Devising, then sharing, OR secrets

Three glaucoma surgeons share the techniques that maximize their surgical outcomes.

The field of glaucoma surgery is constantly changing. Microinvasive glaucoma surgeries (MIGS) as well as traditional glaucoma surgeries are being used worldwide to help prevent patients’ vision loss from glaucoma. Interestingly, glaucoma surgeons have personalized techniques for performing the same surgery, and many of us feel these nuances help maximize our surgical outcomes. In the columns below, we learn from premier glaucoma surgeons about their unique approaches to MIGS and traditional glaucoma surgeries, and the surgical steps they consider essential to maximize safety, efficiency and efficacy outcomes. – Davinder Grover, MD, MPH


  • I create a fornix-based peritomy, leaving a cuff of conjunctiva at the limbus. This approach limits the transection of conjunctival vessels that aid in healing and allows for excellent visualization, and the cuff of limbal conjunctiva improves comfort by covering the sutures used to close the wound.
  • 9-0 nylon is my suture of choice for wound closure since I find it is less inflammatory than Vicryl sutures.
  • I close conjunctival button holes with 9-0 nylon on a BV needle (Ethicon 2829G). But, if sewing to sclera, then I use 9-0 nylon with a spatulated needle (Ethicon 2890G).
  • Ologen collagen matrix can be very useful in bleb revisions for overfiltrating blebs and/or bleb leaks. This is cut to size and placed in the desired location without suture or glue, then the overlying tissue is closed.


  • Early obstruction of flow is imperative in nonvalved implants to prevent hypotony. In most cases, I use a biodegradable 7-0 Vicryl to occlude the lumen of the tube. But, in select cases when a more controlled tube opening is desired, I will either do the surgery in stages (secure the tube to the sclera then at a later date implant the tube in the anterior chamber) or (more commonly) use a ripcord and permanent ligation suture. For the latter, I preplace a 4-0 nylon ripcord and ligate the tube with a 6-0 Prolene posteriorly. When ready, the ripcord can be removed in the office. I prefer this to ligation in the anterior chamber and lasering because it allows for fenestrations and is less bulky, so it slides easily through the scleral tunnel.
  • During the creation of the tube entry tract with a bent 23-gauge needle, my goal is to achieve a good position in the anterior chamber or sulcus while trying to minimize risk of tube exposure. Conjunctival erosions over tubes often occur 1-2 mm posterior to the limbus due to the mechanical forces of the eyelid, desiccation and the physics of a tube wanting to straighten at the bend. So, by keeping the tube covered with a patch graft under the upper lid and protected in a longer scleral tunnel, I reduce the risk of exposures.
  • Fenestrations to aid in early pressure control are unpredictable, but I create about three. If the tube is longer and snaked over the sclera, I target areas where the tube is curved — the increased gape in the fenestration would allow for more flow or take longer to scar over.


  • Patients with neovascular glaucoma often need the quick pressure reduction a valved implant provides. But, I always try to have the patient pretreated with an anti-VEGF agent prior to surgery. I also leave the tube long, since hyphemas are frequent.
  • As people age, aqueous production decreases. So, for octogenarians and older, I often opt for the smaller sized implant (Ahmed FP8, 102 mm2).
  • Though the valve is designed to prevent hypotony, I still use Healon GV to achieve a 30% anterior chamber fill at the conclusion of the case. This helps to buffer the immediate flow in eyes that has acclimated to high pressures and may need time to adjust aqueous production levels.


  • An ab interno trabeculotomy can be completed with a suture or iTrack catheter. I prefer to use the catheter due to the ease in visualization as it transverses Schlemm’s canal and the immediate recognition it gives if it begins to leave the canal to enter the anterior chamber or suprachoroidal space. I can connect the catheter to the viscoelastic, which allows for viscodilation of the canal. This can help break any adhesions and create space and better lubrication, making it easier for the catheter to advance.
  • If the catheter or suture stops and cannot be advanced despite viscodilation, then I remove it and begin to feed it in the opposite direction through the same goniotomy. In the majority of cases, simply changing directions is enough to complete the 360-degree trabeculotomy. In some, the canal cannot be fully cannulated, and I reposition myself and the microscope if necessary. Then, I use a 25-gauge MVR blade to create another goniotomy in the area of the stalled catheter, retrieve it and then attempt to complete the trabeculotomy in the other direction.
  • When doing a combined cataract surgery and GATT, I start with the GATT. If the hyphema is large enough that it interferes with the view for creation of a capsulorhexis, I irrigate the blood and viscoelastic out and refill with fresh viscoelastic before proceeding. I prefer this order because it maximizes the time irrigating fluid is flushed through the canal and the rest of the drainage system, which could help in achieving better pressure control.


  • Once I complete the cataract procedure and rotate the head, and the microscope has been rotated and a gonioprism has enabled a good view of the nasal angle anatomy, I position the CyPass (Alcon) inserter just below the scleral spur. I avoid 3 or 9 o’clock, where I may encounter a dense vascular plexus. Then, the implant should slide easily into the supraciliary space. Once the inserter disengages, I tap the device down with the inserter until 1 to 1.5 of the retention rings are visible.
  • When implanting this device under topical anesthesia, I warn the patient that there is some brief discomfort as I insert the device. I usually request that the patient receive some additional intravenous medications after completion of the cataract portion.
  • In eyes with a short axial length (<22 mm), use caution or consider a different MIGS procedure. These patients are at a higher risk of choroidal effusions and anterior chamber shallowing.


  • Good visualization is necessary, so I rotate the head and microscope and place a gonioprism on the eye. As I work with the blade through the temporal wound, I take care to minimize corneal striae that interfere with the view.
  • I use the inside-out technique that starts with the dual blade in the center of the nasal angle and opens outward in both directions. This results in two scrolls of trabecular meshwork that I can leave in place or remove.
  • Similar to GATT, when I’m doing a KDB (New World Medical) in combination with cataract surgery I often do the dual blade first so that more fluid can be flushed through the newly opened canal. However, there are some exceptions, such as eyes with a large lens causing mild narrowing of the angle. If using the ORA System (Alcon) to guide IOL selection, I will do the cataract first followed by the KDB.


  • Although KDB has largely supplanted it in my practice, I use Trabectome (NeoMedix) for cases in which I want to remove the meshwork but am concerned about bleeding. For instance, I recently treated an older child with Sturge-Weber syndrome with uncontrolled IOP, despite maximum medications. The patient had open angles but with tufts of vascular granulation tissue throughout. A trabeculotomy or dual blade would have likely resulted in a significant hyphema. In this case, I successfully used a Trabectome to open the inner wall of Schlemm’s canal and cauterize those abnormal vessels, which resulted in IOP lowering, minimal hyphema and quick visual recovery.

About the Author

By Jonathan S. Myers, MD


  • After opening the conjunctiva and Tenon’s capsule at the limbus with a single 1-to-2-mm cut, I use almost entirely blunt dissection by spreading with a blunt tipped Stevens tenotomy scissors to clear the conjunctiva and tenons from the sclera in the superior quadrant. Then, I force the Tenon’s up to the limbus, and open at the limbus with one to two cuts of the Stevens scissors — one blade in, one blade out. I find this process minimizes bleeding and is efficient, and freeing up the tissues makes for better exposure during the case and easier closure at the end.
  • Lidocaine jelly 1% has made surgery easier for me and my patients. After the betadine drops, I ask patients to look down. I then coat the superior bulbar conjunctiva with lidocaine jelly. I then hold the lid open, eyeball down, jelly in place, for 20 seconds or more to allow the lidocaine to absorb. This usually results in 20 to 30 minutes of great anesthesia, and I can perform the trabeculectomy without a block. Subconjunctival lidocaine 1% nonpreserved can always be added under the conjunctiva on a blunt cannula later in the case if needed.
  • Inject mitomycin C (MMC) solution, 0.1 cc to 0.2 cc of 50/50 mixture of lidocaine 1% preservative free with MMC (0.2-0.4 mg/cc), at the start of the case 6-8 mm posterior to the limbus. Massaging this into the superior quadrant, up to the limbus, is time efficient, allows diffuse application of MMC, requires less dissection and aids anesthesia. Surgeons first transitioning to this approach should start on the low end of MMC dosage and tie their trabeculectomy flaps a bit tighter to start.


  • Similar to the trabeculectomy anesthesia above, I use topical lidocaine jelly for tubes. This allows dissection back to the equator. After gentle blunt spreading in the superotemporal quadrant, a blunt cannula can be used to place short- or long-acting anesthetic further back. Most patients don’t need a patch, and the risks of retrobulbar injections and heavy sedation are avoided.
  • To facilitate a long tunnel for tube entry, I use a bent 23-gauge needle. First, with the needle bevel up, I bend the first 3 mm to 5 mm of the needle into a gentle sideways curve or a bend of at least 45 degrees using a hemostat. Next, I make a right angle bend upward in, away from the plane of the curve about 10 mm back from the tip. This results in a needle bent to resemble a curved hockey stick, which can be used to tunnel from 3 mm to 6 mm back from the limbus with the needle bevel up. The curve of the stick allows an approach from the superotemporal area towards the limbus. Twisting the hockey stick, by rotating the syringe down to force the needle to dive as it approaches the limbus, can help make the tube entry angle more parallel to the iris, further from the cornea. These long tunnels allow the surgery to be done without a patch graft. For right and left eyes, the needle is bent to curve in the opposite direction: when the needle is at the limbus in either eye, the curve goes straight back away from the limbus and then toward the temporal side. For a right eye, use the right hand to introduce the needle from 3 mm to 6 mm superotemporal to the limbus, making a bit of a curve toward the limbus. For the left eye, use the left hand.
  • In many cases, if the tube is left long, the plate can drape posteriorly behind the limbus without scleral fixation. However, two things are critical. First, the plate must be far enough back that, after pulling on the tube, the plate tends to drift back posteriorly (not continue coming forward). Second, the tube must be securely tied to the sclera at some point to prevent the tube from sliding posteriorly out of the anterior chamber. If these two details are not carefully attended to, plate and tube migration can lead to significant complications. Avoiding the scleral passes makes for a quicker surgery with less risk, especially if you are assisting a less experienced surgeon.
  • For nonvalved tubes, I no longer place a “ripcord” suture. I place a Vicryl ligature suture and wait. Many a time I was tempted to pull the ripcords when it was too early and the risk of hypotony or hemorrhage was high. The vast majority of ripcords were never needed or released. I still use ripcords for special situations, but not routinely.


  • Sometimes, it’s easier to do the iStent (Glaukos) at the beginning of the case when the cornea is clear, the surgeon fresh, the patient attentive and the viscoelastic thicker (for those of us who start with a dispersive viscoelastic). These factors all favor a better view — and an easier insertion.
  • When I watch the most graceful iStent videos, it appears helpful to angle the approach of the stent inserter in the corneal wound to allow a bit of an angle of attack, so that there is no need to indent the meshwork or canal. As soon as the meshwork is pierced, swinging the inserter to a more neutral angle flattens the stent approach allowing an easy glide along the canal. This can involve a surprising amount of wrist motion and flex if you are not accustomed to this approach.


  • As many surgeons will attest, if you encounter resistance halfway into the insertion, an early release of the CyPass from the inserter allows for easy completion of the insertion with a gentle push on the end of the implant. Use care not to release so early or abruptly as to allow the stent to contact the endothelium. Early release can turn a challenging insertion into a very easy completion.
  • Warn the patient before insertion that he may feel some discomfort for a few seconds. Not everyone does, but for some it’s notable.


  • A less experienced surgeon will find it easiest to start with pseudophakic eyes on the same side as the dominant hand. Novice Xen (Allergan) surgeons will be amply rewarded for time spent prior to surgery finding the best hand position that allows control of the needle tip but easy slider movement. Surgeons with smaller hands may find operating from above, rather than temporally, allows for easier hand positions.
  • In eyes with deep orbits, a more temporal and less inferior incision will make the surgery easier. Compensate by entering close to 12 o’clock, but at an angle. Rotate the eye down during the procedure once the sclera is engaged to create a better view of the superior bulbar conjunctiva when exposure is an issue.


  • A self-retaining AC cannula hooked to a high bag of BSS really speeds up fluid egress and aids more complete evacuation of serous choroidals. Choose the location carefully so it’s not in the way. An inferotemporal sclerotomy for choroidal drainage goes well with a superotemporal or inferonasal anterior chamber cannula in many cases.

About the Author

By Inder Paul Singh, MD


  • The key to predictable outcomes with a filtering procedure is to try to standardize, as much as possible, postoperative outflow. There are a few variables to control, including amount and time of MMC used, conjunctival wound creation and closure, scleral flap size and thickness, flap suture tension and size of sclerotomy (for trabeculectomy).
    • MMC. I now prefer to inject the MMC in the subconjunctiva posteriorly (around 4 mm to 6 mm behind the limbus) to promote posterior bleb formation. I use 0.2 mg/cc for older patients and naïve patients and 0.4 mg/cc for younger patients and those with more scarring potential. I find injecting the MMC provides a broader, more diffuse application of the medication. It also allows me to dissect the tissue planes better.
    • Wound construction. I use the least amount of cautery for hemostasis to reduce the risk of fibrosis. I prefer fornix-based wounds for visualization and to decrease posterior fibrosis. I will consider a limbus-based flap if I can rotate the eye downward; provide good visualization posteriorly; and if the conjunctiva is very thin.
    • Scleral flap. I create the same size and shape if possible to allow a better way to predict how much flow is expected. I usually like to keep at least 1 mm of flap on each side of the sclerotomy to protect from excess flow on each side of the flap. I try to maximize flow posteriorly rather than the sides of the flap.
    • Express shunts. I try to keep the flow more aggressive postoperatively than a trabeculectomy since fibrosis starts to occur immediately postop. The better the flow, the less chance of bleb fibrosis.


  • Prime the tube and make sure you test it after tying it off to confirm there is no flow. If there is some flow despite the ligature, you may have a hypotonous eye postoperatively.
  • I create three to four venting slits in the tube anterior to my ligature to provide some immediate postop flow. I use a 15-degree blade to perform these slits. Stop when you see the tip of the blade just crossing the other side of the tube.
  • I create a long scleral tunnel with a 22-gauge needle to pass the tube through. This decreases the chances of delayed erosion.
  • Postop courses may differ based on the type of the nonvalved device. For Baerveldt devices, I tell patients to watch for sudden decreases in vision at four to six weeks postop when the tube typically opens. I tell patients to decrease strenuous activity around that time to prevent a sudden decompression of the eye. For other non-valved devices (Molteno 3), we do not often see a sudden decrease in IOP when the tube opens due to the double reservoir. With this feature, though, it can take longer, sometimes three to four months, to reach target IOP.


  • Prime these devices, and make sure there is good flow through the tube and out the plate/valve mechanism. I use BSS in a syringe using a 27- or 30-gauge cannula. If not primed, the flow may be restricted postoperatively.
  • I do not grab the plate near the valve mechanism since that might disrupt or alter the functionality.
  • I place the plate as far back as possible (suture holes around 9 mm to 10 mm back), helping to create a more posterior flow and decrease the chance of postoperative bleb dysesthesia and fibrosis.
  • Postoperatively, I am aggressive with aqueous suppressants to decrease flow immediately and keep the IOP near the 10 mm to 12 mm Hg range. We have found decreasing immediate flow decreases the lay-down of fibroblasts and therefore decreases the intensity of the hypertensive phase we typically see at four to six weeks postop.


  • Leave some viscoelastic in the anterior chamber (AC) near the angle to help tamponade any refluxing heme. We often see a reflux of blood though the canal intra-op and thus form a hyphema postop.
  • I keep the head above the heart during surgery to help decrease chances of blood reflux. I often do the surgery with the patient in reverse Trendelenburg to help position the head above the heart.
  • I recommend using the iTrack (Ellex) catheter at first, since the blinking light makes it easier to see the catheter travel through the canal. After a few cases, I think it is fine to use a 5-0-nylon or prolene suture, which does save on cost. You need to use a cautery to blunt the tip to pass these sutures through. One advantage with this catheter is that it allows conversion to an ab interno canaloplasty procedure while at the table. One can inject viscoelastic through the catheter instead of performing the trabeculotomy.


  • I prefer to approach the trabecular meshwork (TM) more in a parallel fashion rather than at a 10-to-15-degree angle, like we were initially taught. Once in position, I push posteriorly as I slide the stent into the canal (“push and slide” technique). This allows for less “poking” into the canal and less bleeding in my hands. This also decreases the need to rotate my wrist as much.
  • Before introducing the iStent, I try to decompress the eye again right before I add more viscoelastic into the eye. This allows for a reflux of heme into the canal, which then helps me determine the likelihood of a patent canal. Also, it helps me identify where collector channels might be positioned. Once I see areas of heme, I target those areas for placement of the stent.
  • A subtle but helpful point I have learned has been to hold the loader in place as you press the button to deploy the iStent. I often “pulled back” right away as I pressed the button, but I noticed the stent was only partially in the canal. Once I held the stent an extra second or two in place after pressing the button, I found the stent stayed entirely in the canal more often.


  • There should be very little resistance when placing the stent into the supraciliary space. There is a small amount of initial resistance when piercing through the iris, but after that it should glide in smoothly. If you encounter resistance, change the angle of entry or the angle of your wrist. If you are too flat, you might be hitting the scleral wall; if too steep, you are likely hitting the ciliary body.
  • The visual cue to know the optimal amount of the CyPass stent is in the AC is the TM. I try to keep the stent opening at the TM level. That usually corresponds to two rings of the collar remaining in the AC.
  • Relax the hour hand when entering the angle. I have found it is common to raise the wrist when approaching the entry point, but that can change the angle of entry. Instead, it is better to relax the wrist and drop it slightly.
  • When pressing the button to release the CyPass, again, hold the loader in place for a second or two to decrease the chance of moving the stent once it is in place. Also, I leave the stent a little bit under-implanted and then use the loader to “tap” it further into the supraciliary space.
  • Under-implant the CyPass a bit. It is much easier to tap the CyPass in further than to pull it back with a forceps. If there are three rings or I see the stent collar above the level of the TM, I use the loader and gently tap it in further. Leaving it too long in the AC might increase risk of endothelial cell loss.


  • As with all MIGS procedures, visualization of the angle structure is key and keeping the AC deep and pressurized is helpful in these cases to prevent reflux of heme into the AC.
  • If removing the viscoelastic with irrigation and aspiration (I&A), I use the I&A tip to irrigate and aspirate the strip of TM, which is easier than grabbing the strip with forceps.
  • When excising the strip of TM, I recommend pushing the heel of the blade posteriorly more than you think needed. This adjusts the angle of the blade and keeps it from getting “stuck.” Therefore, if you feel resistance, you just push on the heel and raise the anterior portion of the blade to keep it from getting caught up with tissue. The blade should slide very easily along the curvature of the angle. OM

About the Author