13Tympanic Membrane Reconstruction and Prosthesis Coupling

FWhy the drum is part of the ossiculoplasty

It is tempting to think of ossiculoplasty as a problem of struts and stapes — the bridge between the oval window and whatever lies lateral to it. But a prosthesis is only as good as the two things it touches, and laterally it almost always touches the tympanic membrane. A reconstruction that ignores the drum is a bridge with one abutment built on sand. Wullstein framed the whole enterprise of tympanoplasty not as anatomical closure but as the restoration of a sound-protected, aerated middle ear with a vibrating membrane coupled to the oval window [1956]. That definition is exactly the brief for this module: the drum must be intact, mobile and well-supported precisely where the prosthesis loads it, or the elegant medial coupling counts for nothing.

Two failure modes make the point. First, a perforationor a thin, retracted drum leaks the very pressure the transformer is supposed to concentrate. The conductive loss a perforation causes is frequency-dependent — worst at low frequencies — scales with the size of the hole, and is larger for a given hole in a smaller, poorly aerated middle-ear and mastoid volume [2006]. Second, a rigid alloplastic head plate resting against a bare drum remnant concentrates its load on a thin membrane and, over months, erodes through and extrudes. Both problems are solved at the drum, not at the stapes: the membrane is reconstructed to be intact and aerated, and the lateral interface is supported so the prosthesis is held without being expelled. Rebuilding the drum is therefore not a finishing touch after the ossiculoplasty — it is part of it.

FFascia or cartilage: choosing the graft

Two autografts dominate drum reconstruction. Temporalis fascia (or its cousin, the fascia of the deep temporal muscle) is thin, pliable and acoustically near-transparent; it heals into a supple membrane that vibrates much like the native drum. It is the historic workhorse for a straightforward perforation in a well-aerated ear. Cartilage— harvested from the tragus or concha, usually with its perichondrium — is stiffer and heavier but far more resistant to retraction and resorption. The clinical question is rarely “which graft is better?” in the abstract, but “which graft does this ear need?” A dry, primary ear with good Eustachian-tube function will do well with fascia; an atelectatic, revision, or poorly aerated ear — and any drum that must carry a prosthesis— leans toward cartilage.

The trade-off has been quantified. A meta-analysis of 44 studies and 4,582 patients found that cartilage achieves a higher graft take rate than temporalis fascia, while fascia gives a marginally smaller postoperative air-bone gap; overall hearing gains are comparable, and palisade cartilage actually performed best on postoperative pure-tone average [2025]. In other words, cartilage buys reliability and durability at a small acoustic premium that, in practice, is hard to detect. Dornhoffer’s series of more than 1,000 cartilage tympanoplasties — spanning cholesteatoma, recurrent perforation and atelectasis — reported around 96% closure with air-bone gaps settling into the 11–15 dB range, confirming that cartilage is not an acoustic compromise so much as an insurance policy for the difficult ear [2003].

TThe lateral interface: coupling the prosthesis to the drum

Where the prosthesis head meets the lateral structures is the crux of this module. There are three broad strategies, and they form a clear hierarchy. The best is to couple to the malleus handle: the head plate aligns along, or grips, the manubrium. The retained malleus steadies the whole construct, steers the force vector toward the oval window, and seats the prosthesis near the umbo — the region of maximal drum vibration. Whenever the handle is present and reachable, this is the interface of choice, and clinical series consistently show better, more durable hearing when the malleus is incorporated.

When the malleus is absent, medialised or foreshortened, the prosthesis must couple to the drum directly— and this is precisely where a bare alloplastic head plate fails. The solution is to interpose cartilage: a small plate or a perichondrium-cartilage island is laid between the head plate and the drum. It spreads the point load over a wider area, resists retraction in a poorly aerated ear, and dramatically lowers the extrusion rate, which is the chief reason alloplastic reconstructions are wrapped in cartilage at the interface [2003]. The third option — head plate against a bare drum — is the one to avoid: high early coupling, but a steady march toward erosion and extrusion. The explorer below contrasts the three.

Two refinements make the cartilage interface work. First, a perichondrium islandgraft — a disc of cartilage with a surrounding skirt of perichondrium — can be suspended over the drum remnant so the perichondrium anchors it and the cartilage centre takes the head-plate load. Second, the prosthesis itself is often given a gentle bend (Dornhoffer and Gardner advocated roughly 30° toward the promontory) so its shaft follows the conical geometry of the reconstructed drum rather than tenting it laterally. The reconstructed membrane and the prosthesis are designed together, not in sequence.

TCartilage thickness, palisades and the acoustic cost

Cartilage’s one liability is its mass and stiffness: a heavy, rigid plate damps the high-frequency end of the drum’s response. The fix is to manage thickness and geometry deliberately. Temporal-bone vibrometry showed that a plate of about 0.5 mmgives the least acoustic transfer loss while still resisting deformation under the pressure swings of a working middle ear — and since native tragal or conchal cartilage is 0.7–1.0 mm at harvest, surgeons routinely thin it toward that target [2000]. Below 0.5 mm the graft transmits beautifully but begins to retract; above it, stability rises but the high frequencies are increasingly muffled. The slider below makes this opposition explicit.

Geometry matters as much as thickness. A continuous full plate behaves as one stiff mass; segmentingit — the palisadetechnique, in which several strips of cartilage are laid side by side — reduces the effective mass and stiffness and so preserves more of the high-frequency transfer, which is why palisade reconstructions edge out solid plates acoustically [2002]. Tos catalogued the many variants — palisades and slices, foils and plates, perichondrium-cartilage island grafts, and total pars-tensa composite grafts — into a six-group classification that gives a shared vocabulary for these choices [2008]. The practical rule that emerges is simple: use the thinnest, least massive cartilage construct that will still support the interface and resist retraction in thisear. In a stable, aerated ear that may be a thin island; in a wet, atelectatic, revision ear it may be a sturdier plate, accepting a little high-frequency cost for the security of take.

CAeration, the round window and re-establishing the transformer

A perfectly reconstructed drum and a perfectly seated prosthesis will still disappoint if the middle ear cannot breathe. The transformer is an aerated system: an air cushion behind the drum lets the membrane move, and a free round window lets perilymph displace so the cochlear partition can be driven. Two clinical corollaries follow. First, do not seal the round window. The aim of reconstruction is to preserve the oval–round window pressure difference, not to occlude the round window; packing, an over-large graft, or a misplaced cartilage shelf that damps the round-window niche will blunt the very gradient the transformer depends on. Second, restore aeration. A non-aerated, retracted drum stiffens against the promontory, raises input impedance, and degrades coupling regardless of how good the prosthesis interface looks at the end of the case [2006].

This is why drum reconstruction and ossiculoplasty are judged by the company they keep. Statistical staging of ossiculoplasty outcomes repeatedly finds that the middle-ear environment— mucosal health, drainage, aeration — outweighs the prosthesis material chosen, and a well-supported, well-aerated reconstruction is what converts an anatomically tidy operation into a hearing result [2001]. Re-establishing the transformer, in Wullstein’s original sense, means all of it together: an intact, mobile, supported membrane; a coupled prosthesis driving the oval window; a free round window; and an aerated cleft behind it [1956].

CPutting it together: a reconstruction sequence

A workable intra-operative sequence falls straight out of the principles above. First, assess the anchors and the environment: is the malleus handle present and reachable, is the mucosa healthy, is the ear aerated, is this a primary or a revision ear? Second, choose the lateral coupling: if the malleus is reachable, plan to couple to it; if not, plan a cartilage-supported drum interface. Third, choose the graft: fascia for a simple, dry, well-aerated perforation; cartilage wherever the drum is atelectatic, revision, or must carry a prosthesis — accepting the higher take rate as cheap insurance [2025, 2003].

Fourth, shape the cartilage: thin toward 0.5 mm, and segment into palisades or use a perichondrium island when high-frequency preservation matters and the ear is stable enough to allow it[2000, 2002, 2008]. Fifth, set the prosthesis: seat the head near the umbo, bend the shaft to follow the drum’s cone, and confirm light, even contact without tenting or footplate overload. Sixth, protect the windows and the air space: keep the round window free, avoid over-packing, and respect aeration [2006, 2001]. Done in this order, the drum stops being an afterthought and becomes what Wullstein always intended: the supported, vibrating lateral face of a restored transformer.