METHODS: We performed whole genome sequencing and analyzed 6 199 696 common variants among 113 aromatic ASM-induced SJS/TEN cases and 84 tolerant controls of Han Chinese ethnicity.
RESULTS: In the primary analysis, nine variants reached genome-wide significance (p p = .001; homozygotes: relative risk = .23, p p < 5e-6) identified through the primary and subanalyses (stratified by HLA-B*15:02 status and drug exposure) suggests that genetic variation within regulatory DNA may contribute to risk indirectly by disrupting the regulation of pathology-related genes. The genes implicated were specific either to the primary analysis (CD9), HLA-B*15:02 carriers (DOCK10), noncarriers (ABCA1), carbamazepine exposure (HLA-E), or phenytoin exposure (CD24).
SIGNIFICANCE: We identified variants that could explain why some carriers of HLA-B*15:02 tolerate treatment, and why some noncarriers develop ASM-induced SJS/TEN. Additionally, this analysis suggests that the mixing of HLA-B*15:02 carrier status in previous studies might have masked variants contributing to susceptibility, and that inheritance of risk for ASM-induced SJS/TEN is complex, likely involving multiple risk variants.
OBJECTIVE: To develop and validate a deep learning model using readily available clinical information to predict treatment success with the first ASM for individual patients.
DESIGN, SETTING, AND PARTICIPANTS: This cohort study developed and validated a prognostic model. Patients were treated between 1982 and 2020. All patients were followed up for a minimum of 1 year or until failure of the first ASM. A total of 2404 adults with epilepsy newly treated at specialist clinics in Scotland, Malaysia, Australia, and China between 1982 and 2020 were considered for inclusion, of whom 606 (25.2%) were excluded from the final cohort because of missing information in 1 or more variables.
EXPOSURES: One of 7 antiseizure medications.
MAIN OUTCOMES AND MEASURES: With the use of the transformer model architecture on 16 clinical factors and ASM information, this cohort study first pooled all cohorts for model training and testing. The model was trained again using the largest cohort and externally validated on the other 4 cohorts. The area under the receiver operating characteristic curve (AUROC), weighted balanced accuracy, sensitivity, and specificity of the model were all assessed for predicting treatment success based on the optimal probability cutoff. Treatment success was defined as complete seizure freedom for the first year of treatment while taking the first ASM. Performance of the transformer model was compared with other machine learning models.
RESULTS: The final pooled cohort included 1798 adults (54.5% female; median age, 34 years [IQR, 24-50 years]). The transformer model that was trained using the pooled cohort had an AUROC of 0.65 (95% CI, 0.63-0.67) and a weighted balanced accuracy of 0.62 (95% CI, 0.60-0.64) on the test set. The model that was trained using the largest cohort only had AUROCs ranging from 0.52 to 0.60 and a weighted balanced accuracy ranging from 0.51 to 0.62 in the external validation cohorts. Number of pretreatment seizures, presence of psychiatric disorders, electroencephalography, and brain imaging findings were the most important clinical variables for predicted outcomes in both models. The transformer model that was developed using the pooled cohort outperformed 2 of the 5 other models tested in terms of AUROC.
CONCLUSIONS AND RELEVANCE: In this cohort study, a deep learning model showed the feasibility of personalized prediction of response to ASMs based on clinical information. With improvement of performance, such as by incorporating genetic and imaging data, this model may potentially assist clinicians in selecting the right drug at the first trial.
METHODS: Participants with rifampicin-susceptible pulmonary tuberculosis received a single intravenous infusion of pascolizumab or placebo; and standard 6-month tuberculosis treatment. Pascolizumab dose increased in successive cohorts: [1] non-randomised 0.05 mg/kg (n = 4); [2] non-randomised 0.5 mg/kg (n = 4); [3] randomised 2.5 mg/kg (n = 9) or placebo (n = 3); [4] randomised 10 mg/kg (n = 9) or placebo (n = 3). Co-primary safety outcome was study-drug-related grade 4 or serious adverse event (G4/SAE); in all cohorts (1-4). Co-primary efficacy outcome was week-8 sputum culture time-to-positivity (TTP); in randomised cohorts (3-4) combined.
RESULTS: Pascolizumab levels exceeded IL-4 50% neutralising dose for 8 weeks in 78-100% of participants in cohorts 3-4. There were no study-drug-related G4/SAEs. Median week-8 TTP was 42 days in pascolizumab and placebo groups (p = 0.185). Rate of TTP increase was greater with pascolizumab (difference from placebo 0.011 [95% Bayesian credible interval 0.006 to 0.015] log10TTP/day.
CONCLUSIONS: There was no evidence to suggest blocking IL-4 was unsafe. Preliminary efficacy findings are consistent with animal models. This supports further investigation of adjunctive anti-IL-4 interventions for tuberculosis in larger phase 2 trials.
METHODS: A case-control study was performed to detect HLA loci involved in aromatic antiepileptic drug-induced Stevens-Johnson syndrome in a southern Han Chinese population. Between January 1, 2006, and December 31, 2015, 91 cases of Stevens-Johnson syndrome induced by aromatic antiepileptic drugs and 322 matched drug-tolerant controls were enrolled from 8 centers. Important genotypes were replicated in cases with maculopapular eruption and in the meta-analyses of data from other populations. Sequence-based typing determined the HLA-A, HLA-B, HLA-C, and HLA-DRB1 genotypes.
RESULTS: HLA-B*15:02 was confirmed as strongly associated with carbamazepine-induced Stevens-Johnson syndrome (p = 5.63 × 10(-15)). In addition, HLA-A*24:02 was associated significantly with Stevens-Johnson syndrome induced by the aromatic antiepileptic drugs as a group (p = 1.02 × 10(-5)) and by individual drugs (carbamazepine p = 0.015, lamotrigine p = 0.005, phenytoin p = 0.027). Logistic regression analysis revealed a multiplicative interaction between HLA-B*15:02 and HLA-A*24:02. Positivity for HLA-A*24:02 and/or HLA-B*15:02 showed a sensitivity of 72.5% and a specificity of 69.0%. The presence of HLA-A*24:02 in cases with maculopapular exanthema was also significantly higher than in controls (p = 0.023). Meta-analysis of data from Japan, Korea, Malaysia, Mexico, Norway, and China revealed a similar association.
CONCLUSIONS: HLA-A*24:02 is a common genetic risk factor for cutaneous adverse reactions induced by aromatic antiepileptic drugs in the southern Han Chinese and possibly other ethnic populations. Pretreatment screening is recommended for people in southern China.
METHODS: We performed a meta-analysis of three GWAS comprising 684 patients with type 2 diabetes and 955 controls of Southern Han Chinese descent. We followed up the top signals in two independent Southern Han Chinese cohorts (totalling 10,383 cases and 6,974 controls), and performed in silico replication in multiple populations.
RESULTS: We identified CDKN2A/B and four novel type 2 diabetes association signals with p p meta = 2.6 × 10(-8); OR [95% CI] 1.18 [1.11, 1.25]). In silico replication revealed consistent associations across multiethnic groups, including five East Asian populations (p meta = 2.3 × 10(-10)) and a population of European descent (p = 8.6 × 10(-3)). The rs10229583 risk variant was associated with elevated fasting plasma glucose, impaired beta cell function in controls, and an earlier age at diagnosis for the cases. The novel variant lies within an islet-selective cluster of open regulatory elements. There was significant heterogeneity of effect between Han Chinese and individuals of European descent, Malaysians and Indians.
CONCLUSIONS/INTERPRETATION: Our study identifies rs10229583 near PAX4 as a novel locus for type 2 diabetes in Chinese and other populations and provides new insights into the pathogenesis of type 2 diabetes.