LAUSR

Late-onset Pompe disease (LOPD) is an autosomal recessive disease caused by acid alpha glucosidase deficiency. The phenotype is a progressive proximal myopathy. Respiratory failure is the main life-threatening complication, usually resulting from diaphragm weakness, which may be independent of the severity of motor involvement. Screening for diaphragm function include the assessment of postural drop in forced vital capacity (FVC) moving from sitting to supine position (DVC), and measurement of Maximal Inspiratory Pressure (MIP). We previously observed by Magnetic Resonance Imaging (MRI) that axial muscles involvement (posterior trunk, abdominal wall) represents a feature peculiar to LOPD, otherwise uncommon in other myopathies. Axial involvement may be suspected in patients with chronic lumbar pain, hyper-lordosis, and abdominal prominence, but axial muscles are difficult to assess by clinical examination alone: thus, imaging fills a clinical need. What is the clinical impact, if any, of axial muscles involvement? Are there any functional correlates, beyond lumbar pain and postural changes? Indeed, trunk muscles may be involved in respiration, with posterior muscles participating in inspiration, and anterior abdominal wall muscles contributing to forced expiration. Our hypothesis is that axial involvement may be a sentinel sign of respiratory dysfunction, and that MRI of the axial muscles may represent an effective approach to screen for respiratory impairment in LOPD, to optimize pulmonary evaluation and treatment strategy for these individuals. We investigated prospectively 19 patients (8 females) aged 54.6§ 18.2 years (range 25 76) with genetically confirmed LOPD. Clinical, demographic, genetic data are in Supplementary Table 1. Muscle MRI was performed as previously described by a 1.5T MRI scanner (1.5T Philips Intera and 1.5T Philips Ahieva XR Realeas) using T1-weighted spin-echo axial images from the mid-dorsal segment to the sacrum, using the same parameters (TR=300ms, TE=10ms, thickness =10mm, matrix=640£ 640, in plane resolution 0.6£ 0.6mm). Muscles were graded qualitatively according to the Mercuri score. We considered two muscles of the posterior wall of the lower trunk (Quadratus lumborum, Iliocostalis lumborum), and seven anterior wall muscles (Multifidus, Longissimus, Iliopsoas, Rectus abdominis, Transversus abdominis, Obliquus externus abdominis, Obliquus internus abdominis). Two independent observers blinded to clinical data examined all scans. Respiratory assessment was performed within 48 h ofm MRI, according to standard guidelines. A postural drop of FVC (DVC) 30% was considered expression of diaphragmatic weakness; MIP was measured from the Functional Residual Capacity in the upright position; Maximal Expiratory Pressure (MEP) was measured at the Total Pulmonary Capacity. Both MIP and MEP were repeated at least three times or until two identical readings were obtained, with patients receiving strong verbal encouragement; the best value of both measurements was used. Deviations of quantitative variables from normality were calculated by the Shapiro-Wilk test (p<0.05). Quantitative variables with normal distribution are described as mean § standard deviation or by median (25th 75th percentiles) otherwise. To test for significant differences in terms of normally distributed variables between binary conditions we used the Welch’s t-test, and the Wilcoxon rank-sum test to test for differences in terms of variables deviation from the normal. Pairwise correlations were estimated by the Spearman test, associations between categorical variables by the Fisher’s exact test. The significance threshold was set to p<0.008 based on the Bonferroni correction accounting for the number of muscles for which the MRI score was evaluated (a=0.05/6 tests). Univariate tests were applied to evaluate: a) significant correlations between MRI measurements and: MIP, MEP, FVC, DVC, FVC%; b) presence of significant associations between MRI measurements corresponding to the analyzed muscles and: Diaphragm 20 or Diaphragm 30. Statistical tests were performed by the R software v. 3.1.0 (www.r-project.org/). Summary statistics reporting the characteristics of the analyzed patients are in Supplementary Table 2 (quantitative variables) and 3 (categorical variables). Fig. 1 shows three different patterns of severity of trunk involvement. Involvement of the Internal Oblique and Multifidus correlated with worse MIP (rho=0.85, p = 0.004 and rho=0.75, p = 0.003 respectively). Similarly, Internal Oblique, Multifidus and Longissimus muscles were positively correlated with DVC values (rho=0.86, p<0.001; rho=0.80, p<0.001 and rho=0.73, p<0.001