Endophthalmitis Postintravitreal Bevacizumab Injections: Incidence, Management, and Visual Outcome in Tertiary Care Center in Pakistan.

PURPOSE: The purpose is to study the incidence, characteristics, management, and visual outcome of endophthalmitis postintravitreal bevacizumab (Avastin®) injections.
METHODS: Retrospective cohort study conducted at Fauji Foundation Hospital that is a tertiary care center in Rawalpindi, Pakistan. The study duration was from January 1, 2015 to March 31, 2020. A total of 2321 eyes of 1407 patients were treated with Intravitreal bevacizumab (IVB) in a minor operation theater using standard aseptic measures. We studied the incidence, clinical characteristics, treatment, and visual outcomes of endophthalmitis post-IVB.
RESULTS: Seven eyes (0.30%) out of 2321 eyes developed post-IVB endophthalmitis during the study period. Three eyes (42.8%) were culture positive and four eyes (57.2%) were culture negative. Five eyes (71.4%) showed visual improvement and two eyes (28.6%) did not show visual improvement.
CONCLUSION: Prompt intravitreal antibiotic along with vitreous tap in cases of postintravitreal endophthalmitis may prevent vision loss which provides a useful alternative to immediate pars plana vitrectomy that requires more expertise, equipment, and an operation theater. Copyright:

Introduction

Intravitreal injections (IVIs) have become part of the standard regimen for inhibiting vascular endothelial growth factor (VEGF) and thereby, managing and improving the visual prognosis in cases of diabetic macular edema, retinal vein occlusion, age-related macular degeneration, etc.[12] Whereas IVIs are largely safe to use, they do run a slight risk of unwanted side effects including sterile inflammation, intraocular pressure elevation, endophthalmitis, retinal detachment, and vitreous hemorrhage.[3]

The incidence of postintravitreal endophthalmitis has been reported to be between nil and 1.6%.[4] In this regard, it is important to keep in mind that a given individual may receive monthly IVIs for a period that may be as long as several years. It is, therefore, pertinent to focus attention on post-IVI endophthalmitis cases despite the low incidence.[5] The incidence of endophthalmitis varies widely with protocol used.[6] It has been recommended in the literature to use povidone-iodine, eye speculum, and face masks for reduced contamination. Further, proper sterilization protocols should be followed during administration of IVIs preferably in an operation theatre setting.[7] Employing proper sterilization can keep the incidence of endophthalmitis in check, for instance, a study in France, where using a face mask while administering IVIs is the norm, reported endophthalmitis occurring in as less as 0.021% of cases.[8]

Moreover, whereas the risk of endophthalmitis may be low, the consequences of post-IVI endophthalmitis may be severe visual loss, in particular if treatment is delayed or if the offending pathogen is markedly virulent. Furthermore, there is no clear protocol for dealing with endophthalmitis following intravitreal anti-VEGF administration. Some practitioners follow the guidelines of the endophthalmitis vitrectomy study (EVS) which was done on acute endophthalmitis postcataract extraction or secondary lens implantation that recommends performing vitrectomy only in cases with light perception or worse vision.[9] This, however, may not be the most suitable choice as the etiology and progression of postcataract endophthalmitis may be different from that of post-IVI endophthalmitis. Most recently, focus has shifted toward prevention strategies and early diagnosis of endophthalmitis and its treatment through tap and inject technique or vitrectomy.[10]

The purpose of our study was to study the incidence, characteristics, management, and visual outcome of endophthalmitis postintravitreal bevacizumab (Avastin®) injections. In this regard, we have performed culture sensitivity analysis as well. We utilized the immediate tap and inject approach in all cases regardless of the visual acuity (VA) of the patient. In case no improvement in terms of improvement in VA, reduced pain, reduced cell/flare, or better fundus view was not seen after 48 h, a vitrectomy was performed.

Methods

This retrospective cohort study is based on patients treated at the retina clinic of Fauji Foundation Hospital, a tertiary care facility in Rawalpindi (Pakistan), from January 1, 2015 to March 31, 2020. The study was conducted after obtaining the required approval from the Ethical Review Committee of the Fauji Foundation Hospital. The electronic medical records of the patients were used for required data collection. We focused on the diagnoses, VA before IVI, symptoms and signs of patients documented as cases of post-IVI endophthalmitis, treatment given to patients such as intravitreal vancomycin and ceftazidime and/or pars plana vitrectomy (PPV), culture results, and the outcome of patients as observed and documented in the follow-up.

We included all patients that were administered IVB injections after obtaining written informed consent. Topical povidone iodine (5%) was used both before and after IVB administration. IVB was administered by postgraduate trainees and ophthalmologists in a minor operation theater setting while ensuring sterilization through using disposable masks, caps, sterilized gowns, and gloves. In addition, sterilized shield speculum, caliper, and towel were used. IVB was injected using a 30-gauge needle in the inferotemporal quadrant, 3.5 (4) mm posterior to the limbus for pseudophakic (phakic) patients. After administration, the patients were given topical antibiotic.

Patients who were administered IVB at some other facility were excluded. The diagnosis of endophthalmitis was done through a clinical examination followed by a B-scan as per the standard protocol. Upon confirmation, vitreous tap was performed aseptically and immediately sent for culture and sensitivity. The infected eyes were then given intravitreal vancomycin 1 mg/0.1 ml and ceftazidime 2.25 mg/ml as well as subconjunctival vancomycin 50 mg/ml and ceftazidime 125 mg/ml. In addition to this, oral ciprofloxacin 500 mg BD and diclofenac sodium 50 mg BD were started as well as topical moxifloxacin, prednisolone, and atropine drops. The cases were reviewed after 48 h to evaluate improvement in terms of VA, reduced pain, reduced cell/flare, or better fundus view. In cases lacking improvement, PPV was carried out along with administration of intravitreal vancomycin and ceftazidime.

Results

A total of 2321 eyes of 1407 patients were given IVB injections over a duration of 5 years. Most of the patients treated were females (1326/1407 = 94.2%) while the remaining 81 were males (5.7%). The mean age of the patients was 61.4 years ± 8.1 years. The most common diagnosis (64.2%) was that of nonproliferative diabetic retinopathy (PDR) with clinically significant macular edema (CSME). Around 10.3% of patients suffered from central retinal vein occlusion (CRVO), 8.4% from PDR with CSME, 7% from branch retinal vein occlusion (BRVO). Age-related macular degeneration (ARMD) affected 4.6% of patients while vitreous hemorrhage and neovascular glaucoma (NVG) were present in 4.5% and 1%, respectively. Details of diagnoses are given in Figure 1.

Figure 1

Diagnosis of patients included in the study (n = 1407) NPDR: Nonproliferative diabetic retinopathy, CSME: Clinically significant macular edema, CRVO: Central retinal vein occlusion, PDR: Proliferative diabetic retinopathy, BRVO: Branch retinal vein occlusion, ARMD: Age-related macular degeneration, Vit hmg: Vitreous hemorrhage, NVG: Neovascular glaucoma

Out of 2321 eyes, seven eyes (0.30%) developed post-IVB endophthalmitis. Three eyes (42.8%) were culture positive and four eyes (57.2%) were culture negative. The details of our treatment protocol are shown as a flowchart in Figure 2.

Figure 2

Time frame for this retrospective cohort study begun on March 31, 2020, showing the treatment plan followed. IVB: Intravitreal bevacizumab, IVAB: intravitreal antibiotic, PPV: Pars plana vitrectomy

As shown in Table 1, all cases of endophthalmitis presented within 2 days of IVB administration and with a VA of CF or worse. Five out of these seven returned to their preinjection VA or better. Both cases in which VA did not improve were culture positive (one infected with Streptococcus pneumonia and the other infected with pseudomonas). It can also be seen in Table 1 that four out of seven patients were treated only with intravitreal antibiotic (IVAB) while the remaining three also underwent PPV.

Table 1

Clinical and microbiological characteristics of patients with post IVB endophthalmitis along with visual prognosis after treatment.

Diagnosis	Pre-Inj VA	Time of presentation	VA at presentation	Treatment	VA after treatment	Culture
Mild NPDR with CSME	6/12	01 day	CF	IVAB	6/12	Negative
Mild NPDR with CSME	6/9	02 days	CF	IVAB	6/9	Negative
Moderate NPDR with CSME	6/12	01 day	CF	IVAB	6/12	Negative
PDR with CSME	6/60	02 days	HM	IVAB	NPL	Strep pneumoniae
Severe NPDR with CSME	6/60	02 days	CF	IVAB + PPV	6/24	Staph auerus
Moderate NPDR with CSME	6/18	02 days	CF	IVAB + PPV	6/12	Negative
Severe NPDR with CSME	6/60	01 day	PL	IVAB + PPV	NPL	Pseudomonas

VA: Visual acuity, NPDR: Non proliferative diabetic retinopathy, CSME: Clinical significant macular edema, PDR: Proliferative diabetic retinopathy, IVB: Intravitreal bevacizumab, CF: Counting finger, IVAB: Intravitreal antibiotics, PPV: Pars plana vitrectomy

All the culture negative patients progressed to a VA better than 6/12 after 3 months, while of those having positive cultures, two worsened to NPL, with the third improving to 6/24 vision. There was a positive correlation between severity of diabetic retinopathy (DR) with bacterial cultures. Details of these results are summarized in Table 1.

Discussion

Infectious endophthalmitis after intravitreal anti-VEGF is a potentially devastating complication. With an increased use of anti-VEGF treatments, it is important to be aware of the associated risk of endophthalmitis that will inevitably accompany any such treatment, even though at a low rate. Likewise, due attention must be paid to various methods that may be used to reduce the risk of infectious endophthalmitis.[11] In our 4-year retrospective analysis, we report a postintravitreal endophthalmitis rate of 0.3%. A similar study conducted in Karachi, Pakistan, reported endophthalmitis rate after IVB to be 0.19%.[12]

In our study, patients presented with multiple diagnoses including DR, CRVO, BRVO, ARMD, NVG, and vitreous hemorrhage. However, it is to be noted that all seven patients in our series that developed endophthalmitis suffered from DR. Literature review suggests that elevated levels of glucose in diabetes mellitus seem to weaken the body's natural immunity by impairing the epidermal growth factor receptor signaling and also by repressing the process of basal cell multiplication. In addition, it hinders the function of polymorphonuclear cells and monocytes/macrophages. There is also delayed epithelial healing which leads to vision threatening complications such as infective keratitis and postoperative endophthalmitis due to hampered and poor wound closure.[1314] This correlation of endophthalmitis with diabetes is consistent with the results of Wani et al. in which five patients developed endophthalmitis out of total 5429 injections and four patients had DR while only one had ARMD.[15] Further, we also observe in our study that higher the grade of DR, the lesser is the visual improvement.

All seven patients underwent initial treatment with tap/inject technique. Four (57.1%) cases responded with tap/inject technique whereas three (42.8%) of the unresponsive cases required subsequent vitrectomy. There are no clear guidelines on how to treat infectious endophthalmitis following intravitreal drug administration. Therefore, the EVS guidelines become the route followed by many ophthalmologists.[9] The findings of the EVS have their own limitations with regard to infectious endophthalmitis following IVI. First, PPV has become an indispensable tool in the hands of a trained surgeon with its techniques matured, since the time of EVS, to the extent that many surgeons will be inclined to perform it as the primary option in cases of endophthalmitis after IVI. Second, patients that require IVI usually have poor preinjection VA, resulting for example from subretinal fibrosis in ARMD, compared to those patients undergoing cataract surgery. Chaudhary et al. have suggested that in cases of postinjection endophthalmitis, the outcomes of primary tap/inject are better than vitrectomy alone, as in their study about 90% of the patients improved to their preinjection VA with tap and inject technique but in comparison only 46% of the patients regained their preinjection VA with initial PPV.[16] Furthermore, no clear benefit of vitrectomy over antibiotic injection alone was found.[17]

About 42.8% were culture positive and 57.2% were culture negative. Out of culture positive, one patient had developed Staphylococcus aureus as the causative agent who improved to 6/24 vision from CF after undergoing PPV. The second patient had S. pneumonia as causative agent who developed phthisis bulbi. The third case had Pseudomonas as the causative agent, which unfortunately also ended up with phthisis bulbi. These results are consistent with the literature, for example, Jeong et al. have shown that endophthalmitis caused by Pseudomonas aeruginosa shows poor recovery rate (21.4%) whereas Staphylococcus epidermidis cases show good VA recovery at a rate of 69.7%.[18]

The finding of poor visual recovery in cases of streptococcal endophthalmitis is also demonstrated by Kurniawan et al. with about 77.6% of their cases showed a final VA 6/60 or worse and 24.7% were either enucleated or eviscerated.[19]

Out of culture-negative cases, these improved to 6/12 vision. These four culture-negative patients might be due to potent antibiotic intake before the presentation, low microbial count or it might have been due to sterile endophthalmitis. In our study, VA improvement was more in cases who were culture negative as compared to the culture-positive patients. This finding is consistent with Harry et al. in which they found that while treatment outcomes were variable, they were generally better in the culture negative cases.[20]

We chose to treat every patient of infectious endophthalmitis with immediate vitreous tap followed by an IVAB injection regardless of VA at presentation of the patient. This is contradictory to EVS as we did not take VA as standard to decide our management. We believe that the key to management is timely treatment whether it is done through IVABs or PPV. The earlier we start treatment, the better the outcome. This is in complete concordance with the findings of Negretti et al.[21] In EVS, PPV was done within 6 h while according to FRIENDS, PPV was performed after 2 days on average.[22] This window is rather not possible in our setup due to multiple hindrances. We do not require much preparation to administer IVIs so it can be handled immediately. Whereas, for PPV, we need vitreoretinal team, operation theater, anesthesia fitness (optional) so arranging a PPV can delay our treatment. Clearly, in our circumstances, immediate tap/inject is the preferred option and if no improvement is seen, then PPV is the route to follow after 24–48 h.

Our study has several limitations. First, these data rely upon retrospective review of electronic medical data to identify endophthalmitis cases. Second, the sample size was small. Finally, our study was not a randomized control trial to directly compare management outcomes of primary tap and inject technique and primary PPV.

Conclusion

Prompt IVAB along with vitreous tap in cases of postintravitreal endophthalmitis may prevent vision loss which provides a useful alternative to immediate PPV that requires more expertise, equipment, and an operation theater.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.