Post-mastectomy reconstruction: A risk-stratified comparative analysis of outcomes

Sujata Saha, Armando A. Davila, Jon P. Ver Halen, Umang K. Jain, Nora Hansen, Kevin Bethke, Seema A. Khan, Jacqueline Jeruss, Neil Fine, John Y.S. Kim*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

10 Scopus citations


Introduction: Although breast reconstruction following mastectomy plays a role in the psychological impact of breast cancer, only one in three women undergo reconstruction. Few multi-institutional studies have compared complication profiles of reconstructive patients to non-reconstructive. Methods: Using the National Surgical Quality Improvement database, all patients undergoing mastectomy from 2006 to 2010, with or without reconstruction, were identified and risk-stratified using propensity scored quintiles. The incidence of complications and comorbidities were compared. Results: Of 37,723 mastectomies identified, 30% received immediate breast reconstruction. After quintile matching for comorbidities, complications rates between reconstructive and non-reconstructives were similar. This trend was echoed across all quintiles, except in the sub-group with highest comorbidities. Here, the reconstructive patients had significantly more complications than the non-reconstructive (22.8% versus 7.0%, p<0.001). Conclusion: Immediate breast reconstruction is a well-tolerated surgical procedure. However, in patients with high comorbidities, surgeons must carefully counterbalance surgical risks with psychosocial benefits to maximize patient outcomes. Level of evidence: Level 3.

Original languageEnglish (US)
Pages (from-to)1072-1080
Number of pages9
Issue number6
StatePublished - Dec 2013


  • Breast reconstruction
  • Complication
  • Implant
  • Outcomes
  • Submuscular
  • Tissue expander

ASJC Scopus subject areas

  • Surgery


Dive into the research topics of 'Post-mastectomy reconstruction: A risk-stratified comparative analysis of outcomes'. Together they form a unique fingerprint.

Cite this