A total of 41 older patients participated in this prospective observational study, which was undertaken between September 2017 and November 2018 at our geriatric acute care hospital department. In order to investigate the effect of immobility on mid-thigh cross-sectional muscle area, patients were recruited based on their mobility status at hospital admission. Mobility status was evaluated according to walking ability as described by the respective item of the Barthel-Index (BI) and patients were grouped into two categories as follows:
Mobile group. Older patients with either walking ability score of 15 (who can sit or walk at least 50 m independently without a walker or help but may use any aid except walker, i.e. stick) or 10 (who walks at least 50 m with a walker or with help of one person) were considered as mobile patients.
Immobile group. Older patients with either walking ability score of 5, who can walk the distances in the living area with help or 0, who do not meet these criteria and lost mobility due to acute disease since less than 3 weeks before admission and are expected to have a poor rehabilitation prognosis for the week after admission, were considered as immobile patients.
It is worth noting that mobile and immobile patients were selected from the geriatric day clinic and the geriatric hospital department, respectively. Patients from the geriatric day clinic had better functional and nutritional status compared to those from the geriatric hospital department. The inclusion criteria for participation of both groups were patients of 65 years or older who were expected to be hospitalized for at least 14 days, ability to cooperate and written informed consent. Exclusion criteria were immobility for more than 3 weeks prior to admission, edema, leg amputation, decompensated heart failure and expected change in diuretic dose during hospitalization and metallic implants. Functional status, body weight and mid-thigh MRI measurements were conducted within 24 hours after hospital admission (baseline) and before discharge (follow-up) in the immobile and mobile groups. In addition, geriatric assessment was performed at hospital admission except the Barthel-Index and the muscle strength measurement, which were evaluated on admission and at discharge. The attending physician recorded the clinical routine data. C-Reactive Protein (CRP) was analyzed according to standard clinical procedures at hospital admission All research related data were obtained and recorded by the first author. The study protocol had been approved by the ethical committee of Ruhr-University Bochum (17-6048, approved on 08.08.17).
Nutritional status was evaluated using the Mini Nutritional Assessment Short Form (MNA-SF)  and subjects were categorized as having normal nutritional status (12-14 points) or a risk of malnutrition (8-11 points) and malnourished (0-7 points). Self-caring activities was determined using Barthel-Index (BI) . The point’s range of the German version of the BI is 0-100 pts., with 100 pts. indicating independency in all activities of daily living. Evaluation of frailty was based on FRAIL simple scale  and a score of 0 is considered not frail whereas scores of 1-2 and 3-5 are considered as pre-frail and frail, respectively. The risk of sarcopenia was investigated with the use of SARC-F questionnaire  ranges from 0 to 10 and subjects with score ≥4 were defined as having probable sarcopenia. Medical comorbidities were evaluated using Charlson Comorbidity Index (CCI) .
All patients received the similar nutrition except oral nutrition supplement which was only provided for patients with MNA-SF <8. All patients had physical training for at least 30 minutes twice a day as a routine rehabilitation program. However, immobile patients who were more or less bedridden participated less. Furthermore, individualized training program was provided to all patients according to the deficient in activity of daily living.
Body weight was assessed in light clothing with an accuracy of 0.1 kg and height was measured to the nearest 0.5 cm with a stadiometer during hospitalization. The degree of unintentional weight loss was obtained either by interviewing the patients, if competent, or asking their proxy, where necessary.
Handgrip strength (HGS) was assessed using a Jamar dynamometer (Lafayette Instrument Company, Lafayette, IN). Isometric knee extension strength was measured according to the protocol described by Gandevia . Briefly, knee strength was assessed with the patient in a seated position with a strap around the leg 10 cm above the ankle joint whereas the hip and knee joint angles positioned at 90 degrees (Figure 1a). Handgrip and knee strengths were measured three times at the dominant or unaffected side of hand/leg and the maximum score was recorded.
Mid-thigh MRI cross sectional area
The MRI scan was performed with a Siemens Magnetom Sonata, 1,5 Tesla (Siemens Medical Solutions, Erlangen, Germany) to assess the mid-thigh muscle, subcutaneous and intermuscular fat cross-sectional area (CSA). Briefly, middle length of the non-affected, preferably dominant thigh was measured and marked with a semi-circumferential line, drawn with a permanent marker for replication at the time of follow-up. Directly before the MRI measurement, two MRI-detectable capsules were fixed at both ends of the line (Figure 1b). Five T1-weighted transversal scans with slices thickness of 4 mm were obtained and the single-slice with both landmarks was selected for segmentation. The following imaging parameters were used: 447 ms repetition time (TR), 13 ms echo time (TE), 160° flip angle (FA) and 400 x 300 mm2 field of view (FOV) with a resolution of 1.2 x 0.8 mm2. The field of view and the settings were kept constant throughout all measurements. Single slice CSA of muscle, subcutaneous and intermuscular fat at mid-thigh were manually segmented (Figure 1b) using the SliceOmatic software (version 5.0; Tomovision, Montreal, Canada). The segmentation of MRI images was blinded for group assignment of subjects.
The statistical analysis was performed using SPSS statistical software (SPSS Statistics for Windows, IBM Corp, Version 24.0, Armonk, NY, USA). With the expected 0% change in muscle mass of mobile patients and 2.7% loss of muscle mass in immobile patients with a realistic high standard deviation of ± 3%, a case number of N = 40 in a 1: 1 design with a power of 0.8 and a Type I error of 0.05 is calculated (http://PowerAndSampleSize.com). Continuous variables are reported by means and standard deviations (SDs) for normally distributed variables and median values with interquartile ranges (IQR) for non-normally distributed data. Categorical variables are shown as n (%). Differences between variables and between baseline and follow-up within each group (immobile and mobile groups) were analyzed by using paired samples t test for normally distributed values. Differences in variables at baseline and in magnitude of changes at follow-up between groups were analyzed by using an unpaired t test in normally distributed variables and the Mann-Whitney U test for continuous variables with non-normal distribution. Categorical variables were compared by the Chi square test. A stepwise multiple regression analysis was performed to test the impact of risk factors such as mobility status, age, BMI, malnutrition and inflammation (as independent variables) on changes in mid-thigh muscle CSA as a percentage of initial muscle area as (dependent variable). P < 0.05 was determined as the limit of significance.