An Economic Evaluation of Triptan Products for Migraine

Transcription

1 Blackwell Science, LtdOxford, UKVHEValue in Health ISPOR Original ArticleEconomic Evaluation of TriptansPerfetto et al. Volume 8 Number VALUE IN HEALTH An Economic Evaluation of Triptan Products for Migraine Eleanor M. Perfetto, PhD, MS, 1 Kathleen A. Weis, DrPH, PhD, MSN, ENP, FNP, 2 C. Daniel Mullins, PhD, 3 Prasun Subedi, MS, 3 Paul J. Healey, Sr., MD, MPH, JD, MBA 2 1 Epidemiology and Biostatistics, The Weinberg Group Inc., Washington, DC, USA; 2 Pfizer Global Pharmaceuticals, Pfizer Inc., New York, NY, USA; 3 Pharmaceutical Health Service Research, University of Maryland School of Pharmacy, Baltimore, MD, USA ABSTRACT Objective: A composite outcome measure in migraine treatment assessment is useful to clinical decision-makers and payers as it can provide a more accurate reflection of effectiveness and allows for more complete modeling of economic value. The objective of this study was to compare the total triptan cost to treat 100 migraine patient attacks and the cost per successfully treated patient (CPSTP) for six marketed triptans using a composite measure of effectiveness, the successfully treated migraine (defined as requiring only one triptan dose to treat one migraine attack during a 24-h period). Methods: This analysis was conducted from the perspective of the payer. Clinical data were abstracted from a rigorous, published meta-analysis. Two-hour response and pain-free response were used in conjunction with the recurrence rate reported in the meta-analysis to calculate the number of doses used by treatment successes and failures. The average wholesale price per dose was then used to calculate total triptan cost. Results: Of the nine oral triptan doses compared, eletriptan 40 mg was associated with both the lowest total triptan cost for treating 100 migraine attacks ($1560) and with the lowest CPSTP ($56.39). Conclusions: The relative CPSTP rankings for migraine therapies are dependent on the definition of treatment success and relative pricing. The results of this study support the use of eletriptan for the treatment of acute migraine based on the model assumptions. This study can be used to assist in formulary considerations and offers a model that can be adapted by health-care decision-makers. Keywords: 5-HT receptor agonists (triptans), cost effectiveness, cost, managed care, meta-analysis, migraine. Introduction Migraine is a chronic, episodic condition that places a tremendous burden on the health-care system, patients and their families, and employers [1,2]. It is estimated that $1 billion is spent annually in medical expenditures for migraine treatment in the United States, and that over $13 billion are lost annually because of reduced productivity and employee absences [3]. The 5-HT 1B/1D receptor agonists (triptans) have become the preferred treatment choice for migraine [4]. Although triptans have added to the direct cost for drug expenditures for the treatment of migraine, they have significantly reduced lost productivity and work absences, and have contributed to improved patient quality of life [5 8]. Health-care decision-makers need evidencebased information to make informed formulary Address correspondence to: Paul J. Healey, Outcomes Research, Pfizer Global Pharmaceuticals, Pfizer Inc., 235 East 42nd Street, New York, NY , USA. paul.j.healey@pfizer.com /j x selections from among the marketed triptans. Economic models help to support health-care decisionmaking, and allow decision-makers to base formulary choices on value for dollars spent on therapy. Nevertheless, the migraine economic literature to date has limited utility for decision-makers for several reasons. First, existing economic studies have only used standard efficacy endpoints (e.g., 2-h response rate) in the economic analyses. These measures, although accepted as efficacy endpoints for clinical trials, may not sufficiently capture the experience of the migraine patient and have not considered effectiveness from the patient perspective (e.g., wanting quick return to pain-free status without migraine recurrence). Second, only limited comparisons among various therapy options have been made in the existing literature. Also, most studies do not use methodology that takes into account the placebo response noted in pain and migraine research. These issues are discussed in more detail below. Although clinically useful, the standard 2-h response (R2h) or 2-h pain-free response (PF2) effi- ISPOR /05/

2 648 cacy endpoints do not capture the full effectiveness or reflect the rate of migraine recurrence associated with the various triptans. Recently, Ferrari et al. [9] have recommended the use of the 24-h sustained pain-free response (24-h SPFR), a composite measure that is defined as the proportion of patients who are pain-free by 2 h post dose, who do not experience a recurrence of moderate or severe migraine pain, and who also do not use any rescue medications during the remainder of the 24-h postdose period. Ferrari notes that although the sustained pain-free (measure) is the ideal efficacy endpoint, it is also the hardest to achieve and it may be unrealistic to expect high sustained pain-free rates with the current drugs. [9,10] A more realistic and achievable headache response measure is the 24-h sustained response (24-h SR), defined as a moderate or severe migraine that has a reduction in pain to mild or pain-free within 2 h of one triptan dose, and is sustained with no recurrence of moderate or severe pain for the remainder of the 24-h postdose period, without the use of rescue medications. Nevertheless, most published migraine studies have not used composite endpoints like the 24-h SPFR or 24- h SR. (See Table 1 for a summary of commonly used migraine clinical endpoint definitions.) Another aspect of published migraine economic studies that limits their usefulness to health-care decision-makers is that most existing studies have modeled comparisons of only a few selected treatments. There are several economic analyses that use a variety of cost models to compare one triptan to another triptan, or to compare triptans with usual or customary therapy. For example, Biddle and colleagues [11] compared sumatriptan to usual therapy for a US population. Caro et al. [12] compared naratriptan to customary therapy for a Canadian population. Caro et al. have also compared sumatriptan to usual therapy [12]. Williams et al. [13] compared a stratified-care regimen (which included zolmitriptan) to a more typical stepped-care regimen for a UK population. Mannix Perfetto et al. et al. [14] compared sumatriptan and almotriptan in a model that examined the direct costs of managing adverse chest symptoms. There are a few published economic reports that examine multiple triptans, but these studies do not compare the drugs in head-to-head clinical trials. Adelman and Belsey [15] performed a meta-analysis of efficacy data from 27 oral triptan trials, and calculated the cost to achieve pain-free status within 2 h after initial dosing with a triptan. In a recent analysis, Reeder et al. [16] considered several triptans in an economic study of migraine therapies. Abstracting the efficacy and safety data presented from a published meta-analysis [17], Reeder and colleagues applied price estimates to calculate estimated total drug cost. Although the Reeder study compared multiple triptans, two of the newest triptan products (eletriptan and frovatriptan) were excluded from the study because data from the source meta-analysis or data related to pricing in the United States were not available at the time of their analysis. It should be noted that both Adelman and Belsey [15], and Reeder et al. [16] used the number needed to treat (NNT) methodology in their studies. The use of NNT or a therapeutic gain (TG) approach is particularly useful, as these methods control for variability in the placebo response seen both from trial to trial and from product to product [18 21]. This is an important consideration in pain studies. Given the relative lack of cost-effectiveness studies comparing multiple triptans utilizing a composite measure of effectiveness, more information is needed by health-care decision-makers for appropriate migraine therapy selection. The objective of this study was to compare the triptan costs to achieve treatment success among available triptan therapies in the United States using a composite measure of effectiveness. The composite measure used in this analysis was the successfully treated migraine, defined as requiring only one triptan dose to treat one migraine attack during a 24-h Table 1 Clinical endpoint definitions in migraine treatment Response rate at 2 h (R2h) Pain-free at 2 h (PF2) Sustained pain-free (SPF) Recurrence (Rec) The proportion of patients whose moderate or severe headache at baseline improves to mild or no pain at 2 h postdose (response at 2 h is the traditional primary efficacy endpoint in triptan migraine trials). The proportion of patients whose moderate or severe headache at baseline improves to no pain at 2 h postdose (pain-free at 2 h is now the recommended primary efficacy endpoint in acute migraine trials). The proportion of patients who were pain-free at 2 h and who did not experience a recurrence of moderate or severe headache and did not use any analgesic or other headache medication over the subsequent 22 h postdose. The proportion of patients with headache response at 2 h who experience a return (or relapse) of moderate or severe headache in the subsequent 22 h. Source: Ferrari et al., 2002 [9].

3 Economic Evaluation of Triptans 649 period. Successful treatment was used to calculate the total triptan cost to treat 100 migraine patient attacks and the cost per successfully treated patient (CPSTP). Methods Overview This study examines both the total triptan cost and the CPSTP in treating a hypothetical cohort of 100 migraine patient attacks. The perspective is that of the payer, typically a health-care insurer or managed care organization. Clinical data were taken from the published literature and were used to calculate the number of successfully treated patients and the number of treatment failures. Standardized prices were applied to calculate the total cost for triptans used to treat 100 patient attacks. The CPSTP was subsequently calculated for each triptan product examined. Clinical Data Source Clinical data (e.g., response rates, recurrence rates) for this study were abstracted from the metaanalysis performed by Ferrari et al. [9,17]. In their analysis, Ferrari and colleagues recognized that migraine is a condition associated with a considerable placebo treatment effect. To account for this placebo effect, Ferrari et al. presented the placebosubtracted response rate for each triptan. Specifically, Ferrari et al. subtracted for each individual study the placebo response from the response to the active drug [9], and presented this calculated rate as a more accurate measure of treatment efficacy. Ferrari deemed these adjusted response rates the TG for each triptan. Ferrari s TG-adjusted measures were used in this analysis along with the reported recurrence rates. Frovatriptan was excluded from this analysis because no information was provided by the sponsor company to Ferrari et al. [9,17] for the metaanalysis; thus, data available were too limited to include the product. Additionally, eletriptan 80 mg was excluded from this analysis as it is not an approved dose in the United States. In total, data were abstracted from Ferrari et al. [9,17] for six different triptan products, several with multiple dose options, for a total of nine different drug/dose triptan products. Definition of Outcome: the Successfully Treated Migraine The successfully treated migraine, was used as the outcome measure in this analysis, and is defined as requiring only one triptan dose to treat one migraine attack during a 24-h period. Figure 1 depicts the schematic used to classify patients into categories of treatment success or treatment failure. Time line 0 h > 2 h > 24 h Response at 2 h (R2h) Recurrence (Mod/Sev pain) PF2 x recurrence Take 2 triptan doses Pain-free at 2 h (PF2) Mild pain (Mild pain) PF2 x (100% Rec SPF) Take 1 triptan dose Sust. pain-free (SPF) PF2 x sustained PF Take 1 triptan dose Moderate or Severe Migraine Pain* Mild pain (R2h PF2) Recurrence (Mod/Sev pain) No recurrence (No/Mild pain) Mild pain x recurrence Take 2 triptan doses Mild pain x (100% recurrence) Take 1 triptan dose *Assumes complete data (no missing data) set. Moderate/severe pain (100% R2h) Although patients in this group do not get a response with the first dose, some take a second triptan dose. Base case assumes only one dose taken by this group. Sensitivity analysis assumes that one-half of these patients. take a second dose. Take 1 or 2 triptan doses Figure 1 Model for calculation of treatment success versus failure using clinical data from a published meta-analysis. R2h, response rate at 2 h; PF2, pain-free response rate at 2 h; PF, pain-free rate; SPF, sustained pain-free rate; Rec, recurrence rate in the 24-h period.

4 650 In the Figure, patients experiencing a moderate or severe migraine who take one triptan dose for the initial treatment of their migraine pain experience one of two possible outcomes. Some patients will achieve a response by 2 h (R2h, defined as a decrease in pain to mild or pain-free); other patients will not achieve a response at 2 h, and will remain at moderate or severe pain. Those patients who do not achieve a response by 2 h are automatically classified as treatment failures in this analysis, regardless of subsequent pain relief. Patients who are classified into the R2h group can be further split into two subcategories: those who had mild pain at 2 h, and those who were painfree at 2 h (PF2). The number of patients who had mild pain at 2 h is calculated by subtracting the percentage of patients who were classified as PF2 from the percentage of patients classified as R2h. If patients achieved PF2, but then experienced a recurrence of migraine pain, they were deemed to be treatment failures. Similarly, patients who achieved mild pain at 2 h, but then experienced a recurrence of migraine pain were also classified as treatment failures. Any patient who had achieved R2h (both PF2 and mild pain), and who did not experience a recurrence in the subsequent 22 h were the patients deemed to have achieved 24-h SR, and were classified as treatment successes. These patients only required the use of one triptan dose to eliminate their migraine pain, and did not need any additional medication to manage their migraine. Dosing Assumptions The classification scheme used in Figure 1 was also used to determine how many triptan doses a patient in each category would take. It was assumed that all patients took at least one triptan dose for the Perfetto et al. initial treatment of migraine pain. It was further assumed that those patients who did achieve R2h but then experienced a subsequent recurrence of pain took a second triptan dose. For the cohort of patients who achieved R2h and did not experience a recurrence (24-h SR, the treatment success group), it was assumed that no additional triptans were needed. In the base case analysis, it was assumed that patients who did not achieve R2h (e.g., those who remained at moderate/severe pain at 2 h) took only one triptan dose. Nevertheless, it was recognized that some patients who did not achieve R2h might try to take a second dose of triptan, even though such additional dosing may not be recommended. It is unknown how frequently this might happen. Therefore, a sensitivity analysis was conducted and assumed that one-half of the nonresponders at 2 h attempted to relieve their moderate or severe migraine by taking a second triptan dose. Calculation of Total Triptan Cost and CPSTP The cost for each triptan dose was calculated using the average wholesale price (AWP) for each triptan marketed in the United States [22]. The costs were assumed to be AWP minus 15% for a branded triptan, and are listed in Table 2. The total triptan cost for treating a hypothetical cohort of 100 migraine patient attacks was calculated, as well as the cost to successfully treat one migraine patient. Costs were calculated based on the expected number of doses each of the subgroups depicted in Figure 1 would use and were calculated in Table 3. The total triptan cost was calculated as the sum of the cost of treating both the successfully treated patients and the treatment failures for the 100 patient attacks. The CPSTP was calculated as the ratio of the total triptan cost to treat 100 patient Table 2 Treatment Calculation of triptan cost to treat 100 migraine attacks and cost per successfully treated patient Response at 2-h TG (R2h TG)* (%) Pain-free 2-h TG (PF2 TG)* (%) Recurrence (Rec)* (%) Sustained pain-free (SPF)* (%) Health-care payer Price /Dose (AWP-15%) ($) Total cost to treat 100 patient attacks ($) Cost per successfully treated patient (CPSTP) ($) Almotriptan 12.5 mg Eletriptan 40 mg Naratriptan 2.5 mg Rizatriptan 5 mg Rizatriptan 10 mg Sumatriptan 50 mg Sumatriptan 100 mg Zolmitriptan 2.5 mg Zolmitriptan 5 mg *Abstracted from Ferrari et al., 2001 and 2002 [9,17]. Assumed price per dose in the United States. Average Wholesale Price (AWP) from AnalySource, FirstDataBank August 13, PF2 TG, pain-free response rate at 2 h postdose therapeutic gain; R2h TG, response rate at 2 h postdose therapeutic gain; Rec, recurrence rate in the 24-h period.

5 Economic Evaluation of Triptans 651 Table 3 Description of success and failures in the model Category Description Formula Treatment success: Cost for patients who achieved 24-h sustained response Treatment Failure: Costs for patients who responded in the first 2 h, but subsequently had a recurrence (thus, requiring a second triptan dose) Treatment Failure: Cost for patients who never attained a 2-h response (Proportion of patients who achieved R2h, and did not experience a recurrence) (Cost of one triptan dose) (100 patient attacks) (Proportion of patients who initially achieved R2h, but then experienced a recurrence) (Cost of two triptan doses) (100 patient attacks) (Proportion of patients who did not achieve R2h) (Cost of one triptan dose) AWP, average wholesale price; R2h TG, response rate at 2 h postdose therapeutic gain; Rec, recurrence rate in the 24-h period. R2h TG (1 Rec) (AWP 15% for one triptan dose) 100 patient attacks R2h TG Rec (AWP 15% for two triptan doses) 100 patient attacks (1-R2h TG) (AWP 15% for one triptan dose) attacks divided by the number of successfully treated patients in 100 patient attacks. Results Table 2 summarizes the data abstracted from Ferrari et al. [9,17], the assumed costs for each triptan considered, and the calculated total triptan cost and CPSTP. Eletriptan 40 mg had the lowest total cost to treat 100 patient attacks at $1560, followed by zolmitriptan 2.5 mg with a total cost of $1629. The triptans with the highest total cost to treat 100 patient attacks were rizatriptan 10 mg at $1802; zolmitriptan 5 mg at $1889; and naratriptan 2.5 mg at $1945. Eletriptan 40 mg resulted in the lowest CPSTP ($56.39), followed by zolmitriptan 2.5 mg ($75.62). The highest CPSTPs were associated with naratriptan 2.5 mg ($111.44), rizatriptan 5 mg ($105.72), and almotriptan ($90.52). In the sensitivity analysis where nonresponders to all drugs were assumed to take 1.5 doses, no differences in ranking by CPSTP were noted. Eletriptan remained the lowest ($73.38), followed by zolmitriptan 2.5 mg ($99.51). Naratriptan remained the highest at $ for the CPSTP. Discussion This study provides a comparative analysis of all the triptans marketed in the United States except frovatriptan, calculating both the total triptan cost to treat 100 patient attacks and the CPSTP for each drug and dose. The primary clinical endpoint used in this comparison was successful treatment; the 24-h SR. The 24-h SR rate is a composite outcome measure that better reflects the true effectiveness of a product, as compared with the simpler consideration of PF2 or R2h endpoints typically utilized for product approval [9,23]. When the efficacy data abstracted from Ferrari et al. [9,17] are compared with the composite endpoint and to the cost analysis, patterns emerge that allow for predictions about which products will have a relatively high 24-h SR rate and which products will be associated with lower costs. Products with relatively good efficacy at 2 h, relatively lower placebo-response rates, and relatively lower 24-h recurrence rates, result in the best composite outcome measure. This translates into better results for the migraine patient and lower triptan costs for the third-party payer. In their systematic review of the migraine pharmacological treatment literature, Oldman et al. [20] emphasized the importance of composite endpoints like the 24-h SR rate, highlighting the relationship between sustained relief and low cost. Simply stated, patients who respond quickly to treatment and who do not experience recurrence require no additional care or drugs. Thus, sustained relief may be the most important outcome from the standpoint of cost [20]. A health insurance plan generally can estimate the number of members who suffer from migraine and the number of attacks those sufferers typically experience in a year. The decision-maker needs to be able to estimate the triptan costs to effectively treat those attacks. In this study, the definition of effectiveness a successful treatment is achieving pain relief status at 2 h and maintaining that status for an additional 22 h. These are more stringent criteria than utilized in the past, but they are based on what patients desire from successful migraine therapy [24]. A treatment that can achieve a successful outcome for an attack with one dose is most desirable; that is, one dose of triptan for a single attack of migraine. Those patients who achieved 24-h SR had one migraine treated with one dose, which would be expected to be the most efficient and lowest cost approach. Thus, products with the highest 24-h SR will be expected to have the lowest total costs. Nevertheless, if there is also a high recurrence, then total costs to treat the attacks will increase because it is

6 652 expected that more than one dose will be taken for those who were initially successful at 2 h but who recurred, and also for a proportion of those who did not respond initially (Fig. 1). Eletriptan 40 mg had the highest R2h (35.2%), followed by rizatriptan 10 mg (34.6%). Eletriptan 40 mg had the lowest total costs ($1560) because of its relatively low unit cost and low recurrence rate. Nevertheless, rizatriptan 10 mg had a relatively moderate unit cost and a relatively high recurrence rate; this resulted in a higher total cost ($1802) than would have been anticipated based on its R2h ranking. Rizatriptan 10 mg had total costs higher than that of zolmitriptan 2.5 mg, almotriptan 12.5 mg, and sumatriptan 50 mg and 100 mg even though its R2h was higher (Table 2). Published Meta-Analyses Several meta-analyses have been published examining comparisons among triptan therapies (Table 4). The meta-analysis by Ferrari et al. [9,17] was rigorous and served as the data source for this study. Reeder et al. [16] also used the data from the Ferrari et al. [9] meta-analysis in their cost study. Nevertheless, the authors did not include eletriptan in their comparison study because of a lack of pricing data at that time; thus both eletriptan 40 mg and 80 mg were not considered. The authors also found that naratriptan had the highest cost to attain h sustained pain-free patients at almost $12,000. Almotriptan was lowest in that analysis at just over $4000. In the current analysis, almotriptan total cost and CPSTP were higher than both eletriptan and zolmitriptan. This may indicate that accounting for the placebo effect in the endpoint can make a difference in the results and data interpretation. Reeder et al. [16] supplemented their analysis of 24-h sustained pain-free costs with an additional measure 24-h sustained pain-free without adverse events by using the overall proportion of patients experiencing adverse events. Although this accounting for differences in adverse event rates between products contributes to the calculation of a more ideal composite outcome measure, the data for this calculation required many assumptions. Specifically, the analysis assumed that all adverse events were reported accurately and uniformly, using the same terminology; that each adverse event carried the same weight and severity; and that the distribution of the events was the same for the entire study population (i.e., across successfully treated patients, patients who have a recurrence, and nonresponders). Perfetto et al. In their literature review, Oldman et al. [20] reported that a meaningful assessment of adverse events in a 24-h period could not be completed. They stated that for most studies, the adverse events are reported for 7- to 10-day periods. They also stated that data collection and reporting were not of sufficient quality and that the information was often contaminated by the effects of rescue medications. For these reasons, an attempt to refine the composite measure by including the adverse event rates for the triptans considered was not done in the present study. The Bandolier Library [25] includes a league table comparing sustained response over 24 h for various triptans. The league table compared the triptans in terms of the NNT. Bandolier reports that eletriptan 80 mg and 40 mg have the lowest NNT whereas sumatriptan 50 mg and 100 mg, along with rizatriptan 5 mg, were among the highest. Zolmitriptan, almotriptan, and naratriptan were not reported by Bandolier [25]. Tfelt-Hansen [26] reported a similar trend in a comparison of TG among oral therapies, with eletriptan 80 mg and 40 mg, and rizatriptan 10 mg as the highest in TG, defined as proportion of patients responding to active drug minus proportion responding to placebo. Naratriptan, sumatriptan 50 mg, and rizatriptan 5 mg again fell among the lowest. Zolmitriptan was not reported in that study [26]. Although these meta-analytic studies reported similar trends as compared with the data reported here, there are some differences (Table 4). This was most likely due to the different approaches taken. For example, Reeder et al. [16] used the sustained pain-free response. Adelman and Belsey examined 2-h response, not sustained response [15]. The Bandolier report [25] was similar in design to the current study and accounted for both the placebo effect and the sustained response. The results reported here are similar to the Bandolier results. Nevertheless, the Bandolier report was based on a small number of trials and excluded many of the drugs included in this analysis [25]. Limitations Further work is needed to assess what other medications (e.g., triptan, nontriptan prescriptions, and OTCs) patients are taking. There is inadequate information on patient use of medications in recurrence of migraine. Additionally, some patients may take a second triptan when they have not had a response (mild or no pain) or take a triptan before their migraine is moderate or severe. Further research in this area is required to better

7 Economic Evaluation of Triptans 653 Table 4 Reports of migraine end points from published meta-analyses Study Current analysis Reeder et al. (2002) [16] Bandolier (2002) [25] and Oldman et al. (2002a) [27] Adelman and Belsey (2003) [15] Tfelt-Hansen (2000) [18] Measure CPSTP 24-h sustained pain-free cost* NNT sustained response 2-h pain-free NNT calculated from TG based on response Lowest Eletriptan 40 mg = $56.36 Almotriptan 12.5 mg = $4,100 Eletriptan 80 mg = 2.8 Rizatriptan 10 mg = 3.2 Eletriptan 80 mg = 2.4 Eletriptan 80 mg = $93.31 (base-case assumption) Rizatriptan 10 mg = $6,200 Eletriptan 40 mg = 3.6 Zolmitriptan 5 mg = 4.2 Eletriptan 40 mg = 2.6 Zolmitriptan 2.5 mg = $75.62 Zolmitriptan 5 mg = $7,600 Rizatriptan 10 mg = 5.6 Almotriptan 12.5 mg = 4.7 Rizatriptan 10 mg = 2.6 Sumatriptan 50 mg = $77.59 Sumatriptan 100 mg = $8,000 Sumatriptan 50 mg = 6.0 Sumatriptan 100 mg = 4.7 Zolmitriptan 2.5 mg = 3.1 Rizatriptan 10 mg = $82.53 Naratriptan 2.5 mg = $11,600 Sumatriptan 100 mg = 6.7 Zolmitriptan 2.5 mg = 5.1 Sumatriptan 100 mg = 3.1 Zolmitriptan 5 mg = $84.93 Rizatriptan 5 mg = 8.3 Sumatriptan 50 mg = 6.0 Sumatriptan 50 mg = 3.3 Sumatriptan 100 mg = $85.29 Naratriptan 2.5 mg = 8.2 Rizatriptan 5 mg = 3.8 Almotriptan 12.5 mg = $90.52 Frovatriptan 2.5 mg = 11.3 Almotriptan 12.5 mg = 3.8 Rizatriptan 5 mg = $ Naratriptan 2.5 mg = 4.5 Highest Naratriptan 2.5 mg = $ Frovatriptan 2.5 mg = 6.3 Drugs not included/not calculated Frovatriptan Eletriptan Almotriptan Eletriptan Zolmitriptan 5 mg Frovatriptan Frovatriptan Sumatriptan 50 mg Naratriptan Rizatriptan 5 mg Zolmitriptan Zolmitriptan 2.5 mg *Dollar values are estimated from bar graph in Reeder et al., 2002 [16]. Sustained, indicated that patient did not experience recurrence during the 22 h after meeting the 2-h response criteria. Response, patients met the criteria for responding to therapy but were not necessarily pain-free. Pain-free, patients had to be pain-free at 2 h to be considered a responder.

8 654 understand these patients and treatment effectiveness, as well as the impact of concurrent medications and comorbidities. In the analysis, the assumption was made that recurrence rates were equal across outcome subgroups. Specifically, it is assumed that the recurrence rate for patients who achieved pain-free status at 2 h was the same as the recurrence rate for patients who achieved mild pain at 2 h. Nevertheless, this may not be the case; patients in one outcome group could have lower or higher recurrence rates than patients in the other group. Because Ferrari et al. did not report differential recurrence rates by 2-h pain status, this possibility could not be taken into account in this study. Nevertheless, the potential for differences in recurrence rates between patients with mild or no pain at 2 h, or between treatment and placebo groups should be recognized and investigated in future research. This study did not consider the costs of the rescue medication that might be used by patients who did not respond to treatment or who have a recurrence after an initial 2-h response. These details were not reported in the meta-analysis that was the source for data used in this analysis. Also, the range of choices for rescue medications is very broad and the costs vary widely. This analysis focused only on doses and costs for the triptan products compared. An additional limitation of this type of study is that the data used are not from a head-to-head trial comparing all of the triptans. Although this type of trial would be the ideal source of information for this type of economic analysis, it may not be considered feasible to conduct such a large and expensive randomized, controlled trial. Nevertheless, given the general similarities between clinical trials of the triptan therapies, in terms of outcomes and endpoints considered, patient populations, and other design attributes, meta-analysis is considered a valid and reliable tool for combining similar data from separate trials. Conclusions The relative CPSTP ranking for migraine therapies is dependent on the definition of treatment success and relative pricing. A composite outcome measure for defining migraine treatment success is useful in that it provides a more accurate reflection of effectiveness and allows for more complete modeling of the value of therapy. This measure also must account for both short- and long-term effects, as well as the placebo effect of the treatment. In this study, a composite measure of effectiveness, the Perfetto et al. treatment success, was based on the 24-h SR rate. Of the triptan therapies compared, eletriptan 40 mg was associated with the lowest total triptan cost for treating 100 attacks and with the lowest CPSTP. The results of this study support the use of eletriptan for the treatment of acute migraine based on the model assumptions made. This study can be used to assist in formulary considerations and offers a model that can be adapted by health-care decision-maker. Source of financial support: This study was sponsored by Pfizer Pharmaceuticals, New York, NY, USA. References 1 Ferrari MD. The economic burden of migraine to society. Pharmacoeconomics 1998;13: Lipton RB, Stewart WF, Diamond S, et al. Prevalence and burden of migraine in the United States: data from the American Migraine Study II. Headache 2001;41: Hu XH, Markson LE, Lipton RB, et al. Burden of migraine in the United States: disability and economic costs. Arch Intern Med 1999;159: Silberstein SD. Practice parameter: evidence-based guidelines for migraine headache (an evidencebased review): report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 2000;55: Lofland JH, Johnson NE, Batenhorst AS, Nash DB. Changes in resource use and outcomes for patients with migraine treated with sumatriptan: a managed care perspective. Arch Intern Med 1999;159: Cohen JA, Beall D, Beck A, et al. Sumatriptan treatment for migraine in a health maintenance organization: economic, humanistic, and clinical outcomes. Clin Ther 1999;21: Fishman P, Black L. Indirect costs of migraine in a managed care population. Cephalalgia 1999;19: Coukell AJ, Lamb HM. Sumatriptan. A pharmacoeconomic review of its use in migraine. Pharmacoeconomics 1997;11: Ferrari MD, Goadsby PJ, Roon KI, Lipton RB. Triptans (serotonin, 5-HT 1B/1D agonists) in migraine: detailed results and methods of a metaanalysis of 53 trials. Cephalalgia 2002;22: Lipton RB, Stewart WF. Do doctors understand what patients with migraine want from therapy? Headache 1999;39: Biddle AK, Shih YC, Kwong WJ. Cost-benefit analysis of sumatriptan tablets versus usual ther-

9 Economic Evaluation of Triptans 655 apy for treatment of migraine. Pharmacotherapy 2000;20: Caro G, Getsios D, Caro J, et al. Sumatriptan: economic evidence for its use in the treatment of migraine, the Canadian comparative economic analysis. Cephalalgia 2001;21: Williams P, Dowson AJ, Rapoport AM, Sawyer J. The cost effectiveness of stratified care in the management of migraine. Pharmacoeconomics 2001;19: Mannix LK, Adelman JU, Goldfarb SD, et al. Almotriptan versus sumatriptan in migraine treatment: direct medical costs of managing adverse chest symptoms. Am J Manag Care 2002;8:S Adelman JU, Belsey J. Meta-analysis of oral triptan therapy for migraine: number needed to treat and relative cost to achieve relief within 2 hours. J Manag Care Pharm 2003;9: Reeder CE, Steadman S, Goldfarb SD. Economic comparison of oral triptans for management of acute migraine: implications for managed care. Am J Manag Care 2002;8:S Ferrari MD, Roon KI, Lipton RB, Goadsby PJ. Oral triptans (serotonin 5-HT 1B/1D agonists) in acute migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358: Tfelt-Hansen P. Efficacy and adverse events of subcutaneous, oral, and intranasal sumatriptan used for migraine treatment: a systematic review based on number needed to treat. Cephalalgia 1998;18: Goadsby PJ. A triptan too far? J Neurol Neurosurg Psychiatry 1998;64: Oldman AD, Smith LA, McQuay HJ, Moore RA. Rizatriptan for acute migraine. Cochrane Database Syst Rev 2002; CD Goadsby P. Response to Therapeutic Gain: A Critique. Headache 1999;39: AnalySource Online. WWW document: Available from: AnalLogin.taf [Accessed August 13, 2004]. 23 Tfelt-Hansen P, Block G, Dahlöf C, et al. Guidelines for controlled trials of drugs in migraine: second edition. Cephalalgia 2000;20: Lipton RB, Hamelsky SW, Dayno JM. What do patients with migraine want from acute migraine treatment? Headache 2002;42(Suppl. 1):S Migraine League Table: Acute Treatments and 24 Hour Sustained Response. Oxford, UK: Bandolier Library, Available from: ac.uk/bandolier/booth/migraine/league.html. 26 Tfelt-Hansen P, De Vries P, Saxena PR. Triptans in migraine: a comparative review of pharmacology, pharmacokinetics and efficacy. Drugs 2000;60: Oldman AD, Smith LA, McQuay HJ, Moore RA. Pharmacological treatments for acute migraine: quantitative systematic review. Pain 2002;97: