10The Austin-Kartush Classification of Ossicular Defects

FTwo questions, four types

When a surgeon lifts the tympanic membrane and finds an eroded ossicular chain, the eye is drawn to what has been lost— most often the long process of the incus, the most vulnerable ossicle in chronic ear disease. Austin’s insight in 1971 was that the reconstructive future of the ear is better predicted by what remains, and that the two structures that matter most can each be reduced to a single yes-or-no question[1971]. Is the malleus handle still present? Is the stapes superstructure(the arch and head, the capitulum) still present? Two binary questions generate a 2×2 grid — four combinations, which Austin labelled types A to D. Almost the entire logic of ossicular reconstruction can be hung on this small matrix.

The reason these two structures were chosen, rather than the incus or the drum, is purely mechanical. A prosthesis must be anchoredat two ends: medially it must drive the inner ear, and laterally it must be supported and aligned. The stapes provides the medial coupling — either its head, if the superstructure survives, or its footplate, if it does not. The malleus provides the lateral support and, crucially, steers the force vector toward the cochlea rather than letting the prosthesis tilt or migrate [1971]. Knowing which of these anchors you have therefore tells you, before any prosthesis is chosen, what kind of bridge you are able to build. The matrix is, in effect, a pre-operative inventory of anchor points. One important convention underlies the whole grid: the footplate is assumed to be mobile across all four types. A fixed footplate is a different problem, handled separately, and is one of the gaps that Kartush later filled.

FReading the quadrants A to D

The four types follow directly from the two questions. Type A is the favourable corner: malleus handle present andstapes superstructure present. Here only the incus is missing, so a short partial prosthesis can be slotted from the stapes head to the malleus — a near-anatomical repair with both anchors intact [1971]. Type B keeps the malleus but has lost the superstructure, leaving only the mobile footplate medially; the prosthesis must now be a longer total reconstruction reaching the footplate, but the retained malleus still stabilises it. Type C is the inverse: the superstructure survives but the malleus is gone, so a partial prosthesis can sit on the stapes head, yet without the malleus the construct loses lateral support and the force vector is less ideal. Type Dis the difficult corner — neither malleus nor superstructure remains, only the footplate — demanding a total prosthesis with no malleus to steady it.

It is worth dwelling on why the diagonal matters. The two favourable types, A and B, share a retained malleus; the two less favourabletypes, C and D, share its absence. Austin and later authors came to regard the malleus as the more decisive of the two anchors for stability and durable hearing — a theme Black would make explicit two decades later, arguing that the malleus is the single most important ossicle for prosthesis stability and long-term gain [1992]. The matrix thus does more than label; reading down the malleus column already sorts ears into broadly better and worse prognostic halves before a single measurement of hearing is made.

TWhat Kartush added: fixation and the 0–F scheme

Austin’s grid is elegant but deliberately narrow: it describes erosion defects in an ear with a mobile footplate. It says nothing about a chain that is intact but immobile, nor about the ear whose problem is fixation rather than loss. In 1994 Kartush extended the scheme to close these gaps [1994]. He added three categories to the original four. Type 0 denotes an intact ossicular chain— the baseline against which defects are measured, and a reminder that not every diseased ear needs a prosthesis. Type E denotes malleus-head fixation, typically by tympanosclerosis or epitympanic adhesion, where the chain is continuous but anchored in the attic. Type F denotes stapes fixation, where the footplate cannot move. Together these convert Austin’s erosion grid into a fuller vocabulary spanning continuity, erosion and fixation.

The clinical force of the additions is that fixation changes the operation entirely. A type E ear may need the malleus head to be freed or divided before any reconstruction will work; a type F ear may need a stapedotomy or stapes mobilisation rather than — or in addition to — a strut. Slotting a prosthesis against a fixed anchor simply transmits sound into an immovable wall. Kartush did not stop at anatomy, however. In the same body of work he embedded ossicular status within a broader, weighted Middle Ear Risk Index (MERI), scoring otorrhoea, perforation, cholesteatoma, granulation, prior surgery and ossicular status together to stratify risk [1994]. Later refinement added host factors such as smoking, which independently worsens tympanoplasty outcome and compounds the ossicular score [2001]. The lineage from Austin to Kartush is therefore a move from a pure anatomical snapshot toward a multifactorial prognostic instrument, while keeping the malleus-and-stapes matrix at its core.

TFrom type to prosthesis: turning the matrix into a plan

The practical payoff of the classification is that it maps almost one-to-one onto prosthesis choice. The single most useful distinction it draws is between a partial ossicular replacement prosthesis (PORP), which couples to a surviving stapes head, and a total ossicular replacement prosthesis (TORP), which spans all the way to the footplate. Read directly off the matrix: types A and C retain the superstructure and are PORP territory; types B and D have lost it and are TORP territory. The malleus column then modifies the plan — when the malleus is present (A, B) the prosthesis can be anchored to or aligned along the handle for stability; when it is absent (C, D) a cartilage plate is usually interposed between prosthesis and drum to spread load and reduce the risk of extrusion.

The matrix frames the choice but does not rigidly dictate it, and type A is the clearest example. With both anchors present the obvious reconstruction is a short PORP, yet a prospective randomised trial in type A ears found that a TORP gave comparable hearing gain and air-bone gap closure while offering greater construct stability — so some surgeons elect a total prosthesis even when a partial one would reach [2024]. The lesson is that the Austin type defines the anchors available; the surgeon still weighs stability, extrusion risk and personal experience within that frame. Notice too that the matrix is read intra-operatively as much as pre-operatively: the audiogram suggests the defect, but the definitive type is assigned once the drum is elevated and the malleus, incus remnant, superstructure and footplate are each inspected and probed for mobility.

CPrognosis and the limits of the matrix

For all its utility, the Austin-Kartush matrix is a description of anchors, not a validated outcome predictor, and the experienced surgeon holds it loosely. Its central prognostic assumption — that an absent stapes superstructure (types B and D) worsens the achievable hearing — was directly tested when Dornhoffer and Gardner built the Ossiculoplasty Outcome Parameter Staging (OOPS) index from 200 ears. On multivariate analysis the presence of the superstructure was not an independent predictor of air-bone gap closure, a finding that gently overturns a long-held belief and shows that a well-seated TORP on a healthy footplate can hear as well as a PORP on a stapes head[2001]. What the same analyses repeatedly find does matter is the broader middle-ear environment: drainage, mucosal health, revision status and the presence of the malleus.

This is why modern practice treats the Austin type as one input among several rather than a verdict. Two ears of identical Austin type can have very different prospects if one is dry with healthy mucosa and the other is discharging through a fibrotic, poorly aerated cleft. The scoring systems built on top of the matrix — MERI from Kartush, SPITE from Black, OOPS from Dornhoffer and Gardner — exist precisely to fold these host and disease factors back in, and they consistently show that such factors, together with the malleus, outweigh both the superstructure and the prosthesis material chosen[1992, 2001]. The matrix tells you what you can build with; the staging systems tell you how likely the building is to stand.

CUsing the classification in counselling and audit

In day-to-day otology the classification earns its keep in three places. First, operative documentation: recording “Austin type B, mobile footplate, healthy mucosa, primary surgery” captures in a few words the anatomy that the next surgeon — or the same surgeon at revision — most needs to know. Second, counselling: the type lets you set honest expectations. A type A or B ear with good mucosa can be offered an optimistic prognosis; a type C or D ear, or any type complicated by active discharge, an absent malleus, a fibrotic cleft or a footplate that proves fixed, warrants a frank discussion of more guarded hearing and possible staging[2001, 2001]. Third, audit and research: because every otologist uses the same A-to-D vocabulary, hearing outcomes can be compared across surgeons and centres, and a series can be stratified by ossicular status rather than lumped together.

The mature way to use the matrix, then, is as a shared languageand a starting scaffold rather than an algorithm. Begin with the two questions — malleus handle, stapes superstructure — to fix the type and the broad prosthesis class. Add the Kartush categories when the problem is continuity or fixation rather than erosion. Then layer on the host and disease factors that the staging indices quantify, and the surgeon’s own judgement about stability and extrusion. The classification has endured for half a century not because it predicts outcome perfectly — it does not — but because it asks the two questions that most economically organise the operative decision, and because every otologist in the world can answer them at the same point in the same operation.