SEARCH

SEARCH BY CITATION

Keywords:

  • composite endpoint;
  • response;
  • King’s Health Questionnaire;
  • quality of life;
  • overactive bladder syndrome;
  • solifenacin

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

OBJECTIVE

To evaluate three potential definitions of treatment response in overactive bladder (OAB) syndrome, based on the International Continence Society (ICS) definition of OAB, using pooled data from two large clinical trials of the antimuscarinic agent solifenacin succinate.

PATIENTS AND METHODS

As the ICS definition of OAB does not rely on urodynamic observations for a diagnosis, the use of composite endpoints, such as in the definitions examined here, might provide a more relevant measure of patient benefits. Each definition was tested to assess the impact of pharmacotherapy, comparing treatment response rates based on composite endpoints. Responder definitions were validated based on the impact on quality of life (QoL). Pooled data from two 12-week multinational, multicentre, double-blind randomized studies of solifenacin, a once-daily oral antimuscarinic for treating OAB symptoms, were used for this analysis. Based on the symptom data from these studies (episodes of urgency, incontinence, nocturia and voiding frequency per 24 h), response was defined as being: (i) a reduction by half or more in all baseline symptoms; (ii) a reduction by half or more in urgency and at least one other symptom; or (iii) resolution of urgency episodes and at least one other symptom. Health-related QoL changes were recorded using the King’s Health Questionnaire (KHQ).

RESULTS

Solifenacin at 5 mg or 10 mg once daily resulted in significantly higher response rates than with placebo, irrespective of the definition of a responder (P ≤ 0.001). When assessing the most challenging definition of response (resolution of urgency and at least one other symptom), solifenacin at 5 mg and 10 mg resulted in 20% and 25% of patients, respectively, meeting the response criteria, vs 12% of placebo-treated patients (P ≤ 0.001 for both solifenacin doses). For all three definitions of response, responders showed greater improvements in all KHQ domains than did nonresponders, and the differences were statistically significant.

CONCLUSIONS

The two strictest definitions of response (halving or more in all baseline symptoms or resolution of urgency and more than one other symptom) appeared to capture QoL changes most effectively. Establishing the most appropriate definition to assess treatment-related changes in OAB will require further investigation, but it is clear from this analysis that composite clinical endpoints, based on the ICS definition of OAB, are a valid approach for assessing the effect of treatments for OAB.


Abbreviations
KHQ

King’s Health Questionnaire

OAB

overactive bladder syndrome

QoL

quality of life.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Overactive bladder (OAB) syndrome is a highly prevalent medical condition that is estimated to affect >17 million men and women in the USA [1] and 50–100 million worldwide [2]. One of the underlying limitations in obtaining accurate estimates of the prevalence of OAB has historically been a lack of consensus over what symptoms or evaluations should be used to define this syndrome [3]. To address this issue, raise awareness of the condition and promote standard approaches to treatment, the ICS included a new definition of OAB [4]. The 2002 ICS Terminology Committee report states that, ‘urgency, with or without urge incontinence, usually with frequency and nocturia, can be described as the overactive bladder syndrome’. These terms can be used if there is no confirmed infection or other obvious pathology [5].

The ICS definition highlights three key points. First, the diagnosis of OAB does not rely on urodynamic evaluations, but arises solely from patient symptoms. This allows and encourages a wider range of clinicians, beyond specialists, to identify patients suitable for treatment [4]. Second, urgency is recognized as the fundamental symptom of OAB. While the ICS definition facilitates communication and basic patient care, it does not address clinical research, nor does it provide a direct link to clinical trial entry criteria or outcome measures. There remain many different ways that treatment results are reported, with no apparent universally accepted definition of ‘response’[6–9]. To support research and patient care, trials should report results using common endpoints based on accepted terminology. We contend therefore that as urgency is at the heart of the OAB syndrome, reducing urgency should be a primary target for assessing the effectiveness of any treatment for this condition. It is unfortunate that trials do not normally report reductions in urgency as a primary outcome variable, mainly because there is no commonly agreed method for evaluating this key symptom [10]. Third, the ICS definition highlights that OAB is a syndrome, with several symptoms together determining the severity and state of this condition. However, clinical trials do not consider this definition and typically report single-outcome variables, when a composite endpoint including several key symptoms might more accurately reflect the patient’s response to treatment.

It is also important that outcome measures reflect quality of life (QoL), as OAB profoundly compromises QoL, including social, physical, psychological, occupational and sexual aspects [11]. These negative pressures on QoL can result in the development of complex coping mechanisms, such as ‘toilet mapping’ and defensive voiding, as well as self-imposition of limitations on daily activities [11]. Ultimately, patients are treated because their symptoms affect their QoL and it is logical to conclude that effective therapy, if measured by relevant clinical endpoints, should correlate with improvements in patient QoL.

Based on these three principles, we retrospectively examined three potential new definitions of treatment response that accommodate the ICS definition of OAB, using data generated in two large-scale clinical trials of the antimuscarinic agent solifenacin succinate. We tested each definition for its ability to assess the effect of pharmacological treatment, comparing response rates based on these composite endpoints to the effect on QoL. The ultimate goal of this analysis was to determine the definition of a responder that most accurately reflects changes in QoL.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Pooled data were analysed from two 12-week multinational, multicentre, double-blind randomized studies of solifenacin, the primary results and methods of which were published previously [12,13]. Both studies were conducted with similar protocols, thus enabling pooling of the data for this analysis. After a 2-week placebo run-in period, patients were randomized to either 5 mg solifenacin once daily, 10 mg solifenacin once daily, placebo once daily or tolterodine 2 mg twice daily (in one study only, as an active treatment arm) for 12 weeks. OAB symptom data were collected from 3-day voiding diaries. All patients had above normal levels of voiding frequency (>8 episodes/day; mean number of episodes: solifenacin 5 mg group, 12.1; solifenacin 10 mg group, 12.2), 99% reported an urgency episode, 58% reported an incontinence episode and 88% had an episode of nocturia, during the 3-day baseline diary assessment. Eligible subjects either had pure OAB, or mixed incontinence with a predominant urge component. The presence of stress incontinence was assessed by history, data from the voiding diaries, and a cough-provocation test; if the investigator determined that stress incontinence was the predominant symptom, the patient was excluded.

Data relating to baseline-to-endpoint changes in urgency, incontinence, nocturia and voiding frequency/24 h were collected in the two studies using the voiding diaries. Based on these symptoms, treatment response was defined in three ways: (i) A reduction by half or more in every baseline symptom; (ii) A reduction by half or more in urgency episodes and at least one other symptom; (iii) resolution of urgency episodes and at least one other symptom.

Episodes of urgency, incontinence and nocturia were reported as the number of episodes per day. All data are presented as the mean of data collected over 3 days using the voiding diary. Percentage reductions were used to assess the response ((endpoint mean − baseline mean)/baseline mean × 100). Incontinence episodes included all episodes of incontinence, not just those associated with urgency. All of the proposed definitions require reduction in urgency as a prerequisite for response, in line with the ICS definition of OAB, which places urgency as the defining symptom of OAB.

The analysis was not restricted to patients who had all four symptoms at baseline, as not all four symptoms are required to be present for entry into the clinical trials or for a diagnosis of OAB. However, improvement was only registered in those symptoms that were present at baseline. Normalization of voiding frequency was considered as those patients achieving ≤ 8 voids/24 h. A 50% improvement was defined as a reduction by half in the number of voids above eight that the patients had at baseline. Thus, a patient with 12 voids/24 h at baseline who had 10 voids/24 h at endpoint would be considered to have a 50% improvement. Resolution of all other symptoms was defined as the presence of at least one episode during baseline but no episodes at endpoint, as recorded by the patient diary.

QoL data in the two studies were obtained using the King’s Health Questionnaire (KHQ) [14]. This short, sensitive, disease-specific instrument uses patient-reported outcomes and contains 21 questions distributed among nine domains, plus a 10th domain of 11 questions designed to identify individual bladder problems and their effects on the patient. In the KHQ, each domain is represented individually with no overall score. Each domain is marked from 0 to 100 points, with the highest score denoting the lowest QoL. As such, reductions in score denote an improvement in QoL. Patients were supplied with the questionnaire at the start of treatment, after 4 weeks of treatment and at the end-of-study visit. Questionnaires were completed by patients independently of the investigators, to avoid bias.

Between-treatment comparison of the proportion of patients considered to receive effective treatment (i.e. responders), as per the three definitions of response, was formally assessed using Fisher’s exact test. Within-treatment differences in baseline to endpoint QoL changes for each of the 10 domains of the KHQ were compared for responders vs nonresponders, for each of the three definitions of response in turn. For these QoL comparisons, P values were based on an analysis of covariance model with centre, treatment, responder indicator and treatment-by-responder indicator as terms, and baseline as a covariate.

The number of patients and age range in each treatment group was: solifenacin 5 mg, 552, 19–85 years; solifenacin 10 mg, 554 18–84 years; placebo, 534, 18–82 years; tolterodine 4 mg, 250, 19–79 years. The patient demographics and clinical characteristics were similar at baseline.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Solifenacin treatment at either 5 mg or 10 mg resulted in higher response rates than with placebo, for all three proposed definitions of response (Table 1). Irrespective of the definition used, the between-treatment differences in response rates were statistically significant compared with placebo for both solifenacin dose groups (P ≤ 0.001). When assessing the most challenging composite endpoint (complete resolution of urgency and at least one other symptom) solifenacin at 5 mg and 10 mg resulted in 20% and 25% of patients, respectively, meeting the endpoint response criteria, compared with 12% of patients receiving placebo (P ≤ 0.001 for both 5 mg and 10 mg solifenacin vs placebo; Table 1).

Table 1. Response rates, and mean changes in the KHQ domains in responders and nonresponders receiving 5 mg solifenacin, for each of the three composite endpoint definitions of treatment response, i.e. 1) a reduction by half or more in every baseline symptom; 2) a reduction by half or more in urgency and at least one other symptom; and 3) the resolution of urgency and at least one other symptom
VariableDefinition
123
  • P = 0.001 or

  • <0.001 vs placebo for rates, Fisher’s exact test of benefit vs no benefit for the treatment group vs placebo. For KHQ, responders vs non-responders

  • *

    P ≤ 0.01,

  • §

    P ≤ 0.001.

Response rate, n (%)
Placebo 89 (17)208 (39) 66 (12)
Solifenacin, mg daily
 5147 (27)292 (53)108 (20)
 10165 (30)321 (58)136 (25)
Mean change (baseline; endpoint value) in KHQ domain
General health perception
 responder −9.2 (34.9; 25.7)§ −6.2 (36.4; 30.2)§ −6.9 (36.5; 29.6)§
 non-responder −2.4 (37.9; 35.5) −2.2 (37.9; 35.7) −3.7 (37.2; 33.5)
Incontinence impact
 responder−30.6 (69.9; 39.3)§−28.9 (73.3; 44.4)§−31.7 (69.9; 38.2)§
 non-responder−22.5 (79.9; 57.4)−20 (81.4; 61.4)−23.0 (78.8; 55.8)
Role limitations
 responder−28.2 (55.6; 27.4)§−26.0 (58.1; 32.1)§−30.4 (55.9; 25.5)§
 non-responder−17.6 (61.6; 44.0)−14.4 (62.1; 47.7)−18.2 (61.0; 42.8)
Physical limitations
 responder−27.4 (55.4; 28.0)§−22.8 (55.5; 32.7)§−26.3 (54.7; 28.4)§
 non-responder−13.9 (59.0; 45.1)−11.8 (60.9; 49.1)−15.6 (58.8; 43.2)
Social limitations
 responder−19.8 (31.7; 11.9)§−15.7 (32.7; 17.0)§−16.2 (31.2; 15.0)*
 non-responder −8 (34.2; 26.2) −6.2 (34.5; 28.3)−10.1 (34.1; 24.0)
Personal relationships
 responder−16.5 (29.7; 13.2)§−13.6 (32.1; 18.5)*−17.4 (35.6; 18.2)*
 non-responder −5.5 (31.5; 26.0) −3.5 (29.7; 26.2) −6.7 (29.9; 23.2)
Emotions
 responder−22.5 (43.5; 21.0)§−20.9 (45.5; 24.6)§−24.5 (44.3; 19.8)§
 non-responder−13.5 (47.9; 34.4)−10.4 (48.1; 37.7)−13.9 (47.3; 33.4)
Sleep/energy
 responder−21.2 (47.2; 26.0)§−18.2 (50.2; 32.0)§−20.0 (50.2; 30.2)§
 non-responder−10.9 (53.5; 42.6) −8.7 (53.6; 44.9)−12.2 (52.1; 39.9)
Severity measures
 responder−16.0 (40.1; 24.1)§−13.8 (38.9; 25.1)§−15.4 (34.4; 19.0)§
 non-responder −8.4 (40.3; 31.9) −6.8 (41.7; 34.9) −9.3 (41.7; 32.4)
Symptom severity
 responder −4.4 (10.2; 5.8)§ −4.3 (10.7; 6.4)§ −5.1 (9.9; 4.8)§
 non-responder −3.0 (11.9; 8.9) −2.4 (12.2; 9.8) −3.0 (11.8; 8.8)

For the health-related QoL in responders vs nonresponders, a previously published evaluation of the total population from which this secondary analysis was derived showed statistically significantly greater baseline-to-endpoint improvements in both 5 mg and 10 mg solifenacin groups vs placebo in nine of the 10 domains of the KHQ [15]. In the current analysis, irrespective of which treatment patients received (5 mg or 10 mg solifenacin or placebo), those considered as responders according to the predefined composite endpoints had significantly better KHQ scores in most domains. As all treatment groups had similar clinical endpoints and QoL changes, to simplify the analysis and testing of the composite endpoints, only data from those patients receiving 5 mg solifenacin are presented as representative of QoL changes in responders and nonresponders, measured using the KHQ (Table 1).

The results of this analysis show that, irrespective of the responder definition used, patients who were classed as having responded to treatment had greater improvements in all of the KHQ domains than nonresponders. These within-treatment differences between responders and nonresponders were statistically significant for all domains of the KHQ.

The differences between scores for responders vs nonresponders were statistically significant for all 30 patient groups (P < 0.05) as divided by treatment and responder definition (Table 1). One of the challenges of health-related QoL research is to translate statistically significant health-related QoL changes into interpretable clinical or medically important ones. To establish what can be considered a clinically meaningful change, the developers of the KHQ tested their questionnaire using patients with OAB being treated pharmacologically [16]. This was achieved using an anchor-based approach with both an overall rating of patient-perceived treatment benefit and one of perceived disease impact [16]. From this analysis the minimum important difference for the KHQ was calculated as a 5% change (five or more points) with any improvement above this level representing a real improvement to patients. The authors also suggested that a change of 12–15 points represented a moderately clinically meaningful difference for all domains except symptom severity; this domain only requires a difference of ≥ 2 points to be considered minimally clinically meaningful, and 6–8 points to be considered moderately meaningful [16]. Based on these recommended boundaries of clinical relevance, all of the responders in the current study achieved at least a minimal clinically meaningful improvement in all of the KHQ domains (Fig. 1a,b). Furthermore, responders had a moderate clinical improvement in seven of the 10 KHQ domains (Fig. 1b).

image

Figure 1. Changes in: a, the first five and b, the last five KHQ domains for patients receiving 5 mg solifenacin classed as responders, using the three composite endpoints. Hash markings, a reduction in every baseline symptom by half or more; full circle, a reduction in urgency by half or more and at least one other symptoms; open circle, resolution of urgency and at least one other symptom.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Traditionally, OAB clinical trial outcome data have reported individual symptoms in isolation, with urinary frequency and incontinence episodes being the primary outcome measures. However, this approach might not optimally portray the true therapeutic outcome, as it does not properly reflect the primary symptom and the accepted definition of OAB, nor does it reflect what matters most to patients, i.e. improvements in daily QoL and treatment benefit from the patient’s perspective. The diagnosis of OAB is founded on the presence of urgency, with or without urge incontinence, increased voiding frequency and nocturia. The definition of OAB clearly links these symptoms together, with urgency recognized as the pivotal symptom present in all sufferers. Based on this, the primary clinical outcome measure must first reflect improvements in urinary urgency, preferably linked to changes in the other symptoms of OAB; such measurements might be most closely reflected by a composite endpoint that considers changes in more than one symptom. The dataset used for the current analysis is taken from two large clinical trials of solifenacin, the pooled data from which acted as the source for the current analysis and for previous reporting of single symptom improvements [17,18]. This dataset provides a good starting point from which to assess the suitability of the composite endpoints, and their ability to effectively capture treatment-induced changes.

There are at least three perspectives to assessing the outcome of treatments: those of the patient, those of the physician, and those derived by objective measurements. The ideal outcome should consider all of these perspectives and should recognize that there will be differences among them. Ultimately, patient satisfaction and improved QoL are the definition of success when treating benign conditions such as OAB. Using changes in QoL to validate clinical efficacy measures reflects current views in the field of urology, with current emphasis on analysing both objective measures and patient perspective, as used in the treatment of stress urinary incontinence [19,20]. In the current study, those patients classified as responders according to the composite endpoints used in this analysis had moderately clinically meaningful QoL improvements in most domains of the KHQ.

While all three composite endpoints correlated well with improvements in patient QoL, the two stricter definitions appeared to capture QoL changes most effectively: (i) a reduction by half or more in every baseline symptom; and (ii) resolution of urgency episodes and at least one other symptom (Fig. 1). These two composite endpoints appear to be equivalent at capturing improvements in half of the KHQ domains each, and can thus be considered the most clinically relevant of the three definitions of response analysed here. These two composite endpoints are strongly associated with improved QoL, supporting the conceptual validity of this new approach to OAB outcome measures. Interestingly, complete resolution of urgency was the most difficult endpoint to achieve, in the light of urgency having been identified as the key symptom of OAB and the proposed model for urgency having a significant role in driving the other symptoms of OAB [10]. However, we contend that a composite endpoint is more reflective of the true nature of OAB and will ultimately be shown to correlate better with improved QoL, patient satisfaction and persistence with treatment.

Currently it is impossible to determine the optimum outcome measure for use in clinical trials of OAB. Endpoints should focus on changes in urgency, with or without other symptoms, and QoL, which, if combined, might adequately cover all key aspects important to patients. If a single composite endpoint can be agreed and adopted for use by all clinical trials, this would significantly help to harmonise the reporting of trial data, and remove the current confusion created by individual symptom reporting.

Several of the following key issues must be resolved before this advance can be made. The first is the development of a definitive method to measure urgency, and an agreement by the ICS and developers of clinical trials to use a unified approach for assessment of urinary urgency. A critical limitation of our current knowledge is that there has been very little investigation into the most valid and relevant tool to measure urgency, the most difficult symptom of OAB to measure reliably [10]. The final agreed method must be simple, reproducible and reliable. Second, there must be agreement on the role and importance of disease-specific QoL in assessing the outcome. Assessing QoL alone might be the sole outcome measured, as improvement from the patient perspective is the ultimate aim of treating benign conditions; however, there will always be a need to understand how therapy affects individual symptoms and there is evidence that other measures beyond QoL also need to be assessed to accurately capture treatment-induced improvements [21]. Finally, it remains to be seen whether a composite endpoint including symptoms other than urgency improves the performance of the outcome measure, and agreement needs to be reached on which symptoms (frequency, nocturia, incontinence) should be included, and how they should be optimally assessed.

Establishing which definition is the most appropriate to capture treatment-related changes in OAB will require further investigation. However, it is clear from this analysis that the use of a composite clinical endpoint is a valid approach when assessing OAB symptom changes. The composite endpoints presented here might be considered as a good starting point to base such a measure, accurately reflecting the current ICS definition of OAB, and having been shown to mirror changes in QoL. Further research will be required to establish the most effective measure of treatment efficacy in terms of patient outcomes.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

This study was funded and sponsored by Astellas, including editorial support and statistical analysis. Editorial assistance was provided by Ross Kay, Medicus International. No financial support was provided to Drs Payne and Kelleher for this project. Dr Payne has served as a consultant to Astellas and has been sponsored to present at seminars by Astellas. Dr Kelleher has been sponsored to attend meetings and present at seminars by Astellas.

CONFLICT OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

C. Kelleher has received honoraria for lecturing from Pfizer, Astellas, Novartis, UCB, and Watson. He is on the advisory board for Astellas and Novartis, is involved in research with Avera and Tanabe, and has a limited research grant for six month project from Pfizer.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES