fesoterodine exposed patients (p = 0.73).
 
The observed changes for the small number of patients (39 receiving placebo and 27 receiving fesoterodine) with a baseline MMSE score of <25 are shown in figure 2.  There was no statistically significant difference between placebo and fesoterodine exposed patients (p = 0.87).

For all placebo and fesoterodine treated patients, approximately 1/3 either improved their score, achieved the same score or declined by 1 point or more.  For those with a baseline MMSE score of <25, denoting cognitive impairment, the majority of patients, albeit a small sample improved their score – probably demonstrating a regression to the mean with repeated testing. Another potential explanation for the degree of MMSE change seen is the practice effect. However, this is unlikely, given the fact that the distribution of scores that “improved” on the MMSE from screen to baseline was similar to those that “declined”.  This study used a conservative definition of change, the minimally important difference in change in MMSE has ben reported as between 2.4 – 3.7 in clinical trials of dementia drugs. We could demonstrate no other obvious pattern of either improvement or decline in MMSE regardless of baseline score. The MMSE is most likely insensitive to the degree of change previously observed in dedicated cognitive function assessments for OAB drugs. The sensitivity of the Montreal Cognitive Assessment Score, an alternative short assessment used in testing for mild cognitive impairment and executive dysfunction is likewise not yet known.  In these 2 clinical trials the reported incidence of cognitive adverse events was <1%.

In medically complex older adults with OAB treated with fesoterodine, the MMSE showed no consistent pattern of change over three months. The pattern of change was not statistically different between fesoterodine- and placebo-treated patients. MMSE should not be routinely used to monitor cognitive change in response to OAB therapy.

JAMA Internal Medicine, 2015. 175(3): p. 401-7
BMJ, 2018. 361: p. k1315
Drugs Aging. 2017 Aug;34(8):615-623