This meta-analysis screened 976 records and pooled 72 studies covering 3,783 ears, comparing lateral wall against perimodiolar electrode arrays for hearing preservation in cochlear implantation.
Pooled hearing preservation was 45.9% for lateral wall against 34.2% for perimodiolar arrays. But the head-to-head comparison — 18 studies, 1,710 ears — gave a risk ratio of 1.24 with an interval from 0.99 to 1.56 and a p value of 0.057, which does not clear the conventional threshold.
The sensitivity analysis is what makes this interesting. Excluding studies with a month or less of follow-up, the risk ratio became 1.37 with an interval from 1.06 to 1.77 and a p value of 0.016 — significant. So the difference between the arrays appears with time rather than immediately, which is consistent with delayed loss from intracochlear trauma rather than an acute insertion effect.
The authors are careful not to turn this into a rule, and they are right. Electrode choice depends on preoperative residual hearing, cochlear anatomy and surgeon familiarity, and an unfamiliar array in experienced hands is not obviously better than a familiar one. What this does support is that where hearing preservation is the priority and both are available, lateral wall is the reasonable default.
- Where residual hearing preservation is the goal, lateral wall arrays are the reasonable default.
- The difference emerged only with follow-up beyond a month — short-term audits will miss it.
- Do not override cochlear anatomy or surgeon familiarity on the strength of a pooled risk ratio.
- Audit hearing preservation at six and twelve months, not at discharge.
The statistics, in plain English
This is a good example of why a p value of 0.057 should not be read as no difference. The head-to-head risk ratio of 1.24 had an interval of 0.99 to 1.56 — it barely touches 1.0, so the data lean towards a real advantage without formally establishing it. Excluding studies with a month or less of follow-up raised it to 1.37 with an interval clear of 1.0. A sensitivity analysis that strengthens a result is more persuasive than one that weakens it, particularly when the exclusion has a clinical rationale rather than a statistical one: hearing loss after implantation develops over months, so a study measuring at four weeks was never going to detect it. The overall pooled rates of 45.9% against 34.2% come from all studies including non-comparative ones, and should carry less weight than the head-to-head figure.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for ent & head and neck surgery, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free