Differential Effects of the Antimuscarinic Agents Darifenacin and Oxybutynin ER on Memory in Older Subjects

1. Introduction 

Overactive bladder (OAB) is a widespread condition, the prevalence of which rises with increasing age [1]. As bladder contractions are mediated primarily by cholinergic activation of muscarinic M3 receptors [2], antimuscarinics are used widely as first-line OAB treatments [3]. Some agents may, however, be associated with safety concerns, including effects on the central nervous system (CNS), e.g., memory impairment [3]. The potential for CNS safety issues is of particular concern in older patients, who are more vulnerable because of age-related memory decline [4], reduced brain muscarinic receptor density [5] and comorbidities [6]. Furthermore, clinical studies have demonstrated increased sensitivity of older subjects to antimuscarinics, including effects on memory [7], [8]. Consequently, selecting an appropriate antimuscarinic for OAB requires balancing efficacy with possible effects on memory.

An important differentiator between antimuscarinics is activity at muscarinic receptor subtypes (M1–M5). While oxybutynin binds preferentially to M3 and M1 receptors, darifenacin demonstrates 9.3-fold selectivity for M3 over M1 receptors in-vitro [9]. This observation may be important for effects on memory, as the muscarinic M1 receptor plays a role in memory/cognition [10], [11]. Indeed, it has been proposed that M3 selectivity may confer benefits, as non-M3-receptor–mediated CNS side effects may be avoided (or reduced) [12]. Supporting evidence includes two studies of healthy subjects (one in subjects ≥65 years), in which darifenacin had no effect versus placebo on cognition [13], [14]. In contrast, a small-scale clinical study showed that oxybutynin was associated with cognitive dysfunction [15].

To our knowledge, there are no reports of a single-controlled study investigating the effects of two separate antimuscarinics on memory. We report a comparison of the effects of darifenacin and extended-release (ER) oxybutynin on memory in older subjects.

3. Results 

3.1. Subjects 

One hundred and fifty subjects (darifenacin n=49, oxybutynin ER n=50, placebo n=51) were randomised and comprised the safety population. Of these, 134 completed the study and formed the modified ITT population for the primary analysis (Fig. 2). Of the nine subjects who discontinued in the darifenacin group, six had partial data and were included in secondary analyses. There were six discontinuations in the oxybutynin ER group, of which partial data available for five subjects were included in secondary analyses. One subject in the placebo group discontinued, for whom partial data were not available for inclusion in secondary analyses (Fig. 2). Demographics and baseline characteristics were similar across treatment groups (Table 2).

Fig. 2. Patient flow through the 3-week study.

Table 2. Subject demographics and baseline characteristics

		Darifenacin (n=49)	Oxybutynin ER (n=50)	Placebo (n=51)
	Mean age (yr) (range)	66.4 (60–82)	68.0 (60–81)	67.4 (61–83)
	Female (n [%])	29 (59.2)	31 (62.0)	33 (64.7)
	Mean BMI (kg/m2) (range)	25.9 (19–30)	26.7 (21–30)	25.7 (19–30)
	Race (%)
	Caucasian	93.9	94.0	94.1
	Black	4.1	6.0	3.9
	Pacific Islander	2.0	0	0
	Other	0	0	2.0
	Mean baseline score for delayed recall on:
	Name–Face Association Test*	5.2	5.8	5.4
	First–Last Name Association Test†	1.7	1.8	1.6

BMI=body mass index.

* Modified intent-to-treat population (primary): darifenacin n=40, oxybutynin ER n=44, placebo n=50. † Modified intent-to-treat population (secondary): darifenacin n=46, oxybutynin ER n=49, placebo n=50.

3.2. Assessment of memory – delayed recall 

There was no significant difference between the darifenacin and placebo groups with respect to the primary end point, delayed recall on the Name–Face Association Test at week 3 (mean difference, −0.06, p=0.908). In contrast, scores for delayed recall on the Name–Face Association Test were significantly lower in the oxybutynin ER group than the placebo group (mean difference, −1.30, p=0.011) or darifenacin group (mean difference, −1.24,=0.022), indicating memory deterioration (Fig. 3).

Fig. 3. Effects of darifenacin, oxybutynin ER and placebo on accuracy of delayed recall on the Name–Face Association Test at each time point. ER=extended release; ANCOVA=analysis of covariance.

Results from the Name–Face Association Test at week 2 were consistent with those at week 3, when there also was no significant difference between the darifenacin and placebo groups for delayed recall (Fig. 3, Table 3). For oxybutynin ER, scores at week 2 were significantly lower than for placebo or darifenacin (mean differences, −0.99, p=0.022 and −1.23, p=0.007, respectively), showing that the memory impairment at week 3 also was evident at week 2 (Table 3). There were no significant between-treatment differences at week 1, when lowest doses were administered.

Table 3. Accuracy of delayed recall on the Name–Face Association Test over time*

	Treatment	n	Comparator	Estimated LSM difference	95% CI	p value
	Week 1
	Darifenacin 7.5mg	45	Oxybutynin ER 10mg	0.61	−0.26, 1.49	0.170
			Placebo	0.32	−0.54, 1.19	0.463
	Oxybutynin ER 10mg	49	Placebo	−0.29	−1.13, 0.55	0.497
	Placebo	50	–	–	–	–
	Week 2
	Darifenacin 7.5mg	42	Oxybutynin ER 15mg	1.23	0.35, 2.12	0.007
			Placebo	0.25	−0.62, 1.12	0.576
	Oxybutynin ER 15mg	47	Placebo	−0.99	−1.83, −0.15	0.022
	Placebo	49	–	–	–	–
	Week 3
	Darifenacin 15mg	40	Oxybutynin ER 20mg	1.24	0.18, 2.29	0.022
			Placebo	−0.06	−1.08, 0.96	0.908
	Oxybutynin ER 20mg	44	Placebo	−1.30	−2.28, −0.31	0.011
	Placebo	50	–	–	–	–

ER=extended release; LSM=least square mean.

* Analysis of covariance model adjusted for baseline score, age and gender. Negative differences indicate relatively worse scores.

For darifenacin and placebo, there was a trend for improvement during the study (Fig. 3), reflecting a learning effect whereby subjects improve through practise. Thus, by week 3, mean scores for delayed recall on the Name–Face Association Test had increased by 0.9 and 1.0 in the placebo and darifenacin groups, respectively. In the oxybutynin ER group, in whom a similar learning effect was expected, a decrease in performance by −0.8 was observed.

In delayed recall on the First–Last Name Association Test, oxybutynin ER resulted in significant impairment versus placebo (p<0.05) at weeks 1 and 2 (Fig. 4; Table 4). In contrast, no significant differences were observed between darifenacin and placebo at any time point (Fig. 4; Table 4).

Fig. 4. Effects of darifenacin, oxybutynin ER and placebo on accuracy of delayed recall on the First–Last Name Association Test at each time point. ER=extended release; ANCOVA=analysis of covariance.

Table 4. Scores in additional tests of memory and cognitive function over time

	Test	Treatment	Comparator	Estimated LSM difference
				Week 1	Week 2	Week 3
	Immediate memory recall
	Name–Face Association (accuracy, second acquisition)	Darifenacin	Placebo	0.10	0.48	−0.30
			Oxybutynin ER	−0.13	0.91	0.44
		Oxybutynin ER	Placebo	0.23	−0.43	−0.74
	First–Last Name Association (accuracy, second acquisition)	Darifenacin	Placebo	−0.13	−0.26	−0.00
			Oxybutynin ER	0.15	0.29	0.32
		Oxybutynin ER	Placebo	−0.28	−0.55*	−0.32
	Facial Recognition (accuracy, correct before first miss)	Darifenacin	Placebo	0.36	0.28	0.29
			Oxybutynin ER	1.22	−0.09	1.66
		Oxybutynin ER	Placebo	−0.87	0.37	−1.37
	Delayed memory recall
	First–Last Name Association (accuracy)	Darifenacin	Placebo	−0.26	−0.11	0.18
			Oxybutynin ER	0.27	0.43	0.57
		Oxybutynin ER	Placebo	−0.53*	−0.53*	−0.39
	Misplaced Objects (correct recall at first attempt)	Darifenacin	Placebo	−0.73	−1.13	−0.73
			Oxybutynin ER	−0.42	0.38	0.30
		Oxybutynin ER	Placebo	−0.31	−1.51†	−1.03*
	Visual attention
	Matching to Sample (efficiency)	Darifenacin	Placebo	−0.61	−0.27	−1.94
			Oxybutynin ER	−2.05	−1.04	−1.83
		Oxybutynin ER	Placebo	1.44	0.77	−0.11
	Visual Sequence Comparison (efficiency)	Darifenacin	Placebo	−0.15	−1.06	−2.33
			Oxybutynin ER	−0.81	−1.89	−1.24
		Oxybutynin ER	Placebo	0.65	0.83	−1.09
	Information-processing speed
	Divided Attention (Sequence Comparison Speed, Dual Condition) (s)	Darifenacin	Placebo	0.03	0.06	0.30*
			Oxybutynin ER	−0.07	0.08	0.24
		Oxybutynin ER	Placebo	0.10	−0.02	0.06
	Divided Attention (Sequence Comparison Efficiency) (s)	Darifenacin	Placebo	0.42	0.31	−2.19
			Oxybutynin ER	−0.20	0.19	−1.98
		Oxybutynin ER	Placebo	0.61	0.12	−0.21
	Divided Attention (Sequence Comparison Accuracy) (s)	Darifenacin	Placebo	−1.47	2.21	2.21
			Oxybutynin ER	−0.88	2.38	2.38
		Oxybutynin ER	Placebo	−0.59	−0.17	−0.17
	Visual Sequence Comparison (median reaction to correct response)	Darifenacin	Placebo	0.04	0.07	0.12
			Oxybutynin ER	0.00	0.11	0.10
		Oxybutynin ER	Placebo	0.03	−0.03	0.01
	Divided Attention (Single Task Premature Hits) (s)	Darifenacin	Placebo	0.21	0.50	0.12
			Oxybutynin ER	0.10	0.56*	0.38
		Oxybutynin ER	Placebo	0.11	−0.06	−0.27
	Divided Attention (Task Reaction Time, Dual Condition) (s)	Darifenacin	Placebo	0.05	0.10	0.02
			Oxybutynin ER	−0.01	0.04	−0.04
		Oxybutynin ER	Placebo	0.06	0.06	0.06
	Divided Attention (Premature Hits, Dual Condition) (s)	Darifenacin	Placebo	0.40	−0.20	0.12
			Oxybutynin ER	0.25	0.22	0.39
		Oxybutynin ER	Placebo	0.15	−0.42	−0.27
	Psychomotor/reaction time
	Divided Attention (Response Speed to Visual Monitoring Task Alone) (s)	Darifenacin	Placebo	−0.01	0.00	−0.03
			Oxybutynin ER	−0.01	−0.01	−0.03
		Oxybutynin ER	Placebo	−0.00	0.01	0.01

Modified intent-to-treat population. ER=extended release; LSM=least square mean.

* p<0.05. † p<0.01 (analysis of covariance [adjusted for baseline score, age and gender]).

In the Misplaced Objects Test, oxybutynin ER resulted in significantly lower scores than placebo at weeks 2 and 3 for correct recall at first attempt (Table 4), suggesting a decline in performance, whereas darifenacin was not significantly different from placebo at any time point.

3.8. Adverse events 

The incidence of all-causality AEs is shown in Table 5. The most frequently reported AEs, as expected for this class, were dry mouth and constipation. Dry mouth occurred more frequently during oxybutynin ER than darifenacin treatment (40.0% vs 26.5%). One patient in each of the oxybutynin ER and darifenacin groups discontinued because of dry mouth. The incidence of constipation was higher in the darifenacin than oxybutynin ER group (Table 5). Only one patient (in the darifenacin group) discontinued because of constipation. The total incidence of all-causality nervous system events was similarly low in all groups, with only two severe cases, both in the oxybutynin ER group (Table 5). There was one serious AE (hip fracture following an accident at home in a subject given oxybutynin ER), which was not considered to be related to the study drug. There were no clinically significant findings from assessments of laboratory values or vital signs.

Table 5. Adverse event incidence (safety population)

		Darifenacin (n=49)	Oxybutynin ER (n=50)	Placebo (n=51)
	Subjects with any adverse event (n)	27	26	23
	Treatment-related	26	22	16
	Severe adverse events	4	7	1
	Serious adverse events	0	1	0
	Most common all-causality adverse events (n)
	Dry mouth	13	20	6
	Constipation	10	2	1
	Dyspepsia	3	2	1
	All-causality nervous system events (n)	5	4	3
	Severe	0	2	0

4. Discussion 

This study demonstrated that the M3 selective receptor antagonist darifenacin had no significant effect on memory in older subjects. In contrast, oxybutynin ER resulted in significant memory deterioration, as measured by delayed recall on the Name–Face Association Test at week 3. Comparing these results with normative data for this test [4] indicates that the degree of memory change seen with oxybutynin ER (baseline to week 3) was comparable to a decline that occurs over the course of 10 years in the normal aging process. Additional tests of delayed recall indicated significant memory impairment with oxybutynin ER versus placebo (Name–Face Association at week 2, First–Last Name Association at weeks 1 and 2, and Misplaced Objects at weeks 2 and 3), while darifenacin was not significantly different from placebo in delayed recall at any time point.

The delayed recall tests were selected on the basis of their relevance to daily activities. Recalling the name of someone to whom one is introduced is the most problematic memory task faced on a daily basis in many cultures, and performance declines markedly over the adult life-span [4], [22]. For example, performance on the Name–Face Association Test [20] declines by >65% between age 25 and 75 years [4]. This ‘normal’ decline may be exaggerated by drugs [24], and the combined effect would be expected to be of clear clinical significance. In a similar manner, performance declines with age on the First–Last Name Association [19] and Misplaced Objects Tests [23]; this effect also can be exaggerated by drugs [24]. Thus, the deleterious effects of oxybutynin ER on these tests suggests that this agent, at the dosage tested, may be associated with diminished performance on important tasks of daily life that depend on delayed recall.

Differential outcomes between darifenacin and oxybutynin ER may arise from differences in either CNS penetration or muscarinic receptor-binding profiles. Darifenacin exhibits limited CNS penetration in preclinical studies, which may result from its moderate lipophilicity, relatively large molecular size, polarity and active efflux across the blood-brain barrier via the P-glycoprotein pump [25]. Once in the CNS, muscarinic-binding profiles play a role. Whereas darifenacin shows marked M3 selectivity, oxybutynin demonstrates high affinity for M3 and M1 subtypes [9]. The latter subtype is abundant in the neocortex, hippocampus and neostriatum, in contrast with low levels of M3 receptors in the brain [26], and M1 receptors are known to be particularly important for memory/cognition [10], [11]. This hypothesis is supported by the low incidence of nervous system AEs with darifenacin, both here and in longer-term clinical trials. A pooled analysis of three 12-week, fixed-dose studies with darifenacin showed a profile of nervous system events that was comparable with placebo, both overall and in patients ≥65 years [27], [28]. In contrast, results from four clinical studies of ≤4 months showed that the incidence of somnolence and dizziness with oxybutynin ER 5–30mg/day was 12% and 6%, respectively [16].

This study also measured immediate recall, for which oxybutynin ER showed some impairment versus placebo, while darifenacin and placebo were not significantly different. Major changes were not expected here, as muscarinic receptor activation is thought to be involved primarily in memory consolidation [29] (i.e., how recent recollections are crystallised into memory). Similarly, significant effects were not expected in attention tests, and no effects were observed across multiple tests. Within the Divided Attention Test, however, darifenacin was significantly worse than placebo for sequence comparison speed (mean difference, 0.3s) but did not differ from placebo in accuracy. Darifenacin scored worse than oxybutynin ER in the number of premature hits on a reaction time measure at week 2 (but not week 3) and did not differ significantly from placebo. Given that subjects receiving darifenacin performed no differently than those on placebo/on multiple other tests of attention, these isolated findings are not considered clinically relevant. In contrast, findings with oxybutynin ER on memory were replicated across multiple tests at different time points, and are supported by earlier studies and an identifiable mechanism of action.

Interestingly, there were no reported differences in self-rated memory between treatments. This finding is particularly important as it indicates that memory changes may go unnoticed. This low level of awareness may account for the low rate of reporting of memory impairment with antimuscarinic therapies in clinical practice. In addition, disease- or treatment-related memory/cognitive impairment may be difficult to recognise, and memory decline may be attributed wrongly to aging.

Although relevant for all patients, these findings are particularly important for older patients, since this population may have risk factors for memory/cognitive impairment [30]. Important differences between this study and clinical practice are that frail elderly patients may have been under-represented (since ability to complete computerised cognitive tests and normal MMSE scores were required), and use of multiple medications with anticholinergic activity will be common in older patients in clinical practice [30]. In the study reported here, where patients were not receiving anticholinergic co-medication and had no cognitive impairment at baseline, the lowest oxybutynin ER dose (10mg od) did not cause memory impairment. In clinical practice, however, when patients may be receiving concomitant anticholinergics or have existing memory impairment, it is possible that this dose of oxybutynin ER may have a greater impact.

5. Conclusions 

The results of this study add considerably to our knowledge regarding the differential effects of two antimuscarinics, darifenacin and oxybutynin ER, on memory. The finding that darifenacin does not impair memory is consistent with earlier studies [13], [14], and the observation assumes clinical significance in light of the clear demonstration of memory impairment during oxybutynin ER treatment. These findings highlight a need for further studies to fully establish the effects of all OAB antimuscarinics on memory/cognition.Conflicts of interest

Preparation of this manuscript was supported by Novartis Pharma AG, and editorial and project management services were provided by ACUMED®.