The authors have declared that no competing interests exist.
Conceived and designed the experiments: CW XL. Performed the experiments: XL CW. Analyzed the data: XL CW. Contributed reagents/materials/analysis tools: XL CW. Wrote the paper: XL YT.
To compare the safety and efficacy of adjunctive aripiprazole versus placebo for antipsychotic-induced hyperprolactinemia.
Population: adult patients presenting with antipsychotic-induced hyperprolactinemia diagnosed by prolactin level with or without prolactin-related symptoms. Interventions: adjunctive aripiprazole vs. adjunctive placebo. Outcome measures: adverse events and efficacy of treatment. Studies: randomized controlled trials.
Five randomized controlled trials with a total of 639 patients (326 adjunctive aripiprazole, 313 adjunctive placebo) met the inclusion criteria. Adjunctive aripiprazole was associated with a 79.11% (125/158) prolactin level normalization rate. Meta-analysis of insomnia, headache, sedation, psychiatric disorder, extrapyramidal symptom, dry mouth, and fatigue showed no significant differences in the adjunctive aripiprazole treatment group compared with the placebo group (risk difference (Mantel-Haenszel, random or fixed) −0.05 to 0.04 (95% confidence interval −0.13 to 0.16); I2 = 0% to 68%, P = 0.20 to 0.70). However, sedation, insomnia, and headache were more frequent when the adjunctive aripiprazole dose was higher than 15 mg/day. Meta-analysis of the prolactin level normalization indicated adjunctive aripiprazole was superior to placebo (risk difference (Mantel-Haenszel, random) 0.76 (95% confidence interval 0.67 to 0.85); I2 = 43%, P<0.00001). The subgroup analysis confirmed that the subjects who received adjunctive aripiprazole 5 mg/day showed a degree of prolactin normalization similar to that of all participants. No significant differences between groups in discontinuation and improvements of psychiatric symptoms.
Adjunctive aripiprazole is both safe and effective as a reasonable choice treatment for patients with antipsychotic-induced hyperprolactinemia. The appropriate dose of adjunctive aripiprazole may be 5 mg/day.
Hyperprolactinemia is a common and serious side effect of antipsychotic treatment
Four strategies have been recommended for the treatment of this side effect. (1) Reduction of antipsychotic dose, but maintenance treatments with a reduction to a lower dose have a higher relapse rate than the no-dose-reduction treatment
On the other hand, the role of aripiprazole in antipsychotic-induced hyperprolactinemia has recently gained some attention. Several case reports and a small open label trial showed the addition of aripiprazole reversed antipsychotic-induced hyperprolactinemia to risperidone
As a newer pharmacologic atypical antipsychotic, adjunctive aripiprazole may optimize D2 receptor activity, and thus reduce risk for extrapyramidal symptoms associated with haloperidol or risperidone
The aim of this meta-analysis of randomized controlled trials was to compare adjunctive aripiprazole treatment with placebo for the treatment of antipsychotic-induced hyperprolactinemia, with particular focus on safety and tolerability.
Randomized controlled trials comparing adjunctive aripiprazole with placebo for antipsychotic-induced hyperprolactinemia in adult patients were eligible for inclusion. We included studies which reported at least one of the outcome measures mentioned below and with well defined treatment programs. We excluded case series, retrospective studies, non-randomized studies and system reviews.
We recorded clinical outcomes based on intent to treat (ITT) analysis if available. The primary outcome measures of this system review were adverse events and efficacy of treatment, as reported in the studies (
Outcome measure | Adjunctive aripiprazole or placebo |
Adverse events | |
Prolactin level normalization | Hyperprolactinemia defined as prolactin level ≥60 µg/L, normal prolactin level was defined as a serum prolactin level of <30 ng/ml for the patients in Chinese study. Normal prolactin level was defined as a serum prolactin level of <24 ng/ml for women and <20 ng/ml for men in the study by shim. |
Prolactin-related symptoms recovery | |
Discontinuation rate | All reasons for study discontinuation were included in meta- analysis. |
Improvement in psychiatric symptoms |
Two authors independently searched the Pubmed, Embase, and Cochrane Library databases for randomized controlled trials comparing adjunctive aripiprazole with placebo for antipsychotic-induced hyperprolactinemia. We also searched the Chinese databases (CBM and CNKI databases) using the same key words. The search included all studies published between January 2001 and December 2012, regardless of language. Keywords used included: aripiprazole, hyperprolactinemia, prolactin abnormal, randomized controlled trial, controlled clinical trial, randomized, placebo, drug therapy, randomly, and trial. The keywords were used in combination with the Boolean operators AND, OR, and NOT. We used the “related article” function to supplement the search. We manually searched bibliographies of randomized controlled trials for studies that were missed in the electronic search. Two authors assessed studies that met the inclusion criteria for the meta-analysis independently; the third author would be involved in discussion if any disagreements existed.
We used the method of random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias to assess the methodological quality of included individual randomized controlled trials. Furthermore, we used the grading of recommendations assessment, development, and evaluation (GRADE) system to rate the quality of evidence and strength of recommendations of this meta-analysis. This was recommended by the Cochrane Collaboration. GRADE included systematic assessments of all included trials across six main domains for each outcome: limitations of the study design and execution, inconsistency, indirectness, and imprecision of results, publication bias, and large treatment effect. Accordingly, we graded the recommendation for outcome measure of either adjunctive aripiprazole or placebo as very low, low, moderate, or high.
We performed the meta-analysis according to the recommendations of the Cochrane Collaboration, using the Review Manager Version 5.1.7.0 software. Two authors independently extracted data for analysis. We presented the summary statistic of dichotomous outcomes as risk difference for adverse events, discontinuation, and treatment efficacy. The Mantel-Haenszel and Inverse-Variance methods were used to combine the summary statistic. We used I2 method to assess the statistical heterogeneity. When combining studies for meta-analysis, we assumed that variation existed between trials, so we used a random effect model to provide a conservative estimate of the results if I2>25%
We conducted separate subgroup analysis for adverse events, discontinuation and prolactin level normalization of Chinese studies in this meta-analysis. The dose of aripiprazole was all 5 mg/day in Chinese studies, and the definition of hyperprolactinemia was prolactin level >60 µg/L in all of these studies. We did this subgroup analysis to examine if variations in populations, antipsychotics dosage and definition of hyperprolactinemia across studies may have been responsible for the heterogeneity of the primary outcome. We assessed whether these differences influenced the outcome.
We initially considered seven trials and included five in our analysis (
Xu 2006 |
Shim 2007 |
Ji 2008 |
Chen 2009 |
Kane 2009 |
Randomization methods were reported as computer generated in one study, and random number table generated in another studies. The randomization method was high risk in another study because baseline adaptive random grouping was reported, which was partial randomization. Concealment of allocation was all unclear risk in five studies because it was not reported. Blinding methods were unclear in two studies, the rest of the studies were low risk in blinding. All studies described dropouts and withdrawals (
Study | Trail Design | Participants | Weeks | N | Antipsychotic | Aripiprazole dose(mg/day) | Rate of prolactin level normalization | Prolactin-related symptom recovery |
Xu 2006 | Placebo-controlled,single-blind | Female | 6 | 60 | Risperidone Sulpiride | 5 | Aripiprazole: 83.3%Placebo: 0% | 27/28 patients regained menstruation; 16/16 no longer complained galactorrhea; No change in placebo group. |
Shim 2007 | Placebo-controlled,double-blind | Either gender | 8 | 54 | Haloperidol | 15–30 | Aripiprazole: 88.5%Placebo: 3.6% | 7/11 patients regained menstruation; 1/2 no longer complained galactorrhea; No change in placebo group. |
Ji 2008 | Placebo-controlled,single-blind | Female | 6 | 130 | Risperidone | 5 | Aripiprazole: 82.0%Placebo: 4.0% | _ |
Chen 2009 | Placebo-controlled,double-blind | Male | 8 | 72 | Risperidone | 5 | Aripiprazole: 67.6%Placebo: 6.5% | _ |
Kane 2009 | Placebo-controlled,double-blind | Either gender | 16 | 322 | Risperidone Quetiapine | 5–15 | Aripiprazole: 19.5%Placebo: 9.1% | _ |
Study | Averse events(incidence rate: % or the number of events: n) | Discontinuation(incidence rate: % or the number of events: n) |
Xu 2006 | _ | |
Shim 2007 | Two patients in the aripriprazole group experienced insomnia, anxiety, and irritability, and both opted for withdrawal from the study. | |
Ji 2008 | Adjunctive aripiprazole vs. placebo: Subject was lost to follow-up (2 vs. 4); Switching to other antipsychotics(3 vs. 4); | |
Chen 2009 | Adjunctive aripiprazole vs. placebo: Adverse event (1 vs. 1); Subject was lost to follow-up (0 vs. 1); Switching to other antipsychotics(2 vs. 1); | |
Kane 2009 | Adjunctive aripiprazole vs. placebo: Adverse event (5.4% vs. 10.3%); Subject withdrew consent (8.9% vs. 5.8%); Subject was lost to follow-up (7.1% vs. 7.7%); Poor/noncompliance (7.1% vs. 3.2%); Other reasons (2.4% vs. 3.9%); Lack of efficacy (0.6% vs. 0%). |
Critical outcome | Participants(studies) | Risk ofbias |
Inconsistency |
Largeeffect$ | imprecision | Public bias | Overall quality of evidence# |
566(4) | Serious a, b | Serious c | No | No | undetected | +/+/−/−/; low | |
Studies with aripiprazole 5 mg/day | 190(2) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate |
566(4) | Serious a, b | Serious c | No | No | undetected | +/+/−/−/; low | |
Studies with aripiprazole 5 mg/day | 190(2) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate |
638(5) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate | |
Studies with aripiprazole 5 mg/day | 262(3) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate |
448(3) | No | Serious c | No | No | undetected | +/+/+/−/; moderate | |
394(2) | No | No | No | No | undetected | +/+/+/+/; high | |
126(2) | No | No | No | No | undetected | +/+/+/+/; high | |
376(2) | No | No | No | No | undetected | +/+/+/+/; high | |
316(4) | Serious a, b | No | Very large d | No | undetected | +/+/+/+/; high | |
Studies with aripiprazole 5 mg/day | 262(3) | Serious a, b | Serious c | Very large d | No | undetected | +/+/+/+/; high |
638(5) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate | |
Studies with aripiprazole 5 mg/day | 262(3) | Serious a, b | No | No | No | undetected | +/+/+/−/; moderate |
Meta-analysis of insomnia, headache, sedation, and psychiatric disorder (
A secondary analysis of insomnia, headache, and sedation including the data from the studies with adjunctive aripiprazole dose 5 mg/day (
Meta-analysis of extrapyramidal symptoms, dry mouth, and fatigue (
Meta-analysis of the prolactin level normalization (
A secondary analysis of the prolactin level normalization including the data from the studies with adjunctive aripiprazole dose 5 mg/day (
Two studies reported the change of prolactin-related symptoms (
Meta-analysis of discontinuation (
A secondary analysis of discontinuation including the data from the studies with adjunctive aripiprazole dose 5 mg/day (
In all of the five studies, there were differences in the mean change from baseline to endpoint in either PANSS or BPRS total score in the adjunctive aripiprazole treatment group compared with the placebo group. (
This first meta-analysis of five randomized controlled trials including 639 patients compared the safety and efficacy of adjunctive aripiprazole treatment with that of adjunctive placebo in antipsychotic-induced hyperprolactinemia. It showed that adjunctive aripiprazole is generally safe and well tolerated, with no significant increase in the risk of adverse events and discontinuation compared with placebo. No significant differences in insomnia, headache, sedation, psychiatric disorder, extrapyramidal symptoms, dry mouth, and fatigue were found. Furthermore, adjunctive aripiprazole was superior to placebo in prolactin level normalization. However, augmentation with aripiprazole failed to demonstrate significant improvement in psychiatric symptoms compared with placebo.
To the best of our knowledge, this is the first study to review the safety of adjunctive aripiprazole treatment. Adverse events occurred at similar incidence rates among patients receiving adjunctive aripiprazole and placebo, except for sedation, insomnia, and headache, which were more frequent in the adjunctive aripiprazole (dose >15 mg/day)group than in the placebo groups. These side effects are known to occur in a small proportion of patients given aripiprazole
Although most case reports, open label studies, and placebo-controlled trials in adult patients have shown that adjunctive aripiprazole is associated with lower prolactin levels, unchanged or even elevated prolactin levels have also been reported by some studies
We also searched Chinese database in this meta-analysis, which included randomized controlled trials comparing adjunctive aripiprazole with placebo for antipsychotic-induced hyperprolactinemia in China. Thus, it is the first study to include all trials available without applying any language restrictions. Moreover, we conducted subgroup analysis of Chinese studies in this meta-analysis. This subgroup analysis confirmed the significant superiority for adjunctive aripiprazole 5 mg/day versus placebo regarding the prolactin level normalization; furthermore we also confirmed that there was no significantly difference in insomnia, headache, sedation, and discontinuation between adjunctive aripiprazole 5 mg/day and placebo. Thus, the finding suggest that 5 mg/day dose of adjunctive aripiprazole may be appropriate, in fact adverse events were more frequent when adjunctive aripiprazole dose >15 mg/day in previous discussion.
There was significant heterogeneity of the results in adverse events and prolactin level normalization in this meta-analysis, suggesting the effect of relevant moderator and mediator variables. This may be because meta- analyses combine results from trials that differ in their methodology, study size and year, patient and treatment selection, outcome variables, and study conduct. However, when combining studies of this nature for meta-analysis, we used a random effect model to provide a conservative estimate of prolactin level normalization and adverse events. We used a fixed effect model to analyze discontinuation because no heterogeneity existed. On the other hand, we conducted subgroup analysis of Chinese studies in this meta-analysis that sought to disentangle relevant moderator variables, strengthening the primary finding. In addition, in the meta-analysis of the prolactin level normalization, the study by Kane et al. (2009) was excluded because the presence of hyperprolactinemia was not an eligibility criterion.
This meta-analysis only included 639 patients in five randomized controlled trials. The sample size was relatively small for meta-analysis, which prevented further data exploration. However, we included studies with well defined randomized controlled trials comparing adjunctive aripiprazole with placebo. Although there were about 10 randomized controlled trials published in Chinese language, only 3 trials were included in this meta-analysis after a more strict quality assessment. In addition, random sequence generation was high risk in one study, potentially contributing to selection bias; however, this study was low risk in other bias. Although the GRADE approach showed the quality of evidence were “low” in insomnia and headaches, other outcomes were from “moderate” to “high”, especially the quality of evidence in prolactin level normalization were all “high”. This high quality of evidence will increase our confidence to use adjunctive aripiprazole to manage antipsychotic-induced hyperprlactinemia.
The correlation between aripiprazole dose and the change in prolactin levels may not have been fully analyzed because the aripiprazole doses were all higher than 5 mg/day in this meta-analysis. In contrast, in a previous study, 2 mg/day of aripiprazole acted as a partial dopamine agonist producing a clinical effect
A review by Hoffer et al. (2009) included case studies, reports, and placebo-controlled trials studies to study the effectiveness of aripiprazole in lower prolactin level
Marder et al. (2003) conducted a pooled analysis of safety and tolerability data from five 4- to 6-week double-blind multi-center studies of patients with schizophrenia or schizoaffective randomized to aripiprazole (n = 932) or placebo (n = 416)
Similarly, Deleon and Citrome had reviewed the efficacy and tolerability of aripiprazole in schizophrenia and bipolar disorder respectively
Antipsychotic-induced hyperprolactinemia has both short and long-term consequences including menstrual irregularities, amenorrhea, galactorrhea, gynecomastia, sexual dysfunction, infertility, osteoporosis, and even breast cancer
This meta-analysis of five randomized controlled trials comparing adjunctive aripiprazole and placebo have shown that adjunctive aripiprazole can be used safely and effectively for antipsychotic-induced hyperprolactinemia. Adjunctive aripiprazole is not associated with increased adverse events and discontinuations compared with placebo. Adjunctive aripiprazole is superior to placebo in prolactin level normalization. The recommended dose of adjunctive aripiprazole may be 5 mg/day. An early adjunctive aripiprazole to manage hyperprolactinemia may increase treatment compliance, which may offer benefits in schizophrenia treatment. The efficacy of adjunctive dose <5 mg/day needs further evaluation.
(EPS)
We wish to thank Prof. Phil Wiffen (Training Director of UK Cochrane Center) and Dr. Jun Xia (Cochrane Schizophrenia Group) for their system review courses. We wish to thank Dr. Zhirui Zhou (Radiotherapy Department, The Tumor Hospital of Jilin Province) for his comments on the manuscript.