Incidence of post-cataract surgery endophthalmitis in the United States has been investigated with smaller institution-based studies and larger Medicare-based studies. Medicare is a federal health insurance program in the United States. It provides coverage for approximately 50 million Americans, including virtually all people aged 65 years and older and some younger adults with permanent disabilities or end-stage renal disease. A retrospective study was based on a 5% sampling of the 1994–2001 Medicare claims identifying cataract surgeries and subsequent cases of presumed endophthalmitis occurring within the same or next calendar quarter of surgery. The incidence of endophthalmitis in the United States rose from 1.79 cases per 1000 in 1994 to 2.47 cases per 1000 in 2001, an overall increase of 37%. This increase paralleled the adoption of clear corneal wounds from scleral tunnel incisions for phacoemulsification [8]. In another retrospective study based on the Medicare database from 2006 to 2011, out of 2,261,779 cataract surgery cases, 4416 (0.195%) patients were diagnosed with endophthalmitis within 6 months of the surgery. The 0.195% rate from 2006 to 2011 was comparable to the 0.179% rate observed in 1994 in the previous study, prior to the increase to 0.274% in 2001 associated with the adoption of clear corneal wounds. This suggests that with increased experience of creating clear corneal wounds, the rate of endophthalmitis decreased from 2001 to 2006 returning to that observed with scleral tunnels. This study also reports the incidence of fungal endophthalmitis at 0.0005% (121 cases) [9]. A more recent review of 5% of Medicare claims between 2010 and 2013 revealed that 300 patients were diagnosed with endophthalmitis during the year following 216,703 cataract surgeries, which yielded an endophthalmitis rate of 0.14%, also supporting the return of the incidence of this complication at or below the levels seen at the era of scleral tunnels [10]. A smaller retrospective study based in Utah found that endophthalmitis was diagnosed in 26 of 9079 cataract surgeries (0.286%) performed between 1997 and 2001 at the Moran Eye Center [11]. When looking at a longer period in the same center, from 1997 to 2007, the rate of endophthalmitis decreased to 0.157%, with 46 cases of endophthalmitis out of 29,276 cataract surgeries performed during a 10-year period, once again suggesting that with increased experience with clear corneal wound construction, the incidence of endophthalmitis decreases to a baseline number [12]. The incidence of post-cataract surgery endophthalmitis is shown in Table 6.2.

The majority of cases of postoperative endophthalmitis were caused by gram-positive organisms that are normal flora of the eyelid and conjunctiva. These bacteria may gain access to the intraocular space either through direct inoculation during surgery or due to migration of local flora into an incomplete wound closure postoperatively. In a prospective study consisting of 700 consecutive patients undergoing planned extracapsular cataract extraction, anterior chamber aspirates were culture positive in 14.1% at the beginning and in 13.7% at the end of surgery, despite the use of povidone-iodine 10% antisepsis; coagulase-negative staphylococci and Corynebacterium were the most common isolates [13]. In a smaller study on 113 patients undergoing cataract surgery, two patients (1.8%) showed growth in culture of the aqueous humor sampled at the end of the surgery, despite the use of povidone-iodine antisepsis. Fortunately, no patient developed endophthalmitis [14]. These studies suggest that host factors can clear a low inoculum of bacteria in the anterior chamber after cataract surgery without developing endophthalmitis. The increased endophthalmitis rate with posterior capsular defects suggests that the body cannot clear a bacterial inoculum in the vitreous cavity as effectively as in the anterior chamber.

In 1995, the Endophthalmitis Vitrectomy Study addressed the management of endophthalmitis following cataract surgery, which was performed by extracapsular extraction. It remains today the prospective study with the largest number of endophthalmitis patients. Among the 422 patients, vitreous cultures were positive in 69.3% of cases and 9.3% presented with polymicrobial growth. The most common bacteria were Staphylococcus epidermidis in 70% of bacterial isolates, Streptococcus species in 9.0%, Staphylococcus aureus in 9.9%, and enterococci in 2.2%. Gram-positive bacteria represented 94% of isolates, with 5.9% gram-negative species. All gram-positive species were sensitive to vancomycin [15]. Later studies, during the clear cornea wound phacoemulsification era, reported similar microbiological spectra for endophthalmitis following cataract surgery. A retrospective study of 502 endophthalmitis patients, selected using the 2003–2004 Medicare database, found culture yield to be 58% (lower than 69.3% in EVS), with coagulase-negative Staphylococcus in 45% of isolates and Streptococcus species in 12% of isolates. Gram-positive bacteria represented 93% of isolates. This study also reported that patients with Streptococcus were ten times more likely to have poor visual outcomes than those with coagulase-negative Staphylococcus. Worse visual outcomes were similarly noted when comparing patients with gram-negative bacteria to those with gram-positive ones. Finally, a smaller difference in poor visual outcomes was also noted between patients with culture-positive and culture-negative vitreous [16]. Another retrospective study reviewed 73 patients presenting with endophthalmitis at Bascom Palmer, within 6 weeks of cataract surgery from 1996 to 2005. Coagulase-negative staphylococci were isolated in 68.4% of eyes, Streptococcus species in 8.2%, and Staphylococcus aureus in 6.8%. Worse visual acuity outcomes were noted for infections caused by Staphylococcus aureus or Streptococcus species compared to those caused by coagulase-negative staphylococci [17]. The uniform microbiologic spectrum in these North American studies may sometimes contrast with the spectra reported on other continents. For instance, an institution-based retrospective study in Taiwan from 2004 to 2015 found that among 32 patients that developed endophthalmitis following cataract surgery, the most common isolates were Enterococcus species at 38.1%, Staphylococcus epidermidis at 28.6%, and Staphylococcus aureus at 9.5% of isolates [18]. Microbiological spectrum of post-cataract surgery endophthalmitis is shown in Table 6.3.

Treatment of this sight-threatening disease has historically consisted of administration of intravitreal, subconjunctival, and intravenous antibiotics, with or without intravitreal or oral corticosteroids to minimize inflammatory damage, and drainage of the vitreous abscess by pars plana vitrectomy. The Endophthalmitis Vitrectomy Study (EVS) is the major landmark evidence-based trial, which established treatment criteria for this condition. This prospective multicenter randomized clinical trial studied the treatment of endophthalmitis developed within 6 weeks of cataract surgery in patients who presented with vision between 20/50 and light perception (LP), without a history of comorbidities which could reduce their visual potential. All 420 patients received intravitreal vancomycin to cover gram-positive organisms and amikacin to cover the gram-negative ones, as well as subconjunctival dexamethasone, vancomycin, and ceftazidime. Patients were randomized to receive additional immediate pars plana vitrectomy or administration of intravenous antibiotics. The results determined that immediate vitrectomy would only benefit patients with LP, while in those with hand motions (HM) or better vision, using intravitreal antibiotics without vitrectomy would provide a similar long-term visual outcome. Moreover, the use of intravenous antibiotics provided no additional benefits to the intravitreal treatment. In the subgroup of diabetic patients, however, those who had HM or better vision also appeared to benefit from immediate vitrectomy as 57% of them achieved 20/40 vision, whereas only 40% did so without vitrectomy [19].

The mainstay of post-cataract endophthalmitis treatment in North America remains close to the one recommended two decades ago by the EVS study. Patients presenting with LP vision or worse undergo emergent pars plana vitrectomy, while those presenting with HM vision or better undergo the less invasive vitreous tap instead. All patients receive empiric intravitreal antibiotic injections, which most often include 1 mg vancomycin to cover gram-positive organisms and 2.25 mg ceftazidime for gram-negative organisms. The latter can be substituted with 0.4 mg amikacin in patients allergic to beta-lactams, although there have been reports of retinal infarction with aminoglycosides at therapeutic dosages. While all patients received subconjunctival antibiotics in the Endophthalmitis Vitrectomy Study, these have been dropped from standard treatment in North America over the past 20 years. In one retrospective study between 1991 and 2002, the final visual outcome of 43 patients presenting with HM vision and acute post-cataract endophthalmitis was similar whether subconjunctival antibiotics were added to the intravitreal ones or not. Moreover, the visual outcomes were comparable to those of the EVS patients [20]. Similar findings regarding the use of subconjunctival antibiotics were reported for treatment of endophthalmitis secondary to trauma, cataract, or glaucoma surgery in a retrospective study of 54 patients treated at Bascom Palmer from 1995 to 2002. This lack of additional effect occurred despite the nonrandomized nature of these trials where the subconjunctival antibiotics may presumably have been used in eyes with more severe disease, as the eyes who did not receive them had a lower rate of enucleation or absent LP outcomes [21]. With the improvement of vitrectomy technology over the past 20 years, allowing safer cutting close to the retina and better intraoperative viewing, more complete vitrectomies are performed, contrasting with the limited vitreous removal suggested in the EVS protocol prohibiting posterior vitreous detachment induction and advising “to remove at least 50% of vitreous gel in eyes with no vitreous separation.” In a consecutive series of 47 eyes, which underwent complete vitrectomy for endophthalmitis with similar inclusion/exclusion criteria to the EVS, 91% achieved ≥20/40 final visual acuity, as opposed to a 53% rate in the EVS (p < 0.0001, Fisher’s exact test). No serious adverse effects developed such as retinal detachment and phthisis bulbi or indications for enucleation. There was no case of anatomical failure, as opposed to the EVS with an 11% rate in the nonsurgical group and a 5% rate in the vitrectomy group [22]. Whether early vitrectomy in eyes with hand motions or better vision provides a better outcome by removing harmful agents and inflammatory mediators from the vitreous cavity could benefit from a randomized clinical trial. An indication of expected results could be found in a Medicare-based retrospective study. Across the five states in the study, the use of vitrectomy varied significantly in patients with better than light perception vision. Rates of vitrectomy in such patients ranged from 19% in Michigan to 56% in California, although no evidence was found that this was associated with better visual outcomes [16]. The good bioavailability of oral moxifloxacin following two or five orally administered 400 mg tablets, with obtained intravitreal drug concentrations exceeding the MIC90 (minimal inhibitory concentration in which 90% of isolates were inhibited) of most bacteria responsible for endophthalmitis, would also merit revisiting in future studies addressing the use of systemic antibiotics in the treatment of endophthalmitis [23–25].

In order to reduce the risk of endophthalmitis following cataract surgeries, varied treatments have been attempted pre-, peri-, and postoperatively. Given the low incidence of endophthalmitis, an exceedingly large number of patients would be required for a treatment study to be powered to demonstrate a statistically significant effect. A comprehensive review of studies published between 1966 and 2000 found only perioperative povidone-iodine antisepsis to be effective at reducing endophthalmitis rates. Subconjunctival antibiotics, topical antibiotics, antibiotics inside irrigating solution, and lash trimming did not present conclusive evidence of further reducing this risk [26]. Despite this, many American surgeons prescribe antibiotic drops in the pre- and postoperative period in order to reduce the bacterial load and potential inoculum through the surgical wound. A retrospective study at the Moran Eye Center in Utah found topical ofloxacin postoperative use between 1997 and 2001 was more beneficial than ciprofloxacin. While the use of both antibiotics was equal during that period, 85% of endophthalmitis cases developed in patients under topical ciprofloxacin and 15% of them in patients under ofloxacin. The difference between antibiotics was significant (p < 0.00026) and may have been due to better penetration of topical ofloxacin into the anterior chamber and a lower kill time for this medication [11]. The replacement of these third-generation agents by newer fourth-generation fluoroquinolone antibiotics prompted a second retrospective study at the Moran Eye Center, from 1997 to 2007. The use of moxifloxacin and gatifloxacin eye drops from 2003 to 2007 was associated to a lower rate of endophthalmitis of 0.056% when compared with the 0.197% rate under ciprofloxacin and ofloxacin eye drop use from 1997 to 2003 (p = 0.0011). When looking at individual agents, the 0.015% rate with gatifloxacin was lower than the 0.1% rate with moxifloxacin (p = 0.04) [12]. With the increase in endophthalmitis isolate resistance to fluoroquinolones identified in New York and Florida over the past decades, the benefits of these topical antibiotics as prophylactic treatment may prove to be short-lived however [5, 6]. The use of intracameral cefuroxime at the end of cataract surgery reduced the occurrence of postoperative endophthalmitis by an odds ratio of 4.92 (p = 0.001) in a European prospective randomized study of 16,603 patients undergoing cataract surgery from 2003 to 2006. The study reported rates of culture-proven infectious endophthalmitis at 0.07% in the groups receiving intracameral cefuroxime prophylaxis compared with rates of 0.34% in the control groups not receiving intracameral cefuroxime and was stopped ahead of targeted enrolment once this benefit became apparent [27]. Concerns were raised however with the limited coverage against gram-negative bacteria and poor coverage against methicillin-resistant Staphylococcus epidermidis and Staphylococcus aureus. One consideration to keep in mind is the routine use of intracameral cefuroxime, moxifloxacin, or vancomycin as a prophylactic treatment could lead to increased resistance and sacrifice the benefits of these agents as first-line treatment.