Audiogram shape is a prognostic marker in sudden sensorineural hearing loss

Design

Retrospective cohort with multivariable logistic regression, single tertiary centre, 2012-2019

Population

487 adults with unilateral idiopathic sudden sensorineural hearing loss

Primary outcome

Complete recovery by presenting audiogram configuration

Effect

Complete recovery 50.0 per cent ascending, 25.2 per cent flat, 9.8 per cent descending, 0 per cent total; descending OR 0.25 (95 per cent CI 0.15 to 0.42), total OR 0.05 (0.01 to 0.22); AUC 0.78

Four hundred and eighty-seven adults with unilateral idiopathic sudden sensorineural hearing loss treated at one tertiary centre between 2012 and 2019 were classified by the configuration of the presenting pure-tone audiogram: flat in 43.8 per cent, descending in 29.4 per cent, totally deaf in 21.9 per cent and ascending in 4.9 per cent.

Complete recovery occurred in 80 patients, 16.4 per cent overall, and the spread by configuration is the finding: 50.0 per cent of ascending curves, 25.2 per cent of flat, 9.8 per cent of descending, and none of the totally deaf ears. On multivariable analysis a descending configuration independently predicted failure to recover completely (OR 0.25, 95 per cent CI 0.15 to 0.42) as did total deafness (OR 0.05, 0.01 to 0.22), alongside older age and a worse initial pure-tone average. The model reached an area under the curve of 0.78 with a Nagelkerke R-squared of 0.32.

None of this changes what you give - the audiogram configuration is a prognostic marker, not a treatment selector, and steroids are still started urgently for everyone. What it changes is the conversation. A patient with a descending curve who is told the average recovery figure will hear something considerably more optimistic than their own position; one with an ascending curve told the same figure will be needlessly frightened. The shape is already in front of you at the first visit and costs nothing to use.

• Classify the presenting audiogram as ascending, flat, descending or total, and record it.
• Counsel from the configuration-specific figure, not the pooled 16 per cent.
• Treat a totally deaf ear as the worst prognostic group, where no patient recovered completely here.
• Do not withhold or delay steroids on prognostic grounds; this predicts outcome, not response.
• Add age and initial pure-tone average to the discussion; both were independent predictors.

The statistics, in plain English

The odds ratios describe complete recovery, so values well below 1 mean worse odds; an interval of 0.01 to 0.22 for total deafness is wide because no patient in that group recovered, which makes the estimate unstable even though the direction is unambiguous. An area under the curve of 0.78 is useful discrimination but not enough to predict an individual, and a Nagelkerke R-squared of 0.32 says roughly two-thirds of the variation remains unexplained. Only 24 patients had ascending curves, so the 50 per cent figure rests on a small group.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.