Faculty Bibliography

We carried out the first systematic review to gauge if assessment of blinding was conducted in RCTs investigating medications for ADHD. Put of 161 RCTs form the dataset MED-ADHD (https://med-adhd.org/), we found only one RCT that reported blinding integrity, indicating that assessments of blinding integrity are very rarely conducted in the field. While our findings are not meant to invalidate the evidence on the benefits of ADHD medications, they suggest an opportunity to improve reporting of clinical trials in ADHD.

BACKGROUND:Individuals with attention deficit hyperactivity disorder (ADHD) are at a higher risk of depression and lower quality of life (QoL); however, it is unclear whether disrupted sleep and circadian rhythms mediate this increased risk. OBJECTIVES/OBJECTIVE:We investigated whether disruption of self-reported sleep and circadian factors mediate the associations of ADHD traits with depression symptom severity and QoL. METHODS:1364 participants (mean: 51.86 (SD=0.37) years, 75% women) from a large-scale cross-sectional online survey (Netherlands Sleep Registry) completed a sociodemographic questionnaire, the Adult ADHD Rating Scale, Hospital Anxiety and Depression Scale, Satisfaction With Life Scale (SLS) and Cantril Ladder (CL) (QoL measures), Insomnia Severity Index, Pittsburgh Sleep Quality Index and Munich Chronotype Questionnaire. FINDINGS/RESULTS:Higher ADHD traits were significantly associated with depression symptom severity (p=0.03), lower QoL (p<0.001), insomnia severity (p<0.001), lower sleep quality (p<0.001) and later chronotype (p=0.01). No sleep or circadian factor significantly mediated the association of the severity of symptoms of ADHD and depression (all p>0.1). Conversely, only insomnia severity significantly mediated the association of ADHD traits and QoL (SLS: standardised β=-0.10, 95% CI (-0.12 to -0.04); CL: standardised β=0.103, 95% CI (0.04 to 0.16)). CONCLUSION/CONCLUSIONS:ADHD traits were associated with lower QoL and it was partially mediated by insomnia severity. Future studies targeting insomnia complaints in this population may help mitigate their depression complaints and improve their QoL. CLINICAL IMPLICATIONS/CONCLUSIONS:Our results may help current clinical guidelines that do not typically link sleep/circadian complaints to QoL in ADHD assessment.

BACKGROUND:Attention-deficit/hyperactivity disorder (ADHD) symptomatology in childhood is associated with a high risk of suicide attempt later in life. However, symptom presentation in ADHD is heterogeneous, and little is known about how suicide risk varies according to different profiles of ADHD symptoms and sex. OBJECTIVE:The aim was to investigate the longitudinal associations between childhood profiles of ADHD symptoms (ie, hyperactivity-impulsivity and inattention) and youth suicide attempt in males and females, separately. METHODS:This population-based cohort study used data from three longitudinal cohorts: the Quebec Longitudinal Study of Child Development (QLSCD), the Quebec Longitudinal Study of Kindergarten Children (QLSKC) and the Quebec Newborn Twin Study (QNTS) for a total of 4399 participants (1490 from the QLSCD, 2134 from the QLSKC and 775 from the QNTS; 50% females) followed up from ages 6-23 years. Symptoms of hyperactivity-impulsivity and inattention were assessed by teachers five times from ages 6-12 years. Suicide attempt in adolescence and young adulthood (by age 23) was self-reported. Multitrajectory modelling was used to identify profiles of ADHD symptoms, and regression analysis was used to test their association with suicide attempt, adjusting for childhood socioeconomic and clinical characteristics. FINDINGS/RESULTS:We identified four ADHD symptom profiles with distinct associations with suicide attempt for males and females. Compared with those with persistently low symptoms, females with persistently high inattention and hyperactivity-impulsivity (OR: 2.54, CI 1.39 to 4.63) or high inattention and low hyperactivity-impulsivity (OR: 1.81, CI 1.21 to 2.70) were at higher risk of suicide attempt, while, among males, only those with decreasing hyperactivity-impulsivity and inattention over time (OR: 2.23, CI 1.20 to 4.13) were at higher risk of suicide attempt. CONCLUSIONS:Risk of suicide attempt in children with ADHD symptoms varies according to both symptom profile and sex, the highest risk being for females with high inattention symptoms (with or without hyperactivity), and males with decreasing symptoms. CLINICAL IMPLICATIONS/CONCLUSIONS:Taking into account differences in both sex and ADHD symptoms profile may be relevant to more accurately identify and manage suicide risk in individuals with high ADHD symptoms, though caution is needed when generalising our population-based findings to clinical populations.

IMPORTANCE/UNASSIGNED:Representation of race and ethnicity in randomized clinical trials (RCTs) is critical for understanding treatment efficacy across populations with different racial and ethnic backgrounds. OBJECTIVE/UNASSIGNED:To examine race and ethnicity representation and reporting across RCTs of pharmacotherapies for mental disorders. DATA SOURCES/UNASSIGNED:PubMed (Medline), Embase (Ovid), APA PsycInfo, and Web of Science were searched until March 1, 2024, to retrieve network meta-analyses including RCTs of pharmacotherapies for International Statistical Classification of Diseases and Related Health Problems, Tenth Revision mental disorders. STUDY SELECTION/UNASSIGNED:RCTs that recruited people of any age with a diagnosis of a mental disorder and that tested the efficacy of any pharmacologic intervention vs any control arm. DATA EXTRACTION AND SYNTHESIS/UNASSIGNED:Random-effects logit-transformed proportion meta-analyses were used to estimate prevalence rates of race and ethnicity groups and their temporal trends across RCTs and to compare US RCT prevalence rates with US Census data. The Preferred Reporting Items for Overviews of Reviews was used to report our review. MAIN OUTCOMES AND MEASURES/UNASSIGNED:Reporting of data and percentages of race and ethnicity. The year of publication, type of RCT, geographic location, age group, and sample size were also included. There were no deviations that occurred from the original protocol. RESULTS/UNASSIGNED:Data were obtained from 1683 RCTs (375 120 participants in total). Of these, 1363 (91.7% of participants) included participants aged 18 years or older; 680 RCTs (36.0% of participants) were from the US, 404 (17.1% of participants) were from Europe, and 293 (29.9% of participants) were from multiple geographic locations. Race and ethnicity were reported in 39.2% of RCTs; reporting was the highest in US-based RCTs (58.7%) and lowest in Central and South America (8.7%) and Asia and the Middle East (12.4%). Among participants, 2.7% (95% CI, 2.1%-3.5%) self-reported as Asian, 9.0% (95% CI, 8.1%-10.0%) as Black, 11.0% (95% CI, 9.1%-13.3%) as Hispanic among White, 80.2% (95% CI, 78.8%-81.5%) as White including Hispanic, and 5.8% (95% CI, 5.2%-6.4%) as other race or ethnicity, multiracial, or multiethnic. There was more frequent reporting of race and ethnicity in US RCTs (log odds increased by 0.066 each year) and less frequent reporting in non-US RCTs (log odds increased by 0.023 each year). Studies reporting race and ethnicity did not generally include larger sample sizes (mean sample size, 263.7 [95% CI, 15.0-860.3] participants) compared with those not reporting such data (mean sample size, 196.6 [95% CI, 12.0-601.3] participants), albeit not in all locations. In US RCTs, adults in the other or multiracial and multiethnic category were historically overrepresented, while adults in Asian, Black, Hispanic among White, and White including Hispanic categories were underrepresented; Asian, Black, and Hispanic among White children and adolescents are still currently underrepresented. CONCLUSIONS AND RELEVANCE/UNASSIGNED:The findings of this meta-analysis suggest that differences in reporting race and ethnicity across geographic locations and underrepresentation of certain racial and ethnic groups in US-based RCTs highlight the need for international guidelines to ensure equitable recruitment and reporting in clinical trials.

This umbrella review synthesizes data on the prevalence of mental disorder symptoms among university students worldwide. A systematic search of seven databases (inception-July 23, 2023) followed PRISMA guidelines. We included meta-analyses assessing the prevalence of mental disorder symptoms, evaluating methodological quality with AMSTAR-2. A random-effects meta-analysis was conducted, along with meta-regression and subgroup analyses for moderators (percentage of females, publication date, healthcare-related degrees, COVID-19 pandemic). We included 1,655 primary studies from 62 meta-analyses, encompassing 8,706,185 participants. AMSTAR-2 ratings classified 35 % of meta-analyses as low quality and 65 % as critically low. Pooled prevalence estimates were: depression-mild (35.41 %, CI=33.9-36.93) and severe (13.42 %, CI=8.03-19.92; k=952; n=2,108,813); anxiety-mild (40.21 %, CI=37.39-43.07) and severe (16.79 %, CI=7.21-29.29; k=433; n=1,579,780); sleep disorders (41.09 %, CI=35.7-46.58); eating disorders (17.94 %, CI=15.79-20.20); gambling disorder (6.59 %, CI=5.52-7.75); post-traumatic stress disorder (25.13 %, CI=20.55-30.02); stress (36.34 %, CI=29.36-43.62); and suicide-related outcomes (ideation past 12 months: 10.76 %, CI=9.53-12.06; lifetime ideation: 20.33 %, CI=16.15-24.86; suicide attempt past 12 months: 1.37 %, CI=0.67-2.29; lifetime attempt: 3.44 %, CI=2.48-4.54). Meta-regression analyses identified statistically significant moderators of prevalence such as healthcare academic degrees and the pandemic in the case of depression and studies with more females in the case of sleep disorders. This is the most comprehensive synthesis on the prevalence of mental disorder symptoms in university students, providing crucial insights for clinicians, policymakers, and stakeholders.

BACKGROUND:) that underlies the vulnerability to both physical and mental conditions could have important implications for our approach to health assessment and treatment. OBJECTIVE:in children and adolescents. METHODS:This Swedish registry-based cross-sectional study included children and adolescents born between 1996 and 2003 with follow-up until 2013. We extracted data on 25 mental and physical health conditions according to the ICD-10 system. To determine the optimal dimensional structure of these conditions, several competing measurement models were tested, including correlated factors, one factor, various bifactor specifications and bifactor exploratory structural equation modelling (ESEM). FINDINGS/RESULTS:=0.423; ECV=0.130) factors also indicated additional significant unique contributions. CONCLUSIONS:underlying both mental and physical conditions, alongside distinct domain-specific factors. These findings have important implications for clinical practice, providing evidence that suggests the need for more integrated approaches to health assessment and treatment that consider the interconnectedness of mental and physical health.

IMPORTANCE/UNASSIGNED:Expectancy effects are significant confounding factors in psychiatric randomized clinical trials (RCTs), potentially affecting the interpretation of study results. This narrative review is the first, to our knowledge, to explore the relationship between expectancy effects, compromised blinding integrity, and the effects of active treatment/placebo in psychiatric RCTs. Additionally, we present statistical and experimental approaches that may help mitigate the confounding impact of expectancy effects. The review concludes with recommendations to enhance the reliability of RCTs in psychiatry. OBSERVATIONS/UNASSIGNED:The placebo response comprises both specific and nonspecific elements, with expectation being a key specific component. Evidence from experimental and clinical studies suggests that expectancy can influence treatment responses in RCTs. Blinding integrity may be compromised by perceived treatment efficacy and adverse effects, introducing bias into outcome assessments. Treatment expectations can lead to unblinding during RCTs, and meta-analytic data from studies in the fields of psychedelics and anxiety disorders indicate that this can influence effect sizes. Therefore, controlling for expectancy effects is essential when interpreting RCT results. Novel statistical methods, though still in need of further validation, offer strategies to address this issue. Another approach may involve experimental medicine models, which aim to develop objective improvement markers (readouts) less affected by expectancy effects. CONCLUSIONS AND RELEVANCE/UNASSIGNED:Expectancy effects represent a significant confound in psychiatric RCTs. We recommend collecting data on treatment expectations alongside monitoring blinding integrity to more accurately interpret study results. Additionally, developing objective readouts that are less confounded by expectancy effects offers another promising avenue for mitigating these confounding influences in psychiatric RCTs.

BACKGROUND:Concerns about the cardiovascular safety of medications used for the treatment of attention-deficit hyperactivity disorder (ADHD) remain. We aimed to compare the effects of pharmacological treatments for ADHD on haemodynamic values and electrocardiogram (ECG) parameters in children, adolescents, and adults. METHODS:For this systematic review and network meta-analysis, we searched 12 electronic databases, including Cochrane CENTRAL, Embase, PubMed, and the WHO International Clinical Trials Registry Platform, from database inception to Jan 18, 2024, for published and unpublished randomised controlled trials comparing amphetamines, atomoxetine, bupropion, clonidine, guanfacine, lisdexamfetamine, methylphenidate, modafinil, or viloxazine against each other or placebo. Primary outcomes were change in systolic blood pressure (SBP) and diastolic blood pressure (DBP), measured in mm Hg, and pulse, measured in beats per minute, at timepoints closest to 12 weeks, 26 weeks, and 52 weeks. Summary data were extracted and pooled in random-effects network meta-analyses. Certainty of evidence was assessed with the Confidence in Network Meta-Analysis (CINeMA) framework. This study was registered with PROSPERO, CRD42021295352. Before study initiation, we contacted representatives of a UK-based charity of people with lived experience of ADHD-the ADHD Foundation-regarding the relevance of the topic and the appropriateness of the outcomes chosen. FINDINGS/RESULTS:102 randomised controlled trials with short-term follow-up (median 7 weeks [IQR 5-9]) were included, encompassing 13 315 children and adolescents (aged ≥5 years and <18 years; mean age 11 years [SD 3]; of available data, 9635 [73%] were male and 3646 [27%] were female; of available data, 289 [2%] were Asian, 1719 [15%] were Black, and 8303 [71%] were White) and 9387 adults (≥18 years, mean age 35 years [11]; of available data, 5064 [57%] were male and 3809 [43%] were female; of available data, 488 [6%] were Asian, 457 [6%] were Black, and 6372 [79%] were White). Amphetamines, atomoxetine, lisdexamfetamine, methylphenidate, and viloxazine led to increments in haemodynamic values in children and adolescents, adults, or both. In children and adolescents, mean increase against placebo ranged from 1·07 (95% CI 0·36-1·79; moderate CINeMA confidence) with atomoxetine to 1·81 (1·05-2·57; moderate) with methylphenidate for SBP; from 1·93 (0·74-3·11; high) with amphetamines to 2·42 (1·69-3·15; low) with methylphenidate for DBP; and from 2·79 (1·05-4·53; moderate) with viloxazine to 5·58 (4·67-6·49; high) with atomoxetine for pulse. In adults, mean increase against placebo ranged from 1·66 (95% CI 0·38-2·93; very low) with methylphenidate to 2·3 (0·66-3·94; very low) with amphetamines for SBP; from 1·60 (0·29-2·91; very low) with methylphenidate to 3·07 (0·69-5·45; very low) with lisdexamfetamine for DBP; and from 4·37 (3·16-5·59; very low) with methylphenidate to 5·8 (2·3-9·3; very low) with viloxazine for pulse. Amphetamines, lisdexamfetamine, or methylphenidate were not associated with larger increments in haemodynamic values compared with atomoxetine or viloxazine in either children and adolescents or adults. Guanfacine was associated with decrements in haemodynamic values in children and adolescents (mean decrease against placebo of -2·83 [95% CI -3·8 to -1·85; low CINeMA confidence] in SBP, -2·08 [-3 to -1·17; low] in DBP, and -4·06 [-5·45 -2·68; moderate] in pulse) and adults (mean decrease against placebo of -10·1 [-13·76 to -6·44; very low] in SBP, -7·73 [-11·88 to -3·58; very low] in DBP, and -6·83 [-10·85 to -2·81; very low] in pulse). Only four RCTs informed on effects in the medium term and none on the long term. INTERPRETATION/CONCLUSIONS:Practitioners should monitor blood pressure and pulse in patients with ADHD treated with any pharmacological intervention, and not stimulants only. Given the short duration of available randomised controlled trials, new research providing insights on the causal effects of ADHD medications on cardiovascular parameters in the longer term should be funded. FUNDING/BACKGROUND:National Institute for Health and Care Research.