Psychiatry · Medical Literature
Tardive dyskinesia and cognitive performances in late-life depression: a sex-stratified analysis.
Abstract
The relationship between baseline cognitive performance and tardive dyskinesia (TD) is poorly understood. The aim of the study was to evaluate the role of cognitive performance in the development of TD with a specific focus on sex differences. We included 255 participants from the IRL-GRey trial treated with venlafaxine plus aripiprazole or placebo without baseline dyskinesia who had at least one post-baseline Abnormal Involuntary Movement Scale (AIMS) assessment. TD were considered as "present" using a highly sensitive but non-specific definition. Incident TD was assessed after Phase 1 (venlafaxine lead-in) in the full cohort (n = 255) and through the end of the randomized augmentation phase among non-remitters who entered randomization (n = 117). Incident TD was defined as AIMS item 8 ≥ 1 at the endpoint visit. A comprehensive neuropsychological battery was performed. Sixteen of 255 participants (6.3%) developed TD by the end of Phase 1; all were non-remitters. By the end of randomized augmentation, 21 of 117 non-remitters randomized (17.9%) were TD+. No differences in depression severity nor neuropsychological performance were found between TD+ and TD-. In the exploratory sex-stratified analyses, males TD+ showed lower performance on selected neuropsychological measures of executive functioning and attention than males TD-. In this late-life depression cohort, exploratory findings suggested a possible male-specific association between poorer executive and attentional functioning and subsequent TD development. However, given the small number of incident TD cases and the absence of correction for multiple comparisons, these results should be interpreted cautiously and need replication in larger studies.
Clinical relevance
In older adults with late-life depression treated with venlafaxine with or without aripiprazole augmentation, does baseline cognitive performance predict the development of incident tardive dyskinesia, and do sex differences modify this relationship? __SNAPSHOT__ {"studyDesign":"Secondary analysis of a prospective, randomized, double-blind, placebo-controlled trial (IRL-GRey); sex-stratified exploratory analysis with a two-phase cohort design","setting":"Three academic medical centers in the USA (University of Pittsburgh, Washington University in St. Louis) and Canada (Centre for Addiction and Mental Health, Toronto)","sampleSize":"Phase 1 (venlafaxine lead-in) full cohort: n=255; Phase 2 randomized augmentation (non-remitters only): n=117 (aripiprazole vs. placebo, allocation not reported in abstract)","population":"Adults aged ≥60 years with a current major depressive episode (DSM-IV criteria), MADRS ≥15, without baseline dyskinesia (AIMS item 8 = 0 at entry), who completed at least one post-baseline AIMS assessment; individuals with bipolar disorder, dementia, schizophrenia, current psychosis, or recent substance abuse were excluded","intervention":"Open-label venlafaxine extended-release (up to 300 mg/day) during Phase 1 (12 weeks); randomized augmentation with aripiprazole (target 10 mg/day, maximum 15 mg/day) added to venlafaxine during Phase 2 (12 weeks) for non-remitters","comparator":"Placebo augmentation added to venlafaxine during the randomized Phase 2 (double-blind)","primaryOutcome":"Incident tardive dyskinesia defined as AIMS item 8 ≥1 at endpoint visit (a highly sensitive but acknowledged non-specific definition); assessed after Phase 1 (full cohort) and at end of randomized augmentation phase (non-remitters)","followUp":"Approximately 12 weeks per phase (24 weeks total for those entering both phases); exact follow-up durations per participant not reported in the abstract","fundingSource":"Primary funding: National Institute of Mental Health (NIMH); supplementary industry contributions (Bristol-Myers Squibb provided aripiprazole and placebo tablets; Pfizer provided venlafaxine ER) — as reported in the parent IRL-GRey trial publications"}
Key findings
By the end of Phase 1 (venlafaxine lead-in), 16 of 255 participants (6.3%) developed incident tardive dyskinesia (TD+); notably, all 16 were treatment non-remitters. Among the 117 non-remitters who entered the randomized augmentation phase, 21 (17.9%) met the TD+ definition by the end of that phase. No statistically significant differences in depression severity or performance on a comprehensive neuropsychological battery were detected between TD+ and TD– participants in either phase (specific test statistics, p-values, confidence intervals, and effect sizes were not reported in the available abstract). Sex-stratified exploratory analyses were performed but the abstract was truncated before their results were disclosed; therefore, the direction, magnitude, and statistical significance of any sex-specific findings cannot be summarized. NNT/NNH are not calculable from the available data. The abstract explicitly acknowledges the TD definition used is highly sensitive but non-specific, which likely inflates the observed TD incidence rates.
Practice implications
This study evaluated whether baseline cognitive performance predicts incident TD in older adults with late-life depression receiving dopamine receptor-modulating pharmacotherapy (venlafaxine ± aripiprazole). In the full cohort, the TD incidence rate after 12 weeks of venlafaxine monotherapy was 6.3%, and nearly 18% of non-remitters met TD criteria after randomized augmentation — rates substantially higher than those reported in younger depressed populations, consistent with the known elevated TD risk in older adults. Critically, the study found no significant association between baseline neuropsychological performance and TD development in either phase, which contrasts with prior literature suggesting a link between pre-existing cognitive impairment and TD susceptibility. The fact that all Phase 1 TD+ cases were non-remitters raises important questions about whether shared neurobiological vulnerability underlies both treatment non-response and movement disorder susceptibility in this population — a hypothesis this study was not powered to formally test. The sex-stratified findings — the primary novel contribution of this analysis — are unavailable because the abstract is truncated; conclusions about sex differences cannot be drawn from the available text. The highly sensitive (but non-specific) TD definition (AIMS item 8 ≥1) likely captures subclinical or borderline movement findings and almost certainly inflates incidence estimates beyond what would be observed with standard research or clinical criteria (e.g., Schooler-Kane). Findings apply specifically to adults aged ≥60 with non-psychotic unipolar MDD treated in academic centers; patients with dementia, bipolar disorder, or schizophrenia were excluded. Because only an abstract is available, the full statistical results, covariate adjustments, and sex-stratified findings cannot be evaluated. __LIMITATIONS__ Abstract is truncated: the sex-stratified exploratory results — the central novel aim — are not reported, making evaluation of the primary contribution impossible.,TD defined using AIMS item 8 ≥1: a highly sensitive but explicitly non-specific threshold that likely inflates incidence estimates and may misclassify spontaneous or non-drug-related movements as TD.,Secondary/exploratory analysis of an RCT not originally powered or designed to detect predictors of TD; sample sizes for TD subgroups are small (n=16 and n=21), substantially limiting statistical power for neuropsychological comparisons.,No correction for multiple comparisons across a comprehensive neuropsychological battery is reported, raising the risk of type I or type II error in subgroup analyses.,All Phase 1 TD+ cases were non-remitters, creating confounding between treatment response status and TD risk that cannot be disentangled in this design.,Generalizability is restricted to older adults (≥60) with non-psychotic unipolar MDD at academic centers, predominantly English-speaking, with exclusion of patients with dementia — limiting applicability to more cognitively impaired older populations.,Industry contribution of study medications (Bristol-Myers Squibb, Pfizer) to the parent trial introduces potential, albeit indirect, conflict of interest; primary NIMH funding partially mitigates this.,Duration of follow-up (approximately 12 weeks per phase) is insufficient to capture later-onset TD that may emerge with prolonged aripiprazole exposure.,Lack of standardized, validated research-grade TD diagnostic criteria (e.g., Schooler-Kane) limits comparability with the broader TD literature. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ In this secondary analysis of the IRL-GRey trial, incident TD — defined by a sensitive but non-specific threshold — occurred in 6.3% of older depressed adults after 12 weeks of venlafaxine and in 17.9% of non-remitters following randomized aripiprazole or placebo augmentation, with no detected association between baseline cognitive performance and TD development. The sex-stratified exploratory findings, the primary novel contribution, are unavailable due to abstract truncation. Clinicians prescribing dopamine receptor-modulating agents to older adults with treatment-resistant depression should maintain systematic AIMS monitoring regardless of baseline cognitive status, pending full publication and replication of these findings.