To determine the positive predictive value (PPV) of myelin oligodendrocyte glycoprotein (MOG)-IgG for MOG-IgG-associated disorder (MOGAD) in a tertiary referral center.

MOGAD is a distinct demyelinating disease characterized by attacks of acute optic neuritis, myelitis, brain/brainstem dysfunction, or combinations thereof. MOG-IgG positivity assessed by live cell-based assays is highly specific and confirms MOGAD diagnosis when accompanied by the correct clinical phenotype, but false positives may occur. 

We retrospectively identified Mayo Clinic patients consecutively tested for MOG-IgG as part of their diagnostic work-up (1/1/2018-12/31/2019) with a live cell fluorescence-activated cell sorting assay. Sera with IgG-binding-index ≥2.5 were titrated and an end-titer ≥20 was used as positivity cut-off. The pre-test probability was considered high with a typical MOGAD phenotype and low for atypical phenotypes. Medical records of MOG-IgG-positive cases were independently reviewed and classified by two neurologists blinded to autoantibody titers as “true-positive” or “false-positive” at last follow-up; consensus was reached for discordant cases.

Of 1260 patients tested, 92 (7%) were MOG-IgG positive. Twenty six (28%) cases were designated as false-positive by the two raters, with independent agreement in 25 (96%). Alternative diagnoses included: multiple sclerosis, 11; ischemia, 3; B12 deficiency, 2; neoplasia, 2; genetically-confirmed adrenomyeloneuropathy, 1; others, 7. The overall PPV (number of true positives/total positives) was 72% and was titer dependent: titer ≥1000, PPV 100%; titer=100, PPV 82%; Titer=20-40, PPV 51%. The median (range) end-titer was higher with true positives (100 [20-10000]) than false positives (40 [20-100]); p<0.0001. The PPV was higher for children (94%) than adults (67%) and for high (85%) than low (12%) pre-test probability; p<0.05. The specificity of MOG-IgG was 97.9%.

MOG-IgG is a highly specific biomarker for MOGAD but caution is advised with low titers and atypical phenotypes. Ordering MOG-IgG in low pre-test probability situations increases the risk of false positive results.