Background Voltage-gated potassium channel (VGKC)-complex antibodies can be associated with a range of immunotherapy-responsive clinical presentations including limbic encephalitis, Morvan's syndrome and acquired neuromyotonia. However, there are patients with positive levels in whom the significance is uncertain.

Objective To evaluate the clinical significance associated with positive (>100 pM) VGKC-complex antibodies.

Methods Over a 4-year period, 1053 samples were sent for testing of which 55 were positive. The clinical presentations, final diagnoses and responses to immunotherapies, when given, were assessed retrospectively and the likelihood of autoimmunity was categorised as definite, possible, unlikely or undetermined (modified from Zuliani et al 2012).

Results Only 4 of the 32 patients with low-positive (100–400 pM) levels were considered definitely autoimmune, 3 with peripheral nerve hyperexcitability and 1 with a thymoma; 3 were given immunotherapies. Of the remaining 28 with low-positive levels, 13 (3 of whom had tumours) were considered possibly autoimmune, and 15 were unlikely or undetermined; 1 was given immunotherapy unsuccessfully. Of the 23 patients with high-positive (>400 pM) levels, 12 were given immunotherapies, 11 of whom showed a good response. 11 were considered definitely autoimmune, 10 with limbic encephalitis (antibody specificity: 5 LGI1, 1 contactin2, 2 negative, 2 untested) and 1 with a tumour. In the remaining 12, autoimmunity was considered possible (n=9; most had not received immunotherapies), or unlikely (n=3).

Conclusions As antibody testing becomes more widely available, and many samples are referred from patients with less clear-cut diagnoses, it is important to assess the utility of the results. VGKC-complex antibodies in the range of 100–400 pM (0.1–0.4 nM) were considered clinically relevant in rare conditions with peripheral nerve hyperexcitability and appeared to associate with tumours (12.5%). By contrast high-positive (>400 pM; >0.4 nM) levels were considered definitely (38%) or possibly (49%) clinically relevant, but not all patients had a ‘classical’ limbic encephalitis and some did not receive immunotherapies.