ANN M. CARRASQUILLO and DEBRA A. GOLDSTEIN
Table Of Contents
INFECTIOUS POSTOPERATIVE UVEITIS
NONINFECTIOUS POSTOPERATIVE UVEITIS
|Postoperative inflammation can result from various factors including exacerbation of underlying uveitis, retained intraocular foreign material after surgery, retained lens material, intraocular lens (IOL) chafing, phacoanaphylactic endophthalmitis, sympathetic ophthalmia, and infection. Differentiating noninfectious from infectious endophthalmitis can be difficult. Infectious inflammation may occur concurrently with noninfectious inflammation, making diagnosis even more challenging. If significant inflammation exists, especially in the presence of a hypopyon, possible infectious causes must be investigated. After microbial infection has been ruled out, other causes of postoperative uveitis must be considered.|
|Surgical manipulation of the eye causes breakdown of the blood-aqueous
barrier resulting in protein leakage and cellular reaction in the aqueous
humor, manifested clinically as flare and cell. Inflammation after
surgery is normal. However, what constitutes normal inflammation and
what is considered out of the ordinary inflammation may be unclear. Knowing
patient and surgical characteristics that may predispose individuals
to increased inflammation following surgery is, thus, important for
postoperative management and diagnosis.|
Using anterior segment fluorophotometry, Sanders and associates demonstrated that large limbal incisions caused significant blood-aqueous barrier breakdown.1 Clinically, phacoemulsification has been observed to cause less postoperative inflammation than extracapsular cataract extraction.2Oshika and colleagues compared patients undergoing surgery by planned 11-mm incision extracapsular cataract extraction to patients receiving phacoemulsification with nonfoldable IOLs (7-mm incision) and foldable IOLs (4-mm incision) and found that both aqueous flare and cell count measured with the laser flare cell meter were highest postoperatively in the 11-mm incision group, with less inflammation noted in decreasing order by the 7- and 4-mm incision groups.3 In addition, those authors reported that flare intensity in all incision groups remained significantly higher than those of age-matched normal controls for up to 6 months following surgery.3 However, Corbett and partners reported no difference in postoperative inflammation between extracapsular cataract extraction through a 12-mm incision or phacoemulsification with a 7-mm incision (small incision phacoemulsification was not included).4
Other factors that may increase normal postoperative inflammation include a history of previous intraocular surgery, greater technical difficulty of the operation, nonwhite race and brown iris pigmentation.4 Intraoperative factors such as iris manipulation, anterior vitrectomy, prolonged surgical time, and toxic reactions to irrigating solutions; viscoelastic products or polishing agents on IOLs may also cause greater blood-aqueous barrier breakdown and leakage leading to more pronounced postoperative inflammation.5–9
|INFECTIOUS POSTOPERATIVE UVEITIS|
EARLY BACTERIAL ENDOPHTHALMITIS
Despite the low frequency with which this entity is encountered, bacterial endophthalmitis remains one of the most dreaded and destructive complications following intraocular surgery. The combination of insults from the invading bacteria and the patient's own immune system may produce significant destruction leaving the eye with a poor visual prognosis. Prompt diagnosis and treatment are essential for achieving improved visual and functional potential. A key factor to early diagnosis is a high suspicion of endophthalmitis in postoperative cases with greater inflammation or pain than expected.
The incidence of bacterial endophthalmitis has decreased significantly over the last century; unfortunately, it has not been reduced to zero. In the late 1800s, the incidence of bacterial endophthalmitis was reported to be approximately 10%.10 By the early to mid 1900s, the incidence had dropped to 0.58%11–17 and by the mid 1980s, such incidence had been reduced even further to 0.074%.18 One series published by Kattan and associates in 1990 reviewed 30,003 surgical cases performed at a single teaching hospital between 1984 and 1989.15 They found the incidence of culture-proven acute postoperative endophthalmitis to be 0.072% following extracapsular cataract extraction. Fisch and coworkers reported the incidence of culture proven postoperative endophthalmitis to be 0.31% following all penetrating ocular surgery and 0.32% after cataract surgery.19 In 1991, Javitt and partners reported the results of postoperative endophthalmitis in a nationally representative sample by reviewing all U.S. Medicare beneficiaries undergoing cataract extraction in 1984.20 Among the 338,141 patients analyzed, the risk of endophthalmitis within one year of surgery was 0.17% for intracapsular cataract extraction (ICCE) and 0.12% for extracapsular cataract extraction (ECCE). In 1996, Norregaard and associates reported similar results from a cross national comparison using data from the National Danish administrative hospital register.21 Due to the nature of broad national record collection in these studies, however; not all cases were verified by culture. In 1996, Aaberg and coworkers reported the results of a 10-year review of the incidence of culture-proven acute-onset endophthalmitis.18 They found overall incidence during this period to be 0.093%. As expected, incidence varied by surgical type. Cataract surgery with or without IOL placement had an incidence of 0.082%, pars plana vitrectomy (PPV) 0.046%, penetrating keratoplasty 0.178%, secondary IOL 0.366% and combined cataract extraction and penetrating keratoplasty 0.194%.1
Most studies evaluating the incidence of postoperative endophthalmitis have looked at patients undergoing surgery as inpatients or at major teaching hospitals. As more surgery is being performed at outpatient surgery centers, studies comparing the rates of endophthalmitis in the outpatient versus inpatient setting are of paramount importance. By 1984, about half of all cataract surgery performed on Medicare recipients was completed in an outpatient care setting.22 Between 1983 and 1984, Medicare implemented policy requiring that its recipients undergo outpatient cataract surgery except in special instances. In 1992, Javitt and coworkers published data examining the rates of postoperative endophthalmitis encountered in an outpatient versus inpatient setting.22 By comparing results from their survey of Medicare recipients undergoing cataract surgery in 1984 (before changes in Medicare reimbursement) to a cohort of patients undergoingsurgery between 1986 and 1987, they found a statistically significant decrease in the rate of endophthalmitis. At 1-month postoperative, the risk of endophthalmitis was 0.055% in the group undergoing outpatient surgery compared with 0.085% in the inpatient group. At 1 year, the risks were 0.081% and 0.12%, respectively.22 The two cohorts of patients evaluated in the study, however, were not entirely comparable. Antibiotic prophylaxis and operative technique including the proportion of cases done by small incision versus large incision surgery were not specified and may limit the clinical significance of the results.
What causes some patients to develop endophthalmitis is unclear. Factors such as inoculum size, virulence of bacterial strain, intraoperative complications, and deficits in local or systemic host defenses likely play a role.23
There is significant evidence supporting the theory that bacterial entry through the wound site during surgery is the source for organisms that produce infectious postoperative endophthalmitis.24,25 In 1989, Sherwood and coworkers cultured fluid aspirated from the anterior chamber and drained from the conjunctival sac during cataract surgery in 101 patients.24 They demonstrated bacterial growth by culture of aspirate fluid from 29 patients and from the conjunctival sac of 90 patients.24 In addition, in all patients, conjunctival fluid stained with fluorescein flowed into the anterior chamber during irrigation and aspiration performed during extracapsular cataract surgery. They demonstrated that this fluid is contaminated with potentially pathogenic organisms, yet none of the patients in their study developed endophthalmitis.
Menikoff and associates conducted a case-control study of 24,105 patients undergoing intraocular surgery between 1988 and 1990 to identify risk factors for postoperative endophthalmitis.6 One factor identified as a significant independent risk for the development of endophthalmitis was intraoperative communication with the vitreous cavity.6 This complication was associated with a risk ratio of 13:7.6 Likewise, in their study of Medicare beneficiaries undergoing cataract extraction, Javitt and colleagues found that surgery accompanied by anterior vitrectomy was associated with more than a fourfold increase in the development of postoperative endophthalmitis at 1 month compared with patients undergoing cataract surgery alone.20 Several other studies have also found a higher incidence of endophthalmitis after vitreous communication has been introduced.7,21,23,26,27
Vitreous communication seems to be important in the development of postoperative endophthalmitis for several reasons. Removal of potentially pathogenic bacteria from the vitreous cavity appears to be less efficient than from the anterior chamber.9,23,28 Vitreous cultures have been demonstrated to yield a significantly higher percentage of confirmed growth and higher colony counts compared with aqueous.15,29–31 It has been suggested that vitreous may be more supportive of bacterial growth than aqueous fluid and that the viscous properties of vitreous may retard bacterial clearance.7,23,28,31
ICCE has also been associated with a higher risk of postoperative endophthalmitis compared with findings involving ECCE.20,21 Javitt and coworkers found a 0.11% risk for endophthalmitis 1 month after ICCE compared with a 0.085% risk following ECCE or phacoemulsification.20 This factor may be related to the vitreous communication allowed with intracapsular surgery.
Certain types of IOLs may also be associated with an increased risk of infectious endophthalmitis. Sherwood and associates showed that fluid bathing the conjunctiva during surgery is contaminated with potentially virulent bacteria.24 IOLs often touch various aspects of the ocular adnexa during lens insertion. Vafidis and coworkers demonstrated that factors such as electrostatic charges may cause bacteria to adhere to IOLs,32 which allows bacteria to be carried into the eye with the lens or to adhere after IOL insertion when contaminated irrigation fluid flows through the anterior chamber.24 Several investigators have shown that bacteria adhere preferentially to lenses with haptics constructed of polypropylene compared with polymethylmethacrylate.6,33 Similarly, transscleral fixation of posterior chamber IOLs with exposed polypropylene sutures has also been associated with an increased risk of infectious endophthalmitis.6
Numerous other factors have also been identified as increasing the risk of postoperative endophthalmitis. Increased intraoperative instrumentation and duration of the operative procedure seem to place patients at increased risk.6,7,9 This finding may be partly attributable to complicated cases that require anterior vitrectomy, a factor that is independently associated with an increased risk of infection. Menikoff and colleagues found patients with a history of drug allergy to be at significantly increased risk to develop endophthalmitis.6 The authors hypothesized that these patients may have a form of atopy, which is often associated with denser bacterial colonization of the adnexa and an altered immune response. Several studies have implicated diabetes mellitus with a greater risk of infectious endophthalmitis.15,19,26,34 This finding may be secondary to an impaired immune system in many of these patients. Similarly, advanced age has been associated with a greater risk of endophthalmitis. As in patients with diabetes, this finding may be explained by a relative diminution in host defenses.21
A key factor in the prognosis of early postoperative endophthalmitis is prompt diagnosis and treatment. Although the ophthalmologist rarely encounters this entity, it must be kept in mind in the differential diagnosis of a postoperative patient with excessive inflammation or pain. Early postoperative endophthalmitis usually presents within the first few days after surgery. In their 10-year review of the incidence and outcomes of cataract surgery at a single institution, Aaberg and associates reported a range of 1 to 40 days, with a median of 5 days, from the initial procedure to the development of acute postoperative endophthalmitis.18 Kattan and coworkers reported similar findings in 1990, with a range from 1 to 40 days and an average of 8 days.15 Results from the Endophthalmitis Vitrectomy study, a multicentered randomized prospective clinical trial of 420 patients at 24 centers with clinical evidence of postoperative endophthalmitis following cataract extraction or secondary IOL implantation, reported a median time of 6 days until patient presentation to a study center.34 Of these patients, 24% presented within 3 days of the initiating procedure, 37% within 4 to 7 days, 17% within 8 to 13 days, and the remaining 22% presented within 2 to 6 weeks.34
Common findings include pain, conjunctival injection, lid edema, and decreased vision. Signs and symptoms are variable, however, and they often depend on the virulence of the underlying organism.23 In the Endophthalmitis Vitrectomy study, blurred vision was the most common symptom.34 In that study, 86% of patients presented with a visual acuity of less than 5/200. Pain was reported in 74% of patients.34 Other signs include hypopyon, inflammatory pupillary membrane, vitreous opacification and inflammation, anterior chamber cell and flare, and a diminished red reflex23,34,35 (Fig. 1). Less common findings may include an afferent papillary defect and corneal ulcer or infiltrate.34 Again, these findings vary and likely depend on the underlying microorganism as well as the individual's immune response.
It is generally agreed that the greatest source of bacteria causing postoperative endophthalmitis originates from the patient's own eyelid and adnexal flora.15,19,36–38 The ocular adnexa has been shown to be densely colonized with organisms commonly implicated in the etiology of postoperative endophthalmitis.36,39–41 Ariyasu and associates demonstrated that organisms cultured intraoperatively at the time of wound closure had identical typing and antibiotics sensitivities to organisms isolated from the eyelids and conjunctiva.38 By using restriction enzyme endonuclease studies, Speaker and partners found that 82% of organisms isolated from the vitreous of patients with postoperative endophthalmitis were genetically indistinguishable from isolates recovered from the patient's eyelid, conjunctiva, and nose.37
Gram-positive, coagulase-negative micrococci, including Staphylococcus epidermidis are responsible for most cases of postoperative endophthalmitis.15,18,19,23,26,34,38,42 Kattan and coworkers reported that 78% of organisms cultured following cataract surgery between 1984 and 1989 consisted of gram-positive species.15 More cases were attributed to S. epidermidis than to any other species. In their group of all patients with culture-positive endophthalmitis from 1990 to 1994, Aaberg and coworkers found 76.7% of patients grew gram-positive organisms, 47.8% of which were S. epidermidis.18 In the Endophthalmitis Vitrectomy study (EVS), gram-positive, coagulase-negative micrococci represented 70% of the 323 isolates meeting the study criteria for confirmed growth.42 This group consisted almost entirely of S. epidermidis with a single isolate of Micrococcus kristinae.42 The next most common group consisted of other gram-positive species, including S. aureus (9.9%), 29 isolates of Streptococcus species (9.0%), and 7 isolates of Enterococcus species (2.2%). The least common group of organisms consisted of gram-negative species, representing only 5.9% of isolates.42 Polymicrobial growth occurred in 9.3% of the EVS patients. In this group, two species of gram-positive, coagulase-negative micrococci grew in more than half of patients.42 It is clear that organisms typically found colonizing the ocular adnexa are not so benign as once believed.40,43 They appear to play a significant role in the pathogenesis of postoperative endophthalmitis.
One of the most important aspects in the prognosis of postoperative endophthalmitis is the physician's willingness to make a diagnosis. This requires maintaining a high level of suspicion for patients who are at greater risk preoperatively and for patients who have greater postoperative inflammation than expected. If postoperative inflammation is disregarded and other signs and symptoms of endophthalmitis go unnoticed, critical time may be lost and prompt diagnosis be delayed.
In the EVS, 24% of patients presented within 3 days of the surgical procedure; 61% presented within 1 week.34 These figures emphasize the importance of early and frequent postoperative examinations, especially in complicated surgical cases. Patients should be educated regarding the normal postoperative course and counseled about warning signs such as excessive pain and decreased vision. Patients should also be instructed to call their physicians promptly if these symptoms begin.
In cases of suspected endophthalmitis, a careful history and examination are of utmost importance. Systemic diseases such as diabetes mellitus, which might put the patient at increased risk for the development of endophthalmitis or might point to a source of endogenous infection, should be noted. Factors such as antibiotic and steroid use and systemic immunosuppression should also be determined. Details regarding the operative course such as duration of the case, presence of capsular rupture, or vitreous loss are important.6,44 Physical examination should include a careful determination of visual acuity as this influences the immediate course of treatment.34 The pupils should be examined and any afferent pupillary defect noted. Findings such as the degree of anterior chamber inflammation, the presence of a hypopyon or corneal infiltrate, and clarity of the vitreous should be noted. Loss of vision, vitreous opacification, and the presence of a hypopyon are some of the findings consistent with acute postoperative infection.23 In general, if a hypopyon is present, one needs to assume that the source is infectious until proven otherwise. Integrity of the wound should be checked by looking for signs of wound leak, vitreous wicks, and the presence of any suture material that might provide an infectious tract. The conjunctiva should be checked for evidence of a bleb and the iris examined for rubeosis. IOL position, stability, and capsular integrity should be noted. The capsule should also be examined for evidence of white plaques, which may suggest an infectious process especially in late postoperative endophthalmitis. Retained lenticular material should also be noted as this can cause inflammation, which may mimic infectious endophthalmitis. This is especially important if the posterior capsule is no longer intact. Ultrasound can be useful in the diagnosis of retained nuclear material, which may not be visualized on clinical examination. Examination of the posterior segment should include noting the presence or absence of a red reflex, degree of retinal vascular visualization, retinal detachment, choroidal detachment, or vitreous membrane. Retinal detachments are difficult to repair in patients with underlying endophthalmitis and usually confer a poor visual prognosis.35 When the posterior segment cannot be properly visualized, B-scan ultrasound should be performed.
After endophthalmitis has been suspected and a thorough history and physical examination completed, diagnosis is confirmed by completion of appropriate cultures. External cultures of the eyelid margins and conjunctiva may be obtained. Both aqueous and vitreous cultures should be obtained before injection of any antibiotics or steroids. A vitreous sample is especially important because a large number of aqueous samples demonstrate negative growth. In the EVS, confirmed growth on culture was demonstrated from 54.9% of vitreous samples compared with only 22.5% of aqueous samples.31 In cases in which a vitreous sample was positive, Bode and colleagues demonstrated negative growth in up to 57% of aqueous cultures.30 Given these findings, a vitreous sample is absolutely imperative for adequate diagnosis.
Vitreous specimens can be obtained either by trans-pars plana needle aspiration or by vitreous biopsy with a vitrectomy unit. In the EVS, vitreous was cultured by one of three methods: needle aspiration, vitreous biopsy with the vitrector, or vitreous biopsy followed by vitrectomy (vitreous effluent was also cultured in these cases). No statistically significant difference was found in the yield among the three groups.31 Some authors have suggested that a vitreous culture using a vitrector unit may be superior because a higher inoculum is obtained, and a larger specimen is collected.45 This may aid in culture positivity and assist in clinical recovery by removing a greater number of infectious organisms from the eye. In addition, aspiration of vitreous with a syringe is often difficult and can place traction on the retina.23 If needle aspiration is attempted and fails, a vitreous biopsy using a vitrectomy unit should be performed.
When a vitrectomy is used for biopsy and culture, an initial undiluted vitreous specimen should be obtained after placement of the sclerotomies, but before the infusion is turned on. The cutter can then be introduced in to the vitreous and a vitreous sample removed and aspirated into a syringe. Following sample collection, the infusion can be turned on.34 Vitreous effluent from the vitrectomy cassette can be collected and cultured following the surgical case. The effluent is usually processed through filter paper as in the EVS.31 The dilute effluent can also be centrifuged to produce a concentrated specimen.23
After specimens have been collected, samples are cultured on appropriate media. These include blood, chocolate, thioglycolate broth, and Sabouraud dextrose agar, all incubated at 37°C. Gram and Giemsa staining should also be performed. The presence of a positive result on Gram stain has been associated with a higher incidence of culture positivity.31 In the EVS, the positive predictability of the Gram stain was greater than 94%.31 A positive culture is often defined as at least semiconfluent growth on a solid medium, any growth on two or more media, or growth on one medium with a positive Gram stain.34,46 Detection of bacterial DNA using polymerse chain reaction (PCR) may be considered for cases in which infectious endophthalmitis is clinically suspected, but in which negative cultures have been obtained.47,48 Okhravi and associates reported a 100% concordance rate between PCR- and culture-positive bacterial samples and demonstrated that the technique is useful for detecting infection in cases which are culture negative.47
The aims in treating patients with postoperative endophthalmitis include sterilization of the eye and control of inflammation.
Intravitreal antibiotics are the initial treatment of choice for postoperative endophthalmitis.23 Two drugs are typically used. One should have broad and effective activity against gram-positive species, specifically coagulase-negative staphylococci, and the other should protect against gram-negative bacilli. In the EVS, intravitreal vancomycin and amikacin were administered to all patients and demonstrated effective antimicrobial activity against most pathogens.34,44 Vancomycin is the traditional drug of choice for gram-positive coverage. It is the only agent tested to which all gram-positive species in the EVS were susceptible.44 Development of vancomycin-resistant strains of staphylococcal and streptococcal species has prompted the Centers for Disease Control and Prevention (CDC) to recommend reserving use of vancomycin for treatment of serious infections produced by β-lactam resistant gram-positive microorganisms.49 Cefazolin (Ancef) has activity against many gram-positive organisms. However, in a rabbit model of S. aureus endophthalmitis, cefazolin was less effective than vancomycin and clinical resistance of gram-positive organisms is frequent.50,51 In light of the absence of data suggesting an effective alternative to vancomycin, and the potentially devastating consequences of endophthalmitis, vancomycin remains the agent of choice for the coverage of gram-positive organisms in cases of postoperative endophthalmitis. Based on susceptibility data from the EVS, vancomycin in combination with either amikacin, ceftazidime, or ciprofloxacin are appropriate antibiotic choices.44 Many practitioners have moved away from using gentamicin because it is more toxic than amikacin in experiments using rabbit eyes. Additionally, macular infarction has been reported after its administration in patients.52,53
Periocular antibiotics may be used as an adjunct to intravitreal therapy.34 Subconjunctivally injected antibiotics have poor intravitreal penetration54 but can achieve high concentrations in the anterior segment. Similarly, topical antibiotics provide poor intravitreal penetration but can achieve high concentration in the anterior segment and help sterilize wound sites.44 It is still unclear whether subconjunctival and topical antibiotics offer additional benefit in the treatment of postoperative endophthalmitis.
The role of corticosteroids in the treatment of postoperative endophthalmitis is still somewhat controversial. In the EVS, corticosteroids were administered to all patients subconjunctivally, topically, and orally, although the study was not designed to determine the efficacy of this adjunctive therapy. Corticosteroids play an important role in controlling the inflammatory response after infection, which may be further damaging to the eye. Corticosteroid use intravitreally may help control intraocular inflammation, especially after bacterial death from antimicrobials. Unless fungal endophthalmitis is suspected, some practitioners recommend the use of intravitreal steroids in conjunction with intravitreal antibiotics.23,55 Several studies have suggested that administering intravitreal corticosteroids may be beneficial,56–58 although at least one study found that intraocular steroid use was not associated with a significant effect on visual acuity.18 No prospective, controlled clinical trials have yet proved the efficacy of intravitreal corticosteroids in the management of postoperative endophthalmitis.
The EVS determined that patients who presented with initial vision of only light perception who underwent a vitrectomy had a three times greater chance of achieving 20/40 final visual acuity (33% versus 11%) compared with patients who had only a vitreous tap or biopsy. There was no difference in outcome between vitrectomy and vitreous tap in patients with endophthalmitis who presented with initial visual acuity of hand motions or better.34
Before the results of the EVS, systemic antibiotic use was considered a standard treatment for postoperative infectious endophthalmitis.34 The EVS concluded, however, that there was no difference in visual acuity or media clarity outcome with or without the use of systemic antibiotics.34 These results were important for several reasons. Systemic antibiotics may have significant adverse effects and their use is not without risk. Systemic antibiotics are also expensive. Not only do the drugs incur expense, they usually require either admission to a hospital or home-based nursing for their administration. Thus, the findings of the EVS may save the patient unnecessary risk and may obviate the need for hospital admission. However, results of the EVS may need to be reevaluated in light of new fluoroquinolone antibiotics that have better intraocular penetration compared with that of other antibiotics.9,59
Of utmost importance in the treatment of postoperative endophthalmitis is prompt initiation of therapy. This point cannot be overemphasized. Endophthalmitis is an ocular emergency and initiation of treatment should reflect this. The visual outcome of the patient is strongly connected to the efficiency with which a diagnosis of endophthalmitis is made. In addition, the results of the EVS and other studies have provided significant guidelines for the treatment of postoperative infection, but one must keep in mind that each case must be evaluated on an individual basis. The patient must be monitored closely and decisions to change treatment strategies or initiate a surgical procedure must be considered at each step.
Despite the fear provoked by a diagnosis of postoperative endophthalmitis, final visual acuity can be surprisingly good. In general, the worse the acuity at initial presentation, the worse the final visual outcome will be.44 In the EVS, 41% of patients achieved 20/40 or better visual acuity at 3 months. By the end of the study, more than half of patients (53%) achieved visual acuity of 20/40 or better and 74% achieved 20/100 vision or better. Only 11% of patients had acuity worse than 5/200.34 Other studies have found similar results. Kattan and colleagues found that approximately one third of patients achieved a final visual acuity of 20/60 or better following endophthalmitis after cataract extraction with or without IOL placement.15 Speaker and Menikoff found 64% of patients with endophthalmitis after intraocular surgery had a visual acuity of 20/100 or better after treatment.26
Final visual acuity after postoperative endophthalmitis varies according to surgical type. Aaberg and coworkers found the median visual acuity after endophthalmitis treatment was 20/200 overall. Postcataract extraction endophthalmitis patients had a median final acuity of 20/133, and the best median visual acuity, 20/40, was obtained after secondary IOL placement. As may be expected, patients with post-PPV endophthalmitis fared worst, with a median acuity of no light perception.18
The EVS and other studies have found certain risk factors associated with a worse visual acuity outcome after treatment for postoperative endophthalmitis. According to the EVS, the most significant independent risk factor for decreased vision was visual acuity of light perception or worse at initial presentation.34 Other independent risk factors included advanced age, diabetes, presence of a corneal infiltrate and/or ring ulcer, intraoperative violation of the posterior capsule, low or high intraocular pressure, presence of an afferent pupillary defect, rubeosis, and absent red reflex.34,44 Studies have demonstrated that more virulent organisms, including S. aureus, Streptococcus species, and gram-negative rods are often associated with worse visual prognosis.19,23,60 A shorter time interval between the onset of symptoms and initiation of treatment also plays a role in visual outcome following postoperative endophthalmitis.60 In the EVS, initiation of treatment was begun within 6 hours of the initial exam.34 Visual results after treatment were better in the EVS than in many other studies.15,18,26,60 This may have resulted in part from efficient initiation of treatment, underscoring the importance of prompt diagnosis and treatment. The first few hours in the management of postoperative endophthalmitis are critical.
Preventive measures play an important role in reducing the incidence of postoperative endophthalmitis. The practitioner can play an active role in the prevention of infection by choosing to incorporate appropriate aspects into perioperative management of his or her patients.
Preparation of the surgical field with sterile drapes does not eliminate the risk of infection, but it can reduce it.61 Adhesive drapes, which cover the eyelid margin and lashes, may be particularly important in reducing the chance of infection from microbial contaminants of the ocular adnexa.
Various topical antiseptics have been used for the preoperative preparation of the eye. Argyrol, a mild silver protein solution, was commonly used for many years. In 1983, however, Isenberg and associates demonstrated that the chemical preparation was no more effective than no treatment in reducing the number of bacteria.62 In addition, the same investigators evaluated the effect of conjunctival irrigation in the preoperative chemical preparation of the eye.63 They determined that both the bacteria colony count and species count increased following irrigation with saline solution compared with nonirrigated eyes.
Unlike many more traditional topical antiseptics, povidone-iodine has been shown to have significant antimicrobial effect.26,64–69 Povidone (polyvinylpyrrolidone) is a high-molecular-weight, water-soluble polymer that forms a complex with iodine.68 After instillation, free iodine is released, which has been credited with the compound's antimicrobial effect.68,70 Povidone-iodine possesses minimal tissue toxicity and has a broad antimicrobial effect with activity against bacteria, viruses, fungi, protozoa, and spores.67,68 Maumenee and Michler first studied the use of iodine solution in ophthalmic surgery in the early 1950s.71 They found it possessed an antimicrobial effect on the ocular adnexa when the solution was used concomitantly on the skin. In 1984, Apt and coworkers demonstrated that eyes treated with half-strength povidone-iodine solution preoperatively had a 91% decrease in bacteria colony counts and a 50% decrease in the number of species compared with those found in controls.68 The solution did not cause significant conjunctival hyperemia and no allergic or adverse reactions were reported. Speaker and Menikoff demonstrated that the preoperative application of 5% povidone iodine to the ocular surface significantly lowered the incidence of culture-positive endophthalmitis after intraocular surgery compared with eyes prepared with a silver protein solution.26 Among more than 3000 cases in which povidone-iodine was used, the solution was well tolerated and no adverse reactions were noted.26
In most studies of povidone-iodine, a 5% solution is prepared by diluting full-strength 10% solution 1:1 with saline solution.23,26,68 Two drops of the solution are instilled in the conjunctival sac, and the eyelids are gently manipulated to distribute the solution over the surface of the eye. The solution bathes the eye while the rest of the preparation is completed. The solution should remain in contact with the eye for several minutes to exert its antimicrobial effect. The skin, eyelid margins and lashes are then scrubbed with povidone-iodine-soaked gauze or cotton swabs. Finally, the eye is irrigated with saline solution at the end of the preoperative preparation. This step is important to avoid intraocular toxicity with the solution.26 Povidone-iodine solution should not be confused with povidone-iodine scrub. The latter contains a detergent, which is toxic to the eye and should be avoided.26
Use of perioperative antibiotics in ophthalmic surgery remains controversial. Short-course topical antibiotics administered before surgery reduce the amount of lid and conjunctival bacteria.69,72,73 The amount of pathogen reduction depends on the type of antibiotic used, the frequency and duration of its use, and the type of bacteria present and the susceptibilities of the organism.73
Inasmuch as the major infectious source for postoperative endophthalmitis is the patient's own eyelids and conjunctiva.9, it is reasonable to attempt to reduce the number of organisms with prophylactic antibiotics. Up to three quarters of patients carry organisms on the ocular adnexa capable of causing postoperative endophthalmitis.40 In 1964, Allen and Mangiaracine suggested that preoperative topical antibiotics reduced the incidence of endophthalmitis.74 At that time, chloramphenicol and gentamicin were routinely used as antibiotics. Studies using these antibiotics have shown variable efficacy.75–77 More recently, studies employing fluoroquinolone antibiotics have reported favorable results. This family of antibiotics has broad-spectrum antimicrobial activity, low rates of resistance, and minimal toxicity.78 Donnenfeld and associates demonstrated that patients given topical 0.3% ofloxacin 30 and 90 minutes preoperatively had aqueous humor drug levels that approached the minimal inhibitory concentration of 90% of Staphylococcus epidermidis, S. aureus, and most gram-negative organisms.78 Von Gunten and coworkers reported similar results.59 Snyder-Perlmutter amd and colleagues reported that patients receiving topical antibiotics every 5 minutes for three doses had a statistically significant reduction in bacterial colony-forming units following treatment with ciprofloxacin but not ofloxacin.79 These studies suggest that topically applied antibiotics might decrease the intraocular bacteria count sufficiently to prevent development of infectious endophthalmitis. The low incidence of endophthalmitis, however, makes it difficult to prove that a direct correlation exists between antibiotic use and a reduction in the incidence of infectious postoperative endophthalmitis.
Subconjunctival antibiotics have been commonly used since the 1950s for infectious prophylaxis following ocular surgery.80 During this time their use has been a source of continued controversy. Administration of subconjunctival antibiotics circumvents the corneal epithelial barrier and permits a high concentration of drug to penetrate the corneal stoma and diffuse into the aqueous humor.9 Studies have demonstrated good penetration of antibiotics into the anterior chamber following subconjunctival administration.44,81 A subconjunctival injection of antibiotics can achieve bactericidal levels in the anterior chamber thus eliminating most infection.9 It has been suggested, however, that high intraocular levels of antibiotic achieved with subconjunctival injection might only partially treat an infection in some cases and delay onset of symptoms and diagnosis leading to a poor outcome.82 Additionally, the concentration of anterior chamber antibiotic decreases rapidly after administration and vitreous penetration following subconjunctival administration of antibiotics is also poor, limiting its efficacy.9,23,54,73
Studies evaluating the efficacy of subconjunctival antibiotics in decreasing the incidence of endophthalmitis have been inconclusive. In 1967, Cassady published an uncontrolled series of 1212 cataract extractions in which subconjunctival antibiotics were used.27 No cases of postoperative endophthalmitis were reported. Cassady attributed his success to the use of subconjunctival antibiotics and justified their usage after every cataract extraction. During that same year, a controlled, prospective study was reported evaluating subconjunctival antibiotic use following cataract surgery.83 Two cases of endophthalmitis were reported in both the control and treatment groups. Unfortunately, the small number of infections prevented any conclusions beingdrawn regarding the statistical significance of antibiotic prophylaxis. In another study, Kolker and associates reported a 1.42% incidence of postoperative endophthalmitis in patients given no prophylaxis versus 0.21% in a similar group given subconjunctival penicillin and streptomycin.84 In two series reviewing postoperative endophthalmitis, all but one patient had received postoperative subconjunctival antibiotics.15,26 No definitive evidence has elucidated the appropriate role of subconjunctival antibiotics in surgical prophylaxis. Issues of efficacy, safety, and cost must be considered and a controlled, prospective clinical trial must conducted before a final verdict is reached.
Aminoglycosides were commonly used for subconjunctival injection, but such usage has fallen into disfavor. Gentamicin has been associated with more conjunctival toxicity and greater patient discomfort than cefuroxime.85 Aqueous penetration of gentamicin is poorer than cephalosporins and more species that are implicated in the etiology of postoperative infectious endophthalmitis are sensitive to cephalosporins than aminoglycosides.85 Retinal toxicity has also been demonstrated following aminoglycoside use. In 1991, a survey of members of the Retina, Macula and Vitreous Societies revealed 93 cases of macular infarction following administration of gentamicin sulfate with eight additional cases secondary to other aminoglycosides.53 Most reports followed planned intravitreal injection of antibiotic but twenty-three cases occurred in eyes treated with prophylactic subconjunctival antibiotics after routine ocular surgery. Of these cases, most occurred following known or presumed ocular penetration; at least two cases were thought to be associated with drug leakage into the eye through the cataract wound following injection. Aminoglycosides have an extremely small therapeutic window, making their use for routine surgical prophylaxis unwise.53 If even a small amount of gentamicin dosed for subconjunctival use gains entry to the eye, retinal toxicity could occur.80 Compochiaro and Conway pointed out that endophthalmitis is treatable whereas macular infarction is not; they thus recommended that use of prophylactic subconjunctival aminoglycosides be abandoned following routine ocular surgery and subconjunctival use of other antibiotics be administered only after serious consideration and careful surgical attention.53
Intracameral antibiotic use, including antibiotic administration in irrigating solutions and through direct injection, is another method of infectious prophylaxis that has gained popularity recently. Preliminary toxicity studies have been done in animals, but difficulty has been encountered when equivalent antibiotic doses are applied in humans.86,87 This has resulted in disagreement regarding clinical doses and formulations of prophylactic intracameral antibiotics.88–90 Theoretically, intracameral antibiotics may reduce the incidence of postoperative infectious endophthalmitis, but the clinical evidence supporting their use is largely circumstantial. No prospective, controlled clinical trials have been completed to date to address this question. An uncontrolled study by Gills reported one case of endophthalmitis in 20,000 cataract extraction following gentamicin use in irrigating solution and one case in 9928 after intracameral vancomycin.88 No cases were reported following 25,000 operations in which both gentamicin and vancomycin were used in irrigating solution. In another uncontrolled study, Gimbel and partners reported no cases of endophthalmitis following 4,684 cataract extractions in which both intracameral gentamicin and vancomycin were used.90 Those authors reported no adverse effects on corneal endothelium when cases were compared to similar series following phacoemulsification with lens implantation and no intracameral antibiotics, but endothelial damage may have been spared because vancomycin instead of being infused in the irrigating solution was delivered into the capsular bag by intracameral injection after lens placement. Beigi and colleagues reported that the addition of gentamicin and vancomycin to irrigation solution during phacoemulsification significantly reduced contamination of anterior chamber aspirates with bacteria.89 No patients in the study developed postoperative endophthalmitis, however, making it impossible to determine whether intracameral antibiotic use decreased the incidence of postoperative infection. One should also note that the CDC has cautioned against the routine use of vancomycin.49
DELAYED-ONSET BACTERIAL ENDOPHTHALMITIS
Awareness of delayed onset or late endophthalmitis has increased in recent years. As with acute infectious postoperative endophthalmitis, recognition of this entity is of utmost importance in early diagnosis and improved visual prognosis. The signs and symptoms of delayed onset endophthalmitis can be subtle, often mimicking other causes of inflammation and making diagnosis difficult.
Many terms have been used to describe this entity including delayed-onset, chronic, and late postoperative endophthalmitis. A consensus regarding the time period in which this disorder occurs has also been disputed. Some authors have suggested that symptoms starting more than 1 month postoperatively should be considered chronic endophthalmitis, whereas other authors have defined chronic endophthalmitis as requiring two or more episodes of low-grade postoperative infectious uveitis regardless of the time of onset.91–93 The Endophthalmitis Vitrectomy Study assessed acute postoperative endophthalmitis and evaluated patients with evidence of endophthalmitis within six weeks after surgery.34 For our purposes, we will define delayed-onset postoperative infectious endophthalmitis as uveitis that develops more than 6 weeks after surgery with evidence of an infectious process. Even so, this definition is somewhat arbitrary and merely serves as a guideline for classification purposes.
Most data estimate incidence of acute postoperative infectious endophthalmitis or do not distinguish between acute and chronic infections when reporting statistics of incidence.15,18,19,21,91 Unfortunately, this makes it difficult to assess the frequency with which delayed-onset endophthalmitis is encountered. In addition, studies have evaluated risk factors associated with acute postoperative endophthalmitis, but data are lacking with regard to infection of delayed onset.6,91
Propionibacterium acnes was the first bacterial organism reported to cause chronic low-grade inflammation after intraocular surgery.91,94 In 1986, Meisler characterized a new syndrome of delayed onset endophthalmitis and reported six cases of chronic postoperative endophthalmitis masquerading as chronic iridocyclitis from which Propionibacterium species were cultured.95 Other bacteria can cause a similar picture of chronic low-grade inflammation, including coagulase-negative Staphylococcus,96–98 Staphylococcus aureus,99,100 anaerobic Streptococcus,101 Corynebacterium species,96 Actinomyces,102,103 and Nocardia species.104 Various fungi, particularly Candida species, can produce a similar picture.96,105,106 Despite the array of organisms implicated in delayed-onset postoperative endophthalmitis, P. acnes is most often encountered.
P. acnes is a gram-positive, non-spore-forming anaerobic pleomorphic bacillus that has been implicated in a wide variety of systemic infections.92 It is a ubiquitous organism, which inhabits the human sebaceous follicle.102 It has been cultured in up to 43.8% of normal human conjunctivae.43 In addition to chronic endophthalmitis, ocular and periocular infections from P. acnes have included preseptal cellulitis, canaliculitis, dacryocystitis, conjunctivitis, and keratitis.107 Several cases of P. acnes-induced acute postoperative endophthalmitis have also been described.93,102
P. acnes cell wall resists digestion by macrophagesand polymorphonuclear cells, which allows it to persist sequestered in inflammatory cells within the eye.91,101,108 Some evidence has suggested that P. acnes can also live within the lens capsule, protected from inflammatory cells and attempts to sterilize the eye with antibiotics.109 Release of sequestered organisms may explain the onset of endophthalmitis that has been reported after treatment with capsulotomy using neodymium/yttrium-aluminum-garnet (Nd:YAG) lasers.95,109,110
Patients with delayed-onset postoperative endophthalmitis may present with signs and symptoms typically found in patients with acute endophthalmitis. Most often, however, they present with low-grade chronic and often recurrent inflammation that may simulate noninfectious uveitis. Such inflammation occurs over the course of months compared with the abrupt inflammatory presentation seen in acute postoperative infections. Patients may initially show some improvement following corticosteroid treatment, which hinders the correct diagnosis of endophthalmitis. Another finding that separates this entity from acute endophthalmitis is development of granulomatous precipitates on the corneal endothelium and IOL.92,95,110,111 The finding of white plaques on the posterior capsule is also suggestive, although not pathognomonic for P. acnes infectious postoperative endophthalmitis.91,92,101,112,113 A similar plaque has also been described in other causes of bacterial and fungal endophthalmitis.114,115 The plaque is usually located between the IOL and the posterior capsule or peripherally in the capsule, requiring adequate dilation for visualization.92,101 (Fig. 2). Histologically, the plaque consists of a mixture of lens material and colonies of microorganisms.92,101 Other clinical findings may include hypopyon, a fibrinous reaction in the anterior chamber manifested as beaded fibrin strands, as well as vitritis and cystoid macular edema.95,102,109,110 Patients usually develop symptoms between 1 and 10 months after surgery, although the time elapsed to culture proven infection is often delayed owing to initial misdiagnosis.95,96,102,109,110,116
As in cases of acute postoperative endophthalmitis, aqueous and vitreous cultures should be obtained. Owens and coworkers suggest directing the anterior chamber tap needle into the capsular bag or toward the plaque to increase pathogen yield.117 Likewise, it is desirable to acquire a section of the capsule or plaque when attempting to obtain a vitreous specimen.91,92 This should be sent for culture as well as histopathologic studies.92,110 Specimens should then be sent for both Gram stain and culture, including culture on anaerobic media. The specimens must be incubated for at least 1 to 2 weeks because it may take longer than expected to observe growth.92 As in cases of acute infectious endophthalmitis, PCR may be helpful in establishing a diagnosis.47,48
No consensus exists regarding the best treatment for patients with chronic infectious postoperative endophthalmitis. Various treatment strategies have been suggested but the individual patient must be kept in mind when deciding on a therapeutic plan.
Some authors have suggested that initial treatment should consist of medical therapy such as topical and systemic antibiotics alone, or in combination with intraocular antibiotics.93,102,111 P. acnes has been shown to be sensitive to penicillins and cephalosporins, but vancomycin may be the intraocular antibiotic of choice.91,92,111 Vancomycin is not only effective against P. acnes, but provides coverage against other bacteria such as coagulase-negative streptococcus, which can cause delayed-onset endophthalmitis.51,91 Other antibiotics such as clindamycin and chloramphenicol can also be effective.91,117 Zambrano and coworkers reported successful treatment using a nonsurgical approach consisting of topical, intraocular and/or intravenous antibiotics in four of nine cases reviewed at Bascom Palmer Eye Institute.102 Clindamycin irrigation into the capsular bag has also been reported to be useful in treating P. acnes delayed-onset endophthalmitis,117,118 although another series reported recurrent inflammation in 100% of patients receiving intraocular antibiotic treatment alone.111
Most cases will go on to require some type of surgical treatment due to recurrent or persistent inflammation. Zambrano and colleagues proposed an algorithm to treat patients with delayed-onset infectious postoperative endophthalmitis.102 They suggested that patients who present with mild symptoms on initial examination have cultures performed and be given intraocular vancomycin. If that therapy is ineffective, those authors suggest that PPV with excision of the intracapsular plaque and repeat antibiotic injection be completed while maintaining sufficient capsule to support the IOL. Patients presenting with more severe clinical findings may receive this treatment initially. Finally, if this approach does not alleviate inflammation, IOL removal with complete capsular bag removal and possible IOL exchange can be performed.102 Clark and colleagues111 found that patients treated with intra-ocular antibiotics alone or a combined therapy with PPV and intraocular antibiotics had high rates of recurrent and persistent inflammation (100% and 50%, respectively.) Patients receiving PPV with subtotal capsulectomy and intraocular antibiotics had a 14% incidence of inflammation following treatment. No recurrent or persistent inflammation was reported for those patients who later went on to receive PPV, total capsular bag removal, intraocular antibiotics, and either IOL exchange or removal.111
Despite various available treatment options and the lack of ideal therapeutic alternatives, the visual prognosis for these patients can be relatively good.93,102,111,119 Clark and colleagues reported final visual acuities to be 20/40 or better in 50% of patients and 20/400 or better in 75%.111 Zambrano and coworkers reported final visual acuities for two thirds of patients to be between 20/20 and 20/60 and Winward and coworkers reported final visual acuity of 20/70 or better in 70% of patients.93,102 Poor final visual acuity is often attributed to cystoid macular edema, retinal detachment, retinal vascular events, glaucoma, or phthisis bulbi.92,102,111
The first case of fungal endophthalmitis was reported by Romer in 1902 following a case of penetrating trauma with subsequent Aspergillus fumigatus infection.120 Fungal endophthalmitis is usually considered in the differential diagnosis of indolent postoperative inflammation, although reports of acute postoperative fungal endophthalmitis have also been cited.105,106,121 Fungal infections are relatively rare and can resemble bacterial endophthalmitis, making diagnosis difficult.105,114 When this diagnosis is being entertained, diagnostic measures should be completed expeditiously because a delayed diagnosis may be associated with a poorer visual prognosis.106
Some clinical signs may help distinguish fungal endophthalmitis from other sources of infection. Fungal infections may exhibit “fluff balls” or a “string-of-pearls” infiltrate in the anterior chamber and vitreous.91,101,122,123 The infection tends to be more localized with inflammation often limited to the anterior chamber, pupillary space, or anterior vitreous,105,123 and a mass may be seen on the iris, ciliary body, lens, or adherent to the cornea105,106,124 (Fig. 3). Reports have demonstrated that fungus can infiltrate these ocular tissues, which serve as a reservoir for persistent infection.105 Other findings can include persistent corneal edema, corneal infiltrates or necrotizing scleritis.91,110,111,125 As in chronic bacterial endophthalmitis, a hypopyon or white plaque may be seen, and low-grade persistent and recurrent inflammation is often noted.91,101,106,121,122
Various organisms have been reported to cause postoperative fungal endophthalmitis. These include Candida parapsilosis, Aspergillus flavus, Torulopsis candida, Torulopsis magnoliae, Paecilomyces lilacinus, Acremonium species,105,106,114,122,126,127 and others. Exogenous fungal endophthalmitis typically occurs following cataract surgery, but reports have also documented infection after penetrating keratoplasty, retinal surgery, and glaucoma filtering.24,80,106,128 Pflugfelder and coworkers noted that some fungi appear to be associated with a better visual prognosis including Aspergillus, Paecilomyces, and Candida, whereas infections secondary to Acremonium and Fusarium species typically have poorer outcomes.105
When a fungal infection is suspected, cultures should be obtained expeditiously. Given the variability of clinical presentation and the possibility of fungal infection mimicking bacterial endophthalmitis, it is reasonable to culture intraocular specimens for both bacteria and fungi in all patients with suspected endophthalmitis.105 Cultures should be obtained in the same fashion outlined for bacterial endophthalmitis. When a capsular plaque is present, this should be harvested for histologic examination. Similarly, aspiration or excision of any mass that may be present in the anterior chamber, or adherent to the cornea, iris, or lens is important for both diagnosis and treatment.105 Both Giemsa and Gomori methamine silver stains should be obtained in addition to a Gram stain.105,106 Negative fungal cultures may be found in 50% of cases and do not preclude a diagnosis of fungal disease.106,123 Fungal infections may be localized to one part of the vitreous and specimens obtained for culture may fail to sample these areas.123 Negative bacterial cultures may suggest the diagnosis of a fungal infection.
No consensus exists regarding treatment for patients with fungal endophthalmitis in the postoperative setting. Most reports of endophthalmitis postoperatively are retrospective reviews using various treatment options chosen on an individual patient basis.105,121,122 A combination of therapeutic options is usually employed including intraocular, intravenous, and oral antifungals, corticosteroids and PPV.91,105,106,122
Intraocular amphotericin B is usually the drug of choice for primary therapy until culture and sensitivity results are obtained.105,122 Amphotericin is effective against a wide range of fungal organisms capable of causing postoperative endophthalmitis.105,122 Pflugfelter and coworkers suggested that amphotericin B should be used after obtaining appropriate cultures if clinical evidence such as delayed onset of inflammation, inflammation in the presence of a relatively quiet eye, minimal discomfort, an inflammatory mass in the eye, or fungi on stained smear suggest a fungal infection.105 The drug should be administered in the area of maximum inflammation.105 Amphotericin may increase ocular inflammation after administration and may need to be given repeatedly if persistent or recurrent inflammation occurs.105,122,129 If intraocular amphotericin B therapy fails, miconazole treatment has been suggested.105 This may be true for Paecilomyces lilacinous, which has demonstrated resistance to amphotericin.105,106
Periocular antifungals have been suggested if the cornea, sclera, or anterior chamber is the source of infection.105 Animal studies have demonstrated good anterior chamber penetration of natamycin, amphotericin B, and miconazole following topical administration.96,130 Systemic antifungal use has been widely debated. No prospective controlled studies have evaluated the efficacy of systemic therapy for postoperative fungal endophthalmitis. Many patients receive systemic antifungals as part of their treatment regimen, but good visual outcomes have also been reported without systemic therapy.105,131 Systemic antifungal use is not without risk. Amphotericin B can induce renal toxicity and hematologic side effects and is often not tolerated. Ketoconazole and fluconazole may be better tolerated but still have the potential to induce liver toxicity.91,105,132 Flucytosine seems to be the least toxic, but it is generally not used for monotherapy because it is associated with a high incidence of resistance.105,133,135 The role of oral antifungal therapy in postoperative endophthalmitis has not been clearly defined, and the risks and benefits of their use must be weighed before treatment.
Corticosteroids must be used with caution in suspected cases of fungal endophthalmitis, and their role in treatment is still unclear. Steroids have been used in this setting in an effort to control inflammation and limit tissue destruction incited by the host inflammatory response.134 Pflugfelder and coworkers suggested that topical corticosteroids could be used to control inflammation after appropriate medical and surgical therapy had been employed.105 Shulman and Peyman reported a beneficial role for intraocular steroids in cases of fungal endophthalmitis.135
A surgical approach to postoperative fungal endophthalmitis is often undertaken.91,105,106,121,122 Vitrectomy including excision of involved ocular structures and intraocular injection of antifungals is suggested.105,106,122 Pflugfelder and coworkers reported that seven of eight patients with good visual results in their series had undergone vitrectomy followed by intraocular antifungal administration.105 This therapy may need to be repeated several times before the infection is controlled, and patients need to be monitored closely for recurrence despite removal of involved tissue.122
Inflammation resulting from bleb-related infections is another cause of postoperative uveitis. Infection can develop in eyes after planned glaucoma filtering surgery or following inadvertent blebs after cataract surgery or trauma.136–139 Bleb-related infections can be further divided into localized bleb infection or bleb-associated endophthalmitis. The term blebitis was first coined by Brown and partners to describe a limited infection centered around a bleb without vitreous involvement.140 This is in contrast to bleb-related endophthalmitis, a more virulent infection with anterior chamber inflammation in addition to often-marked vitritis that typically develops months to years after glaucoma-filtering surgery, although cases of acute-onset, bleb-related endophthalmitis have been reported.137,141 Blebitis may be a precursor to the development of fulminant endophthalmitis.140,142 Treatment may be more effective when undertaken earlier; therefore, it is important to recognize bleb-related infections early on and to treat aggressively to avoid more serious ocular complications and a poorer visual prognosis.140
Clinical Features and Treatment
Blebitis and bleb-related endophthalmitis can be differentiated by the extent of ocular inflammation and differences in clinical presentation. Patients with blebitis typically present with a several-day history of redness, pain, photophobia, conjunctival discharge, and significant conjunctival injection surrounding an opalescent bleb (often with an overlying epithelial defect), which may contain inflammatory material.140–143 Patients may have variable anterior chamber reactions, hypopyon, and positive results to Seidel testing.140,141 By definition, patients with blebitis do not demonstrate vitritis. In contrast, patients with bleb-related endophthalmitis usually present with rapidly progressive symptoms including worsening visual acuity, pain, redness, diffuse conjunctival injection with an opalescent bleb (often without an overlying epithelial defect), anterior chamber reaction with variable fibrin or hypopyon and florid vitritis.137,141,142 Seidel testing may also be positive.137
Patients with blebitis develop infection in the setting of thin, cystic blebs, which typically leak on Seidel testing.140,142,143 Thin blebs lead to decreased tissue resistance and infection presumably develops when bacteria gain entry into the eye at the vulnerable bleb site and through overlying epithelial defects.142 This may explain why S. epidermidis and S. aureus, organisms that normally inhabit the ocular adnexa, are typically cultured from patients with blebitis.140–142 Blebitis often responds well to intense topical and subconjunctival antibiotic treatment and patients usually obtain nearly complete visual recovery, factors which may be due in part to the low virulence of the infecting organisms.137,140–142,144
Patients with bleb-associated endophthalmitis may develop infection in the setting of either thin or thick-walled blebs, with or without leakage.142 The organisms responsible for infection are usually more virulent bacteria, such as Streptococcus species or Haemophilus influenzae,101,137–139,141,142 which presumably gain entry to the anterior chamber by penetrating an intact conjunctival bleb.138,139,142,143,145 Endophthalmitis after inadvertent blebs is caused by similar organisms.141 Mandelbaum and colleagues reported a case of endophthalmitis occurring 60 years after inadvertent formation of a bleb.138 Rarely, bleb-associated endophthalmitis may develop in the acute postoperative period.137,139,141 In these instances, it is believed that perioperative introduction of bacteria, presumably from flora of the ocular adnexa, is responsible for infection.139,141,144 This would explain why coagulase-negative Staphylococci are the most common cause of acute bleb-related endophthalmitis, just as in cases of acute endophthalmitis following other ocular surgery.101,141
Treatment of bleb-related endophthalmitis is generally similar to that of acute postoperative endophthalmitis. Anterior chamber and vitreous cultures are usually obtained and intravitreal and periocular antibiotics administered. Vitreous specimens are more effective for identifying infectious organisms than cultures obtained from the aqueous or bleb; therefore, tapping the bleb is generally not recommended.137–139 Given the virulence of the infecting organisms, a vitrectomy is often required early in the treatment course and systemic antibiotics are often added, in contrast to postcataract endophthalmitis.138,139,141 Kangas and colleagues reported a final visual acuity of 20/400 or better in 55% of patients treated by initial PPV compared with 33% treated by initial vitreous tap without vitrectomy.137 Mandelbaum and associates have advocated that a complete vitrectomy be done as part of the initial therapeutic measures for eyes with bleb-associated endophthalmitis, given that a high proportion of these cases are caused by virulent bacteria that may severely damage the eye.138 Vitrectomy removes not only a significant bacterial inoculum but also damaging bacterial toxins and inflammatory factors induced by the infectious cascade.138,139
Risk Factors and Prognosis
In the last decade, the use of antifibrotic agents such as mitomycin C-(MMC) and 5-fluorouracil (5-FU) has increased.137 Histopathologic studies have shown that the use of MMC and 5-fluorouracil leads to thin-walled and cystic avascular blebs, which seem to be at greater risk for infection.146,147 These thin-walled blebs have a tendency to leak and studies have shown that bleb leakage is a significant risk factor for bleb-related infection.143,144,148 Wolner and colleagues found that the incidence of bleb-related endophthalmitis after trabeculectomy with 5-FU therapy resembled that of endophthalmitis after full-thickness glaucoma procedures, suggesting that the incidence of endophthalmitis after trabeculectomy with 5-FU is higher than that performed without adjunctive 5-FU.144 Bleb location may also be a risk factor. Higginbotham and partners found that inferior trabeculectomy performed with adjunctive MMC is associated with a greater than sevenfold increased risk of bleb-related endophthalmitis compared with superior filters with MMC.145 Wolner and colleagues noted a similar trend when 5-FU was used for inferior trabeculectomy. In their study, the relative risk of bleb-related endophthalmitis after inferior trabeculectomy with adjunctive 5-FU was 4.0 compared with a superior filter with 5-FU.144 Blebs associated with antifibrotic agents are thinner and may be more vulnerable to exposure and trauma induced by mechanical irritation from the lower lid margin.144 However, Mochizuki and partners found that the incidence of late bleb-related infection after trabeculectomy with antifibrotics was similar to that after trabeculectomy without these agents.143 Other risk factors that have been associated with an increased risk of bleb-related endophthalmitis include contact lens wear, conjunctivitis, and recent upper respiratory tract infections.138,139,149
Despite aggressive therapeutic maneuvers, patients with late bleb-related endophthalmitis typically have a poor visual prognosis.137–139,142,145,148 Kangas and associates reported a final visual acuity of 20/400 or better in 47% of patients following bleb-associated endophthalmitis.137 The poor visual outcome of patients is thought to result from the virulence of the infecting organisms.138,142,145 In the study by Kangas and associates, 40% of patients with streptococcal infection achieved a final acuity of 20/400 or better versus 52% in the nonstreptococcal group.137 Mandelbaum and colleagues reported that ten of sixteen eyes with Streptococcus-related endophthalmitis lost all light perception and fared worse than those without such infection.138 These results underscore the importance of early and aggressive treatment in patients with suspected bleb-related inflammation.
|NONINFECTIOUS POSTOPERATIVE UVEITIS|
Patients with underlying iridocyclitis undergoing cataract extraction or other intraocular surgery may have prolonged and severe postoperative inflammation. This entity should be kept in mind in the differential diagnosis of patients with inflammation following surgery. A full discussion of this topic is beyond the scope of this chapter.
POSTOPERATIVE LENS-INDUCED UVEITIS
Lens-induced uveitis refers to several mechanisms of intraocular inflammation induced by lens protein. Differentiating between these entities is often difficult clinically and definitive diagnosis may rely upon histopathologic identification of the type of cellular infiltration and pattern of inflammation around the lens.
Phacoanaphylactic endophthalmitis, also known as phacoanaphylactic uveitis or phacoantigenic uveitis, was first reported by Straub in 1919, and an autoimmune mechanism of inflammation was later described by Verhoff and Lemoine in 1922.150,151 The term anaphylactic was used to describe the acuity with which inflammation developed in these patients, rather than a type of immune response. Traditionally, phacoanaphylactic endophthalmitis was believed to result when sequestered lens protein was exposed to the intraocular environment by disruption of the lens capsule, causing a reaction similar to rejection of a foreign tissue graft.5,152 It is now recognized that phacoanaphylactic endophthal-mitis is a rare autoimmune inflammatory response to lens protein resulting from an abrogation of tolerance.5,152,153
We now know that lens tissue is normally recognized by the immune system. This is in contrast to the traditional belief that lens protein was immunologically foreign because it was sequestered within the lens capsule.5,154 Phacoanaphylactic endophthalmitis is not a cell-mediated rejection of tissue, but an immune complex disease that develops following abrogation of normal tolerance to lens protein.5,152,155 Phacoanaphylactic endophthalmitis is therefore a disturbance in the normal recognition of lens protein as self.5
Phacoanaphylactic endophthalmitis has a wide range of clinical manifestations, making it difficult to distinguish from other causes of postoperative inflammation.5,152,156,157 In a clinicopathologic review, it was diagnosed clinically in only five percent of cases with histopathologically diagnosed disease.152 These results must be appreciated when considering its clinical features. By definition, the lens capsule must be disrupted, and onset of inflammation may vary depending on prior sensitization to lens protein.152 Cases of phacoanaphylactic endophthalmitis are typically unilateral although bilateral cases have been reported, making a distinction between phacoanaphylactic endophthalmitis and sympathetic ophthalmia difficult.158,159 Onset of inflammation has been reported to vary from 1 day to 59 years after lens capsule disruption.5,152,156,157 Patients, who are often in the fifth to seventh decades of life, present with ciliary injection, photophobia, and decreased vision.5,152,156 Thach and coworkers reported a visual acuity of light perception or worse in all but one of 144 cases.152 Ocular inflammation may vary from mild to severe, with a large number of anterior chamber and vitreous cells; intraocular pressure can be variable. A hypopyon may be observed, and granulomatous keratic precipitates and posterior synechiae are generally present. Unlike traditional descriptions of inflammation limited to the anterior segment, Thach and colleagues noted inflammation of the choroid in 76% of cases reviewed, and retinal involvement in roughly half.152 Lens fragments may be observed in the anterior chamber or vitreous cavity, and ultrasound can be useful for identifying pieces of lens and surrounding inflammation.152 Inflammation may wax and wane, as in other causes of postoperative endophthalmitis, and a response to corticosteroid treatment is often noted.
The classic histologic description of phacoanaphylactic endophthalmitis is that of a zonal granulomatous inflammatory reaction centered around the lens or lens fragment.5,152,160 Polymorphonuclear leukocytes make up the inner layer, surrounded by an intermediate mantle layer of epithelioid cells, macrophages, and giant cells. The outer layer contains variable amounts of fibrous connective tissue and collagen infiltrated by lymphocytes.5,152,160,161 Histopathologic evidence of zonal granulomas is not pathognomonic, however, as bacteria can produce granulomatous inflammation as may IOL implants.5,152,162 Phacoanaphylactic endophthalmitis should be suspected in eyes with residual lens material, granulomatous inflammation, and a history of surgical or traumatic rupture of the lens capsule. Microbiologic cultures should be negative. Failure to diagnosis this condition may result in chronic inflammation leading to synechiae, retinal detachment, optic atrophy, and phthisis bulbi.152
Treatment for phacoanaphylactic endophthalmitis consists of removing lens material through pars plana vitrectomy. Phacoanaphylactic endophthalmitis can be cured after surgery.5 In some cases, topical and systemic corticosteroids are added to the treatment regimen. These drugs may help reduce inflammation, but definitive treatment is removal of the offending lens material. Compared with therapeutic management, making the correct clinical and pathologic diagnosis of phacoanaphylactic endophthalmitis remains the most challenging aspect of the disease.
Phacogenic uveitis is a nongranulomatous, nonspecific chronic inflammatory reaction to lens protein. Previously, phacogenic uveitis was referred to as phacotoxic uveitis, which is now considered a misnomer.5,163,164 Originally, it was thought to be induced by toxic material released from the lens capsule. It is now believed that phacogenic uveitis is a nongranulomatous variant of phacoanaphylactic endophthalmitis.161
Clinically, patients with phacogenic uveitis present with inflammation following surgical or traumatic rupture of the lens capsule, as in phacoanaphylactic endophthalmitis. Patients can present with inflammation weeks to months following surgery, and synechiae or hypopyon may be present.161,165 A history of retained lens fragment may be obtained and ultrasound can be useful in demonstrating opacities consistent with lens material.165 Some patients may have a significant amount of retained lens material, but inflammation may also be seen with smaller amounts of retained cortex. In contrast to patients with phacoanaphylactic endophthalmitis, keratic precipitates are usually absent, and if they are present they are typically nongranulomatous.161 Increased intraocular pressure may be present, similar to phacolytic glaucoma.161,165
Histologically, phacogenic uveitis demonstrates nongranulomatous inflammation surrounding lens material. Lymphocytes, histiocytes, and some polymorphonuclear leukocytes may be seen but the classic zonal inflammatory reaction noted in phacoanaphylactic endophthalmitis is absent.
Treatment for this condition is similar to cases of phacoanaphylactic endophthalmitis. Removal of the remaining lens material is necessary and patients may sometimes be given adjunctive corticosteroids to control inflammation. It is essential that this entity be differentiated from infectious postoperative endophthalmitis, which requires antimicrobial therapy. Intraocular culture specimens should be obtained in most cases, and intraocular antibiotics are often given if infection cannot be initially ruled out.165,166 Cases of infectious endophthalmitis in the setting of retained lens fragments and inflammation have been reported.166 This underscores the importance of investigating for both infectious and noninfectious causes of inflammation in instances of postoperative uveitis.
INTRAOCULAR LENS-RELATED INFLAMMATION
Inflammation secondary to IOL placement is much less common now than it was when they were first introduced. Early IOL polishing and sterilization techniques occasionally resulted in a toxic inflammatory response. Uveitis glaucoma hyphema syndrome was first reported secondary to poorly finished closed-loop anterior chamber IOLs that caused mechanical injury to the iris. Although this syndrome is much less common in the face of better lens finishing, it may still be seen even with posterior chamber IOLs. Both anterior chamber- and sulcus-placed IOLs can cause chafing of the iris resulting in pigment release and breakdown of the blood aqueous barrier. Liberation of iris pigment granules may simulate uveitis. True inflammation may also occur because of mechanical alteration of the blood aqueous barrier and liberation of mediators of inflammation. Transient microhyphema has also been reported secondary to sulcus placed IOLs.167–169
Clinical examination may reveal iris transillumination, and dragging or tucking of the iris may sometimes be seen. With sufficient transillumination, a portion of the IOL haptic may be seen posterior to the iris. Pigment granules may also be seen in the anterior chamber or on the posterior iris surface. The ultrasound biomicroscope may be helpful for illustrating lens chafing. The lens haptic may be seen rubbing against the iris and the haptics may be seen eroding into the ciliary body170 (Fig. 4).
Concerns about mechanical IOL trauma have resulted in the recommendation for placing IOL haptics in the capsular bag whenever possible.167,171 However, one report has recommended IOLs be placed in the ciliary sulcus as a means of decreasing postoperative posterior synechiae in patients with uveitis.172 The potential benefit of decreased syne-chiae must be weighed against the potential for increased inflammation and pupillary capture.
Inflammation related to IOL placement may be intermittent or chronic. It is important to distinguish pigment dispersion from true inflammation. If inflammation is severe and cannot be controlled with topical corticosteroids, consideration must be given to IOL repositioning or exchange.
Sympathetic ophthalmia is a rare cause of bilateral, granulomatous panuveitis, which is particularly devastating because of its potential to blind both eyes. The exact etiology of sympathetic ophthalmia is unknown, but it is hypothesized that trauma to the eye releases previously sequestered uveal antigens, which stimulate an immunologic response.173 The injured eye has been referred to as the exciting eye, and the fellow eye as the sympathizing eye. Sympathetic ophthalmia may occur following either traumatic or surgical manipulation of the eye but is more frequently encountered following ocular trauma.173–176 It is difficult to know the precise incidence of the disease because of its low occurrence, and estimates often fail to separate cases of sympathetic ophthalmia occurring following trauma or surgery.177 One study, however, reported an incidence of 0.015% following anterior segment surgery.178
Sympathetic ophthalmia has been reported to occur following nearly all types of intraocular surgery and after several nonpenetrating procedures such as cyclocryotherapy, laser retinal photocoagulation, local irradiation and Nd:YAG laser cyclophotocoagulation.173,179–183 It is most often encountered following cataract extraction, surgical manipulation of the iris and vitreoretinal surgery.175,177
The interval between ocular injury and onset of clinical inflammation has been reported from 5 days to 66 years.176,184 However, approximately 65% of patients present within 2 weeks to 3 months after injury, and 90% present within 1 year.175,176
Clinically, most patients present with insidious onset of ciliary injection, pain, photophobia, and decreased vision of the sympathizing eye. Loss of accommodation may be an early symptom in the sympathizing eye. The exciting eye is usually chronically inflamed or phthisical.174 If untreated, the disease may progress to a granulomatous panuveitis characterized by mutton fat keratic precipitates, aqueous cell and flare, iris thickening, posterior synechiae, vitritis, disc edema, retinal vasculitis, and choroidal infiltration and thickening.185,186 Intraretinal inflammation and retinal detachment may be present as well as multiple yellow-white lesions corresponding to Dalen-Fuchs nodules on histopathology.176,186 Dalen-Fuchs nodules are characteristic of sympathetic ophthalmia but are not pathognomonic and are found in about one third of patients.174,176 It has been suggested that their clinical appearance is correlated with a more severe stage of disease.187 Extraocular symptoms including alopecia, vitiligo, poliosis, and dysacusis may occur, but these are more characteristically found in Vogt-Koyanagi-Harada syndrome.176,185,186
The only known prevention of sympathetic ophthalmia is enucleation of the traumatized, exciting eye before initiation of inflammation in the fellow eye.173,174 No prophylactic measures can be taken, therefore, to prevent postoperative sympathetic ophthalmia. Enucleation of the exciting eye after contralateral ocular inflammation remains controversial. Some authors consider that enucleation does not improve final outcome, whereas others argue that enucleation of the exciting eye within 2 weeks of inflammation may improve visual prognosis.173,176,183,187 Caution should be used, however, in enucleating eyes with useful vision as the traumatized eye may later have the better visual acuity.
Medical treatment for sympathetic ophthalmia consists of aggressive immunosuppression. Corticosteroids are the mainstay of therapy. Topical steroids and mydriatics are used to control milder anterior inflammation and systemic steroids are added in more severe cases. Patients who cannot tolerate corticosteroids or who do not respond may be started on other immunosuppressive agents such as cyclosporine, cyclophosphamide, methotrexate, azathioprine, and chlorambucil.173,174,187
The advent of immunosuppressive therapy has changed the visual outlook for sympathetic ophthalmia. Previously, most eyes diagnosed with sympathetic ophthalmia had a poor visual prognosis, but good visual outcomes can now be achieved with early and aggressive therapy.173,174,176 Lubin and colleagues reported that 90% of eyes treated with corticosteroids achieved 20/100 visual acuity or better compared with only 57% of eyes not receiving steroid treatment.176 Those authors also found a correlation between final visual acuity in the sympathizing eye and severity of inflammation in the exciting eye, with over 93% of eyes with mild inflammation retaining 20/70 vision or better.176 Other authors have similarly reported good visual results with early and aggressive immunosuppressive therapy.173,187 Prompt diagnosis is, therefore, key for achieving better visual outcome, and sympathetic ophthalmia must be considered in patients with bilateral ocular inflammation and a history of surgical or traumatic manipulation of the eye.
PSEUDOPHAKIC CYSTOID MACULAR EDEMA
In 1953, Irvine first described this syndrome of macular changes and decreased visual acuity occurring after cataract extraction, and in 1966 Gass and Norton elaborated on it.188,189 Thus, cystoid macular edema (CME) following cataract surgery has been termed Irvine-Gass syndrome. This entity has been recognized as one of the most common and troublesome complications following cataract extraction, occurring after both complicated and uncomplicated cases.190–193
CME is characterized by breakdown of the blood-retinal barrier with accumulation of fluid within the sensory retina of the macula, specifically in the inner nuclear and outer plexiform layers, producing a petaloid pattern.194,195 (Fig. 5). The exact pathogenesis of CME after cataract surgery is unknown. Suggested mechanisms have included diffusion of prostaglandins and other inflammatory factors from the anterior segment to the retina, tractional forces on the peripheral retina, vitreous hyperosmolarity from rupture of the anterior hyaloid face, and excessive light exposure.195–197 Histopathologic studies by Martin and colleagues demonstrated periphlebitis in 70% of patients with Irvine-Gass syndrome, suggesting that a generalized inflammation-induced increase in capillary permeability may occur.198
Angiographic CME must be differentiated from clinical CME, the former representing patients with evidence of CME on fluorescein angiogram, but no visual disturbance, and the latter associated with reduced vision.191 Many patients demonstrate angiographic evidence of CME following uneventful cataract surgery, but most have no biomicroscopic evidence of CME and no decreased visual acuity.193,195 According to Gass, up to 70% of patients may have had angiographic CME following ICCE, with fewer than 10% having clinical CME.195 One prospective study reported the incidence of clinical CME to be 1.7% 6 months following ECCE with posterior chamber IOL.199 IOL implantation at the time of surgery does not seem to influence the incidence of postoperative CME.193
The incidence of postoperative CME has decreased with the movement from ICCE to ECCE and phacoemulsifications.193,200 Intracapsular surgery has a significantly higher rate of postoperative CME than extracapsular surgery, and extracapsular cases complicated by vitreous loss have a significantly higher rate of CME than uncomplicated ECCE cases.193,200 The incidence of CME is also greater in patients with iris prolapse, vitreous adhesion to the wound, diabetic retinopathy, and significant postoperative inflammation.195,201
Clinically significant CME typically presents within 4 to 12 weeks after surgery.195 Patients complain of blurred central vision, which is generally reduced to the range of 20/30 to 20/70.195 Visual symptoms may fluctuate and may not correlate with findings on fluorescein angiogram.191 Circumlimbal injection along with anterior chamber cell and flare and vitreous cells may be noted.195,196 Evidence of intraoperative complications such as posterior capsule rupture and vitreous adhesion to anterior chamber structures may be seen. Fundus examination may reveal foveal granularity and intraretinal cystoid spaces may be demarcated. Alternately, the fundus may appear normal. For cases in which biomicroscopic evidence of CME is not evident, fluorescein angiography is helpful for diagnosis. The practitioner must remember that CME may result from other unrecognized disorders present before cataract extraction or developing after surgery such as diabetic retinopathy, vein occlusions, hypertensive retinopathy, and retinal detachment.190,194,197,202
Most cases of postoperative CME resolve spontaneously without treatment, making it difficult to evaluate the effect of therapeutic agents. Various pharmacologic treatment modalities have been used to treat chronic postoperative CME including topical, periocular, and systemic steroids, topical nonsteroidal antiinflammatory agents, and systemic carbonic anhydrase inhibitors. Only topical ketorolac tromethamine 0.5% has been shown to significantly improve visual acuity in prospective randomized studies.203,204 For eyes in which chronic postoperative CME persists despite medical therapy, the Vitrectomy-Aphakic Cystoid Macular Edema Study Group demonstrated that statistically significant improvement in visual acuity could be achieved after vitrectomy.205 Others investigators have found similar results.200,206 Nd:YAG lasers have also been used to interrupt vitreous strands adherent to the iris or wound.207,208
3. Oshika T, Yoshimura K, Miyata N. Postsurgical inflammation after phacoemulsification and extracapsular extraction with soft or conventional intraocular lens implantation.J Cataract Refract Surg 1992;18:356
22. Javitt JC et al. National outcomes of cataract extraction. Retinal detachment and endophthalmitis after outpatient cataract surgery. Cataract Patient Outcomes Research Team. Ophthalmology 1994;101:100
28. Maylath FR, Leopold IH. Study of experimental intraocular infection. I. The recoverability of organisms inoculated into ocular tissues and fluids. II. The influence of antibiotics and cortisone, alone and combined, on intraocular growth of these organisms. Am J Ophthalmol 1955;40:86
34. Results of the Endophthalmitis Vitrectomy Study: A randomized trial of immediate vitrectomy and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis. Endophthalmitis Vitrectomy Study Group. Arch Ophthalmol 1995;113:1479
36. Bannerman TL et al. The source of coagulase-negative staphylococci in the Endophthalmitis Vitrectomy Study:A comparison of eyelid and intraocular isolates usingpulsed-field gel electrophoresis. Arch Ophthalmol 1997;115:357
76. Bell TA et al. The effect of trimethoprim-polymyxin B sulphate ophthalmic ointment and chloramphenicol ophthalmic ointment on the bacterial flora of the eye when administered to the operated and unoperated eyes ofpatients undergoing cataract surgery. Eye 1988;2(Pt 3):324
90. Gimbel HV, Sun R, DeBroff BM. Prophylactic intracameral antibiotics during cataract surgery: The incidence of endophthalmitis and corneal endothelial cell loss. Eur J Implant Refract Surg 1994;6:280
126. Fridkin SK et al. Acremonium kiliense endophthalmitis that occurred after cataract extraction in an ambulatory surgical center and was traced to an environmental reservoir. Clin Infect Dis 1996;22:222
159. deVeer JA. Bilateral endophthalmitis phacoanaphylactica. Pathologic study of the lesion in the eye first involved and, in one instance, the secondarily implicated, or “sympathizing” eye. Arch Ophthalmol 1953;54:607