Wound Dressing and Post-Operative Complications Following Breast Reduction: A Retrospective Study of 250 Patients
Steven L Zeng, Thomas Ren, Ricardo C Rodriguez, Nina E Ringelman, Salman Choudhry, Gedge RossonAbstract
Background
Post-operative wound dressings are routine but understudied in reduction mammoplasty. While an ideal dressing should promote healing, limit infection, and be well-tolerated, practice varies widely without consensus.
Objectives
This study compares the impact of dressings on post-operative outcomes following reduction mammoplasty.
Methods
A single-institution, retrospective review was performed of reduction mammoplasty patients. Data included demographics, operative details, dressing type, and complications. Dressings included Dermabond (DER), Prineo (PRI), Histoacryl (HIS), Sylke (SYL), Steri-Strips (STE), and combinations (COM). Contralateral breasts in oncoplastic or free-flap cases were excluded.
Results
250 patients (439 breasts) were included. At baseline, cohorts differed in BMI, ASA, diabetes, operative time, and resection weight (all p<0.05), largely due to PRI, which contained higher-risk patients. Analyses were repeated, excluding PRI: differences resolved. On univariate analysis, PRI had the highest complication rates. Excluding PRI, HIS had higher skin reaction rates than SYL (p=0.003) and COM (p=0.049), and higher dehiscence than SYL (p=0.043). On multivariate analysis, SYL predicted lower skin reaction (OR 0.21, p=0.003) and dehiscence (OR 0.47, p=0.02). PRI predicted higher reoperation (OR 6.4, p=0.01) and poor scarring (OR 34.8, p<0.001), though limited by baseline confounding.
Conclusions
Dressing selection may affect wound outcomes after reduction mammoplasty. Sylke tape was associated with lower rates of skin reaction and dehiscence across both analyses, though potentially limited by residual confounding. Prineo’s association with reoperation and poor scarring is notable but difficult to disentangle from patient selection bias. Combination dressings showed lower skin reaction rates but were heterogeneous, making interpretation challenging. Prospective data are needed to confirm these findings.