Criteria for Selection of Cases

A computerized search was performed to identify all dogs that underwent phacoemulsification surgery at Carolina Veterinary Specialists in Greensboro, NC, from December 1995 to March 2002. Dogs were required to have a minimum of 3 months postoperative follow-up and no history of glaucoma to be included in the study.

Procedures

Information recorded for each operated eye included signalment (breed, age at the time of surgery, and sex), eye operated (right vs left), age at the onset of the cataract (congenital, juvenile [1 to 4 years], adult [> 4 to 8 years], or senile [> 8 years]), stage of cataract (immature, mature, or hypermature), cause of the cataract (ie, trauma or diabetes mellitus), presence of lens-induced uveitis (yes or no), preexisting ocular complications, duration of follow-up, and postoperative complications. All preoperative cataract evaluations consisted of a complete ophthalmic examination by a diplomate of the American College of Veterinary Ophthalmologists by use of slit-lamp biomicroscopy^a and applanation tonometry.^b Pupillary dilation was induced with 1% tropicamide ophthalmic solution, and the fundus was examined by use of indirect ophthalmoscopy.^c The presence of aqueous flare was determined and graded as mild (+1), moderate (+2), or severe (≥ +3).¹³ A neuroophthalmic examination was performed to assess menace responses and dazzle reflexes. Clients were questioned regarding the dog's loss of vision, and an assessment was made to correlate the severity of vision loss to the completeness of the cataract. Bright-flash electroretinography^d was performed on all eyes in which the fundus could not be evaluated by use of indirect ophthalmoscopy. If retinal detachment was suspected, ocular ultrasonography^e was performed. Gonioscopy^f was performed only if a risk of primary glaucoma was anticipated on the basis of breed, a narrowed anterior chamber, or high intraocular pressure at preoperative screening. Dogs with preoperative lens-induced uveitis were treated with 1% topically administered prednisolone acetate if no corneal ulcer was present. If a dog had a corneal ulcer, surgery was postponed until the ulcer healed. The necessity of a preoperative CBC and serum biochemical profile was based on the age (> 5 years) and general health status of the dog. For all dogs, measurement of PCV and serum total protein concentration and assessment of renal function were performed prior to surgery. Dogs with diabetes mellitus were assessed via recently determined glucose concentration curves or fructosamine concentrations for adequate diabetic regulation prior to surgery.

If all preoperative variables were within reference ranges, phacoemulsification surgery was performed as described.^2,4,14 Preoperative treatment was instituted approximately 2 hours prior to surgery and consisted of topically administered atropine (1%), phenylephrine (10%), neomycin-polymyxin B-dexamethasone (0.1%), and flurbiprofen (0.3%). Perioperative injections of cefazolin (22 mg/kg [10 mg/lb], IV) and flunixin meglumine (0.5 mg/kg [0.23 mg/lb], IV) were administered at the time of surgical preparation. All eyes received intraocular lens implants^g unless an intraoperative complicating factor precluded placement of the lens (3 eyes: 1 eye with uveitis-induced miosis, 1 eye that required a large anterior capsulectomy, and 1 eye with a large posterior capsule tear requiring vitrectomy). Viscoelastic material was routinely removed. Each surgery was performed by the same surgeon (MPN).

Immediate postoperative medical treatment included an antimicrobial (either orally administered cephalexin or orally administered amoxicillin-clavulanic acid), topical administration of 1% prednisolone acetate every 6 hours, topical administration of 1% tropicamide every 8 hours, and topical administration of antimicrobial (either tobramycin or neomycinpolymyxin B–gramicidin) every 8 hours. Dogs were routinely sent home in the evening of the day of surgery, and reexaminations were planned for the next day and at 1 week, 3 weeks, 7 weeks, and 12 weeks after surgery. Additional and more frequent recheck examinations were scheduled as necessitated by development of complications. The frequency of topical administration of medications was gradually decreased on the basis of the presence and magnitude of uveitis, pupillary dilatation, and wound healing. As a general rule, administration of either 1% prednisolone acetate or a topically administered nonsteroidal anti-inflammatory medication was continued for a minimum of 3 months after surgery. Dogs with postoperative intraocular pressure spikes were treated with an orally administered carbonic anhydrase inhibitor (methazolamide) and a topically administered prostaglandin analog (latanoprost) as needed. Infrequently, an additional surgery was performed to repair broken sutures secondary to self-trauma or wound dehiscence.

Statistical analysis—A Cox proportional hazards model was estimated by use of a software program^h to obtain RRs of 2 outcomes: postoperative development of glaucoma and failure (blindness, enucleation, or evisceration). Risk factors considered included age at surgery (< 4 years, 4 to 8 years, and > 12 years [> 8 to 12 years was considered the reference group]) and sex (female, male, and neutered male [spayed female was considered the reference group]). Also included were cataract hypermaturity, lens-induced uveitis, and diabetes mellitus. To control for breed, indicators were introduced for those breeds in which at least 6 eyes were observed and at least 1 of those eyes had failure or developed glaucoma, depending on the outcome examined. These included Bichon Frise, Boston Terrier, Cocker Spaniel, Dachshund, Miniature Poodle, Miniature Schnauzer, mixed breed, Cocker Spaniel–Poodle cross (Cockapoo), and Shih Tzu. Yorkshire Terriers were also included in the models for glaucoma. All other breeds formed a reference group. Each risk factor was evaluated separately, and risk factors that were significant at P < 0.10 were used in backward stepwise selection (P < 0.05) to identify the best parsimonious multivariate model. In all models, a random effect frailty parameter was introduced to account for the correlation of the errors when dogs had surgery performed on both eyes.

Results

Medical records of 247 dogs with 420 operated eyes were reviewed. Seventy-five dogs were excluded from the study because of inadequate follow-up. The final study population was 172 dogs (290 eyes). Grouping of postoperative follow-up periods was as follows: 3 months, 290 eyes; > 3 to 6 months, 259 eyes; > 6 months to 1 year, 200 eyes; > 1 to 2 years, 132 eyes; > 2 to 3 years, 80 eyes; > 3 to 4 years, 39 eyes; and > 4 years, 17 eyes.

The most common breeds were Miniature Poodle (22/172 dogs, 35/290 eyes), Cocker Spaniel (19/172 dogs, 29/290 eyes), Boston Terrier (16/172 dogs, 29/290 eyes), Miniature Schnauzer (13/172 dogs, 23/290 eyes), Labrador Retriever (11/172 dogs, 18/290 eyes), and Bichon Frise (11/172 dogs, 17/290 eyes). Median age at surgery was 8 years (mean, 7.79 years), and median duration of follow-up was 0.88 years (mean, 1.37 years) with a range of 3 months to 4.75 years. The study population comprised 12 sexually intact females, 83 spayed females, 15 sexually intact males, and 62 neutered males. One hundred twelve bilateral surgeries and 66 unilateral surgeries were performed on 138 right eyes and 152 left eyes. Surgery was performed on 62 hypermature cataracts, preoperative lens-induced uveitis was detected in 86 of 290 (30%) eyes, and dogs with diabetes mellitus had 85 operated eyes.

The most common complication after surgery was posterior capsule opacification, which was graded subjectively as mild (transparency was decreased but the fundic examination was unobstructed), moderate (transparency was decreased and the fundic examination was partially obstructed), and severe (transparency was decreased and the fundic examination was completely obstructed; Table 1). Prevalence of retinal detachment as a postoperative complication was low as well. The incidence (ie, number of new cases) of retinal detachment was low at 1 of 290 (0.34%) at 3 months, 2 of 259 (0.7%) at 3 to 6 months, 2 of 200 (1.0%) at 6 months to 1 year, 1 of 132 (0.75%) at 1 to 2 years, and 1 of 80 (1.25%) at 2 to 3 years.

Glaucoma was defined as increased intraocular pressure and intraocular changes that persisted beyond the immediate postoperative period. Of 290 eyes in the study, 36 (12.41%) developed glaucoma after surgery. Breed was a significant risk factor for the development of glaucoma after surgery. Relative to the reference group, Boston Terriers (P = 0.007; RR, 5.35; 95% CI, 1.59 to 17.90), Cocker Spaniels (P = 0.007; RR, 5.29; 95% CI, 1.57 to 17.78), Dachshunds (P = 0.029; RR, 5.14; 95% CI, 1.18 to 22.41), Cockapoos (P = 0.008; RR, 5.87; 95% CI, 1.59 to 21.69), and Shih Tzus (P = 0.001; RR, 8.67; 95% CI, 2.30 to 32.59) were more likely to develop glaucoma after surgery. Neutered male dogs were less likely to develop glaucoma after surgery than the reference group (P = 0.022; RR, 0.339; 95% CI, 0.13 to 0.86). Cataract hypermaturity increased the risk of developing glaucoma (P = 0.032; RR, 2.08; 95% CI, 1.07 to 4.06), although no increased risk was associated with lens-induced uveitis or diabetes mellitus. In the multivariate model, only 4 breeds—Cockapoos, Boston Terriers, Cocker Spaniels, and Shih Tzus—were retained as significantly and positively associated with the development of glaucoma. All other variables were eliminated from the model because of P > 0.05.

There were few significant risk factors associated with failure. Compared with other breeds, Cockapoos (P < 0.001; RR, 8.29; 95% CI, 2.83 to 24.26) and Shih Tzus (P = 0.001; RR, 8.77; 95% CI, 2.30 to 33.46) were more likely to have failure. Bichon Frise had RR of failure of 5.19 (95% CI, 0.82 to 33.10), but this was not significant (P = 0.081). Compared with older and younger dogs, those 4 to 8 years old at the time of surgery were more likely to have failure (RR, 2.55; 95% CI, 0.88 to 5.99), but this was not significant (P = 0.081). Cataract hypermaturity was significantly (P = 0.008; RR, 2.49; 95% CI, 1.26 to 4.89) associated with increased risk of failure, but lens-induced uveitis and diabetes mellitus were not. In the multivariate model, only 2 breeds—Shih Tzus and Cockapoos—were retained as significant risk factors.

Discussion

Phacoemulsification has become the technique of choice for cataract removal in veterinary ophthalmology. Despite currently reported short-term success rates exceeding 90%,^6,9 a plethora of complications may arise in the immediate and long-term postoperative periods. Uveitis, postoperative ocular hypertension, glaucoma, retinal detachment, endophthalmitis, corneal ulcers, corneal edema, and wound dehiscence are the most important concerns.^6,15–18 Persistent uveitis may result in the formation of pre-iridal fibrovascular membranes, glaucoma, retinal detachment, posterior capsule opacification, and permanent corneal edema, precluding long-term success.^15,19 Glaucoma and retinal detachments may result in permanently blind and often painful eyes, necessitating salvage procedures such as enucleation or evisceration.¹⁵

By far, posterior capsule opacification was the most common complication encountered in the present study. By the 1- to 2-year evaluation period, 69% of the eyes evaluated had either mild or moderate posterior capsule opacification, which is higher than previously reported (22%).⁶ Posterior capsule opacification is caused by proliferation and migration of residual lens epithelial cells. Histologically, both cuboidal anterior epithelial cells and equatorial lens bow cells are present, although equatorial cells are more likely to migrate and form lenticular pearls.²⁰ The resultant fibrosis and wrinkling of the posterior lens capsule compromise the clarity of the visual axis.²¹

The high prevalence of posterior capsule opacification found in our study population emphasizes the importance of developing new ways to treat dogs with posterior capsule opacification. Although phacoemulsification with capsule vacuuming results in a decrease in residual cell density, posterior capsule opacification is not prevented unless almost all of these cells are removed.²² Growth factors such as transferrin, transforming growth factor β-2, and basic fibroblast growth factor have been linked to development of posterior capsule opacification in dogs.^23–25 In humans, attempts to suppress growth factors and chemically ablate or suppress lens epithelial cells have been limited by problems with toxicosis.²⁶ Humans with posterior capsule opacification are presently treated via laser capsulotomy, although this technique has not proven successful in dogs. Improvements in intraocular lens materials and construction are presently thought to offer the best hope of decreasing posterior capsule opacification and have led to a substantial decrease in posterior capsule opacification in humans.^27–29

Retinal detachment was an infrequent complication in our study, with a prevalence of only 0% to 1.25%, which is lower than previously reported (4.7% of eyes).⁶ Six of 172 (3%) dogs and 8 of 290 (2.7%) eyes developed retinal detachments in our study. These numbers were too low to evaluate statistically, although it was noteworthy that none of the dogs with retinal detachment were Bichon Frise. Bichon Frise was one of the more common breeds in the study (11/172 dogs and 17/290 eyes) and has been reported to be at increased risk of retinal detachment after surgery.³⁰ In a recent retrospective study, prophylactic retinopexy in Bichon Frise was recommended because of higher prevalence of rhegmatogenous retinal detachment in Bichon Frise without retinopexy, regardless of whether or not phacoemulsification was performed.³¹ It is unclear why retinal detachment was not found in any of the Bichon Frise in our study. It is possible there may be different subpopulations with different risks. Alternatively, the low prevalence of retinal detachments in our study in general may have been related to patient selection or few intraoperative variables rather than breed or cataract-related factors.

The incidence of postoperative glaucoma in our study was less than has been reported (16.8%).¹⁶ Prevalence in our study remained < 5.0% until the 1- to 2-year follow-up period and even then remained < 10% until a large number of dogs were lost to follow-up. Even so, at the 2- to 3-year and 3- to 4-year intervals, prevalence (17.5% and 25%, respectively) still remained less than what has been reported at 12 months (28.8%).¹⁶ Actual prevalence may have been even lower because the population was skewed by the fact that dogs are much more likely to be returned for long-term follow-up examinations if a problem arises. Intraocular lens placement was previously cited as a factor that lowered risk of postoperative glaucoma,¹⁶ although it was unclear if risk was reduced by the lens or by the absence of intraoperative complications that prohibit lens placement. Because an intraocular lens was not placed in only 3 of 290 eyes in our study, we were unable to evaluate intraocular lens placement as a risk factor for glaucoma or failure. Previous studies used records from teaching institutions where multiple surgeons with various levels of experience perform phacoemulsification. Because all surgeries in this study were performed by an American College of Veterinary Ophthalmologists diplomate with extensive surgical experience, it is possible that the lower prevalence of glaucoma was related to increased frequency of intraocular lens placement or the reduction in surgical trauma and complications that permitted frequent intraocular lens placement.

Previous retrospective studies^16,19 have examined the incidence and risk factors associated with postoperative glaucoma in both the short- and long-term periods. In 1 study,¹⁹ Boston Terriers, dogs with intraoperative intraocular hemorrhage, and dogs with preoperative uveal or retinal abnormalities were at high risk for development of glaucoma in the short- and long-term follow-up periods. In accordance with other studies, we found Boston Terriers to have significantly increased risk for glaucoma after surgery, compared with other breeds. In our study, Cockapoos, Shih Tzus, and Cocker Spaniels also had significantly increased risk. Previous studies have not reported these breeds as having increased risk of postoperative glaucoma. Many theories have been proposed to explain these findings in Boston Terriers, including an inherited predisposition to primary glaucoma, brachycephalic conformation, decreased corneal sensitivity, active lifestyle, and increased rates of ocular trauma. Many of the same factors could explain the increased risk in Shih Tzus, Cockapoos, and Cocker Spaniels.

Glaucoma may develop after surgery as a result of an inherited predisposition (ie, pectinate ligament dysplasia), aqueous misdirection, or ongoing uveitis resulting in pre-iridal fibrovascular membrane formation, anterior or posterior synechia, or pupillary block. Anterior vitreous prolapse secondary to posterior capsule tears may also result in pupillary block and glaucoma. Gonioscopy was not performed routinely prior to phacoemulsification in this study, so we were unable to assess whether pectinate ligament dysplasia was a contributing factor to the development of postoperative glaucoma. A recent histologic study¹⁵ of eyes enucleated or eviscerated because of complications following phacoemulsification surgery found concurrent evidence of peripheral anterior synechiae, iris bombé, or goniodysgenesis in 37 of 44 eyes with glaucoma. A previous histologic study³² found similar abnormalities.

Cataract hypermaturity has been identified as an important risk factor for development of postoperative glaucoma,¹⁶ and results of our study concurred with those findings. With hypermaturity, lens proteins are resorbed and leak across the lens capsule, resulting in a loss of lens volume and stimulation of the intraocular inflammatory response. As a result, the ocular immune system is primed before surgery. Weakening and instability of the zonular fibers, as well as vitreal degeneration, often accompany hypermaturity. These degenerative changes may increase the likelihood of surgical complications and retinal detachment. In addition, although the exact pathogenesis of pre-iridal fibrovascular membrane formation is still unclear, it is conceivable the factors inciting pre-iridal fibrovascular membrane formation may already be present prior to surgery, increasing the likelihood of angle closure and postoperative glaucoma.^33,34 Other potential risk factors such as preoperative lens-induced uveitis and diabetes mellitus did not result in significantly greater risk. This finding is in agreement with some previous studies.^11,19

Short-term failure rate (at 3 months) was low at 1.4% and lower than previously reported.^6,9 Our study did not, however, include dogs with < 3 months of follow-up, so some dogs with severe complications requiring early enucleation would have been excluded. Failure increased with time, although it remained < 8.0% through the 2- to 3-year evaluation period. These values are lower than that reported at 22.6 weeks (14.4%).⁶ Breed was the strongest factor associated with failure, with Cockapoos and Shih Tzus having increased risk. Previous studies^6,9 evaluating success rates did not evaluate breed as a factor, so comparisons were not possible. Eyes with hypermature cataracts also had increased risk of failure in our study. Similarly, in a previous study,⁶ it was reported that a higher percentage of favorable outcomes result when immature cataracts are removed.

Endophthalmitis is generally considered an acute and devastating complication of phacoemulsification and often develops 4 to 7 days after surgery, invariably resulting in blindness. In 1 study,¹⁵ suppurative endophthalmitis was diagnosed histologically in 17 of 21 eyes enucleated in the first 3 months after surgery. In our study, a small percentage of eyes developed endophthalmitis (1.4%), although this was higher than recently reported in humans (0.18%).³⁵ In studies³⁶ in humans, a decrease in the incidence of endophthalmitis during a 20-year period when surgeons converted to extracapsular from intracapsular extraction is reported. A decrease was not detected during a 12-year period when surgeons converted to phacoemulsification.³⁵ Reasons remain unclear, although refinements in surgical technique may not have been long-standing enough to result in a decrease at this time.

As with most retrospective case series, the number of dogs with long-term follow-up decreased severely after 1 to 2 years. Establishing a true measure of longterm success is often hindered by the fact that dogs with ongoing problems or complications will be more likely to be reevaluated for care. In addition, when operating on older patients, the dropout rate may be influenced by patient deaths. For instance, in our study, 13 dogs with 20 operated eyes died during the first 2 years. The actual number of patients that died during the study period may have been larger because all medical records were obtained from the referral institution and not from the primary care veterinarian.

Some of the other problems traditionally associated with retrospective studies were minimized. All surgeries and evaluations were performed by the same American College of Veterinary Ophthalmologists diplomate. As a result, surgical technique, surgical time, and pre- and postoperative treatments did not vary substantially.

One of the most alarming findings in this study was the number of dogs excluded because of inadequate follow-up. Seventy-five dogs had phacoemulsification performed during the study period but had < 3 months of follow-up and were thus excluded from statistical analysis. At this clinic, routine follow-up visits are included in the price of the surgery, so lack of compliance was not caused by cost. Considering the fact that painful complications such as glaucoma, retinal detachments, and protracted uveitis occur even years after surgery, it is imperative that owners be counseled regarding the need for lifelong monitoring.