Glutamic Acid Decarboxylase Antibody Unexpectedly Detected During Recovery Phase in Three Patients With Voltage-Gated Potassium Channel Antibody-Related Autoimmune Encephalitis

Voltage-gated potassium channel (VGKC) and glutamic acid decarboxylase (GAD) antibodies are increasingly tested in patients with suspected autoimmune encephalitis (AE) and other neurological diseases, yet their clinical significance, especially when detected outside of classic contexts, remains poorly defined. This case series of three patients explores an unusual pattern of antibody dynamics in VGKC antibody-positive and leucine-rich glioma-inactivated 1 (LGI1) and contactin-associated protein-like 2 (CASPR2) antibodies-negative AE. All patients underwent immunotherapy based on clinical suspicion and improved, and had delayed emergence of GAD antibodies during recovery phases. The key observations were that VGKC antibody levels were elevated during the acute presentation and either declined or normalized following treatment. GAD antibodies were absent during acute illness but emerged during the post-treatment/recovery phase, when patients were already recuperating. No patient developed syndromes classically associated with GAD antibodies (e.g., worsening encephalitis, stiff person syndrome, and refractory seizures). LGI1 and CASPR2 antibodies were consistently negative throughout all tested phases. These findings bring up the questions of the role(s) GAD antibody may play in the disease processes in AE and other neurological diseases, including the possibility of GAD antibody positivity not necessarily indicating active or causative disease, and suggest it may, in some contexts, represent a phenomenon of immune remodeling or non-specific activation. Additionally, these cases reinforce the potential clinical utility of initiating immunotherapy based on phenotype and clinical suspicion, particularly in VGKC-positive, LGI1/CASPR2-negative presentations, while emphasizing the need for more nuanced interpretation of antibody dynamics over time.