Development of prediction models based on respiratory assessments to determine the need for non-invasive ventilation in patients with myotonic dystrophy type 1.

ABSTRACT

INTRODUCTION: Myotonic dystrophy type 1 is a slowly progressive, multisystem, autosomal dominant disorder, in which the impairments of respiratory systems represent one of the main causes of death. OBJECTIVE: The aim of our study is to develop prediction models to identify the most appropriate test(s) providing indication for NIV. METHODS: DM1 patients attending the NEMO Clinical Center (Milan) between January 2008 and July 2020, who had been subjected to a complete battery of respiratory tests, were retrospectively recruited. Demographic, clinical, and anthropometric characteristics were collected, as well as arterial blood gas (ABG) analysis, spirometry, respiratory muscle strength, cough efficacy, and nocturnal oximetry as respiratory assessments. Patients were stratified in those requiring NIV and those with normal respiratory function. RESULTS: Out of 151 DM1 patients (median age 44 years [35.00-53.00]; male/female ratio 0.80 (67/84)), 76 had an indication for NIV initiation (50.33%). ABG, spirometry, and nocturnal oximetry prediction models resulted in an excellent discriminatory ability in distinguishing patients who needed NIV from those who did not (AUC of 0.818, 0.808, and 0.935, respectively). An easy-to-use calculator was developed to automatically determine a score of NIV necessity based on the prediction equations generated from each aforementioned prediction model. CONCLUSIONS: The proposed prediction models may help to identify which patients are at a higher risk of requiring ventilator support and therefore help in defining individual management plans and criteria for specific interventions early in the disease course.