In most cases of CPE, patients are referred for the management of a cortico-dependent anterior uveitis, as noted in 4/7 of our patients. According to Fardeau et al, the partially response to steroid represented one of the item of the definition of CPE (9). Typical clinical presentation is unilateral granulomatous uveitis involving the anterior chamber with a possible later progression into the vitreous. Most frequent ocular symptoms are a recurrent eye pain, a progressive decrease vision, a red eye, isolated or associatedto eye pain. An anterior chamber inflammation was present in 85,7% of our cases, with different grade of inflammation. Hypopion was noted in one of our cases, this clinical feature was reported by several authors in approximately 46% of the cases of CPE (8, 11-14). A white plaque on the anterior surface of the IOL or on the posterior capsule, was noted in all our patients, this finding was reported in 28.5% to 100% of CPE (11,14,15) and was found to be mostly associated with P.acnes (12,16–20). Others clinical features were noted in literature, suggesting fungal infection, such as a stirringly white infiltrates or clumps in the anterior chamber (16,21,22).
In this series the mean interval between cataract surgery and the diagnosis of CPE was 30.7 weeks. This interval was extremely variable in literature: 2 weeks, 3 weeks, 6 weeks (9,14,23,24), 3 months (12,13) or 6 months (25) . The mean time interval between cataract surgery and the diagnosis of CPE was relatively shorter in Onchobactrum Anthropi (O.anthropi) endophthalmitis and non tuberculous mycobacterium than in P. acnes endophthalmitis or Pseudomonas oryzihabitans, 6.8, 2, 36 and 16 weeks reported by previous studies, respectivelly (21,26) . This difference of delay, and the onset of clinical manifestations are probably related to average time of growth of the microorganisms (18,26,27).
In our study, AH culture was negative in 6 out of 7 patients, and only one case was culture-proven CPE, with an IOL microbiological diagnosis. The most cases series in literature provide from culture-proven cases from clinical and microbiology laboratory database (8,28–31). However, a negative culture result does not necessarily imply a bacteria-free infection (22,26,32). These results might be due to the nature of the cases being referred to our institution. These patients had been treated before being referred, thus the initial microbial profile might have been altered (33) . In addition, most of the microorganisms responsible for CPE are widely distributed in the natural environment, such as the water sources for O.anthropi, water and soil for Alacaligines faecalis or eyelid for P.acnes (22,23,34). They are generally not virulent but the production of a biofilm on the artificial devices, such as in the surface of the IOL or on its haptics, may lead to these sequestration of microorganisms into the CB (1,13,17,17–20,32–36). Furthermore, vitreous taps have a higher rate of culture positivity than AS. However, in cases of negative cultures of both samples, the best result is the culture of the removed CB and IOL, as found in the single culture proven case of our series. Identification of the microorganism may need special culture media and prolonged incubation time. An aerobic and fungal culture is highly recommended (14,15). More recently, the role of molecular testing by polymerase chain reaction or “PCR”, is essential, and allows a microbiological diagnosis in 71% of cases of postoperative acute and delayed-onset endophthalmitis as demonstrated by Chiquet (37–39). However, diagnosis based on clinical findings should be performed, without waiting for the microbiological results the to treat (22,29). E.faecalis was the only identified microorganism in our study. It is a gram positive bacterium that is part of the normal human gastrointestinal track flora (40,41). It is a relatively rare cause of endophthalmitis, found in 1.23% of acute post-cataract surgery endophthalmitis cases in the Endophthalmitis Vitrectomy Study (41,42). In a reported case series of E.faecalis endophthalmitis, the onset of clinical signs was within 4 days in 61,53%, between 4 days and 6 weeks in 7,69% and after 6 weeks in 19,23% of cases (36,41). As we noted in our series, it is usually related to a poor visual outcome with only 15% achieving a visual acuity better than 6/60, probably related to the bacterial virulence. Table 3 describes reported microbiological proven CPE after PKE and IOL implantation in literature (8,23,24,28,30–32,39,43–70)
The management of CPE is controversial. The sequestration of microorganisms into the CB, their different virulence proprieties and the possibility of polymycrobial infection have made it difficult to establish a unique protocol treatment (13,45). However, regardless of the clinical presentation and its severity, its management has to be prompt. A sample of intraocular fluid for microbiological investigation is mandatory in any suspected CPE before initiating treatment. In our series, the treatment of CPE was based on a “step by step approach”. The first line treatment was medical, followed by the IOL-conservative surgical treatment, and the non-conservative-IOL surgical treatment, as recommended in literature (9,10,12). As described by Güler and Aldave, IOAB were based on a “two-compartment approach” that included a simultaneous injection in the humor aqueous and vitreous (71,72). Vancomycine (1mg/0.1 mL) and ceftazidim (2,25mg/0.1 mL) were used for empiric coverage of gram-positive and gram-negative organisms in the primary procedure (12,73,73). Ciprofloxacin (0.2 mg/0.1 mL) was used in non-responding cases and/or resistance to Ceftazidim (49,56). We performed an irrigation and washing of the CB, associated to the IOAB, as recommended in some studies (9,12). The use of SATB remains controversial (13,15,21,74,75). The slow growth of most common microorganisms isolated in CPE and their sequestration in the CB justify the need of repeated IOAB injection as we observed in cases n°3, 5, 6. (9,12,49,71). The association of steroids to IOAB injection and SATB administration depends on severity of inflammation and are proposed in different routes of administration. When inflammation recurs or increases, surgical treatment based on PPV is required to eliminate the contaminating foci, and may include two types of approaches, depending on whether the IOL is explanted or not. The IOL-conservative surgical treatment is generally associated to capsulotomy of the most infiltrated areas of the CB, and the IOL is preserved (1,10). This treatment strategy allows the removal of localized infectious sources while leaving enough capsular support for the IOL(26) . If the conservative-IOL surgical treatment is not efficient, as observed in our case 1, removal of the entire CB and the IOL is required to eradicate all sites of intraocular infection (9,13,15,17,19,23,26,27,34,43–48,73,76). Surgical treatments were observed in 50% of this case series, which agrees with other studies that reported it in approximately 30% to 73% of cases (12,15,16,27,41). A definitive initial surgical procedure should be considered in any patient with strong clinical evidence of severe or refractory CPE, or when the clinical features are suggestive or microbiologically proven aggressive microorganisms, such as fungal infection, Onchobactrum Anthropi, P.Acnes, or polymicrobial infection (15,22,26,34,45,52,72–74,77).
In this series, the mean final BCVA was 20/160 with a gain of 4 lines (ranging from 1/80 to 20/32), poor visual outcome was observed in case 2 and 4 where postoperative complications were noted such as CME (case 4) and ERM (case 2). The visual prognosis of CPE is various from one report to another, with a better visual prognosis than acute-onset endophthalmitis (16). Hsu et al noted that a long incubation (>1 month) would be associated with favorable visual outcomes compared to acute cases (18). However, CPE can lead to poor visual outcome despite the two-compartment approach of IOAB, and safe technics of complete non-conservative IOL surgical therapy, mainly related to the causative organism, specially fungal and/or nontuberculous mycobacterium infection (18,33,74). In another hand, polymycrobial infection has been associated with failure of IOL- conservative treatment (14,26,48,56,78).
Based on the literature, and the results observed in this study, we propose the following algorithm for the management of CPE (Table 3). First line treatment is medical, given at presentation and repeated if needed based on IOAB and associated to an AC and CB wash. If inflammation persists or recurs, the second step is an IOL-conservative surgical treatment, based on PPV, associated with CB partial removal and IOAB injection. The third step is a non-conservative-IOL surgical treatment, based on an additional PPV, associated to residual CB removal and IOL explantation. At any step of treatment, we start with an ocular sample for microbiological analyses, whether it concerns the intraocular fluids, humor and vitreous, CB or/and IOL when they have been removal. The surgical treatment is proposed as first line approach in particular situations (Figure 3).
The limitations of the current study include its retrospective design, a relatively small number of patients, the absence of performance of vitreous tap and inclusion of cases with negative aqueous humor cultures. Nevertheless, we addressed non-proven culture CPE after uneventful PKE, an issue that is encountered in our daily practice and we tried to extrapolate from our experience an algorithm that can help to manage such an entity.