Subjects (n=20) were recruited from among those admitted to a facility for visually impaired people; 16 were men and 4 women; mean age was 25.95±10.98 years (Table 1). All subjects could walk independently under supervision. All subjects had no previous history of regular exercise that included aerobic and anaerobic elements. We excluded people who had disability or injury of upper or lower extremities, history of convulsion, mental retardation sufficiently severe to preclude learning to row, or cardiopulmonary insufficiency. Blood laboratory data, body composition, physical fitness test, whole spine X-ray, and Tetrax posturography were checked before and after 6 weeks of rowing exercise training. All subjects were asked to maintain routine nutrition and lifestyle during training. This study was approved by the Clinical Trial Ethics Committee of the Hallym University of College of Medicine. All patients provided written informed consent.

Blood sampling was conducted to gather laboratory data before and after 6-week rowing exercise training. Laboratory data included hemoglobin, protein, albumin, total cholesterol, triglyceride (TG), high-density lipoprotein (HDL) cholesterol, and LDL cholesterol.

Dual-energy X-ray absorptiometry (Lunar Prodigy, GE, Madison, WI, USA) was used to measure body composition before and after 6-week rowing exercise training. Body composition was comprised of android/gynoid ratio, fat mass, lean mass, bone marrow content, total body fat percent, body mass index (BMI), and body weight.

The physical fitness test was composed of grip strength, back strength, trunk flexion, trunk extension, and vertical jump using physical fitness measurement equipment (Takei Scientific Instruments Co. Ltd., Niigata, Japan). Grip strength (kg) and back strength (kg) were measured to evaluate static strengthening. The flexibility item was trunk flexion (cm) and trunk extension (cm). The swiftness item was vertical jump (cm). Results of the physical fitness test were converted to the percentage of age/sex-matched normal data. Sources for grip and back strength, trunk extension and vertical jump are "Standard Values for Physical Strength of the Japanese 2000" by the Standard Physical Strength Survey Team, Tokyo Metropolitan University; source for trunk flexion is "The Survey of National Physical Fitness 2011" by the Korea Institute of Sport Science, Korea Sports Promotion Foundation.

For testing back strength, the subject was asked to stand upright on the base of the dynamometer with feet shoulder-width apart, arms straight, and fingers extended downward as far as possible on the front of thighs. The bar is then attached to the chain so that it is 1 to 2 inches below fingertips. Then the subject bends forward slightly and grasps the bar. The subject should hold his or her head upright and look straight ahead. Lifting the bar steadily, the subject is required to keep legs straight and feet flat on the base of the dynamometer.

For testing trunk flexion, the subject sits with knees fully extended so that the bottoms of the feet are against the surface of the flex box. Arms are extended forward, with one hand placed on the top of the other. With the instructor holding the subject's knees straight, the subject steadily reaches as far forward as possible and maintains this position for 3 seconds.

Trunk extension measurement requires 2 people in addition to the subject: one measurer and one assistant. The subject lies face down on the floor, clasps hands behind the waist and holds the legs apart so that the toes are about 45 cm from each other. The assistant holds on the back of the thighs to fix the position. The subject raises his or her chin as high as possible by gradually bending upper body backward. The measurer measures the distance of the highest position of the chin from the floor.

For the grip strength measurement, the subject stands upright and squeezes the dynamometer as tightly as possible.

For testing vertical jump, the subject puts on the belt around his or her waist tightly so as not to be shaken provide the main body of Jump-MD. Then, stand on the center of the rubber plate and jump straight up.

All tests were performed twice and the greater values of the 2 results for each measurement were selected.

Using Tetra-ataxiametric posturography (Tetrax, Sunlight Medical Ltd., Tel-Aviv, Israel), the fall index was checked before and after 6 weeks of rowing exercise training.

Whole spine X-rays were taken to measure Cobb's angle before and after 6-week rowing exercise training.

Concept2 rowing model E (Concept2 Inc., Morrisville, VT, USA) was used for 6 weeks of rowing training (Fig. 1). The rowing machine was set up 5 days a week; each session included a 10-minute warm-up, a 20-minute rowing exercise, and a 10-minute cooling-down period. The subjects were given instructions on how to use the rowing machine prior to the exercise program. Each subject was asked to row for 1 minute with maximal effort to measure exercise intensity. Individual's maximal exercise capacity revealing maximum muscular strength for 1 minute was regarded as 1 repetition maximum (RM). Exercise intensity was 40%-50% of 1 RM during the 2-week adaptation period and 50%-60% during next 4 weeks. The subjects rowed at a self-selected stroke rate and resistance level. As all subjects were beginners, the commonly chosen resistant level was from 3 to 5. Concept2 rowing model E has a monitor that provides distance, speed, pace, calories, and watts for each workout, and it was helpful for controlling and supervising each session. The assistant could supervise the entire exercise process including getting on and off the rowing machine. Blood pressure and heart rate were checked before and after the exercise to monitor abnormal cardiopulmonary response.

Data were uploaded into the SPSS ver. 21.0 (IBM SPSS, Armonk, NY, USA). Descriptive statistics were used to describe the subjective characteristics. A paired t-test was performed to compare the health status before and after the exercise. p<0.05 was adopted as the level of statistical significance.

Initially, none of the subjects was anemic; one subject had low protein, and 3 subjects had low albumin levels. Five subjects had LDL-cholesterol levels that were above the upper normal limit.

After the training, the protein and albumin had increased significantly, while the LDL had decreased significantly. Hemoglobin, total cholesterol, and TG levels decreased, while HDL-cholesterol increased; however, the differences were not statistically significant (Table 2).

Using healthy body fat range chart based on National Institutes of Health/World Health Organization guidelines that categorize ranges of body fat percentages, initially, in terms of total body fat percentage, 8 subjects were normal, 1 was overfat, and 11 were obese. According to the criteria determined by the Korean Society for the Study of Obesity, BMI of ≥23 indicates the overweight, and BMI of <18.5 indicates underweight. BMI for the subjects showed that 5 were normal, 5 were underweight, and 10 were obese. Fat mass and total body fat percentage decreased significantly after the training. Lean body mass and bone marrow content increased after the training, though not statistically significantly (Table 3).

Initially, comparing age/sex-matched normal data, the grip strengths were 44.31%±19.08% (right) and 49.45%±23.98% (left), and the back strength was 31.58%±16.43%; while the trunk flexion was 34.92%±71.68%, the trunk extension was 23.71%±15.30%, and the vertical jump was 28.39%±21.51% (Table 4). In addition, the visually impaired subjects seemed to have low physical fitness levels, as compared with the general population, especially in back strength, trunk flexion/extension, and vertical jump.

After training, back strength and trunk flexion increased significantly. Grip strength of right hand, the trunk extension, and the vertical jump also increased, but not significantly.

Based on the initial fall index (risk of fall assessment), 14 subjects were at high risk of falling and 4 were at medium risk, while only 2 were at minimal risk. After 6 weeks of rowing training, 4 subjects had moved to a lower risk group. In all subjects' results, fall index decreased by 4.60±5.12 points, but this result was not statistically significant (Table 4).

The 9 scoliosis patients showed an initial Cobb's angle of 14.20°±5.38°. After the training, this decreased by 1.11°±1.55°, but this result was not statistically significant. Nevertheless, 6 patients showed decreased Cobb's angle after 6-week rowing training.