How to compare doses of different antipsychotics: A systematic review of methods

Schizophrenia Research 149 (2013) 141–148

Contents lists available at SciVerse ScienceDirect

Schizophrenia Research journal homepage: www.elsevier.com/locate/schres

How to compare doses of different antipsychotics: A systematic review of methods Maxine X. Patel a,⁎, Ioli A. Arista a, Mark Taylor b, Thomas R.E. Barnes c a b c

Institute of Psychiatry, King's College London, Dept of Psychosis Studies PO68, 16 DeCrespigny Park, London SE5 8AF, UK Intensive Home Treatment Team, Henderson Unit, Royal Edinburgh Hospital, Edinburgh EH10 5HF, UK Imperial College, Centre for Mental Health, Faculty of Medicine, The Claybrook Centre, 37 Claybrook Road, London W6 8LN, UK

a r t i c l e

i n f o

Article history: Received 20 March 2013 Received in revised form 20 May 2013 Accepted 14 June 2013 Available online 8 July 2013 Keywords: Antipsychotics Dose Efficacy Side effects Equivalence Schizophrenia

a b s t r a c t Background: The ability to calculate equivalent dosage is important when comparing or switching between doses of different antipsychotics in the treatment of schizophrenia. It is also necessary when designing antipsychotic comparator trials which control for dosage. Method: A systematic review to identify and critically evaluate the methods available for the estimation of antipsychotic dose equivalence was conducted. Electronic searches were carried out using Medline and PubMed and additional information was requested from pharmaceutical companies. The identified methods were evaluated against specific criteria regarding scientific rigour, quality of source data underpinning the method, clinical applicability and utility. Results: Eleven articles were identified that described methodologies for antipsychotic dose equivalence. Seven of these referred to calculated methods, including chlorpromazine equivalence, maximum dose and daily-defined dose, and relied on an evidence base from both fixed and flexible dosing data. The remaining four described consensus methods which were based on the knowledge and experience of experts. Chlorpromazine was used as the standard comparator drug in the majority of the calculated equivalence studies, whereas risperidone was used for most consensus methods. Conclusions: Comparison of methods for calculating antipsychotic dose equivalence suggests that different methods yield different equivalencies and the evidence is not sufficiently robust for any of these to be considered as a gold standard method. Thus, choice of method may introduce bias, either an over or underestimate of equivalent dosage, when designing head-to-head, antipsychotic, fixed-dose trials. Consequently, clinical trial reports should routinely include justification of the choice of method for calculating dose equivalence. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Antipsychotic medications are commonly used for the pharmacological treatment of schizophrenia and related psychotic illnesses. While all of these medications share dopamine D2 receptor antagonist properties, they also have varying receptor binding profiles. Despite this pharmacological heterogeneity, with the exception perhaps of clozapine, the differences in efficacy are modest and must be weighed against larger differences in liability for particular side effects (McCue et al., 2006; Leucht et al., 2009a,b). Switching between antipsychotics is commonplace in routine clinical practice (Bitter et al., 2008; Taylor et al., 2008). To minimise any disruptive effects and maximise the likelihood of success, the clinician must choose an appropriate target dose for the new antipsychotic. This is perhaps commonly selected on the basis of dose equivalency data (Lambert, 2007) or by titrating to the dose that

⁎ Corresponding author. Tel.: +44 2078485136; fax: +44 2078480572. E-mail addresses: [email protected] (M.X. Patel), [email protected] (M. Taylor), [email protected] (T.R.E. Barnes). 0920-9964/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.schres.2013.06.030

produces maximum effectiveness and then stopping the titration when tolerance to emergent side effects is no longer maintained. An understanding of antipsychotic dosing in terms of equivalent efficacy is also necessary in clinical research. Drug dose comparisons are necessary in pharmacoepidemiological drug utilisation studies and clinical trials. This is particularly true in randomised controlled trials (RCTs) conducted pre-licensing to demonstrate the superiority of a new antipsychotic over placebo and/or non-inferiority to the current ‘gold standard’ comparator antipsychotic, although choice and dose of a comparator antipsychotic can vary between studies. For most head-to-head clinical trials comparing antipsychotic medication, clinically equivalent dosages are chosen in order to control for dosage. For example, interpretation of the relative efficacy of the antipsychotics tested in one large, pragmatic clinical trial, the CATIE study, was partly confounded by the use of relatively high doses of olanzapine and/or relatively low doses of risperidone (Rosenheck et al., 2009). Antipsychotic dose comparison is important both in clinical practice and for research purposes. For example, assessment of the quality of antipsychotic prescribing practice will include the identification of high-dose prescribing and polypharmacy (Paton et al., 2008; Lin et al.,

142

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

2010; Suzuki, 2011). As it is not realistic to expect all antipsychotics to be compared with one another in fixed-dose RCTs and across the various illness phases, a method for calculating or otherwise extrapolating equivalent doses is required. However, the development of valid and reliable methods of dose comparison is yet to be fully realised. Consequently, we aimed to conduct a systematic review to identify and critically evaluate the methods currently available to compare the doses of individual antipsychotic drugs. 2. Method 2.1. Search strategy A systematic electronic search was conducted using Medline and PubMed in March 2012. For Medline, keywords were used, mapped to MeSH headings as appropriate, for the following terms: 1. 'antipsychotic' OR 'neuroleptic'; 2. 'dose' OR 'dosage' OR 'dosing'; 3. 'equivalen*'; 4. 'consensus'. Combinations were then conducted to form term 5 from '1 AND 2' as well as term 6 from '3 OR 4'. The final combination was formed from '5 AND 6'. For PubMed, an identical search strategy was used other than term 3 being constructed with the key words of ‘equivalents’ or ‘equivalency’ or ‘equivalence’. The resultant abstracts were then examined for duplication and independently reviewed by two authors. The primary inclusion criteria were provision of a description of a method for antipsychotic dose comparison in humans and publication in English. If a method for antipsychotic dose comparison was described in several articles, then the earliest article only was included. Where no initial, definitive, agreed decision could be made on the basis of the abstract alone, the full article was examined and disagreement between the two reviewers was resolved with subsequent discussion. In order to check for other eligible methods, six pharmaceutical companies known to be working in the field of psychosis were contacted to request information regarding which dose comparison method, if any, was referred to when selecting the dose(s) of an active comparator drug in antipsychotic phase II/III clinical trials. 2.2. Classification and analysis The methods of antipsychotic dose comparison were classified, described and evaluated. As no suitable pre-existing tool existed, the quality assessment criteria used included the type of method used to derive equivalency, source data on which the method was based, the key comparator (baseline antipsychotic drug and dose e.g. chlorpromazine 100 mg) applicability for different antipsychotics with various formulations and for their full dose range, and generalisability across different clinical presentations. For consensus methods, description of the sample of Experts was also examined. Using Web of Knowledge the citation rate of the original source articles for the selected methods, as of October 2012, was recorded. An additional table was constructed that included the equivalent values to the most commonly used firstgeneration antipsychotics (FGAs) and second-generation antipsychotics (SGAs), including long-acting injections (LAIs). 3. Results 3.1. Search strategy The Medline search produced 484 initial articles, of which 27 met the eligibility criteria. These were fully examined including hand-searching of their reference lists, which yielded a further 3 eligible methods. Of these 30 articles, 9 were found to describe for the first time a unique methodology for comparing doses between antipsychotic medications. A tenth article (of the original 30) reported a comparison of two methods to standardize antipsychotic doses including one not described previously (Rijcken et al., 2003). As this article did not provide sufficient information regarding the method, the source book (WHO, 2012a) describing

the method was then referred to. Searching on PubMed, 22 out of 422 retrieved articles were found to be relevant, but none described an additional methodology to those identified via Medline. One further eligible method was found following communication with pharmaceutical companies. Overall, a total of 11 original methods for comparing antipsychotics doses were found. Eligible methods were grouped as follows: (i) calculated methods including three on chlorpromazine equivalence (Davis, 1974; Woods, 2003; Andreasen et al., 2010), three using maximum dose (Milton et al., 1995; Yorston and Pinney, 1997; Davis and Chen, 2004), one describing daily defined dose (DDD) (WHO, 2012a); and (ii) four consensus methods (Kane et al., 2003; Buckley, 2005; Simpson et al., 2006; Gardner et al., 2010). 3.2. Calculated methods 3.2.1. Chlorpromazine equivalence Davis (1974) pioneered dose equivalence methodology for FGAs, using chlorpromazine as the standard comparator, and utilising source data from double-blind trials comparing chlorpromazine with other FGAs. In all studies providing source data, the optimal clinical response was determined by the study physician, and this was used to estimate empirically the efficacy equivalence between drugs. Chlorpromazine equivalents were then developed referring to the dose of an antipsychotic in mg/day that was as effective as 100 mg/day of chlorpromazine [Table 1]. Subsequently, Woods (2003) developed dose equivalency tables for SGAs based on the methodology of Davis (1974). The minimum effective dose, which is the lowest dose that is significantly superior to placebo, was derived from various placebo-controlled and fixed-dose trials. Initially haloperidol equivalencies were estimated, since the minimum effective dose of haloperidol was reported widely in the studies under examination and was considered to be 4 mg. Haloperidol equivalent doses were then converted into chlorpromazine equivalents based on the assumption that “2 mg of haloperidol equals 100 mg of chlorpromazine” (American Psychiatric Association, 1997). The most recent method for calculating chlorpromazine equivalence was developed by Andreasen et al. (2010). The source data to estimate new dose equivalents were derived from pre-existing consensus guidelines (Kane et al., 2003), where equivalent dosages of FGAs and SGAs to haloperidol or risperidone, respectively, were estimated. From the derived equivalencies it was concluded that “it would probably be possible to generate linear equations to derive equivalency”. Thus, a linear regression analysis was conducted in which the dose equivalents of haloperidol and chlorpromazine were used for equivalent values of the other antipsychotics. 3.2.2. Maximum dose 3.2.2.1. Near-effective maximum dose. is defined as the threshold dose eliciting clinical response with the least adverse profile, and can be calculated from dose–response curves which were constructed using data from fixed-dose randomised placebo-controlled studies (Davis and Chen, 2004) [Fig. 1]. Equivalence between antipsychotics is then established by comparing the near-effective doses. Dose equivalence tables have also been calculated on the basis of the median effective dose producing a response in half of the population (ED50) (Davis and Chen, 2004). For trifluoperazine and fluphenazine, a single near-effective dose was not identified due to the lack of sufficient data from fixed-dose and placebo-controlled studies. The best dose– response curves were said to be constructed for risperidone, and for oral and intramuscular olanzapine. A single near-effective maximum dose was reported for aripiprazole and risperidone, whilst for amisulpride the near-effective dose was found by extrapolation. For eight antipsychotics, there was a range of near-effective values provided with as much as a 4-fold variation between the lower and upper range values for quetiapine.

Table 1 Calculated methods for antipsychotic dose equivalence. Chlorpromazine equivalents Original method

Adapted method for SGAs

Linear equations

Near-effective

Class of antipsychotic Development

FGAs Davis (1974)

SGAs Woods (2003)

SGAs Andreasen et al. (2010)

FGAs and SGAs Davis and Chen (2004)

Methodology

Equally efficacious drugs identified Minimum effective doses (the from source data. Ratios estimated for lowest dose superior to placebo) 100 mg chlorpromazine of equally efficacious drugs selected from source data. Haloperidol equivalents converted into chlorpromazine equivalents Double blind flexible dose studies Fixed dose trials comparing chlorpromazine with FGAs. Efficacy determined by clinician reporting 15 6 238 633 No standardised efficacy measure in Inconsistency for minimum source data effective doses of haloperidol and chlorpromazine

Source data

No. of antipsychotics No. of citations Key Limitation

Maximum dose

FGA: First generation antipsychotics. SGA: Second generation antipsychotics. SmPC: Summary of Product Characteristics. N/A: Not applicable. a Based on details provided by Mace and Taylor (2005). b citation rate for article by Mace and Taylor (2005). c determining the citation rate for daily defined dose for antipsychotics was not possible.

Regression coefficients and equations for equivalents of 2 mg haloperidol and 100 mg chlorpromazine based on source data

Daily defined dose Licensed

FGAs and SGAs Milton et al. (1995); Yorston and Pinney (1997) Construction of dose–response Prescribed dose is divided by curves from source data. The maximum licensed daily dose near-effective maximum dose was and can then be converted into the dose required to elicit a rea percentage sponse with least adverse profile

Experts responses from Consensus reported by Kane et al. (2003)

Randomised placebo controlled studies

14 57 Linearity of dose equivalency was assumed

19 127 Inadequate data to construct dose–response curves for all antipsychotics

The maximum daily dose licensed by regulatory authorities based on pre-licensing trial data 18 a 3b Maximum dose is based on tolerance rather than efficacy

FGAs and SGAs WHO (2012a,b) Defined as assumed average maintenance dose per day per drug for it is main indication in adults

Choice of average dose for each drug is based on SmPC and available literature 56 N/Ac No information about therapeutic efficacy, it is only a technical metric to measure drug consumption

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

Method for dose equivalence

143

144

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

Plateau

Response (%) 100

has been included here as it has been used in efficacy studies which compare low versus medium doses of antipsychotics (Uchida et al., 2011) and in studies comparing methods of dose equivalence (e.g. Rijcken et al., 2003). The DDD is the assumed, average, maintenance dose per day per drug for its main indication in adults (WHO, 2012a). The calculation of the DDD is a compromise based on the dosage recommendations in each drug's description in its ‘summary of product characteristics’ (SmPC) and in the available literature. For the majority of antipsychotics, the upper region of the licensed dose range is used. For example, the DDDs for chlorpromazine and risperidone are 300 mg and 5 mg respectively (WHO, 2012b).

Near- effective maximum dose

Z 75

X 50

25

Y ED50

3.3. Consensus methods

Log Dose Fig. 1. Dose–response curve and dose equivalency estimation based on the near-effective maximum doses*. *If a given single antipsychotic of dose X was more effective than dose Y, it was assumed to be in the linear portion, and if X was equally efficacious with dose Z, it was likely to be on the plateau. If dose X was not superior to placebo, then it was concluded to have near placebo efficacy (adapted from Davis and Chen, 2004).

3.2.2.2. Maximum licensed doses. for different antipsychotics can also be used as a basis for comparing doses and two methods were found. Milton et al. (1995) proposed using a ratio of maximum equivalents, where the prescribed dose is divided by the maximum licensed daily dose as described in the British National Formulary (BNF, Joint Formulary Committee, 2012). If the ratio exceeds 1 it is said to indicate maximum, whilst 0 reflects minimum. Subsequently, Yorston and Pinney (1997) described this ratio as simply a percentage of the maximum licensed dose and further detail was provided by Mace and Taylor (2005). For example, the maximum daily dose of chlorpromazine is 1000 mg, which constitutes 100%BNF. Additionally, dosage sums of two or more antipsychotics can be calculated, allowing total antipsychotic exposure during polypharmacy to be estimated (Mace and Taylor, 2005; Barnes et al., 2006). It has been also shown that %BNF can be used to help identify patients at high risk of toxicity (Yorston and Pinney, 2000). 3.2.3. Defined daily dose (DDD) Although the primary purpose of the DDD system is not to guide therapeutic equivalence but to aid comparison of drug utilisation, it

Kane et al. (2003) described the results of a consensus of 47 experts from the USA who were asked to estimate equivalents of different antipsychotics for fixed doses of haloperidol and risperidone. The experts considered 10 mg of haloperidol to be equivalent to 491.9 mg of chlorpromazine [Table 2]. Buckley (2005) evaluated the opinions of 375 attendees at an American Psychiatric Association AGM. Participants were asked to estimate the equivalent doses of olanzapine, quetiapine, ziprasidone and aripiprazole to 4 mg risperidone. There was a tendency to report higher doses of quetiapine (600 mg) and ziprasidone (150 mg) as equivalent to 4 mg risperidone. Clinicians estimated that 4 mg of risperidone was equivalent to 15 mg of olanzapine (42% of clinicians); 600 mg of quetiapine (51%); 160 mg ziprasidone (51%); 15 mg of aripiprazole (41%). Simpson et al. (2006) developed risperidone equivalents for a study investigating the effects of risperidone LAI. Seven experts evaluated systematic reviews of clinical and preclinical to estimate risperidone equivalencies for 16 antipsychotics. Gardner et al. (2010) reported on the most recently conducted consensus meeting involving 46 participants from 18 countries which was funded by the University, and grants from the Bruce J. Anderson Foundation and the McLean Private Donors Psychopharmacology Research Fund. The dosing equivalencies were estimated using a two-stage Delphi Method. Initially, the participants were asked to estimate equivalencies for 20 mg olanzapine corresponding to a specific clinical scenario. For oral antipsychotics and LAIs, respondents considered a “moderately symptomatic adult man with DSM-IV schizophrenia with N 2 years of antipsychotic treatment and not considered treatment refractory”. For short-acting injections, the reference drug was 5 mg haloperidol and the case scenario was “an adult man with DSM-IV schizophrenia not treated for 2 weeks, presenting with delusions,

Table 2 Consensus methods for antipsychotic dose equivalence. Kane et al., 2003

Buckley, 2005

Simpson et al., 2006

Gardner et al., 2010

Methodology

Open-ended questions to estimate equivalencies for fixed doses of haloperidol and risperidone

Estimated equivalence to 4 mg risperidone for 4 SGAs

Risperidone equivalents were estimated but method not further described

Baseline comparator drug

Haloperidol and risperidone

Risperidone

Risperidone

Clinical indications

Not specified

Not specified

No. of antipsychotics % of agreement

15 90%

4 Not available

Patients with stable presentation of schizophrenia 16 Not available

No. of experts Funding

49 Sponsored by pharmaceutical industry

375 Educational grant from AstraZeneca

Delphi method - part I: estimate equivalency of 20 mg olanzapine for antipsychotics oral and LAIs, and for one 5 mg haloperidol short-acting intramuscular injection. Part II: experts ranked their level of agreement to their ranked responses Olanzapine and haloperidol in clinical scenarios but, later, also calculated for chlorpromazine Case scenarios considered which differed for the formulations 62 90% for LAIs; 83% oral agents; 43% short acting injections 43 Sponsored by non-industry organisations

No. of citations

134

2

a

7 Main study reported in article supported by Janssen. Authors from same company 33a

41

citation of Simpson et al. (2006) is more likely to reflect citation of the study outlined in the article rather than the dose comparison method.

Table 3 Summary of dose equivalents from calculated and consensus methods (doses or ratios can only be compared within a single column)a. Method for dose equivalence Maudsley Guidelinesb CPZ equivalent

Daily defined dose Consensus

Original

Linear equations Near-effective

Licensed (%BNF) -

Kane et al. (2003) Buckley (2005) Simpson et al. (2006) Gardner et al. (2010)

Unit

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Dose (mg/day)

Ratioa

Baseline comparator

N/A

Chlorpromazine Chlorpromazine 100 mg 100 mg

N/A

N/A

N/A

Risperidone 4 mg

Risperidone 4 mg

Risperidone 4 mg

Chlorpromazine dose ratio

FGA Chlorpromazine Haloperidol Fluphenazine Perphenazine Prochlorperazine Trifluoperazine Sulpiride Zuclopenthixol

300 N4 – – – 15 800 –

100 2 1.2 8.9 14.3 2.8 – –

100 1.6 1.8 6.8 – 5.7 – –

400–450 3.5–10 b6.9 – – 10–15 – –

1000 30 – – – 50 2400 150

300 8 10 30 100 20 800 30

350 7.5 7.5 24 – 15 – –

– – – – – – – –

500 10 10 50 25 30 – 20

1 60 50 20 6.9 30 0.75 12

SGA Amisulpride Aripiprazole Paliperidone Olanzapine Quetiapine Risperidone Sertindole Ziprasidone Zotepine

800 10 – 10 300 3 12 80 –

– 7.5 – 5 75 2 – 60 –

– 8 – 5.3 175.5 1.2 – 62.6 –

200 10 – N16 150–600 4 12–20 80–160 –

1200 30 – 20 750 16 – – 300

400 15 6 10 400 5 16 80 200

– 15 – 15 450 4 – 120 –

– 15 – 15 600 4 – 160 –

– – – 15 600 4 – 120 –

0.9 20.0 66.7 30 0.80 100 30 3.8 2

Clozapine

–

–

138.8

N400

900

300

350

–

–

–

LAI Fluphenazine decanoate Haloperidol decanoate

– –

– –

7.7 44.2

50 mg/wk 75 mg/wk

1/day 3.3/day

25 mg/2–3 wk 100 mg/2–3 wk

– –

– –

– –

Risperidone LAI

–

–

–

25 mg/2 wk 100–200 mg/ mth 50 mg/mth

25 mg/wk

2.7/day

–

–

–

–

Doses or ratios cannot be compared across a row as the baseline comparator varies as does the unit

b

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

a

Maximum dose

Minimum effective

Maudsley Prescribing Guidelines: minimum effective dose for relapse exacerbation.

145

146

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

auditory hallucinations, agitation, poor cooperation, threatening behaviour and who is refusing oral medication”. In the second stage of the process, the participants were asked to rank their level of agreement for the averaged responses, derived from the first stage. Equivalency ratios were later calculated for chlorpromazine. 3.4. Equivalence values Table 3 presents the equivalent dosages for the most commonly used FGAs and SGAs, as calculated by the identified methodologies, as well as the minimum effective doses reported in the Maudsley Prescribing Guidelines (D. Taylor et al., 2012). The %BNF values were taken from Mace and Taylor (2005). 4. Discussion 4.1. Calculated methods Calculated methods most commonly used chlorpromazine as the standard comparative drug and relied on fixed and flexible dosing data of variable quality and clinical generalisability (Andreasen et al., 2010; M. Taylor et al., 2012) as there is an inconsistent evidence-base regarding the minimum effective dose of chlorpromazine or haloperidol. Davis and Chen (2004) highlighted that 31 out of 41 placebo-controlled studies have shown that a daily dosage of chlorpromazine greater than 300 mg is superior to placebo, with only 4 studies finding that a dose less than 300 mg a day was superior to placebo. The minimum effective daily dose of haloperidol is thought to be 4 mg (Woods, 2003) but this conclusion was drawn from small studies, each of which included fewer than 30 participants. Other data analysis suggests that daily doses of 3.3–4.0 mg of haloperidol elicit similar clinical responses to higher doses of 4.0–8.0 mg (Davis and Chen, 2004) and doses as low as 2–3 mg may be equivalent for clinical response in first episode cases (McEvoy et al., 1991; Schooler et al., 2005). Davis himself later criticized the methodology of chlorpromazine equivalence on the grounds that it was not based on dose–response curves and thus equivalencies were likely to be ‘inaccurate’ (Davis and Chen, 2004). Alternatively, the linear regression method for chlorpromazine equivalence used by Andreasen and colleagues did allow for estimation of equivalents over both low and high doses (Andreasen et al., 2010) but as dose–response curves are sigmoid, a method based on such a linear relationship is open to the accusation of oversimplification. Other calculated methods were based on maximum dose and included the ‘near-effective maximum dose’ method which allows the comparison of drugs that may not be equally efficacious, but overall there is, perhaps, insufficient, consistent data available for most antipsychotics for the near-effective maximum dose to be indisputably established. Moreover, other maximum dose methods such as %BNF did not consider relative efficacy in converting antipsychotic doses, as drugs with identical %BNF values might not be equally efficacious. As the BNF is UK-based, international dose comparisons may also be problematic unless licensed maximum doses are the same elsewhere. Lastly, the DDD method did not provide any information about the efficacy or relative efficacy of antipsychotics. Further, the DDD does not consider the wide dose range available for antipsychotics as seemingly arbitrary doses are chosen as a global standard for each drug which are not necessarily meaningful in all local clinical settings and for all illness phases. The validity of the calculated methods has been considered through comparison studies. The DDD, %BNF and chlorpromazine equivalents were shown to be positively correlated in study participants who were treated with low-medium antipsychotic doses and belonged to a homogeneous patient group (Nosè et al., 2008). Similarly, a 93% concordance between %BNF and chlorpromazine equivalents for identifying patients prescribed high doses has also been shown (Yorston and Pinney, 2000). In contrast, discrepancy between DDD-equivalents and chlorpromazine equivalents has been reported by others (Rijcken et al.,

2003; Barr et al., 2010). On balance, the maximum dose methods seem to have most utility with the %BNF for clinical practice and perhaps the near-effective maximum dose method for clinical research. However, it could be argued that the near-effective maximum dose method is also more suitable for optimising dosage in clinical practice, potentially allowing achievement of maximal therapeutic effect while minimising dose-related side effects. 4.2. Consensus methods The lack of dosing guidelines has also resulted in consensus studies (level 4 evidence), which mostly used risperidone as the standard comparative drug and relied on the knowledge and experience of experts in the field. For the consensus reported by Kane et al. (2003), no particular clinical scenario was described to the experts, so both the type and phase of illness being considered were not explicit. Further, outlier responses were eliminated and means and standard deviations of the estimations were adjusted to the nearest dose of available pill strength; it is not known whether there was any bias present in rounding up or down to the next available pill strength. Buckley (2003) reported that approximately half of their responders were not confident in estimating dose equivalence, and were not willing to provide an equivalence estimate. Alternatively, Simpson et al. (2006) provided no information on the data used to support the rationale of dose equivalency and the method used to reach consensus among the experts. Further, these three consensus methods appear to have been supported by the pharmaceutical industry, although the influence of this on the process, if any, can only be surmised. The most recent consensus appears to provide the most complete estimation of equivalencies, and specific clinical scenarios and the equivalents reported were found to be strongly correlated (r = 0.89) with the near-effective dose equivalents for 10 antipsychotics (Gardner et al., 2010). However, a bias may have been introduced in stage two of this consensus, as it is more likely that people will agree with a dose that is said to reflect the opinion of the majority, particularly when participants may be uncertain of their personal perspective. This is reflected by the reduction of the variation coefficient from 57% in the first stage to 33% in the second stage. Nonetheless, this latter consensus is perhaps the most comprehensive of the four reviewed. 4.3. Definition of ‘equivalence’ For the most part, ‘dose equivalence’ is thought to relate to equivalence of efficacy for symptom reduction rather than equivalence for tolerability and safety, the latter being pertinent but largely ignored. Consequently, the measurement of efficacy is important for dose comparison methods but the absence of established and universally agreed symptom reduction threshold criteria for response complicates the assessment of efficacy or therapeutic equivalence. A 20% reduction in the Positive and Negative Syndrome Scale (PANSS, Kay et al., 1987) total score is commonly used to define response in clinical trials (Lepping et al., 2011) although it is equivalent to only ‘minimal improvement’ on the Clinical Global lmprovement scale (Leucht et al., 2005a). For patients with an acute relapse of psychosis, which is often an entry criteria for licensing RCTs, a 50% reduction in PANSS has been considered as an appropriate response threshold (Leucht et al., 2005b). Moreover, different response margins are used in non-inferiority trials (e.g. Martin et al., 2002; Fleischhacker et al., 2009). Agreement on clinically meaningful reduction in PANSS scores and a PANSS-derived, universally-agreed, threshold margin for non-inferiority trials would be most helpful. Alternatively, ‘all cause discontinuation’ is an outcome measure, suitable for double-blind studies, that encompasses more than just symptom efficacy and may result in different conclusions being drawn regarding the overall efficacy and dose–response relationship of a particular antipsychotic dose (Citrome et al., 2009). Ultimately, however, symptom reduction equivalence may not be as meaningful

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

as the consideration of dose equivalence in terms of achieving clinical treatment response or clinical remission, which no currently available method addresses. 4.4. Conclusions and implications Dose equivalence for various antipsychotics was found to be determined by both calculated and consensus methods. Based on the findings of this review, no single method stands above the rest and the evidence is not sufficiently strong for one to be recommended more than the others. Thus, none can currently be considered as a ‘gold standard’. Further, the evidence-base underpinning all of the methods was not sufficiently robust to yield uniformly consistent and reliable dose equivalence across all drugs (both FGAs and SGAs), the different formulations available and the full range of prescribed doses. Use of an individual method also has limitations and the extent to which dose equivalence may change with different illness phases (for example, first-episode psychosis versus established illness), different clinical indications (e.g. positive symptoms in schizophrenia or mood stabilization in acute mania) and different patient groups is yet to be determined. Comparison of the different methods suggests that they will yield different equivalencies and the choice of currently available method may subsequently introduce bias in either direction when designing head-to-head, antipsychotic, fixed-dose studies. Consequently, trials should routinely report the reasons why a particular method of dose equivalence was chosen and how this yielded the chosen dose(s). The limitations of this review include the use of only two electronic data sources and that the search was constrained to publications in English. The quality assessment criteria were predetermined and were considered to be the most relevant to both clinical and research settings. Direct dose comparisons between methods were not possible, partly because the reference drug varied. Nonetheless, this systematic review did allow for a synthesis overview for estimates of dose equivalence and methodologies for the most commonly prescribed antipsychotics. Ideally, a robust method of estimating dose equivalence for efficacy should be developed that is valid across different diagnoses and formulations. This should begin with an agreed, unified measure to assess clinical response efficacy in clinical drug trials, which would also adequately reflect response in clinical practice, whilst acknowledging that not all antipsychotics may be equally efficacious or effective (Leucht et al., 2009a,b). Clinical response could then possibly be externally validated by D2 dopamine receptor occupancy studies as it is known that a clinical response requires approximately 65–70% D2 receptor occupancy in the striatum (Tauscher and Kapur, 2001). For example, 5 mg risperidone yields identical D2 receptor occupancy to 20 mg olanzapine (Kapur et al., 1999). No published dose equivalence methods have been based explicitly on D2 occupancy data, although an important caveat here is that the other receptor systems may play a role in the mechanism of action of antipsychotics (Rijcken et al., 2003). Moreover, new antipsychotics in development rely on the glutamatergic system and comparison between these newer drugs and older D2 antipsychotics will be likely in the future. Role of funding source Own account. Contributors MXP formulated the research question and designed the systematic review. MXP and IAA conducted the methods and analyses. All authors, including MT and TREB, interpreted the findings, co-wrote the manuscript and have approved the final manuscript. Conflict of interest MXP holds a Clinician Scientist Award supported by the National Institute for Health Research and has also received consultancy fees, lecturing honoraria, and/or research funding from Janssen, Lilly, Endo, Lundbeck, Otsuka and Wyeth. MXP has previously worked on or is currently working on clinical drug trials for Janssen and Amgen. MT has received consultancy or speaker fees from BMS, Endo, Lundbeck, and Janssen in the last 3 years. TREB has received lecturing honoraria from Lilly and Roche. IAA has no conflicts of interest. The views expressed in this publication are those of the authors and not

147

necessarily those of the NHS, the National Institute for Health Research, or the Department of Health.

References⁎ American Psychiatric Association, 1997. Practice guideline for the treatment of patients with schizophrenia. Am. J. Psychiatry 154 (Suppl. 4), 1–63. ⁎Andreasen, N.C., Pressler, M., Nopoulos, P., Miller, D., Ho, B.C., 2010. Antipsychotic dose equivalents and dose-years: a standardized method for comparing exposure to different drugs. Biol. Psychiatry 67 (3), 255–262. Barnes, T.R.E., Davison, S., Ferrier, I.N., Howard, R., Kerwin, R., King, D.J., et al., 2006. Consensus statement on high-dose antipsychotic medication. Council Report CR138. Royal College of Psychiatrists, London. Barr, A.M., Honer, W.G., Johnson, J.L., Wu, T.K., Procyshyn, R.M., 2010. A comparison of antipsychotic drug-defined daily doses versus chlorpromazine equivalent doses in patients with or without extrapyramidal motor symptoms. J. Clin. Psychopharmacol. 30 (6), 741–743. Bitter, I., Treuer, T., Dyachkova, Y., Martenyi, F., McBride, M., Ungvari, G.S., 2008. Antipsychotic prescription patterns in outpatient settings: 24-month results from the Intercontinental Schizophrenia Outpatient Health Outcomes (IC-SOHO) study. Eur. Neuropsychopharmacol. 18, 170–180. ⁎Buckley, P.F., 2005. Dosing equivalency of second-generation antipsychotic medications. J. Clin. Psychopharmacol. 25 (5), 501–502. Citrome, L., Reist, C., Palmer, L., Montejano, L.B., Lenhart, G., Cuffel, B., et al., 2009. Impact of real-world ziprasidone dosing on treatment discontinuation rates in patients with schizophrenia or bipolar disorder. Schizophr. Res. 115 (2–3), 115–120. ⁎Davis, J.M., 1974. Dose equivalence of the antipsychotic drugs. J. Psychiatr. Res. 11, 65–69. ⁎Davis, J.M., Chen, N., 2004. Dose response and dose equivalence of antipsychotics. J. Clin. Psychopharmacol. 24 (2), 192–208. Fleischhacker, W.W., McQuade, R.D., Marcus, R.N., Archibald, D., Swanink, R., Carson, W.H., 2009. A double-blind, randomized comparative study of aripiprazole and olanzapine in patients with schizophrenia. Biol. Psychiatry 65 (6), 510–517. ⁎Gardner, D.M., Murphy, A.L., O'Donnell, H., Centorrino, F., Baldessarini, R.J., 2010. International consensus study of antipsychotic dosing. Am. J. Psychiatry 167 (6), 686–693. Joint Formulary Committee, 2012. British National Formulary, 64th edition. Royal Pharmaceutical Society and BMJ Publishing Group Ltd., London (September 2012). ⁎Kane, J.M., Leucht, S., Carpenter, D., Docherty, J.P., 2003. The expert consensus guideline series. Optimizing pharmacologic treatment of psychotic disorders. J. Clin. Psychiatry 64 (Suppl. 12), 1–100. Kapur, S., Zipursky, R.B., Remington, G., 1999. Clinical and theoretical implications of 5-HT2 and D2 receptor occupancy of clozapine, risperidone, and olanzapine in schizophrenia. Am. J. Psychiatry 156 (2), 286–293. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13 (2), 261–276. Lambert, T.J., 2007. Switching antipsychotic therapy: what to expect and clinical strategies for improving therapeutic outcomes. J. Clin. Psychiatry 68 (Suppl. 6), 10–13. Lepping, P., Sambhi, R.S., Whittington, R., Lane, S., Poole, R., 2011. Clinical relevance of findings in trials of antipsychotics: systematic review. Br. J. Psychiatry 198 (5), 341–345. Leucht, S., Kane, J.M., Kissling, W., Hamann, J., Etschel, E., Engel, R., 2005a. Clinical implications of Brief Psychiatric Rating Scale scores. Br. J. Psychiatry 187, 366–371. Leucht, S., Kane, J.M., Kissling, W., Hamann, J., Etschel, E., Engel, R.R., 2005b. What does the PANSS mean? Schizophr. Res. 79 (2–3), 231–238. Leucht, S., Corves, C., Arbter, D., Engel, R.R., Li, C., Davis, J.M., 2009a. Second-generation versus first-generation antipsychotic drugs for schizophrenia: a meta-analysis. Lancet 373 (9657), 31–41. Leucht, S., Komossa, K., Rummel-Kluge, C., Corves, C., Hunger, H., Schmid, F., et al., 2009b. A meta-analysis of head-to-head comparisons of second-generation antipsychotics in the treatment of schizophrenia. Am. J. Psychiatry 166, 152–163. Lin, C.H., Kuo, C.C., Chou, L.S., Chen, Y.H., Chen, C.C., Huang, K.H., et al., 2010. A randomized, double-blind comparison of risperidone versus low-dose risperidone plus low-dose haloperidol in treating schizophrenia. J. Clin. Psychopharmacol. 30 (5), 518–525. Mace, S., Taylor, D., 2005. A prescription survey of antipsychotic use in England and Wales following the introduction of NICE guidance. Int. J. Psychiatry Clin. Pract. 9 (2), 124–129. Martin, S., Lĵo, H., Peuskens, J., Thirumalai, S., Giudicelli, A., Fleurot, O., et al., 2002. A double blind randomized comparative trial of amisulpride versus olanzapine in the treatment of schizophrenia: short-term results at two months. Curr. Med. Res. Opin. 18 (6), 355–362. McCue, R.E., Waheed, R., Urcuyo, L., Orendain, G., Joseph, M.D., Charles, R., et al., 2006. Comparative effectiveness of second-generation antipsychotics and haloperidol in acute schizophrenia. Br. J. Psychiatry 189, 433–440. McEvoy, J.P., Hogarty, G.E., Steingard, S., 1991. Optimal dose of neuroleptic in acute schizophrenia. A controlled study of the neuroleptic threshold and higher haloperidol dose. Arch. Gen. Psychiatry 48 (8), 739–745. ⁎Milton, J., Lawton, J., Buckley, A., 1995. Neuroleptic prescribing practice. Psychiatr. Bull. 19, 575–576. Nosè, M., Tansella, M., Thornicroft, G., Schene, A., Becker, T., Veronese, A., et al., 2008. Is the Defined Daily Dose a reliable tool for standardizing antipsychotic dosages? Int. Clin. Psychopharmacol. 23 (5), 287–290. Paton, C., Barnes, T.R., Cavanagh, M.R., Taylor, D., Lelliott, P., 2008. POMH-UK project team. High dose and combination antipsychotic prescribing in acute adult wards in the UK: the challenges posed by p.r.n. prescribing. Br. J. Psychiatry 192 (6), 435–439. ⁎

denotes included articles.

148

M.X. Patel et al. / Schizophrenia Research 149 (2013) 141–148

Rijcken, C.A., Monster, T.B., Brouwers, J.R., de Jong-van den Berg, L.T., 2003. Chlorpromazine equivalents versus defined daily doses: how to compare antipsychotic drug doses? J. Clin. Psychopharmacol. 23 (6), 657–659. Rosenheck, R.A., Davis, V.G., Davis, S.M., Stroup, S., McEvoy, J., Swartz, M., et al., 2009. Can a nonequivalent choice of dosing regimen bias the results of flexible dose double blind trials? The CATIE schizophrenia trial. Schizophr. Res. 113 (1), 12–18. Schooler, N., Rabinowitz, J., Davidson, M., Emsley, R., Harvey, P.D., Kopala, L., et al., 2005. Risperidone and haloperidol in first-episode psychosis: a long-term randomized trial. Am. J. Psychiatry 162, 947–953. ⁎Simpson, G.M., Mahmoud, R.A., Lasser, R.A., Kujawa, M., Bossie, C.A., Turkoz, I., et al., 2006. A 1-year double-blind study of 2 doses of long-acting risperidone in stable patients with schizophrenia or schizoaffective disorder. J. Clin. Psychiatry 67 (8), 1194–1203. Suzuki, T., 2011. An importance of dose in antipsychotic combination therapy for schizophrenia. J. Clin. Psychopharmacol. 31 (3), 399. Tauscher, J., Kapur, S., 2001. Choosing the right dose of antipsychotics in schizophrenia: lessons from neuroimaging studies. CNS Drugs 15 (9), 671–678. Taylor, M., Shajahan, P., Lawrie, S., 2008. Comparing the use and discontinuation of antipsychotics in clinical practice – an observational study. J. Clin. Psychiatry 69 (2), 240–245.

Taylor, D., Paton, C., Kapur, S., 2012. The Maudsley Prescribing Guidelines in Psychiatry, 11th Edition. Wiley-Blackwell, Chichester. Taylor, M., Cavanagh, J., Hodgson, R., Tiihonen, J., 2012. Examining the effectiveness of antipsychotic medication in first-episode psychosis. J. Psychopharmacol. 26 (5 Suppl.), 27–32. Uchida, H., Suzuki, T., Takeuchi, H., Arenovich, T., Mamo, D.C., 2011. Low dose vs standard dose of antipsychotics for relapse prevention in schizophrenia: meta-analysis. Schizophr. Bull. 37 (4), 788–799. ⁎WHO Collaborating Centre for Drug Statistics Methodology, 2012a. Guidelines for ATC classification and DDD assignment (Oslo). WHO Collaborating Centre for Drug Statistics Methodology, 2012b. ATC index with DDDs (Oslo). ⁎Woods, S.W., 2003. Chlorpromazine equivalent doses for the newer atypical antipsychotics. J. Clin. Psychiatry 64 (6), 663–667. ⁎Yorston, G., Pinney, A., 1997. The use of high dose antipsychotic medication. Psychiatr. Bull. 21, 566–569. Yorston, G., Pinney, A., 2000. Chlorpromazine equivalents and percentage of British National Formulary maximum recommended dose in patients receiving high-dose antipsychotics. Psychiatr. Bull. 24, 130–132.