Multidimensional Analysis Predicts Clinical Severity in CDKL5 Deficiency Disorder: Readiness for Clinical Trials
Abstract number :
1.485
Submission category :
16. Epidemiology
Year :
2023
Submission ID :
1287
Source :
www.aesnet.org
Presentation date :
12/2/2023 12:00:00 AM
Published date :
Authors :
Presenting Author: Jenny Downs, BApplSci (physio) GradCertPubHlth MSc PhD – Telethon Kids Institute
Peter Jacoby, MSc – Biostatistician, Child Disability, Telethon Kids Institute; Eric Marsh, MD – Children’s Hospital of Philadelphia; Jacinta Saldaris, Phd – Telethon Kids Institute; Helen Leonard, MBChB – Telethon Kids Institute; Bernhard Suter, MD – Texas Children’s Hospital; Elia Pestana Knight, MD – Cleveland Clinic; Heather Olson, MD – Boston Children’s Hospital; Dana Price, MD – NYU Langone Health; Judith Weisenberg, MD – Washington University of St. Louis; Raj Rajaraman, MD – UCLA Mattel Children’s Hospital; Scott Demarest, MD – Children’s Hospital Colorado; Tim Benke, MD – University of Colorado School of Medicine
Rationale: CDKL5 deficiency disorder (CDD) is a developmental epileptic encephalopathy characterized by early onset seizures and significant global developmental delays. Upcoming trials of precision therapies for CDD require reliable, valid, and sensitive clinical outcome assessments (COAs) to demonstrate impacts across the affected domains in CDD. Our team has undertaken a comprehensive program of research investigating the psychometric properties of multiple COAs for CDD. This study investigated how our suite of measures of functional ability and comorbidities combine to reflect overall severity in CDD.
Methods: We recruited families with a child with CDD enrolled in the International CDKL5 Clinical Research Network (ICCRN), comprising eight Centers of Excellence (CoE) and the International CDKL5 Disorder Database (ICDD). Data on multiple COAs for functional abilities and comorbidities were collected. A structural equation model (SEM) was constructed with measured domains within the COAs grouped a priori to form three latent endogenous variables: (1) remotely collected video measures of gross motor and hand function and clinician-observed movement disorder; (2) parent-reported severity of seizures, alertness, insomnia, and excessive daytime sleepiness; and (3) clinician and parent reported communication. Domains measuring vision, feeding and behavior were included separately as observed endogenous variables. There was a latent exogenous variable representing overall severity. Full Information Maximum Likelihood (FIML) was used to fit the model and a factor score for overall severity was calculated for each participant. Overall severity scores were correlated with Alpha/Delta power ratios for those participants with EEG data. The contribution of each COA to overall severity was evaluated using linear regression.
Results: Data from 235 individuals enrolled (ICCRN: 150 from the CoE, 85 from the ICCD) was used for SEM. Participant ages ranged from three months to 41 years with a median age of 7.1 years. A total of 192 (81.7%) were female. For this severe disorder, there was an approximately normal distribution of overall clinical severity scores. The model fit was good (Figure 1). The Pearson correlation coefficient between overall severity and the EEG Alpha/Delta power ratio was 0.60. Clinician-observed data explained 98.6% of variance of severity and video data explained 90.6% of variance. Less variance was explained by the caregiver-reported domains (74.6%), communication (66.9%), and the sleep questionnaire (12.5%).
Conclusions: The analysis provides evidence that our COAs form a coherent set of measures which capture clinical severity in CDD and which, taken together, correlate with an EEG biomarker. This study suggests that the increments and range in our CDD measures are appropriate to CDD and could indicate whether new precision therapies are successful.
Funding:
NIH/NINDS U01NS114312 (Benke/PD)
Epidemiology