1a. In head to head trials of anticholinergic OAB drugs what is the comparative efficacy?

We found 24 head-to-head trials of oxybutynin, tolterodine, trospium, flavoxate, solifenacin, and/or darifenacin. All included studies are summarized in Evidence Table 1. Study quality assessments are presented in Evidence Table 2.

No good quality study was found. The only two flavoxate studies^{23, 24}, one study comparing oxybutynin IR and tolterodine IR²⁵ and two studies comparing oxybutynin immediate and extended release²⁶, were assessed as poor quality, and all others were fair quality. The poor quality studies suffered from lack of details on randomization, allocation concealment and baseline characteristics or lack of randomization and differences in potentially important baseline characteristics. Ten studies used an intention to treat analysis overall, three studies used an intention to treat analysis for adverse events but not efficacy. The poor quality studies are not discussed here (see Evidence Tables 1 and 2). Since no fair or good quality head-to-head study of flavoxate was found, no results are presented for that drug.

The included studies had similar eligibility and exclusion criteria, largely enrolling patients with exclusively or predominantly urge incontinence. One trial involving trospium and oxybutynin only included patients with a spinal cord injury.²⁷ Some studies enrolled patients with combined stress and urge incontinence, with symptoms of urge predominant. The studies enrolled significantly more women than men and although the age ranges of enrolled patients were wide the mean age for most studies was approaching 60 years. These gender and age trends reflect the typical characteristics of the population with urge incontinence. One study included only female patients²⁸, and another analyzed the female subgroup of a larger trial.²⁹ Eight of 16 fair quality studies were conducted at least in part in the US, while the others were conducted primarily in European countries, except for a few that were conducted in Asia and Canada.

We found six fair quality studies comparing an immediate release formulation of one anticholinergic OAB drug to another.^{28, 30, 31, 32} Four of these studies compared oxybutynin to tolterodine and were all sponsored by Pharmacia (the makers of tolterodine). Tolterodine was dosed at 2mg twice daily in all studies, while oxybutynin was dosed at 5mg twice daily in two studies^{28, 32} and 5mg three times daily in two.^{31, 33} Two compared immediate release formulations of oxybutynin to trospium. One trial was sponsored by a company that makes trospium, the other did not report sponsorship. The study durations ranged from 2 to 54 weeks.

We found six fair quality studies comparing an extended release formulation of an anticholinergic urinary incontinence drug to an immediate release formulation.^34–39 Four studies compared oxybutynin ER to oxybutynin IR,^{36–38, 40} one tolterodine ER to tolterodine IR,³⁴ one study oxybutynin ER to tolterodine IR,⁴¹ and one tolterodine ER to oxybutynin IR. Oxybutynin doses ranged from 5mg to 30mg a day, while tolterodine was dosed at 4mg a day.

Of the four studies comparing oxybutynin ER to oxybutynin IR, one was 6 weeks in duration, and compared 10mg per day of oxybutynin either ER once daily or 5mg IR twice daily ER to IR. Patients getting the ER formulation received a single daily 10mg dose; patients on the IR formulation got 2 daily 5mg doses.³⁶ The other three studies^{37, 38, 40} used a dose titration up to the threshold of either intolerable side effects (in which case the dose was reduced by 5mg per day) or maximum efficacy. In the latter case, the "optimal" dose was then maintained for 7 days. Total, mean or range duration of trial actually experienced were not reported. In one study the "optimal" dose was maintained for 4 weeks and the efficacy analysis was performed with an "efficacy" population defined as patients, who completed ≥ 2 weeks in the stable-dose phase and did not have major protocol violations.⁴⁰ All four studies were funded by or had authors from the companies who make the extended release formulations involved.

We found only one study comparing tolterodine ER 4mg once daily to tolterodine 2mg twice daily (a placebo arm was also included) for 12 weeks.³⁴ This was the largest study included, with over 500 patients per treatment, and used an intention to treat analysis. In a subgroup analysis of women only from this study, 1235 women were included.

One study of tolterodine ER compared to oxybutynin IR in 600 Asian patients (Japanese or Korean) was found. Doses compared were 4mg tolterodine and 9mg oxybutynin daily. The manufacturer of tolterodine ER provided funding, and the formulation of oxybutynin IR used does not appear to be available in the US.

One study compared oxybutynin ER to tolterodine IR.⁴¹ This study compared oxybutynin 10mg once daily and tolterodine 2mg twice daily for 12 weeks. The funding was provided by the manufacturer of the extended release form of oxybutynin (Alza) and one of the authors was employed by this company.

Two studies comparing extended release formulations of oxybutynin and tolterodine to each other were found.^{21, 42} The OPERA trial enrolled 790 women to take either tolterodine ER 4mg or oxybutynin ER 10mg daily for 12 weeks.²¹ The manufacturer of oxybutynin ER provided the funding for this study. In the second study (the ACET trial), oxybutynin was dosed at 5 to 10mg once daily and tolterodine at 2 to 4mg once daily.⁴² Funding of this study was not reported. The study design was unusual and problematic, in that it consisted of two separate trials. One trial randomized patients to one of two doses of tolterodine in an open label (unblended) fashion. The other randomized patients to one of two doses of oxybutynin. Other than the two drugs, the same protocol was used at each center, however the choice of which trial (drug) each center was assigned appears to have been at the discretion of the investigators, therefore this could not be considered a purely randomized trial. They state that centers were assigned based on 1) geographic location, and 2) prescribing patterns for both drugs with an effort to produce balance.

The transdermal (TD) form of oxybutynin (which received FDA approval in late February 2003) was studied compared to oxybutynin IR and tolterodine ER in separate studies.^{43, 44} The study of oxybutynin TD versus oral IR allowed dose titration, from 1.3 to 5.2mg per day via patch, or 5 to 15 mg per day orally.⁴³ The other study randomized patients to 3.9mg/day TD or 4mg/day tolterodine ER. The manufacturer of the TD system funded both studies.

Two studies comparing trospium chloride to oxybutynin IR were found. The first trial conducted in multiple German centers compared trospium 20mg twice daily (plus a mid day placebo dose) to oxybutynin IR 5 mg three times daily. All the subjects in this trial had spinal cord injuries. No included outcomes were reported. The trial is discussed in the section on subpopulations (Key Question 3).²⁷ The second trial was conducted in multiple European centers comparing trospium 20 mg twice daily to oxybutynin IR 5 mg twice daily. One author of this study was from a pharmaceutical company that manufactures trospium in Europe. Data was collected over an average of 54 weeks at multiple intervals.⁴⁵

One fair quality systematic review²⁰ reported efficacy differences between antimuscarinics (oxybutynin, tolterodine, trospium, darifenacin and solifenacin) using clinical outcomes. This review concluded that solifenacin resulted in significantly greater reductions in urgency episodes and micturition frequency when compared to tolterodine IR. The original study⁴⁶ compared the active medications to placebo in the primary analysis, and conducted only "exploratory" analyses of tolterodine versus solifenacin. The systematic review concluded that oxybutynin ER caused a significantly greater mean reduction in incontinence episodes and a significant increase in the number of patients who returned to continence when compared to tolterodine ER. This change in incontinence episodes was not reported as a significant difference in original OPERA trial²¹ and the authors of the systematic review appear to have used a per protocol analysis to calculate relative risk values, resulting in a significant difference. The proportion returned to continence was not an outcome measure included to assess efficacy in this systematic review.

Two studies compared solifenacin to tolterodine (one tolterodine IR and the other tolterodine ER). The first, a fair-quality study, compared solifenacin 5mg, solifenacin 10mg, tolterodine IR 2mg twice daily, and placebo.⁴⁶ This study was not powered to show treatment differences between the active treatment arms. Thus, the authors did not do a statistical analysis of the drugs compared to one another, but they did statistical analyses of each drug compared to placebo.

The second study, the STAR trial⁴⁷ was designed as a "non-inferiority trial" with respect to its primary outcome (mean number of micturitions/24 hours), such that claims of superiority are not intended to be drawn from this data. In this trial, Chapple et al.⁴⁷ compared tolterodine ER 4mg to a "flexible" dose of solifenacin (5mg or 10mg) over a total of 12 weeks. Patients were randomized to either tolterodine 4 mg ER or solifenacin 5mg for the first four weeks. Then, at week 4, patients were allowed to increase their dose if they were not satisfied with treatment efficacy. The resulting dose was maintained for the remaining eight weeks of the study. The investigators stated that using flexible dosing allowed the trial to mirror clinical practice as closely as possible. One problematic aspect of this trial in its efficacy and safety analysis should be noted: data for both doses of solifenacin are combined in the analyses of outcome results. Thus, it is not possible to determine which dose of solifenacin provided greater efficacy or if the different doses caused a difference in rates of adverse events. Because the authors did not do a statistical analysis of the adverse events between solifenacin and tolterodine, we did a statistical analysis of the adverse events rates of the STAR trial ourselves, using the StatsDirect program.

Incontinence episodes and micturitions per 24 hours

Immediate release vs Immediate release

The objective measures in these studies were mean change in numbers of incontinence episodes per day or micturitions per day. Four studies examined oxybutynin versus tolterodine immediate release formulations. One study by Leung²⁸ did not report the actual data for these outcomes, but reported that by analysis of variance there were no significant differences between the groups. In the other three studies, the range of mean change in micturitions per day in the tolterodine groups was −1.7 to −2.7 and in the oxybutynin groups −1.7 to −2.3. The range of mean change in number of incontinence episodes per day for tolterodine was −1.3 to −2.2, and for oxybutynin −1.4 to −1.8. The difference in standardized effect sizes of the mean change (from baseline to end of study) reported in these studies is plotted in Figures 2 and 3. One study examined trospium versus oxybutynin IR formulations. Significant differences were not found for micturition frequency, incontinence episodes or urgency episodes.⁴⁵ No significant differences were found between the drugs, by intention to treat analysis, in any study.

Figure 2

Incontinence episodes per day; IR versus IR

Figure 3

Micturitions per day; IR versus IR

Immediate release vs Extended release

Two studies^{37, 38} using a dose titration of oxybutynin ER or IR to adverse effects or efficacy reported no significant difference between groups in the mean change in incontinence episodes per week (rather than per day), but not enough data was reported to allow graphing. Converted to mean change in incontinence episodes per day, the mean change in the ER groups was −3.2 and −2.2, and in the IR groups was −2.9 and −2.2 in the Anderson and Versi studies, respectively. Time period from baseline to assessment was not reported. Neither study used an intention to treat analysis. The withdrawal rate for extended release and immediate release groups was 13% (ER) and 12% (IR)³⁸ and 6% (ER) and 8% (IR).³⁷ Alza, manufacturer of the oxybutynin ER formulation, funded both studies.

A study of 10mg oxybutynin ER once daily or 5mg oxybutynin IR twice daily³⁶ used an ER formulation used in this study is not available in the US and different outcome measures than the other studies: the proportion with day and nighttime continence, day/night micturition, and day/night incontinence episodes. For these reasons, we did not evaluate this study further here.

An additional study comparing titrated "optimal" doses of 10–30mg oxybutynin ER once daily (increasing in 5mg increments) or 5mg oxybutynin IR twice daily (also maximum of 30mg/day) showed that the mean titrated doses were similar, with 15.2 mg for the controlled release version and 14.0 mg for the immediate release formulation.⁴⁰ Baseline reductions in incontinence episodes/24h were 2.0 and 2.4 for the controlled release and immediate release forms respectively. Similarly, reductions in micturitions/24h at the end of treatment were 1.8 and 2.4 for ER and IR forms respectively. These differences were not statistically significant. This was a study of a controlled-release product newly introduced into the Canadian market by Purdue Pharma and is currently not available in the U.S.

In the OPERA study of tolterodine ER 4mg once daily versus tolterodine 2mg twice daily,³⁴ no significant differences were found in mean change (absolute) in micturitions or incontinence episodes per week (see Figure 4). Converted to per day, the mean change in incontinence episodes was −1.6 (ER) and −1.5 (IR) and mean change in number of micturitions per day was −3.5 (ER) and −3.3 (IR). Mean change in the number of urinary pads used per day was −3.3 in both groups. The median percent change in incontinence episodes was also reported. The percent reduction was 71% ER, 60% IR, 33% placebo. The authors state that they present this outcome because the data were positively skewed, and that they believe the relative change is more relevant than the absolute change. Few other studies report data in this way. This outcome measure and the fact that the data were skewed limit comparability of these results to those in other trials. Overall withdrawal was 12%, with similar rates in the two drug treatment groups. A subgroup analysis of only women in this study found similar results, and is discussed under Question 3.²⁹

Figure 4

Tolterodine ER vs IR (Van Kerrebroeck 2001)

Oxybutynin ER was compared to tolterodine IR in one study.³⁵ Based on an analysis of covariance, adjusting for baseline and severity of symptoms, oxybutynin ER was significantly more effective at reducing the number of incontinence episodes per week (p = 0.03), and number of micturitions per week (p = 0.02). This analysis was not intention to treat; the proportions of patients excluded from the analysis are 14% in the oxybutynin ER group and 11% in the tolterodine group. Therefore, due to dropouts, the analysis presented may not reflect actual reductions in efficacy. Insufficient data were presented to calculate the mean change in incontinence or micturitions based on intention to treat for this review.

Tolterodine ER was compared to oxybutynin IR in Japan and Korea.³⁹ No significant differences were found in percent change in median incontinence episodes, pad use, or micturitions per day. The median percent change in incontinence episodes was 78.6% for tolterodine, and 76.5% for oxybutynin. The absolute change is not reported, and again the data were reported to be skewed. The changes in voids per day were −2.1 and −2.0 for tolterodine and oxybutynin, respectively. There was no change in pad use, however.

A study of solifenacin 5mg or 10mg once daily and tolterodine IR 2mg twice daily demonstrated that both doses of solifenacin and tolterodine produced significantly lower mean number of micturitions when compared with placebo.⁴⁶ Solifenacin at both doses, but not tolterodine, resulted in statistically significant improvements in urge and overall incontinence episodes per 24 hours, and episodes of urgency. "Exploratory" analyses of solifenacin versus tolterodine resulted in significant differences favoring solifenacin at both doses in reducing the episodes of urgency, and solifenacin 10mg was superior to tolterodine in reducing micturitions per 24 hours.

The STAR trial, which was designed as a "non-inferiority" trial for its primary outcome,⁴⁷ examined the difference between a "flexible" dose of solifenacin (5mg or 10mg) and tolterodine ER 4 mg (all once daily), which revealed significantly higher rates of efficacy in decreasing urgency episodes, incontinence, urge incontinence and pad usage for the solifenacin patients.⁴⁷ However, the study did not demonstrate statistically significant between-treatment differences in the primary endpoint, micturitions/24h, nor in nocturia episodes. Thus, the finding of no between-treatment statistically significant difference for the primary outcome qualifies as evidence that solifenacin was "non-inferior" compared to tolterodine ER in reference to this primary endpoint. Based on the flexible-dose study design of this trial, both doses of solifenacin were combined and analyzed as one group for outcomes data and rates of adverse events.

The trials of solifenacin are reported here because although the dosing is once daily, this is based on elimination pharmacokinetics, not an extended release formulation.

Extended release vs Extended release

The OPERA trial²¹ randomized 790 patients to 10mg daily of oxybutynin ER or 4mg daily of tolterodine ER for 12 weeks. 47% of patients had prior anticholinergic drug therapy for urge incontinence. There was no difference between the groups in the mean change in urge incontinence episodes (−26.3 vs −25.5 per week, oxybutynin vs tolterodine) which was the primary outcome measure. There were also no differences found based on mean change in total incontinence episodes. Differences were found in the proportion of patients with no incontinence (23% vs 17%; p = 0.03) and the mean change in micturitions per week (28.4% vs 25.2%, p=0.003) at week 12, in favor of oxybutynin ER. This study was fair quality, and used the last observation carried forward technique to conduct an intention to treat analysis on these efficacy measures.

The other study comparing the two extended release formulations did not report these outcomes.⁴²

Transdermal vs Immediate release

A six-week study comparing oxybutynin TD to oxybutynin IR assigned the starting dose depending on the previous dose of oxybutynin (patients were required to have been on oxybutynin for at least 6 weeks, and to have had symptomatic improvement).⁴³ Dose was then titrated to effect or side effects over the 6-week study period. 76 patients were enrolled. No significant differences were found in this small study in the percent change in mean incontinence episodes (66.7% vs. 63.9%) or the number with zero incontinence episodes in week 6 (21% vs.26%).

Transdermal vs Extended release

A study of 361 patients randomized to oxybutynin TD 3.9mg per day or 4mg tolterodine ER per day or placebo.⁴⁴ All patients had to have been taking an anticholinergic drug for incontinence, with symptomatic improvement prior to enrollment. The distribution of those taking oxybutynin (oral) and tolterodine prior to enrollment was about even in all groups. No significant differences were found between these drugs based on mean change in incontinence episodes per day at 12 weeks (oxybutynin TD −2.9, tolterodine ER −3.2 p=0.5878) or mean decrease in urinary frequency per day (oxybutynin TD −1.9, tolterodine ER −2.2, p=0.2761).

1b. In trials of anticholinergic OAB drugs compared to non-drug therapy, other drug therapy, or placebo what is their comparative efficacy?

We found six trials of one of the three anticholinergic incontinence drugs compared to another drug not currently used or not on the market in the U.S. (Evidence Table 3).^48–53 Two used oxybutynin^{52, 53} and four used flavoxate.^48–51 The mean change in micturitions per day in the two oxybutynin studies were −2.5 and −2.4, within the range of change seen in the head-to-head trials. The flavoxate studies used 200mg three or four times a day. Two studies were for only 7 days, one for 14 days and one for 6 weeks. The ability to compare the results of these studies to results found with oxybutynin or tolterodine is extremely limited, as only one study (20 patients, 14 day duration) used outcome measures similar to the head-to-head trials.⁴⁸ This study enrolled women, mean age 51 years, with cystometry-proven detrusor instability in a randomized crossover study of flavoxate, emepronium or placebo. The mean change in number of micturitions per day for flavoxate was +1 (emepronium −0.5, placebo −1). The mean change in the number of incontinence episodes per day was −1 for flavoxate (emepronium −1, placebo −2). This is also the only study that met fair quality criteria; all others were poor primarily because of lack of important details such as eligibility and exclusion criteria.

We found five studies comparing oxybutynin to non-drug therapy (bladder training, electrostimulation therapy)^54–58 These trials are summarized in Evidence Table 4. Four of these appear to be reporting different outcomes from the same trial and will be treated as one study.^{54, 56, 58} Two studies reported the mean change in number of micturitions per day (oxybutynin IR −2, and −2.1) or mean change in incontinence episodes per week (oxybutynin IR −10.2).^54–56 These data are within the range reported in the head-to-head trials. The other studies report outcomes such as proportion with clinical cure (IR 73%) or change on Global Severity Index (IR 2.1), which were not used by other studies of oxybutynin, tolterodine or flavoxate. These studies include an extension phase of a trial that reports the results of combination therapy (IR and behavioral therapy) following the completion of the original randomized trial comparing oxybutynin IR and behavioral therapy.⁵⁹

We found 24 placebo-controlled trials (one is an unpublished study that was provided by the manufacturer)^{46, 60–81} and three systematic reviews^{19, 20, 82} of OAB drugs. Only one of the systematic reviews did not present enough data to assess individual drugs; it assessed the effectiveness of any anticholinergic OAB drug compared to placebo.⁸² Nine of the trials that met inclusion criteria assessed tolterodine compared to placebo.^62–68 The results for these trials on mean change in number of micturitions or incontinence episodes per day are presented in Evidence Table 5. The range of mean change in micturations/24h for tolterodine 4mg daily is −0.1 to −2.3, while the placebo rates range from 0 to 1.4. In the head-to-head trials, the range of mean change in micturations/24h for tolterodine 4mg/day was −1.7 to −3.5. The range of mean change in incontinence episodes/24h is −0.7 to −1.7 and placebo ranges from 0 to −1.9. In the head-to-head trials, the range of mean change in incontinence episodes/24h for tolterodine was −1.3 to −2.2. The findings of the placebo-controlled trials show a lower reduction in micturitions and incontinence episodes than the head-to-head trials, but are consistent with each other.

Two additional studies were found comparing tolterodine ER to placebo. Change in micturition frequency and incontinence episodes were consistent with previous tolterodine ER versus placebo trials and with results of comparative trials involving tolterodine ER.^{77, 78}

Only one study each was found comparing oral oxybutynin⁶¹ and flavoxate⁶⁰ to placebo. Other studies did not meet the inclusion criteria. The oxybutynin study used a dose of 2.5mg twice daily, and compared median change in daytime incontinence and frequency. Actual data were not reported, but the analysis showed oxybutynin to be better than placebo at reducing daytime frequency (p = 0.0025), but not incontinence. The flavoxate study compared 200mg flavoxate three times daily to placebo. The difference between flavoxate and placebo in the mean change in number of micturitions per day was not statistically significant (−0.292, p = 0.95).

Transdermal oxybutynin delivered at doses of 1.3, 2.6, and 3.9mg per day was compared to placebo for 12 weeks.⁷⁶ Less than half of those enrolled in a 4 week screening process were randomized in this study, and there was a 14% drop-out rate during the trial. The mean change in incontinence episodes per day was −2.7 with 3.9mg TD, vs −2.1 with placebo (p <0.05). The other doses were not significantly different from placebo on primary the outcome measure. The degree of change see in both groups is larger than seen in other placebo-controlled trials, or head-to-head trials.

Three trospium vs. placebo trials were found.^{74, 75, 79} Only one, a 12 week trial, reported change in micturitions per day, −2.4 for trospium, −1.3 for placebo and change in incontinence episodes, −2.3 for trospium and −1.9 for placebo.⁷⁹ In a head to head trial similar results for reduction in micturitions per day were found for trospium at 26 weeks, −2.9.⁴⁵

A very small placebo-controlled trial evaluating scopolamine transdermal patch in 20 women diagnosed with detrusor instability showed greater improvements in diurnal frequency, nocturia, urgency, and urge incontinence with scopolamine than placebo.⁷² It is important to note that this trial was limited to two weeks of treatment only.

One study comparing solifenacin 5mg and 10mg to placebo to assess efficacy and also safety and tolerability demonstrated statistically significant greater reductions in micturitions per 24 hours, number of incontinence episodes, and episodes of urgency in both doses of solifenacin compared to placebo.⁷⁰ This study also showed that approximately 50% of patients from both dosage groups who had at least one incontinence episode at baseline returned to continence by the end of the study. Nocturia reduction was only seen in the 10mg dose.

Three trials comparing darifenacin to placebo were found in the article searches,^{69, 71, 73}, and one trial was identified by the manufacturer. The first study with darifenacin 30mg⁶⁹ examined warning time as defined by the time from the first sensation of urgency to voluntary micturition or incontinence, which was increased under active treatment compared to placebo. Darifenacin 30mg is twice the currently FDA approved maximum daily dose of 15mg per day. Authors note that the 30mg dose used in this trial was chosen during the investigation phase and was assumed to be clinically effective at the time of study design. Secondary efficacy endpoints included clinical monitoring urgency severity, subjective urgency severity recorded at home, and home diary number of urgency episodes. This study was very short with a two-week treatment period, and reported the outcome data as median changes, not mean changes.

The second trial involved a dose titration of 7.5mg to 15mg darifenacin (when requested and tolerated) compared to placebo.⁷³ This study reported a median change of number micturitions/24h and change in incontinence episodes/week for darifenacin and placebo at 12 weeks of −18.9% vs. −10% and −67.3% vs. −42.9%, respectively. An additional darifenacin trial showed similar results at three different doses, 3.75mg, 7mg and 15mg and also reports improvements in terms of medians.⁷¹ The median percent change from baseline of micturition frequency/24h was −13.8%, −16.3%, −15.3% and −8.0% for 3.75mg, 7.5mg, 15mg and placebo respectively and the median percent change from baseline of number of incontinence episodes per week was −58.8%, −67.7%, −72.8% and −55.9% for 3.75mg, 7.5mg, 15mg and placebo respectively.

The final study identified was submitted by the manufacturer.⁸³ This study included darifenacin 15 and 30 mg, tolterodine IR 2 mg twice daily, and placebo arms. A total of 680 patients were enrolled. Only results from the darifenacin and placebo arms were reported. The primary outcome was number of micturitions per week, but median change in micturitions per day is also reported. The median percent changes were −19.1% for darifenacin 15mg, −19.7% for darifenacin 30mg, and −3.2% for placebo. The difference compared to placebo was significant only for the darifenacin 30mg comparison (p = 0.047), but not the darifenacin 15mg comparison (p = 0.075). The fact that all four of these studies used medians to report results instead of means makes it impossible to make indirect comparisons of efficacy and safety for darifenacin.

Long-Term Studies

There are no long-term head-to-head studies designed to assess adverse events of tolterodine, darifenacin, solifenacin, or flavoxate. We found one head to head study⁴⁵ comparing adverse events in trospium and oxybutynin over an average of 54 weeks. This study compared trospium at 20 mg twice daily to oxybutynin IR at 5 mg twice daily. Significant differences were found favoring trospium for adverse events taken as a whole, adverse events having probable or possible connection with trial medications, and for dryness of the mouth. Subjective appraisal of tolerability also favored trospium at 26 and 52 weeks. Overall rates of adverse events were high in both groups.

We found three study of prescription claims data that evaluated the discontinuation rate of new prescriptions for tolterodine or oxybutynin (see Evidence Table 7).¹⁰⁷ One study evaluated the proportion of patients discontinuing treatment (not refilling prescription) in a 6-month period in 1998. Thirty-two percent of patients prescribed tolterodine, compared with 22% on oxybutynin were still refilling their prescriptions at 6 months (p<0.001, this difference remained significant after adjusting for age and co-payment). The mean time to discontinuation was 45 days for oxybutynin and 59 days for tolterodine; 68% on oxybutynin never refilled the original prescription, compared to 55% on tolterodine. While the differences are significant, the numbers apparently discontinuing treatment are high in both groups.

In another study of a pharmacy claims database, patient records were evaluated over a 12-month period following the initial prescription for tolterodine ER, or oxybutynin ER or IR.¹⁰⁸ Inpatients were included. The researchers identified 33,067 patient records for the study, with 50% taking tolterodine ER, and 25 and 26% taking tolterodine ER and IR, respectively. Compliance (based on prescription refills) was not found to be different between tolterodine ER and oxybutynin ER, but oxybutynin IR is stated to be lower (no statistical analysis presented). Persistence rates were low overall, but highest for tolterodine ER (mean 139 days) compared to oxybutynin ER (mean 115 days) or oxybutynin IR (mean 60 days). Statistical analyses of these differences were not conducted. The difference was statistically significant at months 1, 2, 3 and 12 (p< 0.001) for the comparison of tolterodine ER to either formulation of oxybutynin. Differences at other months are presumed to be nonsignificant (data not reported).

The third study used a Medicaid claims database, excluding those eligible for Medicare or in institutions.¹⁰⁹ The researchers identified 1637 patient records for the study. In this study, only 11% were taking tolterodine ER, 13% oxybutynin ER, and 76% oxybutynin IR. Notably, 30% of oxybutynin IR users were under 18 years old. In this study, only 32% on oxybutynin IR, and 44% on either tolterodine or oxybutynin ER continued to take the drugs after 30 days (p<0.001). The 1-year persistence rates were 9%, 6% and 5% for tolterodine ER, oxybutynin ER, and oxybutynin IR, respectively (p=0.086). In a Cox-regression model adjusting for age, sex and race, persistence was not difference between oxybutynin IR and tolterodine ER. In this analysis, oxybutynin ER had a higher risk of nonpersistence after 30 days on drug that tolterodine ER (no difference in the first 30 days). An analysis of risk for "nonpossession" (similar to compliance measures based on days supply provided) indicated no differences between the drugs. Similarly an analysis of switching from the index drug showed "little difference", with 6% switching drug.

We found five open label studies of tolterodine, one 12-week uncontrolled study¹¹⁰ and four 9 to 12 month extension studies following RCTs.^{35, 111–113} Overall adverse event reporting was high (see Evidence Table 7). Dry mouth was the most common adverse event reported, ranging from 13 to 41% of patients. In the short-term study, 8% of these were classified as severe; while in longer-term studies 2–3% reported severe dry mouth. Other adverse events reported included urinary tract infection, headache, and abdominal pain. The longer studies reported 3 to 5 adverse events rated as serious and classified as possibly or probably related to tolterodine. These included urinary retention, worsening of multiple sclerosis, pulmonary edema, tachycardia, hernia, abdominal pain, constipation, and dyspepsia/reflux. Between 8 and 15% of enrolled patients withdrew because of adverse reactions. Two of these studies^{35, 113} reported that dry mouth accounted for only 1–2% of patients withdrawing overall.

An uncontrolled 12-month open-label extension of 4 randomized placebo-controlled trials for tolterodine IR evaluated a total of 714 patients.¹¹² Total number of withdrawals due to adverse events was 105 (15%) with dry mouth reported by 41% of the patients. Dose reduction was offered for those patients with tolerability problems. In a 12-month open-label extension of the previously cited head-to-head comparison of tolterodine ER to oxybutynin IR study, patients were offered tolterodine ER 4mg. The most frequent adverse event in this extension was dry mouth, reported by 33.5% of patients during the 12 months, which is lower than levels (36.8%) found in the 12-week original study. There was a 1% withdrawal rate due to adverse events over the 1ong-term study. It is not clear if the patients in either of these 2 studies overlap patients in previously reported studies (above) that also combine data from patients followed after participating in RCTs.

In addition to these open label prospective studies, we reviewed two uncontrolled studies identifying patients by new tolterodine prescriptions.^{114, 115} One study evaluated adverse events and tolerability over 12 weeks.¹¹⁴ Only 4% of patients reported any adverse event, with dry mouth being the most common (2%). The other study¹¹⁵ identified all new prescriptions for tolterodine in the UK in a six month period, and asked the prescribing general practitioner to retrospectively complete a standard form assessing adverse events at 3 and 9 months. Overall, the physicians reported 3634 events; 13% of these were classified as an "adverse drug reaction (ADR)". Dry mouth was the most common, accounting for 2.9% of all events and 0.5% of all ADRs. Dry mouth was followed by unspecified adverse events, headache or migraine, and UTI. Withdrawals due to adverse events occurred in 4.8% overall, with 1.7% due to dry mouth.

One observational study evaluating implementation of a toileting program that included tolterodine for nursing home residents who did not respond to the non-drug protocol, did not meet our criteria for efficacy, but did report adverse events data.¹¹⁶ This study found that 4% (2 patients) of participating residents had their dosage of tolterodine reduced due to dry mouth (1 patient) and nausea (1 patient). One patient was taken off tolterodine because of increased confusion along with increased back and leg pain.

An open label 12-week study of oxybutynin reported 59% of patients with dry mouth, 23% moderate to severe.¹¹⁷ Similar to the open-label tolterodine studies, withdrawals due to adverse events were 8% overall, 1.6% due to dry mouth.

Solifenacin safety and tolerability was studied in a long-term, 40-week open-label extension study¹¹⁸ that included patients who had completed one of two different trials: one was a placebo-controlled 12-week trial by Cardozo and colleagues⁷⁰ that compared solifenacin 5mg and 10mg to placebo, and the second trial⁴⁶ compared solifenacin 5 mg, solifenacin 10mg, tolterodine 2 mg IR twice daily, and placebo. In the 40-week extension study, 81% of patients who began the study completed all 40 weeks; 4.7% of patients withdrew due to adverse events. Of the patients who completed this study, 20.7% reported dry mouth, 9.6% reported constipation, and 6.9%, blurred vision.

Open-label extension studies are only generalizable to the patient populations included in the trials and to those patients who responded adequately to the drug used in the extension study.

Two poor quality observational studies of tolterodine and oxybutynin are not discussed here.^{119, 120} In addition, one extension study of darifenacin was identified by the manufacturer, but data provided were not adequate for quality assessment of the study.

Short-Term Trials

Adverse events reported in short-term head-to-head trials are summarized in Evidence Table 8. The overall adverse event rate was high in all the studies, ranging from 49 to 97%. The most common adverse event in all studies was dry mouth. Comparisons of the rates of adverse events and dry mouth are presented in Figures 5 and 6. The risk of dry mouth was 28% lower with tolterodine IR than oxybutynin IR (pooled risk difference −0.28, 95% CI −0.34, −0.21). Two of these studies^{32, 33} reported the incidence of severe dry mouth with tolterodine and oxybutynin, 1% vs 5% (NS), and 4% vs 15% (p = 0.01). The other study reported that more patients on oxybutynin than those on tolterodine reported severe dry mouth, but numbers were not reported. One additional study²⁸ assessed dry mouth using a xerostomia questionnaire. They found a significant deterioration on all measures of the scale (except denture fit) for both drugs, with no difference between them.

Figure 5

Risk difference in Overall Adverse Event Rates

Figure 6

Risk difference in Dry Mouth Rates

One short-term trial of trospium versus oxybutynin IR found a higher incidence of severe dry mouth in oxybutynin IR, 23% vs. 4% though overall adverse events were comparable.²⁷ Overall incidence of adverse events was high.

The three studies comparing oxybutynin IR and oxybutynin ER showed differing results, with the two studies using an extended release formulation made by ALZA reporting lower incidence of dry mouth and adverse events with the ER than IR formulation.^{37, 38} These studies also reported a higher incidence of severe dry mouth with the IR formulation, especially as doses increase. Both studies showed a larger difference in moderate to severe dry moth at the 10 and 15 mg levels than at 5 mg per day levels, but at 20 mg/day dose one study³⁷ showed a small difference, and the other³⁸ showed a much larger difference. This second study also allowed 25 and 30 mg/day doses of the ER formulation, which resulted in higher proportions of patients with moderate to severe dry mouth than the lower doses, but similar to each other. The study using an extended release formulation made by a Finnish company reported higher rates of dry mouth, but lower rates of overall adverse events in the ER group³⁶ However, this version of extended release oxybutynin is not currently available in the US. The one study comparing tolterodine ER to IR reported no difference in overall adverse event rates, but a slightly lower rate of dry mouth (risk difference −7% 95% CI −12, −1.6) with the ER form. An additional study comparing oxybutynin IR and oxybutynin ER showed slightly higher withdrawal rates due to adverse events for the immediate release form (20% vs. 17%) and similar numbers of reported dry mouth, however these differences were not statistically significant.⁴⁰ Most other adverse events were reported in greater numbers for the oxybutynin IR, but again differences were not statistically significant.

The study of tolterodine ER vs oxybutynin IR found significantly fewer patients reporting dry mouth with tolterodine ER (33.5%) compared to oxybutynin IR 53.7%, p<0.001.³⁹ Overall adverse events were not reported in a way that could be directly compared.

The study of oxybutynin ER versus tolterodine IR found no difference in overall reports of adverse events, and a non-significant reduction in the proportion with dry mouth.

In the better quality study of the two ER formulations, dry mouth was the most common adverse event noted, and was significantly more frequent in the oxybutynin ER group than the tolterodine ER group (29.7% vs. tolterodine 22.3%, p=0.02).²¹ While not reaching statistical significance, the number of patients with mild or moderate to severe dry mouth were greater in the oxybutynin group.

The other study comparing the ER formulations of tolterodine and oxybutynin used a visual analog scale to assess change in severity.⁴² The authors reported a dose-dependent change for both drugs, but a statistically significant increase only for oxybutynin 10mg once daily compared to tolterodine 4mg once daily (p = 0.03). Other adverse events reported include headache, abdominal pain, constipation, micturition disorders, UTIs, dizziness, somnolence, and vision disturbances. The rate of occurrence of these varied from study to study, and the overall rates of adverse events varied from study to study, reflecting differences in approach to identifying and classifying adverse events.

A small study of oxybutynin TD versus IR found a much higher rate of dry mouth in the IR group (39% vs 82%, p<0.001).⁴³ The rate of dry mouth reported with the IR form in this study was the highest incidence reported in any study. Based on an unvalidated questionnaire, the severity of dry mouth appeared worse in the IR group, but few rated these side effects "intolerable". All patients had been taking oxybutynin IR prior to enrolment, and 67% on TD reported a reduction in dry mouth, compared to 33% on IR. However, overall adverse event rates were not reported. This was a short, 6-week study.

A 12-week study of oxybutynin TD versus tolterodine ER found fewer systemic adverse events among patients in the oxybutynin TD group, including dry mouth, but these did not reach statistical significance.⁴⁴ Application site reactions were reported in 26% of the oxybutynin TD group and 5.7% in the tolterodine-placebo patch group.

In a comparison of varying doses of darifenacin ER and IR and oxybutynin IR, visual nearpoint (a measure of the anticholinergic effect on vision) was not statistically different between the drugs.¹²¹

The STAR trial, which was designed as a non-inferiority trial for its primary outcome measure, compared solifenacin (5mg or 10mg) and tolterodine ER 4mg, reported adverse events data as both solifenacin dosage groups combined compared to the tolterodine group. Because the authors did not do a statistical analysis comparing the rates of the adverse events, we conducted our own statistical analysis. The adverse events that were most commonly reported for both groups were dry mouth (30% for solifenacin, 24% for tolterodine, p<0.05), constipation (6.4% for solifenacin, 2.5% for tolterodine, p = 0.009), and blurred vision (0.7% for solifenacin, 1.7% for tolterodine, NS).⁴⁷ Withdrawals due to adverse events were not significant between groups and were 3.5% for patients receiving solifenacin compared to 3.0% for those receiving tolterodine .

In a trial by Chapple et al.⁴⁶ the comparisons of solifenacin 5mg, solifenacin 10mg, and tolterodine IR to placebo the rates of patients reporting dry mouth were as follows: 14% of the solifenacin 5 mg group, 21.3% of the solifenacin 10mg group, 18.6% of the tolterodine group, and 4.9% of the placebo group.⁴⁶ These differences were not statistically significant by chi square analysis. The reports of constipation were 7.8% for solifenacin 10mg, 7.2% for solifenacin 5mg, 2.6% for tolterodine 4mg, and 1.9% for placebo. The comparisons of tolterodine versus either solifenacin dose are statistically significant, favoring tolterodine (p < 0.05 for both). Similarly, blurred vision was reported by 5.6% of solifenacin 10mg patients, 3.6% of solifenacin 5mg patients 1.5% of tolterodine 4mg patients, and 2.6% of the placebo patients. The comparison of tolterodine and solifenacin 10mg is statistically significant by chi square analysis (p = 0.0115) The percentages of withdrawals due to adverse events were lowest for patients receiving tolterodine (1.9%) followed by solifenacin 10mg (2.6%), solifenacin 5mg (3.2%) and lastly by the placebo patients (3.7%), however none of these differences were statistically significant by chi square analysis.5.

One fair quality systematic review reported adverse event differences between antimuscarinics using only data from short term randomized controlled trials.²⁰ The review concludes that oxybutynin IR (based on one study), tolterodine IR and ER have the most favorable adverse event profiles in placebo-controlled trials. However, the authors also find that oxybutynin IR has the highest rate of adverse events in active control trials. Oxybutynin IR was found to have a greater rate of dry mouth compared with oxybutynin ER, oxybutynin TD, and tolterodine ER and IR in the meta-analysis. Further, they report that there is evidence to show that oxybutynin TD has lower rates of dry mouth and in one study, greater rates of withdrawal due to adverse events (skin reactions at application site) when compared with tolterodine ER.