The primary end point was reached in 7.1%, 31.3%, and 82.4% of the patients in the air, carbon dioxide, and water insufflation groups, respectively (
P<0.001), water being significantly better than air (
P<0.001) and carbon dioxide (
P=0.003). Mean pain scores were 5.28, 5.12, and 3.64 for air, carbon dioxide, and water insufflation groups, respectively (
P<0.001), with water being significantly better than both air (
P<0.001) and carbon dioxide (
P<0.001). Sedation was required in 28.6%, 25.0%, and 11.8% of the patients in the air, carbon dioxide, and water insufflation groups, respectively (
P=0.470). At the rectosigmoid junction, the VAS score were 6.00, 6.00, and 4.18 for the air, carbon dioxide, and water insufflation groups, respectively (
P<0.001), with water being significantly better than both air (
P=0.002) and carbon dioxide (
P=0.001). At the splenic flexure, the VAS scores were 7.29, 6.81, and 4.63 for air, carbon dioxide, and water insufflation groups, respectively (
P<0.001), with water being significantly better than both air (
P<0.001) and carbon dioxide (
P=0.001). At the hepatic flexure, the VAS scores were 3.60, 2.73, and 2.43 for air, carbon dioxide, and water insufflation groups, respectively (
P=0.090). During withdrawal, the VAS scores were 1.90, 1.20, and 1.73 for air, carbon dioxide, and water insufflation groups, respectively (
P=0.080). Pressure was required in 92.8%, 68.8%, and 35.3% of the patients in the air, carbon dioxide, and water insufflation groups, respectively (
P=0.010), with significantly less pressure required with water as compared to air (
P=0.003). Change of position was required in 28.6%, 43.8%, and 29.4% of the patients in the air, carbon dioxide, and water insufflation groups, respectively (
P=0.600) (
Fig. 2,
Supplementary Table 3).