Systematic review and meta-analysis on outcomes for endoscopic versus external dacryocystorhinostomy

Background: Dacryocystorhinostomy (DCR) is commonly performed for epiphora, dacryocystitis and during tumor surgery. External (EXT-DCR) and endoscopic DCR (END-DCR) are both practiced. END-DCR was initially performed with laser (EL-DCR) but has shifted to careful bone removal with mechanical drills (EM-DCR). High level evidence from comparative cohorts was sought to compare outcomes. Method: Medline (1966 - January 28th, 2013) and Embase (1980 - January 28^th, 2013) were searched for comparative studies (RCT/cohorts) of END-DCR to EXT-DCR for acquired nasolacrimal duct (NLD) obstruction. Primary outcome was DCR success, defined as resolution of symptoms and/or patent NLD on irrigation or dacroscintography. Secondary outcomes were scarring, infection and post-operative bleeding. Meta-analysis was performed with the Mantel-Haenszel Method and presented as Risk Ratios (RR) with Confidence Intervals (CI). Results: The search identified 3582 studies and 355 were reviewed after screening. Full text review yielded 19 studies (4 RCTs and 15 cohorts). Overall, EXT-DCR had slightly better success rates than END-DCR (RR 0.96, CI 0.93-1.00). However, EM-DCR outcomes were comparable to EXT-DCR (RR 1.02, CI 0.98-1.06), whereas EL-DCR had poorer outcomes (RR 0.85, CI 0.79-0.91) when compared separately. The RR for scarring, bleeding and infection with END-DCR versus EXT-DCR was 0.07 (CI 0.02-0.22), 0.72 (CI 0.46-1.13) and 0.24 (CI 0.11- 0.54), respectively. The rates of reported revision surgery were similar. Conclusion: DCR is a procedure with high success rates. Endoscopic procedures differ greatly by technique with EM-DCR offering comparable results to EXT-DCR, without the risk of cosmetically unacceptable scars.