Management of Problems with Descemet’s Membrane

35 Management of Problems with Descemet’s Membrane

Randall J. Olson

The basement layer of the corneal endothelium is quite inelastic and extremely strong. It is clearly much thicker and tougher than the anterior capsule, and therefore resistant to tearing in most cases. Nonetheless, tears near the wound or the paracentesis and a significant tearing or stripping of Descemet’s with large detachments are not uncommon clinical problems associated with cataract surgery (Box 35.1). Like so much else in medicine, prevention is the key in that, just like capsulorrhexis, small tears tend to extend into large tears!¹

To gain access to the anterior chamber (AC) for surgery with a self-sealing incision, phacoemulsification incisions, by definition, cut through Descemet’s membrane. A clean cut through Descemet’s membrane with a sharp instrument, be it metal or diamond, is one of the keys to avoiding further complications.

Almost all tears occur anterior to the entry wound into the AC. Tears that occur at the time of the incision or paracentesis are almost always due to an instrument or technique failure. One cause is sweeping, either anteriorly or posteriorly, as we start to enter Descemet’s membrane. This is easier to do than one might imagine in our creation of a relatively long tunnel and self-sealing incision. With a sharp instrument, even with movement as we enter Descemet’s, it is hard to create a tear; therefore, when a tear occurs it is most commonly created by a dull blade, be it diamond or metal. However, reused metal instruments are probably the most common cause (Fig. 35.1). If tears occur at the time of the incision or paracentesis, even with a sharp instrument, videotaping the technique so it can be critiqued should help in making changes in the future that obviate this problem. Even small tears are a potentially much bigger problem later.

Most anterior tears are an insertion problem created by catching the anterior Descemet’s lip whenever an instrument is inserted into the eye. Due to the small size of the irrigation and aspiration (I/A) tip, it has been our experience that it is difficult to create a tear during this phase, and therefore it usually occurs when the phacoemulsification instrument or intraocular lens (IOL) is inserted into the eye. The unfortunate scenario is a small tear created at either the time of incision or with the phacoemulsification instrument insertion (Fig. 35.2) and then significantly extended during phaco, with the movement of the phaco tip or irrigation, or later, with IOL insertion (Fig. 35.3). This is also much more common with a very tight wound and difficult insertion at any time.

The following are general rules for avoiding anterior tears

Do not create a very tight wound! If insertion is difficult, it is much better to enlarge the wound prior to IOL insertion. An exceedingly tight wound is also a problem in regard to in inflow difficulties (pinching the sleeve), which can result in chamber depth fluctuations and phaco burns! As we torque the instrument inside the eye, tight wounds also often result in considerable visual distortion due to wrinkling of the cornea; therefore, there is no advantage to an overly tight wound.
Insert the phaco instrument parallel to the wound plane. Although this may be self-evident with chambers that have collapsed, it is not uncommon to lift up the instrument to avoid the iris. We think this is a very common way that small anterior tears occur. A little viscoelastic just inside the wound, allowing room to enter the eye parallel to the incision, is one way to avoid this problem, especially when there is a small tear that we do not want to enlarge!
Place posterior pressure on the wound or paracentesis. Because catching on the anterior lip is the real problem, it is our practice always to purposely push slightly posteriorly as we enter the wound to avoid putting pressure on that anterior cut edge of Descemet’s.

Box 35.1 Descemet’s Membrane Problem

Prevention

Use sharp instruments (diamond or new metal) to create incision and incise Descemet’s membrane.

Management

Prevent enlargement during insertion of the phaco tip and the intraocular lens (IOL).
If the wound is tight, enlarge it.
If the tear is on one side, enlarge the opposite side. Insert the phaco tip parallel to the wound, but do not lift the tip on insertion.
Use viscoelastic in the wound to open it and provide room for phaco tip insertion.
Prior to IOL insertion, enlarge the wound.

Treatment

Make a single tear on one side of the wound less than 2 mm from the visual axis.
Irrigate closed through the paracentesis and ignore.
A double tear will scroll.
Attempt irrigation; if successful, ignore.
If unsuccessful, do viscoelastic reattachment.
Use three full-thickness sutures.
Leave viscoelastic; treat for potential intraocular pressure (IOP) spike with topical β-blockers and carboxy anhydrase inhibitor (CAO) (Diamox).
Giant (or more than three quarters of Descemet’s membrane detachment) is treated by a 20% sulfur hexafluoride (SSF6) bubble.

Fig. 35.1 A dull blade piercing Descemet’s membrane and pushing it away from the stroma.

We can often feel the resistance as we enter and engage the anterior Descemet’s edge. Relieve this pressure by pushing posteriorly and actually feel Descemet’s pop over the front edge of the sleeve, thereby avoiding a tear.

Remember that it is the small tear that becomes the large tear, so we need to be extra cautious if we notice a tear. I find it well worth opening the incision on the side of the wound opposite the tear to avoid additional pressure on that edge with the wound. Now is the time to be extra careful about each maneuver as we go through the wound to avoid extending this tear.

Creating a tear with IOL insertion also comes from trying to put too large a lens through too small a wound. Videotapes of insertion techniques in which surgeons seesaw their way through an extremely tight wound demonstrate that is a potential recipe for a large tear disaster. Generally, even under these circumstances, just a small tear is created, and a large tear occurring means there has already been a small one in place. All the previous rules already mentioned apply, but in particular be concerned about the overly tight wound. It only takes a second if the wound is too tight to enlarge it slightly.

What do we do when we already have a tear and therefore it is too late for prevention? At our center, we have found that most often it is a single tear on one edge of the wound (Fig. 35.4), and as long as it is more than 2 mm from the visual axis, we have been able to irrigate it closed through the stab incision using balanced salt solution at the end of the procedure. It always looks fine the next day, and the endothelial pump will keep the tear attached. The only damage done under these circumstances is usually to the surgeon’s ego, except for the fact that such wounds do not “self-seal” as easily, and, rarely, they also require a suture.

A double tear on both ends of the wound is fortunately harder to create but unfortunately presents a bigger problem (Fig. 35.5). Often one side of the tear will scroll up near the visual axis, and a single tear into the visual axis also acts in a similar way with some scrolling so that it will not easily irrigate into place. Such wounds rarely self-seal. If, however, we can irrigate the flap closed with a self-sealing wound, and it stays this way for a couple of minutes after the procedure, we still do nothing else, and it is fine the next day. Fortunately, we have not had a Descemet’s tear that looks fine at the table but has been a problem the next day. However, those cases with extensive scroll detachment that will not irrigate into place require a definitive treatment on the surgical table (Fig. 35.6).²

One approach is using an ophthalmic viscosurgical device (OVD) to hold the torn area into position against the cornea (Fig. 35.7). Although OVD is certainly effective, it is a double-edged sword. If we are not careful, we can get the OVD between Descemet’s and the stroma, and then achieving resolution on the table is exceedingly difficult.³ Vigorous I/A to try to remove the sub-Descemet’s viscoelastic can create considerable endothelial damage, and under such circumstances very gentle I/A, allowing time for the OVD to reabsorb on its own, is best. Spontaneous reattachment without needing an additional surgical procedure is still common in such cases.