Objective:To explore the effects of bariatric metabolic surgery on body composition.Methods:The retrospective cohort study was conducted. The clinicopathological data of 66 patients with metabolic diseases who were admitted to the Third Xiangya Hospital of Central South University from January 2013 to December 2014 were collected. There were 42 males and 24 females, aged (40±11)years, with a range from 17 to 63 years. Of the 66 patients, 27 undergoing laparoscopic sleeve gastrectomy (LSG) and 39 undergoing laparoscopic Roux-en-Y gastric bypass (LRYGB) were allocated into LSG group and LRYGB group, respectively. The body composition of all patients was determined by dual-energy X-ray absorptiometry at preoperation and postoperative 6 months. Observation indicators: (1) the changes of anthropometric parameters, glucolipid metabolism, body fat mass percentage (BF%) and the ratio of Android BF% and Gynoid BF% (A/G ratio) from preoperation to postoperative 6 months; (2) the changes of whole and local body composition from preoperation to postoperative 6 months; (3) analysis of the correlation between BF% and anthropometric parameters, glucolipid metabolism. (4) Follow-up. Follow-up was conducted using outpatient or hospitalization examination to detect the changes of body composition at the time of postoperative 6 month. The follow-up time was up to July 2015. Measurement data with normal distribution were represented as Mean± SD, paired-samples t test was used for intra-group comparison, and independent-samples t test when baseline data were consistency or covariance analysis when baseline data were not consistency was used for inter-group comparison. Measurement data with skewed distribution were represented as M ( P25, P75), and comparison between groups was analyzed using Wilcoxon signed rank test. The correlation test was undertaken with the Pearson bivariate analysis. Results:(1) The changes of anthropometric parameters, glucolipid metabolism, BF% and A/G ratio from preoperation to postoperative 6 months: for patients in the LSG group, the body mass, body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), diastolic blood pressure (DBP), systolic blood pressure (SBP), fasting plasma glucose (FPG), HbA1c, high density lipoprotein cholesterol (HDL-C), triglyceride (TG), whole BF%, arms BF%, legs BF%, trunk BF%, Android BF%, Gynoid BF% and A/G ratio at preoperation and postoperative 6 months were (102±17)kg, (37±5)kg/m 2, (118±14)cm, 1.01±0.06, (94±14)mmHg(1 mmHg=0.133 kPa), (137±15)mmHg, (8.1±4.2)mmol/L, 7.3%±2.4%, (1.11±0.26)mmol/L, 2.14 mmol/L(1.73 mmol/L, 2.59 mmol/L), 40%±6%, 46%±10%, 36%±8%, 42%±6%, 45%±6%, 37%±7%, 1.23±0.18 and (82±15)kg, (29±4)kg/m 2, (101±13)cm, 0.95±0.08, (76±10)mmHg, (118±16)mmHg, (7.2±1.2)mmol/L, 5.4%±0.8%, (1.26±0.32)mmol/L, 1.21 mmol/L(0.88 mmol/L, 1.55 mmol/L), 36%±8%, 41%±9%, 34%±10%, 38%±8%, 41%±8%, 35%±10%, 1.20±0.17, respectively. There was no significant difference in the intra-group comparison of the Gynoid BF% and A/G ratio ( t=1.903, 1.730, P>0.05) and there were significant differences in the intra-group comparison of the rest of above indicators ( t=12.748, 13.283, 9.013, 3.804, 6.031, 6.226, 2.393, 4.287, -2.900, 3.193, 2.932, 5.198, 2.167, 3.357, 3.116, P<0.05). For patients in the LRYGB group, the body mass, BMI, WC, WHR, DBP, SBP, FPG, HbA1c, HDL-C, TG, whole BF%, arms BF%, legs BF%, trunk BF%, Android BF%, Gynoid BF% and A/G ratio at preoperation and postoperative 6 months were (80±12)kg, (28±4)kg/m 2, (98±9)cm, 0.96±0.05, (85±10)mmHg, (134±17)mmHg, (8.6±2.8)mmol/L, 8.3%±1.7%, (1.13±0.26)mmol/L, 2.06 mmol/L(1.15 mmol/L, 3.30 mmol/L), 30%±8%, 29%±11%, 23%±9%, 37%±7%, 40%±7%, 29%±8%, 1.42±0.26 and (69±9)kg, (24±3)kg/m 2, (91±8)cm, 0.93±0.05, (80±9)mmHg, (129±18)mmHg, (7.4±1.8)mmol/L, 7.0%±1.5%, (1.18±0.29)mmol/L, 1.29 mmol/L(0.85 mmol/L, 2.02 mmol/L), 25%±8%, 23%±12%, 20%±9%, 29%±9%, 32%±10%, 25%±9%, 1.29±0.25, respectively. There was no significant difference in the intra-group comparison of the SBP and HDL-C ( t=1.733, -1.073, P>0.05) and there were significant differences in the intra-group comparison of the rest of above indicators ( t=10.525, 10.200, 7.129, 2.887, 2.805, 2.517, 3.699, 2.608, 7.997, 8.018, 6.029, 8.342, 8.069, 5.813, 6.391, P<0.05). There were significant differences in DBP, SBP, HbA1c, trunk BF%, Android BF% and A/G ratio at postoperative 6 months between LSG group and LRYGB group ( F=6.408, t=2.641, F=20.673, 5.140, 5.735, 4.714, P<0.05). (2) The changes of whole and local body composition from preoperation to postoperative 6 months: for patients in the LSG group, the whole fat mass, muscle mass, fat-free mass at preoperation and postoperative 6 months were (38.74±9.68)kg, (57.71±11.62)kg, (60.14±11.95)kg and (26.64±8.29)kg, (48.65±13.80)kg, (51.00±14.27)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=5.256, 5.413, 5.315, P<0.05); the arms fat mass, muscle mass, fat-free mass were (5.19±1.67)kg, (5.78±1.58)kg, (6.10±1.64)kg and (3.73±1.19)kg, (5.10±1.53)kg, (5.43±1.57)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=7.564, 5.405, 5.363, P<0.05); the legs muscle mass and fat-free mass were (19.05±4.19)kg, (19.93±4.35)kg and (15.93±4.71)kg, (16.81±4.87)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=5.623, 5.568, P<0.05); the trunk fat mass and fat-free mass were (21.93±4.90)kg, (29.7±5.94)kg and (14.69±4.79)kg, (24.78±7.02)kg respectively, showing significant differences in the intra-group comparison of the above indicators ( t=8.903, 5.421, P<0.05); the Android fat mass and fat-free mass were (4.16±1.19)kg, (5.01±1.12)kg and (2.57±0.90)kg, (3.83±1.20)kg respectively, showing significant differences in the intra-group comparison of the above indicators ( t=8.288, 7.637, P<0.05); the Gynoid fat mass and fat-free mass were (5.51±1.42)kg, (9.27±1.86)kg and (3.85±1.16)kg, (7.65±2.31)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=7.461, 5.672, P<0.05); the skeletal muscle index were (8.86±1.38)kg/m 2 and (7.49±1.71)kg/m 2, respectively, showing a significant differences in the intra-group comparison ( t=5.724, P<0.05). For patients in the LRYGB group, the whole fat mass, muscle mass, bone mineral content, fat-free mass at preoperation and postoperative 6 months were (23.58±7.80)kg, (51.76±8.35)kg, (2.55±0.48)kg, (54.31±8.63)kg and (16.88±6.86)kg, (49.41±7.70)kg, (2.47±0.50)kg, (51.88±8.05)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=9.001, 3.974, 4.354, 4.075, P<0.05); the arms fat mass were (2.72±2.37)kg and (1.73±1.02)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=3.470, P<0.05); the legs fat mass, muscle mass, fat-free mass were (5.21±2.46)kg, (16.68±3.50)kg, (17.60±3.66)kg and (4.01±2.12)kg, (15.63±2.90)kg, (16.54±3.05)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=6.592, 3.372, 3.319, P<0.05); the trunk fat mass were (14.87±4.11)kg and (10.38±4.00)kg, respectively, showing a significant difference in the intra-group comparison of the above indicators ( t=8.431, P<0.05); the Android fat mass and fat-free mass were (2.61±0.86)kg, (3.96±0.87)kg and (1.81±0.79)kg, (3.78±0.67)kg respectively, showing significant differences in the intra-group comparison of the above indicators ( t=8.032, 2.153, P<0.05); the Gynoid fat mass and fat-free mass were (3.14±1.17)kg, (7.89±1.58)kg and (2.44±0.96)kg, (7.43±1.26)kg, respectively, showing significant differences in the intra-group comparison of the above indicators ( t=6.112, 3.207, P<0.05); the skeletal muscle index were (8.04±1.22)kg/m 2 and (7.43±1.13)kg/m 2, respectively, showing significant differences in the intra-group comparison ( t=4.953, P<0.05). There were significant differences in whole muscle mass, whole fat-free mass, arms fat mass, legs muscle mass, legs fat-free mass, trunk fat-free mass, Android fat-free mass, Gynoid fat-free mass and skeletal muscle index at postoperative 6 months between LSG group and LRYGB group ( F=13.846, 13.614, 23.696, 7.100, 7.127, 15.243, 16.921, 8.625, 5.497, P<0.05). (3) Analysis of the correlation between BF% and anthropometric parameters, glucolipid metabolism: the whole BF% of 66 patients was positively correlated with body mass, BMI, WC and WHR ( r=0.405, 0.663, 0.625, 0.331, P<0.05); the arms BF% was positively correlated with body mass, BMI, WC and WHR ( r=0.432, 0.682, 0.639, 0.309, P<0.05); the legs BF% was positively correlated with body mass, BMI and WC ( r=0.366, 0.646, 0.564, P<0.05); the trunk BF% was positively correlated with body mass, BMI, WC and WHR ( r=0.332, 0.560, 0.554, 0.335, P<0.05); the Android BF% was positively correlated with body mass, BMI, WC and WHR ( r=0.327, 0.537, 0.543, 0.336, P<0.05); the Gynoid BF% was positively correlated with BMI and WC ( r=0.561, 0.488, P<0.05), and negatively correlated with FPG ( r=-0.491, P<0.05); the A/G ratio was negatively correlated with BMI ( r=-0.334, P<0.05), and positively correlated with FPG ( r=0.506, P<0.05); the skeletal muscle index was positively correlated with body mass, BMI, WC and WHR ( r=0.757, 0.641, 0.609, 0.519, P<0.05), and negatively correlated with HDL-C ( r=-0.369, P<0.05). (4) Follow-up: 66 patients were followed up at the time of postoperative 6 month. Conclusions:Both LSG and LRYGB significantly change body composition. LRYGB is superior to LSG in reducing trunk BF% and Android BF%. The effects of the two surgical methods on fat mass and bone mineral content are similar. LSG lead to a more significant decrease in whole muscle mass, and LRYGB lead to a more significant decrease in legs muscle mass and skeletal muscle index.