02-06-heres.fm

Reviews and Overviews
Why Olanzapine Beats Risperidone, Risperidone Beats
Quetiapine, and Quetiapine Beats Olanzapine: An
Exploratory Analysis of Head-to-Head Comparison
Studies of Second-Generation Antipsychotics
Stephan Heres, M.D.
Objective: In many parts of the world,
the results in favor of the sponsor’s drug.
John Davis, M.D.
largely replaced typical antipsychotics as Results: Of the 42 reports identified by
Katja Maino, M.D.
nia. Consequently, trials comparing twodrugs of this class—so-called head-to- head studies—are gaining in relevance.
studies, the reported overall outcome was Elisabeth Jetzinger, M.D.
in favor of the sponsor’s drug. This pat- tern resulted in contradictory conclusions Werner Kissling, M.D.
across studies when the findings of stud- Stefan Leucht, M.D.
sp o nso rs we re co m p are d. Po te nt ial and the drug favored in the study’s over- doses and dose escalation, study entry cri- teria and study populations, statistics and Method: The authors identified head-to-
head comparison studies of second-gen-
Conclusions: Some sources of bias may
limit the validity of head-to-head compar- ison studies of second-generation antipsy- bias identified in this review were subtle all studies fully or partly funded by phar- rather than compelling, the clinical use- fulness of future trials may benefit from mask the names and doses of the drugsused in the trial, and two physicians minor modifications to help avoid bias.
sources of bias can be addressed by study initiators, peer reviewers of studies under were not blinded to the study sponsor re- viewed the entire report of each study for (Am J Psychiatry 2006; 163:185–194)
A scientific debate about the effectiveness of second- paring the same two drugs have had contradictory conclu- generation antipsychotics, compared to conventional an- sions (1, 2). This effect may not be totally unrelated to the tipsychotics, has been going on for several years. Although funding sources of the trials. Conflicts of interest arising all questions have not as yet been answered, second-gen- from a pharmaceutical company’s sponsorship of clinical eration antipsychotics are now defined as the gold stan- trials of a drug it manufactures are obvious (3), and the as- dard in most aspects of treatment, at least in highly indus- sociation of funding and conclusions is found in numer- trialized countries. As a result, so-called head-to-head ous medical specialties (4). In this article, we present a comparisons, i.e., randomized, controlled clinical trials summary of head-to-head comparison studies in psychia- with two or more active second-generation antipsychotic try in which we focus on various aspects of potential bias comparators, have become increasingly important as new that may arise from such conflicts of interest. To our knowledge, this work is the first examination of potential Somewhat confusing is the fact that different trials com- bias related to study sponsorship of head-to-head com- Am J Psychiatry 163:2, February 2006 COMPARISON STUDIES OF ANTIPSYCHOTICS
FIGURE 1. Scores Assigned by Two Physicians in Blind Rat-
tion of a trial, unless the abstract stated otherwise. Secondary ings of 30 Abstracts Reporting the Outcome of Head-to-
publications were excluded in order to avoid multiple inclusions Head Comparison Studies of Second-Generation Antipsy-
of the source trial in the analysis. We also screened proceedings of choticsa
selected conferences for the period from 1999 to February 2004.
The conference reports we reviewed were limited to materials from events attended by members of our work group.
Blinded Rating of Abstracts
On the basis of the hypothesis that funding by a pharmaceuti- cal company may influence the outcome of a trial, we checked the reports for information on sponsorship by a “profit-making orga-nization.” The abstract of each study was modified to mask the Abstracts 10
names and doses of the drugs used in the trial, and two physicians(a psychiatrist [K.M.] and an internist [E.J.]), both of whom were blinded to the funding source for the trial and were not involved in the design of the evaluation, read the complete abstract and Number of
rated which drug was favored in the overall conclusion. The rat- ings were made on a 6-point scale proposed by Gilbert et al. (5) and previously used in studies evaluating the association of fund-ing and conclusions in drug trials (4, 6). The scoring method is de- scribed in the footnote to Figure 1. For blinding, the second-gen- eration antipsychotic names in the abstracts were replaced by “DRUG A” and “DRUG B” (“DRUG A” was not always the sponsor’s a The abstracts were modified for blind ratings by replacing the drug drug and vice versa), and the total dose/dose range was replaced names with “DRUG A” and “DRUG B” and replacing the total dose/ by “x.” A separate sensitivity analysis that included only peer-re- dose range by “x.” Scores from 4 to 6 favored the sponsor’s drug, viewed publications was carried out. Two-sided binomial sign and scores from 1 to 3 favored the comparator. Scores were as- tests were used to test the hypothesis of potential influence of the signed according to the following scale: 1=DRUG B is highly pre- sponsor on the study outcome, and Cohen’s kappa was used for ferred and is the best alternative; should be considered the stan- measuring interrater reliability. Statistical significance was de- dard intervention in all patients, or the like; 2=DRUG B preferred toDRUG A, but DRUG A might be promising under certain circum- fined at an alpha level of <0.05.
stances or the like; 3=DRUG A and DRUG B about equal, but DRUG Identifying Potential Sources of Bias
A is disappointing, as DRUG B had some minor advantages; 4=DRUG A and DRUG B about equal, but DRUG A is successful because The trial reports were read independently by two authors who of minor advantages; 5=DRUG A preferred to DRUG B, but further were not blinded to the sponsor of the trial (S.H., S.L.) to identify trials still indicated; may be more costly or similar disclaimer; 6= potential sources of bias that could have influenced the results in DRUG A highly preferred and should be considered the standard in- favor of the sponsor’s drug. We focused on several factors that tervention for all patients, or the like. In this example, “DRUG A” have been discussed as potential sources of bias, including fea- designates the study sponsor’s drug and “DRUG B” designates the tures of study design, dose ranges, titration schedules, statistics, comparator, although in the actual abstracts modified for blind rat-ings, “DRUG A” was not always used to designate the sponsor’s drug reporting of results, and wording of findings (4, 7, 8). If the con- clusions of the two reviewers differed, consensus was achieved bydiscussion. The second author (J.D.) checked and approved thefindings. As a reference for dose ranges, we used the following parison studies of antipsychotic medications. We also ex- range recommendations included in the American Psychiatric amined the association of the conclusions of head-to- Association Practice Guideline for the Treatment of Patients With head comparison studies with the source of funding. Con- Schizophrenia, second edition (9): 10–30 mg/day of aripiprazole, sequently this study is not a review or a meta-analysis in 150–600 mg/day of clozapine, 10–30 mg/day of olanzapine, 300– which the efficacy or tolerability of different second-gen- 800 mg/day of quetiapine, 2–8 mg/day of risperidone, 120–200mg/day of ziprasidone, and 5–20 mg/day of haloperidol. For eration antipsychotics is examined but an exploratory ap- amisulpride, we used the following dose ranges suggested in the proach to clarifying partly contradictory study results in drug company’s product information: 400–800 mg/day for the field of schizophrenia treatment.
acutely ill patients and 50–300 mg/day for patients with predom-inantly negative symptoms.
Search Strategy
Search Results
We searched MEDLINE (1966–September 2003) for random- From 146 publications found in the MEDLINE search, ized, controlled trials comparing the second-generation antipsy- we excluded 61 reviews, 22 reports of additional data from chotics aripiprazole, amisulpride, clozapine, olanzapine, que-tiapine, risperidone, sertindole, and ziprasidone. The search previously published trials or preliminary results, 17 re- terms were paired combinations of the second-generation anti- ports of laboratory or electrophysiological data, five re- psychotics and the term “rand*” (for “random,” “randomized,” ports of add-on therapy with other drugs, four reports on etc.). We excluded reviews, meta-analyses, reports focused solely alternative diagnoses, 11 reports of studies that did not in- on laboratory or electrophysiological data, trials with combined clude a direct head-to-head comparison, and one report drug treatment, and reports on patient populations with diag-noses other than schizophrenia or schizoaffective disorder. Re- on combined antipsychotic treatment, which left 25 pub- ports on drug efficacy were considered to be the primary publica- lications for analysis. The complete trial report for one of Am J Psychiatry 163:2, February 2006 HERES, DAVIS, MAINO, ET AL.
the 25 publications could not be obtained, and that study TABLE 1. Number of Reports That Favor the Study Spon-
was excluded. Thirteen conference presentations of head- sor’s Drug or the Comparison Drug in Industry-Sponsored
Head-to-Head Comparison Studies of Second-Generation
to-head drug comparisons were identified, and during the Antipsychotics
analysis, another four publications and one report in press were identified, for a total of 42 trial reports. Of the 42 re- ports, 32 were fully or partly funded by a pharmaceutical company that manufactured one of the drugs used in the trial (1, 2, 10–39). One of the 42 studies was conductedwith supplemental funding from a pharmaceutical com- pany, although the acquisition and reporting of the data were implemented with no input from the company (40); this study was not included in the blinded rating of ab- stracts, but it was included in the analysis of sources of bias. Nine of the 42 studies were not funded by a pharma- ceutical company (41–49). Two reports of sponsored stud- ies did not include an abstract (10, 36). Thus, 30 trials were included in the blinded rating of study abstracts.
Sponsorship and Outcome as Reported in Study
Abstracts
According to the ratings by the two physicians, the over- of bias are summarized in a separate table (Data Supple- all outcome reported in the study abstracts was in favor ofthe sponsor’s drug in 90.0% of the abstracts (N=27 of 30) ment 1) available from the first author and available with (p<0.001, binomial sign test) (Figure 1). For each abstract, the online version of this article at http://ajp.psychiatry- the scores of the two raters were the same or differed by online.org. We identified potential sources of bias as de- only 1 point, and the two raters did not differ in whether batable or clear. For example, in several instances, we the outcome was judged to be in favor of the sponsor’s identified debatable sources of bias in dose ranges for ris- drug (a score of 4, 5, or 6) or the comparator (a score of 1, peridone, for which the appropriate range may still be ar- 2, or 3). According to the criteria of Landis and Koch (50), guable. We identified clear sources of bias in instances in- the interrater agreement was “moderate” (kappa=0.44, volving obviously inappropriate choices of dose, design, p≤0.001) for the numeric rating and “almost perfect” reporting, etc. We emphasize that although at least some (kappa=1.0, p<0.001) for the outcome category. Figure 1 of the biases we identified seemed very obvious, our anal- shows the distribution of the scores for both raters. In the ysis remains speculative, and there is no proof that the fac- sensitivity analysis that included only the abstracts that tors we identified really influenced the results. The biases underwent peer review (N=21), the result was virtually we identified are described in the following sections.
identical, with 90.5% (19 of 21) rated as having an outcome Doses and Dose Escalation
in favor of the sponsor’s drug (p<0.001, binomial sign test).
The interrater agreement was “substantial” (kappa=0.61, Dose ranges and dose escalation are crucial factors that p<0.001) for the numeric rating and “almost perfect” potentially influence trial outcome. In numerous trials, (kappa=1.0, p<0.001) for the outcome category. Table 1 dose ranges are scheduled according to the manufac- summarizes the ratings for studies comparing pairs of turer’s package insert, which is problematic with antipsy- drugs by whether one or the other manufacturer spon- chotic drugs. For example, in trials with risperidone, doses sored the study. Only three of these 21 reports did not fa- up to 10 mg/day or even 12 mg/day are frequently possible vor the sponsor’s drug. These pairwise comparisons re- in flexible titration schedules, although this dosing level vealed contrasting outcomes, depending on the sponsor may diminish both the efficacy and tolerability of the of the study, although the outcomes were derived from tri- drug. After the introduction of risperidone to the market, several studies in the mid-1990s yielded evidence of anoptimal dose range of 4–8 mg/day, with an increasing risk Possible Effects of Sponsorship on Trial Outcome
of extrapyramidal side effects at higher doses without any and Reporting
gain in efficacy (51, 52). At the time of the earliest studies Two authors who were not blinded to the sponsor of the included in this summary (1), these data were presumably trial reviewed the study reports and identified potential not yet accessible, but in more recent trials, the dose sources of bias in the following areas: dose and dose esca- ranges should have been adapted to maintain a fair level lation, entry criteria and study population, statistics and of comparison. Trials that did not include the 4-mg/day methods, and reporting and wording of results. The char- dose, recently referred to as the advisable dose (53), and acteristics of the individual trials and the potential sources trials that allowed doses of up to 12 mg/day (10, 12, 34, 37, Am J Psychiatry 163:2, February 2006 COMPARISON STUDIES OF ANTIPSYCHOTICS
40) are problematic. Choosing 4 mg/day as the lower limit sertindole (11), and ziprasidone (10) became comparators of the dose range is also problematic, as downward dose in trials. Exclusion of patients who previously were nonre- adjustment in case of side effects is not possible. Although sponders to risperidone or any other comparator (16) is a dose range of 2–6 mg/day was used in trials sponsored seldom explicitly stated in reports of head-to-head trials, by the manufacturer (2, 18), and even lower doses were although this feature could have a critical effect on obser- used in elderly patients with schizophrenia in trials spon- vations of the efficacy of or response to antipsychotic sored by the manufacturer (19, 21), competitors consis- For trials involving schizophrenic patients with pre- Dose ranges are also problematic in comparisons in- dominantly negative symptoms, questions about the ac- volving other drugs. Dose ranges of clozapine, especially curate definition of the study population may be raised.
in trials that included patients with treatment-resistant Even if appropriate scales for measuring negative symp- schizophrenia, often appear to be too strictly limited (53), toms, such as the negative syndrome subscale of the Posi- resulting in relatively low mean daily doses (<400 mg/day) tive and Negative Syndrome Scale (PANSS) or the Scale for (13, 14, 39). These levels are in contrast to data revealing the Assessment of Negative Symptoms, are applied, there that doses up to 600 mg/day (54) or even 900 mg/day (55, is still the need for information on positive symptoms, as 56) of clozapine proved highly efficacious in treatment-re- they might also be present at study entry. An entry crite- sistant schizophrenia. In comparisons involving olanza- rion of a difference of 6 points between the PANSS nega- pine, the upper limit of the dose range is often set at 15 tive and positive subscale scores may ensure that subjects mg/day (16, 20, 38), thus excluding the most effective 20- have a predominance of negative symptoms, but it leaves mg/day dose. Use of this limited dose range possibly re- room for speculation about the effect of positive symp- duces olanzapine’s efficacy and may result in a misleading toms if baseline information about positive symptoms is conclusion of the competitor’s therapeutic superiority or not presented (30). Correspondingly, in trials involving pa- equality. The optimum dose range of amisulpride in pa- tients with treatment-resistant illness, transparent criteria tients with predominantly negative symptoms ranges for inclusion and exclusion of participants are also re- from 50 to 300 mg/day (57), but in a study comparing quired (54), although no universally accepted definition of amisulpride with another antipsychotic, it should have treatment-resistant schizophrenia exists (59). Studies in been ensured that the patients did not have significant which antipsychotic treatment nonresponse and intoler- positive symptoms at study entry because higher amisul- ance are allowed as alternative entry criteria (14) may have pride doses (400–800 mg/day) are necessary for treatment results that are difficult to interpret. If results derived from such studies are presented in terms of efficacy in treat- Finding the optimum dose escalation schedules for ment-resistant patients, even if the study is not explicitly both compounds in a study is difficult and may be another focused on this population, misunderstandings are fore- source of bias (2, 12, 16, 18–20, 24, 28, 34, 40, 58). In some cases, the bias may derive from the fact that titration ismandatory for some drugs (risperidone, clozapine, sertin- Statistics and Methods
dole), while the comparator (for example, olanzapine) In recent years, studies with a noninferiority design does not require a stepwise dose escalation. Slow titrationcan prolong the time to the full onset of the therapeutic ef- have become a reasonable alternative to placebo-con- fect of a drug, and the optimal dose of the comparator may trolled trials for comparison of the efficacy of antipsy- be reached earlier. This difference plays a major role in chotic agents (60). In a study designed to prove a drug’s su- studies evaluating efficacy over a brief period of time. On periority over an active comparator, large sample sizes are the other hand, side effects might be more likely to appear usually required. However, equivalence can be shown in a with fast dose escalation. The attempt to escape the esca- one-sided noninferiority design with less effort, depend- lation problem by using a fixed-dose regimen raises other ing on the predicted threshold for equivalence, although it problems. Studies with fixed-dose regimens lack natural- is important to note that in a noninferiority design with a istic plausibility because the unrealistic limits imposed do narrow range of equivalence, the sample size required not reflect the therapeutic flexibility required in the treat- may exceed that necessary for a superiority design. Conse- ment of schizophrenia (16, 23, 32, 33, 44, 45).
quently, a basic requirement is to define a priori the extentof the difference between the treatments that is consid- Entry Criteria and Study Population
ered acceptable for declaring noninferiority (61). It seems Because the second-generation antipsychotics became very arguable to assume an equivalent antipsychotic effi- available on the market one by one over the last decade, a cacy of a drug at a threshold of just over 60% of the treat- trial’s entry criteria with respect to previous drug treat- ment effect achieved by the active comparator as mea- ment have to be chosen carefully. Risperidone had been in sured by the reduction in the PANSS total score (10) or the use for more than 5 years when newer drugs such as PANSS negative subscale score (30). Other equivalence amisulpride (32, 37), quetiapine (24, 29), olanzapine (17), thresholds yield findings of more clinical relevance, but Am J Psychiatry 163:2, February 2006 HERES, DAVIS, MAINO, ET AL.
the thresholds differ between comparable studies (28, 32, provided. If this information is omitted, the reported fre- quency of occurrence of extrapyramidal symptoms gives For multiple comparisons, such as those that occur with only a vague impression of the likelihood of these side ef- the use of test batteries in cognition studies, an adjust- ment for multiple testing may be necessary, but no gener- Poster Reports and Multiple Publishing
ally accepted approach toward this statistical problem ex-ists. One work group may confuse the reader by applying Phrasing of abstracts is difficult, because much infor- an adjustment for multiple testing in one study (18, 20) mation has to be made transparent to the reader in only a and not in a comparable trial (19). In some studies, the ap- few lines. Although the abstracts of many head-to-head plication of an adjustment was not explicitly mentioned or studies adhere to widely accepted structural standards adequately discussed, despite the presence of multiple (65), the results stated are often highly selective. For exam- ple, in the abstract of one study (29), a significant differ-ence in rates of extrapyramidal symptoms that favored the Another source of potential bias is a study design in sponsor’s drug is reported in detail, but the side effects un- which an acute-phase trial of up to 8 weeks is followed by favorable to the drug were mentioned without corre- a continuation phase of up to 12 months that is focused on long-term maintenance of the treatment effect. After the Preliminary results of trials are often presented as poster acute phase, patients who are nonresponders are discon- reports at conferences. Presentation of multiple poster re- tinued from the study and only those who meet the re- ports on the same trial with different first authors can lead sponse criteria are included in the maintenance phase to the impression that independent studies have been (63). This design may be acceptable for relapse studies but conducted (10, 66). If data from a previously published leads to problems in response trials. Selecting only re- trial are later used as the basis for reports focusing on sub- sponders for continuation in a trial that is focused on re- populations or secondary objectives, the abstracts of the sponse (as measured, for example, with the mean reduc- later studies should contain a cross-reference to disclose tion of the PANSS score from baseline to endpoint) as well the source of the data at a glance (62–64, 67, 68). Stand- as further improvement alters the study population radi- alone publication of data deriving from another trial with- cally, necessitating careful interpretation of the results in out a reference to the earlier trial gives the impression that separate trials have been conducted (18, 19).
Reporting and Wording of Results
A complete disclosure of all results of the head-to-head Discussion
comparison would appear to be mandatory but is not al- The first part of our analysis revealed a clear link be- ways provided. Results favoring the drug manufactured by tween sponsorship and study outcome as reported in the the sponsor are often presented in detail, and unfavorable abstract, as 90.0% of the abstracts were rated as showing results often are mentioned in a brief sentence at the very an overall superiority of the sponsor’s drug. This finding is end of the report’s results section or not mentioned at all in accordance with numerous previous reports of a similar (1, 12). Accordingly, the report’s authors may choose to effect in other medical fields (3, 4, 6, 69). Even more strik- present only data from observed cases or only data from a ing were our findings for pair-wise comparison of different last-observation-carried-forward analysis, depending on trials that examined the effects of the same two drugs (Ta- the resulting outcomes. If the last-observation-carried- ble 1). We found that different comparisons of the same forward design showed no significant difference between two antipsychotic drugs led to contradictory overall con- drugs, the results from the observed cases may be dis- clusions, depending on the sponsor of the study. On the played in detail and presented as a significant outcome of basis of these contrasting findings in head-to-head trials, the study (11). The relevant population for evaluation of it appears that whichever company sponsors the trial pro- the primary outcome should be stated a priori in the pro- duces the better antipsychotic drug. This peculiar result tocol and made transparent to the reader.
led us to take a closer look at various design and reporting Furthermore, reporting of adverse events seems to be features. Indeed, a number of potential reasons for the as- selective (34, 36, 38, 62), and the corresponding level of sociation between drug-company-sponsored trials and fa- significance for comparisons of rates of adverse events may not be consistently stated (21, 29). Information onside effects that are very likely to occur, such as sedation Limitations to Our Approach
and weight gain with olanzapine (15, 64) or elevation of A first limitation is that we did not retrieve all trials that prolactin levels with amisulpride (28), may be lacking. In were presented at conferences. Because no databases for addition, in reports of extrapyramidal symptoms, detailed such presentations exist, we were limited to the posters information on the mean daily dose of anticholinergic from conferences attended by members of our work medication and the number of patients who received at group. The conference presentations we included are least one dose of anticholinergic medication should be therefore not necessarily representative of all conference Am J Psychiatry 163:2, February 2006 COMPARISON STUDIES OF ANTIPSYCHOTICS
publications. We did not, however, want to exclude this Sponsorship and outcome as reported in the ab-
material completely, because conferences are an impor- stract. Our results show that reading only the abstract of
tant way for companies to distribute information. We a study is insufficient for a complete understanding of the made no selection among the available reports. The main study findings. However, lack of time makes it difficult limitation of our exploratory analysis is that it must re- even for scientific experts to read all trial reports in detail.
main speculative by nature. Although in some cases—for Therefore, peer reviewers of studies being considered for example, the trial in which the optimal risperidone dose of publication should pay close attention to the conclusions 4 mg/day was explicitly excluded (10)—it is quite obvious stated in study abstracts. Overall, we found that the struc- that the factor we identified may have biased the results, ture of the abstracts in the current review adhered to there is no proof that it really did. Only a “remake” of the widely accepted standards (65), but the selection of the re- study factoring out the source of bias could test the im- sults and the phrasing used to convey the results needed pact. Furthermore, other readers may have different opin- to be carefully scrutinized. To avoid bias in this crucial sec- ions, especially about the more subtle potential sources of tion of trial reporting, we suggest that peer reviewers verify bias. Finally, we emphasize that most of the identified fac- whether the abstract really summarizes the overall results tors were indeed rather subtle and did not reflect an at- of the trial in a balanced way. Detailed guidelines in this tempt by the drug trial sponsors to intentionally misinter- area for peer reviewers would be useful.
pret their findings or to willfully mislead readers.
Dose and dose escalation. In head-to-head trials, dose
Benefit From Industry-Sponsored Trials
ranges and escalation schemes have a major effect on theoutcome. To avoid potential bias, study initiators could In many respects the industry-sponsored studies in- ask the competitor to provide a suggested dose range and cluded in our review met high methodological standards titration schedule for its compound, as the manufacturer (26, 27) and often surpassed non-industry-sponsored tri- of a drug knows its properties best. Alternatively, external als in the quality of research methods (6, 70). Industry-in- experts could function as independent advisers, but they dependent studies are not necessarily free of bias and are should then be named in the report as a source of infor- often too underpowered to find statistically significant dif- mation on the dosing regimen. In addition, responsible ferences or to allow any generalization (46, 47, 71). In our agencies such as the U.S. Food and Drug Administration review, the sample size per group of the nine studies not (FDA) or the European Medicines Agency (EMEA) might funded by a for-profit organization ranged from nine to be given the chance to look at the protocol before the 113 patients. Other factors that contribute to the excellent study is begun in order to allow the correction of obvious methodological standards of industry-sponsored trials are valid central randomization, the high quality of data ac-quisition and management, regular auditing processes, Entry criteria and study population. Regarding study
and the pharmaceutical company’s researchers’ detailed population and inclusion criteria, study initiators should knowledge about the drug (6, 70). There is also no doubt follow broadly accepted standards in the characterization that the development of the second-generation antipsy- of the eligible patients. Diagnostic validity is hardly ever chotics was a major step forward. For the first time anti- mentioned in sponsored trials, and theoretically heteroge- psychotic drugs with clearly defined dose ranges were neous outcomes may be partly due to the heterogeneity of made available, while the optimum dose, even of the stan- the study population. The use of structured clinical inter- dard conventional antipsychotic haloperidol, is still in views may help identify the proper study population. For doubt. Industry-organized trials also markedly improved example, a characterization of patients with predomi- our knowledge about general clinical questions such as nantly negative symptoms has been proposed (73). Defin- medication switching strategies (72), the treatment of pa- ing a valid study population is essential in studies of pa- tients with refractory disorders (34), and the overall effec- tients with treatment-resistant illness that focus on the tiveness of new and conventional antipsychotics for treat- efficacy of antipsychotics, and other aspects of previous ment of negative symptoms (73). However, if all studies by treatment discontinuation, such as medication intoler- drug companies report positive outcomes, the findings ance, should not be used as alternative inclusion criteria.
Otherwise it is unclear which aspect is related to the supe-riority of a compound (14).
Suggestions for Potential Improvement
Statistics and methods. A comprehensive assessment
Given the unique opportunities of industry for organiz- of the statistical methods applied in the studies we re- ing methodologically sound, large-scale trials, the associa- viewed is beyond the scope of this article. We therefore tion between outcome and sponsor found in the rating of comment only on two points that came up several times abstracts in our study is unsatisfactory. We believe, how- during our review. In the last 5 years, noninferiority de- ever, that in the case of many of the problematic points signs have become more common, leading to a major raised in the Results section, relatively simple measures problem with the threshold of equivalence (74). It is hardly could improve the situation to an appreciable extent.
acceptable to consider the lower margin of the 95% confi- Am J Psychiatry 163:2, February 2006 HERES, DAVIS, MAINO, ET AL.
dence interval at a level of only 60% of the efficacy of the staff. The international Current Controlled Trials meta- competitor to be a sign of noninferiority. As the trend to- register (www.controlled-trials.com) combines national ward this type of statistical design is likely to endure, an as well as disease-specific registers, and each trial in- expert consensus on methods for setting the thresholds is cluded in the register is assigned a specific number. The needed. Other confusing aspects include the use of vari- U.S. Freedom of Information Act mandates publicly ac- ous test methods and lack of the correction for multiple cessible “electronic reading rooms” for materials available statistical tests in trials in which effects on cognitive func- through the Freedom of Information Act, such as, for ex- tion are examined. Recently, a guideline for standard test ample, information about studies registered with the FDA.
batteries for measuring cognition became available (75), However, in our experience, the registers are not easy to and it could soon be followed by a consensus on the statis- tical methods that should be used in this field of research.
Poster reports and multiple publishing. Publication
In general, study initiators should define outcome param- of findings on different aspects of the same trial in several eters a priori and choose the appropriate correction reports has been criticized as the “salami strategy” of sci- method for multiple testing. If the correction method is entific reporting. This criticism may not always be justi- applied to a subset of tests only, this fact should be ex- fied, because it is simply not feasible to report in one pub- lication all the data from a large trial with several aspects Reporting and wording of results. Wording and phras-
of interest or a huge sample size. Readers’ understanding ing of study results are surely the most debatable sources of of the different aspects covered by the study can be en- bias. The CONSORT (consolidated standards of reporting hanced if the masses of data are split into several reports.
trials) statement, developed in the mid-1990s, proposed a However, authors should always clearly state the source checklist to ensure completeness of reporting and assess- reference of the data that are presented (78). Otherwise, ment of the validity of trial results (76). In addition, the Inter- the reader might get the impression that several trials were national Committee of Medical Journal Editors set up a list undertaken, although in fact there was only one. A similar of uniform requirements for manuscripts, including trial problem occurs if different researchers from the same trial registration and complete reporting of all acquired data (77).
are listed as the first author of various conference presen- The recommendations leave a considerable margin for tations or publications by the work group. Because many wording and interpretation of the findings. Therefore, it is scientists have only limited time and choose the abstract again the responsibility of peer reviewers for scientific jour- as the primary information source, the underlying core nals to demand balanced reporting of the results.
study should always be mentioned in the abstract. More-over, data presented exclusively in conference poster ses- Readers of the trial reports should pay close attention to sions or symposia, which normally do not undergo peer the choice of the primary outcome variables and to the review, must be considered problematic (70).
presentation of the results in order to obtain a realistic im-pression of whether a new and unknown aspect of drug Is It All a Matter of Sponsoring?
treatment, following the “uncertainty principle” (6), was The need for more industry-independent studies has observed or whether the study was designed to yield pre- been recognized, and some have already been conducted dictable results in favor of the sponsor’s drug. The uncer- and published (80). Although reports from industry-inde- tainty principle states that a patient should be enrolled in pendent trials may not include biased reporting and a randomized, controlled trial only if there is substantial wording, specific design features such as dose ranges and uncertainty about which of the treatments would benefit study populations can still remain problematic. For exam- the patient most. For example, the appropriateness of a ple, the design of a recent industry-independent study of trial focused on weight gain is debatable if a sponsor’s Alzheimer’s disease patients (81) has been criticized (82, drug that is already known for its minor impact on weight 83). The treatment of schizophrenia has many different as- is compared to a treatment previously shown to be highly pects, and numerous studies will be needed to advance treatment. It is unlikely that public funding will cover The observation that only studies with significant find- them all. We therefore believe that the chance for further ings tend to be published led Melander et al. (78) to coin improvement of current industry-supported trials should the phrase “evidence b(i)ased medicine.” It is noteworthy that a guideline for “good publication practice” has beenproposed to help avoid further publication bias (79). Each Received Nov. 22, 2004; revisions received April 28 and May 31, protocol registered with the European Clinical Trial Data- 2005; accepted Sept. 19, 2005. From Klinik und Poliklinik für Psychi- base is issued a unique number, making trials traceable atrie und Psychotherapie der Technischen Universität München amKlinikum rechts der Isar; the Department of Psychiatry, University of and missing reports conspicuous. Unfortunately, access to Illinois at Chicago; the Department of Psychiatry, Ludwig-Maximilian this information is limited to the study initiator and EMEA Universität, München, Germany; and the Department of Internal Am J Psychiatry 163:2, February 2006 COMPARISON STUDIES OF ANTIPSYCHOTICS
Medicine, Technische Universität München, Germany. Address corre- 16. Cornblatt B: Neurocognitive effects of aripiprazole vs olanza- spondence and reprint requests to Dr. Heres, Klinik und Poliklinik für pine in stable psychosis (abstract). Int J Neuropsychopharma- Psychiatrie und Psychotherapie der Technischen Universität München am Klinikum rechts der Isar, Moehlstrasse 26, 81675 17. Gureje O, Miles W, Keks N, Grainger D, Lambert T, McGrath J, München, Germany; s.heres@lrz.tum.de (e-mail).
Tran P, Catts S, Fraser A, Hustig H, Andersen S, Crawford AM: No source of funding or any grant was used to finance this study.
Olanzapine vs risperidone in the management of schizophre- Additional information on this study accompanies the online ver- nia: a randomized double-blind trial in Australia and NewZealand. Schizophr Res 2003; 61:303–314 18. Harvey PD, Green MF, McGurk SR, Meltzer HY: Changes in cog- References
nitive functioning with risperidone and olanzapine treatment:a large-scale, double-blind, randomized study. Psychopharma- 1. Tran PV, Hamilton SH, Kuntz AJ, Potvin JH, Andersen SW, Beas- ley C Jr, Tollefson GD: Double-blind comparison of olanzapine 19. Harvey PD, Napolitano JA, Mao L, Gharabawi G: Comparative versus risperidone in the treatment of schizophrenia and other effects of risperidone and olanzapine on cognition in elderly psychotic disorders. J Clin Psychopharmacol 1997; 17:407–418 patients with schizophrenia or schizoaffective disorder. Int J 2. Conley RR, Mahmoud R: A randomized double-blind study of risperidone and olanzapine in the treatment of schizophrenia 20. Harvey PD, Siu CO, Romano S: Randomized, controlled, dou- or schizoaffective disorder. Am J Psychiatry 2001; 158:765– ble-blind, multicenter comparison of the cognitive effects of ziprasidone versus olanzapine in acutely ill inpatients with 3. Bekelman JE, Li Y, Gross CP: Scope and impact of financial con- schizophrenia or schizoaffective disorder. Psychopharmacol- flicts of interest in biomedical research: a systematic review.
21. Jeste DV, Barak Y, Madhusoodanan S, Grossman F, Gharabawi 4. Als-Nielsen B, Chen W, Gluud C, Kjaergard LL: Association of G: International multisite double-blind trial of the atypical an- funding and conclusions in randomized drug trials: a reflection tipsychotics risperidone and olanzapine in 175 elderly patients of treatment effect or adverse events? JAMA 2003; 290:921– with chronic schizophrenia. Am J Geriatr Psychiatry 2003; 11: 5. Gilbert JP, McPeek B, Mosteller F: Statistics and ethics in surgery 22. Breier A, Berg PH, Thakore JH, Naber D, Gattaz WF, Cavazzoni P, and anesthesia. Science 1977; 198:684–689 Walker DJ, Roychowdhury SM, Kane JM: Olanzapine versus 6. Djulbegovic B, Lacevic M, Cantor A, Fields KK, Bennett CL, Ad- ziprasidone: results of a 28-week double-blind study in pa- ams JR, Kuderer NM, Lyman GH: The uncertainty principle and tients with schizophrenia. Am J Psychiatry 2005; 162:1879– industry-sponsored research. Lancet 2000; 356:635–638 7. Bodenheimer T: Uneasy alliance—clinical investigators and 23. Kinon B: Improvement of comorbid depression with olanza- the pharmaceutical industry. N Engl J Med 2000; 342:1539– pine versus ziprasidone treatment in patients with schizophre- nia or schizoaffective disorder, in Abstracts of the XIIth Biennial 8. Safer DJ: Design and reporting modifications in industry-spon- Winter Workshop on Schizophrenia, Davos, Switzerland, Feb 7– sored comparative psychopharmacology trials. J Nerv Ment Dis 13, 2004. Schizophr Res 2004; 67(suppl 1):163 24. Knegtering R, Castelein S, Bous H, Van Der LJ, Bruggeman R, 9. American Psychiatric Association: Practice Guideline for the Kluiter H, van den Bosch RJ: A randomized open-label study of Treatment of Patients With Schizophrenia, second edition. Am the impact of quetiapine versus risperidone on sexual func- tioning. J Clin Psychopharmacol 2004; 24:56–61 10. Addington D, Pantelis C, Dineen M, Bermattia I, Romano SJ, 25. Lecrubier Y, Bouhassira M, Olivier V, Lancrenon S, Crawford Murray SR: Ziprasidone versus risperidone in schizophrenia: 52 AM: Olanzapine versus amisulpride and placebo in the treat- weeks of comparative data, in 2003 Annual Meeting New Re- ment of negative symptoms and deficit states of chronic search Program and Abstracts. Arlington, Va, American Psychi- schizophrenia, in Abstracts of the 12th Congress of the Euro- pean College of Neuropsychopharmacology, London, Sept 21– 11. Azorin J, Toumi M, Sloth-Nielsen M: Sertindole is well tolerated 25, 1999. Eur Neuropsychopharmacol 1999; 9(suppl 5):288 and superior to risperidone with respect to efficacy in patients 26. McQuade RD, Jody D, Kujawa MJ, Carson WH Jr, Iwamoto T, with schizophrenia (abstract). Schizophr Res 2003; 60(suppl 1): Archibald DG, Stock EG: Long-term weight effects of aripipra- zole versus olanzapine, in 2003 Annual Meeting New Research 12. Azorin J-M, Spiegel R, Remington G, Vanelle J-M, Péré J-J, Program and Abstracts. Arlington, Va, American Psychiatric As- Giguere M, Bourdeix I: A double-blind comparative study of clozapine and risperidone in the management of severe 27. Meltzer HY, Alphs L, Green AI, Altamura AC, Anand R, Bertoldi chronic schizophrenia. Am J Psychiatry 2001; 158:1305–1313 A, Bourgeois M, Chouinard G, Islam MZ, Kane J, Krishnan R, 13. Bitter I, Dossenbach MR, Brook S, Feldman PD, Metcalfe S, Lindenmayer JP, Potkin S: Clozapine treatment for suicidality in Gagiano CA, Furedi J, Bartko G, Janka Z, Banki CM, Kovacs G, schizophrenia: International Suicide Prevention Trial (In- Breier A: Olanzapine versus clozapine in treatment-resistant or terSePT). Arch Gen Psychiatry 2003; 60:82–91 treatment-intolerant schizophrenia. Prog Neuropsychophar- 28. Mortimer A, Martin S, Loo H, Peuskens J: A double-blind, ran- domized comparative trial of amisulpride versus olanzapine 14. Bondolfi G, Dufour H, Patris M, May JP, Billeter U, Eap CB, Bau- for 6 months in the treatment of schizophrenia. Int Clin Psy- mann P (Risperidone Study Group): Risperidone versus cloza- pine in treatment-resistant chronic schizophrenia: a random- 29. Mullen J, Jibson MD, Sweitzer D: A comparison of the relative ized double-blind study. Am J Psychiatry 1998; 155:499–504 safety, efficacy, and tolerability of quetiapine and risperidone 15. Ciudad A: Olanzapine and risperidone: results of a one-year in outpatients with schizophrenia and other psychotic disor- randomized trial in outpatients with schizophrenia with prom- ders: the Quetiapine Experience With Safety and Tolerability inent negative symptoms, in Abstracts of the XIIth Biennial (QUEST) study. Clin Ther 2001; 23:1839–1854 Winter Workshop on Schizophrenia, Davos, Switzerland, Feb 7– 30. Olie JP, Spina I, Benattia I: Ziprasidone versus amisulpride in 13, 2004. Schizophr Res 2004; 67(suppl 1):161 the treatment of negative symptoms of schizophrenia: a 12- Am J Psychiatry 163:2, February 2006 HERES, DAVIS, MAINO, ET AL.
week, double-blind trial (abstract). Schizophr Res 2002; 45. Klieser E, Lehmann E, Kinzler E, Wurthmann C, Heinrich K: Randomized, double-blind, controlled trial of risperidone ver- 31. Oliemeulen EAP: Is olanzapine a substitute for clozapine? the sus clozapine in patients with chronic schizophrenia. J Clin Psy- effects on psychomotor performance, in Abstracts of the 10th Biennial Winter Workshop on Schizophrenia, Davos, Switzer- 46. Svestka J: A double-blind comparison of olanzapine and que- land, Feb 5–11, 2000. Schizophr Res 2000; 41(1):187 tiapine in the treatment of acute exacerbations of schizo- 32. Peuskens J, Bech P, Moller HJ, Bale R, Fleurot O, Rein W: Amisul- phrenic disorders, in Abstracts of the 16th Congress of the Eu- pride vs risperidone in the treatment of acute exacerbations of ropean College of Neuropsychopharmacology, Prague, Sept schizophrenia. Psychiatry Res 1999; 88:107–117 20–24, 2003. Eur Neuropsychopharmacol. 2003; 13(suppl 4): 33. Potkin SG, Saha AR, Kujawa MJ, Carson WH, Ali M, Stock E, Stringfellow J, Ingenito G, Marder SR: Aripiprazole, an antipsy- 47. Svestka J: Olanzapine versus risperidone in first-episode schizo- chotic with a novel mechanism of action, and risperidone vs phrenic and schizoform disorders: a double-blind comparison, placebo in patients with schizophrenia and schizoaffective dis- in Abstracts of the 16th Congress of the European College of order. Arch Gen Psychiatry 2003; 60:681–690 Neuropsychopharmacology, Prague, Sept 20–24, 2003. Eur 35. Ritchie CW, Chiu E, Harrigan S, Hall K, Hassett A, Macfarlane S, Neuropsychopharmacol 2003; 13(suppl 4):S291 Mastwyk M, O’Connor DW, Opie J, Ames D: The impact upon 48. Wahlbeck K, Cheine M, Tuisku K, Ahokas A, Joffe G, Rimon R: extra-pyramidal side effects, clinical symptoms and quality of Risperidone versus clozapine in treatment-resistant schizo- life of a switch from conventional to atypical antipsychotics phrenia: a randomized pilot study. Prog Neuropsychopharma- (risperidone or olanzapine) in elderly patients with schizophre- nia. Int J Geriatr Psychiatry 2003; 18:432–440 49. Daniel DG, Goldberg TE, Weinberger DR, Kleinman JE, Pickar D, 34. Purdon SE, Jones BD, Stip E, Labelle A, Addington D, David SR, Lubick LJ, Williams TS: Different side effect profiles of risperi- Breier A, Tollefson GD: Neuropsychological change in early done and clozapine in 20 outpatients with schizophrenia or phase schizophrenia during 12 months of treatment with olan- schizoaffective disorder: a pilot study. Am J Psychiatry 1996; zapine, risperidone, or haloperidol. Arch Gen Psychiatry 2000; 50. Landis JR, Koch GG: The measurement of observer agreement 36. Sacchetti E: Comparison of quetiapine, olanzapine and risperi- for categorical data. Biometrics 1977; 33:159–174 done in patients with schizophrenia: interim results of a ran- 51. Chouinard G, Jones B, Remington G, Bloom D, Addington D, domized, rater-blinded study, in Abstracts of the 16th Congress MacEwan GW, Labelle A, Beauclair L, Arnott W: A Canadian of the European College of Neuropsychopharmacology, Pra- multicenter placebo-controlled study of fixed doses of risperi- gue, Sept 20–24, 2003. Eur Neuropsychopharmacol. 2003; done and haloperidol in the treatment of chronic schizo- phrenic patients. J Clin Psychopharmacol 1993; 13:25–40 37. Sechter D, Peuskens J, Fleurot O, Rein W, Lecrubier Y: Amisul- 52. Peuskens J: Risperidone in the treatment of patients with pride vs risperidone in chronic schizophrenia: results of a 6- chronic schizophrenia: a multi-national, multi-centre, double- month double-blind study. Neuropsychopharmacology 2002; blind, parallel-group study versus haloperidol. Br J Psychiatry 38. Simpson GM, Weiden PJ, Pigott TA, Romano SJ, Siu C: Ziprasi- 53. Davis JM, Chen N: Dose response and dose equivalence of an- done versus olanzapine in schizophrenia: 6-month blinded tipsychotics. J Clin Psychopharmacol 2004; 24:192–208 continuation study, in 2002 Annual Meeting New Research 54. Kane J, Honigfeld G, Singer J, Meltzer H: Clozapine for the treat- Program and Abstracts. Arlington, Va, American Psychiatric As- ment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry 1988; 45:789–796 39. Tollefson GD, Birkett MA, Kiesler GM, Wood AJ: Double-blind 55. Simpson GM, Josiassen RC, Stanilla JK, de Leon J, Nair C, Abra- comparison of olanzapine versus clozapine in schizophrenic ham G, Odom-White A, Turner RM: Double-blind study of clo- patients clinically eligible for treatment with clozapine. Biol zapine dose response in chronic schizophrenia. Am J Psychia- 40. Volavka J, Czobor P, Sheitman B, Lindenmayer J-P, Citrome L, 56. Rosenheck R, Cramer J, Xu W, Thomas J, Henderson W, Frisman McEvoy JP, Cooper TB, Chakos M, Lieberman JA: Clozapine, K, Fye C, Charney D: A comparison of clozapine and haloperi- olanzapine, risperidone, and haloperidol in the treatment of dol in hospitalized patients with refractory schizophrenia. N patients with chronic schizophrenia and schizoaffective disor- 41. Jerrell JM: Cost-effectiveness of risperidone, olanzapine, and 57. Leucht S, Pitschel-Walz G, Engel RR, Kissling W: Amisulpride, an conventional antipsychotic medications. Schizophr Bull 2002; unusual “atypical” antipsychotic: a meta-analysis of random- ized controlled trials. Am J Psychiatry 2002; 159:180–190 42. de Haan L, Beuk N, Hoogenboom B, Dingemans P, Linszen D: 58. Thornley B, Adams C: Content and quality of 2000 controlled Obsessive-compulsive symptoms during treatment with olan- trials in schizophrenia over 50 years. Br Med J 1998; 317:1181– zapine and risperidone: a prospective study of 113 patients with recent-onset schizophrenia or related disorders. J Clin Psy- 59. Miller AL, Hall CS, Buchanan RW, Buckley PF, Chiles JA, Conley RR, Crismon ML, Ereshefsky L, Essock SM, Finnerty M, Marder 43. Breier AF, Malhotra AK, Su TP, Pinals DA, Elman I, Adler CM, La- SR, Miller DD, McEvoy JP, Rush AJ, Saeed SA, Schooler NR, Shon fargue RT, Clifton A, Pickar D: Clozapine and risperidone in SP, Stroup S, Tarin-Godoy B: The Texas Medication Algorithm chronic schizophrenia: effects on symptoms, parkinsonian side Project antipsychotic algorithm for schizophrenia: 2003 up- effects, and neuroendocrine response. Am J Psychiatry 1999; date. J Clin Psychiatry 2004; 65:500–508 60. Laster LL, Johnson MF: Non-inferiority trials: the “at least as 44. Conley RR, Kelly DL, Richardson CM, Tamminga CA, Carpenter good as” criterion. Stat Med 2003; 22:187–200 WT Jr: The efficacy of high-dose olanzapine versus clozapine in 61. D’Agostino RB, Massaro JM, Sullivan LM: Non-inferiority trials: treatment-resistant schizophrenia: a double-blind crossover design concepts and issues—the encounters of academic con- study. J Clin Psychopharmacol 2003; 23:668–671 sultants in statistics. Stat Med 2003; 22:169–186 Am J Psychiatry 163:2, February 2006 COMPARISON STUDIES OF ANTIPSYCHOTICS
62. Sajatovic M, Mullen JA, Sweitzer DE: Efficacy of quetiapine and 73. Moller HJ: Amisulpride: efficacy in the management of chronic risperidone against depressive symptoms in outpatients with patients with predominant negative symptoms of schizophre- psychosis. J Clin Psychiatry 2002; 63:1156–1163 nia. Eur Arch Psychiatry Clin Neurosci 2001; 251:217–224 63. Wirtz HS, Kinon BJ, Zhao Z, Barber B: Acute response to olanza- 74. Gomberg-Maitland M, Frison L, Halperin JL: Active-control clin- pine but not to risperidone predicts the likelihood of contin- ical trials to establish equivalence or noninferiority: method- ued improvement over time in patients with schizophrenia ological and statistical concepts linked to quality. Am Heart J (abstract). Schizophr Res 2002; 53(suppl 1):181 64. Edgell ET, Andersen SW, Johnstone BM, Dulisse B, Revicki D, 75. Green MF, Nuechterlein KH, Gold JM, Barch DM, Cohen J, Es- Breier A: Olanzapine versus risperidone: a prospective compar- sock S, Fenton WS, Frese F, Goldberg TE, Heaton RK, Keefe RS, ison of clinical and economic outcomes in schizophrenia.
Kern RS, Kraemer H, Stover E, Weinberger DR, Zalcman S, Marder SR: Approaching a consensus cognitive battery for clin- 65. Haynes RB, Mulrow CD, Huth EJ, Altman DG, Gardner MJ: More ical trials in schizophrenia: the NIMH-MATRICS conference to informative abstracts revisited. Ann Intern Med 1990; 113:69– select cognitive domains and test criteria. Biol Psychiatry 2004; 66. Benattia I, Addington D, Pantelis C, Dineen M, Murray S: 76. Moher D, Schulz KF, Altman DG: The consort statement: re- Ziprasidone versus risperidone in schizophrenia: an eight- vised recommendations for improving the quality of reports of week, double-blind trial with forty-four-week continuation (ab- parallel-group randomised trials. Clin Oral Investig 2003; 7:2–7 stract). Schizophr Res 2003; 60(suppl 1):273 77. ICMJE (International Committee of Medical Journal Editors): 67. Feldman PD, Kaiser CJ, Kennedy JS, Sutton VK, Tran PV, Tollef- Uniform Requirements for Manuscripts Submitted to Biomedi- son GD, Zhang F, Breier A: Comparison of risperidone and olan- cal Journals: writing and editing for biomedical publication.
zapine in the control of negative symptoms of chronic schizo- phrenia and related psychotic disorders in patients aged 50 to 78. Melander H, Ahlqvist-Rastad J, Meijer G, Beermann B: Evidence 65 years. J Clin Psychiatry 2003; 64:998–1004 b(i)ased medicine—selective reporting from studies sponsored 68. Tollefson GD, Andersen SW, Tran PV: The course of depressive by pharmaceutical industry: review of studies in new drug ap- symptoms in predicting relapse in schizophrenia: a double- plications. Br Med J 2003; 326:1171–1173 blind, randomized comparison of olanzapine and risperidone.
79. Wager E, Field EA, Grossman L: Good publication practice for pharmaceutical companies. Curr Med Res Opin 2003; 19:149– 69. Montgomery JH, Byerly M, Carmody T, Li B, Miller DR, Varghese F, Holland R: An analysis of the effect of funding source in ran-domized clinical trials of second generation antipsychotics for 80. Stroup TS, McEvoy JP, Swartz MS, Byerly MJ, Glick ID, Canive JM, the treatment of schizophrenia. Control Clin Trials 2004; 25: McGee MF, Simpson GM, Stevens MC, Lieberman JA: The Na- tional Institute of Mental Health Clinical Antipsychotic Trials of 70. Lexchin J, Bero LA, Djulbegovic B, Clark O: Pharmaceutical in- Intervention Effectiveness (CATIE) project: schizophrenia trial dustry sponsorship and research outcome and quality: system- design and protocol development. Schizophr Bull 2003; 29: atic review. Br Med J 2003; 326:1167–1170 71. Casey DE, Carson WH, Saha AR, Liebeskind A, Ali MW, Jody D, In- 81. Courtney C, Farrell D, Gray R, Hills R, Lynch L, Sellwood E, Ed- genito GG: Switching patients to aripiprazole from other anti- wards S, Hardyman W, Raftery J, Crome P, Lendon C, Shaw H, psychotic agents: a multicenter randomized study. Psychop- Bentham P: Long-term donepezil treatment in 565 patients with Alzheimer’s disease (AD2000): randomised double-blind 72. Leucht S, Pitschel-Walz G, Abraham D, Kissling W: Efficacy and extrapyramidal side-effects of the new antipsychotics olanza- 82. Holmes C, Burns A, Passmore P, Forsyth D, Wilkinson D: pine, quetiapine, risperidone, and sertindole compared to AD2000: design and conclusions. Lancet 2004; 364:1213-1214 conventional antipsychotics and placebo: a meta-analysis of 83. Akintade L, Zaiac M, Ieni JR, McRae T: AD2000: design and con- randomized controlled trials. Schizophr Res 1999; 35:51–68 Am J Psychiatry 163:2, February 2006

Source: http://psychiatryonline.org/data/Journals/AJP/4034/185.pdf