Accuracy of MSBase criteria to diagnose secondary progressive multiple sclerosis in large German real-world patient cohort

Abstract

Background and ObjectivesAccurate diagnosis of secondary progression in multiple sclerosis (MS) remains a challenge since standardized criteria are missing. In 2016, the MSBase registry presented an algorithm that enabled the diagnosis of secondary progressive multiple sclerosis (SPMS) more than three years earlier compared to diagnosis by neurologists. This work aimed to test whether this approach is equally effective in a real-world cohort of MS patients. MethodsThis longitudinal retrospective study analyzed clinical data of outpatients with MS recorded until October 2020 in the NeuroTransData registry, a Germany-wide network of 153 certified neurologists. Patient data had been captured in time during clinical visits employing a defined standardized clinical data set in the webbased NeuroTransData patient management platform DESTINY®. The time between the diagnosis of relapsing-remitting multiple sclerosis (RRMS) to SPMS onset was compared with one determined using MSBase criteria (MSBC). Group 1 consisted of patients diagnosed with SPMS during the observation period, whereas group 2 included RRMS patients who did not convert to SPMS during the observation period. ResultsOf 21,281 patients with MS included in our registry, 194 and 9506 patients were allocated to groups 1 and 2, respectively. 10.3% of patients with RRMS were diagnosed with SPMS simultaneously, whereas 60.8% were diagnosed with SPMS at least 3 months earlier by treating neurologists compared to the MSBC. In group 1, the MSBC showed a low sensitivity of 32.0% and an accuracy of 61.4% but a high specificity of 89.6%. In group 2, the MSBC identified 7.8% of patients with SPMS at some point during the observation time. Moreover, test-retest variability remains a challenge since 29.4% of patients diagnosed with SPMS by treating physicians did not fulfil the MSBC at a later point in time. DiscussionThese results are inconsistent with earlier SPMS diagnosis using the MSBC compared to clinical diagnosis by treating physicians. Therefore, there remains a need for an operational, structured, and validated approach to SPMS diagnosis.