05wang 77.85

The Journal of Mental Health Policy and EconomicsJ Ment Health Policy Econ 7, 77-85 (2004) Should Clozapine Continue to be Restricted to Third-Line Status for Schizophrenia?: A Decision-Analytic Model Philip S. Wang,1* David A. Ganz,2 Joshua S. Benner,3 Robert J. Glynn,4 Jerry Avorn5 1MD, DrPH, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA 2MD, MPH, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA 3PharmD, ScD, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA 4PhD, ScD, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA 5MD, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA treatments for treatment-sensitive as well as treatment-resistantschizophrenia.
Background: Clozapine is currently restricted to patients who have Received 28 April 2003; accepted 4 April 2004 failed at least two trials of other antipsychotic medications becauseof concerns that its use as a first–line agent would lead to greatermortality, mainly through agranulocytosis.
Aims of the Study: We sought to determine the cost-effectivenessof allowing clozapine to be a first-line treatment versus the current policy of restricting clozapine to third-line status.
Methods: We performed a cost-effectiveness analysis usingpublished data from randomized controlled trials and epidemiologic In 1989, the U.S. Food and Drug Administration (FDA) studies. The target population was patients with schizophrenia in an approved clozapine as a treatment for schizophrenia.
acute psychotic episode, with a lifetime time horizon and societal However, the FDA restricted clozapine to third-line status, perspective. Outcome measures included life expectancy, quality- only allowing it to be used after a patient has failed at least adjusted life expectancy, costs, and cost-effectiveness ratios.
two different antipsychotic medications for lack of response Results: Using clozapine as a first agent would lead to modest gains or intolerable side effects.1-3 This restriction of clozapine to in life-expectancy as well as quality-adjusted life expectancy, use in treatment-resistant patients arose out of concern that if relative to restricting its use to patients who failed 2 conventionalantipsychotics. The cost-effectiveness ratio of using clozapine first clozapine were used as a first-line agent, it would lead to vs. using clozapine third would be $24,100 per quality-adjusted life increased mortality. In pre-marketing studies clozapine was year (QALY). In 1-way and probabilistic sensitivity analyses, these found to cause potentially fatal agranulocytosis in findings were robust to a variety of assumptions.
approximately 1% of patients.4,5 Because this risk was Discussion: Allowing clozapine to be a first-line agent may lead to reversible if detected early, weekly monitoring of patients' small gains in life expectancy at moderate but acceptable costs.
white blood cell counts was required;5 this further added to Implications: While these results do not shed light on whether the cost associated with clozapine treatment, which already clozapine should be the preferred first-line strategy, they do suggest greatly exceeded that of conventional antipsychotic therapy that clozapine should be added to the armamentarium of possible due to higher drug costs.6 Restricting clozapine to treatment-resistant patients was also supported by the fact that earlyclinical trials showing greater efficacy for clozapine vs.
* Correspondence to: Philip Wang, M.D., Dr.P.H., Division of conventional antipsychotics were conducted mainly in Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, 1620 Tremont Street, Suite 3030 Boston, MA 02120, USATel.: +1-617-278 0930 However, in the decade since this restriction on clozapine Fax: +1-617-232 8602 to third-line status was imposed, much additional evidence has emerged concerning its risks and benefits. A recent meta- Source of Funding: Financial support for this study was provided entirely analysis8 and other randomized trials9 conducted among by a Research Career Award from the National Institute of Mental Health treatment-sensitive as well as treatment-resistant patients (K01-MH0165 to Wang). No pharmaceutical company was involved in anyway in the initiation, conduct, or support of this research.
have found that clozapine is significantly more likely than Copyright g 2004 ICMPE Figure 1. Schematic Diagram of the Markov Model.
Note: Figure 1 presents a schematic diagram of the Markov model used in this analysis. Health states are in ovals. Arrows represent allowed transitions. Atinception all patients start in the psychosis state. From there, patients can remain psychotic, recover from their psychotic episode, go on to have intolerable sideeffects (i.e., tardive dyskinesia if on a conventional antipsychotic or agranulocytosis if on clozapine), or die.
conventional antipsychotics to improve psychotic episodes dyskinesia.11 Clozapine use has also been associated with and prevent relapse. More recent data from the Clozaril lower rates of suicide attempts and completed suicides.12-15 National Registry in the U.S. has also shown that the Generic forms of clozapine have now become available, incidence of agranulocytosis on clozapine, and fatality further lowering its cost.16 resulting from it, are substantially lower than originally This evidence suggesting greater benefits and potentially feared.10 This has led the FDA to relax its requirement for lower risks and costs for clozapine has led to the question of weekly white blood cell (WBC) monitoring, somewhat whether the indications for clozapine should be expanded to reducing the costs associated with clozapine therapy.
include use as a possible first-line agent in treatment- Clozapine has been shown to be relatively free of the sensitive patients.14,17 Unfortunately, formal analyses of this extrapyramidal side effects associated with conventional question are lacking. Cost-effectiveness analyses performed antipsychotics, and in fact may be a treatment for tardive to date have generally been limited to comparisons of P.S. WANG ET AL.
Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) clozapine vs. conventional antipsychotics among exclusively By contrast, the clozapine-first strategy consisted of the treatment-resistant populations.18-29 These have consistently shown clozapine to have favorable cost-effectiveness ratios initiating treatment with clozapine; for treatment-resistant patients.
Our first aim was to assess whether clozapine should be a antipsychotics if the patient fails to recover to the point possible first-line treatment for schizophrenia, relative to of being dischargeable from the hospital, or relapses restricting clozapine for only patients who have failed two trials of other antipsychotics. Using data from randomized controlled trials and epidemiologic studies, we modeled theclinical and economic outcomes of these two strategies in a We also examined a third strategy that consisted of using hypothetical cohort of patients with schizophrenia only conventional antipsychotics (i.e., never using clozapine undergoing an acute psychotic episode. Because clozapine is even in treatment-resistant patients). We employed this currently underutilized even among treatment-resistant conventional antipsychotic only strategy in subanalyses to patients,30 we also compared strategies using clozapine to the estimate the cost-effectiveness of strategies using clozapine common strategy of never using it.
in mental health care systems which do not utilize clozapine Finally, it is important to distinguish these aims from a and offer only conventional antipsychotics to treatment- question regarding clozapine that this study was not designed resistant patients.
to answer. Newer atypical antipsychotics (e.g., risperidone, We then developed a Markov model35 of transitional olanzapine, and quetiapine) are now widely available and probabilities in which patients could occupy one of seven used,31 raising the question of whether these newer agents or types of health states: acute psychosis while taking clozapine should be the preferred first-line treatment in clozapine, acute psychosis while taking conventional schizophrenia. Trials completed to date indicate the antipsychotic, recovered with clozapine, recovered with possibility that clozapine may have greater efficacy.15,32 conventional antipsychotic, serious TD with conventional However, definitively answering what should be the antipsychotic, agranulocytosis with clozapine, or dead. We preferred first-line treatment will have to wait until sufficient then conducted a cohort simulation to track transitions head-to-head trials comparing clozapine to newer atypical between states representing the expected clinical and antipsychotics in treatment-sensitive patients, such as the economic effects in patients hypothetically randomized to recent InterSePT trial,15 are completed.
one of the three antipsychotic strategies (see Figure 1). Acycle length of 3 months was used. We adopted the societalperspective in our analysis. The model was constructed using the DATA 3.5 decision analysis program (TreeAgeSoftware; Williamstown, MA).
Data Analytic Procedures Transition Probabilities Each state of the model has a mortality rate associated with it, We developed a computer model based on all available data which in turn determines the probability of death in any from the clinical literature to estimate the outcomes of three given cycle (see Table 1). The mortality hazard for each state treatment strategies in a hypothetical 30-year old patient with was calculated by combining the effects of suicide while on a schizophrenia hospitalized with an acute psychotic episode specific antipsychotic regimen, death by agranulocytosis (30 years of age was chosen, rather than younger, to while on clozapine, and age-adjusted all-cause mortality rates accommodate the later ages of onset in women)33,34. The from U.S. life tables.36 To reproduce the life-expectancies clozapine-third strategy described the likely outcomes of observed in published record linkage studies of treated using clozapine only after a patient had failed 2 trials of schizophrenia patients,37 we multiplied age-adjusted all- conventional antipsychotics; it consisted of the following: cause mortality rates in U.S. life tables by a factor of four.
We modeled death due to suicide or agranulocytosis as initiating treatment with a conventional antipsychotic; additive effects to age-adjusted mortality. The rate of death (ii) switching to a second conventional antipsychotic if the by suicide while on conventional antipsychotics was derived patient fails to recover to the point of being from published record-linkage data for treated patients with dischargeable from the hospital, or relapses after schizophrenia.37 The rate of death by suicide while taking recovery, or develops serious tardive dyskinesia (TD) clozapine was obtained from registry data on the cumulative on the first conventional antipsychotic; incidence of suicide among 51,333 patients.13 The rate of (iii) following this, switching to clozapine if the patient fails death from agranulocytosis while taking clozapine was to recover to the point of being dischargeable from the obtained from Clozaril National Registry data from 1990- hospital, or relapses after recovery, or develops serious 1994.10 For sensitivity analyses, we used the upper and lower TD on the second conventional antipsychotic; bounds of the 95% confidence interval around the estimates (iv) switching back to a conventional antipsychotic if the of the rates of death by suicide and agranulocytosis.
patient fails to recover to the point of being Data on the effectiveness of clozapine vs. conventional dischargeable from the hospital, or relapses after antipsychotics in treatment-sensitive and treatment-resistant recovery, or develops agranulocytosis on clozapine.
patients with schizophrenia were obtained from a recent SHOULD CLOZAPINE CONTINUE TO BE RESTRICTED TO THIRD-LINE STATUS? Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) Table 1. Assumptions Governing Transition Probabilities, Costs, and Utilities for the Markov Model 3-Month Probabilites on Conventional Antipsychotic Psychosis to recovered Recovered to relapse 3-Month Probabilities on Clozapine Psychosis to recovered Recovered to relapse No agran. to agran.
Agranulocytosis to death Clozapine (425 mg/d) Haloperidol (15 mg/d) Inpatient psych. hospital.
Residential treatment Quality Weights Assigned to Health States Recovered from psychosis Actively psychotic Discount Rate (annual) meta-analysis of all data from randomized controlled trials.8 upper and lower bounds of the 95% confidence interval To estimate the probability of recovery from psychosis under (C.I.) around each estimate of effectiveness from the meta- different regimens, we used data on the end point of analysis were used in sensitivity analyses of the probabilities dischargeability from the hospital at the end of short-term of recovery and relapse. We conservatively used only the trials. Although clozapine demonstrated greater effectiveness probability of developing the most serious extrapyramidal on the endpoint of clinical improvement (based on change in symptoms such as TD,38 without considering the symptom severity rating scores), we used dischargeability probabilities of developing other extrapyramidal symptoms from the hospital because it may be a better proxy for such as Parkinsonism, dystonias, or akathisia.
clinically meaningful improvement, and is likely to produce In estimating the proportion of recovered patients who a more conservative estimate of the relative effectiveness of would relapse within 3 months if their antipsychotic was withdrawn due to side effects, we used a published The meta-analysis was also the source of data on the short- estimate39; its 95% CI was used in sensitivity analyses.
antipsychotics to prevent relapse of recovered patients.8 The P.S. WANG ET AL.
Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) The 3-month cost of WBC monitoring was based on a To measure the health effects of strategies, we started by published estimate for weekly WBC testing.6 Weekly WBC recording the unadjusted life expectancies (i.e., life years) testing was assumed rather than the currently recommended associated with each (see Table 1). We then calculated our monitoring strategy because our estimates of the occurrence main measure of the health effects of each strategy in terms and fatality from agranulocytosis derive from a time when of quality-adjusted life years (QALYs). QALYs are weekly WBC testing was required. This assumption also calculated by multiplying the life years spent in specific creates a more conservative estimate of the cost of the health states by the quality of life weights associated with clozapine-first strategy if agranulocytosis is equally well those states. We used published quality-of-life weights for detected by WBC monitoring every two weeks. The cost of the health states in schizophrenia, derived from standard treating agranulocytosis is based on published estimates and gambles, rating scales, and paired comparison questions40 includes hospitalization costs.45 (see Table 1 for the actual quality weights assigned to health In an attempt to manage TD, clinicians frequently try states). The decrement in quality of life assigned to TD was pharmacologic treatments which may or may not be estimated from standard gambles and rating scales carried successful;43 a cost for 3 months of pharmacologic treatment out among patients with schizophrenia.25 Since no quality of (e.g., benztropine) was assigned to those who developed TD life weight for agranulocytosis has been calculated, we as a result of conventional antipsychotic use.46 employed values assigned to severe infection while To define a range for the sensitivity analyses of each cost immunosuppressed from cancer chemotherapy;41 to the estimate, we subtracted 25% of the cost for the lower bound extent that agranulocytosis is a less serious clinical condition, and added 25% for the upper bound.
this biased our results against the clozapine-first strategy. To Probabilistic Sensitivity Analyses define a range for sensitivity analyses, we subtracted 25% to In addition to conducting sensitivity analyses on individual estimate a lower bound and added 25% for the upper bound.
variables, we used Monte Carlo simulation to vary transition We also conducted a sensitivity analysis in which health probabilities, costs, and utilities simultaneously.47 A states were not adjusted for such quality-of-life differences.
probability distribution was created for each variable on the basis of the 95% CI or other range used in 1-way sensitivityanalyses described above. New values from each probability In the base case, we included the following direct medical distribution were randomly selected during each of 1000 costs for each strategy (see below for sources): hospitalization iterations, and cost and effectiveness of each strategy were for psychotic episodes, outpatient care, residential treatment costs, and antipsychotic medication costs (see Table 1). To beconservative, we assigned to clozapine users who developed agranulocytosis the expense of hospitalization for this We discounted all costs and health effects at an annual rate of condition. All clozapine users also were assigned a weekly 3% for the base case, with sensitivity analyses performed cost for WBC monitoring. Conventional antipsychotic users between 0% and 5%.
who developed serious TD experienced the cost ofpharmacologic treatment for this side effect. We inflated allcosts to 1999 U.S. dollars using the medical care component of the Consumer Price Index (CPI-M).42 To calculate drug costs, our base case assumed that Base-Case Analysis clozapine users took 425 milligrams daily while conventionalantipsychotic users took 15 mg of haloperidol daily; these For 30-year-old patients with schizophrenia, the undiscounted dosages were chosen because they correspond to the median annual costs under the clozapine-first, the clozapine-third, and of recommended therapeutic dose ranges.1,2,43 We calculated the conventional antipsychotics only strategies were $26,650, the costs of the drugs based on the average wholesale price $26,640, and $26,530, respectively. The undiscounted life of generic clozapine and generic haloperidol.16 expectancies under these three strategies were 31.03, 31.01, Psychiatric hospitalization costs are for a stay of mean and 30.92 years, respectively.
length (23 day) and based on the average cost for all stays in After discounting at 3% annually, the total discounted costs inpatient mental health settings in the U.S. as identified by for the clozapine-first, clozapine-third, and conventional the Inventory of Mental Health Organizations and General antipsychotic only strategies were $514,100, $513,800, and Hospital Mental Health Services.44 Three-month outpatient $509,200, respectively (see Table 2, column 1). After and residential treatment costs were obtained by first discounting and quality-adjusting for time spent in each calculating the average number of units of treatments and health state, the three strategies yielded 14.59, 14.58, and services used by patients with schizophrenia followed for a 14.51 discounted quality-adjusted life years (QALYs)(see 2-year period;21 average numbers of units were then Table 2, column 2).
multiplied by median costs per unit identified from Thus under base case assumptions, our main finding is that psychiatric records and administrative (Medicaid) data.23 the hypothetical strategy of using clozapine first versus the Hospitalized patients were assigned outpatient and residential currently approved strategy of using clozapine third, would treatment costs for the proportion of 3-month cycles during cost $24,100 per additional QALY gained (see Table 2, which they were not inpatients.
column 3). Results of our subanalyses exploring the cost- SHOULD CLOZAPINE CONTINUE TO BE RESTRICTED TO THIRD-LINE STATUS? Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) Table 2. Base Case Results Life Years (QALYs)* Conventional Only Conventional only * QALY = quality adjusted life year effectiveness of strategies in settings where the current care failure, and relapse on clozapine, as well as freedom from involves only the use of conventional agents, even for extrapyramidal side effects. These benefits outweigh the rare treatment-resistant patients, are shown in Table 2, column 4.
adverse consequences caused by agranulocytosis. The gains The clozapine-first vs. conventional only strategy would cost seen in quality-adjusted life expectancy are obtained at the $58,000/QALY; the clozapine-third vs. conventional only acceptable costs of $24,100/QALY. This cost-effectiveness strategy would cost $64,400/QALY.
ratio is comparable to the ratios for many other commonlyaccepted medical interventions.48 In health care systems Sensitivity Analyses where clozapine is currently never used even for treatment-resistant patients, employing it as a first-line agent also 1-Way Sensitivity Analyses appears to lead to modest gains in QALYs at a fairly In one-way sensitivity analyses (see Table 3), the cost- reasonable cost.
effectiveness ratio comparing clozapine-first vs. clozapine- In all sensitivity analyses, using clozapine first continued third strategies was most sensitive to estimates of the rate of to result in slightly higher quality-adjusted life expectancy recovery from acute psychotic episodes on conventional when compared to using clozapine only after failure of two antipsychotics and on clozapine; assigning the value most conventional agents. The cost-effectiveness ratio of the unfavorable to the clozapine-first strategy for either of these clozapine-first vs. the clozapine-third strategy was most yielded ratios over $30,000/QALY. The analysis was also sensitive to estimates of the efficacy of clozapine and sensitive to the quality of life weight assigned to being conventional antipsychotics in producing recovery from recovered from psychosis, rates of TD on conventional acute psychotic episodes. For this reason, it is noteworthy antipsychotics, the cost of clozapine, the cost of inpatient that we used dischargeability from the hospital as our hospitalizations, the discount rate, relapse rates on clozapine measure of recovery,8 rather than improvement in symptom and conventional antipsychotics, and the cost of residential severity as in earlier studies.18-29 Using this endpoint treatment, in descending order. The cost-effectiveness ratio underestimates the advantage of clozapine over conventional was relatively insensitive to all remaining variables.
antipsychotics because it does not account for the greaterfrequency of partial improvements observed for clozapine vs.
Probabilistic Sensitivity Analysis When transition probabilities, costs, and utilities were The analysis was also sensitive to the quality of life weight allowed to vary simultaneously in the Monte Carlo assigned to recovery from psychosis. We conservatively simulation, the 25,th 50,th and 75th percentiles of the cost- assumed that recovery from psychosis led to the same effectiveness ratio of the clozapine-first compared with the quality-of-life with either clozapine or conventional clozapine-third strategy were $16,700, $23,500, and $31,100 antipsychotics, even though other investigators have found per QALY gained, respectively.
greater preferences in standard gambles25 as well as higherpatient satisfaction ratings8 for clozapine as compared withconventional agents.
The probability of developing TD on conventional antipsychotics was another important parameter. Again, it is The FDA originally restricted clozapine use to only patients important to point out that we conservatively modeled TD to who failed two other antipsychotics, out of concern that be a temporary condition lasting only 3 months rather than as using clozapine as a first-line agent would lead to greater loss the chronic form often encountered in clinical practice; in of life, largely through death from agranulocytosis.1-5 addition, we conservatively only considered tardive Contrary to this view, we found that employing clozapine as dyskinesia but not other extrapyramidal side effects from a first-line agent was no more hazardous and may actually lead to small gains in life expectancy and quality-adjusted symptoms, dystonias or akathisia.11,39,43,49 Both assumptions life expectancy relative to waiting for two failures. This are likely to have biased our results against the clozapine- results from decreased likelihoods of suicide, treatment first strategy.
P.S. WANG ET AL.
Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) Table 3. One-Way Sensitivity Analyses of the Most Influential Parameters on the Cost- Effectiveness Ratio Comparing Clozapine-First to Clozapine-Third Strategies Prob. Recovering On Conventional Antipsychotic (%) Utility of Being Recovered Prob. Recovering on Clozapine (%) Prob. TD on Conventional Antipsychotic (%) Cost of Clozapine (425 mg/d) ($) Cost of Inpatient Psychiatric Hospitalization ($) Discount Rate (%) Prob. Relapse on Clozapine (%) Prob. Relapse on Conventional Antipsychotic (%) Cost of Residential Treatment ($) The fact that the model's results varied with the cost of atypical antipsychotic olanzapine rather than conventional clozapine is also important in light of clozapine's recent neuroleptics).15 Furthermore, our sensitivity analyses patent expiration and the availability of generic products.
indicate that the decision to use clozapine as a first-line agent The price of generic clozapine has been declining and should does not depend critically on a benefit on suicide for continue to come down in the future, potentially making the use of clozapine as a first-line agent more economical. We Second, even estimates drawn from randomized clinical also assumed weekly WBC monitoring for clozapine in our trials may have limited internal and external validity. For base case analysis, rather than the current practice of example, medication non-compliance in clinical trials can monitoring every 2 weeks after 6 months. All of these attenuate the apparent efficacy of antipsychotic medications assumptions are likely to have biased our results against the and affect cost-effectiveness analyses in unpredictable ways.51 Furthermore, any attenuation due to non-compliance To our knowledge, this is the first study to quantitatively may have been differential between clozapine vs.
compare the strategy of using clozapine as a first-line agent conventional antipsychotics, because weekly WBC to the current strategy of restricting clozapine use to only monitoring required of clozapine patients may enhance those who have failed at least two trials of other compliance. Non-compliance with WBC monitoring could antipsychotics. Prior analyses have generally compared the itself lower costs and effectiveness of clozapine regimens; the net impact on cost-effectiveness ratios, while uncertain, antipsychotics for the treatment of patients with exclusively may be lessened by recent reductions in FDA monitoring treatment-resistant schizophrenia. While these have requirements. The generalizability of estimates from clinical consistently shown that clozapine is cost-effective for trials to ‘‘real-world'' practice is also unknown, due to the treatment-resistant patients,18-29 they have not addressed the greater likelihood of non-compliance in typical settings.
larger question of managing treatment-sensitive as well as Third, to avoid ‘‘state explosion'', the model does not completely describe the range of health states that could be These results have some limitations. First, some estimates experienced by patients with schizophrenia. For example, were drawn from observational studies, introducing the clozapine and conventional antipsychotic users can possibility of confounding bias. For example, we used experience other side effects from their medications that were observational data12-14 to estimate the suicide rates on not modeled as health states in our analyses.43 However, our clozapine and conventional neuroleptics because sample model does capture those side-effects that have been clearly sizes in clinical trials or even meta-analyses were too small to established and are most burdensome (due to their generate stable estimates. Such estimates could be prevalence and clinical seriousness). Furthermore, we confounded by the fact that current clozapine users (who had conservatively did not model several potential benefits of to comply with weekly WBC monitoring to receive the clozapine, including improvements in: psychotic symptom drug), may be in a more stable clinical phase than either past severity that may be clinically important but not enough to users or non-users.50 However, one recent randomized trial lead to discharge from the hospital;8 cognitive function;52 found significantly reduced suicidal behavior among clozapine users (although it compared clozapine to the newer SHOULD CLOZAPINE CONTINUE TO BE RESTRICTED TO THIRD-LINE STATUS? Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) Fourth, it was not possible to construct 95% confidence 7. Kane J, Honigfeld G, Singer J, Meltzer H. Clozaril Collaborative Study intervals for some of our parameter estimates due to the lack Group. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry 1988; 45: of information on their underlying variability. While we relied on conventions commonly employed under such 8. Wahlbeck K, Cheine M, Essali A, Adams C. Evidence of clozapine's circumstances (e.g., adding and subtracting 25% of base case effectiveness in schizophrenia: a systematic review and meta-analysis of estimates) to construct ranges for sensitivity analyses, we do randomized trials. Am J Psychiatry 1999; 156: 990-999.
not know how such ranges would relate to true 95% 9. Lieberman JA, Phillips M, Gu H, Stroup S, Zhang P, Kong L, Ji Z, Koch G, Hamer RM. Atypical and conventional antipsychotic drugs in confidence limits. Furthermore, while our Markov model treatment-naı¨ve first episode schizophrenia: a 52-week randomized trial assumes constancy in parameter estimates, it is possible (but of clozapine vs. chlorpromazine. Neuropsychopharmacology 2003; 28: unknown) that transition probabilities, costs, and utilities change with time and the occurrence of clinical events. We 10. Honigfeld G, Arellano F, Sethi J, Bianchini A, Schein J. Reducing believe the ranges employed in sensitivity analyses are clozapine-related morbidity and mortality: 5 years of experience with theClozaril National Registry. J Clin Psychiatry 1998; 59(suppl 3): 3-7.
sufficiently broad to contain the average in parameters 11. Lieberman JA, Saltz BL, Johns CA, Pollack S, Borenstein M, Kane J.
experienced over the time horizon, however we cannot be The effects of clozapine on tardive dyskinesia. Br J Psychiatry 1991; 158: 503-510.
Finally, clinicians and patients still face a critically 12. Walker AM, Lanza LL, Arellano F, Rothman KJ. Mortality in current and former users of clozapine. Epidemiology 1997; 8: 671-677.
important question about clozapine that this study cannot 13. Meltzer HY, Fatemi H. Suicide in schizophrenia: the effect of clozapine.
answer: what should the preferred first-line treatment be in Clin Neuropharmacol 1995; 18: S18-S24.
schizophrenia? This question has become more pressing as 14. Meltzer HY, Okayli G. Reduction of suicidality during clozapine newer atypical agents become widely available and used.31 treatment of neuroleptic-resistant schizophrenia: impact on risk-benefit Unfortunately, we were unable to answer this question assessment. Am J Psychiatry 1995; 152: 183-190.
15. Meltzer HY, Alphs L, Green AI, Altamura A, Anand R, Bertoldi A, because there are insufficient head-to-head trials comparing Chouinard G, Islam MZ, Kane J, Krishnan R, Lindenmayer JP, Potkin S.
clozapine to newer atypical antipsychotics in treatment- Clozapine treatment for suicidality in schizophrenia: International sensitive patients.54-56 Trials completed to date leave open Suicide Prevention Trial (InterSePT). Arch Gen Psychiatry 2003; 60: 82- the possibility that clozapine may possess superior efficacy 16. Drug Topics Red Book. Medical Economics Company, Montvale, NJ.
on several important clinical outcomes.15,32 Clearly, additional data from large trials comparing clozapine to 17. Meltzer HY. Suicide in schizophrenia: risk factors and clozapine newer atypical agents in treatment sensitive patients are treatment. J Clin Psychiatry. 1998; 59 (Suppl 3): 15-20.
needed in the future. With publication of such results, the 18. Rosenheck R, Cramer J, Xu W, Thomas J, Henderson W, Frisman L, structure of the present analysis can be updated to shed light Fye C, Charney D. A comparison of clozapine and haloperidol inhospitalized patients with refractory schizophrenia. N Engl J Med 1997; on this critically important question.
337: 809-815.
In summary, we found that using clozapine as a first-line 19. Rosenheck R, Cramer J, Allan E, Erdos J, Frisman LK, Xu W, Thomas J, agent vs. as a third-line agent, rather than leading to loss of Henderson W, Charney D. Cost-effectiveness of clozapine in patients life, actually leads to small gains in life-expectancy and with high and low levels of hospital use. Arch Gen Psychiatry 1999; 56:565-572.
quality-adjusted life-expectancy at the acceptable cost of 20. Rosenheck R, Cramer J, Xu W, Grabowski J, Douyon R. Thomas J, $24,100/QALY. While results from this study do not shed Henderson W, Charney D. Multiple outcome assessment in a study of light on whether clozapine should be the preferred first-line the cost-effectiveness of clozapine in the treatment of refractory strategy, they do suggest that clozapine should not schizophrenia. Health Serv Res 1998; 33: 1237-1261.
necessarily be confined to its role as a third-line agent. Both 21. Meltzer HY, Cola P, Way L, Thompson PA, Bastani B, Davies MA, Snitz B. Cost effectiveness of clozapine in neuroleptic-resistant clinical and economic reasons appear to justify its addition to schizophrenia. Am J Psychiatry 1993; 150: 1630-1638.
the armamentarium of possible treatments for treatment- 22. Honigfeld G, Patin J. A two-year clinical and economic follow-up of sensitive as well as treatment-resistant schizophrenia.
patients on clozapine. Hosp Commun Psychiatry 1990; 41: 882-885.
23. Revicki DA, Luce BR, Weschler JM, Brown RE, Adler MA. Cost- effectiveness of clozapine for treatment- resistant schizophrenic patients.
Hosp Commun Psychiatry 1990; 41: 850-854.
24. Jonsson D, Wa˚linder J. Cost-effectiveness of clozapine treatment in therapy-refractory schizophrenia. Acta Psychiatr Scand 1995; 92: 199-201.
1. Lehman AF, Steinwachs DM. Translating research into practice: the 25. Glennie JL. Pharmacoeconomic evaluations of clozapine in treatment- Schizophrenia Patient Outcomes Research Team (PORT) treatment resistant schizophrenia and risperidone in chronic schizophrenia.
recommendations. Schizophrenia Bull 1998; 24: 1-10.
Ottawa, Canada: Canadian Coordinating Office for Health Technology 2. American Psychiatric Association. Practice guideline for the treatment Assessment (CCOHTA), July 1997.
of patients with schizophrenia. Washington, D.C.: American Psychiatric 26. Palmer CS, Revicki DA, Genduso LA, Hamilton SH, Brown RE. A cost- Association, 1997.
effectiveness clinical decision analysis model for schizophrenia. Am J 3. Physicians' Desk Reference. 54th Edition. Medical Economics Company, Managed Care 1998; 4: 345-355.
Montvale, NJ. 2000.
27. Davies LM, Drummond MF. Assessment of costs and benefits of drug 4. Griffith RW, Saameli J. Clozapine and agranulocytosis. Lancet 1975; 2: therapy for treatment-resistant schizophrenia in the United Kingdom. Br J Psychiatry 1993; 162: 38-42.
5. Alvir JMJ, Lieberman JA, Safferman AZ, Schwimmer JL, Schaaf JA.
Clozapine-induced agranulocytosis. Incidence and risk factors in the Pharmacoeconomic evaluation of clozapine in treatment-resistant United States. N Engl J Med 1993; 329: 162-167.
schizophrenia: a cost-utility analysis. Can J Clin Pharmacol 2001; 8: 6. Meltzer HY, Cola PA. The pharmacoeconomics of clozapine: a review. J Clin Psychiatry 1994; 55:161-165.
29. Essock SM, Frisman LK, Covell NH, Hargreaves WA. Cost- P.S. WANG ET AL.
Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004) effectiveness of clozapine compared with conventional antipsychotic 43. Marder SR. Antipsychotic medications. In: Schatzberg AF, Nemeroff medication for patients in state hospitals. Arch Gen Psychiatry 2000; 57: CB, eds. Textbook of psychopharmacology, 2nd edition. American Psychiatric Press, Wash., D.C. 1998.
44. Weiden PJ, Olfson M. Cost of relapse in schizophrenia. Schizophr Bull schizophrenia. Schizophr Bull 1997; 23: 663-674.
1995; 21: 419-429.
31. Wang PS, West JC, Tanielian T, Pincus HA. Recent patterns and 45. Zhang M, Owen RR, Pope SK, Smith GR. Cost-effectiveness of predictors of antipsychotic medication regimens used to treat clozapine monitoring after the first 6 months. Arch Gen Psychiatry 1996; schizophrenia and other psychotic disorders. Schizophr Bull 2000; 26: 53: 954-958.
46. Glazer WM, Ereshefsky L. A pharmacoeconomic model of outpatient 32. Azorin JM, Spiegel R, Remington G, Vanelle JM, Pere JJ, Giguere M, antipsychotic therapy in revolving door schizophrenic patients. J Clin Bourdeix I. A double-blind comparative study of clozapine and Psychiatry 1996; 57: 337-345.
risperidone in the management of severe chronic schizophrenia. Am J 47. Doubilet P, Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabilistic Psychiatry 2001; 158: 1305-1313.
sensitivity analysis using Monte Carlo simulation: a practical approach.
33. Reicher A, Maurer K, Loffler W, Fatkenheuer B, An Der Heiden W, Med Decis Making 1985; 5: 157-177.
Munk-Jorgensen P, Stromgren E, Hafner H. 1991. Gender differences in 48. Graham JD, Corso PS, Morris JM, Segui-Gomez M, Weinstein MC.
the age of onset and course of schizophrenic disorders. In: Hafner H, Evaluating the cost-effectiveness of clinical and public health measures.
Gattaz WF, eds. Search for the causes of schizophrenia. Berlin: Ann Rev Pub Health. 1998; 19: 125-152.
Springer, pp. 14-33.
49. Sachdev P. The epidemiology of drug-induced akathisia: Part II.
34. Bromet EJ, Dew MA, Eaton WW. 2002. Epidemiology of psychosis Chronic, tardive, and withdrawal akathisias. Schizophr Bull. 1995; 21: with special reference to schizophrenia. In: Tsuang MT, Tohen M, eds.
Textbook in psychiatric epidemiology. New York: John Wiley, pp. 365- 50. Sernyak M, Desai R, Stolar M, Rosenheck R. Impact of clozapine on completed suicide. Am J Psychiatry 2001; 158: 931-937.
35. Sonnenberg FA, Beck JR. Markov models in medical decision making: a 51. Hughes DA, Bagust A, Haycox A, Walley T. The impact of non- practical guide. Med Decis Making 1993; 13: 322-338.
compliance on the cost-effectiveness of pharmaceuticals: a review of the 36. National Center for Health Statistics. Vital statistics of the United States, literature. Health Economics 2001; 10: 601-615.
1995. Volume II, Mortality, Part A, Section 6: life tables. DHHS 52. Hagger C, Buckley P, Kenny JT, Friedman L, Ubogy D, Meltzer HY.
Publication No (PHS) 98-1147. Hyattsville, MD: US Dept. HHS, 1998.
Improvement in cognitive functions and psychiatric symptoms in 37. Newman SC, Bland RC. Mortality in a cohort of patients with treatment-refractory schizophrenic patients receiving clozapine. Biol schizophrenia: a record linkage study. Can J Psychiatry 1991; 36: 239- Psychiatry 1993; 34: 702-712.
53. Palmer CS, Revicki DA, Halpern MT, Hatziandreu EJ. The cost of 38. Kane JM, Woerner M, Borenstein M, Wegner J, Lieberman J. Integrating suicide and suicide attempts in the United States. Clin Neuropharmacol incidence and prevalence of tardive dyskinesia. Psychopharmacol Bull 1995; 18: S25-S33.
1986; 22: 254-258.
54. Klieser E, Lehmann E, Kinzler E, Wurthmann C, Heinrich K.
39. Baldessarini RJ, Viguera AC. Neuroleptic withdrawal in schizophrenic Randomized, double-blind, controlled trial of risperidone versus patients. Arch Gen Psychiatry 1995; 52: 189-192.
40. Revicki DA, Shakespeare A, Kind P. Preferences for schizophrenia- Psychopharmacol 1995; 15 (1 suppl 1): 45S-51S.
related health states: a comparison of patients, caregivers and 55. Bondolfi G, Dufour H, Patris M, May JP, Billeter U, Eap CB, Baumann psychiatrists. Internat Clin Psychopharmacol 1996; 11: 101-108.
41. Weeks JC, Tierney MR, Weinstein MC. Cost-effectiveness of schizophrenia: a randomized double-blind study. Am J Psychiatry 1998; prophylactic intravenous immune globulin in chronic lymphocytic 155: 499-504.
leukemia. N Engl J Med 1991; 325: 81-86.
56. Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical 42. Gold MR, Siegel JE, Russell LB, Weinstein MC, eds. Cost-effectiveness antipsychotics in the treatment of schizophrenia: systematic overview in health and medicine. New York: Oxford University Press, 1996.
and meta-regression analysis. Br Med J 2000; 321: 1371-1376.
SHOULD CLOZAPINE CONTINUE TO BE RESTRICTED TO THIRD-LINE STATUS? Copyright g 2004 ICMPE J Ment Health Policy Econ 7, 77-85 (2004)

Source: http://www.icmpe.org/test1/journal/issues/v7pdf/7-077_text.pdf