Vascular endothelial growth factor (VEGF)-A is a key mediator of angiogenesis. The discovery of VEGF-A, like many other major discoveries in medicine, happened partly by observations and partly by chance. Napoleone Ferrara and his team were working on a population of non-hormone-secreting cells from the anterior pituitary of cows. (1) One day Ferrara mixed some isolates from cultures of follicular cells with endothelial cells. Ferrara was surprised to see the rapidly proliferating endothelial cells started rapidly. Ferrara understood that the pituitary cells were secreting an angiogenic protein. With this observation, Ferrara worked on the isolation and cloning of this angiogenic protein. In 1989, Napoleone Ferrara and Hazel at Genentech laboratory were the first to isolate and clone vascular endothelial growth factor (VEGF). Ferrara recalls, “I worked on the isolation of VEGF in my spare time during my first six months to a year at Genentech. Once we cloned VEGF in 1989, the company became interested and this became more and more my full-time pursuit.” In 1993, Ferrara reported that inhibition of VEGF-induced angiogenesis by specific monoclonal antibodies led to dramatic suppression of the growth of a variety of tumors in vivo. Thus bevacizumab (Avastin), a recombinant humanized anti-VEGF monoclonal antibody was developed in Ferrara’s laboratory and Genentech, the company with which he used to work licensed ranibizumab for ophthalmic use (2).

BEVACIZUMAB:

Bevacizumab consists of a full-length monoclonal antibody. It was approved for the treatment of metastatic colorectal cancer in 2004.

STRUCTURE OF BEVACIZUMAB:

Bevacizumab consists of a full-length monoclonal IgG1-antibody that binds to all isoforms of VEGF-A. Bevacizumab is produced in Chinese hamster ovary (CHO) cells, which is the cellular production pathway, which includes glycosylation of the proteins. The resulting bevacizumab anti-body contains 93% of the human amino acid sequence. Bevacizumab contains the Fab and Fc portions of a regular antibody, with a molecular weight of 148 kDa and binds to the identical epitope on VEGF molecules

Acute Intraocular Inflammation / Sterile Endophthalmitis with intravitreal injection of bevacizumab:

Sterile endophthalmitis characterized by significant anterior chamber and/or vitreous inflammation following intravitreal injection of bevacizumab, in the absence of infection. Sterile endophthalmitis is similar to the toxic anterior segment syndrome (TASS) that can occur after cataract surgery. Slit-lamp examination reveals marked anterior chamber reaction, mostly without hypopyon. Vitritis is usually mild to moderate and characterized by large cellular aggregates. This kind of post-injection inflammation differs from infectious endophthalmitis in following features: they tend to have more acute presentation (typically within 24–48 h) with minimal pain, more severe anterior chamber inflammation. Usually they are associated with more favorable visual outcomes.

The exact cause of inflammation induced by intravitreal bevacizumab remains largely unclear. An immunogenic mechanism has been attributed owing to the larger protein load and size of bevacizumab molecule which make it more immunogenic than ranibizumab. (6) Bevacizumab is derived from the cultures of mammalian ovarian cancer cells. This cellular production pathway yields glycosylated proteins with a higher immunogenic potential than the nonglycosylated pure proteins produced by a bacterial pathway (E. coli) in ranibizumab.

Such a massive immunological response following intravitreal injection of bevacizumab is a possible event, described in the literature and may occur even after preceding uneventful interventions. It is similar to the toxic anterior segment syndrome (TASS) that can occur after cataract surgery

Inflammation-induced by bevacizumab is usually an early onset associated with a painless drop in VA mostly without conjunctival injection. Patients present with a marked anterior chamber reaction, mostly without hypopyon. Vitritis is characterized by large cellular aggregates of pseudogranulomatous aspect.

Inflammation-induced by intravitreal bevacizumab differs from infectious endophthalmitis in the following features: (4,5)

• More acute presentation (typically within 24–48 h),
• More severe anterior chamber inflammation,
• Minimal pain
• More favorable visual outcomes

The cause of inflammation induced by intravitreal bevacizumab remains unclear. However various possible mechanism has been suggested. It is believed to occur due to an immunogenic mechanism in which foreign epitopes are presented by antigen-presenting cells to B and T cells. (6) Bevacizumab can diffuse into the anterior chamber more readily, and remain in the anterior chamber and vitreous for significantly longer than ranibizumab (7). Also larger protein load and size of bevacizumab molecule with an additional Fc fragment, make it more immunogenic than ranibizumab. Bevacizumab is harvested from cultures of mammalian ovarian cancer cells, that is the cellular production pathway, which includes glycosylation of the proteins with a higher immunogenic potential than the nonglycosylated pure proteins produced by a bacterial pathway (E. coli) in ranibizumab. It has been estimated that bevacizumab preparations contain about 70% of nonhumanised material (5) and this excess particulate matter, which is still within normal limits for intravenous use, maybe one of the causes for such inflammation. Further, bevacizumab solution has a pH of 6.2, well below the normal vitreous pH of 6.5–8.5.(8) Also, commercially bevacizumab is manufactured for intravenous infusion and therefore possibly undergoes a relatively less tight purification procedure than drugs formulated for intraocular use.

Endophthalmitis with intravitreal injection of bevacizumab:

An exhaustive Medline search of studies citing endophthalmitis (between 2011-15) after usage of Bevacizumab (Avastin, Genentech, South San Francisco, California, USA) revealed a varying incidence. Smaller studies appeared to have a higher (small study effect) incidence compared to studies having a larger number. Similarly, prospective studies had a different incidence as compared to one retrospective in nature. Although there were many studies looking into endophthalmitis after use of anti-VEGF, the studies specifically focusing on Avastin were less. Overall endophthalmitis rate post Avastin usage varies between 0% - 0.54% (9,10)

Avastin had received a negative media hype in West a few years back and is again under scrutiny following some untoward incidences in India. As cited “Avastin doesn’t blind people, people blind people” (11). The use of an “off label” drug has been prompted by the severe prohibitory cost of its sibling Ranibizumab (Lucentis, Genentech). In an Indian scenario where the cost is largely borne by the patient, having a cheaper drug which works just as fine as its costly sibling in terms of efficacy and side effects seems like a legitimate option.

Avastin is generally available as a 4ml or a 16 ml single-use glass vial. Ophthalmological practices involve either a single puncture and dividing into multiple aliquots (to be used in single or divided days), or involves multiple puncture for individual cases. The procedure of aliquoting involves stringent measures and United States Pharmacopeia (USP) ‘chapter 797’ details the correct way of doing it. A gist of the procedure can be found in the reference titled “Avastin doesn’t blind people” (11). In India where the concept of compounding pharmacy doesn’t exist in all places, the drug is hence divided in house. The patients are either pooled in, whereby a single puncture is made into the rubber stopper and the drug divided into multiple tuberculin syringes to be used on the same day or the glass vial and the rubber stopper is punctured at multiple occasions as and when the need arises. The former is more common in bigger institutes and private chains where a continuous flow of patients exists. The latter is commoner in practices involving smaller numbers.

The role of antibiotics in the peri-injection period has been a topic of constant debate. The points in favor of not using any prophylaxis are many, the most important one being induction of drug resistance. The use of anti-VEGF in a patient is generally not an onetime affair, hence increased chances of resistance in subsequent uses. The use of antibiotics as a pre or post-injection prophylaxis on a regular basis allows the selection of a more virulent strain altering the normal ocular flora (12,13). Repeated use of a particular antibiotic not only induces resistance to the particular agent but also against others. (14). The bactericidal effect if used immediately in the pre-injection period is also of questionable value (12). Studies involving a large number of subjects have found no difference in endophthalmitis rates when comparing antibiotics and non-use of the same (12,15, 16, 17). A recent study by the CATT study group (18) states no particular advantage of antibiotics in reducing the endophthalmitis rate. On the contrary use of antibiotics may increase the endophthalmitis rate (15). Although many surgeons still use antibiotics, the current trend is not using one.

Since the procedure of intravitreal injection involves puncturing of the ocular coats, an ideal scenario would be the standard operating room. Injections in the West and few parts of Europe employ the procedure room or the office setting for giving an intravitreal injection. This consists of cleanroom in the outpatient area which, although not up to operating room standards serves the purpose. In fact, most of the current studies employ office scenario to inject and studies specifically looking into this aspect have not found any increase in endophthalmitis incidence when given as an office procedure (19)

Although the process of an intravitreal injection involves multiple steps, the only steps which are almost universally agreed upon are the usage of topical povidone-iodine and sterile eye speculum (bladed or wired). The use of surgical gloves, sterile caps, surgical masks, sterile drapes, displacement of the conjunctiva is not universally agreed upon. Bactericidal effect of iodine varies between 0.005% to 10% with its effect seen within sixty seconds (20). The effect of a drop of 5% povidone-iodine is supposed to last for 24 hrs and a single application has a bactericidal effect equivalent to antibiotic use for 3 days (21). A recently cited paper brings up an interesting point of the pathology for which an antiVEGF is being given, playing a part in the etiology of endophthalmitis (22). According to their paper, patients receiving injections for retinal vein occlusions had lowest chances of endophthalmitis. Patients receiving injections for diabetic macular edema and AMD had more chances of infection. A poor immune status along with altered conjunctival flora were cited as causes for increased predisposition in diabetics whereas age was a factor in patients with AMD.

Although bacterial infections are much more common post anti-VEGF, fungal infections can also occur (23). The most common organism causing post-injection endophthalmitis is the coagulase-negative staphylococci (24) followed by viridans streptococci. Compared to normal intraocular surgery, the chances of post streptococcal infections are three times higher in an intravitreal setting (25). The role of viridans streptococci in the context of intravitreal scenario is mainly due its commensalism with the human nasopharyngeal tract and its possible role as a contaminant through aerosol route. Study by authors (26,27) have shown the increased chances of infection by the act of talking and not wearing a mask. Although many studies don’t mention the use or non-use of face masks, few that report of not using one don’t have a higher incidence of infection (12,28). The absolute diagnosis of the causative organism is to get a positive culture. Half of the cases leading to endophthalmitis are culture-negative (29). The usual culture positivity varies between 57% to 66% (30).

The visual outcome of patients with post-injection endophthalmitis depends on many factors such as the causative organism, how soon the patient presents, his general overall condition. Before labeling one as endophthalmitis it is important to rule out ‘sterile inflammation’. Although reported mainly with Bevacizumab; sterile reactions can also occur with ranibizumab (30). It is important to remember that Avastin has been meant for intravenous use. We use it intraocularly which has a smaller volume and confined space compared to the overall body when given intravenously, hence effects seen intraocularly may not be seen when administered intravenously. Endotoxins which are normally occurring byproducts of commercially made immunoglobulins are supposed to play a major role in these sterile reactions. The role of immune reaction to the antibody (since seen more commonly on sequential injections) have also been implied. Furthermore, the role of silicone oil contaminants and temperature fluctuations which supposedly increases the immunogenicity may also have a role to play ( 31, 32).

The literature provides us many different variations of the clinical presentation of post-injection endophthalmitis. Severe pain, lid edema, signs of severe intraocular inflammation generally denotes an infective etiology although it is not a rule. A sterile reaction can present almost in a similar way. Generally sterile inflammation presents early with less signs of pain, hypopyon, does well with topical steroids and visual recovery is faster. These features cannot be generalized as hypopyon with pain can also be a presenting feature (32). The best thumb rule is to have a high index of suspicion and treat any case presenting with intraocular inflammation after injection as infective until proven otherwise. The initial management involves as any other case of endophthalmitis which needs immediate principle of tap and inject. Aqueous or a vitreous tap generally yields material for staining and culture. A negative culture does not rule out an infection. Visual recovery is generally poor in culture-positive cases especially with streptococci (25, 33). Since results of culture may be delayed, its ideal to start treatment depending on the clinical and microbiological results. Polymerase Chain Reaction (PCR) when available is a useful guiding tool but like any other diagnostic modality it not a hundred percent sensitive in ruling out an infective agent. It is ideal to wait for some time to see the clinical progression and the effect of topical medications before shifting to surgery. Early surgical intervention in terms of vitrectomy may not be helpful in all cases. Worsening of visual acuity, pain, and intraocular inflammation can be used as guiding principles to decide on surgery (34).

Endophthalmitis is a dreaded post-injection complication which no surgeon wants to have. In spite of doctors having the best intention for the patient, sometimes such unfortunate accidents do happen. It is best to be vigilant when such a scenario arises and try to find the cause for the accident rather than shying away. Informing the right authorities and also keeping the patient abreast helps regain patient confidence and trust.

Checklist before Avastin

Pre Procedure

1) Check the systemic status: glycemic index, blood pressure, any drug allergy, cardiac function, h/o anticoagulants

2) Check locally: rule out any ocular infections like blepharitis, external or internal hordeolum, conjunctivitis, dacryocystitis, keratitis, scleritis

3) Perform a detailed retinal examination to rule out any untreated retinal holes, tears. Inform patients about the risk of inducing a PVD in high myopes. Presence of lattice is not a contraindication, however, if same/other eye has a history of retinal detachment, it’s better to do laser to concerned lesion before antiVEGF

4) Examine the optic nerve head to rule out any glaucoma

3) Avoid injecting without an attendant

When preparing Bevacizumab

1) The person withdrawing it should wear sterile gloves, face mask, and ideally should be done in a sterile environment.

2) Best to puncture once and divide into multiple doses which should be used up on the same day. Pooling patients offers the most secure method

3) Once a vial is opened the drug should ideally be used up in 1-3 weeks*. Remember to clean the rubber stopper with isopropyl alcohol and betadine before withdrawing again.

4) If the drug has been divided and kept, remember to send a few samples for microbiological analysis before you inject it.

5) Microbiological analysis can be done before each injection when vial kept for a longer time

6) Always inspect the bottle before withdrawing.

7) Maintaining cold chain is MUST ( 2-8 degrees). DO NOT FREEZE

Procedural

1) Place of injection: Office/OT (surgeon preference)

2) Pre injection antibiotics: surgeon preference

3) Use of povidone-iodine is MUST

4) The topical anesthetic in the form of drops or jelly

5) A sterile speculum is MUST

6) Use of face mask, sterile gloves, surgical cap, the patient drape is surgeon and institution preference. In the Indian scenario, it’s better to have all of the above.

7) Post injection antibiotic: surgeon preference

8) Eye patch: not must

9) Head bath: withhold for 24 hrs

10) Review within a week

Summary of studies citing post bevacizumab endophthalmitis between 2011- 2015

NM:Not Mentioned,Variation:varied between surgeons, no fixed protocolAB:Antibiotic

Reference

1. Ribatti, D. Napoleone Ferrara and the saga of vascular endothelial growth factor. Endothel. J. Endothel. Cell Res. 15, 1–8 (2008).
2. Ferrara, N. From the discovery of vascular endothelial growth factor to the introduction of avastin in clinical trials - an interview with Napoleone Ferrara by Domenico Ribatti. Int. J. Dev. Biol. 55, 383–388 (2011).
3. Gillies, M. C. What we don’t know about avastin might hurt us. Arch. Ophthalmol. Chic. Ill 1960 124, 1478–1479 (2006).
4. Fielden, M., Nelson, B. & Kherani, A. Acute intraocular inflammation following intravitreal injection of bevacizumab--a large cluster of cases. Acta Ophthalmol. (Copenh.) 89, e664–665 (2011).
5. Georgopoulos, M., Polak, K., Prager, F., Prünte, C. & Schmidt-Erfurth, U. Characteristics of severe intraocular inflammation following intravitreal injection of bevacizumab (Avastin). Br. J. Ophthalmol. 93, 457–462 (2009).
6. Hermeling, S., Crommelin, D. J. A., Schellekens, H. & Jiskoot, W. Structure-immunogenicity relationships of therapeutic proteins. Pharm. Res. 21, 897–903 (2004).
7. , M. W. Pharmacokinetics, pharmacodynamics and pre-clinical characteristics of ophthalmic drugs that bind VEGF. Expert Rev. Clin. Pharmacol. 7, 167–180 (2014).
8. Mamalis, N. et al. Toxic anterior segment syndrome. J. Cataract Refract. Surg. 32, 324–333 (2006).
9. Casparis H, Wolfensberger T J, Becker M, Eich G, Graf N et al. Incidence of Presumed Endophthalmitis after intravitreal injection performed in the operating Room: A Retrospective Multicenter Study. Retina. 2014 Jan; 34(1):12–17.
10. Falavarjani K G, Modarres M, Hashemi M, Parvaresh M, Naseripour M. Incidence of Acute Endophthalmitis After Intravitreal Bevacizumab injection in a single clinical center. Retina. 2013 May; 33(5):971–974
11. Gonzalez S, Rosenfeld PJ, Stewart MW, Brown J, Murphy SP. Avastin doesn't blind people, people blind people. Am J Ophthalmol. 2012 Feb;153(2):196-203
12. Bhavsar AR, Sandler DR. Eliminating antibiotic prophylaxis for intravitreal injections: a consecutive series of 18,839 injections by a single surgeon. Retina. 2015 Apr;35(4):783-8
13. Yin VT, Weisbrod DJ, Eng KT, et al. Antibiotic resistance ofocular surface flora with repeated use of a topical antibioticafter intravitreal injection. JAMA Ophthalmol 2013;131456–461.
14. Kim SJ, Toma HS. Antimicrobial resistance and ophthalmic antibiotics: 1-year results of a longitudinal controlled study of patients undergoing intravitreal injections. Arch Ophthalmol. 2011 Sep;129(9):1180-8
15. Cheung CS, Wong AW, Lui A, Kertes PJ, Devenyi RG et al. Incidence of endophthalmitis and use of antibiotic prophylaxis after intravitreal injections. Ophthalmology. 2012 Aug;119(8):1609-14
16. Park Y, Kim KS, Park YH. Acute endophthalmitis after intravitreal injection and preventive effect of preoperative topical antibiotics. J Ocul Pharmacol Ther. 2013 Dec;29(10):900-5
17. Bhatt SS, Stepien KE, Joshi K. Prophylactic antibiotic use after intravitreal injection: effect on endophthalmitis rate. Retina. 2011 Nov;31(10):2032-6
18. Meredith TA, McCannel CA, Barr C, Doft BH, Peskin E et al; Comparison of Age-Related Macular Degeneration Treatments Trials Research Group. Postinjection endophthalmitis in the comparison of age-related macular degeneration treatments trials (CATT). Ophthalmology. 2015 Apr;122(4):817-21
19. Abell RG, Kerr NM, Allen P, Vote BJ. Intravitreal injections: is there benefit for a theatre setting? Br J Ophthalmol. 2012 Dec;96(12):1474-8
20. Shimada H, Hattori T, Mori R, Nakashizuka H, Fujita K et al. Minimizing the endophthalmitis rate following intravitreal injections using 0.25% povidone-iodine irrigation and surgical mask. Graefes Arch Clin Exp Ophthalmol. 2013 Aug;251(8):1885-90
21. Isenberg SJ, Apt L, Yoshimori R, Khwarg S. Chemical preparationof the eye in ophthalmic surgery. IV. Comparison ofpovidone-iodine on the conjunctiva with a prophylactic antibiotic.Arch Ophthalmol 1985;103:1340–1342
22. Rayess N, Rahimy E, Shah CP, Wolfe JD, Chen E. Incidence and clinical features of post-injection endophthalmitis according to diagnosis. Br J Ophthalmol. 2015 Nov 19. pii: bjophthalmol-2015-307707.(epub ahead of print)
23. Sheyman AT, Cohen BZ, Friedman AH, Ackert JM. An outbreak of fungal endophthalmitis after intravitreal injection of compounded combined bevacizumab and triamcinolone. JAMA Ophthalmol. 2013 Jul;131(7):864-9
24. McCannel CA. Meta-analysis of endophthalmitis after intravitreal injection of anti-vascular endothelial growth factor agents: causative organisms and possible prevention strategies. Retina. 2011 Apr;31(4):654-61
25. Chen E, Lin MY, Cox J, Brown DM. Endophthalmitis after intravitreal injection: the importance of viridans streptococci. Retina. 2011 Sep;31(8):1525-33
26. O’Kelly SW, Marsh D. Face masks and spinal anaesthesia. Br J Anaesth 1993;70:239
27. Philips BJ, Fergusson S, Armstrong P, Anderson FM, WildsmithJA. Surgical face masks are effective in reducingbacterial contamination caused by dispersal from the upperairway. Br J Anaesth 1992;69:407–408.
28. Tabandeh H, Boscia F, Sborgia A, Ciracì L, Dayani P. Endophthalmitis associated with intravitreal injections: office-based setting and operating room setting. Retina. 2014 Jan;34(1):18-23
29. Goldberg RA, Flynn HW Jr, Isom RF, Miller D, Gonzalez S. An outbreak of streptococcus endophthalmitis after intravitreal injection of bevacizumab. Am J Ophthalmol. 2012 Feb;153(2):204-208.e1
30. Cunningham MA, Tlucek P, Folk JC, Boldt HC, Russell SR. Sequential, acute noninfectious uveitis associated with separate intravitreal injections of bevacizumab and ranibizumab. Retin Cases Brief Rep. 2013 Fall;7(4):355-8
31. Marticorena J, Romano V, Gómez-Ulla F. Sterile endophthalmitis after intravitreal injections. Mediators Inflamm. 2012;2012:928123.
32. Orozco-Hernández A, Ortega-Larrocea X, Sánchez-Bermúdez G, García-Aguirre G, Cantón VM, Velez-Montoya R. Acute sterile endophthalmitis following intravitreal bevacizumab: case series. Clin Ophthalmol. 2014 Sep 9;8:1793-9.
33. Matthews JL, Dubovy SR, Goldberg RA, Flynn HW Jr. Histopathology of streptococcus mitis/oralis endophthalmitis after intravitreal injection with bevacizumab: a report of 7 patients. Ophthalmology. 2014 Mar;121(3):702-8
34. Chaudhary KM, Romero JM, Ezon I, Fastenberg DM, Deramo VA. Pars plana vitrectomy in the management of patients diagnosed with endophthalmitis following intravitreal anti-vascular endothelial growth factor injection. Retina. 2013 Jul-Aug;33(7):1407-16
35. Chen YH, Wu PC, Shiea J, Lo LH, Wu YC, Kuo HK. Evaluation of the sterility, stability, and efficacy of bevacizumab stored in multiple-dose vials for 6 months. J Ocul Pharmacol Ther. 2009 Feb;25(1):65-9
36. Mithal K, Mathai A, Pathengay A, Jalali S, Relhan N et al. Endophthalmitis following intravitreal anti-VEGF injections in ambulatory surgical centre facility: incidence, management and outcome. Br J Ophthalmol. 2013 Dec;97(12):1609-12
37. VanderBeek BL, Bonaffini SG, Ma L. Association of compounded bevacizumab with postinjection endophthalmitis. JAMA Ophthalmol. 2015 Oct;133(10):1159-64
38. Shah CP, Garg SJ, Vander JF, Brown GC, Kaiser RS, Haller JA; Post-Injection Endophthalmitis (PIE) Study Team. Outcomes and risk factors associated with endophthalmitis after intravitreal injection of anti-vascular endothelial growth factor agents. Ophthalmology. 2011 Oct;118(10):2028-34
39. Falavarjani KG, Aghamirsalim M, Modarres M, Hadavandkhani A, Hashemi M et al. Endophthalmitis after resident-performed intravitreal bevacizumab injection. Can J Ophthalmol. 2015 Feb;50(1):33-6
40. Terzi? S1, Pilavdi? A, Nadarevi? Voden?arevi? A. Incidence of endophthalmitis after intravitreal application of anti VEGF therapy at the University Clinical Center in Tuzla, Bosnia and Herzegovina. Med Glas (Zenica). 2015 Aug;12(2):202-5
41. Inman ZD, Anderson NG. Incidence of endophthalmitis after intravitreal injection of antivascular endothelial growth factor medications using topical lidocaine gel anesthesia. Retina. 2011 Apr;31(4):669-72
42. Cheema RA, Alshihry AM, Cheema HR. Incidence of endophthalmitis following intravitreal Bevacizumab injection at a tertiary care hospital in Eastern Province of Saudi Arabia. Saudi J Ophthalmol. 2015 Apr-Jun;29(2):135-6
43. Englander M, Chen TC, Paschalis EI, Miller JW, Kim IK. Intravitreal injections at the Massachusetts Eye and Ear Infirmary: analysis of treatment indications and postinjection endophthalmitis rates. Br J Ophthalmol. 2013 Apr;97(4):460-5
44. Moshfeghi AA, Rosenfeld PJ, Flynn HW Jr, Schwartz SG, Davis JL et al. Endophthalmitis after intravitreal anti-vascular endothelial growth factor antagonists: a six-year experience at a university referral center. Retina. 2011 Apr;31(4):662-8
45. Storey P, Dollin M, Rayess N, Pitcher J, Reddy S et al. The effect of prophylactic topical antibiotics on bacterial resistance patterns in endophthalmitis following intravitreal injection. Graefes Arch Clin Exp Ophthalmol. 2016 Feb;254(2):235-42