In this roundtable discussion, expert vitreoretinal surgeons discuss management of cases where complications occurred. Below are the case presentations, videos, and discussion of the management and outcomes for these cases.
Christina Weng, MD, MBA: Any surgeon knows that if you operate enough, you will experience a complication. We often tend to showcase our best cases, but we learn twice as much when we share our worst complications. I am honored to have 3 surgeons joining me today for this roundtable who I respect very much: Lejla Vajzovic, MD, with Duke Department of Ophthalmology; Yoshihiro Yonekawa, MD, from Wills Eye Hospital; and Michael Jumper, MD, of West Coast Retina. Welcome, friends — I appreciate your willingness to teach others about how you handled some extremely complex retinal surgery situations.
Case 1: Retinal Folds
Lejla Vajzovic, MD: I’m thrilled to share with you a case of retinal fold after superior retinal detachment (RD) repair. I want to credit my fellow Ellie Zhou, MD, for helping with this case. This is a 56-year-old male referred for recurrent RD in the left eye. His primary concern besides blurry vision is that his image is shifted down and to the right. There is history of RD in the family, but otherwise nothing remarkable. The RD repair was initially in office pneumatic with gas. However, gas made its way subretinally, leading to a giant retinal tear which was then repaired with vitrectomy, laser, and oil. Unfortunately, there was retinal fold formation after that surgery.
On presentation, his visual acuity was counting fingers, pinholing to 20/250. He is phakic, with cataract forming. There was a large area of bare retinal pigment epithelium in the superotemporal area, and several retinal folds (Video 1). There was evidence of recurrent RD inferiorly, and the oil fill is about 80%. Optical coherence tomography (OCT) showed that this was multilayered folding with evidence of recurrent RD. This was a guarded prognosis given the fact that this was the third surgery, but the patient’s primary concern was shifting the image, so that was our goal.
The approach to relocating the retina was to first place a buckle. Because this was a phakic patient and I wasn’t planning on doing cataract surgery at the same time, I like a little bit of a broader element. There was proliferative vitreoretinopathy and multiple retinal holes in the inferior retinal area, about 1 month from initial surgery and a couple of weeks from the second. We sometimes talk about beautiful surgeries where we can detach this retina in one go with a subretinal technique. Even in this case, the retina was very detached inferiorly, so although it would have been nice to advance the fluid where we wanted it to go, the fluid will go where it wants to go.
We tried to create multiple blebs to detach the fold, but the fluid headed in the direction we didn’t want it to, escaping through the preexisting tear or adding to already bullous RD. We tried to advance fluid from inferiorly but realized that would be too difficult as well. So, we used a soft-tipped cannula to gently lift up the edges and then switched to the MaxGrip Forceps (Alcon) and continued to lift up the edges. We used a Flex Loop (Alcon) to help push some of the fluid around to less traumatically lift the macular area, but even that was unsuccessful. So, we used the forceps to lift the entire retina and then we proceeded with shaving, really identifying all the areas of retinal tearing, unfolding with the Flex Loop. I like to use the Flex Loop to maneuver the retina around gently to where we think it should be fitting. I recommend a very slow exchange with air and then proceeding with oil or gas. I often will do direct exchange. I opted for a slower fluid-air exchange, primarily because we had a fairly posterior tear to begin with that will allow for slower exchange of that fluid.
We got images 1 day and 1 week postoperatively. Optos images demonstrated that so far, the retina continued to be attached. The patient noted that the shifting image was much improved after even 1 week. An OCT follow-up showed that the fold was no longer there. There was some retinal thinning, and hopefully there will be continued improvement in terms of vision.
Case 1 Discussion
Christina Weng, MD, MBA: That was a nice OCT. I think one of the important distinctions to make is whether you’re dealing with a partial-thickness or a full-thickness retinal fold, and clearly that's a full-thickness retinal fold. What's the timing or urgency when you see this in a postoperative patient? Have you encountered a fold that doesn't easily flatten out? Is it ever totally adherent and unable to be separated? How would your management change in that situation?
Lejla Vajzovic, MD: Yes, I think there are different types of folds, and this was one where spontaneous resolution is unlikely given the degree of folding seen on OCT and throughout the superior macula. This patient was 2 weeks postoperative, and when I get a severe fold as referral, I tend to be aggressive if there are large areas involved and significant folding. In patients with less fold, there can be spontaneous resolution, and there are some great case reports of that. In this patient, I opted for relatively quick surgery because of the double layering of the retina that did not look like it would resolve, as well as a significant shift of the entire macula, so I thought that needed to be corrected as well.
Yoshihiro Yonekawa, MD: I think you did a phenomenal job. I also will peel the internal limiting membrane, which I feel makes the retina a little more elastic. Then I do the submacular BSS and steamroll the PFO. However, with significant folds, it is sometimes best to redetach the retina significantly and do a peripheral retinectomy. But often with peripheral retinectomy, when you want the retina to detach it doesn't, and when you don't want the retina to detach, it does. In this case there was already a detachment, so that worked nicely. I also loved how you used your Flex Loop instrument to very gently manipulate the retina.
J. Michael Jumper, MD: Lejla did a great job in this case. Had you relied on subretinal fluid, even if the retina were otherwise attached right around this fold and then detached inferiorly, it would be very difficult to get this loose enough to unfold, so recreating the retinotomy superiorly was a must. I also think putting the buckle on was important, because you had 2 problems: (1) this was an inferior detachment that wasn't going to resolve if you’d gotten this unfolded, and (2) there's a foreshortening of the retina, so putting a buckle on was very helpful in keeping this retina reattached.
The fluid-air exchange in a giant retinal tear case is a time to be very deliberate and careful. You want to dry that edge. At the beginning of the fluid air exchange with PFO in the eye, there's going be a sandwich of PFO below, infusion fluid on top of that, and then air coming in. With careful observation, you can see the infusion fluid go away and then you're left with only PFO and the air. And even then, I very slowly aspirate with the soft-tipped cannula along the retinotomy edge until I know it's bone dry before I take all the PFO out. That hopefully enables me to avoid big folds like this.
Christina Weng, MD, MBA: One subtle part of the video I thought you did beautifully was where you ducted the eye as superiorly as possible. And you got right into the fluid sandwich that Mike was referring to. Now let’s discuss positioning. Our colleague Cynthia Toth has described temporal positioning after RD repair to help the macula flatten out and prevent folds from forming in that area, which is very common in, for example, superior macula-splitting RDs. Do you follow that teaching?
Lejla Vajzovic, MD: I do follow that teaching. I ask the patient to spend an entire day temporal side down to allow all the fluid to drain out and then position them.
Christina Weng, MD, MBA: This all started from another surgery where during pneumatic retinopexy, the gas went into the subretinal space. I'm sure that's happened to all of us. Any tips from the group for how to avoid that?
Lejla Vajzovic, MD: I think the key is steady and slow injection to avoid any fish eggs or smaller portions of the gas bubble. You want one uniform bubble. For me, as I am injecting and visualizing, I go a little deeper and pull back to see my needle as I visualize and slowly and steadily inject.
J. Michael Jumper, MD: There is a judgment call on which eye is a good candidate for pneumatic retinopexy. In an eye with a large, stiff break, there may be more traction than pneumatic retinopexy can overcome. You have to develop a sense for when not to do a pneumatic procedure in such cases.
Christina Weng, MD, MBA: Speaking of meticulous drying, especially in these cases that are more susceptible to developing these retinal folds: colleagues like Rajiv Muni have described minimal drainage techniques, where you're deliberately leaving some fluid in place and letting it resorb more naturally. Do you think that would make a retinal fold more likely?
Yoshihiro Yonekawa, MD: Superior RD cases tend to have better single-surgery success rates, but they can be very challenging because of situations like this, especially as Christina mentioned the fovea-splitting superior detachments. Even with a microscopic fold or subtle retinal shifting, the patient can be very symptomatic. In these eyes, I push the envelope on pneumatics more than I normally would, because that tends to minimize inferior retinal displacement, or a primary scleral buckle if possible, as buckles without vitrectomy also have lower rates of retinal displacement. And if I have to do a vitrectomy-based procedure, what I've been doing more of is what Rajeev Muni talks about. After completing the vitrectomy with shaving and draining anteriorly to be able to retinopexy, I inject a partial bubble of 100% gas. This is essentially a “super pneumatic,” and the break is tamponaded and pexied, but the RPE pumps the fluid out more slowly, which minimizes retinal displacement.
CASE 2: Dislocated Intraocular Lens
Yoshihiro Yonekawa, MD: This is a case from one of my colleagues. This is an older female patient who is pseudophakic and had a dislocated intraocular lens (IOL), and the plan was to remove the IOL and suture in a secondary IOL (Video 2). As the IOL was removed, a segment of the Soemmering’s ring fell posteriorly. The surgeon attempted to engage the Soemmering’s ring with the vitreous cutter, but the cutter hits the retinal surface just outside of the arcade, as the instrument was used to try to engage the Soemmering’s ring from the side.
The IOP was immediately elevated for hemostasis, and the Soemmering’s ring was engaged from on top in a safer manner and the light pipe was used to push the Soemmering’s ring into the cutter. This is difficult because these rings can be really hard. Laser retinopexy was applied around the break, which already had some subretinal fluid, and a small piece of the Soemmering’s ring was still there. The surgeons shifted to lasering before getting this piece out to prevent the RD from progressing. After this, the smaller piece was engaged with the cutter again, removed, and the laser was applied where there was missing laser, the eye was left aphakic, and gas tamponade with SF6 was used.
Case 2 Discussion
Christina Weng, MD, MBA: All of us here teach fellows and train the next generation of retinal surgeons, and unfortunately, sometimes those types of things happen in the operating room. Anything you would have done differently in managing a case like this?
Lejla Vajzovic, MD: These things happen to all of us, and we learn tips and tricks as we are teaching our fellows. Visualization and depth perception are key, and sometimes focusing too closely on the object causes you to forget where your instrumentation is. In these cases, I like to teach my fellows to stay up higher and start by aspirating to feed the piece into the cutter and then switch it to cut mode while you have purchase. In this case, there is a significantly large piece that won’t swirl in the eye like smaller pieces, so I teach fellows to engage the piece from the top, bring it up to the mid-cavity area, and then cut.
J. Michael Jumper, MD: In this video, you'll see that immediately after recognition, the optic nerve vessels start pulsating because they've gone to a higher pressure to stop the bleeding. Also, the surgeon is working on the edge and outside of where the light is. I recommend keeping the light on the subject, and even leading the light to where you are headed. Illumination and better visualization in this case would have made a difference. Also, it is important to use the proper fragmatome setting when you are retrieving lens material from the surface of the retina. Linear fragmentation allows for initial aspiration followed by phacofragmentation once the foot pedal reaches the second position. With a setting in which fragmentation begins as aspiration commences, there is a greater risk of retinal damage.
Christina Weng, MD, MBA: Soemmering’s rings are deceiving because they can look fluffy and light, but they aren't always, and you can struggle with those for a long time. And the longer you're in the eye, the more chances there are for things like this to happen. In any lens case, I try to remove as much anteriorly as possible, instead of deliberately displacing it posteriorly. I also love using a 23-gauge platform in any sort of lens case because the larger port of the vitrector makes for more efficient lens removal. I always remind fellows to optimize their fluidic settings. Maximize your vacuum and slow down your cut rate. Your fellow did use the light pipe to feed the Soemmering’s ring into the vitrector port, which is a good maneuver, but if you hit a calcification and/or it’s simply taking too long, don’t waste your time. Take a minute to make a localized conjunctival peritomy, create a sclerotomy, use the fragmatome to remove the residual lens particles, and you’ll be out in no time.
When I am trying to pick something up from the retinal surface and it won’t come up with just vacuum, I like to jostle the eye and then supraduct or infraduct it to displace it as far away from the macula as possible, just in case you touch down, which everyone has done. Of note, when jostling the eye like this, you do need to be mindful of where your instruments are to avoid iatrogenic trauma.
Yoshi, how do you mentor trainees through situations like this, during and after surgery?
Yoshihiro Yonekawa, MD: For those with less experience, just pausing for a second is OK. If there is bleeding, I recommend pressing IOP 60, for hemostasis, and then having a conversation about what the next maneuver should be. Our patients are under local anesthesia, so we have these conversations very delicately. In debriefing, I remind them that we've all been in these situations and it’s difficult seeing these complications for the first time. We have to be very supportive in how we provide constructive feedback.
J. Michael Jumper, MD: You mentioned the very difficult lens remnants that can sometimes be calcific. The most extreme example of this is an adult patient with a 2 mm pupil and a detachment who had congenital cataract. Invariably, the lens remnants have turned calcific and are very prominent, and even with iris hooks there isn’t much view. In those cases, I’ll sometimes make a superior corneal scleral wound and deliver it much like I would a lens.
CASE 3: Sudden Patient Movement During Vitreous Hemorrhage Removal
J. Michael Jumper, MD: This is a 61-year-old male with diabetes, very severe vision loss, and proliferative diabetic retinopathy (Video 3). He had panretinal photocoagulation (PRP) in this eye, and he had some good regression of the proliferations. He developed a dense and nonclearing vitreous hemorrhage 6 months later.
This is a case I did with a fellow. She did an incredible job of delaminating fibrovascular proliferations off the nerve and arcades. I had taken over and was doing the last bit of work on the superotemporal arcade when the patient moved despite having his head taped down to the gurney. The video shows the patient whipping back and forth, which, in my opinion, is a greater risk in younger male patients. I removed the instruments from the eye and raise the infusion pressure. After he settled down and I was able to get back into the eye, there was blood and I was not sure where we stood. We could see the pulsatile nerve because of the elevated infusion pressure. It can be difficult to raise the pressure high enough to stop choroidal bleeding. After stopping the bleed with elevated pressure and endodiathermy, I was able to remove a clot from underneath the macula, very close to the fovea. After the blood was removed and the bleeding stopped, I slowly brought the infusion pressure back to normal and immediately addressed any recurrent bleeding. I also worked with the anesthesiologist to ensure that the blood pressure was not elevated, which can also make controlling the hemorrhage more difficult. I then did a fluid-air exchange.
Case 3 Discussion
Christina Weng, MD, MBA: It looked really nice at the end and yet could have gone a very different direction, so it goes to show that dealing with complications requires reflexive thinking and probably some muscle memory. The action you take in those critical seconds can really direct how a case goes and the ultimate prognosis. Mike, was the PFO bubble more to help with tamponade or did you use it to express fluid from that break?
J. Michael Jumper, MD: There was a lot of fluid around that break for me to laser. I didn't want to switch to air and worry about more bleeding, so I thought putting the PFO in would allow me to laser with good pressure and good visibility and then once I was happy with the tamponade and the diathermy, I could then change over.
Yoshihiro Yonekawa, MD: I had a similar case with a choroidal strike and we had to just stop, close up, and come back another day and fix the detachment again, and ultimately the patient did all right.
Christina Weng, MD, MBA: Mike, you mentioned younger males sometimes moving unexpectedly during surgery, and I’ll show a similar case in a moment. I haven’t found a great way to deal with that or reliably predict which patients will do that, so I find that I am more likely to put younger males under general anesthesia for surgery. Your patient was under monitored anesthesia care, right?
J. Michael Jumper, MD: Yes. I am very quick to use general anesthesia for these cases. I am quick to convert over if the patient is thrashing. Our anesthesiologists that I work with understand, and putting in a laryngeal mask airway usually is not a big issue.
Yoshihiro Yonekawa, MD: We're lucky that I think all of us sit superiorly and we can have our hands on the patient’s forehead. We can hold the patients somewhat in place and usually sense if they're feeling uncomfortable. It would be much harder to sense this for surgeons who sit temporally, like most of our cataract surgeon colleagues. In our practice at Wills, we tape the head of every single patient whether it's local or general anesthesia.
One pearl that I’ve learned is when patients are snoring, you just follow the eye and it’s usually fine. But if the patient stops snoring, you know they're about to wake up. When that happens, I come out of the eye, tap the patient’s forehead, and touch base with them before proceeding.
CASE 4: Retained Subretinal PFO After RD Repair
Christina Weng, MD, MBA: This is also a case where a patient’s head movement led to some unfortunate results. This is a 36-year-old male with a distant history of a staple gun injury to his left eye for which he had an open-globe repair and a scleral-fixated IOL placed about 10 years ago (Video 4). He experienced some vision changes recently in his left eye and was referred to me for an RD in that affected eye.
On examination, he had a macula-on rhegmatogenous RD with a 4 o’clock hour giant retinal tear (GRT) superiorly. For GRTs, I generally like to put on a low-lying buckle and then do vitrectomy. I almost never use PFO, but I do use it for GRT cases. And this is a young male who was also a bit anxious, so I decided to put him under general anesthesia, which I otherwise almost never do.
During the surgery I did a core and shave vitrectomy. Injecting the PFO bubble went smoothly, and the retina flattened nicely. I injected until it was beyond the posterior edge of the GRT and then I started my laser. Suddenly, the patient started bucking on the table and despite having his head taped, his head violently moved off the bed, which caused the infusion line to slip out. I replaced it and the anesthesia team stabilized him and readjusted his LMA (laryngeal mask airway). Unfortunately, when I went back in, there were all these small fish-egg PFO bubbles that had slipped beneath the superior GRT.
I tried tipping the eye, but it was too peripheral for the bubbles to come out. I used a soft tip to try to aspirate them out, but there were too many bubbles. I even tried using the infusion line to blow them out, to no avail. So, I performed an air-fluid exchange using the soft-tip cannula to pick out as many bubbles as I could from under the edge of the GRT and got the retina as flat as possible. There was still some shallow fluid in the superior periphery and a plethora of small trapped subretinal PFO bubbles in the area, but I had to call it a day. I trimmed the buckle and closed up, hoping that those tiny bubbles would freeze in place as the retina reattached given that the macula was on the entire time. Immediately after surgery the retina was attached, but at postoperative week 1, I saw that a small PFO bubble had worked its way into the nasal macula. At that point, it wasn’t involving the fovea, so I talked to the patient about the options of either observation or going back in to remove that PFO bubble. He elected to observe, and I felt this was reasonable as I was hesitant to detach the macula in a macula-on case. I monitored him weekly and asked him to nasally position, hoping that the small bubble would work its way nasally and that gravity would pull it down inferiorly.
Luck was not on our side, because at postoperative month 1, he presented with a central circular scotoma. Imaging confirmed that the PFO bubble had migrated subfoveally. So, we went back to the OR for removal of this PFO bubble and possible gas tamponade. In the second vitrectomy, the bubble was in the center of the macula, and there were also trapped bubbles frozen in place under the attached retina in the superior periphery. I used the same 41-gauge beveled cannula as I do for gene therapy, and injected a subretinal bleb of BSS until it incorporated the PFO bubble. You would think it would just float within the bleb once incorporated, but it didn’t. I ended up using my jostling technique, sort of banging on the eye with a scleral depressor, and you can see the PFO bubble drop down towards the superior pocket of the bleb. I then made a draining retinotomy and aspirated it out with the soft tip cannula. I decided to drop a PFO marble on the macula to make sure I got everything out from under the macula through the retinotomy, and then did an air-fluid exchange and applied endolaser around the retinotomy. I tamponaded with 18% SF6, which may have been overkill, but I wanted to make sure everything stayed in place. His retina has remained attached, and he’s about 6 months out now. Unfortunately, there is outer retinal and RPE damage, presumably from where that PFO bubble was sitting, and obviously that’s limited his vision.
Case 4 Discussion
Christina Weng, MD, MBA: For discussion, I rarely use PFO and this is one of the reasons why. Do you have any tips for avoiding retained subretinal PFO? Is there a threshold for when you decide to intervene? And how do you remove it?
Lejla Vajzovic, MD: I do use perfluorocarbon (PFC) quite often in surgery and I think with small gauge surgery, especially valved trocars, the incidence of subretinal PFO has decreased significantly. When I do see it, it’s in cases like this where there was patient movement, or maybe a valve isn’t working well. When I do use it, I tend to use a sandwich technique where I fill up the PFC to the area of retinal tearing and then go to air and sandwich from above vs covering completely and then lasering all the tears and coming back. I find if I do that, there's more chance of PFC hiding anteriorly in the sulcus, or if there's some residual vitreous, smaller bubbles can stay there.
A technique I’ve seen presented is to let PFC just evaporate. So if I'm concerned about some PFC bubbles retained in the eye that are not subretinal, because that wouldn't work in the subretinal location, I'll just add a vent at the end. Then, while we're going to cut the gas or prepare something else, I'll let the eye be air flushed a little bit to allow for any PFC to evaporate.
Yoshihiro Yonekawa, MD: There is a higher risk for PFO bubbles to make their way underneath the retina with large breaks, so it's most common in GRT cases. Your placement of PFO was textbook perfect and what you did nicely was inject the PFO into the PFO rather than dripping it on top, which often causes smaller bubbles to form. Also, if you're doing 23-gauge surgery, you want to use a 25-gauge cannula. If you're doing 25-gauge surgery, use a dual-bore cannula so that you prevent the pressure buildup. If you're fighting resistance as you're putting the PFO in, you're more likely to cause turbulence, and that will create tiny bubbles. It's also good practice to lower the infusion pressure, if you feel like there's any resistance. You also want to make sure that the eye is sealed well. Thank goodness for valved cannulas. Minimize going in and out of the eye when you have PFO in the eye, because more air movement and turbulence causes PFO to slosh around.
As far as removing it, when it's stuck underneath the fovea it’s often because of the surface tension of the thin overlying tissue. When it's farther out in the macula outside of the fovea, the bubble will move around more, and during surgery, you can easily do a subretinal submacular BSS and the bubble will move. What I've done is peel the ILM and suck it out through the retina. I learned from my colleagues in Japan that you can use a soft-tip cannula to aspirate it right through the retina.
Christina Weng, MD, MBA: That transretinal aspiration technique is amazing. How close do you need to get to the retinal surface?
Yoshihiro Yonekawa, MD: Right on the retinal surface.
J. Michael Jumper, MD: What you did for that size of a bubble is a great idea. In fact, a bigger bubble’s better in some ways — to be able to see it, to know where it's moving, and to get it out. PFO under the retina is very similar to exudate in Coats disease. It kind of migrates its way to the fovea due to the stronger pumping effect of the thicker and denser RPE in that area. The PFO can migrate right to the center of the fovea, and it's dastardly. So, when you see PFO anywhere in the subretinal space, you have to watch for migration and do what you did. As far as timing, I might have considered taking the PFO out at the time I noticed it with the retina attached.
Another observation I wanted to make is that it's an awkward maneuver getting PFO in the eye. I inject it into the eye with a 10 mL syringe. I've learned how to really pin my elbow, hold it in one hand, and I put my index finger on the end of the plunger. The problem with that is sometimes getting that initial movement of the fluid out of the cannula can require a great deal of force. Just as you can get the PFO out from under the retina with aspiration, I think you can inject it under the retina with a brisk push on the syringe. When injecting PFO, it is best to start slowly and away from the central macula.
Christina Weng, MD, MBA: Great point. My hands are smaller and just reaching the end of that 10 mL syringe can be difficult. The other thing I teach fellows is when you're injecting PFO—whether it’s with a needle or metal cannula—you need to be careful not to accidentally advance your hand, as this can cause iatrogenic trauma. And stay away from the macula, as you said.
J. Michael Jumper, MD: Cynthia Toth taught me to use the pediatric IV extension tubing because of the smaller bore and lower volume. If you do that, then the surgeon is holding cannula which is separated from the syringe containing the PFO. This means that an assistant or an automated pump on the vitrectomy machine is necessary to infuse the PFO. We all find the technique that works best for us.
Christina Weng, MD, MBA: These were all terrific and unique cases from which we all learned a lot. Thank you all so much for joining.
Hear discussion of this article at retinapodcast.com.
