Objective: To explore factors influencing hearing outcomes in children treated by canal wall up (CWU) and canal wall down (CWD) mastoid surgery.

Methods: Retrospective cohort study including three units in Bristol and Bath, UK. Ninety consecutive children underwent cholesteatoma mastoid surgery, with the first procedure between 1998 and 2001; minimum follow up was three disease-free years.

Results: The CWU and CWD cohorts significantly differed in pre-operative stage and hearing. After disease eradication, air conduction (AC) thresholds changed by +4.0 dB (95 per cent confidence intervals (95%CI) −2.0, 10.1) in the CWD group and −5.3 dB (95%CI −9.3, 1.3) in the CWU group (p=0.029). Using multiple linear regression to account for cohort differences, AC thresholds were increased by: pre-operative AC threshold (p<0.0001), initial ossicular stage (p=0.013), and CWD-surgery (p=0.005).

Conclusion: Disease-free hearing was better with CWU-surgery, less initial ossicular damage, and better pre-operative hearing. Worse initial disease increased the likelihood of CWD surgery. Wider use of ossiculoplasty in the CWU cohort (51 per cent vs 5 per cent) may partially explain the superior results.