Cost-effectiveness of single-dose tamsulosin and dutasteride combination therapy compared with tamsulosin monotherapy in patients with benign prostatic hyperplasia in the UK

Authors


Abstract

What's known on the subject? and What does the study add?

  • UK clinical guidelines for treating male patients with moderate to severe LUTS associated with BPH recommend treatment with an alpha-blocker (such as tamsulosin) in cases where conservative management options have not been successful or are not appropriate. An alpha-blocker plus 5-alpha-reductase inhibitor (such as dutasteride) is recommended for those patients with moderate to severe symptoms and prostate volume >30 mL.
  • The present study evaluates the cost-effectiveness of a new, single-dose combination of tamsulosin and dutasteride (Combodart®) from the perspective of the UK National Health Service. The results show that the combination therapy has a high probability of being cost-effective compared with either monotherapy, and compared with the two therapies taken separately. The probability of the combination therapy being cost-effective at an incremental cost-effectiveness ratio threshold in the range £25 000–£30 000 per quality-adjusted life year is 78–88%.

Objective

  • To estimate the long-term cost-effectiveness of single-dose dutasteride/tamsulosin combination therapy as a first-line treatment for benign prostatic hyperplasia (BPH) from the perspective of the UK National Health Service (NHS).

Methods

  • A Markov state transition model was developed to estimate healthcare costs and patient outcomes, measured by quality-adjusted life years (QALYs), for patients aged ≥50 years with diagnosed BPH and moderate to severe symptoms.
  • Costs and outcomes were estimated for two treatment comparators: oral, daily, single-dose combination therapy (dutasteride 0.5 mg + tamsulosin 0.4 mg), and oral daily tamsulosin (0.4 mg) over a period up to 25 years.
  • The efficacy of comparators was taken from results of the Combination of Avodart and Tamsulosin (CombAT) trial.

Results

  • Cumulative discounted costs per patient were higher with combination therapy than with tamsulosin, but QALYs were also higher.
  • After 25 years, the incremental cost-effectiveness ratio for combination therapy was £12 219, well within the threshold range (£20 000–£30 000 per QALY) typically applied in the NHS.
  • Probabilistic sensitivity analysis showed that the probability of combination therapy being cost-effective given the threshold range is between 78% and 88%.

Conclusion

  • Single-dose combination dutasteride/tamsulosin therapy has a high probability of being cost-effective in comparison to tamsulosin monotherapy in the UK‘s NHS.
Abbreviations
5-ARI

5-alpha-reductase inhibitor

AUR

acute urinary retention

ICER

incremental cost-effectiveness ratio

QALYs

quality-adjusted life years

TWOC

trial without catheter

Introduction

Benign prostatic hyperplasia is a common cause of morbidity in men in later life [1-3]. BPH impacts quality of life directly through development of LUTS such as incontinence, incomplete voiding and nocturia. Patients with BPH also experience a number of anxieties that indirectly impact quality of life, notably a fear of cancer, sleep disruption, discomfort and embarrassment [4]. A postal study (n = 1115) of men aged over 50 in the UK found that 41% of those surveyed had an IPSS indicating moderate or severe BPH, but that only 18% had been diagnosed [1]. More than half of patients with BPH in the same study expressed a fear that it would lead to surgery or more severe symptoms such as acute urinary retention (AUR) [1].

The main aim of medical treatment for BPH is to provide symptom relief and to prevent progression. Alpha-blockers and 5-alpha-reductase inhibitors (5-ARIs) constitute the two main pharmacological agents for the management of BPH/LUTS. 5-ARIs block the conversion of testosterone to dihydrotestosterone, thus reducing cellular growth and in turn reducing the size of the prostate [5-8]. Dutasteride is a 5-ARI which inhibits both type I and type II 5-alpha-reductase [5-8]. Alpha-blockers work by relaxing smooth muscle within the prostate, bladder and blood vessels, increasing urinary flow rates [5-8]. Tamsulosin is a uroselective alpha-blocker which targets alpha adrenoreceptors in the prostate and bladder, and is available as a generic [9].

United Kingdom clinical guidelines for treating male patients with moderate to severe LUTS recommend treatment with an alpha-blocker in cases where conservative management options have not been successful or are not appropriate. An alpha-blocker plus 5-ARI is recommended for patients with moderate to severe symptoms and prostate volume >30 mL [10].

The Combination of Avodart and Tamsulosin (CombAT) trial investigated the effect of single-dose tamsulosin/dutasteride combination therapy compared with tamsulosin and dutasteride monotherapies on the 4-year incidence of AUR, BPH-related surgery and overall clinical progression in patients ≥50 years old with moderate to severe BPH (IPSS ≥12) and prostate volume ≥30 mL [11]. The time to first AUR or BPH-related surgery was significantly shorter (P < 0.001) in patients treated with tamsulosin than in those treated with dutasteride or combination therapy. In patients with a baseline prostate volume ≥40 mL, the incidence of clinical progression (defined as one of the following: symptom deterioration [four or more points on the IPSS on two consecutive visits]; BPH-related AUR; BPH-related urinary incontinence; recurrent BPH-related UTI or urosepsis; or BPH-related renal insufficiency) was higher in patients receiving tamsulosin monotherapy than in those receiving dutasteride or combination therapy [11]. In the same study, single-dose combination therapy was shown to provide significantly greater improvements in patient-reported quality of life than either tamsulosin or dutasteride alone [12].

The aim of the present study was to estimate the long-term cost-effectiveness of oral, daily, single-dose dutasteride/tamsulosin combination therapy (Combodart®), compared with oral, daily tamsulosin (0.4 mg) in the UK. Combodart is indicated for treating moderate to severe symptoms of BPH, and for reduction in the risk of AUR and surgery in patients with moderate to severe symptoms of BPH, and is available as a combination of dutasteride 0.5 mg and tamsulosin 0.4 mg. Tamsulosin is a widely used treatment for BPH in the UK, and is the lowest priced of the treatment comparators in the CombAT trial. The weighted mean cost of generic and branded tamsulosin is £6.29/month, compared with dutasteride alone (£19.80/month), and the single-dose combination (£19.80/month).

Patients and Methods

Cost-effectiveness is assessed by comparing the additional (incremental) cost involved in choosing the combination therapy rather than tamsulosin, and the incremental improvement in patient outcomes. Patient outcomes are measured by quality-adjusted life years (QALYs) – a combination of gains in life expectancy weighted by improvements in health-related quality of life. Costs and outcomes are discounted at 3.5% as recommended by NICE [13]. The incremental cost-effectiveness ratio (ICER) is the ratio of incremental cost to incremental QALY gains. Cost-effectiveness is judged by comparing the estimated ICER with a threshold value, which in the UK is typically in the range £20 000–£30 000 per QALY [13].

Expected costs and patient outcomes for the two comparators were estimated by modelling the incidence of AUR, BPH-related surgery and clinical progression over the patient's lifetime (in practice up to 25 years). A model was developed to simulate clinical events over time, based on a discrete Markov process with six mutually exclusive health states, including one temporary state (AUR), and annual cycle length. Health states were defined to be clinically meaningful and relevant in terms of their cost consequences. The six health states were mild BPH, moderate BPH, severe BPH, AUR, post-surgery and death. Mild, moderate and severe symptom states were defined by IPSS (0–11, mild; 12–23, moderate; and 24–36, severe), in line with the definitions used in the CombAT trial [11]. Although patients with mild symptoms (IPSS <12) were not included in the eligibility criteria for the CombAT trial, a small proportion of patients (7%) had improved to the mild state after recruitment but before administration of the trial drugs. These patients were included as part of the analysis. Figure 1 shows the transitions possible between health states.

Figure 1.

Health states.

Acute urinary retention was modelled as a temporary health state to reflect the fact that it is a short-term complication of BPH which requires emergency treatment. The patient pathway after AUR depends on the success of treatment (emergency catheterization or trial without catheter [TWOC]). Successful treatment leads to a return to the previous health state. Unsuccessful treatment requires BPH-related surgery, which leads to the post-surgery health state. The AUR state was modelled in this way to reflect the assumption that successful catheterization and TWOC have no effect on disease progression.

Although there are a number of surgical options for BPH, the model assumes all patients have TURP when surgery is indicated. Patients undergoing TURP enter the post-surgery health state, where they remain until the end of the model time horizon or death. A patient can undergo up to two TURP procedures (after failure of the first procedure or relapse). A patient might experience a surgery-related adverse event regardless of the success of the TURP procedure.

The model cohort comprised men aged ≥50 years diagnosed with BPH and with moderate to severe LUTS (IPSS ≥12). To reflect the fact that BPH is a chronic condition, the analysis is carried out over a long-term horizon, up to the patient's lifetime. Results are shown at 10 and 25 years. Efficacy data were extrapolated from 4 years to lifetime, based on the 4-year CombAT trial [11].

The model is driven by probabilities of events, such as AUR, BPH-related surgery or death, and by assumptions about disease progression over time. These assumptions are reflected in annual probabilities of transition between health states (Table 1). Most of the transitions populating the model were derived directly from the clinical study report or from individual patient-level data from the CombAT trial [11].

Table 1. Treatment-specific transition probability matrices for patients aged ≥50 years with moderate to severe BPH symptoms.
Annual cyclePost-TURPMildModerateSevereAURDead
Combination therapy      
Post-TURP0.9910.0000.0000.0000.0000.009
Mild0.0000.5390.4090.0370.0060.009
Moderate0.0060.3980.5230.0580.0060.009
Severe0.0060.3460.5570.0770.0060.009
AUR0.0000.0000.0000.0001.0000.000
Death0.0000.0000.0000.0000.0001.000
Tamsulosin      
Post-TURP0.9910.0000.0000.0000.0000.009
Mild0.0000.4260.4840.0630.0170.009
Moderate0.0200.2910.5690.0930.0170.009
Severe0.0200.2410.5850.1270.0170.009
AUR0.0000.0000.0000.0001.0000.000
Death0.0000.0000.0000.0000.0001.000

Transition probabilities between BPH symptom states were calculated using the MSM package of R [14], based on patient-level data from the CombAT trial [11]. Probabilities of transition from cycle 1 (year 1) were based on months 0–12 of the trial. Transition probabilities for remaining model cycles were based on months 13–48 of the trial. Patient-level information on health states was available at each 3-monthly study visit, and this information was used to derive probabilities for transitions between each of the symptom severity states.

Surgery and AUR transition probabilities were calculated based on both the CombAT clinical study report and patient-level data. Three-month transition probabilities to the AUR and post-surgery health states were calculated from the number of yearly events using standard methods as described by Briggs et al. [15], and transformed to annual probabilities for use in the model.

Rates of onward transition from AUR – either back to the previous BPH symptom state or to the post-surgery state – were determined by the success rate of a TWOC. The care pathway for patients experiencing AUR was not reported in the CombAT trial. For the model it was assumed that 50% of TWOC procedures are successful based on a clinical review by Emberton et al. [16]. Similarly, the care pathway of patients who underwent BPH-related surgery was based on the published literature [17-19].

For the cost-effectiveness analysis, the cost and utility of both surgical and medical treatment-related adverse events were included in the analysis. The probability of any adverse event associated with TURP was determined from the European Urology Association BPH treatment guidelines [20]. This total probability was applied to all patients in the post-surgery state, regardless of the success or failure of the procedure. Adverse events associated with medical therapy were based on the CombAT trial. Only those drug-related adverse events reported to have occurred in more than 1% of the population in any treatment arm were considered for inclusion in the model. Of those, adverse events considered to be sufficiently mild (such as semen volume decrease, breast tenderness, nipple pain) were excluded from the analysis. Adverse events of a similar nature (such as retrograde ejaculation and ejaculation failure) were grouped together. The percentage of patients experiencing serious drug-related adverse events was <1% in all treatment arms of the CombAT trial, so serious adverse events were excluded from the analysis.

Resource use included resources associated with disease severity states (maintenance healthcare visits or routine care), resource use associated with AUR and resource use associated with a TURP procedure (Table 2). Since no newly diagnosed patients were considered in the analysis, the annual number of visits to health professionals varied with health state but not with time. Guidelines for the care of BPH patients in primary care and referral to secondary care were used as a basis for the type and frequency of medical treatment [21]. Assumptions about treatment patterns after AUR and BPH surgery were validated by expert clinical advisers. Unit costs were principally derived from the Department of Health, Personal Social Services Research Unit (PSSRU), the British National Formulary and data provided by the manufacturer (GlaxoSmithKline) [22-25].

Table 2. Unit costs.
CostValueSource
Cost per district nurse visit£27.00Curtis [23]
Cost per hospital nurse visit£40.50Curtis [23]
Cost per GP visit£46.73Curtis [23]
Cost per urologist visit£96.00Department of Health [22]
Cost per urodynamic test£144.00Armstrong [32]
Cost per flexible cystoscopy£687.00Department of Health [22]
Cost of prostate-related surgery with complications£2633.16Department of Health [22], GlaxoSmithKline [11]
Cost of prostate-related surgery without complications£1949.76Department of Health [22], GlaxoSmithKline [25]
Cost per episode of AUR (non-elective)£2576.84Department of Health [22]
Monthly cost of combination therapy (Combodart™)£19.80British National Formulary [24]
Monthly cost of dutasteride (Avodart™)£19.80British National Formulary [24]
Monthly cost of tamsulosin£6.29Communication from GlaxoSmithKline, weighted costs of generic and branded drugs (updated October 2010)

Utility values are patient-derived weights representing relative preferences between different health states, where full or perfect health is rated at 1 and dead is assigned a value of 0. Utility weights for each of the six health states were obtained from a meta-analysis of published studies (Baladi et al. [26]). These weights are shown in Table 3. It was assumed that utility after a successful TURP is the same as the utility of moderate BPH symptoms, and that utility after an unsuccessful TURP is the same as the utility of severe BPH symptoms. The disutility of AUR was calculated from utilities presented by Ackerman et al. [27] by weighting the utility values associated with urinary retention in the risk-adverse and non-risk-adverse groups according to sample size and then taking the difference between this and the weighted mean utility of patients with mild and moderate BPH symptoms reported in the study. It was assumed that the disutility of being catheterized while waiting for TURP was the same as the disutility of AUR. For the purposes of this analysis, it was assumed that the disutility of bladder neck contracture and urethral stricture were similar enough that they could be considered as one adverse event. This is in agreement with key opinion leaders and other published studies [28].

Table 3. Utility values.
Health stateUtility valueSource
Mild0.99Baladi et al. [26]
Moderate0.90Baladi et al. [26]
Severe0.79Baladi et al. [26]
Post-TURP, successful0.90Assumption: same as moderate health state
Post-TURP, unsuccessful0.79Assumption: same as severe health state
Post-TURP, totally incontinent0.70Baladi et al. [26]
Disutility of AUR−0.14Ackerman et al. [27]
Disutility of being catheterized while waiting for TURP−0.14Assumption: same as AUR

The disutilities for different adverse events associated with TURP (with the exception of blood transfusion) were calculated from Ackerman et al. [27]. These are shown in Table 4. This was done by taking a weighted mean of the utility associated with each adverse event and then taking the difference between this value and the weighted mean for moderate and severe patients [27]. UK clinical expert opinion indicated there was no disutility associated with a blood transfusion. The disutilities associated with adverse events in relation to the drug therapies are shown in Table 5.

Table 4. Disutility of adverse events associated with TURP.
Adverse event associated with TURPDisutilitySource
TUR syndrome−0.17Ackerman et al. [27]
Blood transfusion−0.00UK clinical expert opinion
Stress incontinence−0.04Ackerman et al. [27]
Bladder neck contracture/urethral stricture−0.02Ackerman et al. [27]
Table 5. Disutility of adverse events associated with drug treatment.
Adverse event associated with treatmentDisutilitySource
Dizziness−0.22Vera Llonch et al. [33]
Ejaculatory abnormality−0.01Ackerman et al. [27]
Impotence/erectile dysfunction−0.06Ackerman et al. [27]
Breast enlargement−0.05Penson et al. [34]

Costs and outcomes were compared between treatment options by estimating the ICER, which is a reflection of the additional cost involved in choosing one option rather than another, relative to the additional benefits which can be gained. In this case:

display math

where TCc = mean cost per patient of the combination therapy; TCt = mean cost per patient of tamsulosin monotherapy; QALYc = QALYs per patient with the combination therapy; and QALYt = QALYs per patient with tamsulosin monotherapy.

All of the main inputs to the cost-effectiveness model are subject to uncertainty. To evaluate the impact of this uncertainty on the outcomes of the model (costs and QALYs), and on the judgment about cost-effectiveness, a probabilistic sensitivity analysis was conducted. Input variables (transition probabilities, cost and utilities) are characterized by their distributions, and the model is run multiple times, each taking a different combination of variable values drawn from the underlying distribution. We present results for 1000 Monte Carlo simulations in the form of a scatterplot on a cost/QALY axis (cost-effectiveness plane; Fig. 2); and as a probability of cost-effectiveness for different values of the ICER threshold (cost-effectiveness acceptability curve; Fig. 3).

Figure 2.

Cost-effectiveness plane for combination therapy compared with tamsulosin.

Figure 3.

Cost-effectiveness acceptability curve for combination therapy compared with tamsulosin.

Variable distributions were selected in line with recommended practice in cost-effectiveness modelling [15]. Cost inputs were varied according to a log-normal distribution. Utility inputs were varied according to a normal distribution. Default means came from the associated default utility inputs for the deterministic model. Binomial probabilistic inputs for the sensitivity analysis were varied according to a beta distribution. Default means came from the associated default inputs for the deterministic model.

Results

After 10 and 25 years, cumulative discounted costs per patient are higher with combination therapy (£3354 and £5499, respectively) than with tamsulosin (£2777 and £4275). Costs are higher with the combination therapy because the higher cost of the drug is not offset by savings in the costs associated with AUR, BPH-related surgery or improvements in clinical progression. However, these improved outcomes do lead to gains in quality of life – cumulative QALYs are higher with combination therapy (7.363 and 13.665 QALYs at 10 and 25 years, respectively) than with tamsulosin (7.318 and 13.565 QALYs at 10 and 25 years, respectively) (Table 6).

Table 6. Cumulative costs and QALYs, incremental costs and QALYs, and ICER for combination therapy compared with tamsulosin.
 Discounted cumulative cost per patientDiscounted cumulative QALY per patientIncremental costIncremental QALYICER
After 10 years     
Combination£33547.363+£577+0.045£12 854
Tamsulosin£27777.318   
After 25 years     
Combination£549913.665+£1224+0.100£12 219
Tamsulosin£427513.565   

Table 6 shows incremental cost per patient, incremental QALY gains and the ICER after 10 and 25 years. After 10 years, the incremental cost-effectiveness ratio for combination therapy compared with tamsulosin is £12 854. The ICER declines with time and this is a reflection of the fact that the QALY gain increases over time slightly more than the incremental cost. After 25 years the ICER is £12 219.

Irrespective of the time period, the deterministic ICER estimates are well within the usual threshold range of £20 000–£30 000 per QALY deemed to be a cost-effective use of NHS resources. On this basis, the single-dose combination therapy would be judged to be more effective and more cost-effective than tamsulosin alone.

Probabilistic sensitivity analysis was undertaken as described earlier. Results are shown for a 25-year time horizon. Each of the points on the cost-effectiveness plane (Fig. 2) represents the combination of incremental costs and incremental QALYs from a separate model iteration with a different combination of input values. There are 1000 iterations in total.

Each iteration gives rise to a slightly different ICER value, and Fig. 3 shows the probability that combination therapy is cost-effective at different thresholds. The probability reflects the proportion of iterations that produce an ICER value less than or equal to the threshold. For a threshold in the range £20 000–£30 000, the probability that combination therapy is cost-effective relative to tamsulosin is between 78% and 88%.

Discussion

Clinical practice guidelines in the UK recommend the use of an alpha-blocker (such as tamsulosin) for patients with moderate to severe LUTS [10]. Our analysis suggests that a single-dose combination of tamsulosin and dutasteride is likely to be a more cost-effective treatment option for patients in this group. The benefits of combination therapy could be achieved by taking dutasteride and tamsulosin separately, but in the UK the single-dose presentation has a significant cost advantage. The price of the single-dose combination therapy is £19.80 per month, compared with £26.09 per month for dutasteride plus tamsulosin (Table 2). Including dispensing fees for two separate items, the cost difference will be more.

The analysis also assessed the cost-effectiveness of combination therapy compared with tamsulosin in a subgroup of the population with BPH with moderate to severe symptoms and baseline prostate volume >50 mL. In this subgroup, the ICER at 10 years was £9271. At 25 years the ICER was £10 297, indicating that combination therapy is even more cost-effective in this more severe patient group than in the general population of the CombAT trial.

Other studies have examined the cost-effectiveness of combination therapy involving an alpha-blocker and dutasteride, compared with alpha-blocker monotherapy. The results of these studies are consistent with the results of the present analysis. Takayama et al. [29] developed a seven-state Markov model to evaluate incremental QALYs and incremental costs from a payer perspective in Japan, over 4-year and 10-year time horizons. Cost-effectiveness was assessed against an ICER threshold of 6–7 million Japanese yen (JPY). In the base case analysis, the ICER for combination therapy compared with alpha-blocker monotherapy at 10 years was 5 974 495 JPY/QALY. Sensitivity analysis suggested that the ICER decreased with greater BPH severity. Bjerklund et al. [30] carried out an analysis comparing combination therapy with alpha-blocker, 5-ARI monotherapy and watchful waiting from the perspective of the Norwegian healthcare system. Overall, the combination therapy was expected to provide the greatest net monetary benefit at willingness-to-pay thresholds at or above €6000 (£5400). Results were sensitive to variability in estimates of health state utility associated with symptom severity. Antonazas et al. [31] developed a semi-Markov model to compare tamsulosin + dutasteride combination therapy with tamsulosin alone from the perspective of the Spanish National Healthcare Service over both 4 and 35 years. At 4 years, the ICER for combination therapy was €14 023/QALY, reducing to €8750/QALY at year 35.

The present study has a number of limitations. In particular, the model assumes 100% compliance with pharmacotherapy and that withdrawals due to treatment-related adverse events are not significantly different between treatment regimes (based on results of the CombAT trial). This might not be entirely consistent with clinical practice where rates of discontinuation with alpha-blocker treatment are typically high. Finasteride is an alternative to dutasteride and it would have been useful to be able to include this as an additional comparator in the model. This has not been possible because of a lack of appropriate head-to-head clinical data. Inevitably, lack of data on some model inputs and the need for simplifying assumptions to make the modelling tractable have introduced additional sources of uncertainly to the results.

In conclusion, the combination of dutasteride and tamsulosin has been shown to significantly reduce the risk of AUR, BPH surgery and clinical progression compared with tamsulosin monotherapy. Current UK clinical guidelines recommend the two treatments together for patients with bothersome moderate to severe symptoms and prostate >30 mL. This analysis has demonstrated that a single-dose combination has a high probability of being more cost-effective than tamsulosin monotherapy, as well as being more effective, and is currently cheaper than a combination of tamsulosin and dutasteride taken separately.

Acknowledgements

The authors wish to acknowledge the work of Tim Reason, Kevin Lock and Kerry Gairy who were involved in the development of the original economic model.

Conflict of Interest

Source of Funding: GlaxoSmithKline. Scott Doyle and Manjit Hunjan are employees and shareholders of GlaxoSmithKline. All listed authors meet the criteria for authorship set forth by the ICMJE.

Ancillary