Purpose
To compare the incidence and clinical outcomes of endophthalmitis following intravitreal injections of bevacizumab, ranibizumab, and aflibercept.
Design
Multicenter, retrospective cohort study.
Methods
All included patients had received intravitreal injections of bevacizumab, ranibizumab, or aflibercept between January 1, 2009 and September 30, 2013 at 5 retina practices. Billing records were used to identify the total number of anti–vascular endothelial growth factor (VEGF) injections administered. Patients who developed endophthalmitis were ascertained from endophthalmitis logs and billing records. Chart review of these patients was performed to confirm that the endophthalmitis was related to the antecedent anti-VEGF injection. Visual outcomes, causative organisms, and clinical course were also recorded.
Results
A total of 503 890 anti-VEGF injections were included, from which 183 cases of presumed endophthalmitis were identified. The rate of endophthalmitis for bevacizumab was 0.039% (60/153 812), which was similar to ranibizumab 0.035% (109/309 722; P = .522) and aflibercept 0.035% (14/40 356; P = .693). Similarly, there was no difference in the rates between ranibizumab and aflibercept ( P = .960). The culture-positive rate of the vitreous/aqueous tap was 38% for both bevacizumab and ranibizumab and was 43% for aflibercept. Furthermore, visual acuity remained decreased at 3 months follow-up for bevacizumab ( P = .005), ranibizumab ( P < .001), and aflibercept ( P = .07) compared to vision at causative injection.
Conclusions
Endophthalmitis following intravitreal bevacizumab, ranibizumab, and aflibercept injection appears to occur at similar rates and have comparable visual outcomes. This study suggests that the choice of anti-VEGF agent should be primarily based on efficacy and patient response rather than concern for risk of infection.
Anti–vascular endothelial growth factor (VEGF) injections have become the standard of care for treating patients with neovascular age-related macular degeneration (AMD), diabetic macular edema (DME), and macular edema owing to retinal vein occlusion (RVO). The 3 most commonly used agents are bevacizumab (Avastin; Genentech Inc, South San Francisco, California, USA), ranibizumab (Lucentis; Genentech Inc), and aflibercept (Eylea; Regeneron Inc, Tarrytown, New York, USA). Endophthalmitis is a severe potential complication following intravitreal injection and can cause significant visual loss. Fortunately, the rate of endophthalmitis is low, with reports in the literature ranging from 0.01% to 0.08%. A recent meta-analysis of 43 articles reported an overall incidence of endophthalmitis at 0.056% following anti-VEGF injections. The study found that the most commonly isolated organisms were coagulase-negative Staphylococcus and Streptococcus species.
Despite the growing use of anti-VEGF agents, there is limited evidence as to the relative safety in regard to risk of endophthalmitis among the 3 commonly used anti-VEGF agents. The purpose of this multicenter study was to compare the incidence of postinjection endophthalmitis among bevacizumab, ranibizumab, and aflibercept, as well as to assess visual outcomes and causative organisms.
Methods
This multicenter, retrospective cohort study received approval from the institutional review board (IRB) at Wills Eye Hospital and central Western IRB. The participating centers in this study include: The Retina Service of Wills Eye Hospital, Mid Atlantic Retina, Philadelphia, Pennsylvania, USA; Associated Retinal Consultants at William Beaumont Hospital, Royal Oak, Michigan, USA; Retina Consultants of Houston, Houston, Texas, USA; Ophthalmic Consultants of Boston, Boston, Massachusetts, USA; and Southeastern Retina Associates, Chattanooga, Tennessee, USA. Billing records and endophthalmitis logs were used to identify patients who developed endophthalmitis following anti-VEGF injections between January 1, 2009 and September 30, 2013. The total number of injections of bevacizumab, ranibizumab, and aflibercept administered for neovascular AMD, DME, proliferative diabetic retinopathy (PDR), branch RVO with macular edema, or central RVO with macular edema was determined from billing records.
Patients’ charts were subsequently reviewed to confirm that the endophthalmitis was linked to the preceding anti-VEGF injection. Furthermore, data on each patient’s visual acuity (VA), indication for anti-VEGF injection, date of causative injection, anterior/vitreous chamber tap and injection of antibiotics or pars plana vitrectomy (PPV), and culture results were recorded.
Inclusion and Exclusion Criteria
All patients diagnosed with presumed infectious endophthalmitis following an intravitreal injection of either bevacizumab, ranibizumab, or aflibercept were included in this study. Endophthalmitis was defined as patients who presented with a clinical suspicion that was high enough to warrant either a vitreous tap (or anterior chamber tap if vitreous fluid was not able to be obtained) and injection of antibiotics or PPV. In general, these patients had presented with decreased visual acuity and pain, and had signs of intraocular inflammation on examination (generally ≥2+ anterior segment cellular reaction and/or posterior segment vitritis) within 7 days of the causative injection. Furthermore, patients were only included if the indication for anti-VEGF injection was 1 of the following diagnoses: neovascular AMD, DME/PDR, or macula edema owing to branch or central RVO. Patients were excluded if they had postinjection inflammation treated with topical steroids rather than an intraocular tap with injection of antibiotics.
Injection Technique
All anti-VEGF injections were administered in an office-based setting. Bevacizumab was repackaged into syringes at compounding pharmacies and distributed to the participating sites. On the other hand, ranibizumab and aflibercept syringes were loaded from single-use vials during the office visit. Eyes typically received 2 cycles of topical anesthesia and 5% povidone-iodine (Betadine 5%; Alcon Labs, Fort Worth, Texas, USA). In some patients, additional 1%–2% subconjunctival lidocaine followed by 5% povidone-iodine was also administered. The use of a sterile speculum, choice of quadrant for injection, and conjunctival displacement were performed at the discretion of the treating physician. A 30 or 31 gauge needle was used to perform the injection 3.5–4.0 mm from the limbus. During the initial period of the study, patients were routinely prescribed postinjection topical antibiotics for prophylaxis. This was followed by a transition period in which some physicians continued prescribing prophylactic topical antibiotics. Finally, toward the end of the study frame, antibiotics were not routinely prescribed.
Statistical Analysis
The primary outcome of the study was to compare the incidence of postinjection endophthalmitis among bevacizumab, ranibizumab, and aflibercept. Pearson χ 2 analysis was used to determine if there was a statistically significant difference in the rates of endophthalmitis among the 3 agents. Odds ratios (OR) were also calculated to compare the odds of developing endophthalmitis among the anti-VEGF agents. Visual acuity was converted to logarithm of the minimal angle of resolution (logMAR) values for analysis. Patients with visual acuity of counting fingers or worse were converted to logMAR units as previously described. Subsequently, paired 2-tailed t test analysis was performed to determine visual outcomes at the time of diagnosis and at 3 months follow-up for each agent. Statistical analysis was performed using GraphPad software (GraphPad, La Jolla, California, USA).
Results
A total of 183 cases of endophthalmitis were identified from 503 890 anti-VEGF injections (1/2753 injections, 0.036%). Baseline characteristics for the 3 anti-VEGF agents are summarized in Table 1 . From a total of 153 812 bevacizumab injections, 60 cases (1/2563 injections, 0.039%) of endophthalmitis occurred. For patients receiving ranibizumab injections, a total of 109 cases of postinjection endophthalmitis were reported from 309 722 ranibizumab injections (1/2841 injections, 0.035%). In the aflibercept group, 14 cases of endophthalmitis were identified from a total of 40 356 injections (1/2882 injections, 0.035%). The OR of developing endophthalmitis following bevacizumab compared to ranibizumab was 1.11 (95% confidence interval [CI], 0.81–1.52; P = .522). The OR comparing bevacizumab to aflibercept was 1.12 (95% CI, 0.63–2.01; P = .693). Similarly, the OR comparing ranibizumab with aflibercept was 1.01 (95% CI, 0.58–1.77; P = .960).
| Baseline Characteristics | Bevacizumab (n = 60) | Ranibizumab (n = 109) | Aflibercept (n = 14) |
|---|---|---|---|
| Age (y) | |||
| Mean (SD) | 75 (12.5) | 80 (9.2) | 82 (6.1) |
| Number of prior injections | |||
| Mean (SD) | 10.1 (9.6) | 13.9 (11.1) | 13.3 (10.8) |
| Mean VA at causative injection | 20/110 | 20/91 | 20/107 |
| Number of injections by diagnosis (%) | |||
| Neovascular AMD | 108 707 (70.7%) | 274 209 (88.5%) | 33 217 (82.3%) |
| Diabetic eye disease | 24 702 (16.1%) | 16 234 (5.3%) | 46 (0.1%) |
| Retinal vein occlusion | 20 403 (13.2%) | 19 279 (6.2%) | 7093 (17.6%) |
Table 2 summarizes the microorganisms isolated from the culture-positive cases for each anti-VEGF agent. Coagulase-negative Staphylococcus was the most commonly isolated organism for both bevacizumab (69.6%) and ranibizumab groups (43.9%). The second most common causative organism was Streptococcus species, which represented 21.7% of the organisms in the bevacizumab group and 22.0% in the ranibizumab group. For the aflibercept group, the culture-positive cases were due to either coagulase-negative Staphylococcus (50%) or Streptococcus species (50%). Table 3 reports visual outcomes according to culture results. Overall, visual outcomes were better in culture-negative cases than in culture-positive cases at 3 months follow-up for the 3 anti-VEGF agents. Furthermore, culture-positive cases due to coagulase-negative Staphylococcus had better visual outcomes at 3 months than those related to Streptococcus species for all groups.
| Bevacizumab | Ranibizumab | Aflibercept | |
|---|---|---|---|
| Positive culture, n (%) | 23 (38%) | 41 (38%) | 6 (43%) |
| Staph. epidermidis | 10 | 3 | 1 |
| Coagulase-negative Staph . | 4 | 12 | 1 |
| Strep. pneumonia | 3 | 2 | 1 |
| Strep. mitis | 0 | 4 | 1 |
| Staph. aureus | 0 | 5 | 0 |
| Staph. lugdunesis | 2 | 2 | 0 |
| Strep. viridans | 1 | 2 | 1 |
| Enterococcus fecalis | 1 | 4 | 0 |
| Staph. auricularis | 0 | 0 | 1 |
| Staph. homininis | 0 | 1 | 0 |
| Strep. sanguis | 0 | 1 | 0 |
| Strep. salivarius | 1 | 0 | 0 |
| Candida parapsicolosis | 0 | 1 | 0 |
| Lactobacillus | 1 | 0 | 0 |
| Nondifferentiated gram-positive cocci | 0 | 1 | 0 |
| Propionibacterium | 0 | 1 | 0 |
| Haemophilus influenzae | 0 | 2 | 0 |
| Bevacizumab | Ranibizumab | Aflibercept | |
|---|---|---|---|
| Culture-positive cases | |||
| Mean VA (causative injection) | 20/96 | 20/73 | 20/103 |
| Mean VA (3 months) | 20/455 | 20/968 | 20/1222 |
| P value a | .035 | <.001 | .161 |
| Culture-negative cases | |||
| Mean VA (causative injection) | 20/120 | 20/109 | 20/120 |
| Mean VA (3 months) | 20/224 | 20/400 | 20/203 |
| P value a | .042 | <.001 | .119 |
| Coagulase-negative Staphylococcus cases | |||
| Mean VA (causative injection) | 20/90 | 20/69 | 20/317 |
| Mean VA (3 months) | 20/257 | 20/446 | 20/341 |
| P value a | .158 | .007 | .423 |
| Streptococcus cases | |||
| Mean VA (causative injection) | 20/100 | 20/118 | 20/33 |
| Mean VA (3 months) | CF | HM | CF |
| P value a | .149 | .001 | .174 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
