ORIGINAL ARTICLE Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 Published online 4 March 2020 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle magnetic resonance imaging reveals increase in fat replacement in skeletal muscles Claudia Nuñez-Peralta1, Jorge Alonso-Pérez2, Jaume Llauger1, Sonia Segovia2,3, Paula Montesinos4, Izaskun Belmonte5, Irene Pedrosa5, Elena Montiel5, Alicia Alonso-Jiménez2, Javier Sánchez-González4, Antonio Martínez-Noguera1, Isabel Illa2,3 & Jordi Díaz-Manera2,3,6* 1 Radiology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain, 2Neuromuscular Disorders Unit, Neurology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Spain, 3Centro de Investigación en Red en Enfermedades Raras (CIBERER), Barcelona, Spain, 4 Philips Healthcare Iberia, Madrid, Spain, 5Rehabilitation and Physiotherapy Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain, 6John Walton Muscular Dystrophy Research Center, University of Newcastle, UK Abstract Background Late-onset Pompe disease (LOPD) is a genetic disorder characterized by progressive degeneration of the skeletal muscles produced by a deficiency of the enzyme acid alpha-glucosidase. Enzymatic replacement therapy with recombinant human alpha-glucosidase seems to reduce the progression of the disease; although at the moment, it is not completely clear to what extent. Quantitative muscle magnetic resonance imaging (qMRI) is a good biomarker for the follow-up of fat replace- ment in neuromuscular disorders. The aim of this study was to describe the changes observed in fat replacement in skeletal muscles using qMRI in a cohort of LOPD patients followed prospectively. Methods A total of 36 LOPD patients were seen once every year for 4 years. qMRI, several muscle function tests, spirometry, activities of daily living scales, and quality-of-life scales were performed on each visit. Muscle MRI consisted of two-point Dixon studies of the trunk and thigh muscles. Computer analysis of the images provided the percentage of muscle degenerated and replaced by fat in every muscle (known as fat fraction). Longitudinal analysis of the measures was performed using linear mixed models applying the Greenhouse–Geisser test. Results We detected a statistically significant and continuous increase in mean thigh fat fraction both in treated (+5.8% in 3 years) and in pre-symptomatic patients (+2.6% in 3years) (Greenhouse–Geisser p < 0.05). As an average, fat fraction increased by 1.9% per year in treated patients, compared with 0.8% in pre-symptomatic patients. Fat fraction significantly increased in every muscle of the thighs. We observed a significant correlation between changes observed in fat fraction in qMRI and changes observed in the results of the muscle function tests performed. Moreover, we identified that muscle performance and mean thigh fat fraction at baseline visit were independent parameters influencing fat fraction progression over 4 years (analysis of covariance, p < 0.05). Conclusions Our study identifies that skeletal muscle fat fraction continues to increase in patients with LOPD despite the treatment with enzymatic replacement therapy. These results suggest that the process of muscle degeneration is not stopped by the treatment and could impact muscle function over the years. Hereby, we show that fat fraction along with muscle func- tion tests can be considered a good outcome measures for clinical trials in LOPD patients. Keywords Enzymatic replacement therapy; Fatty replacement; Muscle degeneration; Muscle MRI; Muscle wasting; Pompe disease Received: 5 May 2019; Revised: 25 December 2019; Accepted: 30 January 2020 *Correspondence to: Jordi Díaz-Manera, Neuromuscular Disorders Unit, Hospital de la Santa Creu I Sant Pau, Sant Antoni Maria Claret 167, Barcelona 08025, Spain. Phone: +34-935565986, Fax: +34-935565602, Email:jdiazm@santpau.cat © 2020 The Authors. Journal of Cachexia, Sarcopenia and Muscle published by John Wiley & Sons Ltd on behalf of Society on Sarcopenia, Cachexia and Wasting Disorders This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Follow-up of late-onset Pompe disease patients with muscle MRI 1033 Introduction over a 1 year period.19,20 However, it is not known if the rate of progression of fat replacement in LOPD is stable over time Pompe disease is a rare autosomal recessive disorder or which clinical and/or demographic factors can influence produced by mutations in the GAA gene, which encodes the this progression. enzyme acid alpha-glucosidase. This enzyme metabolizes gly- The aim of our study was to describe the changes observed cogen to glucose inside lysosomes1. Consequently, mutations in fat replacement in skeletal muscles over a period of 4 years in the GAA gene lead to glycogen accumulation in numerous in a long cohort of LOPD patients. tissues, but clinical symptoms are primarily because of cardiac and skeletal muscle involvement.2 Pompe disease has two distinct phenotypes. The classic in- fantile phenotype is characterized by the onset of symptoms Methods in the first months of life, with muscle weakness, hypertro- phic cardiomyopathy, and respiratory insufficiency.3 Patients Study design and participants die early if they are not treated.4 In contrast, late-onset Pompe disease (LOPD) is more heterogeneous in its clinical This study was a prospective, open-label cohort study presentation.5,6 Patients may present with asymptomatic involving 36 patients with a confirmed diagnosis of LOPD hyperckemia or develop a slowly progressive weakness in- performed at the Hospital de la Santa Creu i Sant Pau (HSCSP) volving respiratory, axial, and limb muscles.7,8 in Barcelona, from December 2013 to June 2018. All patients Enzymatic replacement therapy (ERT) with alglucosidase were evaluated once every 12 months (±2 months) for a total alfa (Myozyme®, Sanofi-Genzyme, Cambridge, MA, USA) is ac- of four visits. At each visit, we performed muscle function cepted as the standard treatment for Pompe disease. Several tests, spirometry, quality-of-life scales, and quantitative open label studies have demonstrated its efficacy in maintain- muscle MRI. The study was registered in ClinicalTrials.gov ing muscle function in LOPD since the original clinical trial was with the identifier NCT01914536. The HSCSP ethics commit- published in 2010.9–11 However, a recent single-centre follow- tee approved the study, and all participants signed an up study suggests that the initial improvement lasts for up to informed consent form. All study procedures were performed 2 or 3 years and is followed by a progressive deterioration in in accordance with Spanish regulations. skeletal and respiratory muscle function.12 Clinical trials with Inclusion criteria for the study were (i) diagnosis of LOPD new therapeutic strategies, including improved enzymes or based on recommendations proposed by the European genetic therapies, are being designed and will probably start Pompe Consortium, reduced enzymatic activity in leukocytes, in the next few years.13,14 These trials will probably rely on fibroblasts, or skeletal muscle and/or the presence of two functional outcome measures to demonstrate their efficacy. mutations in the GAA gene21; (ii) no contraindications to The changes that can be found in common muscle function MRI; and (iii) willingness to complete all muscle function tests in LOPD patients treated with Myozime® for several tests, respiratory assessment, and patient-reported out- years have not been clearly established. Functional tests have comes measures. We included in the study both patients with the disadvantage of being subject to motivation, are influ- skeletal muscle weakness and respiratory involvement and enced by several external factors, and often display limited patients without symptoms. interoperator reproducibility.15 Moreover, the process of All patients were studied by three physiotherapists (I.B., I. muscle degeneration and fatty substitution may have started P., and E.M) with considerable experience in neuromuscular without yet influencing the results of the muscle function disorders at HSCSP. All patients were evaluated by the same tests. Therefore, there is a need for robust biomarkers sensi- physiotherapist throughout the whole study. The physiother- tive to disease progression and treatment efficacy.16 apists evaluated muscle function using the following tests: In recent years, muscle magnetic resonance imaging (MRI) the 6 minute walking test, time to walk 10 m, timed up- has been proposed as a valuable tool for both the diagnosis and-go test, time to climb up and down four steps, and the and follow-up of patients with muscle disorders. The identifi- Motor Function Measure 20-item scale (MFM-20). All timed cation of muscles replaced by fat can be easily assessed using tests were performed asking the patient to not use aids for T1-weighted sequences, which is helpful in the diagnosis pro- walking. Muscle strength was studied using both the Medical cess of patients.17 However, patient follow-up requires a Research Council (MRC) scale and hand-held myometry. Daily more specific technique, such as the Dixon MRI sequence, life activities were studied using the activity limitations scale which is able to quantify the amount of fat present in an area for patients with upper and/or lower limb impairments, and of skeletal muscle of interest.18 This test is useful in natural quality of life was analysed using both the Individualized history studies or clinical trials. In the case of Pompe disease, Neuromuscular Quality of Life Questionnaire and the Short two studies applied Dixon to a large cohort of LOPD patients Form 36 questionnaire. We obtained forced vital capacity, and in both cases demonstrated that the sequence is able to both seated and lying down, using the Carefusion Microlab identify an increase in fat replacement in skeletal muscles ML 3500 MK8 spirometer (Carefusion, Yorba Linda, CA, USA). Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1034 C. Nuñez-Peralta et al. Muscle imaging different outcome measures, including muscle function tests and muscle FF quantified by MRI, was performed using linear All patients were examined in a 1.5 magnetic resonance mixed models applying the Greenhouse–Geisser test. To system (1.5 Achieva dStream; Philips, Eindhoven, NL) at evaluate the magnitude of the changes observed in the re- HSCSP as previously described.22 We used the same position- sults of outcome measures over the 3 year follow-up period, ing protocol for all patients: a supine position with the legs we provide the standardized response mean (SRM). The SRM stretched out. is calculated by dividing the mean score change (final visit Axial 3D Dixon Fast Field Echo was performed on the pelvis minus initial visit) by the standard deviation of the change. and thighs by using a thirty-two body coil, with the following The estimation of the magnitude of change over time using parameters: the repetition time (TR) the echo time (TE) = this formula is based on Cohen’s rule for effect size (ES): 5.78/1.8, 4 m, flip angle = 15°, field of view = 520 × 340 × ‘trivial’ (ES < 0.20), ‘small’ (ES ≥ 0.20 <. 50), ‘moderate’ (ES 300 mm, voxel size = 1 × 1 × 3 mm for thighs, and field of view ≥ 0.50 < 0.80), or large (ES ≥ 0.80).22 = 520 × 320 × 200 mm and voxel size = 1.3 × 1.7 × 5 mm for To study the influence of clinical and demographic factors pelvis. This Dixon sequence follows a seven peak fat model- in the progression of skeletal muscle FFs, we performed a ling and provides separate images for fat and water. The ac- two-step process. In the first exploratory step, we used a quisition time was 45 min per patient. Pearson test to analyse the correlation between every factor Three investigators (C.N-P., A.A-P, and J.D-M.) analysed the of interest and the increase in thigh FF. Those factors with Dixon MR images using a Philips Research Image Development both a statistically significant correlation and a correlation Environment tool developed for this project. Regions of coefficient higher than 0.4 were selected for the second step. interests (ROIs) were manually drawn on four slices of the This second step consisted of an ANCOVA analysis in which following muscles: rectus femoris, vastus intermedius, vastus the influence of every selected factor on mean thigh FF lateralis, vastus medialis, adductor magnus, sartorius, gracilis, increase, as detected by MRI, was assessed separately. The semitendinosus, and semimembranosus; on three slices of significance level for all statistical studies was set at p < biceps femoris long head, biceps femoris short head, and 0.05, and post hoc Bonferroni corrections were used when adductor longus; and on one slice for psoas and lumbar needed. Statistical studies were performed with SPSS for paraspinal muscles. Slices covered the proximal, middle, and Mac computers (Version 21, SPSS Inc, Chicago, IL). Heatmaps distal part of the muscles. We used anatomical landmarks to were developed using R version 3.1.1 (R Foundation for ensure consistency between annual analyses. For every ROI, Statistical Computing, Vienna, Austria). the total area and area covered by fat were calculated automatically using the Philips Research Image Development Environment tool. The fat fraction (FF) coefficient was defined as fat/(fat + water), where fat and water were the image inten- Results sity values over the ROI for the fat and water Dixon images, re- spectively. From those two parameters, and assuming that Description of the cohort water content corresponds mainly to muscle, fat and muscle areas were estimated. Cumulative values across all slices were Thirty-six LOPD patients (20 women, 55.5%) were enrolled in also computed, and as a final index, muscle FF was calculated this study. Mean age at the start was 43.9 ± 14.8 years old. as follows: FF = (muscle fat area × 100)/muscle area. Once we Table 1 provides the main clinical data of the patients at visit obtained the values from every muscle, we calculated the 0 (baseline). At Visit 0, 23 patients were already being treated thigh FF as follows: thigh FF = (sum of fat area of all thigh with ERT with Myozyme® for a mean period of 4.1 years. All muscles × 100)/sum of muscle area of all thigh muscles. For these patients continued the treatment without interruptions every muscle, we calculated the muscle fatty transformation during the whole follow-up. Out of the 13 remaining patients, rate = FF final visit FF initial visit/(one-initial FF). four started treatment during the follow-up because of the The time required to quantify FFs in Dixon images was 45 to development of muscle weakness impairing the flexion 60 min per patient. A high degree of reliability was found be- and/or extension of the hip, as demonstrated in clinical tween investigators. The Intraclass coefficient (ICC) coefficient examination. These four patients have been excluded from was 0.982, with a 95% confidence interval from 0.977 to 0.987. the statistical studies, and their results are shown separately. All these patients started treatment 1 to 3 months after Visit 1. The remaining nine patients were considered pre- Statistics symptomatic because they did not report any relevant symp- toms or had no muscle weakness found by clinical We confirmed that the variables were normally distributed examination, at any time during the study. At Visit 0, 10 pa- using a Kolmogorov–Smirnov test, and therefore, parametric tients needed walking aids such as a cane or a stick for walk- statistical studies were used. Longitudinal analysis of the ing outdoors, and 11 patients needed nocturnal non-invasive Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1035 Table 1 Clinical features of the patients included in the study Age at onset Age at Age at symptoms ERT Visit Baseline Baseline Last visit Last visit P G (years) (years) 0 (years) walking aids ventilation walking aids ventilation 1 F 38 47 50 N N Y (stick) N 2 F 30 39 48 Y (stick, wheelchair) Yes (non-invasive) Y (fully wheelchair) Yes (non-invasive) 3 F — — 26 N N N N 4 F 49 59 63 Y (crouch) N Y (crouch) N 5 F 30 42 45 N N N N 6 F 27 47 51 N N N N 7 M 62 67 66 N Y (non-invasive) N Y (non-invasive) 8 F 48 52 59 Y (stick) N Y (stick) Y (non-invasive) 9 F 40 48 55 N N N N 10 M 36 39 42 N Y (non-invasive) N Y (non-invasive) 11 F 15 24 31 Y (Walker, wheelchair) Y (non-invasive) Y (Walker, wheelchair) Y (non-invasive) 12 F 27 39 46 N N N N 13 M 35 45 47 N Y (non-invasive) N Y (non-invasive) 14 M 41 45 51 N Y (non-invasive) N Y (non-invasive) 15 F 39 46 51 Y (crouch) Y (non-invasive) Y (crouch) Y (non-invasive) 16 M — — 22 N N N N 17 M — — 51 N N N N 18 M — — 14 N N N N 19 F 40 64 65 Y (stick) N Y (stick) N 20 F 24 29 35 Y (crouch) N Y (crouch) N 21 F 36 40 41 N Y (non-invasive) N Y (non-invasive) 22 F 14 45 52 N N Y (stick) N 23 M 22 57 64 Y (Stick) Y (non-invasive) Y (stick) Y (non-invasive) 24 M — — 8 N N N N 25 F 42 55 57 Y (Stick) Y (non-invasive) Y (stick) Y (non-invasive) 26 M 35 43 46 N Y (non-invasive) N Y (non-invasive) 27 M 34 51 51 N N N Y (non-invasive) 28 M — — 51 N N N N 29 M 32 43 44 N N N N 30 F 39 48 54 N N Y (stick) N 31 M — — 12 N N N N 32 M — — 51 N N N N 33 M 38 43 42 Y (stick) N Y (stick) N 34 F 20 22 21 N N N N 35 F — — 35 N N N N 36 F 24 40 29 N N N N F, female; G, gender; M, male; N, not needed; P, patient number; Y, yes. ventilation support (Table 1). During the follow-up period, differences. We did not observe significant differences in three patients started using walking aids, and two more timed tests, MFM-20, or daily life activity scales after this pe- started using non-invasive ventilation at night.20 riod of time. In the case of pre-symptomatic patients, we did not observe significant differences in any of the results of the motor function tests, spirometry, or daily life activity scales. Functional assessments The muscle function tests that showed significant differences Quantitative muscle magnetic resonance imaging: in symptomatic-treated patients after the follow-up period progression of fat fraction in late-onset Pompe were lower limb muscle strength, measured using MRC and disease patients hand-held myometry, and respiratory muscle strength (forced vital capacity seated), with p values lower than 0.01 We performed a total of four visits for the whole cohort of and high SRMs as shown in Table 2. Specific details of the patients using two-point Dixon. In every case, we calculated myometry assessments are detailed in Table 3. In summary, the area of the muscle, the muscle fraction, and the FF of all myometry assessments of the lower limbs but knee flexion psoas, paraspinal muscles, and each muscle in the thighs, showed worsening of muscle strength, with knee extension and from that, we derived the mean thigh FF. In the case of and hip flexion being the two that worsened the most symptomatic patients treated with Myozyme®, we observed ( 4.9 and 4.3 kg, respectively). In contrast, myometry as- a statistically significant increase in mean thigh FF of 5.7 ± sessments of the upper limbs did not show significant 3.7% at the end of the follow-up, which corresponds to an Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1036 C. Nuñez-Peralta et al. Table 2 Change between Visits 0 and 3 evaluation in muscle function tests, spirometry, quantitative muscle magnetic resonance imaging, and patient reported outcome measures in treated and pre-symptomatic late-onset Pompe disease patients Symptomatic treated patients n = 22 Pre-symptomatic patients n = 9 Test Visit 0 Visit 3 P SRM Visit 0 Visit 3 P SRM Time to walk 10 m (s) 7.9 ± 3.8 7.8 ± 6.1 0.30 0.3 3.1 ± 0.5 2.8 ± 0.8 0.39 0.1 6MWT (m) 394.4 ± 150.4 422.3 ± 140.1 0.16 0.1 599.6 ± 85.4 596.2 ± 69.5 0.29 0.03 Timed up & go (s) 7.5 ± 7.2 10.2 ± 6.2 0.20 0.4 3.8 ± 1.5 4.1 ± 0.9 0.44 0.4 Time to climb up four stairs (s) 4.8 ± 3.5 4.7 ± 2.9 0.59 0.3 1.4 ± 0.3 1.3 ± 0.3 0.20 0.3 Time to go down four stairs (s) 3.6 ± 2.4 3.3 ± 2.1 0.40 0.3 1.3 ± 0.1 1.3 ± 0.2 0.61 0.3 MRC total score 95.5 ± 11.1 90.8 ± 11.8 0.0001 0.9 110 ± 0 110 ± 0 0.99 0 MRC LL score 47.1 ± 10.2 42.5 ± 10.3 0.001 0.8 60 ± 0 60 ± 0 0.99 0 Myometry total (kg) 203.7 ± 118.8 180.9 ± 107.7 0.005 0.6 312.3 ± 91.3 354.9 ± 82.3 0.21 0.2 Myometry LL (kg) 117.5 ± 67.7 97.3 ± 58.3 0.0001 0.8 172.3 ± 81.6 201.6 ± 47.0 0.35 0.3 Myometry UL (kg) 95.2 ± 50.4 91 ± 48.4 0.27 0.2 130 ± 10.2 145.8 ± 9.4 0.08 0.5 MFM20 score 48 ± 8 47.6 ± 8.7 0.32 0.1 59 ± 2.4 59 ± 1.9 0.66 0 FVC seated (L) 75.9 ± 22.9 69 ± 25.4 0.0001 0.8 93.6 ± 12.8 90.1 ± 10.3 0.48 0.2 FVC supine (L) 67.6 ± 25.3 65.7 ± 24.1 0.74 0.4 89.3 ± 17.6 88.3 ± 22.5 0.37 0.1 ACTIVLIM score 28 ± 7 27.7 ± 6.8 0.84 0.1 36 ± 0 36 ± 0 0.99 0 Mean thigh FF (%) 36.5 ± 18.6 42.31 ± 19.39 0.0001 1.6 11.9 ± 4.4 15.1 ± 5.1 0.01 1.2 Mean Ps FF (%) 81.8 ± 10.2 82.9 ± 8.6 0.32 0.3 38.3 ± 24.4 45.8 ± 24.3 0.26 0.6 Mean value and standard deviation are shown. Greenhouse–Geisser test analysing the data from Visits 1, 2 and 3 were used to find out whether the differences observed were statistically significant. ACTIVLIM, activity limitations; FF, fat fraction; FVC, forced vital capacity; LL, lower limbs; MFM20, Motor Function Measure 20-item scale; MRC, Medical Research Council; Ps, paraspinal muscles; SRM, standardized response mean; UL, upper limbs; 6MWT, 6 min walking test. Table 3 Change between Visits 0 and 3 evaluation in hand-held myometry assessments in symptomatic-treated and pre-symptomatic late-onset Pompe disease patients Symptomatic treated patients n = 22 Pre-symptomatic patients n = 9 Test Visit 0 Visit 3 p SRM Visit 0 Visit 3 p SRM Neck flexion 7.3 ± 5.3 8.1 ± 6.1 0.1 0.15 15.5 ± 5.7 15.1 ± 6.7 0.2 0.05 Shoulder abduction 23.2 ± 10.3 22.4 ± 10.2 0.52 0.12 27.6 ± 9.7 27.7 ± 6.7 0.98 0.01 Shoulder adduction 15.4 ± 10.3 16.2 ± 2.2 0.37 0.14 26.5 ± 8.7 29.9 ± 7.8 0.02 0.37 Elbow flexion 25.3 ± 12.1 22.9 ± 2.3 0.06 0.38 29.7 ± 9.4 37.4 ± 7.7 0.02 0.82 Elbow extension 19.2 ± 9.9 17.8 ± 8.2 0.13 0.34 24.2 ± 7.4 26.3 ± 6.2 0.22 0.41 Hip flexion 21.4 ± 13.8 17.1 ± 12.7 0.007 0.61 33.9 ± 11.9 32.5 ± 12.3 0.46 0.5 Hip extension 11.3 ± 9.1 8.7 ± 8.9 0.001 0.76 33.2 ± 25.1 35.1 ± 8.1 0.06 0.19 Hip abduction 18.3 ± 12.1 14.6 ± 10.9 0.005 0.63 33.4 ± 17.7 30.3 ± 10.7 0.46 0.24 Hip adduction 11.5 ± 7.3 8.8 ± 6.3 0.001 0.76 20.8 ± 6.3 19.8 ± 5.7 0.27 0.42 Knee extension 33.6 ± 19.8 28.7 ± 16.4 0.01 0.54 48.1 ± 15.7 47.1 ± 11.5 0.88 0.06 Knee flexion 20.1 ± 14.2 19.1 ± 11.4 0.31 0.18 35.2 ± 13.2 34.7 ± 8.2 0.91 0.04 Mean value and standard deviation are shown. Greenhouse–Geisser test analysing the data from Visits 1, 2, 3 and 4 was used to find out whether the differences observed were statistically significant, and p values are shown. All units are expressed as kilogrammes. SRM, standardized response mean. Bold emphasis mean statistical significance increase of 1.9% per year (Greenhouse–Geiser test, p < some pre-symptomatic patients at Visit 0, especially those 0.0001 and SRM 1.6) (Figure 1A). Increase in mean thigh FF older than 25 (Figure 1D). In fact, we observed a positive varied substantially from one patient to another, with a max- correlation between the age at study of pre-symptomatic imum of +13.2% and a minimum of +0.8% (Figure 1B). We ob- patients, and the percentage of FF in these two muscles served a significant increase in FF in all muscles analysed, (adductor major: ρ = 0.71, p = 0.02; paraspinal muscles: except the paraspinal muscle, as described in Table 4 and ρ = 0.96, p = 0.0001; Pearson test). On the basis of these displayed in Figure 2. An example of this increase can be data, we decided to divide our cohort of pre-symptomatic found in Figure 3. patients into younger than 25 years old at baseline (n = 5 In the case of pre-symptomatic patients, we observed a patients) and older than 25 years old at baseline (n = 4 statistically significant increase in mean thigh FF of 2.6 ± patients). Mean thigh FF did not change significantly in pre- 1.6% at the end of the follow-up, which represents an in- symptomatic patients younger than 25 (+0.7 ± 0.3% in 4 crease of 0.8% per year (Greenhouse–Geiser test, p < years; Greenhouse–Geisser p = 0.33), while it increased 0.0001) (Figure 1B). However, we identified higher values significantly in patients older than 25 (+3.2 ± 0.4% in 4 years; of FF in the adductor major and the paraspinal muscles of Greenhouse–Geisser p = 0.004). Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1037 Figure 1 (A) Increase in mean thigh fat fraction at year 1, year 2 and year 3 related to the baseline value. Red bars show data from symptomatic pa- tients treated while blue bars show data from presymptomatic patients. Mean increase and standard error is provided. (B) Increase in mean thigh fat fraction in every single patient of the cohort at visit 3 related to baseline. Red bars represent ERT treated patients. Blue bars represent presymptomatic non-treated patients older than 25, and yellow bars represent patients younger than 25 years old. (C) Involvement of paraspinal and adductor major muscles in presymptomatic patients depending on the age. Fat replacement and atrophy of both muscles was observed in patients older than 25 years old. *Paraspinal muscles **Adductor major muscles Correlation between changes in muscle function min walking test, time to climb up four steps and time to and fat fraction identified by magnetic resonance go down four steps (Table 5). In contrast, we did not identify imaging a significant correlation between increase in mean thigh FF and results of the MFM-20 score or the activity limitations We observed a significant correlation between increase in test. mean thigh FF and changes in muscle strength as measured by hand-held myometry and MRC, as shown in Table 5. The strongest correlation was between increase in FF in the vasti Analysis of patients who started treatment during muscles and decrease in muscle strength for knee extension follow-up (ρ = 0.54, p = 0.02). We also found significant correlations between increase in mean thigh FF and the changes ob- A total of four patients started treatment with Myozyme® served in the following timed tests: time to walk 10 m, 6 during follow-up, because of the onset of muscle weakness. Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1038 C. Nuñez-Peralta et al. Table 4 Increase in fat fraction described by each muscle analysed in treated and pre-symptomatic patients Symptomatic treated patients (n = 23) Pre-symptomatic patients (n = 9) Muscle Visit 0 FF Visit 3 FF P SRM Visit 0 FF Visit 3 FF P SRM Rectus femoris 19.5 ± 18.7 21.1 ± 17.1 0.001 0.84 8.8 ± 2.6 11.3 ± 2.7 0.007 1.02 Vastus medialis 31.64 ± 26.9 31.35 ± 27.1 0.009 0.71 10.2 ± 2.2 12.3 ± 2.4 0.14 1.01 Vastus lateralis 22.6 ± 16.4 27.2 ± 21.1 0.0001 1.07 8.5 ± 1.8 12 ± 2.6 0.007 1.34 Vastus intermedius 38.7 ± 27.2 40.6 ± 28.5 0.0001 1.14 9.7 ± 3.1 14.3 ± 4.1 0.01 1.76 Biceps short head 28.3 ± 20.5 32.7 ± 21.5 0.07 1.18 12.8 ± 4.3 15.7 ± 4.2 0.33 0.61 Biceps long head 55.8 ± 29.1 64 ± 26 0.0001 0.96 11.8 ± 5.1 15.4 ± 7.2 0.03 2.91 Semimembranosus 66.7 ± 26.4 70 ± 24 0.02 0.63 16.9 ± 12.7 21.6 ± 14.5 0.69 0.47 Semitendinosus 46.7 ± 29.1 51.3 ± 32.8 0.0001 0.98 12 ± 4.2 14.2 ± 5.8 0.14 1.12 Adductor major 79.4 ± 17.2 79.2 ± 17.3 0.02 0.69 22.1 ± 20.1 27.2 ± 26.3 0.37 0.69 Adductor longus 51.7 ± 33.6 51.9 ± 32.5 0.0001 0.86 11 ± 3.2 14.8 ± 3.2 0.01 3.43 Sartorius 25.3 ± 17.6 27.8 ± 17 0.0001 0.89 19.8 ± 5.3 20.7 ± 5.5 0.49 0.02 Gracillis 22.8 ± 13 27 ± 12.9 0.0001 0.91 15.2 ± 3.8 17. 2 ± 4.4 0.27 0.4 Paraspinalis 80.4 ± 9.7 83.3 ± 6 0.29 0.3 38.3 ± 24.4 45.8 ± 24.3 0.26 0.6 Psoas 54.8 ± 30.6 60.6 ± 28.7 0.051 0.22 24.7 ± 16.2 29.2 ± 15.9 0.27 0.26 Total thighs 39.5 ± 18.2 42.6 ± 19 0.0001 1.69 12.1 ± 4.2 15.1 ± 5 0.0001 1.26 Mean fat fraction and standard deviation at Visits 0 (baseline) and 3 (Year 3) are provided. Linear mixed model applying Greenhouse– Geisser test was performed, and p is provided. Bold numbers in p values are related with significant changes. The standardized response mean (SRM) for every muscle is also provided. FF, fat fraction. Figure 2 Heatmaps showing changes in fat fraction of thigh muscles studied in symptomatic-treated late-onset Pompe disease patients. Patients and muscles are ordered according to hierarchical clustering with increasing replacement severity from bottom to top (patient-rows) and from left to right (muscles-columns). The increase in fat fraction over a 3 year period of time in the muscle of a patient is indicated by the colour of the square. Red colours mean increased fat fraction, while blue colours mean decreased fat fraction. Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1039 Figure 3 Images illustrate the change in fat replacement throughout the follow-up. Fat fraction maps acquired from the thigh at Visits 0 and 3 (0–100% scale). Increase in fat replacement was visible in most of the muscles, with changes most obvious in the following muscles: (A and A’) vastus lateralis (in this patient fat fraction increased from 15.5% to 35.7%), (B and B’) long head of the biceps femoris (fat fraction increased from 57.6% to 79.4%), (C and C’) semitendinosus (fat fraction rose from 20.3% to 51.4%), and (D and D’) semimembranosus (fat fraction increased from 74.2% to 82.4%). Table 5 Correlation between increase in fat fraction during the follow-up and +1.71% respectively). However, we did not apply any sta- and changes observed in muscle function tests in late-onset Pompe dis- tistical tests to these data because they only correspond to ease patients (n = 32) four patients and can therefore be considered merely de- Significance Correlation scriptive, requiring further evaluation. Variable (p value) coefficient Time to walk 10 m (s) 0.011 0.43 Factors influencing increase in fat fraction 6MWT (m) 0.024 -0.48 Timed up & go (s) 0.09 0.32 Time to climb four stairs (s) 0.017 0.46 We analysed which demographic, clinical, and muscle- Time to go down four stairs 0.015 0.47 functional factors could influence the progression of mean (s) MRC LL score (kg) 0.31 0.15 thigh FF in Pompe disease patients. In a first step, we studied Myometry LL (kg) 0.032 -0.4 which variables had a significant correlation with mean in- MFM-20 (score) 0.56 -0.13 crease in thigh FF, using Pearson correlation tests (Table 7). FVC seated (L) 0.023 -0.39 FVC supine (L) 0.20 -0.24 After that, in a second step, we confirmed which of these var- ACTIVLIM score 0.29 0.18 iables were influencing progression of FF using an ANCOVA Correlation was studied using Pearson test. p value was considered analysis. Table 8 shows the variables that correlated with significant if lower than 0.05. the progression of FF in our cohort of patients after this anal- ACTIVLIM, activity limitations; FVC, forced vital capacity; LL, Lower ysis, and Figure 4 displays the graphics showing correlations limbs; MFM20, Motor Function Measure 20-item scale; MRC, Med- ical Research Council; 6 min walking test. between variables. We observed that at an earlier start of Bold emphasis means variables that were selected for the ANCOVA the symptoms, a longer period of treatment, muscle function study at baseline, and muscle FF at baseline were associated with a higher increase in FF in the muscle MRI. All cases started the treatment shortly after Visit 1 (1 to 3 months later). Table 6 shows the mean thigh FF at Visit 0 (baseline) and the changes observed at every visit in these Muscle fat fraction progression: recommendations four patients. We identified a reduction in the rate of FF in- for the follow-up crease between Visits 0 and 1 (before treatment was started, mean increase +3.39%), Visits 1 and 2, and Visits 2 and 3 We wanted to identify which muscles could be best for (both after starting the treatment, mean increase +1.52% following up fatty replacement in LOPD patients in clinical Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1040 C. Nuñez-Peralta et al. Table 6 Changes in mean thighs fat fraction observed in the four patients who started the treatment during the follow-up due to the presence of muscle weakness Mean thighs fat fraction Patient Visit 0 (%) Visit 1 (%) Visit 2 (%) Visit 3 (%) Patient 1 25.26 30.66, +5.4 31.74, +1.08 32.25, +0.51 Patient 2 9.85 9.43, 0.42 9.02, -0,41 10.53, +1.51 Patient 3 57.69 62.11, +4.42 65.5, +3.39 67.93, +2.43 Patient 4 26.05 30.22, +4.17 32.27, +2,05 34.67, +2.4 Group +3.39 +1.52 +1.71 Mean thighs fat fraction and increase over previous visit is shown for every patient. Table 7 Correlation between increase in fat fraction after 4 years of fol- Table 8 A one-way analysis of covariance was conducted to determine low-up and demographic, clinic, and muscle function tests at baseline which factors influenced the progression of mean thighs fat fraction visit Factor Effect on increase in FF Significance Correlation Time of progression 0.49 Variable (p value) coefficient Age at onset of ERT 0.001 Sex (M/W) 0.43 0.15 Time on ERT 0.045 Age at onset of symptoms 0.14 0.12 Ventilation 0.56 (year) Age at ventilation 0.007 Time of progression (year) 0.049 0.34 Time to walk 10 m 0.052 Age at onset of ERT (year) 0.002 0.7 6MWT 0.009 Time on ERT (year) 0.019 0.49 Time to climb four steps 0.004 Antibody titers 0.87 0.05 Time to go down four steps 0.12 Use of aids for walking (Yes/No) 0.31 0.15 MRC lower limbs 0.0001 Use of ventilation (Yes/No) 0.016 0.42 Myometry lower limbs 0.21 Time to walk 10 m (s) 0.06 0.37 MFM20 0.001 6MWT (m) 0.005 0.52 Baseline FF 0.0001 Time to climb four steps (s) 0.004 0.53 Time to go down four steps (s) 0.07 0.35 Results of Greenhouse–Geisser test are shown and were considered Timed up & go test (s) 0.38 0.22 significant if p < 0.05. MFM20 (score) 0.002 0.57 ERT, enzymatic replacement therapy; FF, fat fraction; MFM20, Mo- MRC lower limbs (score) 0.0001 0.65 tor Function Measure 20-item scale; MRC, Medical Research Coun- Myometry lower limbs (kg) 0.1 0.39 cil; 6 min walking test. FVC (L) 0.73 0.15 Bold emphasis: statistical significance ACTIVLIM (score) 0.65 0.16 Baseline mean thigh FF (%) 0.001 0.58 Correlation was studied using Pearson test. Antibody titers were In this group, the muscle with a higher progression was the studied at baseline visit. paraspinal muscles. The second group (15% to 30% baseline ACTIVLIM, activity limitations; ERT, enzymatic replacement ther- apy; FVC, forced vital capacity; M, men; MFM20, Motor Function mean thigh FF) included 11 symptomatic patients. In this Measure 20-item scale; MRC, Medical Research Council; 6 min group, both paraspinal and adductor major were the two walking test; W, women. muscles showing the highest values of FF at baseline Bold emphasis:patients with statistical significance (78.8% and 75.3% respectively). The muscle with a more uni- form progression was the long head of the biceps femoris trials or natural history studies. Increase in FF in most of the long head. The third group (31% to 45% baseline mean thigh muscles analysed was statistically significant in symptomatic- FF) included six individuals, all of whom were symptomatic. treated patients at the end of the follow-up period. How- In this group, paraspinal, adductor major, semimembranosus, ever, we observed a high variability in the amount of in- semitendinosus, and biceps femoris long head muscles were crease from one patient to another. In order to classify replaced by fat to a large extent, and the muscles identified patients in similar groups, we divided our cohort into four with a higher increase in FF were the biceps femoris short groups, depending on the mean thigh FF at Visit 0 (baseline), head, the vastus intermedius, and the vastus lateralis. The which we had previously described as correlating with the last group (more than 45% baseline mean thigh FF) had 11 clinical situation of the patient.23 Table 9 shows the four individuals, all of them symptomatic. In this group, groups and the changes observed in FF during the follow- paraspinal, adductor major and longus, biceps femoris long up. The first group was ‘baseline FF lower than 15%’. All indi- head, semimembranosus, semitendinosus, and psoas were viduals in that group (n = 9) were pre-symptomatic. Although replaced by fat and therefore not useful for the follow-up. these patients had never had muscle weakness or any symp- In contrast, we identified vastus lateralis as the best candi- toms of muscle weakness, we already detected high values date for the follow-up. A summary of all these findings is of FF in the paraspinal muscles at baseline visit (FF = 35%). displayed in Figure 5. Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1041 Figure 4 (A) Correlation between age at start of ERT and increase in mean thigh fat fraction. (B) Correlation between age at start of ventilation and increase in mean thigh fat fraction. (C) Correlation between results of the 6 min walking test (6MWT) at baseline and increase in mean thigh fat frac- tion. (D) Correlation between time to climb four steps at baseline and increase in mean thigh fat fraction. (E) Correlation between MRC for lower limbs at baseline and increase in mean thigh fat fraction. (F) Correlation between mean thigh fat fraction at baseline and increase in mean thigh fat fraction. ERT, enzymatic replacement therapy; MRC, magnetic resonance imaging. Discussion follow-up of patients in natural history studies or clinical trials. In this study, we followed-up a large cohort of patients with Late-onset Pompe disease is a metabolic disorder charac- LOPD, both symptomatic-treated and pre-symptomatic pa- terized by a progressive accumulation of glycogen inside the tients, for a period of 4 years using Dixon quantitative muscle lysosomes of the tissues. The increasing number of MRI and several clinical assessments including muscle func- glycogen-charged lysosomes in the sarcoplasm of the muscle tion tests, spirometry, and activities of daily living scales. fibres triggers a set of processes that eventually lead to mus- We detected a significant and continuous increase in muscle cle fibre death and replacement by fat and fibrotic tissue.24,25 fat replacement in symptomatic patients despite being The loss of muscle fibres is related to the muscle weakness of treated with ERT, which correlated to a decrease in muscle these patients. In recent years, quantitative muscle MRI has strength. In the case of pre-symptomatic patients, we identi- been implemented as one of the best candidate biomarkers fied an increase in FF, especially in paraspinal muscles, that to follow-up disease progression in patients with different was not associated to any change in motor function tests. muscle disorders, such as Duchenne muscle dystrophy, The study has allowed us to confirm that quantitative muscle facio-scapulo-humeral muscle dystrophy, limb girdle muscle MRI is able to identify subtle changes in muscle structure and dystrophy 2I, or Pompe disease.26–28 Most of the previously it can be considered a promising outcome measure for the published studies using qMRI have focused on quantifying Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1042 C. Nuñez-Peralta et al. +6.66% +1.81% paraspinal muscle; Pso: psoas; RF, rectus femoris; VL: Vastus lateralis, VM: Vastus Medialis, Sa, sartorius; SM, semimembranosus; SRM, standardized response mean; ST, semitendinosus; AM, adductor major; AL, adductor longus; BFSH, biceps femoris short head, BFLH, biceps femoris long head; FF, fat fraction; Gra, gracilis; MFTR, muscle fatty transformation rate; Ps, +4.1% +16.9% +20.6% +17.1% +12.8% +24.3% +8.8% +18.8% +5.3% +13.9% +36.4% +17.1% the amount of fat and/or water present in the skeletal MFTR muscles of the patients. Two-point or three-point Dixon, global T2, and 1H spectroscopy are the main techniques used Visit 0 FF >45% (n = 10) for this purpose.18 Here, we have used two-point Dixon, 0.50 0.82 1.52 1.46 1.35 0.90 0.76 0.34 1.28 0.94 1.03 0.59 0.46 0.09 SRM which is a qMRI sequence that can be done in 1.5 Teslas MRIs, available in most of the clinical centres that follow Pompe patients. Our results support the use of Dixon +7.3% +11.5% +5.4% +6.7% +4.1% +7.7% +9.2% +2.2% +1.4% +0.1% V0-V3 +1% +4% +3% +3.7 Patients were divided in four groups depending on Visit 0 mean thighs fat fraction. Per every muscle, fat fraction at Visit 0 and increase at Visit 3 are shown. technique as a reliable outcome measure for the follow-up of patients with LOPD disease. Using this technique, we have shown that skeletal muscle FF increases in symptomatic 34.3% 56.8% 37.9% 66.8% 44.5% 81.2% 75.9% 36.5% 32.1% 72.9% 86.8% 79.1% 87.7% Visit 0 patients despite ongoing treatment with ERT. 91% The yearly progression ratio of mean FF, close to 2%, was stable in our cohort. This increase is slightly higher than the one found in previous studies. For example, the group of Pro- +3.9% +5.7% +9.8% +15.2% +13.1% +7.1% +23.1% +5.1% +4.8% +8.3% +13.1% +7.1% +15.5% +0.5% MFTR fessor Carlier performed a retrospective analysis of 14 LOPD patients treated with ERT and found a significant increase Visit 0 FF 30-45% (n = 6) of 0.9% in FF of all lower limb muscles during a 1 year pe- riod.19 However, that study included lower leg muscles, such 1.44 1.50 1.28 1.01 1.47 1.23 0.91 0.74 1.82 1.49 1.29 1.63 0.06 SRM 0.8 as tibialis anterior or soleus, that are commonly less involved in LOPD, probably reducing the final average increase in FF +3.4% +4.1% +9.5% +2.4% +3.6% +2.9% +3.8% +3.3% +1.3% +7.7% +0.8% V0-V3 +8% +9% +5% observed. When comparing with other studies, it is important to take into account the acquisition parameters used. For example, in our Dixon sequence, we used a flip angle of 15° that is higher than the one used in previous studies.19 A 16.6% 27.9% 18.4% 40.6% 65.6% 84.5% 42.4% 20.7% 20.1% 61.6% 80.7% 85.1% Visit 0 27% 50% higher T1 weighting, due to the use of a higher flip angle, can lead to an overestimation of the FF, especially for higher FF values.29 We have observed considerable variability in the +2.9% +3.3% +4.7% +6.1% +5.6% +8.3% +4.1% +5.3% +23.1% +14.6% +10.7% +16.1% +11.4% +17.7% MFTR increase of FF from one patient to another, ranging from a 14% increase to less than 1%, suggesting that there are clini- Visit 0 FF 15-30% (n = 11) cal or genetic factors than can influence the progression of 1.56 SRM 0.84 1.51 2.51 1.73 1.51 0.64 0.98 0.79 1.43 0.88 0.94 0.58 1.26 the disease. We have observed that the following correlate with a larger increase in FF over a 4 year period: (i) starting ERT at a younger age, suggesting that patients started symp- +2.7% +2.8% +3.9% +5.2% +4.7% +15.6% +3.6% +5.7% +3.3% +2.6% +8.8% +8.1% +5.9% +3.7% V0-V3 Table 9 Increase in fat fraction depending on visit 0 mean thigh fat fraction toms earlier; (ii) starting mechanical ventilation early; (iii) worse baseline muscle performance; and (iv) higher baseline fat replacement detected by MRI. 8.6% 12.9% 12.5% 15.1% 15.6% 32.5% 75.3% 30.7% 18.6% 15.9% 18.3% 47.6% 78.8% Visit 0 We have also detected significant changes in two function 49% tests: muscle strength and forced vital capacity. These results are in line with other recently published evidence showing decrease of muscle strength in LOPD patients despite ERT +1.7% +1.3% +3.4% +3.4% +2.1% +2.2% +2.5% +2.5% -1.1% +0.9% +0.8% +1.5% +1.8% +8.7% MFTR treatment.12 Moreover, in our study, changes in muscle strength is correlated with increase in FF, which has a clear VI, vastus intermedius; VL, vastus lateralis. Visit 0 FF <15% (n = 9) explanation. Increase in FF comes with a decrease in contrac- 0.87 0.72 1.12 2.32 0.55 1.01 1.07 0.75 0.22 0.23 0.29 0.31 0.25 SRM 0.7 tile muscle mass. It is therefore not unusual that those patients in whom FF increased the most had a greater +1.6% +1.2% +2.5% +3.1% +1.8% +1.9% +2.3% -0.9% -1.4% +0.7% +1.3% +1.4% +5.3% decrease in muscle strength, as has been previously shown V0-V3 +2% in other muscle disorders such as limb-girdle muscular dystro- phy type 2I.30 However, decrease in muscle strength did not impact timed test results in our cohort of patients, probably 8.8% 9.8% 8.1% 9.2% 12.9% 10.5% 17.1% 19.1% 15.1% 11.5% 12.3% 20.7% 38.5% Visit 0 11% because these tests are not only influenced by muscle strength, but also for other factors such as respiratory involvement, fatigue, muscle pain, or personal motivation. Muscle Quantification of FF using Dixon analysis can be a compli- BFSH BFLH Gra AM VM Pso SM cated task that requires a high degree of specialization.31 AL VL Sa RF ST Ps VI Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1043 Figure 5 Progression of fat fraction depending on the baseline mean thigh fat fraction. (A) Baseline image showing the localization of thigh muscles (VL: vastus lateralis, RF: rectus Femoris, VM: vastus medialis, VI: vastus intermedius, BFSH: biceps femoris short head, BFLH: biceps femoris long head, ST: semitendinosus, SM: semimembranosus, AM: adductor major, AL: adductor longus, Sa: sartorius, and Gra: gracilis). (B) Increase in fat fraction over a 4 year period in patients with baseline mean thigh fat fraction of 15% to 30%. (C) Increase in fat fraction over a 4 year period in patients with baseline mean thigh fat fraction of 30% to 45%. (D) Increase in fat fraction over a 4 year period in patients with baseline mean thigh fat fraction >45%. FF, fat fraction. No automatized software is yet available that can distinguish scoliosis in juvenile patients and lumbar pain or changes in one muscle from the others, and so the analysis has to be the walking pattern in adults.34,35 The adductor major muscle performed manually.32 In this study, all MRIs were separately is also involved early in the progression of the disease, and in analysed by three investigators with considerable experience fact, we have observed that these two muscles are almost in this field. We have observed that if anatomical landmarks completely replaced by fat in most of our symptomatic pa- are well established, the variability in the measurements be- tients.36 The muscles of the posterior compartment of the tween investigators is very low. Previous studies have shown thighs are affected later, and so they can be considered good that fat replacement is not homogeneous in muscles in pa- candidates for monitoring FF progression during intermediate tients with muscle diseases, and therefore, analysing one stages of the disease. Finally, vasti muscles are involved in slice only is probably not sufficient to estimate the mean FF more advanced cases. In fact, only two patients in our cohort of that muscle.33 It is therefore necessary to include several had clear quadriceps weakness leading to difficulties walking slices of every muscle in the analysis, increasing significantly downstairs or standing. Vastus lateralis is, in the light of our the time needed to complete the study. The identification results, a good candidate for monitoring FF progression in of single muscles that could act as reporters of the increase patients that are at an advanced clinical stage. in FF is desirable because it could make studies more effi- We had the opportunity to study four patients before cient. Our study allows a description of the typical order in and after starting ERT. Interestingly, we observed that ERT which different muscles are involved, thereby sketching out decreased the rate of increase of FF from 3% to 1.7%, which a hypothetical natural history of fat replacement progression. is a value close to the mean observed in the whole cohort, In our opinion, this is an important step that should be made suggesting that ERT with Myozyme® is effective in slowing to select the best muscle to identify changes in FF over a down the progression of muscle degeneration in Pompe dis- period of time. If a muscle that has already been completely ease. These results are in accordance with those previously replaced by fat is selected, no changes will be observed as reported by the group of Professor Carlier. This group re- there is no more muscle to lose. On the other hand, if a ported that increase in FF was lower in patients treated muscle that is not involved until very late in the progression with ERT than in those not treated after 1 year of follow- of the disease is selected, no changes will likely be recorded, up (19). even though FF could actually be increasing in other muscles. Through our study, we realized that Dixon has some clear We have observed that paraspinal muscles are the first limitations in the study of pre-symptomatic Pompe patients. muscles involved in LOPD from a radiological point of view, Increase in mean thigh FF was only significant in those pre- probably leading to the frequently reported symptoms of symptomatic patients older than 25 years old but not in the Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1044 C. Nuñez-Peralta et al. younger patients whose age ranged from 8 to 21 years old. importance of the radiological findings. As fat replacement is This can be related to the fact that the process of muscle de- an irreversible process, we suggest that muscle MRI should generation and fatty replacement was not active in younger be part of the initial assessment of patients with LOPD and patients. However, we identified two different situations that taken into account in deciding when to start treatment in should be discussed and analysed in further studies. We al- pauci-symptomatic patients. ready identified fat replacement in paraspinal and adductor To conclude, the results of our longitudinal study (i) major muscles of some pre-symptomatic patients at baseline establish the rate of annual progression of fat replacement visit, suggesting that the process of muscle degeneration pro- in LOPD patients treated with ERT with Myozyme; (ii) gresses very slowly in Pompe, and is subclinical in some pa- pinpoint clinical factors influencing the increase in muscle tients, until a moment in which patients experience muscle FF; and (iii) show the clinical relevance of the changes ob- weakness. However, we identified two young patients that served in muscle MRI. developed muscle weakness, one during the study, and the other just at the end, in which increase in FF was not detected despite those patients complaining of worsening Acknowledgements muscle weakness. In the first case, we observed mild psoas and glutei weakness in a 19-year-old girl. At that moment, We would like to thank the Spanish Association of Patients her MRI did not show muscle fat replacement. The second with Glycogenosis (www.glucogenosis.org) for their support case was a 12-year-old boy who developed scoliosis and for our investigation. We thank Neil McMillan for editorial glutei weakness 6 months after Visit 3 (end of the study). support and Ignasi Gich for help with statistics. We thank We performed a new MRI that did not show muscle fat the entire MRI technician team: Esther Alemany, Nieves replacement. It is therefore probable that in some cases, Campillos, Elisenda Mestres, Eugenia Torres, Ricard Cullell, the accumulation of glycogen-charged lysosomes, or the Ingrid Rubio, Alberto Fernández, and Mario González, for presence of free glycogen in the sarcoplasm, disrupts muscle their patience and support for the study. We also thank fibre contraction, leading to muscle weakness before induc- Mrs Concepción Escolá for her assistance during the visits of ing necrosis of the fibre and fatty substitution.37 In these the study. This investigation was sponsored by the following cases, Dixon may not be useful, because there are no grants, one from Sanofi Genzyme and another from the Span- changes in fat tissue content. Therefore, a combination of ish Ministry of Health, Fondos FEDER-ISCIII PI15/01822 and Dixon with a muscle MRI sequence able to identify glycogen PI18/01525 to Dr Jordi Díaz-Manera. Isabel Illa has received in the muscle fibres of weak muscles may more informative speaker honorarium from Grifols and Sanofi-Genzyme. Jordi in this subset of patients. Muscle glycogen can be identified Díaz-Manera has received speaker honorarium from PTC and quantified using some specific muscle MRI sequences, Therapeutics and Sanofi-Genzyme. The authors of this such as GlycoCEST or (13)C-MR spectroscopy, but this tech- manuscript certify that they comply with the ethical guide- nology is not routinely available on standard clinical scan- lines for authorship and publishing in the Journal of Cachexia, ners.38,39 An increased signal intensity in Short Tau Sarcopenia, and Muscle.42 Inversion Recovery (STIR) sequence in muscles from LOPD pa- tients has been described, suggesting that this is related to Contributions of authors the presence of water molecules retained by glycogen.40 In a similar way, increase in the water T2 signal in some muscles CNP, JL, SS, PM, IB, IP, EM, JSG, AMN, II, and JDM carried out was identified in LOPD patients, which correlated with a the concept and design of the study. CNP, JAP, JL, IB, IP, EM, higher increase in FF after 1 year follow-up, suggesting that and AAJ performed the acquisition and analysis of data. CNP, water T2 could be a potential marker of disease activity.19 JAP, JL, SS, PM, AAJ, JSG, AMN, II, and JDM drafted of the However, increased signal intensity in STIR or changes in wa- manuscript. SS performed the acquisition of data. II and ter T2 is not specific and can be observed in muscle fibre ne- JPM obtained funding for the study. Spanish Pompe Study crosis, inflammation, or denervation.41 In our opinion, for Group: see supplemental Table S1. Pompe patients, studying MRI sequences that are able to identify changes related to the accumulation of glycogen is likely to be especially useful for the study of juvenile patients Online supplementary material who are experiencing muscle symptoms. Our results support the use of muscle MRI in the decision- Additional supporting information may be found online in the making process in patients with Pompe disease. We have Supporting Information section at the end of the article. shown that a high value in FF in skeletal muscles at baseline is associated with a quicker progression of the disease. Moreover, there is a correlation between increase in FF and Table S1 Members of the Spanish Pompe Study Group changes in muscle function tests, which translates the clinical Placement Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 Follow-up of late-onset Pompe disease patients with muscle MRI 1045 Conflicts of interest Nuñez-Peralta, Jorge Alonso-Pérez, Jaume Llauger, Sonia Segovia, Paula Montesinos, Izaskun Belmonte, Irene Pedrosa, The company has neither reviewed the data nor suggested Elena Montiel, Alicia Alonso-Jiménez, Javier Sánchez-González, changes in any of the conclusions of the paper. Claudia and Antonio Martínez-Noguera have no conflict of interest. References 1. Lim J, Li L, Raben N. Pompe disease: from 13. Byrne BJ, Geberhiwot T, Barshop BA, 25. Schänzer A, Kaiser AK, Mühlfeld C, Kulessa pathophysiology to therapy and back Barohn R, Hughes D, Bratkovic D, et al. A M, Paulus W, von Pein H, et al. Quantifica- again. Front Aging Neurosci 2014;6:1–14. study on the safety and efficacy of tion of muscle pathology in infantile 2. Kohler L, Puertollano R, Raben N. Pompe reveglucosidase alfa in patients with late- Pompe disease. Neuromuscul Disord disease: From basic science to therapy. onset Pompe disease. Orphanet J Rare Dis 2017;27:141–152. Neurotherapeutics 2018;15:928–942. 2017;12:1–10. 26. Willis TA, Hollingsworth KG, Coombs A, 3. Case LE, Beckemeyer AA, Kishnani PS. In- 14. Puzzo F, Colella P, Biferi MG, Bali D, Paulk Sveen ML, Andersen S, Stojkovic T, et al. fantile Pompe disease on ERT: update on NK, Vidal P, et al. Rescue of Pompe disease Quantitative muscle MRI as an assessment clinical presentation, musculoskeletal man- in mice by AAV-mediated liver delivery of tool for monitoring disease progression in agement, and exercise considerations. Am secretable acid a-glucosidase. Sci Transl LGMD2I: a multicentre longitudinal study. J Med Genet C Semin Med Genet Med 2017;9:418. PLoS ONE 2013;8:e70993. 2012;160C:69–79. 15. Lachmann R, Schoser B. The clinical 27. Andersen G, Dahlqvist JR, Vissing CR, Heje 4. Kishnani P, Corzo D, Nicolino M, Byrne B, relevance of outcomes used in late-onset K, Thomsen C, Vissing J. MRI as outcome Mandel H, Hwu WL, et al. Recombinant hu- Pompe disease: can we do better? measure in facioscapulohumeral muscular man acid α-glucosidase: major clinical ben- Orphanet J Rare Dis 2013;8:160. dystrophy: 1-year follow-up of 45 patients. efits in infantile-onset Pompe disease. 16. Straub V, Bertoli M. Where do we stand in J Neurol 2017;264:438–447. Neurology 2007;68:99–109. trial readiness for autosomal recessive limb 28. Willcocks RJ, Rooney WD, Triplett WT, 5. Schuller A, Wenninger S, Strigl-Pill N, girdle muscular dystrophies? Neuromuscul Forbes SC, Lott DJ, Senesac CR, et al. Multi- Schoser B. Toward deconstructing the Disord 2016;26:111–125. center prospective longitudinal study of phenotype of late-onset Pompe disease. 17. Diaz-Manera J, Llauger J, Gallardo E, Illa I. magnetic resonance biomarkers in a large Am J Med Genet C Semin Med Genet Muscle MRI in muscular dystrophies. Acta duchenne muscular dystrophy cohort. Ann 2012;160C:80–88. Myol [Internet] 2015;34:95–108. Neurol 2016;79:535–547. 6. Hagemans ML, Winkel LP, Van Doorn PA, 18. Burakiewicz J, Sinclair CDJ, Fischer D, 29. Bydder M, Yokoo T, Hamilton G, Middleton Hop WJ, Loonen MC, Reuser AJ, et al. Clin- Walter GA, Kan HE, Hollingsworth KG. MS, Chavez AD, Schwimmer JB, et al. Re- ical manifestation and natural course of Quantifying fat replacement of muscle by laxation effects in the quantification of fat late-onset Pompe’s disease in 54 Dutch pa- quantitative MRI in muscular dystrophy. J using gradient echo imaging. Magn Reson tients. Brain 2005;128:671–677. Neurol 2017;264:2053–2067. Imaging 2008;26:347–359. 7. Echaniz-Laguna A, Carlier RY, Laloui K, 19. Carlier PG, Azzabou N, de Sousa PL, Hicks 30. Murphy AP, Morrow J, Dahlqvist JR, Carlier P, Salort-Campana E, Pouget J, A, Boisserie JM, Amadon A, et al. Skeletal Stojkovic T, Willis TA, Sinclair CD, et al. et al. Should patients with asymptomatic muscle quantitative nuclear magnetic Natural history of limb girdle muscular pompe disease be treated? A nationwide resonance imaging follow-up of adult dystrophy R9 over 6 years: searching for study in France. Muscle Nerve Pompe patients. J Inherit Metab Dis trial endpoints. Ann Clin Transl Neurol 2015;51:884–889. 2015;38:565–572. 2019;6:1033–1045. 8. Hagemans ML, Winkel LP, Hop WC, 20. Figueroa-Bonaparte S, Llauger J, Segovia S, 31. Carlier PG, Marty B, Scheidegger O, Reuser AJ, Van Doorn PA, Van der Ploeg Belmonte I, Pedrosa I, Montiel E, et al. Loureiro de Sousa P, Baudin PY, Snezhko AT. Disease severity in children and adults Quantitative muscle MRI to follow up late E, et al. Skeletal muscle quantitative nu- with Pompe disease related to age and onset Pompe patients: a prospective study. clear magnetic resonance imaging and disease duration. Neurology 2005;64: Sci Rep 2018;8:1–11. spectroscopy as an outcome measure for 2139–2141. 21. van der Ploeg AT, Kruijshaar ME, Toscano clinical trials. J Neuromuscul Dis 2016;3: 9. van der Ploeg AT, Clemens PR, Corzo D, A, Laforêt P, Angelini C, Lachmann RH, 1–28. Escolar DM, Florence J, Groeneveld GJ, et al. European consensus for starting and 32. Ogier A, Sdika M, Foure A, Le Troter A, et al. A randomized study of alglucosidase stopping enzyme replacement therapy in Bendahan D. Individual muscle segmenta- alfa in late-onset Pompe’s disease. N Engl adult patients with Pompe disease: a 10- tion in MR images: a 3D propagation J Med 2010;362:1396–1406. year experience. Eur J Neurol 2017;24: through 2D non-linear registration ap- 10. Schoser B, Stewart A, Kanters S, Hamed A, 768–e31. proaches. Proc Annu Int Conf IEEE Eng Jansen J, Chan K, et al. Survival and long- 22. Cohen J. Statistical power analysis for the Med Biol Soc 2017;2017:317–320. term outcomes in late-onset Pompe dis- behavioural sciences. revised edition ed. 33. Chrzanowski SM, Baligand C, Willcocks RJ, ease following alglucosidase alfa treat- New York: Academic Press; 1977. Deol J, Schmalfuss I, Lott DJ, et al. Multi- ment: a systematic review and meta- 23. Figueroa-Bonaparte S, Segovia S, Llauger J, slice MRI reveals heterogeneity in disease analysis. J Neurol 2017;264:621–630. Belmonte I, Pedrosa I, Alejaldre A, et al. distribution along the length of muscle in 11. Toscano A, Schoser B. Enzyme replacement Muscle MRI findings in childhood/adult Duchenne muscular dystrophy. Acta Myol therapy in late-onset Pompe disease: a onset Pompe disease correlate with 2017;36:151–162. systematic literature review. J Neurol muscle function. PLoS ONE 2016;11: 34. Taisne N, Desnuelle C, Juntas Morales R, 2013;260:951–959. e0163493. Ferrer Monasterio X, Sacconi S, Duval F, 12. Kuperus E, Kruijshaar ME, Wens SCA, de 24. Thurberg BL, Lynch Maloney C, Vaccaro C, et al. Bent spine syndrome as the initial Vries JM, Favejee MM, van der Meijden Afonso K, Tsai AC, Bossen E, et al. Charac- symptom of late-onset Pompe disease. JC, et al. Long-term benefit of enzyme terization of pre- and post-treatment Muscle Nerve 2017;56:167–170. replacement therapy in Pompe disease: a pathology after enzyme replacement 35. Herbert M, Case LE, Rairikar M, Cope H, 5-year prospective study. Neurology therapy for Pompe disease. Lab Invest Bailey L, Austin SL, et al. Early-onset of 2017;89:2365–2373. 2006;86:1208–1220. symptoms and clinical course of Pompe Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555 1046 C. Nuñez-Peralta et al. disease associated with the c.-32–13 T > G 38. van Zijl PC, Jones CK, Ren J, Malloy CR, Asymptomatic Pompe disease: can muscle variant. Mol Genet Metab 2019;126: Sherry AD. MRI detection of glycogen magnetic resonance imaging facilitate 106–116. in vivo by using chemical exchange satura- diagnosis? Muscle Nerve 2016;53:326–327. 36. Carlier RY, Laforet P, Wary C, Mompoint D, tion transfer imaging (glycoCEST). Proc Natl 41. Azzabou N, Carlier PG. Fat quantification Laloui K, Pellegrini N, et al. Whole-body Acad Sci U S A 2007;104:4359–4364. and T2 measurement. Pediatr Radiol muscle MRI in 20 patients suffering from 39. Baligand C, Todd AG, Lee-McMullen B, 2014;44:1620–1621. late onset Pompe disease: involvement Vohra RS, Byrne BJ, Falk DJ, et al. 13C/ 42. von Haehling S., Morley JE, Coats AJS., patterns. Neuromuscul Disord 2011;21: 31P MRS metabolic biomarkers of disease Anker SD. Ethical guidelines for publishing 791–799. progression and response to AAV delivery in the Journal of Cachexia, Sarcopenia and 37. Griffin J. Infantile acid maltase deficiency. I. of hGAA in a mouse model of Pompe dis- Muscle: update 2019. J Cachexia Muscle fiber destruction after lysosomal ease. Mol Ther 2017;7:42–49. Sarcopenia Muscle 2019; 10: 1143-1145. rupture. Virchows Arch B Cell Pathol Incl 40. Pichiecchio A, Berardinelli A, Moggio M, Mol Pathol 1984;45:23–36. Rossi M, Balottin U, Comi GP, et al. Journal of Cachexia, Sarcopenia and Muscle 2020; 11: 1032–1046 DOI: 10.1002/jcsm.12555