Specific clinical traits have been identified that may be early symptoms of multiple sclerosis, according to a study published in Multiple Sclerosis and Related Disorders. These traits, when identified, were able to accurately distinguish patients with multiple sclerosis from those without.
This study drew from administrative data in British Columbia and Saskatchewan and included 8669 cases of multiple sclerosis and 40,867 matched controls from the general population. A training dataset, comprising 66.6% of the British Columbian cohort, was used to arrange L1 penalized logistic regression models to predict multiple sclerosis from filled prescriptions and encounters with physicians and hospitals for 5 years preceding the first demyelinating event of the case. Logistic regression was utilized to perform internal/external validation of the predictors in both sets of data.
Physician data was associated with good predictive performance, with an internal/external validation area under the curve of 0.81/0.79. Examples of multiple sclerosis-related International Classification of Diseases codes that were generated by physicians and validated in both British Columbia and Saskatchewan included eye disorders, central and peripheral nervous system disorders, and cerebrovascular disease (adjusted odds ratio, 1.3 to 7.0). The predictive power of hospital and prescription data ranged from very poor to poor, with respective internal/external validation areas under the curve of 0.54/0.55 and 0.66/0.61. Hospitalizations due to spinal cord diseases, urinary system diseases, or urinary antispasmodic or anti-vertigo prescriptions correlated with odds of multiple sclerosis that were 2 to 3 times higher (adjusted odds ratio 2.3 to 3.3).
The study researchers conclude that they “identified specific clinical characteristics that may be part of the [multiple sclerosis] prodrome. Further, we found that when identified by physician encounters, these characteristics were capable of differentiating between people with and without [multiple sclerosis] with good accuracy. Our findings provide more detailed insight into the prodromal phase of [multiple sclerosis], based on the ‘real life’ clinical care setting, and may contribute to the earlier recognition and diagnosis of [multiple sclerosis].”
Högg T, Wijnands JMA, Kingwell E, et al. Mining healthcare data for markers of the multiple sclerosis prodrome. Mult Scler Relat Disord. 2018 Aug 8;25:232-240. doi: 10.1016/j.msard.2018.08.007