The present study used a double-blind, placebo-controlled, parallel-group, randomized design. After acceptance into the study, participants were randomly allocated to either of two supplemented groups. All participants then started a 6-month exercise and supplement intervention. Before and after the intervention, body composition assessment, physical performance tests and blood sampling were performed. The trial was conducted at Obihiro University of Agriculture and Veterinary.
Medicine (Hokkaido, Japan) between February and November 2018. The data were analyzed between December 2018 and March 2019. The target number of participants was set at 128 (i.e., 64 participants per group) to detect a difference of 0.5 kg in the lean body mass change , assuming a standard deviation of 1.0 kg , to attain 80% statistical power using a two-sided α of 0.05.
Participants were recruited from senior citizens’ clubs in Obihiro, Hokkaido, Japan. All participants completed health history and physical activity questionnaires and met the following inclusion criteria: 60 years or older and physically independent. Individuals were excluded if they (1) had current or previous histories of significant liver, cardiovascular, pulmonary, renal or digestive diseases; (2) had significant orthopedic injuries; (3) had food allergies; (4) restricted protein intake due to medical reasons; (5) were involved in regular resistance training (> 2 times/week); (6) were participating in other clinical studies; (7) or were judged as inappropriate for the study by the principal investigator due to abnormal blood pressure or parameters, or other reasons.
Randomization and blinding
Stratification based on sex and age was performed after eligibility assessment for study participation. Participants were divided into four strata: (1) males less than 70 years old; (2) females less than 70 years old; (3) males 70 years or older; (4) females 70 years or older. In each strata, computer-generated random numbers were assigned to the participants who were then sorted and divided into two equal groups. The groups were randomly assigned to either the acidified milk (MILK) group or placebo (PLA) group, by an individual who was accountable for preparing the test drinks but was not involved in the plan, enrollment, evaluation, intervention, or analysis. The participants, investigators, and all staff members involved in the trial were blinded to group allocation. The randomization code was opened after the study data were checked, collated, and finalized.
Participants ingested a 200 mL polyethylene terephthalate bottle drink daily of either an acidified milk protein drink  or a placebo drink for 6 months. The acidified milk protein drink was made from milk protein concentrate, trehalose, soybean polysaccharide, pectin, fermented cellulose, citric acid, malic acid and food flavors, and contained 7.0 g of carbohydrate, 10.1 g of protein and 0.2 g of fat, providing 68 kcal of energy per bottle. The amino acid profile of the acidified milk protein drink is shown in Table 1. The placebo drink was made from maltodextrin, trehalose, soybean polysaccharide, pectin, fermented cellulose, citric acid, malic acid and food flavors, and contained 16.0 g of carbohydrate, 0.1 g of protein and 0.5 g of fat, providing 68 kcal of energy per bottle. The test drinks had an identical pH (4.2) and appearance. The participants were instructed to keep the test drinks in their refrigerator and consume one bottle per day within an hour after each exercise training session. If the participants did not perform exercise training for some reason, they were instructed to consume a test drink anytime they wanted.
Exercise training program
The exercise training program is shown in Table 2. All participants were instructed to perform an exercise training program daily for 6 months. The exercise training program was composed of 6 body weight exercises and 5 medicine ball exercises. The criteria for using low-to-moderate intensity exercise training  was that the participants were able to repeat the exercises at least 12 times (< 70% of 1-repetition maximum ). Part of the program was changed every 2 months to maintain participants’ motivation. Participants followed a monthly exercise training lesson throughout the 6-month intervention period (seven lessons in total).
For the medicine ball exercises, participants used one or two 1 kg soft medicine balls.
Release and catch: Participants stood on the floor holding a medicine ball in front of their body in one hand. They released the ball and immediately caught the ball in their opposite hand with the palm facing down. They repeated the motion a given number of times (Table 2).
Biceps curl: Participants stood on the floor holding a medicine ball in each hand with their arms hanging by their sides. They performed a biceps curl a given number of times (Table 2).
Upper body rotation: Participants stood on the floor holding a medicine ball in front of their body in each hand. Then they rotated their upper body once with outstretched arms. They repeated the motion 10 times alternating from clockwise to counterclockwise.
Leg flexion and extension: Participants sat on the floor and put a medicine ball between their feet with their knees bent and their feet flat on the floor. They extended and bent their legs without dropping the ball. They repeated the motion 10 times.
Participants counted and recorded the number of steps daily using a triaxial accelerometer (FB-736, TANITA corporation, Tokyo, Japan) during the intervention period.
Body composition assessment, physical performance tests and blood sampling were performed at baseline and after 6 months of intervention by experienced staff members who were blinded to group allocation. On the day before the measurement, participants were barred from drinking alcohol and had dinner between 6 and 10 PM. After the dinner, participants were allowed to drink only water until the measurements started.
Body composition assessments
A direct segmental multifrequency (5 kHz, 50 kHz and 250 kHz) bioelectrical impedance analysis (DSM-BIA) device using an 8-point tactile electrode system (InBody 430, Biospace, Seoul, Korea) [25, 26] was used to measure body weight, lean body mass and fat mass. Participants were measured wearing the same shirt and pair of shorts, which weighed 0.6 kg, so the weight adjustment for clothing was set to 0.6 kg. Body composition assessments were performed between 8.30 and 11:30 AM in fasted states before physical performance tests.
Physical performance tests
On the measurement days, participants practiced each test twice with submaximal efforts to minimize the learning effect before making the actual test. They practiced the push-up motion twice.
Grip strength: A digital handgrip dynamometer (Grip-D, Takei Scientific Instruments Co. Ltd., Niigata, Japan) was used to measure grip strength. Subjects were allowed to adjust the grip on the apparatus, and then performed a maximum force grip with the right hand (isometric exercise) while the left arm was hanging free by the side. The test was carried out twice at an interval of 1 min and the best result was recorded as the grip strength.
Maximal walking speed: Participants walked on a flat, straight, 7-m-long walkway two times at their maximum speed. The time to walk 5-m was measured using diffuse-reflective photoelectric sensors (Yagami Inc., Aichi, Japan) which were placed at the 1-m and 6-m points. For maximum walking speed, the faster time recorded was used and converted into speed (m/s).
Knee extension and flexion strength: Knee extension and flexion strength were evaluated using an isokinetic dynamometer (Biodex System 4, Biodex Medical Systems, Inc. NY, USA) at an angular speed of 60°/s. Full knee flexion (start-position) was set to 100° and full knee extension was set to 30°. Following a familiarizing practice trial, participants extended and then flexed their right knees with full strength while crossing their arms in front of their chests. The test was repeated 3 times at an interval of 30 s and the best result was recorded as the knee extension and flexion strength.
Timed Up and Go test: The Timed Up and Go test measures speed during several functional maneuvers, which include standing up, walking, and turning and sitting down. The test is a reliable and valid test for quantifying functional mobility in older adults . Participants were seated in a normal chair (41 cm high) with their back against the chair and their hands on their thighs. They were instructed to stand up, walk 3-m as quickly and safely as possible, past a cone on the floor, turn around, walk back to the chair, and sit down with back against the chair again . A stop-watch was used to time one test. The test was carried out twice and the best time was recorded.
Sit-to-stand test: The sit-to-stand is often used as a measure of lower-limb strength in older adults . The test measures the time taken to stand from a seated position either one, three, five or 10 times. In this study, participants were seated in a normal chair (41 cm high) and rose from the chair five times as fast as possible while crossing their arms in front of their chests. A stopwatch was used to time the test. The test was carried out twice and the best time was recorded.
Push-up: Participants fully extended their elbows and placed hands on the exercise mat directly under the shoulders, with knees together also touching the mat and trunk and thighs fully stretched. From this position, the elbows were flexed until just touching the mat with their chest and then immediately extended again. They repeated the push-up as many times as possible, and the number of repetitions was recorded.
A blood sample was drawn from each participant in a fasted state at baseline. After 6 months of intervention, blood sampling was performed twice; the first sampling was performed in a fasted state, and the second sampling was performed 30 min after ingestion of 200 mL of the test drink (either an acidified milk or a carbohydrate drink). Immediately before ingestion of the test drink, participants performed the exercise training program (4–6 month version as described in Table 2). The second blood sampling was performed after body composition assessments and before physical performance tests. Plasma leucine was measured using liquid chromatography tandem-mass spectrometry (ACQUITY TQD, Waters Corporation, MA, USA). Other blood parameters were assayed by an independent laboratory (BML, Inc., Tokyo, Japan). Plasma albumin was measured using a bromocresol green method (Clinimate ALB, Sekisui Medical Co., Ltd., Tokyo, Japan). Plasma creatinine (Sikarikid-S CRE, Kanto Chemical Co., Inc., Tokyo Japan), uric acid (Pureauto S UA, Sekisui Medical Co., Ltd., Tokyo, Japan), low-density lipoprotein cholesterol (LDL-C) (Cholestest LDL, Sekisui Medical Co., Ltd., Tokyo, Japan), high-density lipoprotein cholesterol (HDL-C) (Cholestest N HDL, Sekisui Medical Co., Ltd., Tokyo, Japan) and triglycerides (Pureauto S TG-N, Sekisui Medical Co., Ltd., Tokyo, Japan) were measured using enzymatic methods. Plasma glucose was measured using a hexokinase assay (Pureauto S GLU, Sekisui Medical Co., Ltd., Tokyo, Japan). Plasma insulin was measured using a chemiluminescence immunoassay (Chemilumi Insulin, Siemens Healthcare Diagnostics, Inc., Tokyo, Japan). Plasma insulin-like growth factor-1 (IGF-1) was measured using an electro chemiluminescence immunoassay (Elecsys IGF-1, Roche Diagnostics K.K., Tokyo, Japan).
Before and during the 3rd and 6th month of the intervention period, participants recorded total food consumption for 3 consecutive days to determine their daily macronutrient intake exclusive of the test drinks. Food quantities were measured by using standard measuring glasses, spoons, and digital scales. All dietary data were analyzed by an independent laboratory (THF Co., Ltd., Ibaraki, Japan) using Excel Eiyo-kun, version 7.0 (Kenpakusha Co., Ltd., Tokyo, Japan).
All statistical analyses were performed by an independent organization (edihas K.K., Hokkaido, Japan) using IBM SPSS Statistics 24 (IBM Japan, Ltd., Tokyo, Japan). Analyses were done on the full analysis set (FAS) with missing values imputed by the last observation carried forward. All values are expressed as mean ± standard error of the mean (SEM). The Student t-test for independent samples was used to compare differences between groups in participant characteristics at baseline, compliance rate for exercise training and consumption of the test drinks, daily average of the number of steps, and changes of values during the intervention period. A two-factor, repeated-measures analysis of variance (ANOVA) with time as the within-participants factor and group as the between-participants factor were carried out for dietary intake, body composition, physical performance, and blood parameters. When the F-ratio was significant, the Bonferroni post hoc test was employed to identify mean differences. For all statistical analyses, significance was set at p < 0.05.