Introduction

Operative complications can occur at any stage of cataract surgery, anywhere from the administration of local anesthesia right up to the injection of intracameral antibiotics. There are many bear traps and pitfalls along the way that can trip up the inexperienced, unwary, or inattentive surgeon. We need to better understand precisely what those pitfalls are and to anticipate and react early to them. If we do not, then nothing changes. We carry on getting the same complications with the same incidence for the same reasons, ad infinitum. Without wanting to sound too sanctimonious I believe that we need to approach every complication as an opportunity to learn something new. Each event then adds cumulatively to our library of cognitive and technical skills.

This chapter is intended to provide a clear guide to the practical management of the more common and important intraoperative complications and will cover them in sequence through the different stages of phaco surgery.

Local anesthesia

Peribulbar hemorrhage

This can occur with any type of infiltrative anesthesia and is usually fairly immediate but can occasionally be delayed until later during surgery. The orbit becomes tense and incompressible but will usually soften fairly quickly over 10–20 minutes if it is due to the far more common venous bleed. The lids can become tight and difficult to open but again will usually settle down given a little time and some direct orbital pressure. An oculopressive balloon can help limit bleeding and lower the orbital tension. Occasionally the orbit will remain tense, in which case surgery should be abandoned because the risk of chamber collapse, iris prolapse, posterior capsule rupture, and vitreous loss is high. Most patients are relieved that you are putting safety first, once they have understood the risks of continuing. In practice only a small minority (around 5%) of patients with peri- or retrobulbar hemorrhage will need to have their surgery rescheduled1.

Very rarely an orbital arterial bleed may occur. This has happened to the author only once in 20 years of eye surgery. In this case there is rapid and dramatic proptosis accompanied by a rock hard orbit and infiltrative staining of skin and conjunctiva by pressurized blood (Fig. 19.1). Despite intravenous acetazolamide and mannitol plus lateral canthotomy/cantholysis the visual prognosis following arterial orbital hemorrhage is often poor.

Fig. 19.1 Retrobulbar arterial hemorrhage resulting in NPL.

Incisions

Problems with wound construction and architecture are fairly common as a reproducibly ‘perfect’ incision is difficult to achieve. The ‘Rule of Too’s’ describes most of the root causes: too short, too long, too wide, too narrow, too central, too peripheral, too superficial, or too deep.

If it is too short then intraoperative wound leakage with chamber instability will be a problem, as well as an increased risk of iris prolapse and a leaking wound at the end that requires suturing. It will be clear from the start that there is a problem and the best option is to suture the wound and make a better incision at a different site.

A peripheral wound will have breached the conjunctival insertion and is therefore prone to progressive conjunctival ballooning. This leads to fluid pooling on the cornea with resulting poor visibility from distortion and reflections. This is best resolved by extending the conjunctival incision and retracting the back edge of the conjunctiva so that the incisional outflow is no longer sequestered into the subconjunctival space. Firmly massaging already ballooned conjunctiva with a squint hook effectively disperses excess fluid.

Wounds that start peripherally also usually enter the chamber peripherally, so they are additionally prone to iris prolapse. Best to reposit the iris, suture the wound, and make a new incision.

A wound that is too long or too central makes the rhexis difficult and usually undersized as well as eccentric. It also leads to corneal distortion when pointing the phaco tip downwards (so-called ‘roofing’) causing impaired visibility. Such distortion can even tear the back edge of the roof of the incision. A long wound as well as a narrow one will increase the likelihood of a corneal burn, despite modern power modulations such as pulse, burst, and torsional modes. Most significant wound burns result in tissue contraction and wound-gape and need a mattress suture to close them on the table, which may be difficult. The majority will settle over subsequent weeks with conservative management on topical steroids and antibiotic cover. Very rarely scleral patching may be required.

A small dehiscence of Descemet’s membrane related to the internal front edge of the main wound is fairly common. This produces a hinged flap which spontaneously falls back into place and is of no consequence. A larger flap can, on most occasions, be persuaded to lie flat again. If not then an air bubble, SF6 gas, viscoelastic, or rarely suturing have all been used to successfully re-attach hanging flaps. The problem arises when a sizeable flap is stripped off and then completely detaches, never to be seen again. This is rare. The defect is usually triangular in shape with its base the width of the incision and its apex towards the center of the cornea (Fig. 19.2). Nothing to do here except wait and see how well the remaining endothelial cells respond to the greater demand on them. If over one fifth of the endothelium is gone then grafting is likely to be needed. Wait at least 3 months before considering endothelial grafting since slow recovery is the usual course with smaller defects.

Fig. 19.2 Large Descemet’s membrane triangular defect.

When using scleral tunnels caution is needed to avoid cutting the groove too deep otherwise the choroid and suprachoroidal space can be exposed. The sclera should be sutured and a new site used. Scleral tunnels also carry a risk of tunnel hemorrhage, which occurs in up to 10% of such incisions. These are usually obvious at the time of surgery but can be delayed and associated with large postoperative hyphemas. Hemorrhage can be limited by rapid tamponade using viscoelastic. Fastidious hemostasis at the time of surgery is mostly successful in avoiding further problems.

Capsulorrhexis

The correct sizing of the capsulorrhexis is important not only for trapping the anterior edge of the implant optic with an overlapping rim of capsule, but also for avoiding a number of problems and complications. A rhexis that is too small increases the risk of capsular block and possible posterior capsule rupture during hydrodissection. It also is more likely to suffer a secondary radial tear from the phaco tip or second instrument. It should be enlarged at the time and not left to chance. An oversized rhexis is also problematical. The nucleus is likely to tumble out during hydrodissection and also the optic of the IOL will be prone to forward prolapse or decentration from asymmetric capture by the large rhexis. The usual cause of an oversized rhexis is an oversized pupil. We seem to be irresistibly drawn to the edge of the pupil. Nothing you can do about an oversized rhexis except avoid it in the first place.

Anterior capsule tears can be either primary (a tear-out of the rhexis happening as it is being performed) or secondary (when an intact rhexis is subsequently torn). The main practical difference between the two is that with a secondary tear the hydrodissection has already been done and the nucleus is mobile, making life a lot easier. With a primary radial tear-out the first priority is to try and retrieve it before it has gone too far out. This is best achieved using the rhexis rescue maneuver, which comprises firstly overfilling the chamber with a cohesive OVD, unfolding the flap to lie flat on the surface of the nucleus, grasping the free edge of the flap with forceps close to the apex of the tear and pulling it circumferentially backwards and then centrally in order to re-direct the tear inwards4. You may have to make a second stab incision positioned so that you can apply the forceps and exert traction from the most advantageous angle. If this works (and it does more often than not) then your troubles are over. If not then you have to complete the phaco without extending the tear out through the zonules and round to the posterior capsule.

You have four options to choose from, and you must select the one that is safest in your hands:

Hydrodissection

Every single current endocapsular phaco technique requires that the nucleus freely rotates in the capsular bag. The commonest problem encountered with hydrodissection is that it is often incomplete and thereby prevents this free rotation. The most reliable way to achieve a complete subcapsular cortical cleaving hydrodissection, so eloquently first described by Dr Howard Fine, is to make two injections of BSS directly opposite each other. At 3 and 9 o’clock through a superior incision or 6 and 12 o’clock for a temporal incision. Slide the cannula peripherally keeping it tight under the capsule. Tent the capsule slightly forwards then deliver a short but forceful burst of BSS to initiate the wave, followed by a slower more sustained jet to continue the wave. Decompress the capsule block so created, then repeat this on the other side. This is nearly always successful in mobilizing the nucleus – more so than a single injection given diametrically opposite the incision, which is the usual technique seen in practice.

A more worrying complication of hydrodissection happens when there is an unrecognized notch or small tear in the capsulorrhexis. As the bag is inflated and pressurized with fluid this weak point gives way resulting in a rapid radial tear-out, which may indeed extend around to involve the posterior capsule. Use one of the techniques described above to remove the nucleus under these conditions.

The most serious complication of hydrodissection is posterior capsular rupture. The cycle of events leading to this is clearly understood. Firstly capsular block is created by the initial posterior fluid wave. This is associated with a nuclear lift characterized by the lens bulging through the rhexis as it is pushed forwards by the accumulating fluid behind. The capsular block should be decompressed at this stage. It this is not done then continued injection of fluid leads to overinflation of the capsular bag and explosive rupture of the posterior capsule. The sudden rupture of the posterior capsule is accompanied by a rapid deepening of the chamber and narrowing of the pupil as the fluid escapes and the nucleus falls back. This is the so-called pupil ‘snap’ sign that is pathognomonic of PC rupture during hydrodissection6. It happens very quickly and is easily missed or ignored. If you see it then you should stop and decide what best to do. If you are a less experienced surgeon then closing the eye and handing over to an experienced colleague or vitreoretinal surgeon is safest, more so because these cases often involve dense large cataracts and a relatively small rhexis, which makes anterior dislocation and removal of the lens more hazardous. If the pupil snap is unrecognized and the case continues then the nucleus invariably drops into the back of the eye and then there is no choice but to involve the vitroretinal surgeons.

Phacoemulsification

Zonular dialysis

Intraoperative zonular dialysis results from either excessive gouging and pushing with the phaco tip or aspiration and pulling of the peripheral capsule during cortical clean-up, and more rarely from over-zealous insertion of the lens implant. Immediate recognition and prompt action are needed to limit further zonular damage and achieve a good final outcome with a well-centered lens.

Regardless of the stage of surgery in which it happens, the first step is to support the remaining zonules with the insertion of a capsule tension ring (CTR). The vitreous face may or may not have been breached. Use a cohesive OVD to tamponade the defect and minimize the risk of vitreous prolapse. Then inflate the bag and open up the subcapsular space using OVD. Insert the ring in the subcapsular plane in order to minimize trapping of cortical material against the equator of the bag. Insert the leading end of the ring towards the zonular defect to support it whenever practically possible. Using an injector to introduce the ring is a lot easier than freehand bimanual insertion. For the more challenging situations when the bag is either empty or already contains an implant, the ‘fish-tail’ technique is a useful method that avoids the need to linearly dial the ring into the bag (Fig. 19.3). Next check for vitreous prolapse using triamcinolone acetonide and, if present, perform an anterior vitrectomy before continuing. If the zonular defect is diffuse or extensive then additional intraoperative zonular support in the form of capsule anchors can be useful. These effectively stabilize the bag/CTR complex to the sclera and thereby protect the remaining zonules from excessive traction throughout the remainder of the operation.

Fig. 19.3 Fishtail technique for CTR insertion.

For more extreme cases where despite containing a CTR the bag remains de-centered there are two sutured devices available. One is the Cionni modified CTR and the other is the Ahmed sutured capsule tension segment (CTS). Both have eyelets that sit in front of the anterior capsule for suturing to the sclera via the ciliary sulcus. There are still issues regarding the long-term durability of suture material such that many surgeons would opt for a tried and proven angle-supported open-looped anterior chamber implant on these rare occasions. Rather this than risk the initially complex surgery then have to relocate or remove and exchange a dislocated sutured IOL/CTR/bag complex later down the line, particularly in younger patients in whom suture degradation is more likely in their lifetime.

Vitreous loss

The presence of a PC tear or vitreous loss collectively carries with it a significantly increased risk of potentially sight-threatening complications such as retinal detachment, endophthalmitis, cystoid macular edema, suprachoroidal hemorrhage, and glaucoma (UK National Cataract Dataset). However, the impact and potential complications can be mitigated by the use of appropriate vitrectomy techniques and IOL implantation. The desired and wholly achievable result is that the outcome is the same as if vitreous loss had never happened.

The techniques deployed will naturally vary somewhat depending upon the stage of the surgery in which vitreous loss had occurred. The following straightforward menu guide to performing a vitrectomy embraces six key principles and might hopefully be a useful aide memoire:

1 Don’t panic.

2 Don’t deny it.

3 Don’t pull it.

4 Visualize it.

5 Remove it.

6 Follow-up.

These points are dealt with individually below.

Don’t panic

It is so important to remain calm as you need to concentrate and be the one to remain in firm control of what is happening inside the eye as well as in theater. Avoid succumbing to the withdrawal reflex, as mentioned above. Just keep the instruments in the eye with the irrigation running initially and then remove them using the OVD-BSS exchange method described earlier to retain positive pressure to tamponade any further vitreous prolapse.

Don’t deny it

Unfortunately when any disaster strikes our innate tendency is to wish so intensely that it has not really happened that we manage to deny that it has. Vitreous loss is no exception to this. Immediate acceptance that it has happened as soon as it has happened will ensure that the surgeon does not continue to phaco when there is vitreous present. The phaco tip acts as a large aspiration cannula. It does not and never will cut vitreous. But it will certainly pull on it excessively and increase greatly the risk of a retinal tear and detachment through vitreoretinal traction. So that is why if vitreous prolapse is suspected during phaco then the surgeon must immediately accept it then stop both phaco and aspiration. Just leave the irrigation running before completing an OVD-BSS exchange in peace and quiet.

Don’t pull on it

Only very tiny forces are needed to overcome the retinal pigment epithelial (RPE) pump and lift the neurosensory retina off the RPE. The RPE pump exerts a pressure of only 0.27 mmHg. This is so small as to be below the threshold for sensing force feedback. Tearing the retina requires only a little more force so vitrectomy has to be performed whilst exerting as little traction as possible. This means using the highest cut rate your machine can deliver and a low vacuum of about 100–150 mmHg.

Visualize it with triamcinolone

The technique for using triamcinolone acetate suspension (40 mg/ml) to visualize vitreous for anterior vitrectomy was first described by Scott Burke and colleagues in 2003 and is probably one of the most significant advances in surgical technique over the last decade⁷. It is safe and effective and is the only method by which a surgeon can meticulously and confidently remove all prolapsed vitreous from the anterior segment to ensure that none remains to become incarcerated in any of the wounds or occlude the drainage angle at a later date. It is now available commercially in unpreserved formulation from a number of companies so every cataract surgeon should eventually become familiar with this technique as there is none better for visualizing vitreous in order to ensure its complete removal. This minimizes the risk of wound incarceration, vitreous wick, endophthalmitis, and cystoid macular edema.

Remove it

If the nucleus has dropped or is dropping then it is best left to your vitreoretinal colleagues to remove it. All attempts and techniques available to the anterior segment surgeon to retrieve a falling nucleus are a compromise and second best to a controlled pars plana approach by an experienced vitreoretinal surgeon.

If there are nuclear fragments remaining in the anterior segment then any surrounding vitreous needs to be removed first under triamcinolone guidance before attempting their extraction from the eye. The technique of ‘dry’ vitrectomy is useful for this purpose. There is no irrigation in the eye in order to avoid excessive vitreous hydration and further prolapse. The cutter is used through one side-port and simultaneously OVD is injected through a second side-port to replace the volume of fluid aspirated through the cutter (Fig. 19.4). Once the surrounding vitreous has been cleared then the incision should be enlarged to avoid undue force and distortion of the globe during removal of the fragments. To use phaco for their removal has a high risk of vitreous aspiration and vitreoretinal traction.

Fig. 19.4 Triamcinolone guided dry anterior vitrectomy with engulfed nuclear fragment.

Otherwise any residual vitreous is removed using bimanual anterior vitrectomy with the irrigation on, but reduced bottle height as needed to maintain positive pressure, and the foot-pedal set to cut-aspiration sequence for positions two and three. Keep the cutter in the center of the eye away from the edge of the pupil and the rhexis as these both need to remain intact. Once vitreous has been removed from the anterior chamber move the cutter posteriorly through the center of the rhexis to complete a core vitrectomy; this avoids dragging vitreous forwards into the chamber and putting traction on the vitreous base.

Once this core vitrectomy is complete then an appropriately sized and powered lens implant can be placed in the sulcus and the optic captured in the rhexis. A final flush with diluted triamcinolone (fourfold or more) will identify any lingering strands of vitreous, which are fairly rare.

Follow-up

These patients need to have their peripheral retina indented and examined by an experienced clinician early on following surgery. The relative risk of retinal detachment after vitreous loss is around 15–20 times that of an uncomplicated phaco procedure. Cystoid macular edema, endophthalmitis, and glaucoma are also more common so these patients need to be watched vigilantly.

Suprachoroidal hemorrhage

Suprachoroidal hemorrhage is by a long way the most feared and potentially devastating intraoperative complication, but fortunately it is fairly rare at around 0.1%. The good news is that with raised awareness of its early clinical features in a semi-closed eye, as is the situation during phaco, we are hopefully becoming better at rapidly responding by halting the operation and closing the eye. By doing so the hemorrhage is limited as it self-tamponades, therefore the outcome can be excellent. If we do not recognize and respond to the signs then this complication can evolve fairly rapidly into a catastrophe.

Suprachoroidal hemorrhage can occur at any stage during phaco from shortly after the incision is made up until the lens is being implanted. The commonest timing by far is after the nucleus has been removed, just before implantation of the IOL, when the eye is at its softest and has been manipulated during the preceding surgery. This contributes to micro-shearing between the choroid and overlying sclera, which traumatizes and stretches the perforating long posterior ciliary vessels where they bridge the suprachoroidal space near the equator.

Most bleeds are venous and develop over many minutes to produce aggressive chamber shallowing. This is progressively more difficult to overcome so that even a high molecular weight cohesive OVD may fail to deepen the chamber and be spontaneously expelled from the eye. So at the first sign of such a rapid change in chamber stability think of suprachoroidal hemorrhage because there is nothing that can produce quite such acute and progressive signs. At the moment that the thought arises just withdraw all instruments from the eye and seal the wounds. Then have a look with the binocular indirect ophthalmoscope on the operating table. If there is a reasonable view then you may well see a localized subretinal hemorrhage in the mid-periphery. Never attempt to drain a suprachoroidal at the time as this can accelerate you along the pathway to catastrophe.

If the view is poor then abandon the surgery and get a B-scan ultrasound done promptly. This will reveal any significant hemorrhage (Fig. 19.5).

Fig. 19.5 Suprachoroidal hemorrhage B-scan at 24 hours.

The blood will clot and liquefy within 7–10 days and the smaller bleeds will resorb. At this stage the surgery can be safely completed and an IOL implanted as needed. Larger hemorrhages can be drained after such time has elapsed, but if the blood has tracked under the macula then the visual prognosis is predictably poor.

Summary

It is easy to forget that modern cataract surgery is a very safe and successful procedure with an uncomplicated outcome in greater than 95% of cases8. However we still encounter complications that are potentially sight-threatening, but nearly all of these are thankfully very rare and, more importantly, avoidable.

By carefully planning cases and anticipating predictable intraoperative problems we can reduce the risk of most of the significant complications. For those that we run into, we need to be quick to recognize the early signs and react to them promptly without panicking.

By doing so and then managing them systematically we should be able to achieve results that are as good as if they had never happened.