Analysis of sternal healing after median sternotomy in low risk patients at midterm follow-up: retrospective cohort study from two centres

BACKGROUND: For low risk patients undergoing median sternotomies, no midterm follow-up studies involving sternal healing have been conducted. In this study we evaluated sternal healing in low risk patients by chest CT scan and the risk factors associated with poor healing were analyzed. METHODS: Patients who underwent sternal median incision heart surgery from September 2014 to March 2015 were recruited. The clinical information of these patients during hospitalization was collected, and the CT scan data were submitted to the two chief physicians of the Radiology Department for radiographical sternal healing score determination. Based on the method of wound closure, the patients were divided into sternum plate (Plates) and wire groups (Wires). RESULTS: Forty-four patients were recruited. The mean CT examination time was 17.27 ± 2.30 months postoperatively. Twenty-nine (65.9%) patients met the criteria for radiographic sternal healing. Three segments, including the aortopulmonary window, the main pulmonary artery, and the aortic root, had healed less in comparison to the manubrium segment. Compared to patients in whom 6–7 metal wires were used for sternal closure, healing of the lower sternum was worse in patients in whom five wires were used, but the difference in healing was not statistically significant. Univariate analysis of sternal healing showed that patient age was a risk factor for sternal non-healing. When the patient age was > 45 years, the predicted risk of radiographic sternal non-union was 1.833 (95% CI: 1.343–2.503). CONCLUSIONS: At the mid-term follow-up, 65.9% of patients undergoing median sternotomies demonstrated radiographic sternal healing. Age, but not closure device, was a risk factor for sternal non-healing in low risk patients. Use of more wires had a positive impact on sternal healing. TRIAL REGISTRATION: researchregistry4918, registered 28 May 2019, retrospectively registered.