Are antipsychotic drugs effective against acute psychosis?

Trond F. Aarre

doi:10.4045/tidsskr.25.0047

Perspectives

Are antipsychotic drugs effective against acute psychosis?

Norwegian

Trond F. Aarre

See All Articles

Trond F. Aarre

trond.aarre@helse-forde.no

Trond F. Aarre, specialist in drug and addiction medicine and psychiatry. Head of Department at Nordfjord Psychiatric Centre and medical director of Mental Health Care, Førde Hospital Trust.

The author has completed the ICMJE form and declares no conflict of interests.

Article

The scientific evidence suggests that the efficacy of antipsychotic drugs against acute psychosis is so marginal that they hardly deserve their name.

Illustration: Journal of the Norwegian Medical Association / iStock

In January 2025, the Norwegian Directorate of Health published their guidelines for psychopharmacological treatment of psychosis (1). They write: 'Experience tells us that refusal to offer antipsychotics is felt to be irresponsible'. The reason why they say 'is felt to be irresponsible' rather than 'is irresponsible' may well be that the evidence to support this treatment is weaker than one would think. Although the treatment of acute psychosis with antipsychotic drugs is considered routine, or even an obligatory course of action, research shows that few patients benefit significantly from these drugs (2–6).

Knowledge about the benefits of antipsychotics is essential. These drugs are even used for coercive treatment. For that to be acceptable, we need to impose particularly high evidential requirements and be particularly well informed about the evidence.

These drugs are even used for coercive treatment. For that to be acceptable, we need to impose particularly high evidential requirements

The recommendations issued by the Directorate of Health

The Directorate of Health recommend antipsychotic drugs for any first-episode psychosis, although they admit that there are no placebo-controlled trials to support this recommendation. 'Historically, randomisation to placebo has been seen as ethically questionable for people with a first-episode psychosis because trials that involve groups with recurring psychotic episodes have showed that antipsychotics are clearly effective. In practice, placebo groups would therefore have their effective treatment delayed, and this was considered to be ethically problematic vis-a-vis patients who experience the onset of a serious mental disorder' (1). In other words, the Directorate believe that while there is no empirical support for their recommendation, such support is not required because the effect is likely to be as good for first-episode psychoses as for subsequent episodes. I will go on to demonstrate that the efficacy of the drugs in treating recurring episodes of psychosis is not as unequivocal as the Directorate maintain.

Methodology problems

Randomised trials have a number of inclusion and exclusion criteria. For instance, the criteria may exclude people who also meet the criteria for other mental health or substance use disorders, and this will systematically increase the probability of the drugs being effective or reduce the rate of improvement in the placebo group (7).

A more serious problem is that trial participants will normally be taking an antipsychotic drug to start with. In order to be eligible for the trial, they will need to suddenly come off these drugs (8), which is not a recommended course of action in clinical practice. It can be difficult to 'differentiate between symptoms that are caused by reducing or discontinuing a treatment and symptoms of a relapse or new psychotic episodes. There may also be a rebound effect or even a withdrawal syndrome at play' (9). In cases of randomisation to placebo, sudden discontinuation can trigger worsening of the condition, which increases the difference between the placebo group and the medication group (8).

Such methodological problems can lead to exaggerated perceived benefits of the drugs. The efficacy of antipsychotic drugs in controlled trials is greater than in clinical practice (10).

What is meaningful improvement?

I believe that for a therapeutic effect to be meaningful, it needs to be greater than the placebo effect. The difference must be of a certain size to be of clinical interest. If neither the doctor nor the patient can discern the difference, it is of little value.

Efficacy tends to be rated on the symptom rating scale referred to as PANSS (Positive and Negative Syndrome Scale). Analyses of the correlation between PANSS scores and symptom improvement have concluded that: Minimal improvement, which is far less than what we are looking for in clinical practice (11), equates to a 20 % reduction in the PANSS score or at least a 15-point decrease. If there is a reduction in the PANSS score of 50 % or more, or a decrease of at least 33 points, the patient is considered to be 'much improved' (12).

Improvement in the placebo groups

Many get better without antipsychotics. We see improvements in the placebo groups because a crisis often passes with time, and because life events may intervene to ease the problems, irrespective of the treatment. Additionally, the participants receive attention from dedicated professionals who communicate an explanation to their problems, boost the hope of improvement and provide systematic follow-up and emotional support. These non-specific effects produce improvement in both the placebo and the medication groups. Only approximately 40 % of the improvement we see with antipsychotics stems from their active substance (13).

We see improvements in the placebo groups because a crisis often passes with time, and because life events may intervene to ease the problems, irrespective of the treatment

In a meta-analysis, placebo had an effect size of 0.59 for schizophrenia, which is considerably higher than the effect of adding antipsychotics: 'Pre-post effect sizes, whether for placebo or active treatments, are in a different order of magnitude than effect sizes resulting from the comparison of an intervention group with a control group' (14).

How many experience meaningful effect?

The question is whether the efficacy of antipsychotics is clinically interesting, rather than only statistically significant. In my opinion, minimal improvement is not good enough.

A meta-analysis of treatments for people with multiple psychotic episodes (167 trials, 28,102 patients) showed that approximately half saw at least a 20 % reduction in their PANSS score (minimally improved) (3). Only 23 % of the participants achieved a 50 % reduction (much improved). The 'number needed to treat' was five for minimal improvement and eleven for much improvement. Leichsenring et al. maintain that the effect sizes for most of the drugs in this meta-analysis were so small that they were likely to have no clinical significance (6).

In a meta-analysis of trials that compared olanzapine or amisulpride to other antipsychotic drugs or placebo (16 studies, 6221 patients), the condition of 20 % of the patients on antipsychotics remained unchanged or deteriorated (5). Even worse, 43 % of the patients did not even see a minimal improvement, and only a third were 'much improved'. The authors concluded that there is a high nonresponse rate when stringent criteria are set for improvement.

The Norwegian Directorate of Health state that in cases of first-episode psychosis, more than 80 % will have a response, defined as at least a 20 % improvement of the PANSS score (1). This equates to a 'minimal improvement'. A 50 % reduction, which means that the individual has much improved, was seen in approximately 50 % of participants. There were no placebo groups in the referenced study, so we do not know how much of this improvement was caused by the drugs (15).

Staff at the Food and Drug Administration (FDA), which approves new drugs for use in the US, have studied the size of the difference between antipsychotics and placebo. Trials conducted in North America between 1999 and 2007 showed, on average, a reduction in PANSS score that was only six points greater with medication than with placebo (4). In other words, this difference is not even considered a 'minimal improvement'. It is so minute that it would be difficult to perceive and hardly of clinical interest.

For schizophrenia, a standardised mean difference (SMD) in effect sizes between antipsychotics and placebo of 0.73 will equate to a minimal clinical improvement of 15 points on the PANSS score. In their analysis, Leichsenring et al. found a standardised mean difference of 0.38–0.45, which in their opinion cannot be detected by clinicians and therefore can have no clinical benefit (6).

Because there is only a minute effect, approximately 25 % of patients take more than one antipsychotic drug (16). Polypharmacy is associated with severe problems and poor outcomes, but does not resolve these problems. It is also associated with many undesirable effects and a higher mortality rate (16).

Good enough for coercive treatment?

It is clear from the Directorate of Health's comments on the mental healthcare legislation that 'for the expected impact of treatment, the probability of a positive response will have to be higher than usual. Additionally, it is a requirement that the treatment must lead to a considerable change in the patient's condition' (17). If more than minimal improvement is required, there is a conspicuously large proportion of patients who see no meaningful effect from antipsychotic drugs. In my opinion, the higher-than-usual probability requirement for a positive response to coercive treatment, as set out in section 4 - 4 of the Mental Healthcare Act, is not met.

If more than minimal improvement is required, there is a conspicuously large proportion of patients who see no meaningful effect from antipsychotic drug

Nonetheless, some individuals undoubtedly have a manifest and worthwhile benefit from these drugs. While some have little effect against delusional thoughts and hallucinations, they have a favourable effect on irritability, aggressiveness and restlessness. Such improvements can decide whether an individual is a risk to themselves or others, and if the individual can receive the necessary assistance. Even partial efficacy may be what an individual needs to live a dignified life outside of an institution. These are factors we need to consider when we interpret the body of evidence.

Minimal efficacy in most cases

Low medication efficacy and high placebo efficacy suggest that we should continue to conduct placebo-controlled trials of antipsychotics. The disappointing findings of the scientific studies mean that we can no longer take for granted that the drugs have a meaningful effect in the majority of those who experience psychosis.

The problem is not a lack of evidence of effect, but evidence that a majority of patients experience no meaningful effect. The drugs have a series of adverse effects that will have to be balanced against their questionable benefits.

After more than 60 years of administering antipsychotic drugs, we have yet to demonstrate more than minimal efficacy in the majority of patients. Those who believe that antipsychotic drugs are more effective than this, should produce equally good or better studies that arrive at the opposite conclusion. It is no longer enough to wish or believe that these drugs work the way we want them to work. Time has come to demonstrate that they do. Until somebody is able to do that, I will stop referring to them as antipsychotics. Instead, I will use the name introduced by the pharmaceutical industry itself: neuroleptics. For there is reason to doubt their efficacy against acute psychosis.

Comments ( 4 )

Dette kommentarfeltet modereres, men kommentarer blir ikke redaksjonelt behandlet ut over å sikre at de følger retningslinjer for vårt kommentarfelt.

02.04.2025:

Vi takker Trond Aarre for en modig og tankevekkende kronikk. Det krever mot å utfordre det etablerte paradigmet hvor nevroleptika er dominerende ved psykosebehandling. Vi støtter Aarres vurdering av at nytten er utilstrekkelig dokumentert og at kunnskapsgrunnlaget – særlig ved tvangsbehandling og langvarig behandling – er mangelfullt.

For oss som arbeider med barn og unge er det avgjørende å få bedre kunnskap om hvem som har nytte av medikamentell behandling ved sykdomsdebut, og hvem som klarer seg like godt eller bedre uten. Det er dokumentert at seponering etter langvarig bruk øker risikoen for tilbakefall, men det er fortsatt uklart om medikamentene påvirker langtidsprognosen positivt (1).

Smedslund og Stoltenberg (FHI) publiserte i 2018 en litteraturgjennomgang som konkluderte med at effekten av langtidsbruk med nevroleptika er dårlig dokumentert, og påpekte at det på daværende tidspunkt ikke fantes noen RCT-er av behandlingseffekt ved førstegangspsykose (1). Tidligere studier har inkludert pasienter med tidligere episoder og etablert medikamentbruk, noe som gjør resultatene vanskelige å tolke – slik Aarre også påpeker.

Siden 2018 er det publisert noen få studier på medikament-naive pasienter med førstegangspsykose. Francey et al. (2020) fant at psykososial behandling ga like gode langtidsresultater uavhengig av om pasientene fikk nevroleptika eller placebo (2). Morrison et al. (2020) rapporterte lignende funn i en RCT hvor tillegg av nevroleptika ikke ga bedre effekt enn kognitiv terapi alene (3). Folkehelseinstituttet fant i sin nye systematiske gjennomgang i 2021 kun disse to forskningsmiljøene som hadde publisert RCT-er på denne gruppen (4). Selv om studiene er små, gir de grunnlag for å stille spørsmål ved om nevroleptika gir vesentlig bedring for flertallet ved førstegangspsykose.

Tidlige, upubliserte data fra HAMLETT-studien (Nederland, 2024) tyder på at pasienter i remisjon etter seks måneder hadde bedre langtidsutfall ved nedtrapping og seponering enn ved vedlikeholdsbehandling (5).

Denne kunnskapen gjør oss ydmyke i møte med pasienter og foreldres ambivalens. Kanskje er det for en gruppe pasienter bedre å friskne til uten medikamentbruk og vi har per i dag ikke sikker kunnskap om for hvem vi skal insistere på nevroleptikabehandling. Undertegnede Taran Buran Nærdal har selv erfaring med at godt behandlingsresultat er mulig uten bruk av nevroleptika (6).

Vi er bekymret for at Helsedirektoratets reviderte retningslinjer for behandling av psykose kun omtaler medikamentell behandling, utelater pasienter under 18 år, og ikke drøfter det svake kunnskapsgrunnlaget for behandling av førstegangspsykose.

Litteratur:

Smedslund G, Stoltenberg C. Hva vet vi om langtidsvirkninger av antipsykotika? Tidsskr Nor Legeforen 2018; 138. doi:10.4045/tidsskr.18.0729
Francey SM, O’Donoghue B, Nelson B et al. Psychosocial intervention with or without antipsychotic medication for first-episode psychosis: A randomized noninferiority clinical trial. Schizophrenia Bulletin Open 2020; Volume 1(1), sgaa015. https://doi.org/10.1093/schizbullopen/sgaa015
Morrison AP, Pyle M, Gumley A et al. Antipsychotic medication versus psychological intervention versus a combination of both in adolescents with first-episode psychosis (MAPS): a multicentre, three-arm, randomised controlled pilot and feasibility study. The Lancet Psychiatry 2020; 7(9): 788 - 800.
Folkehelseinstituttet (FHI). Systematic review. The effect of antipsychotics on first episode psychosis. 17.11.2021. https://www.fhi.no/en/publ/2021/The-effect-of-antipsychotics-on-first-episode-psychosis/ Lest 1.4.2025.
Johnson K. Antipsychotic Tapering in FEP Beneficial in the Long-term. Medscape Psychiatry. 27.9.2024. https://www.medscape.com/viewarticle/
antipsychotic-tapering-fep-beneficial-long-term-2024a1000hm2?ecd=WNL_mdpls_240927_mscpedit_psych_etid6862511&uac=
105447PY&spon=12&impID=6862511 Lest 1.4.2025.
Nærdal TB. Førstegangspsykose behandlet uten nevroleptika. Tidsskr Nor Legeforen 2021; 141. doi: 10.4045/tidsskr.21.0317

28.04.2025:

Opptil majoriteten av alle som blir gitt antipsykotika utvikler tardiv dyskenesi (1). Burde ikke dette kalkuleres inn i tvangsmedisineringstiltakene?

Litteratur:
1) Wikipedia. Tardive dyskinesia. https://en.wikipedia.org/wiki/Tardive_dyskinesia Lest 28.4.2024.

28.04.2025:

Trond Aarre takkes for at han har tatt opp et viktig og aktuelt tema i sin kronikk «Har antipsykotika effekt mot akutt psykose?» Hans konklusjon er gjennomgående et nei. Kronikken har imidlertid noen manglende definisjoner av begreper og referanser til litteratur som gjør at hans konklusjoner kan misforstås.

Vi kan i denne sammenheng anse begrepet «antipsykotika» som rimelig entydig, selv om andre preparater som har positive effekter ved ulike typer psykoser, ikke har «dopamin-blokkade». «Effekt», «akutt», og «psykose» er imidlertid vanskeligere å definere. Det er likevel ganske selvsagt at antipsykotika har «en slags effekt». De er biologisk aktive stoffer slik Aarre også selv påpeker.

Det finnes ingen unisont akseptert definisjon på begrepet «psykose». Det som imidlertid ofte skjer i daglig, klinisk kommunikasjon, er at begrepet og derav «antipsykotisk effekt» oppfattes som en «anti-schizofren effekt». Med det tenker man seg at «antipsykotika» gjør noe positivt for mennesker med kroniske alvorlige sinnslidelser i et langtidsperspektiv.

Hva mener så Aarre med «akutt»? Er det en nyoppstått tilstand som krever akutt intervensjon, eller er det en tilstand som klinikeren i akuttsituasjonen vurderer prognostisk flere år fremover?

Det er få klinikere som betviler den akutte, klinisk gode effekt av antipsykotika på symptomer, funksjon og adferd ved gjennombrudd av flere typer alvorlig psykose (1). Det er god effekt på agitasjon, psykotisk «katastrofeangst», søvnløshet og uro ved akutte forverringer ved kronisk, schizofren grunnlidelse. Dette er også symptomer på «akutt psykose».

Alvorlige affektive lidelser eventuelt kombinert med umiddelbar selvmordsfare er klinisk mye hyppigere tilstander enn det vi liker å tro. Det er et flytende skille mellom «lite håp», «uttalt pessimisme», «depressivt forvrengt selvbilde», og depressiv psykose. Hvilke symptomer behandlere får fram vil være avhengig av samarbeidsklima, om pasienten «åpner seg», tid på døgnet mm. Ved alvorlige affektive tilstander har antipsykotika en åpenbar plass som adjuvans i den farmakologiske behandlingen. Det gir umiddelbar reduksjon av agitasjon, reduserer angst og «tankekjør», og det får den søvnløse til å sove. Dermed reduseres eventuell selvmordsfare raskt. Antipsykotika har også en dokumentert plass i behandlingen av pasienter med akutte organiske psykoser. Det gjelder også post-ictale, epileptiske psykoser til tross for at antipsykotika senker krampeterskelen.

En helt annen vurdering er effekter av antipsykotika på lang sikt eller i et livsløp. Hvilke faktorer er da mest betydningsfulle for pasienten? Mange kan enes om at terapi som endrer det «naturlige forløp» av tilstandene til bedret livskvalitet er det essensielle. Her er dokumentasjonen for antipsykotika svært sparsom eller fraværende. Ghaemis artikkel anbefales som bakgrunn for en videre diskusjon (2).

Litteratur:
1. Garriga M, Pacchiarotti I, Kasper S, Zeller SL, Allen MH, Vazquez G, et al. Assessment and management of agitation in psychiatry: Expert consensus. World J Biol Psychiatry. 2016; 17(2): 86-128. doi: 10.3109/15622975.2015.1132007.
2. Ghaemi SN. Symptomatic versus disease-modifying effects of psychiatric drugs. Acta Psychiatr Scand. 2022;146(3):251-257. oi: 10.1111/acps.13459.

05.05.2025:

Eg takkar Arne Vaaler for ein tankevekkjande kommentar og vil her utdjupe litt av det han tek opp.

Vaaler viser til at det er mange ulike definisjonar av omgrep på dette feltet. Eg har ikkje problematisert alle desse, men har teke forskarane på ordet og prøvd å formidle deira funn på ein uhilda måte. Når eg nyttar omgrepet «akutt» om behandling med nevroleptika, dreier det seg ikkje om nyoppstått psykose. Eg viser i kronikken til at det ikkje finst placebokontrollerte studiar å stø seg på når det gjeld fyrstegongspsykose. «Akutt psykose» vert, som i den vitskaplege litteratuen, nytta for å skilje bruken ved meir eller mindre akutte forverringar frå den tilbakefallsførebyggjande langtidsbehandlinga som også går føre seg når personen det gjeld er heilt utan plager.

Vaaler skriv av «Det er få klinikere som betviler den akutte, klinisk gode effekt av antipsykotika på symptomer, funksjon og atferd ved gjennombrudd av flere typer alvorlig psykose. Det er god effekt på agitasjon, psykotisk 'katastrofeangst', søvnløshet og uro ...». Han har tvillaust rett i at det er få som tvilar, men spørsmålet eg stiller er ikkje om fagfolks tru og tvil. Det eg er oppteken av, er kva vitskapleg grunnlag me har for overtydingane våre. Det kan godt vere at midla verkar både mot søvnløyse og uro, men dei kontrollerte studiane viser altså at skilnadene mellom medikamenta og placebo er små.

Konteksten for kronikken min er Helsedirektoratet sine retningsliner for medikamentell behandling ved «psykoselidingar». Det er dokumentasjonen for direktoratet sine tilrådingar eg har skrive om. Vaaler tek opp spørsmålet om bruk av nevroleptika ved andre tilstandar, som alvorlege affektive lidingar og akutte organiske psykosar. Kronikken min dreier seg ikkje om dette, så slike spørsmål må vurderast for seg. Vaaler har rett i at eg kunne presisert det klårare.

Skriv Kommentar

Published: 9 April 2025

Tidsskr Nor Legeforen 9 April 2025 Vol. 145.

doi:

10.4045/tidsskr.25.0047

Received 22.1.2025, first revision submitted 16.2.2025, accepted 24.2.2025.

Published: 9 April 2025

Tidsskr Nor Legeforen 2025 Vol. 145.

doi: 10.4045/tidsskr.25.0047

Received 22.1.2025, first revision submitted 16.2.2025, accepted 24.2.2025.

PDF

Print

Reply to article

Are antipsychotic drugs effective against acute psychosis?

The recommendations issued by the Directorate of Health

Methodology problems

What is meaningful improvement?

Improvement in the placebo groups

How many experience meaningful effect?

Good enough for coercive treatment?

Minimal efficacy in most cases

Hva betyr dette for behandling ved førstegangspsykose?

Tardiv dyskenesi

Antipsykotika ved akutte psykoser.

T. Aarre svarer A. Vaaler

Recent Articles