In a retrospective analysis of the Optum database, overactive bladder patients treated with mirabegron bladder patients persisted longer with treatment than patients treated with tolterodine, and had reduced resource utilization and costs.
Objectives: To investigate treatment persistence with 2 treatments for overactive bladder (OAB): the β3-adrenoceptor agonist, mirabegron (25 and 50 mg once-daily), and the antimuscarinic, tolterodine extended release (4 mg once-daily).
Study Design: Claims records of patients in the Optum database, with a valid prescription of mirabegron or tolterodine for OAB, were analyzed retrospectively.
Methods: Patients were included in a per-protocol analysis set (PPS) if they had a minimum of 6-months insurance coverage pre- and post index. Patients were classified into treatment-naïve or treatment-experienced cohorts. Kaplan-Meier estimates were used to assess duration of persistence.
Results: A total of 380 patients received mirabegron (195 treatment-naïve, 185 treatment-experienced) and 2248 received tolterodine (1859 treatment-naïve, 389 treatment-experienced). Median persistence was statistically significantly longer for mirabegron (90 days) than tolterodine (30 days) in the PPS. In treatment-experienced patients, persistence was 125 days for mirabegron and 90 days for tolterodine; in treatment-naïve patients, persistence was 80 days for mirabegron and 30 days for tolterodine. Between the 6-month pre- and post index periods, patients receiving mirabegron experienced a median reduction in unique number of prescription medications (from 12 to 10), while those receiving tolterodine experienced an increase (from 8 to 10). Mirabegron outpatient visits were reduced by a median of 1 per patient, and reduction in overall costs was $122 more per patient with mirabegron than tolterodine.
Conclusions: Patients receiving mirabegron during the first year post launch persisted on therapy longer than those receiving tolterodine. Treatment of OAB with mirabegron was associated with a reduction in overall costs and resource utilization.
Am J Pharm Benefits. 2016;8(2):e25-e33
Persistence with antimuscarinics, the current mainstay of overactive bladder (OAB) treatment, is poor. In this retrospective analysis of claims records of patients in the Optum database, patients receiving mirabegron persisted on therapy significantly longer than those receiving tolterodine extended release (ER).
Overactive bladder (OAB) is a chronic condition defined by symptoms of urinary urgency, usually accompanied by frequency and nocturia, with or without urgency incontinence, in the absence of urinary tract infection or other obvious pathology.1 Its prevalence has been estimated at 11.8%.2
OAB is a highly bothersome condition that has detrimental effects on health-related quality of life,3-6 may cause anxiety and depression,7 and is associated with impairment of work productivity and the ability to perform daily activities.5
Antimuscarinic medications have been the mainstay of pharmacological treatment for the symptoms of OAB since the introduction of oxybutynin over 40 years ago. However, while their efficacy8 and positive effects on quality of life outcomes9-11 are well known, they are associated with poor rates of medication persistence12-17 that hinder effective OAB management.
One study reported rates of discontinuation at 6 months of 71% with oxybutynin, 61% with tolterodine tartrate, 57% with oxybutynin extended release (ER), and 54% with tolterodine ER.13 Unmet treatment expectations and adverse events (AEs), notably dry mouth and constipation, have been cited as reasons for treatment discontinuation.18,19
In order to provide optimal management of patients with OAB, clinicians are increasingly required to consider the impact of therapeutic decisions on quality, cost, and value. The availability of effective treatments that patients continue to take for the prescribed duration is essential in providing value-driven care.
Mirabegron is the first in a new class of agents for the treatment of OAB—the β3-adrenoceptor agonists—that has a mechanism of action different from that of antimuscarinic agents. In phase 3 clinical trials with durations of 12 weeks20-22 and 1 year,23 the effectiveness of mirabegron in improving objective outcomes, including number of incontinence episodes and micturitions per 24 hours, as well as a number of quality-of-life outcomes, has been demonstrated.24
The incidence of treatment-emergent AEs seen with mirabegron was low in the clinical trial program; the majority were mild in severity and few were serious.25 In a pooled analysis of the 3 phase 3 trials, the incidence of dry mouth was 5 times lower with mirabegron than with tolterodine ER 4 mg.25
While mirabegron’s efficacy and tolerability profile are well defined in the clinical trial setting, it is not known whether this translates into improved persistence compared with antimuscarinics in a real-world clinical practice setting. A retrospective analysis of claims records of patients with OAB was conducted using a large prescription and medical claims database in the United States (the Optum database).
The patients had claims for mirabegron (25 and 50 mg once daily) or tolterodine ER (4 mg once daily). The objectives were (a) to compare the time to end of treatment persistence, (b) to investigate baseline demographic and clinical characteristics that may be predictive of increased persistence, and (c) to compare time to end of persistence in treatment-naïve and treatment-experienced patients.
Tolterodine ER was selected as the comparator due to its common use as a first-line treatment for OAB, comparable formulary access to mirabegron across the vast majority of insurance plans within the Optum database, and its history as an active control in the mirabegron clinical trial program.
Optum is a large (114 million fully-insured patient lives covered), nationally representative, commercial prescription and medical claims database in the United States, composed of enrollment information for administrative-services-only lives, Medicaid lives, and Medicare supplemental insured lives from 1993 to the present. It contains hospitalization, medical, pharmacy, and laboratory claims that can be linked anonymously to patient identification numbers.
Study Design and Sample Selection
Patients were included if at any time between January 1, 2010 and September 30, 2013, they had (a) been diagnosed with OAB, according to the International Classification of Diseases, Ninth Revision, Clinical Modification codes summarized in Table 1, (b) had prescription claims for either mirabegron or tolterodine ER (defined as index drugs), and (c) had an insurance coverage eligibility period of at least 6 months pre- and post index (defined as time of first claim for index drug).
The study period was based on the most recently available Optum dataset that coincided with the first year post US launch of mirabegron. Eligible patients were classified into treatment-naïve or treatment-experienced cohorts (no claims record or at least 1 claims records for any other OAB drug in the 6 months prior to first index drug, respectively) for a minimum 6-month look-back period within the database. No patient’s identity or medical records were disclosed for the purpose of this study, except in compliance with applicable law.
The primary endpoint was time to end of persistence, defined as a continuous supply of index drug until any 30-day period during which the patient did not have a supply of index drug. This was determined by summing all the days of supply of index drug.
Overlapping prescriptions of the same medication were assumed to be early refills and analyzed on the assumption that they had been taken sequentially.26 Gaps between prescriptions—number of days between the end of 1 prescription supply and the start of the next—were calculated after application of the shifting method proposed by Leslie (2008).26
Adherence—the proportion of days covered by the prescription—was calculated using prescription fill dates and number of days’ supply for each fill of a prescription. The calculation was interval-based with a fixed post index follow-up period of 6 months.
Baseline demographics summarized were age (continuous and categorical [18-45 years; 46-64 years; and ≥65 years]), region, and insurance plan description. Patient comorbidities for the pre-index period and baseline clinical characteristics were summarized using Charlson comorbidity index (CCI) scores as estimated via claims.
Polypharmacy is often seen as an effective predictor of utilization and healthcare charges over 1 year,27 and was assessed by summarizing the number of unique prescriptions for which claims appeared in the database from any provider from 12 months before until 12 months after the date of the first index drug. OAB-related resource use and polypharmacy, as well as prior and subsequent therapies, were summarized.
Overall costs were calculated for any inpatient or outpatient visits or procedures and prescriptions which began after the index date and for which claims were paid prior to the end of the 6-month follow-up period. Each of the cost centers (inpatient, outpatient, procedures, prescriptions) was also calculated individually within each cohort to identify those costs that were responsible for any meaningful differences in the overall cost. Cost comparisons between the mirabegron and tolterodine cohorts were summarized via descriptive statistics. Pricing is in US dollars adjusted to 2013 values.
The rate of patient persistence with their index drug was estimated using Kaplan-Meier methods for the entire post index period, from first prescription fill until disenrollment from the health plan. If a patient was persistent until the end of the latest available Optum dataset, they were considered persistent until September 30, 2013. Kaplan-Meier plots were stratified by cohort groups, and by patient clinical and demographic factors including sex, and were used to estimate the median time to end of persistence (with 2-sided 95% CIs).
Unstratified and stratified Cox proportional hazard models (stratified by age, sex, and naïve/experienced, with and without comorbidity status and year of first index drug) were fitted to compare persistence with mirabegron and tolterodine. Hazard ratios for all factors in the univariate models were reported.
Significance was evaluated at an α of 0.05. Groups were analyzed retrospectively in the per-protocol set (PPS; patients who meet the study inclusion criteria) and intent-to-treat (ITT; all patients who used the index drug; inclusion criteria not applied) analysis set (the PPS was the primary analysis population).
Study Sample Description
A total of 380 patients received mirabegron (195 treatment-naïve and 185 treatment-experienced) and 2248 received tolterodine (1859 treatment-naïve and 389 treatment-experienced; Table 2). The mean age of patients across the 4 cohorts ranged from 57 to 59 years. The proportion of patients aged <65 years was lowest in the treatment-experienced tolterodine cohort (62.2%; Table 2). Of the patients who received tolterodine, 79.6% were female compared with 62.1% of those who received mirabegron, which was a statistically significant difference.
Additional baseline parameters that were significantly different between the total mirabegron and tolterodine cohorts included each of the health plan regions and insurance plans, and pre-index polypharmacy. Mean CCI score was the same for both treatment-naïve and treatment-experienced patients receiving mirabegron (1.5 in both groups) and was slightly higher than for patients receiving tolterodine (1.2 and 1.3 in the treatment-naïve and treatment-experienced cohorts, respectively).
Median persistence was statistically significantly longer for mirabegron (90 days) than tolterodine (30 days) in the PPS population (P = .0002 stratified, P <.0001 unstratified; Figure A and Table 3). The percentage of patients who remained treatment-persistent was greater with mirabegron than tolterodine at 30, 90, and 180 days post index (Table 4).
Median persistence was also longer with mirabegron than tolterodine in the ITT population (170 days and 90 days, respectively; Figure A). A similar trend was seen in all subgroups: In treatment-experienced patients, persistence was 125 days with mirabegron and 90 days with tolterodine; and in treatment-naïve patients, persistence was 80 days with mirabegron and 30 days with tolterodine (Figure B).
Age and treatment history had significant effects on treatment persistence with both mirabegron and tolterodine (Figure B and Table 3), while sex did not (Figure C). Thus, patients aged ≥65 years and treatment-experienced patients had significantly better treatment persistence with their index drug compared with those aged <65 years and treatment-naïve patients, respectively.
Persistence in patients receiving mirabegron was not affected by total weighted CCI scores or the absence of comorbidity, but both variables had a significant effect on persistence with tolterodine (Table 3).
Adherence, as measured by median proportion of days covered by index drug, was higher with mirabegron than tolterodine, and was higher among treatment-experienced than treatment-naïve patients receiving either drug (Table 5).
Pre-index, polypharmacy (number of unique prescription medications) was higher among patients receiving mirabegron (median of 12 drugs) than tolterodine (median of 8 drugs). However, post index, patients receiving mirabegron experienced a reduction in median number of prescription medications (from 12 to 10), while those receiving tolterodine experienced an increase (from 8 to 10).
The median number of outpatient visits between the 6-month pre- and post-index periods was reduced by 1 in patients receiving mirabegron and remained unchanged in patients receiving tolterodine. A concomitant reduction in total median costs was observed with both mirabegron and tolterodine between the pre- and post index periods; this reduction was primarily the result of the reduction in outpatient visits and was $122 per patient greater with mirabegron than tolterodine.
Medication persistence refers to the act of continuing treatment for the prescribed duration and may be defined as the duration of time from initiation to discontinuation of therapy.28 Persistence with antimuscarinics, the current mainstay of OAB treatment, is poor, leaving many patients without pharmacological treatment options.12-17
It is widely accepted that there is a need for new therapies and strategies to increase treatment persistence and improve outcomes in OAB. This study represents the first examination of persistence with mirabegron in a real-world setting. After examining multiple patient characteristics, data from the Optum database demonstrated that patients prescribed mirabegron persisted on therapy significantly longer than those to whom tolterodine ER was prescribed.
It is notable that the median time to end of persistence with tolterodine ER in this study (30 days in treatment-naïve patients in the PPS population) is similar to the values of 33.0 and 34.0 median days reported previously with tolterodine ER and tolterodine immediate release, respectively.29
The rate of treatment discontinuation at 6 months was approximately 80% versus 65% in the tolterodine and mirabegron groups, respectively. The rate reported with tolterodine is relatively high compared with a previous retrospective database study, which reported rates of 54% to 61% in a female population.13
However, it is important to note the wide discrepancy in persistence rates reported with antimuscarinic agents in medical claims studies (43% to 83% within the first 30 days and rates continuing to rise over time).30 In addition, on a population level, treatment of OAB with mirabegron was associated with a reduction in resource utilization as reflected in fewer outpatient visits and reduction in number of concomitant medications) and a concomitant reduction in total costs that was $122 per patient greater than the reduction seen with tolterodine.
While these results suggest that mirabegron may reduce healthcare resource use, the fact that pre- and post index CCI score was similar in both cohorts highlights the difficulty in assuming a direct causal relationship between mirabegron and healthcare resource utilization. The difference in resource utilization between the 2 cohorts was largely driven by fewer outpatient office visits in the mirabegron cohort.
It is possible that comorbidity status was not sensitive enough to differentiate OAB medication use between the 2 cohorts, but that outpatient visits and polypharmacy correlated well. Polypharmacy is often used as an indicator of utilization and healthcare charges over time, and it is possible that a reduction in concomitant medications unrelated to the treatment of OAB could have also contributed to the decrease in healthcare resource utilization in the mirabegron cohort.
Age and previous treatment were the only patient characteristics that affected persistence, with treatment-experienced patients and those aged ≥65 years demonstrating significantly better persistence rates than treatment-naïve patients and those aged <65 years, respectively, both in patients treated with mirabegron and tolterodine. It may be that persistence is greater in treatment-experienced than treatment-naïve patients because the latter are less inclined to tolerate the side effects associated with treatment during their first prescription.
It has been suggested that older patients may be more likely to respond to OAB medication than younger patients31 and that this might result in better adherence. Although patients in the tolterodine cohort were on average 1.6 years younger than the mirabegron patients, the proportion of patients aged <65 years was identical in both cohorts, which suggests the improved persistence observed in the mirabegron cohort is independent of age.
The greater tendency of older patients to adhere to treatment has been observed previously; D’Souza and colleagues (2008)29 reported 50% better adherence among patients aged ≥65 years than for those aged <65 years. In a retrospective analysis of United Kingdom prescription data, Wagg and colleagues (2012)17 found that patients aged ≥60 years were more likely to persist with prescribed therapy over a 12-month period than those aged <60 years.
However, this is not necessarily the consequence of improved efficacy in these patients. Indeed, a recent subgroup analysis of pooled data from the 3, 12-week phase 3 trials of mirabegron found that efficacy in 2 subgroups of older patients (≥65 years and ≥75 years) was similar to that in the younger subgroups.32
Instead, these results may be reflective of the improved tolerability profile of mirabegron compared with antimuscarinics, which allows patients to remain persistent for longer. The early experience with mirabegron demonstrated within this study may be associated with greater effectiveness in a real-world setting compared with antimuscarinics; however, this needs to be confirmed by larger studies with greater longitudinal surveillance, as patient and prescriber experience with mirabegron expands.
This study has a number of limitations. First, as it was not prospective or randomized, drug selection biases were likely. The post prescription observation period was short and the number of patients in the mirabegron group was small. There is no information regarding the rationale for drug selection by any given health provider, and the 2 drugs may have been targeted to different points in the sequence of therapy or to different patient populations.
Sociodemographic information known to affect persistence or OAB symptoms, such as educational status and alcohol use, were also not available in the Optum database. The large discrepancy in sample size between the mirabegron and tolterodine cohorts may have contributed to the statistically significant differences for some demographic and/or baseline characteristics, although many of these do not represent clinically important differences.
Additionally, no information is available about the clinical rationale employed by prescribers when making decisions to initiate or discontinue treatment; hence, there is no way to identify clinically justified discontinuations. Likewise, we have no information on why patients chose to discontinue their prescriptions—eg, lack of perceived efficacy, intolerance, cost, or some other reason.
Prescription refills by other routes besides pharmacy refills were not capable of being captured, while the filling of a prescription does not necessarily indicate that the medication was taken. There is no data on concurrent nonprescription pharmaceutical interventions that may have impacted persistence with the index drug, and the results are generalizable only to patients with commercial insurance. In addition, Optum claims were not adjudicated against actual medical records, a limitation inherent in all claims database analyses.
While the models controlled for differences between the mirabegron and tolterodine cohorts in such characteristics as age, sex, comorbidities, naïve versus experienced, and year of study entry, the statistical estimates adjusted for these individual patient characteristics on within- and between-drug differences may still contain some degree of bias.
Finally, it should also be noted that the results of this study, from the early wave of patients treated with this new therapy, may not be indicative of persistence patterns once patients and physicians have gained more experience with this agent in the market. Despite these inherent limitations with all medical claims database analyses, the Optum database is a well-recognized and validated method for investigating disease prevalence and medication persistence rates in approximately one-third of the US population.
Mirabegron was associated with statistically significantly longer treatment persistence, and a reduction in overall costs and resource utilization compared with the antimuscarinic agent, tolterodine ER, over a 6-month period. This may be the consequence of mirabegron’s improved tolerability profile compared with the antimuscarinics, or the consequence of the selection bias that is one of the study limitations. The size, diversity, and validity of the Optum database suggest that the results of this study can be generalized to the adult population in the United States.
Editorial assistance was funded by Astellas Pharma Inc and provided by Aideen Young, PhD, and Stuart Murray, MSc, of Envision Scientific Solutions.
Author Affiliations: Department of Urology, New York University Langone Medical Center (VWN), New York, NY; Department of Urology, Medical University of South Carolina (ESR), Charleston, SC; Astellas Scientific and Medical Affairs (BF, GNM, TB), Northbrook, IL; Astellas Pharma Global Development, Northbrook, IL (DBN, AF).
Source of Funding: This study was funded by Astellas Pharma Inc.
Author Disclosures: The authors report the following: investigator for Astellas (VWN); investigator/advisory board member for Astellas (ESR); investigator for Allergan, Cook MyoSite (VWN); fulltime employee at Astellas (BF, GNM, TB, AF, DBN).
Authorship information: Concept and design (VWN, ESR, BF, TB, AF, DBN); analysis and interpretation of data (VWN, ESR, BF, GNM, TB, AF, DBN); drafting of the manuscript (VWN, ESR, BF, GNM, TB); critical revision of the manuscript for important intellectual content (VWN, ESR, BF, GNM, TB, AF, DBN); approval of the manuscript for publication (VWN, ESR, BF, GNM, TB, AF, DBN); statistical analysis (BF, AF).
Address correspondence to: Daniel Ng, Astellas Pharma Global Development, Medical Affairs, Americas, 1 Astellas Way, Northbrook, IL 60062. E-mail: email@example.com.