To develop a clinical predictive scoring model to distinguish autoimmune nodopathies (AN) from other demyelinating neuropathies.

AN represent a distinct group of immune-mediated neuropathies with unique pathobiology and therapeutic and prognostic implications.

Patients with AN and other demyelinating neuropathies (controls) were included to develop a clinical prediction model. Penalized least absolute shrinkage and selection operator (LASSO) regression identified candidate predictors, which subsequently entered multivariable Firth penalized logistic regression. An integer-based clinical score was generated.

Sixty four (median age 47.5 years; range 4-82) cases with AN and 72 controls (median age 47.5years; range 9-81) were included: chronic inflammatory demyelinating polyneuropathy (n=20), multifocal acquired demyelinating sensory and motor neuropathy (n=8), Charcot-Marie-Tooth neuropathy (n=12), polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, and skin changes syndrome (n=11), distal acquired demyelinating symmetric neuropathy (n=11), and multifocal motor neuropathy with conduction block (n=10). Of the 10 variables identified by LASSO regression, 5 were selected for multivariable Firth penalized logistic regression based on clinical relevance. Independent predictors of AN included prominent sensory ataxia (S) (odds ratio (OR) (95% CI) 6.12 (2.34–17.16)), cranial neuropathy (N) (OR (95% CI) 10.13 (3.40–35.33), acute to subacute onset (A) (OR (95% CI) 4.58 (1.59–14.58) and cerebrospinal fluid protein (P) > 200 mg/dl (OR (95% CI) 4.48 (1.63–13.36)). A 4 variable score (SNAP score) was derived, assigning one point to each variable. A cut off score of ≥ 2 demonstrated a sensitivity of 78%, specificity of 93%, positive predictive value 91%, and negative predictive value of 83% with an area under the curve (AUC) of 0.898 for predicting AN. Bootstrap validation (1000 samples) yielded an AUC of 0.894 with an optimism of 0.018.

The Autoimmune Nodopathy Predictive (SNAP) score is a simple bedside tool that distinguishes AN from other demyelinating neuropathies and may help identify patients in whom nodal/paranodal antibody testing should be considered.