Abstract Background The reoperation rate after recurrent groin hernia surgery is more than twice that recorded for primary groin hernia procedures. The aim was to define the outcome from routine redo hernia surgery by analysing a large population-based cohort from a national hernia register. Methods All recurrent groin hernia operations registered in the Swedish Hernia Register from 1992 to 2008 were analysed using multivariable analysis with stratification for preceding repair. Results Altogether 174 527 hernia operations were recorded in the Swedish Hernia Register between 1992 and 2008, including 19 582 reoperations. The preceding repair was included in the register for 5565 of these recurrent repairs. With laparoscopic repair as reference standard, the hazard ratio for recurrence was 2·55 (95 per cent confidence interval 1·66 to 3·93) after sutured repair, 1·53 (1·20 to 1·95) after Lichtenstein repair, 2·31 (1·76 to 3·03) after plug repair, 1·36 (0·95 to 1·94) after open preperitoneal mesh and 3·08 (2·22 to 4·29) after other repairs. Laparoscopic and open preperitoneal repair were associated with a lower risk of reoperation following a preceding open repair (P < 0·001), but no technique differed significantly from the others following a preceding preperitoneal repair. Conclusion The laparoscopic and the open preperitoneal mesh methods of repair for recurrent groin hernias were associated with the lowest risk of reoperation. Although the method of repair in previous surgery must be considered, these techniques are the preferred methods for recurrent groin hernia surgery.