Table 4 Summary of included randomised controlled trials
Author/Year, Country | Study Design | Participants (n, age, gender, ethnicity, level of intellectual disabilities) | Setting | Intervention (including medication that was targeted, Duration of deprescribing intervention and length of follow up) | Outcome Measures | Summary of Findings |
|---|---|---|---|---|---|---|
Research Units on Pediatric Psychopharmacology Autism Network. 2005 USA [20] | Randomised Controlled Trial (RCT) | n: 38 mean age: 9 years old (5 to 17) gender: 87% male ethnicity: not reported ID = mild 8 (21%) moderate 6 (16%) severe 7 (18%) profound 6 (16%) plus 4 incomplete data and remainder borderline or above average IQ | Community | Intervention: Part of a two-stage study. Maintenance dose reduced by 25% per week in the experimental group. Control group continued risperidone. Medication: risperidone Duration: reduced over 3 weeks Length of follow up: 5 weeks after planned discontinuation | ABC CGI | 84% completed discontinuation After 32 participants completed the study, trial was stopped. Relapse rates were 62.5% for gradual placebo substitution and 12.5% for continued risperidone; difference was found to be statistically significant. |
Ahmed et al. 2000 U.K. [21] | RCT | n = 56 (Includes participants also reported in Smith et al., 2002 [22]) mean age = 43(20 to 78) gender: 48% male ethnicity: not reported ID = incomplete data | 45% NHS hospitals 9% NHS Community unit 46% Community residential homes | Intervention: Thirty-six participants randomly allocated to the experimental group underwent four, monthly 25% drug reduction stages. There were no planned drug changes for the control group Medication: Participants received 12 different antipsychotic drugs, most frequently thioridazine (18 people, 12%), haloperidol (13, 23%) and chlorpromazine (8, 14%). Duration: Reduced over 4 months Length of follow up: 1 month after planned discontinuation | ABC, DISCUS, weighing scales, direct observation using palm-top Psion 3a portable computers, Number of participants Successfully deprescribed | 12 participants (33%) completed full withdrawal 7 participants (19%) achieved and maintained at least a 50% reduction. Drug reduction was associated with increased dyskinesia and higher activity engagement but not increased maladaptive behaviour. Some setting characteristics were associated with drug reinstatement. |
de Kuijper et al. 2014 [23] The Netherlands | RCT | n: 98 (Includes participants also reported in de Kuijper et al.,2013 [24], de Kuijper et al. 2014 [25] and de Kuijper et al. 2018 [26]) mean age: 49.8 (15 to 66) gender: 64% male ethnicity: not reported ID: profound 35 (36%) severe 26 (27%) moderate 30 (31%) mild 7 (7.1%) | Community | Intervention: Participants underwent 12.5% antipsychotic dose reduction every 2 or 4 weeks Medication: 65 pipamperone, 18 haloperidol, 15 risperidone, 8 olanzapine, 7 levomepromazine 1 pimozide Duration: Reduced over 14 or 28 weeks Length of follow up: 12 weeks after planned discontinuation | Primary outcome: ABC (irritability subscale) Secondary outcomes: other ABC subscales, CGI, SCOPA-AUT, Epworth Sleepiness Scale, AIMS, Barnes, Unified Parkinson’s Disease Rating Scale Physical Health parameters- weight, BP, lipids, waist circumference, pulse, prolactin, testosterone, Number of participants Successfully deprescribed | Of 98 participants, 43 achieved complete discontinuation; at follow-up 7 had resumed use of antipsychotics. Mean ABC ratings improved significantly for those who achieved complete discontinuation and at follow-up for those who had not achieved complete discontinuation. Similar results with respect to most ABC sub-scales, including the ‘irritability’ subscale. No significant differences in improvement of ABC ratings between both discontinuation schedules. Higher ratings of extrapyramidal and autonomic symptoms at baseline associated with less improvement of behavioural symptoms after discontinuation; higher baseline ABC rating predicted higher odds of incomplete discontinuation. |
de Kuijper et al. 2013 [24] The Netherlands | RCT additional reporting of [23] | (Includes participants also reported in de Kuijper et al., 2014 [23, 24]) | ’ |  | Fasting glucose, triglycerides, high density lipoproteins, low-density lipoproteins, and total cholesterol in blood; height, weight, and waist circumference and systolic and diastolic blood pressure | Discontinuation of anti-psychotics led to a significant decrease in waist circumference, weight, BMI, and systolic blood pressure. Higher baseline dosage associated with a larger decrease in waist circumference, weight, and BMI in these participants. No significant difference between discontinuation in 14 or 28 weeks. Dosage reductions associated with decrease in weight and BMI, negatively associated with metabolic outcomes such as fasting plasma glucose levels. |
de Kuijper et al. 2014 The Netherlands [27] | RCT additional reporting of [23] | (Includes participants also reported in de Kuijper et al.,2013 [24] and de Kuijper et al.,, 2014 [23]) | Â | , | Plasma measurements included prolactin, testosterone (only in male participants), 25-OH vitamin D, PTH, and bone turnover markers, ie, bone alkaline phosphatase (BALP), aminopropeptide type I collagen (PINP), and C-telopeptide type I collagen (CTX). | Both complete discontinuation and dosage reduction led to decrease in prolactin plasma levels and to increase in levels of CTX, the bone resorption marker. Dose reductions associated with a significant decrease in 25-OH vitamin D levels, with less weight loss and higher BMI compared with those who had completely discontinued. More weight loss associated with less difference in baseline/follow-up CTX levels and with less difference in baseline/follow-up 25-OH vitamin D levels. |
Hassler et al. 2007 Germany [28] | RCT | n: 39 (Includes 31 participants also reported in de Hassler et al., 2011 [29]) mean age: 36.4 (SD 10.4) gender: 54% male ethnicity: 100% white ID: severe 28 (72%) moderate 9 (23%) mild 2 (5%) | Inpatient | Intervention: Random allocation of withdrawal of medication after a 6 week period of open treatment Medication: zuclopentixol Duration: Sudden discontinuation Length of follow up: 12 weeks after discontinuation | Primary outcome: MOAS Secondary outcome: withdrawal symptoms, extrapyramidal signs, vital signs, weight, and routine laboratory tests of prolactin and serum levels of zuclopenthixol were conducted. | The placebo group was associated with more aggressive behaviour as indicated by outcomes observed by external raters. |
Hassler et al. 2011 Germany [29] | RCT additional reporting of [28] | n: 31 (The participants were also reported in Hassler et al., 2007 [28]) mean age: 38.4 (adults) gender: 55% male ethnicity: 100% white ID = not reported | Inpatient | Intervention: Prospective follow up of an RCT in which participants who remained on medication were compared to those participants who discontinued during a 2 year period Medication: zuclopentixol Duration: Not reported Length of follow up: Variable | MOAS DAS CGI-I Body weight | Patients still treated with zuclopentixol after 2 years (n = 21) benefitted, compared to the patients who discontinued (n = 10) For continually treated patients, no adverse events, side-effects, or treated extrapyramidal symptoms were reported. They lost on average 1.8 kg body weight. Patients who discontinued zuclopentixol on average gained 2.6 kg body weight. |
Heistad et al. 1982 USA [30] | RCT | n: 100 mean age: 28.5 (13 to 65) ethnicity: not reported gender: 54% male ID: >  50% profound | Inpatient | Intervention: 5 separate arms to assess effect of withdrawal. Details not reported. Patients selected were rank ordered by current drug dose, and one member of each successive pair was randomly assigned to either the drug-placebo or the placebo-drug sequence. Medication: thioridazine Duration: variable Length of follow up: 4 to 5 weeks after discontinuation | Unvalidated adapted behaviour coding system, NOSIE, AIMS | Time-sampling of behaviour showed significant increase in self-stimulation and active negative behaviour and decreased work and life skills while receiving placebo. Most patients’ behaviour was better while on active medication, some showed significant improvement when medication was temporarily discontinued. Favourable long-term progress among those who had medication restored was greater for patients whose behaviour had worsened to the greatest degree during the placebo (discontinuation) trial. |
McNamara et al. 2017 U.K. [31] | RCT | n = 22 mean age: 43 (21 to 68) gender: 68% male ethnicity: not reported ID = not reported | Community | Intervention: Treatment in the intervention group was gradually reduced over a 6-month period and then maintained at the same level for a further 3 months. In the control group, baseline level of medication was maintained throughout the 9-month period. Medication: risperidone Duration: Reduced over 6 months Length of follow up: 6 and 9 months after planned discontinuation | Feasibility outcomes: the Number and proportion of general practices/Community learning disability teams that progressed from initial approach to recruitment of participants and the Number and proportion of recruited participants who progressed through the various stages of the study. Clinical outcomes: MOAS, ABC, PAS-ADD, The ASC, DISCUS, the CSRI, use of other interventions to manage challenging behaviour, use of PRN medication and level of psychotropic medication use. | Of the 22 participants randomised (intervention, n = 11; control, n = 11), 13 (59%) achieved progression through all four stages of reduction. Follow-up data at 6 and 9 months were obtained for 17 participants (intervention, n = 10; and control, n = 7; 77% of those randomised). No clinically important changes in participants’ levels of aggression or challenging behaviour reported. |
Ramerman et al. 2019 The Netherlands [32] | RCT | n:25 (11 participants also reported in Kuijper et al., 2018 [33] and Ramerman et al.,2019 [34]) mean age: 30 gender: 76% male ethnicity: not reported ID: mild 52% moderate 24% severe 24% profound 0% | Inpatient | Intervention: In the discontinuation group, Risperidone was gradually replaced by a placebo over 14 weeks, while the control group maintained their existing dosage. Medication: risperidone Duration: Reduced over 14 weeks Length of follow up: 8 weeks after planned discontinuation | ABC CGS-I, SCOPA-AUT, Epworth Sleepiness Scale, AIMS, Barnes, Unified Parkinson’s Disease Rating Scale Physical Health parameters- weight, BP, lipids, waist circumference, pulse, Prolactin, testosterone, Number of participants Successfully deprescribed | In the discontinuation group, 82% completely withdrew from risperidone. No significant change in irritability, compared with the continuation group, although there was Groupa Time effects on stereotypical behaviour in favour of the continuation group. Significant GroupaTime effects were also found for weight, waist, body mass index, prolactin. Levels and testosterone levels, with beneficial effects for the discontinuation group. 2 participants had severe dyskinesia |
Smith et al. 2002 U.K. [22] | RCT additional reporting of [21] | (Participants also reported in Ahmed et al., 2000 [21]) |  |  | ABS, ABC DISCUS direct observation | High Yule’s Q-value results pre- and post-baseline were found, indicating that clients were highly responsive to staff interaction. Yule’s Q-value did not significantly increase following drug withdrawal. |