Sudden Sensorineural Hearing Loss: Oral Versus Intratympanic Steroids When Systemic Risk Is Real
A single patient with sudden sensorineural hearing loss and poorly controlled diabetes. Oral and intratympanic steroids work about the same — the real question is which route she can actually start today, and what the other one would cost her.
Two nights ago, hosting the mahjong group that has rotated between four kitchens on her block for six years, O.R. — a 58-year-old woman who has run the same neighborhood bakery for nineteen years — asked her guests twice to repeat themselves before admitting something was actually wrong. Her left ear had gone, in her words, “like it was full of water,” and a ringing started within the hour that hasn’t stopped since. There was no vertigo, no ear pain, nothing that felt like an infection. On exam this morning, Weber lateralizes to the right and Rinne shows air conduction greater than bone in both ears — the combination that puts the problem in the nerve rather than the middle ear. The audiogram confirms it: a pure-tone average of 55 decibels on the left, a 45-decibel drop across three consecutive frequencies, well past the 30-decibel mark that separates sudden sensorineural hearing loss from an ordinary complaint about hearing.
She has had type 2 diabetes for eight years, on metformin and glipizide, and her control has been slipping: the HbA1c drawn three months ago came back at 9.4%, high enough that her primary care doctor had already booked today as a medication review before any of this started. That number is what makes the choice of route live rather than academic. Rauch and colleagues’ 2011 randomized trial in JAMA compared the two head to head, and it enrolled patients presenting within fourteen days of onset with a pure-tone average of 50 decibels or worse; at two days and 55 decibels she clears both bars rather than sitting at the edge of either. That trial found intratympanic methylprednisolone non-inferior to a fourteen-day oral prednisone course, with a two-decibel edge to the oral arm, well inside the ten-decibel margin set in advance. Where the finding softened was a subgroup with loss of 90 decibels or more, dizziness, and presentation inside a week, where an advantage for oral could not be excluded. She has no vertigo and nothing approaching 90 decibels, so the subgroup that would argue for the systemic drug is not the one she falls into.
What that leaves is the decision the trial deliberately declined to make for anyone. Its own authors concluded oral prednisone should remain the default when all else is equal, on the grounds that it costs less and is easier than a needle through an eardrum. All else is not equal here. Whether her hearing comes back has as much to do with which door down the hall she gets sent through next as with the pharmacology behind either option — and today, unlike most patients with this diagnosis, that door happens to be two steps away.
In the ENT clinic, within the treatment window
I want to do the intratympanic injection today rather than start oral prednisone. Rauch’s 2011 trial is the one that actually compared the two routes head-to-head, and she sits inside the population it enrolled — inside fourteen days, worse than a 50-decibel average — so this isn’t a case where the systemic drug has a proven edge we’d be giving up. Given where her glucose already sits, that makes the injection the better default, not just an acceptable alternative.
I don’t doubt the trial data — I’m not questioning that the two routes perform comparably.
What I’d push back on is treating “comparable efficacy” as though it settles the whole decision, when the thing that actually determines outcomes in sudden hearing loss is how fast treatment starts, not which molecule delivers it. If getting her the injection meant even a two- or three-day wait for someone available to do it, I’d take oral prednisone today over a better plan next week, and manage her glucose around it — a short, actively-monitored steroid burst on top of a 9.4% A1c is a real but solvable problem, and I’ve solved it plenty of times. Rauch himself concluded oral should stay the default when everything else is equal, and I’d rather argue about whether it’s equal than assume it isn’t.
That would be the right call if the delay were real, but it isn’t here — this clinic can do the injection this afternoon, same visit, no referral needed. Once that’s true, the timing argument doesn’t actually have anything left to argue against; it was never really oral versus injection, it was fast versus slow, and the fast option and the lower-systemic-burden option happen to be the same choice today. And I’d be careful how we describe the fallback: the salvage evidence in this disease runs the other direction, injection after oral failure, not the reverse. Holding oral prednisone for an inadequate follow-up audiogram is a sensible extension of the comparable-efficacy data, not a sequence anyone has trialed. So — injection now, oral held and labeled honestly as untested in that role, and not both at once — there’s no reason to take on the glucose risk of a drug we’re not using as first-line just to feel like we’re doing more.
Agreed: intratympanic methylprednisolone given today, same visit, with a follow-up audiogram scheduled at two weeks to judge whether recovery is adequate. Oral prednisone held in reserve should that recovery fall short — recorded in the plan as an extension of the comparable-efficacy data rather than a tested salvage sequence — and not started alongside the injection, since adding a second route wouldn’t add benefit the trial evidence supports, only systemic risk she doesn’t need to take on. Her glucose was checked before she left, logged as a baseline for comparison at follow-up.
Not fully settled, though it didn’t change today’s plan: the primary care physician’s broader concern — that defaulting to intratympanic injection assumes same-day access that most patients, and most clinics, don’t actually have — was not answered by the fact that this particular patient happened to have it. The otolaryngologist and pharmacologist’s reasoning holds for her specifically; whether it should become the default recommendation for a patient without today’s access remains, by the primary care physician’s own account, a genuinely different question.