Ai-Augmented Management of Pharmacoresistant Insomnia and Restless Legs Syndrome in Park2 Parkinson’s Disease: an N-Of-1 Case Report Integrating Pharmacogenetics, Chronopharmacology, Digital Monitoring, and an Llm Co-Pilot

We report a case of young-onset PARK2 Parkinson’s disease complicated by severe, persistent insomnia, restless legs syndrome (RLS), and REM sleep behavior disorder (RBD), refractory to multiple conventional strategies. The clinical course included features consistent with levodopa-related rebound and augmentation, culminating in a systemic crisis with near-total sleep deprivation and night eating syndrome (NES). An N-of-1 precision approach was implemented by combining pharmacogenetics, molecule- and time-specific repositioning of medications (chronopharmacology), and continuous monitoring using consumer wearable sleep metrics and continuous glucose monitoring (CGM). Crucially, the optimization was conducted within an iterative human–AI loop in which a large language model (LLM) acted as a cognitive co-pilot to integrate multiplexed data streams, surface interaction risks, and generate prioritized hypotheses that were then clinically validated and implemented under continuous neurologist supervision. Key steps included discontinuation of evening levodopa, a strict pre-dinner ‘gastric-sparing’ window on an empty stomach to optimize absorption, and an evening synergy between low-dose pramipexole, pregabalin, and clonazepam. From September 2025 to February 2026, monthly wearable estimates showed deep sleep increasing from 5–8 min to 42–43 min and REM sleep increasing from 7–10 min to 45–50 min. In parallel, metabolic markers improved (HbA1c 7.2% to 6.8%, ALT/GPT 61 to 34 U/L within 12 months), with a 3 kg weight loss. This case illustrates a pragmatic framework for AI-augmented clinical reasoning in a complex multimorbid patient, demonstrating a path from pharmacoresistance to sustained clinical improvement.