Long‐term strength and functional status in inclusion body myositis and identification of trajectory subgroups

Abstract Introduction Objective information on longitudinal disease progression in inclusion body myositis (IBM) is lacking. Methods Longitudinal dynamometry and functional status data were collated from a cohort of IBM patients. Annual change was calculated by means of linear modeling. Trajectories of change in grip, knee extension, IBM Functional Rating Scale (IBM‐FRS) and Neuromuscular Symptom Score (NSS) were identified by means of latent growth mixture modeling. Results Data were collated from 75 IBM patients (348 person‐years follow‐up). Annual strength loss was greatest for pinch (−10%) and knee extension (−4%). Functional deterioration was greatest for males. Three distinct trajectory groups were identified. Rapid deterioration trajectory for grip strength was associated with younger diagnosis age. Rapid deterioration for knee extension strength was associated with older age of diagnosis. Discussion This study has quantified strength change in IBM and identified distinct trajectory groups, which will aid prognostication and stratification for inclusion into future clinical trials.

For determination of anti-CN1A autoantibody status, sera were analyzed in the Department of Biomolecular Chemistry in Nijmegen as part of a previous research project from which a subset of patients were recruited. 2 Briefly, enzyme-linked immunosorbent assay (ELISA) was performed with the three synthetic peptides containing CN1A auto-epitopes previously identified by overlapping peptide microarray analyses. 12 Signals were quantified by determining optical densities at 450 nm (OD450) using methods previously described and defined as seropositive if the OD450 value was greater than or equal to the established cut-off value for the corresponding peptide. 13 Patients were followed prospectively according to clinical need from their time of presentation. Per routine clinical care, quantitative muscle strength measurements using a single CITEC dynamometer (Haren, The Netherlands) were obtained at each clinic visit, which occurred at variable intervals. This was performed by dedicated neurophysiotherapists who received training provided by local muscle trials experts, based at the John Walton Muscular Dystrophy Centre (Newcastle, UK), to minimize inter-rater variability. The force recorded was that required by the examiner to overcome maximum force exerted by the subject (break point), or the maximum exerted force if unable to reach break point. The measurement was repeated until two values within 10% were recorded, and the mean then calculated. The strength of up to 11 different left-sided movements (6 upper-limb and 5 lower-limb) were measured at each visit (see Table 2). Patients also completed two questionnaires at each visit: the Neuromuscular Symptom Score (NSS) 14 and IBM-Functional Rating Score (IBM-FRS). 15 The mean baseline strength of each movement was calculated for the female and male cohorts and compared against reference values of healthy controls. 16 17 This methodology has not previously been applied to longitudinal dynamometry data in IBM. Identification of separate trajectory groups within the cohort was carried out by means of latent GMM for grip, knee extension, IBM-FRS, and NSS. Groups were identified according to strength/score change over time. Grip and knee extension were selected post hoc due to their known particular predilection for strength loss in IBM and high number of available strength measurements. 18 Pinch was not included in GMM analysis due to sufficient longitudinal data only being available in 25 patients. Each study participant was assigned to a single group for grip, knee extension, IBM-FRS, and NSS separately. The number of trajectory groups within the cohort that provided the best fit was identified by standard model comparison techniques: Bayesian Information Criterion and Akaike Information Criterion. 19 Baseline demographics (age of IBM onset, gender) and anti-CN1A autoantibody positivity were compared between trajectory groups using statistical inference tests: Wilcoxon Hip flexion    Ankle dorsiflexion demonstrated strength loss in both cohorts; however, this was greater for the male cohort, compared with the female cohort.
Reduction of both the NSS and IBM-FRS were observed over time. For both, the rate of reduction was more pronounced in the male cohorts, and was particularly apparent for the IBM-FRS.
Identification of latent groups for grip, knee extension, IBM-FRS and NSS was performed, with best fit for models with three groups in each case. The estimated trajectory of each identified group is displayed in Table 4 and Figure 1.
For grip, the largest proportion of the cohort demonstrated gradual strength loss, while smaller proportions followed "rapid deteriorator" and "gradual deteriorator" trajectories. Median baseline strength measurements within each group were similar. Age at time of IBM diagnosis significantly varied across the three groups, with the "rapid deteriorators" demonstrating the youngest age and "nondeteriorators" demonstrating the oldest age. Where tested, anti-CN1A antibody positivity varied between each of the three identified groups, with the majority of the "rapid deteriorator" group being anti-CN1A positive, compared with a minority in the "nondeteriorator" group, although no significant difference was identified.
For knee extension, the majority followed a "variable" trajectory, with no overall strength deterioration after 4 years of follow-up. The median baseline strength measurement was substantially higher for the "gradual deteriorator" group and similar for both the "variable" and "rapid deteriorator" groups. Age significantly varied across the three groups, with younger patients belonging to the "gradual deteriorator" group and older patients belonging to the "rapid deteriorator" group.
For IBMFRS, the majority followed a "rapid deteriorator" trajectory and only a minority demonstrated no deterioration. The "rapid deteriorator" IBM-FRS trajectory group demonstrated the oldest age, followed by the "gradual deteriorator" and "nondeteriorator" groups.
For NSS, the majority followed a rapidly deteriorating NSS trajectory. The "rapid deteriorator" trajectory group were the oldest, followed by the "gradual deteriorator" and "late-deteriorator" groups, similar to that observed with IBM-FRS.

| DISCUSSION
This study has quantified annual dynamometry-derived muscle strength and functional status change in a large "real-world" IBM cohort with long follow-up duration and also identified distinct subgroups, according to  In conclusion, our study has quantified muscle strength and functional status change in a large real-world IBM cohort and identified differences according to gender, age of onset, and possibly CN1A sta-