Overactive bladder (OAB) is a disorder that afflicts 17%1 of women in the United States, making it more prevalent than asthma or diabetes mellitus.2 The prevalence of OAB increases with advancing age.3 Lower urinary tract symptoms associated with OAB include urgency, urinary frequency, nocturia, and urge incontinence.4 These lower urinary tract symptoms have a significant effect on a woman’s quality of life.5 Several anticholinergic medications are available for treatment of lower urinary tract symptoms suggestive of OAB.
Multiple studies have investigated the efficacy and side effects of anticholinergic medications.6–8 Most of these have been clinical trails in which the anticholinergic medication was compared with a placebo or another drug. Clinical trials, although superior in estimating efficacy, are not designed to determine long-term adherence rates to these drugs. Additionally, findings of a clinical trial may not be generalizable to the population as a whole. The purpose of this study was to estimate an overall discontinuation rate in women treated with anticholinergic medications used for the treatment of lower urinary tract symptoms. Our secondary aim was to estimate discontinuation rates among the different classes of anticholinergic medications. Using The Health Improvement Network database gave us a unique opportunity to conduct a population-based study with a large sample size that can answer these important questions.
MATERIALS AND METHODS
The Health Improvement Network9 was established in 2003 as a large, anonymous, electronic medical record database in the United Kingdom. In the United Kingdom, patients’ care is coordinated by general practitioners (GPs) through the National Health Service. When patients are referred for specialty care or nursing care, a consultant initiates a treatment plan, but ultimately chronic therapies are prescribed and monitored by the GP. The Health Improvement Network contains information on diagnoses, prescribed medications, and laboratory values recorded by the GP as part of the patient’s medical record. The GPs are formally trained in data entry, and administrators at The Health Improvement Network cyclically review their data to confirm internal validity for research studies. Most prescriptions given to patients are at 30-day intervals. The refill orders recorded in The Health Improvement Network database also serve as the refill prescriptions given to patients. The cost of the prescriptions is the same regardless of medication type. At the time of our analysis, the database contained more than 4.77 million patient records from more than 300 general practices throughout the United Kingdom. The database represents approximately 4% of the United Kingdom’s population. A number of studies have validated the strength of The Health Improvement Network database.10–12
Our cohort consisted of women 18 years and older who were prescribed anticholinergic medications for the treatment of lower urinary tract symptoms. Women initially prescribed anticholinergic therapy between the period of January 1, 1991, and December 31, 2005, were eligible for inclusion into the study cohort. An episode of drug therapy represented the time from initial prescription to drug discontinuation. For our primary aim, which addressed an overall discontinuation rate of an episode of anticholinergic therapy, discontinuation was defined as when no anticholinergic prescriptions were issued within 90 days after the end of the last anticholinergic drug prescription. Patients were allowed to have multiple episodes of anticholinergic therapy if they restarted. Follow-up time began on the date of the first anticholinergic prescription for the episode.
For our secondary aim, which addressed discontinuation of a specific type of anticholinergic therapy, follow-up time began on the date of first prescription of the particular anticholinergic medication. Specific anticholinergic medications were considered discontinued at the time a patient switched to another medication or when no further prescriptions for the same medication were issued within 90 days of the end of the last prescription. Time to discontinuation was estimated for the following nine anticholinergic medications: tolterodine tartrate, tolterodine tartrate extended-release, oxybutynin, oxybutynin extended-release, flavoxate, terodiline, trospium, propiverine hydrochloride, and solifenacin succinate.
For both aims, therapy switching was defined as a second prescription of an anticholinergic medication from another anticholinergic class within 90 days of the discontinuation date. Patients were censored in both analyses at the time the patient transferred from the GP’s practice, the date of patient death, or the end of the study (February 1, 2007).
The potential confounders examined were age, calendar year of first anticholinergic prescription, number of prior episodes of medication used, number of prior drug classes used, switch to another anticholinergic class, and history of tobacco use. The time interval between the earliest time of diagnosis of lower urinary tract symptoms and treatment (prescription of anticholinergic medication) also was determined. Diagnostic codes used to define lower urinary tract symptoms were nocturia, urgency, urgency of micturition, urge incontinence, urinary symptoms, urinary frequency, and detrussor instability.
For the first aim, the overall discontinuation rate was based on the number of women who discontinued treatment, taking into account the person-month time at risk regardless of switch status. The cumulative incidence of discontinuation was estimated using the Kaplan-Meier method.
For the second aim, drug-specific discontinuation rates were determined. We first plotted the estimated hazard function for each of the nine anticholinergic medications as well as performed a quantitative test of proportional hazards using the Schoenfeld residuals as described in Hosmer and Lemeshow.13 Because the estimated hazard function of the nine medications crossed, the proportional hazard assumption was not satisfied and we were unable to compare the median time to discontinuation between the different drug classes directly. We therefore estimated adjusted survival probabilities using a Cox regression model stratified by anticholinergic drug class. Specifically, we used STATA’s (StataCorp., College Station, TX) stcox estimation command with the vce(robust) option to obtain a robust sandwich variance estimate that adjusts for within-woman correlation stratified by drug class. Cumulative incidence estimates for drug discontinuation then were estimated as one minus the estimated adjusted survival probability. We used the delta method to estimate the variance of our adjusted cumulative incidence measures. All estimates were calculated adjusting for age, number of previous anticholinergic medications prescribed, switch status, number of anticholinergic episodes, year of initiation of anticholinergic treatment, and history of tobacco use. Probabilities presented in this article as cumulative incidence estimates were adjusted to the mean of the potential confounding covariates, and 95% confidence intervals (CIs) were constructed using robust variance estimates to control for the within-woman correlation resulting from multiple observations per woman. Sensitivity analyses using alternative definitions of discontinuation were performed (30 days and 180 days). All analyses were conducted using STATA 10 (StataCorp.).
There were a total of 52,112 eligible episodes of prescribed anticholinergic medications during the cohort period. Of those, 2,693 episodes (5.2%) were excluded because participants were prescribed two anticholinergic medications on the same day for a particular episode. We therefore had 49,419 episodes of anticholinergic drug therapy available for analysis for our primary aim to determine an overall anticholinergic discontinuation rate. The total person-time contribution for this study was 530,498 months from 29,369 women. For the class-specific analysis, where each prescribed drug represented a drug episode regardless of switch status, there were 57,014 episodes.
The mean age of women prescribed anticholinergic medication was 63.9±16.8. Oxybutynin represented the most common initial prescription (51.3%), followed by tolterodine tartrate (20.2%). Only 13.3% of the anticholinergic episodes were started in the month of lower urinary tract symptom diagnosis.
Table 1 displays the overall discontinuation rate of anticholinergic drug therapy. Ninety-one percent of all episodes were discontinued in the study cohort. The median time to overall anticholinergic discontinuation was 4.76 months. Ninety-two percent of episodes of drug therapy were discontinued by 36 months.
The average number of treatment episodes and drug classes prescribed per patient was 1.65±1.31 and 1.54±0.57, respectively. Of the 29,369 women included in the study cohort, 59% tried one episode of anticholinergic treatment, 21% tried two episodes of treatment, 9% tried three episodes of treatment, and 11% tried more than three episodes of treatment. Table 2 shows the characteristics of the 57,014 anticholinergic drug-specific episodes. In the drug-specific analysis, 74% of the episodes represented treatment in women with no prior history of treatment with other drug classes. In 20% of the 57,014 drug-specific episodes, women had a history of prior treatment with one drug class; in 5% of the drug-specific episodes, women had a history of prior treatment with two different drugs. Oxybutynin and tolterodine were the most common drug therapies of the nine drug types, and solifenacin succinate was the least frequent prescription. Solifenacin succinate was the most commonly prescribed medication (19.6%) in women with a history of previous treatment by three or more drug classes. On average, one in five episodes (22.2%) of anticholinergic therapy was in a current smoker. The mean time from lower urinary tract symptom diagnosis to initiation of drug therapy was 28.7 months. The longest time from diagnosis to treatment was with solifenacin succinate therapy (43.1 months), and the least time was with teroldine (6.5 months).
Table 3 describes the time to discontinuation for the nine types of anticholinergic therapy. The median time for drug discontinuation was highest for tropsium and tolterodine (5.47 months each), followed by properivine hydrochloride (5.43 months), extended-release oxybutynin (5.13 months), and solifenacin succinate (5.0 months). The drugs with the shortest median time to discontinuation were teroldine and flavoxate (4 months each).
The adjusted cumulative incidence of discontinuation increased with duration of use for all drug classes. At 6 months, both extended-release oxybutynin (57%, 95% CI 55.1–59.2) and extended-release tolterodine tartrate (54%, 95% CI 52.3–57.4) had lower incidences of discontinuation in comparison with the their multiple-dosing drug classes (71%, 95% CI 68–73.1 and 61%, 95% CI 59–64, respectively). Flavoxate and teroldine had higher rates of discontinuation up to 12 months in comparison with the other seven drug classes. At 24 months of follow-up, the cumulative incidence of discontinuation was greater than 90% for all drug classes and approached 100% by 36 months of use. Solifenacin succinate had the lowest incidence of discontinuation at 6 months in comparison with the remaining eight drug classes.
Overall, the percentage of episodes in which women switched to another anticholinergic medication before discontinuation was 15.8% (95% CI 15.4–16.1). Switching rates declined with increase in number of episodes of drug therapy: 20.9% after one episode, 10.1% after two episodes, and 5.2% after three episodes. For women who used two prior drugs classes, the rate of switching was high, with an average switch rate of 56.2%. The rate of switching was even higher (66.2%) for women with a history of more than two prior drug class uses. When looking at the drug-specific analysis, episodes of treatment with oxybutynin and solifenacin succinate had the lowest rates of switch to another medication (9.8% and 12.3%, respectively, Table 2). Episodes of treatment with properivine hydrochloride and extended-release oxybutynin had the highest rates of switching to another medication (Table 2).
Sensitivity analyses were conducted using an alternative 30-day and 180-day time period to define discontinuation of medication use. The use of a 30-day period increased the overall rate of discontinuation at 9 months to 88.4% (95% CI 88.2–88.6) from 75.9% (95% CI 77.5–76.3). The use of a 180-day period decreased the overall discontinuation rate at 9 months to 61.2% (95% CI 60.7–61.6). When evaluating the anticholinergic class-specific discontinuation rates, the rank order of the nine medications was unchanged with the alternative discontinuation time periods. Sensitivity analysis also was performed, evaluating the difference in drug-specific discontinuation rates in episodes of treatment in those with a lower urinary tract symptom diagnosis compared with those without. Only 32% of the drug-specific episodes had a lower urinary tract symptom diagnosis at the time of therapy. The variation in the adjusted drug-specific cumulative incidence of discontinuation was less than 5% for all nine anticholinergic medications between the two groups.
The most important finding of our study was that the median time to discontinuation for all drugs of anticholinergic therapy was 4.76 months after initial therapy use. Half of women prescribed medication discontinue therapy at 6 months, and three out of four women discontinue therapy by 1 year. Furthermore, the rates of discontinuation increased with duration of use. Very few women switched to another medication after initial prescribed treatment before drug discontinuation. These high discontinuation rates suggest that women have poor adherence to prescribed anticholinergic therapy for lower urinary tract symptoms.
Relatively few studies have assessed adherence to anticholinergic medications used for the treatment of lower urinary tract symptoms. Yu et al14 and Shaya et al15 have reported the discontinuation rates of three medications in two different subsets of U.S. Medicaid populations. Both studies were small (1,500 and 1,637 participants) and examined only three drugs (tolterodine tartrate extended-release, oxybutynin extended-release, oxybutynin). The discontinuation rates reported in those two studies (80–90%) were higher than ours (50–75%) in a similar time period. These dissimilar results could be attributed to differences in the study populations and the size of the study cohorts. Additionally, time-varying covariates were not accounted for in the two studies.
At 6 months, the adjusted cumulative incidence of discontinuation for oxybutynin was 71%, whereas it was 61% for tolterodine tartrate. The adjusted cumulative incidences of discontinuation for extended-release oxybutynin and extended-release tolterodine tartrate were 57% and 54%, respectively. Prior clinical trials have reported a similar trend in lower discontinuation rates for the extended-release drugs in comparison with multidosing formulations. Fewer side effects,7 such as a decreased rate of constipation and dry mouth, also have been reported for tolterodine tartrate as compared with oxybutynin. This may explain the lower discontinuation rates for tolterodine tartrate and extended-release formulations in our study. However, because our data represent analysis of a database, we are unable to identify reasons for discontinuation in our study.
The reported discontinuation rate for extended-release oxybutynin and extended-release tolterodine in clinical trials16,17 is 8–12%, whereas discontinuation rates of 15–20%7 have been reported for oxybutynin and tolterodine. In our study, discontinuation rates for these four drugs were much higher in the 50–70% range at 6 months. The higher discontinuation rates in our study could be attributed to two factors. Participants in clinical trials are not population-based and represent a select group of study participants. These women are likely to be highly motivated and are followed closely by investigators such that medication compliance is high. This could lead to higher drug adherence rates and, hence, lower drug discontinuation rates. Secondly, because clinical trials are investigating specific comparisons, they cannot account for the time-varying factors that may affect drug discontinuation. Our study was population-based, and we were able to adjust for the time-varying covariates that affect drug discontinuation such as the effect of number of prior episodes of anticholinergic therapy, prior number of drug classes used, and therapy switch status.
Terodiline and flavoxate had higher discontinuation rates in comparison with the other drug classes. This result was not surprising because teroldine is no longer prescribed owing to its reported association with heart arrhythmias18; it subsequently was discontinued from the market. Also, a recent meta-analysis19 comparing the efficacy of the different classes of anticholinergic drugs found that flavoxate had reduced efficacy in comparison with other commonly prescribed classes of anticholinergic (oxybutynin, oxybutynin extended-release, tolterodine tartrate, tolterodine tartrate extended-release) medications.
The overall switch rate to another anticholinergic medication in this population was low (15%). For the subset of women who had a history of prior use of two or more anticholinergic medications, the switch rate was high (56–66%). The low overall switch rate could be because women discontinue treatment without attempting additional drug therapy, likely because of inadequate relief of symptoms and/or a high rate of side effects. The small subset of women with high switch rates probably represents women with severe lower urinary tract symptoms. These switch and discontinuation rates highlight the need for more effective treatment for lower urinary tract symptoms.
Surprisingly, in our study cohort, the time interval between the diagnosis of lower urinary tract symptoms as recorded by the diagnosis codes in the electronic medical record and the initial prescription of anticholinergic therapy was 2 years. The significance of this finding is not clear. It is possible that women experienced lower urinary tract symptoms for 2 years before being prescribed anticholinergic treatment. This delay may represent initial trial of nonpharmacologic therapy such as behavioral therapy or dietary and fluid modification. It also may represent an actual delay in pharmacologic therapy until the patient’s symptoms worsened. The cause of the lag between the day the prescription was issued and its start by the patient is unclear.
Further studies in which more detailed information from medical records is available will be needed to delineate the cause for our findings.
There are several strengths to this study. This is a large, population-based study describing anticholinergic discontinuation rates. The large sample size has allowed us to determine drug discontinuation rates across all nine anticholinergic drug classes. Furthermore, the electronic medical record database has allowed us to incorporate the effect of time-varying factors that affect drug-discontinuation rates. The fact that these data represent actual initial and refill prescriptions given to patients improves the generalizability of the results in comparison with those captured from clinical trials.
There are some limitations as well. In the present study, the diagnosis of lower urinary tract symptoms is as recorded in the electronic medical record. We are therefore not certain whether a diagnosis of OAB, as defined by the International Continence Society,4 was confirmed in these women. To the best of our knowledge, the drugs investigated in this study are used only for the treatment of lower urinary tract symptoms suggestive of OAB. This study also does not provide information about the cause for drug discontinuation. It is not known whether reduced efficacy, side effects, or other confounding factors were the cause for drug discontinuation. Lastly, we used a 90-day interval to define drug discontinuation. Although there is no clear consensus on the interval used to define drug discontinuation, many studies have used a 90-day interval.20,21 Some studies have used a 30-day or 180-day22 interval to define drug discontinuation as well. We performed a sensitivity analysis using the alternative definitions of discontinuation and found no difference in the rank order of the nine anticholinergic classes with respect to the incidence of discontinuation.
Clinical trials in general report higher medication-adherence rates in comparison with population-based studies. Large, population-based studies are required to determine the cause for poor adherence to anticholinergic medications in clinical practice. Our high discontinuation rates across all anticholinergic drug classes also highlight the need for more effective therapies for lower urinary tract symptoms. Therefore, we must be vigilant regarding alternative forms of treatment (fluid modification, bladder training, and pelvic floor rehabilitation) for OAB and increase our awareness that this group of women is being treated inadequately.
1. Stewart WF, Van Rooyen JB, Cundiff GW, Abrams P, Herzog AR, Corey R, et al. Prevalence and burden of overactive bladder in the United States. World J Urol 2003;20:327–36.
2. Wein AJ, Rovner ES. Definition and epidemiology of overactive bladder. Urology 2002;60 suppl:7–12.
3. Thomas TM, Plymat KR, Blannin J, Meade TW. Prevalence of urinary incontinence. Br Med J 1980;281:1243–5.
4. Abrams P, Cardozo L, Fall M, Griffiths D, Rosier P, Ulmsten U, et al. The standardization of terminology of lower urinary tract function: report from the Standardisation Sub-committee of the International Continence Society. Neurourol Urodyn 2002;21:167–78.
5. Abrams P, Kelleher CJ, Kerr LA, Rogers RG. Overactive bladder significantly affects quality of life. Am J Manag Care 2000;6:S580–90.
6. Thuroff JW, Bunke B, Ebner A, Faber P, de Geeter P, Hannappel J, et al. Randomized, double-blind, multicenter trial on treatment of frequency, urgency and incontinence related to detrusor hyperactivity: oxybutynin versus propantheline versus placebo. J Urol 1991;145:813–6.
7. Drutz HP, Appell RA, Gleason D, Klimberg I, Radomski S. Clinical efficacy and safety of tolterodine compared to oxybutynin and placebo in patients with overactive bladder. Int Urogynecol J Pelvic Floor Dysfunct 1999;10:283–9.
8. Jonas U, Hofner K, Madersbacher H, Holmdahl T. Efficacy and safety of two doses of tolterodine versus placebo in patients with detrusor overactivity and symptoms of frequency, urge incontinence, and urgency: urodynamic evaluation. The International Study Group [published erratum appears in World J Urol 1997;15:210]. World J Urol 1997;15:144–51.
9. EPIC. EPIC GPRD: a guide for researchers. London (UK): EPIC; 2003.
10. Jick SS, Kaye JA, Vasilakis-Scaramozza Garcia Rodriguez LA, Ruigomez A, Meier CR, et al. Validity of the general practice research database. Pharmacotherapy 2003;23:686–9.
11. Soriano JB, Maier WC, Visick G, Pride NB. Validation of general practitioner-diagnosed COPD in the UK General Practice Research Database. Eur J Epidemiol 2001;17:1075–80.
12. Lewis JD, Schinnar R, Bilker WB, Wang X, Strom BL. Validation studies of the health improvement network (THIN) database for pharmacoepidemiology research. Pharmacoepidemiol Drug Saf 2007;16:393–401.
13. Hosmer DW, Lemeshow S. Applied survival analysis: regression modeling of time to event data. New York (NY): John Wiley & Sons; 1999.
14. Yu YF, Nichol MB, Yu AP, Ahn J. Persistence and adherence of medications for chronic overactive bladder/urinary incontinence in the California Medicaid program. Value Health 2005;8:495–505.
15. Shaya FT, Blume S, Gu A, Zyczynski T, Jumadilova Z. Persistence with overactive bladder pharmacotherapy in a Medicaid population. Am J Manag Care 2005;11:S121–9.
16. Diokno AC, Appell RA, Sand PK, Dmochowski RR, Gburek BM, Klimberg IW, et al. Prospective, randomized, double-blind study of the efficacy and tolerability of the extended-release formulations of oxybutynin and tolterodine for overactive bladder: results of the OPERA trial. Mayo Clin Proc 2003;78:687–95.
17. Appell RA, Sand P, Dmochowski R, Anderson R, Zinner N, Lama D, et al. Prospective randomized controlled trial of extended-release oxybutynin chlorine and tolterodine tartrate in the treatment of overactive bladder: results of the OBJECT study. Mayo Clin Proc 2001;76:358–63.
18. Roden DM. Drug-induced prolongation of the QT interval. N Engl J Med 2004;350:1013–22.
19. Roxburgh C, Cook J, Dublin N. Anticholinergic drugs versus other medications for overactive bladder syndrome in adults. The Cochrane Database of Systematic Reviews 2007, Issue 4. Art. No.: CD003190. DOI: 10.1002/14651858. CD003190.
20. Burke TA, Sturkenboom MC, Lu SE, Wentworth CE, Lin Y, Rhoads GG. Discontinuation of antihypertensive drugs among newly diagnosed hypertensive patients in UK general practice. J Hypertens 2006;24:1193–200.
21. Bourgault C, Senecal M, Brisson M, Marentette MA, Gregoire J-P. Persistence and discontinuation patterns of antihypertensive therapy among newly treated patients: a population-based study. J Hum Hypertens 2005;19:607–13.
22. van Wijk BL, Avorn J, Solomon DH, Klungel OH, Heerdink ER, de Boer A, et al. Rates and determinants of reinitiating antihypertensive therapy after prolonged stoppage: a population-based study. J Hypertens 2007;25:689–97.