6Carhart's Notch and Bone Conduction Artifacts

FWhat the notch looks like

Plot a bone-conduction (BC) audiogram for an ear with a long-standing, purely conductive hearing loss behind a normal drum, and something odd sometimes appears: the curve runs along comfortably and then dips, sharply and symmetrically, at 2 kHz before recovering again at the next octave. That dip is Carhart’s notch, described by Raymond Carhart in 1950 when he noticed that bone conduction in otosclerotic ears was depressed in a stereotyped way — and that much of the depression vanished after surgery on the stapes [1950].

The classic shape has remarkably consistent dimensions. Relative to where the BC curve “should” sit, Carhart reported offsets of roughly 5 dB at 500 Hz, 10 dB at 1 kHz, 15 dB at 2 kHz and 5 dB at 4 kHz— a notch deepest in the mid-frequencies and tapering on either side [1950]. The single most important fact about this notch is in its history: it improves after surgery. A threshold that genuinely reflected a damaged cochlea could not recover when you mobilise an ossicle. The notch, therefore, is largely a mechanical artifact rather than a measure of the inner ear. Toggle the tracing below between the fixed chain and the post-operative ear and watch the 2 kHz point lift.

Before going further, it is worth separating three things that are easy to blur together. The air–bone gap (ABG)measures the conductive component — how much of the loss the middle ear is causing. The bone-conduction threshold is supposed to measure the cochlear reserve— how well the inner ear can hear once the conductive obstacle is removed. The Carhart notch is the inconvenient case in which the bone-conduction reading is itself contaminated by the middle-ear lesion, so the two measurements are no longer independent.

FWhy it fools the cochlear reserve

Every surgeon planning an ossiculoplasty or a stapes operation wants to know the same thing: how well will this ear hear if I close the gap? The ceiling on that result is the cochlear reserve, and we estimate it from the bone-conduction thresholds. If those thresholds are artifactually depressed at 2 kHz, the ear looks as though it has an early sensorineural loss it does not actually have. A trainee who takes the notched BC at face value will underestimate the reserve, under-promise the result, and may even talk a candidate out of an operation that would have served them well.

The honest counselling point flows directly from the artifact: in an ear with a Carhart notch, the true cochlear reserve at 2 kHz is generally better than the audiogram shows. After successful stapes surgery the bone-conduction thresholds tend to improve — the phenomenon of overclosure, in which the post-operative BC sits above its pre-operative level precisely because the artifact has been removed. In one large laser-stapedotomy series, 81% of patients showed overclosure at one or more frequencies, maximal at 1 and 2 kHz, with mean gains of about 7.5 and 8.3 dB respectively [2009]. Decades of stapedectomy data tell the same story of gap closure with recovered bone conduction [1980].

TThe mechanism: a borrowed resonance

Why 2 kHz, and why a notch rather than a uniform shift? The answer lies in how bone conduction actually works. Bone-conducted sound is not a single pathway but a sum of several, classically attributed to Tonndorf. A compressional component squeezes the cochlear capsule and fluids directly and dominates the low frequencies; an osseo-tympanic component radiates from the vibrating canal walls; and crucially an ossicular inertial component arises because, when the skull vibrates, the ossicular chain lags behind by its own inertia and moves relative to the cochlea, driving the stapes in and out of the oval window [2003].

That inertial contribution is frequency-dependent and peaks near the resonant frequency of the ossicular chain, which sits around 1.5–2 kHz. When the footplate is fixed, the chain can no longer move freely against the cochlea, so this resonance contribution is suppressed exactly where it was largest— producing a localised loss of bone-conducted sensitivity at 2 kHz while the compressional pathway carries on unchanged. Finite-element modelling that includes both the compressional and the inertial pathways reproduces the notch only when the ossicular resonance is removed, confirming that the dip is the audiometric shadow of a missing resonance, not a cochlear lesion [2013]. Step through the pathways below to see which survive fixation and which fall silent.

This framing makes overclosure intuitive. Free the footplate and you restore the ossicular resonance; the 2 kHz contribution returns, and the bone-conduction threshold “recovers” toward the true cochlear level. The cochlea never changed — the measurement did.

THow specific is it, really?

It is tempting to treat the notch as a pathognomonic sign — see a 2 kHz dip behind a normal drum, diagnose otosclerosis, plan a stapedotomy. The evidence is more sobering. Any lesion that stiffens or fixes the chain can blunt the same ossicular resonance, so the notch is a marker of a stiff chainin general, not of stapes fixation in particular. When investigators compared the incidence of a 2 kHz BC dip across ossicular pathologies, they found it in 31.4% of stapes fixation ears, 26.3% of incudostapedial joint detachment ears and 30.0% of malleus or incus fixation ears — no significant difference between the groups [2011].

Two lessons follow. First, the notch cannot localise the lesion: a malleus-head fixation or a tympanosclerotic chain can mimic the otosclerotic notch. Second, it is neither perfectly sensitive nor specificfor otosclerosis — many otosclerotic ears show no notch at all, and the dip is not always centred on 2 kHz. In one stapedotomy cohort the “notch” sat at 0.5 kHz in 31% of cases, 1 kHz in 32% and 2 kHz in only 37% [2009]. A systematic review reached the blunt conclusion that the current evidence is insufficient to use the Carhart notch as a diagnostic test for otosclerosis [2013]. It is a supportive sign, not a verdict.

CReading the artifact at the planning table

For the surgeon planning ossicular reconstruction, the notch is most useful when it changes how you read the rest of the audiogram. Practically, it does three things. It flags a fixed or stiff chain behind a normal drum, prompting you to think about footplate mobility, malleus-head fixation and tympanosclerosis before you open the ear. It warns you not to trust the 2 kHz bone-conduction pointas the cochlear ceiling — mentally “lift” that threshold toward its neighbours when you counsel about likely outcome. And it lets you predict overclosure: if the dip is a true Carhart notch, a successful operation should recover several decibels of bone conduction at 2 kHz, so the achievable hearing is better than the raw thresholds imply [1950, 2009].

The notch also sharpens an old clinical dictum. A conductive loss with a normal, mobile drum and a type A tympanogram is ossicular fixation until proven otherwise; a Carhart notch strengthens that suspicion. But it does not tell you whichossicle is fixed, and it does not substitute for the operative findings. The decision to perform a stapedotomy versus an ossiculoplasty, and the choice of prosthesis, still rests on what you see at the microscope — the notch simply tilts your pre-operative expectation toward a stiff chain with recoverable reserve.

CPitfalls, masking and honest counselling

Two technical traps can manufacture or exaggerate a “notch” that is not Carhart’s. The first is inadequate masking. Bone conduction has almost no interaural attenuation, so a thresholds read without proper masking of the non-test ear can reflect the better cochlea, or can wobble around 2 kHz in ways that imitate a notch. Any reported notch should rest on correctly masked bone-conduction thresholds; an unmasked dip is an artifact of testing, not of anatomy. The second trap is conflating a Carhart notch with a noise notch: occupational noise injury classically dips at 4 kHz (sometimes 3 or 6 kHz) and is a genuine sensorineural loss that will notoverclose. Centred frequency and clinical context separate the two — 2 kHz and conductive versus 4 kHz and sensorineural.

Used wisely, the notch is a small but genuine asset in counselling. When an anxious patient points at the 2 kHz dip and asks whether their “nerve hearing” is failing, you can explain honestly that much of that depression is mechanical, that it commonly lifts after surgery, and that their cochlear reserve is probably better than the figure on the page. That is a more accurate and more reassuring message than a literal reading of the audiogram allows. The discipline is to hold the sign at its true worth: a supportive, non-specific marker of a stiff ossicular chain with recoverable reserve— never a stand-alone diagnosis of otosclerosis, a localiser of the fixed ossicle, or a reliable measure of the cochlea [2011, 2013].