Objective To assess the efficacy and summarize the treatment experience in correction of recurrent pectus excavatum by the non-thoracoscopic minimally invasive Nuss procedure. Methods From July 2003 to November 2007, 21 patients with recurrent pectus excavatum were treated with surgical repairs. Among them, there were 15 patients who underwent a Nuss procedure (the recurrent Nuss group), 13 males and 2 females, aged (13.31 ± 4.21) years old, and the preoperative CT scan showed the mean Haller index was 3.98 ± 0.94. The recurrence time of the depression after the primary operation was different: 2 patients with less than 1 year, 4 with 1-3 years, 7 with 3-5 years and 2 with over 5 years. There were 6 patients who underwent a modified Ravitch procedure (the recurrent Ravitch group), 5 males and 1 female, aged (13.67 ± 2.23) years old, and the mean Haller index was 3.92 ± 1.01. The recurrence time of the depression after the primary operation was as follows: 2 patients with less than 1 year, 1 with 1-3 years and 3 with 3-5 years. Another 119 patients had a primary Nuss procedure in the same period (the primary Nuss group), 95 males and 24 females, aged (7.79 ± 3.59) years old, and the mean Haller index was 4.61 ± 1.36. In terms of the age and the Haller index, there was significant difference between the recurrent Nuss group as well as the recurrent Ravitch group and the primary Nuss group (Plt; 0.05), but there was no significant difference between the recurrent Nuss group and the recurrent Ravitch group (Pgt; 0.05). Results All the patients were performed with the surgery successfully without any severe compl ications. The operation time, blood loss and postoperative hospital ization time of the recurrent Nuss group and the primary Nuss group were significantly less than those of the recurrent Ravitch group (P lt; 0.05), but there was no significant difference between the recurrent Nuss group and the primary Nuss group (P gt; 0.05). All the patients were followed up for 1.5-37.0 months (11.2 months on average). Except for 1 with plate displacement and 2 with suture reaction in the primary Nuss group who were readmitted to hospital, the other patients recovered well. The patients had perfect thoracic appearances with increased exercise tolerance and had a significantly improved psychological condition and an active social participation because of their improved appearances. The postoperative mean Haller index was 2.58 ± 0.31, 2.77 ± 0.48 and 2.52 ± 0.34 in the recurrent Nuss group, the recurrent Ravitch group and the primary Nuss group, respectively, with significant difference compared with the preoperative mean Haller index in each group (Plt; 0.05). Conclusion Although the recurrent pectus excavatum repairs are technically more challenging than the primary surgery, the non-thoracoscopic Nuss procedure for reoperative correction is a safe, effective and rel iable method.