Chronic post-op endophthalmitis is diagnosed in cases that present beyond 6 weeks of the inciting event. The clinical picture is usually more subtle and is often misdiagnosed and treated as a sterile anterior uveitis. The signs are usually more subtle, including anterior chamber reaction responsive to corticosteroids but recurrent after cessation of treatment, granulomatous keratic precipitates, IOL precipitates, endocapsular hypopyon (Fig. 10.3), vitritis, or simply as an isolated disk edema [1, 22, 28]. The most common microorganisms involved are coagulase-negative S. epidermidis, Propionibacterium acnes (P. acnes), fungi (e.g., Candida parapsilosis), anaerobic Streptococcus species, Nocardia asteroides, Actinomyces sp., and Corynebacterium sp. [22]. Visual outcomes are generally better than acute cases because these are lower-virulence microorganisms [1, 3, 22, 29]. In P. acnes chronic endophthalmitis, the endocapsular hypopyon represents colonies of the microorganism and inflammatory cells situated between the posterior surface of the IOL and the posterior capsule (Fig. 10.3). This peculiar location is thought to protect the bacteria from antibiotics and host defense mechanisms and could explain the high recurrence rate in cases where the lens and capsule are left in place [22]. In large series evaluating the best treatment option, the following are the recurrence rates of different techniques [22, 29, 30]:

The antibiotic of choice is intravitreal vancomycin 1 mg/0.1 ml. The average time for growth in cultures was between 8 and 10 days [22].

The result of trabeculectomy surgery for glaucoma is a filtering bleb over the trabeculectomy site providing an alternative drainage route of aqueous humor, thus reducing IOP. This bleb should be water tight to function properly and prevent bacterial access to the intraocular environment (Fig. 10.4). The conjunctiva carries normal flora capable of causing post-op endophthalmitis including S. epidermidis, S. aureus, group C Streptococcus, Bacillus sp., Haemophilus influenzae (H. influenzae), and P. aeruginosa [31]. It has been documented that leaky filtering blebs allow access of the tear film into the AC [32] along with the ocular surface bacteria.

Blebitis is diagnosed when the filtering bleb becomes infected without an associated vitritis; the risk of progressing into an endophthalmitis is high if not treated urgently and properly. Bleb-associated endophthalmitis (BAE) is diagnosed when blebitis is associated with vitritis and intraocular inflammation [1]. In such cases, the most common symptoms are pain and redness. Other symptoms include blurred vision, tearing, purulent discharge, lid edema, and photophobia [1, 33] (Fig. 10.5). Risk factors include the use of 5-fluorouracil or mitomycin-C, bleb leaks, conjunctivitis, bleb manipulation, inferior location of bleb, trauma, and contact lens use [33–37]. The incidence of BAE ranges from 0.2 % to 9.4 % which is higher than other intraocular surgeries and is believed to be largely due to the use of antifibrotic agents [33–35]. Bleb-associated endophthalmitis is divided into early, if presented within 6 weeks post-op, or late (may happen even many years after surgery). In the early period, the most common microorganisms are S. epidermidis and S. aureus, while in late cases Streptococcus sp. and H. influenzae are more frequently cultured [34, 36, 37]. Cultures were found to be positive in 67–97 % of cases [34]. Visual acuity outcome is generally poor with final VA better than 20/400 ranging between 31 % and 65 %. Outcomes post-Staphylococcal endophthalmitis seem to be more favorable than post-Streptococcalcendophthalmitis [34, 36]. The EVS did not include patients undergoing filtering glaucoma surgery, and extrapolation of its data should be done carefully. Although no randomized controlled clinical trial has been conducted to date, the surgical outcomes for BAE seem to be better with prompt pars plana vitrectomy and intravitreal antibiotics than with vitreous tap and intravitreal antibiotics regardless of time of onset and pre-op VA [36]. Besides intravitreal antibiotics, topical fortified drops (vancomycin 50 mg/ml and gentamicin 15 mg/ml) and oral fourth-generation fluoroquinolones are used to treat the infected bleb [35, 37, 38]. If a leaky bleb is present in association with blebitis and endophthalmitis, it is recommended to perform a bleb excision and conjunctival advancement in addition to pars plana vitrectomy with topical and intravitreal antibiotics [36, 39].

The number of intravitreal injections performed has increased significantly over the last decade, and, accordingly, the rate of post-intravitreal injection endophthalmitis (IVIE) has also increased. In a recent review by Moshfeghi et al., the rate of IVIE was 0.02–0.14 % [40]. The most common organisms are S. epidermidis, S. aureus, Streptococcus sp., and, usually in case of drug contamination, fungal species. The rate of postinjection Streptococcal endophthalmitis is three times higher than that in post-op cases [41]. This could be explained by the fact that surgical masks were not routinely worn during the office-based procedure. In an experiment simulating the setting of intravitreal injections, oral flora was found to grow on culture plates significantly higher when the surgeon and the patient were not wearing surgical masks than in cases when they wore them, avoided talking, or turned their faces away while talking; the rate of Streptococcal sp.-positive cultures was found to be 67–83 % [42, 43]. In another study, they compared bacterial growth on culture plates while wearing surgical masks, in silence, and with pretreatment of the plates with 5 % povidone-iodine without surgical masks or silence. The least growth was found on the povidone-iodine pretreated plates [43]. In comparing office-based setting versus operating room-based setting, there does not seem to be a significant difference in the rate of endophthalmitis or the causative microorganisms when surgical masks are used during office-based intravitreal injections [44]. Another important cause of IVIE is the contamination of the medications prepared at compounding pharmacies. Several outbreaks have been reported with mainly Streptococcal or fungal species resulting in very poor outcomes [45–47]. If contamination is suspected, testing of unused syringes containing the medication from the same lot is warranted, and it should be reported to the local authorities.

We conducted a retrospective chart review of acute endophthalmitis (AE) post-intravitreal bevacizumab injection in cases done by one surgeon. The surgeon, nurses, and patients wore surgical masks during the procedures done in the minor procedure room at the Department of Ophthalmology ambulatory clinic, Maisonneuve-Rosemont Hospital, University of Montreal, Quebec, Canada. The rate of IVIE was found to be 0.13 % (4 of 3,088 injections); two were culture-positive for coagulase-negative S. epidermidis, and two had no growth (Table 10.1). All patients underwent tap and injection of vancomycin 1 mg/0.1 ml and ceftazidime 2.25 mg/0.1 ml upon diagnosis of IVIE. Three of them underwent a later 25-gauge (25g) MIVS to clear vitreous opacities. The fourth patient developed a central retinal artery occlusion during the period of recovery from IVIE and refused to undergo further procedures. All four patients improved after tap and injection of antibiotics, two out of four (culture-negative) patients improved better than their VA before endophthalmitis had developed, and three out of four continued to receive intravitreal bevacizumab (IVB) without any further infectious complications [48].

The outcomes of IVIE also depend on the causative organism, where coagulase-negative S. epidermidis cases or culture-negative cases have a better final VA outcome than post-Streptococcal cases. The final VA post-IVIE when surgical mask is used was found to be as favorable as post-phaco and IOL implantation acute endophthalmitis with a good percentage of patients returning to baseline VA [49]. Treatment may be guided by the EVS study, but surgeons should carefully extrapolate these results and decide on each patient’s treatment individually. In this particular entity, we prefer to start the treatment with tap/inject to try to preserve the vitreous as the pharmacokinetics of intravitreal injections may be significantly altered in vitrectomized eyes (Fig. 10.6) [50].

This clinical entity is rare and carries poor prognosis for the affected eye. More detailed discussion will follow in the MIVS section.

Type-1 Boston permanent keratoprosthesis is a polymethylmethacrylate device in a collar button shape that is used to restore the clarity of the visual axis in cicatricial corneal diseases that are not amenable to conventional corneal transplant [51] (Fig. 10.7). This procedure allows a connection between the sterile anterior chamber environment and the ocular surface with its flora and external pathogens [52]. In a review of literature from 2001 to 2011 by Robert et al., the prevalence of post-permanent KPro endophthalmitis (KPE) was found to range from 0.0 % to 12.5 %. Risk factors include Steven-Johnson syndrome, ocular cicatricial pemphigoid, noncompliance with antibiotics, nonuse of vancomycin, poor contact lens hygiene, corneal melting, and infectious keratitis. The most common causative agents are gram-positive bacteria (60 %) [52] with one study reporting that 10 out of 15 gram-positive cultures grew Streptococcus viridans [53]. Other organisms include Staphylococcal sp., gram-negative bacteria, and fungi (Fig. 10.8). The most important prevention method is using vancomycin drops in addition to the traditional quinolones; this has significantly reduced the prevalence of KPE [52, 53]. This entity tends to present in the late post-op period, and the outcomes depend on the causative microorganism but are usually poor due to the higher prevalence of Streptococcus sp. A high index of suspicion is important in cases of retroprosthesis membrane and vitritis with or without pain, and B-scan ultrasound is an important diagnostic tool in cases of poor visualization [54, 55].

The diagnosis of endophthalmitis is mainly done on clinical examination, but valuable information can be obtained from A- and B-scan ultrasonography of the affected eye at presentation. Echographic findings include vitreous opacities; vitreous membranes (Fig. 10.9); vitreous incarceration; location of persistent vitreous attachments; vitreoretinal traction; retinal detachment (RD) (Fig. 10.10) and its extent; subretinal opacities; presence of choroidal detachment (CD) and its size, location, and type (serous vs. hemorrhagic); retinochoroidal thickening; choroidal calcification; optic nerve swelling; scleral thickening; lens status; IOL location; retained nuclear fragments; globe size; and orbital inflammation [19, 56, 57]. The gain should be increased between 84 and 90 dB, as inflammatory cells and membranes have low reflectivity [58]. Several study groups found significant negative correlation between these signs and the final VA [19, 57, 59].

A study by Dacey et al. reviewed records of 137 endophthalmitis patients. They found that 69 % (24/35) of eyes with vitreous opacities of severe density had a final VA worse than 1/200 versus 33 % (10/30) of eyes with mild opacities. Also, 91 % (30/33) of eyes with final VA of NLP had moderate to severe opacities density on the initial echography. When they compared the final VA of eyes that had vitreous membranes on the initial echography (20/137) to those that had none, 55 % (11/20) versus 25 % (29/116) had final VA of NLP, respectively. In eyes that had RD on initial echography (25/137), none had final VA better than 20/200. Fifty-six percent of eyes with RD had final VA of NLP versus 21 % of eyes without RD. Of note, the status of the macula (macula-on vs. macula-off) had no statistical significance on final VA. The coexistence of choroidal detachment on presentation had a significant impact on final VA. Eighty-one percent (22/27) of eyes with CD on the initial ultrasound had final VA worse than 1/200 versus 47 % (50/106) of eyes that had no CD. Choroidal detachment was the strongest indicator of final VA worse than 20/100. They also found a correlation between B-scan findings and the infecting microorganism. Streptococcal endophthalmitis cases were at least twice more likely to develop severe vitreous opacities or vitreous membranes on the initial presentation than other infecting microorganisms (56 % and 45 %, respectively). Gram-negative endophthalmitis was significantly correlated with the presence of choroidal detachment but with less significant vitreous opacities [19]. Another valuable use of pre-op ultrasound is to confirm the location of retinal and/or choroidal detachment in relation to the area of intended sclerotomy placement, for either a vitreous sample or surgery, thus avoiding more potential serious complications [58].