The need for routine diverting ileostomy following restorative total mesorectal excision (TME) is increasingly debated as the benefits might not outweigh the disadvantages. This study evaluated an institutional shift from routine (RD) to highly selective diversion (HSD) after TME surgery for rectal cancer. Patients having TME with primary anastomosis and HSD for low or mid rectal cancer between December 2014 and March 2017 were compared with a historical control group with RD in the preceding period since January 2011. HSD was introduced in conjunction with uptake of transanal TME. In the RD group, 45/50 patients (90%) had a primary diverting stoma, and 3/40 patients (8%) in the HSD group. Anastomotic leakage occurred in 10 (20%) and three (8%) cases after a median follow-up of 36 and 19 months after RD and HSD, respectively. There was no postoperative mortality. An unintentional stoma beyond 1 year postoperative was present in six and two patients, respectively. One-year stoma-related readmission and reoperation rate (including reversal) after RD were 84% and 86%, respectively. Corresponding percentages were significantly lower after HSD (17% and 17%; P < 0.001). Total hospital stay within one year was median 11 days (IQR 8–19) versus 5 days (IQR 4–11), respectively (P < 0.001). This single institutional comparative cohort study shows that highly selective defunctioning of a low anastomosis in rectal cancer patients did not adversely affect incidence or consequences of anastomotic leakage with a substantial decrease in 1-year readmission and reintervention rate, leading to an overall significantly reduced hospital stay.