antipsychotic medication
Transcription
antipsychotic medication
Matti Isohanni, Professor of Psychiatry (emeritus). Department of Psychiatry, Institute of Clinical Medicine, University of Oulu, Finland Midlife progression in brain tissue loss and clinical outcomes of schizophrenia in the Northern Finland 1966 Birth Cohort (NFBC 1966) Conflicts of interests No conflicts of interest to declare This work was supported by unrestricted grants from • • • • • the Stanley Medical Research Institute, the Isaac Newton Trust, the Academy of Finland, the Sigrid Juselius Foundation, The Brain & Behavior Research Foundation General aims: We analyzed midlife progression of schizophrenia at ages 34 to 43 years (9 year follow-up) and its correlates to changes in brain morphometric changes, antipsychotic medication, cognition, and clinical outcomes: 1) What happens in outcomes (clinical, cognitive)? 2) What happens in the brain? 3) The effects of antipsychotic medication? 4) Any clinical conclusions? Is recovery possible in schizophrenia? • Clinical and social recovery is possible, median 13.5% (meta-analysis by our group; Jääskeläinen et al. Sch Bull 2013); • In old age sz more common: 18-27% (Jääskeläinen et al Sch Bull 2013) • Recoveries have not increased during the last decades, not even during the antipsychotic era – but decreased! Jääskeläinen et al. Meta-analysis Sz Bull 2013 Why so few recover? • • • • (Partly) treatment resistant disease? Poor treatments and medications? Poor compliance and adherence? Demanding modern social environment and occupational life? • Iatrogenic hypothesis: antipsychotics prevent benign natural course in a subsample and cause poor recovery (Harrow et al Psychol Med 2014)? Methods Northern Finland 1966 Birth Cohort (NFBC 1966) (N=12058) has been followed serially since mid-pregnancy at birth, at ages 1,14, 31 and 46 years. Data linkage to national case etc registers, hospital charts Psychiatric follow-up studies at age 34- and 43-years: structural and functional MRI, cognitive, and clinical analyses were performed at the beginning and end of this 9-year followup. 1. at age 34 years: 73 DSM schizophrenic psychoses and 104 controls (mean time since illness onset 11 years) 2. At age 43 years: 55 schizophrenic psychoses and 192 controls (mean time since illness onset 20 years) 41 sz cases and 75 controls participated in both surveys Methods * Dallas * Augusta * Oulu Brisbane * General aims: We analyzed midlife progression of schizophrenia at age 43 years and its correlates to changes in brain morphometric changes, antipsychotic medication, cognition, and clinical outcomes: 1) What happens in outcomes (clinical, cognitive)? 2) What happens in the brain? 3) The effects of antipsychotic medication? 4) Clinical conclusions 1) What happens in outcomes? Cross-sectional outcome data at age 34 years in NFBC 1966 (Lauronen et al. 2007, Miettunen et al. 2006) The outcome of schizophrenic psychoses is heterogeneous: •minority of individuals experience recovery (3.4%) •some but not many achieve at least periods of symptomatic remission (23%) – 77% not. •many are on disability pension (54%) •most require re-hospitalisation (81%) 1) What happens in outcomes at age 43? Cross-sectional outcome data from the latter follow-up at age 43 years in NFBC 1966 (Jääskeläinen et al. manuscript): (at age 34 years similar as at age 43; Lauronen et al. Sch Res 2008) ▪ 2.8% (one person) recovered ▪ 22% achieved remission ▪ 36% were not on disability pension ▪ 54% used no or at most 300mg/day antipsychotic medication (as chlorpromazineequivalents) Prodictors of poor outcomes • • • • • • • Early onset age Insidious onset Long DUP Suicidal ideations at the beginning Poor scholastic performance Deviant birth weight and length Reduced density (volume) in frontal and temporal areas and white matter Cognitive performance • Neurocognitive performance, especially visual learning and memory associated to poor functional or global outcomes both in the longitudinal and cross-sectional analyses (Juola et al in review). • Cognitive performance decreased in 9 years between 34 and 43 years, and this was also the case in controls • High antipsychotic dose years were related to poor cognition (adjusted with multiple clinical aspects) Husa et al Sch Res 2014) “The results do not support the view that antipsychotics in general prevent cognitive decline or promote cognitive recovery in schizophrenia” (Husa et al Sch Res 2014) General aims: We analyzed midlife progression of schizophrenia at age 43 years and its correlates to changes in brain morphometric changes, antipsychotic medication, cognition, and clinical outcomes: 1) What happens in outcomes (clinical, cognitive)? 2) What happens in the brain? 3) The effects of antipsychotic medication? 4) Clinical conclusions Results (Veijola et al PLoS ONE 2014) Differences in total brain volume change between patients and controls between ages 34 to 43: •The mean annual total brain volume reduction was 0.69% in schizophrenia and 0.49 % in controls. The reduction was significant after adjusting for sex, educational level, alcohol use and weight change. Does change in brain volume predict clinical outcomes in schizophrenia? •In linear regression, total brain volume reduction was not associated with the mean total PANSS score, change in the total PANSS score, mean SOFAS score or change in SOFAS score in schizophrenia. Brain volume reduction also did not associate with means of PANSS sub-scores or changes in PANSS sub-scores. General aims: We analyzed midlife progression of schizophrenia at age 43 years and its correlates to changes in brain morphometric changes, antipsychotic medication, cognition, and clinical outcomes: 1) What happens in outcomes (clinical, cognitive)? 2) What happens in the brain? 3) The effects of antipsychotic medication? 4) Clinical conclusions How to measure lifetime use of antipsychotic medication? • Current and lifetime use of antipsychotic medication was ascertained from interviews, registers, hospital and outpatient case notes. • We measured the chlorpromazine equivalents (Kroken et al BMC Psych 2009) and used daily dose of 100 mg of chlorpromazine for calculating dose years of lifetime antipsychotic medication (Andreasen et al. Biol Psych 2010, Husa et al Sz Res 2014, Veijola et al PLoS ONE 2014). • Daily 100 mg chlorpromazine eq during 9 years follow-up is 9 dose years • Daily 350 mg chlorpromazine (mean approx dose in NFBC 1966) during 9 years follow-up is 32 dose years Results (cont): Is exposure to antipsychotic medication associated with loss of total brain volume over time? The annual reduction in brain volume was associated with higher number of dose years of antipsychotic medication (linear regression; p=0.003). After adjusting for symptom severity and social functioning (mean total PANSS score, change in total PANSS score, mean SOFAS score, change in SOFAS score), logarithm of alcohol use and weight gain the association still remained significant. (Veijola et al PLoS ONE 2014) In separate linear regression analyses, the annual reduction in brain volume was associated with higher number of dose years of typical and atypical antipsychotic medication (typical p=0.034; atypical p=0.005); no interaction by type. Linear regression between dose years (equivalent to daily 100 mg chlorpromazine) of antipsychotics and annual change in brain volume (%) in subjects with schizophrenia (beta=-0.47, t=-2.98, p=0.006). High exposure to antipsychotic medication associated with loss of brain volume over time. Daily 350 mg chlorpromazine (mean approx dose in NFBC 1966) during 9 years follow-up is 32 dose years (Veijola et al PLoS ONE 2014). Annual brain volume reduction (%) in control subjects and in subjects with schizophrenia in three dose year (equivalent to daily 100 mg chlorpromazine) categories of antipsychotic medication. E.g. daily 350 mg chlorpromazine during 9 years is 32 dose years. (Veijola et al PLoS ONE 2014) Results (cont): Regional voxel-based analyses FSL voxelwise SIENA (Guo et al in review) Progressive brain volume reduction in schizophrenia during the 9-year follow-up was widespread and found especially in temporal lobe and periventricular area. Total antipsychotic medication exposure was associated with reductions around ventricles No significant differences in the effects of typical and atypical antipsychotic medication. However, power limited Decline in social and occupational function was associated with right supramarginal gyrus reduction and was independent from antipsychotic medication exposure Regions with greater loss of brain volume in schizophrenia during 9-year follow-up compared to controls (adjusted for sex, handedness and interscan interval, p < 0.05 family-wise error corrected across brain edge): right frontal pole, right middle and inferior frontal gyrus, bilateral central gyrus, bilateral supramarginal gyrus, bilateral temporal lobe (including superior, middle and inferior divisions), right occipital cortex inferior division, right lingual gyrus (mid-line occipital lobe), and bilateral precuneous cortex. (Guo et al, in review) Total antipsychotic medication exposure during 9 year follow-up was associated with reductions around ventricles (Guo et al in review) Guo et al in preparation Discussion • The disease process itself may cause the decrease (Lieberman 1999) • Do iatrogenic factors such as antipsychotic medication contribute to the excessive brain tissue loss? (Ho et al. 2011, Van Haren et al. 2011). Key question in clinical psychiatry! Discussion (cont) Progressive volume changes have previously been demonstrated in schizophrenia patients selected from clinic or ward based cohorts, in comparison to controls who have been recruited from the community by advertisement (Hulshoff Pol and Kahn 2008, Olabi et al 2011). Our study is the first to report progressive loss of brain volume in patients and controls recruited in a population-based sample and midlife (where most patients are). There was a dose-response relationship in schizophrenia between cumulative exposure to antipsychotic medication and brain volume loss over time, providing some evidence that antipsychotic medication may cause loss of brain volume over Discussion (cont) We addressed potential confounding factors by controlling for illness severity, yet the association between brain volume change and medication dose remained statistically significant. However, in naturalistic setting there is possibility for residual, latent confounding – very sick patients get higher doses for longer duration. Do typical and atypicals differ? • Not in this study. Both typical and atypical medications predicted brain volume loss. However, number of cases and statistical power was limited. • They differ from some results, which found that typical, but not atypical, antipsychotic medication predicted brain cortical thinning in schizophrenia (Lieberman et al AGP 2005, Van Haren AGP 2012). • Both typical and atypical antipsychotic agents have been shown to produce loss of brain volume in experimental rodents (Vernon et al Biol Psych 2011) and monkeys (Dorph-Peterson et al Neuropsychopharm 2005). Meta-analysis: Longitudinal studies on antipsychotics and brain volumes. Huhtaniska et al in submission Heterogeneous methods and findings! In total 29 publications from 16 different samples fulfilled our inclusion criteria: scan interval over 2 years In meta-analysis higher antipsychotic exposure associated statistically significantly with decrease in parietal lobe (n=4; r=-0.14, p=0.013) and with increase in basal ganglia volume (n=4; r=0.10, p=0.044). In general, most of the reported correlations between brain volume alterations and antipsychotic dose were statistically non-significant. Meta-analyses brain volumes and antipsychotic exposure by tissue type and brain area combinations. (Huhtaniska et al in submission) studies n GM - total - cerebrum - frontal lobe WM - cerebrum CSF - CSF and ventricles Volume - total brain - cerebrum - frontal lobe - temporal lobe - parietal lobe - occipital lobe - cerebellum - limbic system - basal ganglia random effect meta-analysis correlation (95% z test CI) sign. n 3 3 3 0.04 (-0.10-0.17) 0.08 (-0.29-0.45) -0.04 (-0.22-0.13) 0.55 0.41 0.46 0.58 0.68 0.64 1 2 2 3 0.04 (-0.07-0.16) 0.72 0.47 1 6 0.10 (-0.09-0.28) 1.03 0.31 4 77 -0.06 (-0.23-0.10) -0.08 (-0.19-0.02) -0.16 (-0.34-0.02) -0.12 (-0.27-0.03) -0.14 (-0.26- -0.03) -0.06 (-0.24-0.11) 0.01 (-0.09-0.11) -0.03 (-0.13-0.08) 0.10 (0.002-0.19) 0.73 1.54 1.71 1.54 2.49 0.68 0.20 0.48 2.01 0.47 0.12 0.09 0.12 0.013 0.50 0.84 0.63 0.044 3 2 6 6 3 2 3 2 4 80 0 67 64 0 0 0 0 0 6 3 8 10 4 3 5 3 4 hete Meta-analysis was done for categories with at least three samples. 90 81 General aims: We analyzed midlife progression of schizophrenia at age 43 years and its correlates to changes in brain morphometric changes, antipsychotic medication, cognition, and clinical outcomes: 1) What happens in outcomes (clinical, cognitive)? 2) What happens in the brain? 3) The effects of antipsychotic medication? 4) Clinical conclusions What is the significance of loss of brain volume in schizophrenia? • At present, the clinical significance of antipsychotic induced brain volume reduction is uncertain (harmful loss? harmless loss? delayed pruning?) • Do we scare - without scientific arguments – numerous patients and clinicians that antipsychotics shrink the brain? • Loss of brain volume occurs throughout adult life in the healthy population. • Optimising and minimising antipsychotic medication might, however, be one targeted intervention. • Main reason are neurological, metabolic and cognitive side effects, not speculative brain effects Clinical conclusions (cont) • Antipsychotic medications are used increasingly on-label or off-label (Rissanen et al Human Psychopharm 2013), including dementia, bipolar disorder, unipolar depression, and in a variety of childhood psychiatric conditions • Resposes to antipsychotics are heterogeneous and individualized. Group-level effect size is about 0.4 but individual responses vary from 0 to 0.8. • Antipsychotics may have potentially serious side-effects (in addition to neurological and metabolic) and should be used with considerable caution, especially in conditions where their efficacy has not been established • It is possible that some schizophrenia cases with small relapse risk are overmedicated Clinical conclusions (cont) • “Nothing is known about the effects of antipsychotic drugs compared to placebo after three years” (Leucht et al Cochr Database 2012) • A subsample of sz patients may show good outcomes without prolonged antipsychotic treatment (Bleuler 1978, Fenton and McGlashan 1987, Harding et al 1987, Wunderink et al 2013, Harrow et al 2014) • Reduced use or dose is one alternative (Isohanni 1983, Ciompi and Hoffman 2004, Bola et al 2009) • In a subsample antipsychotics may cease to have positive effects and long-term efficacy: excess or supersensitive dopamine receptors may cause this This is very common phenomenon in somatic medicine (cancers, asthma, PD) Wunderink 2013; reduction Wunderink 2013; maintenanc e 14 (160mg/day: Isohanni 1983 cases) 13,5 (150mg/day) 33 (370mg/day: Isohanni 1983 controls) 81 (900mg/day) Linear regression between dose years (equivalent to daily 100 mg chlorpromazine) of antipsychotics and annual change in brain volume (%) in subjects with schizophrenia (beta=-0.47, t=-2.98, p=0.006). High exposure to antipsychotic medication associated with loss of brain volume over time. Daily 350 mg chlorpromazine (mean approx dose in NFBC 1966) during 9 years follow-up is 32 dose years (Veijola et al in revision). Is it possible to reduce dose or discontinue medication? • The first long-term (7-year) follow-up of a randomized early dose reduction/discontinuation clinical trial was published recently by Wunderink et al (2013). • The study demonstrates superior long-term functional outcome in the dose reduction/discontinuation group, even though the initial relapse rate was about two-fold higher. • In dose reduction group, the mean daily dose (as haloperidol equivalents) was quite low (2.79 mg) compared to the maintenance group (4.08 mg). • In most treatment algorithms, the recommend range in maintenance treatment varies between 150-900 mg/day as chlorpromazine equivalents or 3-18 mg as haloperidol equivalents, to be continued at least 1 year after the first episode and 5 years after multiple episodes (Gaebel et al 2005). Difficulties to predict patients advancing tapering • There seems to be a subgroup of patients (15-20%?) with schizophrenia who manage well especially at the stabilized phase, especially with regards to functional remission, with small doses of antipsychotics, and even without permanent antipsychotic medication. Other (poorly known and partly overlapping) subgroups may be: treatment-resistants, ex-responders, noncompliance patients, etc. • Presently, there are no replicated predictors of patients who respond well to a dose reduction/discontinuation strategy. • Maximal use of psychosocial interventions reduced the dose of antipsychotics (as chlorpromazine equivalents) in an acute psychosis ward from 370 mg/day to 160 mg/day (Isohanni 1983), which is the level identical to Wunderink’s et al (6) dose reduction/discontinuation group. • This result suggests that in a well-organized treatment entity and individualized treatment program with a lot of non-pharmacological care (McGorry et al 2013, McGurk et al 2013, Mueser et al 2013), patient and family guidance, and accurate diagnostics, dose reduction may be a more relevant strategy, contrasted to a situation where antipsychotics are the only active treatment. Future research is needed • Most studies supporting maintenance treatment on antipsychotic medication are performed with relatively short follow-ups and initial phase. • However, most patients with schizophrenia are in midlife and the duration of illness is 10-20 years. • Randomized controlled trials have methodological problems: selection and attrition biases, poor compliance, short follow-up periods, slow study processes. • Case-control studies do not detect the heterogeneity of effects and different individual/subgroup effect sizes • In an observational, naturalistic setting (e.g registers and cohorts in Nordic Countries), there are many methodological challenges. The patients are not treated randomly. This may cause residual, latent confounding: very sick patients often get higher doses and a longer duration of antipsychotic medications. This problem can often be controlled Will clinical decision-making change? • Ethical principles provide guidance, especially Hippocrates´ principle of rejecting mischief. He teaches us: primum non nocere (first, do not harm). • This is one reason to consider the minimizing antipsychotic doses and maximizing psychosocial treatments. • However, even a very experienced clinician may be prone to “pessimism or clinical bias”: he often meets selectively difficult, treatment-resistant and complex patients, but cases with good response and recovery are lost to follow-up. • All this may reduce his courage to taper or discontinue antipsychotic medication and leave these decisions solely to the patients. • Current care guidelines are undifferentiated and vague in regards to dose tapering and discontinuation, and to recommended psychosocial interventions. • Current guidelines make low doses and discontinuation possible, but do not specifically encourage this strategy or suggest how to go about tapering antipsychotic doses (i.e., at what point in the clinical course of illness, over what time period, etc.). Take home message • We do not know the real effectiveness and benefits/risks of antipsychotic drugs from longitudinal (midlife schizophrenia) and lifespan view. • Stochastic factors and chance (e.g. non-compliance) influence (too) much to antipsychotic practices in longitudinal perspective • Scientific understanding of many major somatic disorders (e.g. cancers, cardiovascular diseases) has increased dramatically. Such a breakthrough has not been achieved in schizophrenia. • To minimise risks we may overmedicate many patients which harms some • Most patients do not have optimal non-pharmacological or pharmacological care (In NFBC 1966 one third of overtly psychotic patients had no contact to any treatment systems; Nykänen et al in preparation) • New data does not change current care guidelines but may expand clinical decision making into lower doses, stopping medication, and biopsychosocial care. Acknowledgements: This work was supported by grants from • the Stanley Medical Research Institute, • the Isaac Newton Trust, • the Academy of Finland, • the Sigrid Juselius Foundation, • NARSAD Software development for image analysis was supported by a • Human Brain Project grant from the National Institute of Biomedical Imaging & Bioengineering and • The National Institute of Mental Health. Thanks! For giving data and slides: Erika Jääskeläinen, Jouko Miettunen, Graham Murray, Joyce Guo, Juha Veijola, Sanna Huhtaniska, Salla Nykänen UOulu: Matti Isohanni, Juha Veijola, Jouko Miettunen, Erika Jääskeläinen (nee Lauronen), Sanna Huhtaniska, Päivikki Tanskanen, Hannu Koponen, Antti Alaräisänen, Marianne Haapea, Salla Nykänen, Pirjo Mäki, Matti Penttilä, David Cowling, Irina Rannikko, Irene Isohanni, Kristiina Moilanen, Johanna Koskinen, Marjo-Riitta Järvelin, Anja Husa, Vesa Kiviniemi, Osmo Tervonen. UEastern Finland/Helsinki: Hannu Koponen, Heli Koivumaa-Honkanen, Jari Tiihonen UCambridge: Graham Murray, Joyce Guo, Peter Jones, Khanum Ridler, Edward Bullmore, Anna Barnes, J. Suckling Medical College of Georgia: Brian Miller, Brian Kirkpatrick UTSW, Dallas: Carol Tamminga, Subroto Ghose, Elena Ivleva UQueensland: John McGrath, Joy Welham LMU, Munich: Hans-Jurgen Möller Finland: about 11000 Cohort members and their relatives, about 100 scientific partners