Skip to main content
NCBI home page
The BMJ logoLink to The BMJ
. 2001 Apr 14;322(7291):924.

Atypical antipsychotics in the treatment of schizophrenia

Users' experiences of treatments must be considered

Cliff Prior 1,2,3, Judi Clements 1,2,3, Michelle Rowett 1,2,3
PMCID: PMC1120080  PMID: 11302924

Editor—Geddes et al highlight the poor quality of research into antipsychotic drugs.1 People who use these medicines, and their carers, would strongly echo these views.

Yet the authors use these weak data, putting them through the sophisticated statistical technique of meta-regression analysis, to produce a didactic response: “conventional drugs should remain the first treatment.” The analysis combines studies of six different atypical antipsychotics, including one (clozapine) that is seen by most clinicians as quite different, conducted over 21 years in several different countries. No examination of likely confounding effects is reported; insufficient figures were included in the paper to check. The team's draft report in July 1999, however, suggests that the correlation between effect difference and control dose of haloperidol may be an artefact of including the different atypical antipsychotics in a single analysis.

This is bad science, and worse medicine. Geddes et al have identified important hypotheses: that the optimal dose of conventional antipsychotics may be lower than previously believed and that this dose range may give comparable average efficacy and tolerability to those of atypical antipsychotics. These hypotheses are worthy of study but are not shown by the method used.

Each antipsychotic has a different range of benefits and side effects, and these are of different importance to each user. Given the highly disabling effects of side effects on people's lives, users must have informed choice. They should also expect that their doctor is free to prescribe the best. If the person is too ill to discuss options and there are no other factors, current evidence—repeated by Geddes et al—points to atypical antipsychotics as the first choice.

The National Schizophrenia Fellowship, in partnership with Mind and the Manic Depression Fellowship, has published the largest ever survey of users' experiences of treatments. The results confirm the seriousness of the side effects, poor information about treatments, and denial of choice. Full data are available on www.nsf.org.uk/information/research/ and are being submitted to the National Institute for Clinical Excellence for its technology appraisal of antipsychotics.

We urge clinicians to stop and think before accepting Geddes et al's results. It can be profoundly traumatic to be rendered rigid, trembling, unable to rest, or obese by drug treatment and still coming to terms with the diagnosis, the stigma, and the loss of hopes and expectations that the diagnosis carries.

Footnotes

Competing interests: The National Schizophrenia Fellowship has accepted grants and contracts from central and local government and the NHS; it has also accepted much smaller grants from pharmaceutical companies. Mind will not solicit or accept money from pharmaceutical companies but will participate in conferences that are funded by such companies. The Manic Depression Fellowship has received grants from central and local government and the NHS; it has also occasionally received small grants from pharmaceutical companies (less than 0.5% of its income in any one year) but not in the past two years.

References

  • 1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Pragmatic considerations are important when considering which drug to prescribe

David Taylor 1

Editor—Geddes et al suggest that atypical antipsychotic drugs are no better tolerated than low dose typical antipsychotics and go on to recommend starting treatment with 6-12 mg/day haloperidol or equivalent.1-1 It should not be assumed, however, that typical antipsychotics are anything like well tolerated at these doses. Acute extrapyramidal effects, hyperprolactinaemia, and tardive dyskinesia are frequent problems even at lower, essentially subtherapeutic, doses.1-2 Of course, the last two of these develop some time after the standard six weeks of treatment used in trials. This, and the fact that withdrawals from trials are not a precise measure of patient acceptability, might explain why no clear difference in tolerability was discerned.

The authors also recommend using atypical antipsychotics when patients do not respond adequately to a standard dose of conventional antipsychotic. This is not evidence based. Only clozapine seems to be effective in patients unresponsive to previous treatment, with data relating to other atypical antipsychotics being at best equivocal.1-3 Moreover, none of the data presented in the review seems to support the authors' recommendation.

Arguments surrounding the relative tolerability and efficacy of the different types of drugs are rendered largely redundant by more pragmatic considerations. Atypical antipsychotics are rarely prescribed alone in everyday practice, and concomitantly prescribed conventional antipsychotics greatly increase the frequency of acute movement disorders1-4 and, presumably, worsen tolerability.

Footnotes

Competing interests: DT has received fees for speaking and research funding from AstraZeneca, Eli Lilly, Novartis, and Pfizer. A senior pharmacist in his department is paid for by AstraZeneca. All these companies market or intend to market an atypical antipsychotic.

References

  • 1-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 1-2.Taylor D. Low dose typical antipsychotics—a brief evaluation. Psychiatr Bull. 2000;24:465–468. [Google Scholar]
  • 1-3.Taylor D, Duncan-McConnell D. Refractory schizophrenia and atypical antipsychotics. J Psychopharmacol. 2000;4:409–418. doi: 10.1177/026988110001400411. [DOI] [PubMed] [Google Scholar]
  • 1-4.Taylor D, Mace S, Mir S, Kerwin R. A prescription survey of the use of atypical antipsychotics for hospital inpatients in the United Kingdom. Int J Psych Clin Pract. 2000;4:41–46. doi: 10.1080/13651500052048749. [DOI] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Validity of dropout rates as proxy measure of tolerability is unknown

Rhiannon Rowsell 1,2,3, Christopher Link 1,2,3, John Donoghue 1,2,3

Editor—Geddes et al's analysis of atypical antipsychotic drugs contains sources of bias that they did not address.2-1 Atypical antipsychotics are a heterogeneous group of drugs with important differences. The most transparent measure of atypicality is the degree to which adverse effects mediated by dopamine receptor blockade are reduced.2-2 Thus generalisations that include all of the atypical antipsychotics introduce a bias that dilutes the benefits of the two most atypical compounds, clozapine and quetiapine.

Another source of bias in this paper is the use of dropout rates from clinical trials as the measure of tolerability. In many trials of antipsychotics the dropout rate is linked less with tolerability than with lack of efficacy; thus their use as a measure of tolerability is unreliable.

Geddes et al accept that atypical antipsychotics cause fewer extrapyramidal side effects, which are well known phenomena that add to morbidity and stigma in schizophrenia. But other, absolute measures of adverse effects mediated by dopamine receptor blockade are available.

An increase in plasma prolactin concentration is a serious adverse effect of antipsychotics and can be measured in absolute terms, but Geddes et al did not use it as a measure of tolerability. Modest doses of conventional antipsychotics rapidly increase plasma prolactin concentrations to abnormal levels, and symptomatic hyperprolactinaemia has been reported with conventional antipsychotics well below their effective therapeutic doses in schizophrenia.2-3

In the comparison between quetiapine and chlorpromazine included in the analysis, a significant reduction in prolactin concentrations from baseline values was seen in patients treated with quetiapine but not in those treated with chlorpromazine.2-4 In another study prolactin concentrations in patients treated with quetiapine were no different from those in patients treated with haloperidol or placebo and remained within the normal range even at the highest dose.2-5 In contrast, plasma prolactin concentrations in patients treated with haloperidol at a dose of 12 mg/day increased to more than twice the highest level for patients treated with quetiapine (P=0.0075)2-5 and were in the range that in otherwise healthy women should prompt a search for a prolactinoma.

Geddes et al conclude that treatment with haloperidol at a dose of ⩽12 mg/day will optimise both tolerability and effectiveness. However, it is difficult to have any confidence in their findings. The validity of dropout rates as a proxy measure of tolerability is unknown. Moreover, they do not correlate with objective measures of adverse effects and their use in this way cannot be considered to be evidence based.

Footnotes

Competing interests: RR is an employee and shareholder of AstraZeneca. CL is an employee of AstraZeneca. JD has received fees from AstraZeneca for both consulting and speaking at meetings and has also received reimbursement from the company for attending several conferences. AstraZeneca manufactures quetiapine (Seroquel).

References

  • 2-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2-2.Kerwin R, Taylor D. New antipsychotics: a review of their current status and clinical potential. CNS Drugs. 1996;6:71–82. [Google Scholar]
  • 2-3.Hamner MB, Arana GW. Hyperprolactinaemia in antipsychotic-treated patients: guidelines for avoidance and management. CNS Drugs. 1998;10:209–222. [Google Scholar]
  • 2-4.Peuskens J, Link CJ. A comparison of quetiapine and chlorpromazine in the treatment of schizophrenia. Acta Psychiatr Scand. 1997;96:265–273. doi: 10.1111/j.1600-0447.1997.tb10162.x. [DOI] [PubMed] [Google Scholar]
  • 2-5.Arvanitis LA, Miller BG Seroquel Trial 13 Study Group. Multiple fixed doses of ‘Seroquel’ (quetiapine) in patients with acute exacerbation of schizophrenia: a comparison with haloperidol and placebo. Biological Psychiatry. 1997;42:233–246. doi: 10.1016/s0006-3223(97)00190-x. [DOI] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

“Informed relationship between doctor and patient” does not exist in many parts of the world

Sing Lee 1

Editor—Geddes et al recommended that conventional antipsychotic drugs should be the first line of treatment for schizophrenia.3-1 They did not address the fact that the choice of drug is shaped by a complex of factors that vary across societies. These factors include mode of healthcare financing, patients' status as consumers, strength of advocacy, press freedom, political structure, and the degree of social stigma that psychiatric patients experience.

It is hardly an empirical issue. Where patients and family members have greater access to medical information and a louder voice in the development of health policy, such as in the United States, atypical antipsychotics have become the first line of treatment for psychotic disorders. An opposite situation occurs in much of Asia, including some of the wealthiest societies in the world, where severe discrimination prevails and the mental health service is disproportionately underfinanced. In Asia there is extreme asymmetry of knowledge between doctors and patients. Patients and family members are typically unaware, and would not be told, that atypical antipsychotics can be a treatment option.

Although patients may experience distressing extrapyramidal side effects and their adverse impacts, doctors alone decide whether a patient's side effect is excessive and therefore what drug is to be used. Few patients are told of the risk of tardive dyskinesia, and patients with this irreversible condition do not even know it is drug related.

My experience with patient groups in Hong Kong shows that they are usually less concerned with treatment efficacy than are the authors of meta-analyses. What frighten them most are, in descending order of importance, social stigma, extrapyramidal side effects (themselves a cause of visible stigma), and relapse of the mental condition. There is a long way to go before the “informed relationship between doctor and patient” that Geddes et al alluded to will happen in many parts of the world.

Footnotes

Competing interests: SL directs an anti-discrimination programme in Hong Kong that is partly supported by Janssen-Cilag.

References

  • 3-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Users' views are important

Ian Anderson 1

Editor—Geddes et al's meta-analysis of atypical versus typical antipsychotics must be interpreted carefully.4-1 The overall advantage of atypical compared with typical antipsychotics as used in current clinical practice should not be dismissed. It is extremely optimistic to imagine that doses of typical antipsychotics will be reduced in practice, given the difficulty in getting doctors to change prescribing—for example, to increase the dose of tricyclic antidepressants.4-2

Even if lower doses of typical antipsychotics are used, the efficacy advantage for atypical antipsychotics is not necessarily abolished: the 95% confidence interval encompasses a clinically important advantage to atypical antipsychotics but only a small disadvantage. With regard to tolerability, I take issue with equating it to the number of dropouts due to all causes. This is a measure of several very different factors, comprising a mixture of protocol deviation, withdrawal of consent, stopping due to treatment failure or success, as well as adverse events. It is therefore sensible to analyse discontinuations attributed to adverse events alongside total dropouts; even then we are looking at only one aspect of tolerability, and using this measure alone will not detect the patients who continue with treatment despite appreciable side effects.

This analysis found that atypical antipsychotics cause fewer extrapyramidal side effects even than lower dose typical antipsychotics, and the strong clinical impression, supported by the accompanying editorial,4-3 is that patients choose to avoid these side effects if they can. What is missing from the interpretation of this analysis is any acknowledgement that the users' view may be important. It is difficult therefore to accept the authors' conclusion that atypical antipsychotics are no better tolerated than typical antipsychotics; on the evidence given it just depends on what you mean.

Finally, and perhaps most importantly, the basis of the value judgment used to arrive at the conclusion that typical antipsychotics should be used as first line drugs is not explicit. It seems, despite the authors' protests, to be based implicitly on cost effectiveness by weighing a large cost difference against a perceived absent or minor clinical benefit. I have argued here about the uncertainty of the evidence; how much more uncertain is the translation to recommendation? My own interpretation is that we are condemned to continue living with the uncertainty about whether to use typical or atypical antipsychotics as first line treatment. Might this be a situation in which we should listen to our patients?

Footnotes

Competing interests: IA has received honorariums and research funding from several pharmaceutical companies.

References

  • 4-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4-2.Lawrenson RA, Tyrer F, Newson RB, Farmer RD. The treatment of depression in UK general practice: selective serotonin reuptake inhibitors and tricyclic antidepressants compared. J Affect Disord. 2000;59:149–157. doi: 10.1016/s0165-0327(99)00147-0. [DOI] [PubMed] [Google Scholar]
  • 4-3.Kapur S, Remington G. Atypical antipsychotics. BMJ. 2000;321:1360–1361. doi: 10.1136/bmj.321.7273.1360. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Cost is a crucial issue

Richard Duggins 1,2, David Rhinds 1,2, William Hall 1,2

Editor—The impressive meta-regression analysis by Geddes et al was disappointing in its sidestepping of current issues.5-1 The authors believe that “cost is not a crucial issue at present” because most outcome scores for atypical antipsychotic drugs are the same as those for typical antipsychotics. Their predominant message is, however, that typical antipsychotics should be first line treatment. We strongly disagree that cost is not a crucial issue as it seems to be the only differentiator between the new and old antipsychotic drugs. If cost were not a crucial issue why would Geddes et al recommend which type of antipsychotic should be prescribed first?

As trainee psychiatrists, perhaps we have too much of what Kapur and Remington describe in their editorial as clinical hope,5-2 but we do see a key first line role for atypical antipsychotics. We appreciate being flexible in working collaboratively with patients and able to discuss factors such as weight gain, akathisia, parkinsonism, sexual dysfunction, and risk of tardive dyskinesia. In being offered a choice of drug treatment our patients effectively have a choice about their potential side effects. Prescribing only typical antipsychotics would reduce this choice. The authors state that no single antipsychotic is to be favoured over another and we would like to have all treatment options available in our first line armoury.

We encourage Geddes et al to recognise that cost is a crucial issue and to realise that in effect they have presented a cost minimisation analysis.

Footnotes

Competing interests: None declared.

References

  • 5-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5-2.Kapur S, Remington G. Atypical antipsychotics. BMJ. 2000;321:1360–1361. doi: 10.1136/bmj.321.7273.1360. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Paper underrates patients' experience of extrapyramidal symptoms

Robert Kerwin 1

Editor—Geddes et al studied 52 randomised trials of atypical antipsychotic drugs in the treatment of schizophrenia.6-1 They conclude that there are a few differences in efficacy measures and overall tolerability of conventional versus atypical antipsychotics, the main difference being in the propensity for extrapyramidal symptoms.

I don't think that there is anything new here; many of us who have followed the literature have come to this conclusion. I do think, however, that the paper underrates patients' experience of extrapyramidal symptoms. It is precisely because of patients' experience of these symptoms—not for the relatively unproved superior efficacy—that patients, carers, and charitable groups have campaigned for wider use of atypical antipsychotics. Kapur and Remington's accompanying editorial rightly states that data pooling misses important and proved niche effects,6-2 such as the superiority of clozapine in patients with refractory disease.6-3

A further missed point is that apart from clozapine there is no evidence for a role of atypical antipsychotics in schizophrenia resistant to treatment.6-4 Thus the suggestion that atypical antipsychotics should be reserved for patients in whom treatment with conventional drugs has failed seems spurious. In reality, most of our prevalent group of patients are already taking conventional drugs. Therefore most episodes of schizophrenia should probably be treated anyway with atypical antipsychotics according to the guidelines.

As with all guidelines, one should really focus on the first episode of the disease, as there will be many disease modifying factors that will influence drug response. Nothing in this paper informs practice here. Virtually all 52 trials studied were registration trials in patients with multiple episodes or chronic disease. What evidence there is for the first episode strongly suggests that it is exquisitively sensitive to extrapyramidal symptoms and that atypical antipsychotics may be superior to low dose haloperidol.6-5 Therefore, on the basis of current data, atypical antipsychotics seem to be the drugs of choice at first presentation.

Finally, the authors use the dropout rate as a surrogate for overall tolerability. What is the evidence for this? This is not an unfair question to ask in such an evidence based exercise. My experience of clinical trials in schizophrenia is that patients often drop out because they are deluded, hallucinated, disorganised, and unable to maintain consent. It is premature to conclude that this paper gives support to reversing the trend for wider use of atypical drugs.

Footnotes

Competing interests: RK has received fees for speaking from a range of companies manufacturing both typical and atypical antipsychotics, including Novartis, AstraZeneca, Pfizer, and Sanofi Synthelabo. Primary sources of research funding are peer reviewed by the Medical Research Council and Wellcome Trust. He also receives non-peer reviewed funding from Novartis to perform genetic studies of drug response and studies of suicidality in schizophrenia. He has shares in AstraZeneca and GlaxoSmithKline, which are independently managed by a financial adviser.

References

  • 6-1.Geddes J, Freemantle N, Harrison P, Bebbington P. Atypical antipsychotics in the treatment of schizophrenia: systematic overview and meta-regression analysis. BMJ. 2000;321:1371–1376. doi: 10.1136/bmj.321.7273.1371. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6-2.Kapur S, Remington G. Atypical antipsychotics. BMJ. 2000;321:1360–1361. doi: 10.1136/bmj.321.7273.1360. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6-3.Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment resistant schizophrenic: a double blind comparison with chlorpromazine. Arch Gen Psychiatry. 1988;45:789–796. doi: 10.1001/archpsyc.1988.01800330013001. [DOI] [PubMed] [Google Scholar]
  • 6-4.Conley RR, Tamminga CA, Bartko JJ, Richardson C, Peszke M, Lingle J, et al. Olanzapine compared with chlorpromazine in treatment-resistant schizophrenia. Am J Psychiatry. 1998;155:914–920. doi: 10.1176/ajp.155.7.914. [DOI] [PubMed] [Google Scholar]
  • 6-5.Remington G, Kapur S, Zipursky RB. The pharmacotherapy of first episode schizophrenia. Br J Psychiatry. 1998;172(suppl):66–70. [PubMed] [Google Scholar]
BMJ. 2001 Apr 14;322(7291):924.

Paper corrupts concept of evidence based medicine

Clive Adams 1,2, Lorna Duggan 1,2

Editor—The BMJ champions evidence based medicine,7-1 but Geddes et al's review of atypical antipsychotics in the treatment of schizophrenia7-2 leaves clinicians “on a tightrope act between the persuasiveness of marketing claims, the precise but somewhat myopic results of idealised clinical trials, and the complex realities of clinical practice.”7-3

The reviewers report results of meta-regression, a hypothesis generating technique by which observational epidemiological techniques are applied to trials. They then state, as if hypotheses had been tested, that “conventional antipsychotics should usually be used in the initial treatment of an episode of schizophrenia unless the patient has previously not responded to these drugs or has unacceptable extrapyramidal side effects.”7-2 This statement ignores basic epidemiological principles.

Was the study the strongest that could have been performed? No. Only a subset of atypical drugs is studied,7-4 and these data are out of date; the 1988 search is supplemented by inclusion of data on one study available only in 1999, yet other trials available by that time were ignored.7-5 Only continuous, composite measures of effect on mental state were used. The tables in the paper state that 32% of people left these studies early, and in such trials people who leave early are often assumed to remain stable from their point of exit. Therefore, one third of the data in the meta-analyses, the basis of the meta-regression, probably carry this unlikely assumption.

Did the review assure readers that it minimised the inclusion of bias? No. Exactly how studies were selected, outcomes chosen, and data extracted and entered and analysed is not explained. As consistently more people allocated to typical antipsychotics leave studies early, the assumption of stability beyond the point of exit favours haloperidol and chlorpromazine.

Was the strength of the association, both clinical and statistical, considered? No. Clinical importance was hardly considered, and the reproducibility of results, using different end points, and the effect of controlling for the inevitable confounding were not discussed.

Geddes et al's study could alienate clinicians from practical gains evident from open, well conducted systematic reviews. Smith's editorial stated that it is a lie to suggest that “which interventions to make available can be decided with some data and a computer” and that this “corrupts the concept of evidence based medicine.”7-1 We are concerned that Geddes et al's work, disseminated through the BMJ, the champion of evidence based medicine, corrupts the concept of evidence based medicine and could harm people with schizophrenia.

Footnotes

Competing interests: CEA has attended, and presented information at, functions sponsored by Janssen-Cilag and Eli Lilly. These companies have provided travel, accommodation, and speaker's expenses, but no funds have been paid directly to CEA. Payments related to participation in meetings have been paid to an account to support schizophrenia research. LD has attended functions sponsored by Lundbeck, Janssen, Novartis, Pfizer, and Zeneca and has accepted sponsorship from Eli Lilly for internal flights in the United States and from Janssen for talks. The Cochrane Schizophrenia Group has multiple, and hopefully balancing competing interests. Potential competing interests of the group and individuals are described on http://cebmh.warne.ox.ac.uk/csg/ and sources and quantities of all funding listed.

References

BMJ. 2001 Apr 14;322(7291):924.

Authors' reply

John Geddes 1,2,3, Paul Harrison 1,2,3, Nick Freemantle 1,2,3, Paul Bebbington 1,2,3

Editor—Despite the time that has elapsed since we completed our review, none of the correspondents challenges our findings with new randomised evidence. The disagreements concern our methodology and conclusions.

We agree with several authors that the current randomised evidence is inadequate, but it is important that decisions are based on the best available (albeit imperfect) randomised evidence rather than opinion or selective citation of individual trials. We emphasised the importance of integrating the evidence we presented with clinical judgment and patient preference.

Existing trials have usually been conducted by drug companies for regulatory purposes and have excluded or underrepresented clinically important subgroups of patients such as elderly people and those at an early stage of the illness. It is certainly possible, as Kerwin says, that differences exist between subgroups of patients; rather than missing these, meta-regression can be a powerful way of detecting them.8-1

We disagree with both Taylor and Kerwin about the role of atypical antipsychotic drugs in patients who have not responded to treatment with conventional antipsychotics. This is one subgroup that was well represented in the trials because many of the trial participants had had a suboptimal response to conventional antipsychotics. Definitions of resistance to treatment have broadened since the landmark clozapine study,8-2 and the most clinically useful approach is to consider a change of treatment in any patient in whom response or tolerability is suboptimal.8-3 From a clinical perspective, we consider that it is reasonable to use an atypical antipsychotic if results with one (or more) conventional antipsychotics have been unsatisfactory.

Meta-regression is an established technique for addressing heterogeneity (systematic differences) in the results of randomised trials.8-4 Exploring heterogeneity according to protocol driven criteria makes appropriate use of the available evidence.8-5 Our analysis not only showed the effect of dose on continuous symptom measures with haloperidol but also replicated the finding both for dropout and for the group of trials using chlorpromazine as the comparator. These analyses provide strong evidence for the importance of dose of typical antipsychotic, and are supported by the analyses in which doses were simply dichotomised.

Since higher doses of conventional drugs lead to an increased probability that patients will abandon treatment in trials it is not surprising to find that those same patients have poorer outcomes—the two factors are inextricably linked. In most of the studies, when patients withdrew from the allocated treatment they were considered to have dropped out of the trial.

“Dropout” is a composite outcome including elements of both efficacy and tolerability, but because of the poor reporting of specific adverse events in the studies we reviewed we considered it to be the least biased estimate of the overall acceptability of a treatment. The “last observation carried forward” method was widely used to deal with the resulting missing data to allow an intention to treat analysis. As patients taking typical antipsychotics dropped out earlier on average, this approach will tend to overestimate the comparative efficacy of the atypical drug, because patients leaving the study early will tend to have more symptoms, in contrast to the statement of Adams and Duggan.

Rather than ignoring clinical significance, our work was located firmly in the clinical practice of a group of concerned and interested clinicians and was extensively peer reviewed before submission to the BMJ. Adams and Duggan suggest that our review could harm people with schizophrenia, but they do not state how this might happen or how their interpretation of the data differs from ours.

Kerwin considers that we understated the burden of extrapyramidal side effects, while Prior et al imply that we suggested patients should “be rendered rigid, trembling, unable to rest, or obese by drug treatment.” In fact, we stated that no patient should have to tolerate extrapyramidal side effects, as has all too often been the case. Unfortunately, the benefits on extrapyramidal side effects achieved by atypical antipsychotics are relatively modest and should be considered in the context of the similar overall dropout rates which, despite the caveats mentioned above, show at the very least that the atypical antipsychotics are not unequivocally more tolerable.

Our view remains that the available randomised evidence does not support a wholesale change from conventional drugs as standard first line treatment. It is the gaps in the evidence rather than the cost that should lead clinicians to consider conventional drugs first, though we accept that in many cases a patient's preference or clinician's judgment may make the first line use of atypicals appropriate. Many of the correspondents' criticisms seem to be based on citation of individual trials or secondary outcomes, or to stem from an implicit belief that the burden of proof is on the established drugs, not the new ones, to show superiority. A serious consequence of a premature or uncritical shift to atypical antipsychotics is that it would be difficult to conduct the randomised trials required to answer the many outstanding questions (the existence of which is agreed by all correspondents) because clinicians and patients will not participate.

Footnotes

Competing interests: JG, as director of the Centre for Evidence Based Mental Health, has run workshops around the United Kingdom, organised independently, but often sponsored by pharmaceutical companies. The centre has therefore indirectly received fees and expenses from several of the companies that manufacture antipsychotic drugs. NF has received funds for research, fees, and expenses from several pharmaceutical companies that manufacture antipsychotic drugs and from the Department of Health in England. PH has received support from pharmaceutical companies to attend conferences. He has also received fees for educational lectures to psychiatrists on the psychopharmacology of schizophrenia. PB has received fees for presentations at meetings sponsored by various pharmaceutical companies that manufacture typical and atypical antipsychotics. In addition, he is one of the lead investigators of the European schizophrenia cohort funded by Lundbeck.

References

  • 8-1.Kapur S, Remington G. Atypical antipsychotics. BMJ. 2000;321:1360–1361. doi: 10.1136/bmj.321.7273.1360. . (2 December.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8-2.Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic. A double-blind comparison with chlorpromazine. Arch Gen Psychiatry. 1988;45:789–796. doi: 10.1001/archpsyc.1988.01800330013001. [DOI] [PubMed] [Google Scholar]
  • 8-3.Marder SR. An approach to treatment resistance in schizophrenia. Br J Psychiatry. 1999;37(suppl):19–22. [PubMed] [Google Scholar]
  • 8-4.Smith TC, Spiegelhalter DJ, Thomas A. Bayesian approaches to random-effects meta-analysis: a comparative study. Stat Med. 1995;14:2685–2699. doi: 10.1002/sim.4780142408. [DOI] [PubMed] [Google Scholar]
  • 8-5.Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ. 1994;309:1351–1355. doi: 10.1136/bmj.309.6965.1351. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES