Baseline cognitive scores predict antidepressant non-response in Chinese adults with MDD

BackgroundSubjective cognitive dysfunction is common in patients with major depressive disorder (MDD), but its relationship with depressive symptoms and treatment response remains unclear, especially in routine clinical settings in China.MethodsWe conducted an 8-week retrospective observational study using routinely collected data from the “Good Sleep 365” smartphone platform at the Sleep Disorders Diagnosis and Treatment Center of Hangzhou Seventh People’s Hospital (China) between 1 November 2017 and 10 October 2024. Adults aged 18–65 years who met ICD-10 criteria for a depressive episode, had baseline PHQ-9 ≥10, and completed the PDQ-D-20 were included. PHQ-9 and PDQ-D-20 were assessed at baseline and at weeks 2, 4, 6 and 8. Correlations between depressive symptoms and subjective cognition were examined. Among 233 patients who completed all follow-ups, three hierarchical logistic regression models were built to predict antidepressant response (PHQ-9 reduction ≥50% at week 8), and their performance was evaluated using ROC curves, calibration, decision curve analysis and a nomogram.ResultsIn 321 patients with MDD, higher PHQ-9 scores were consistently associated with more severe PDQ-D-20 total and domain scores at all time points. Higher baseline PDQ-D-20 scores predicted smaller reductions in PHQ-9. Baseline depressive and cognitive measures did not show marked associations with age, sex, education level, occupation or illness duration. The fully adjusted logistic model that combined baseline PDQ-D-20, symptom scales and demographic/clinical variables showed good discrimination (AUC ≈0.91) and acceptable calibration, and baseline PDQ-D-20 remained an independent predictor of non-response.ConclusionIn this smartphone-based cohort of Chinese outpatients from a sleep-disorders clinic, subjective cognitive dysfunction was closely related to depressive symptom burden and predicted poorer antidepressant response, suggesting that routine cognitive assessment may help identify patients at risk of suboptimal outcomes.