Primum non nocere . From when we attended medical school we learned: “first, do no harm.” We first learn how to do an operation; we then learn when to do it; and finally, from experience, we learn when not to do it. Patients presenting for keratoprosthesis (KPro) surgery have typically gone through multiple operations and come with significant psychological burden. It is therefore of utmost importance that they be assessed thoroughly prior to being offered keratoprosthesis surgery.

A keratoprosthesis is an artificial cornea. While a temporary KPro is used for performing vitreoretinal surgery, permanent KPros are used for visual rehabilitation of corneal blindness not amenable to conventional corneal grafting. Currently, there are 2 mainstream devices for 2 different groups of patients. For “wet blinking eyes” with an adequate tear film, fully present and functioning lids, and no keratin on the ocular surface, a Boston Type 1 KPro could be considered. Typical indications include multiple previous failed corneal grafts, vascularized corneas, aniridia, and, more recently, Peter’s anomaly. For dry eyes and eyes with absent lids, the osteo-odonto-keratoprosthesis (OOKP), which uses the patient’s own canine tooth root and surrounding jaw bone to frame a plastic optical cylinder, is the procedure of choice.

Patient assessment should include assessment of the eye and of the person. The latter should include formal psychological assessment, checking the patient’s understanding of the gravity of the surgical program including financial implications, regular life-long follow-up, preparedness for further surgery, preparedness for device failure and permanent sight loss, acceptance of cosmetic appearance, the level of desire to see better (especially if the fellow eye is seeing), and the level of support from family and friends. In the case of the OOKP, there is also oral assessment. Keratoprostheses should only be offered in specialist centers that could deal with complications speedily and properly as and when they arise.

One of the most important factors to consider during preoperative assessment is the visual potential of the eye in question. Will the eye see much better following KPro surgery? In this issue Farias and associates demonstrate the use of preoperative intraocular endoscopic fundal examination prior to offering KPro surgery. How predictive and how safe is it? When should it be used, and when not? Endoscopic video evaluation is itself a procedure, so primum non nocere should also apply. It is a matter of weighing up potential benefits and risks.

More often than not, history and clinical examination already inform us of an eye’s visual potential. For example, we may already know of advanced glaucoma or macular degeneration from history taking. On the other hand, if a patient saw very well initially before failure of his or her last corneal graft, then the visual potential is likely to be good. During examination, we look for brisk projection of light in quadrants. Color sensation and the absence of a relative afferent pupil defect may help. For eyes with no light perception, it goes without saying that KPro surgery should not be offered. There will also be no justification for carrying out endoscopic video evaluation. Office investigations should include ultrasonography to identify retinal detachment. Flash electroretinogram and flash visual evoked potential help evaluate retinal and optic nerve function. The authors’ experience, however, is that good vision can still be had with poor electrodiagnostic results, especially when there is significant symblepharon or surface keratinization.

If after history taking, ocular examination, ultrasonography, and electrodiagnostic tests, there is still doubt with regard to visual potential, and the patient wants to find out more and is prepared to undergo video endoscopy, then this could be carried out. However, there is no guarantee that fundal appearance alone will provide firm guidance. Does the imaging technique provide sufficient resolution? Also, a perfect-looking fundus may belong to an amblyopic eye. On the other hand, a deeply cupped optic disc and significant maculopathy or atrophy may still be compatible with good central vision. In Farias and associates’ study, they proceeded with KPro surgery in just 3 out of 10 cases and obtained good visual results for the 3 eyes of 3 patients. We do not know how the other 7 would have done following KPro surgery. Farias and associates have proven a concept and safety of video endoscopy in eyes being considered for Boston KPro surgery. But to correlate fundal appearance with visual acuity, a better study could simply be a cohort of clinic patients without media opacity (corneal opacity and cataract) or a cohort of pseudophakic patients without corneal pathology.

How safe is the technique? In this small group of patients, the technique has proven to be safe, but there will be complications in a larger study population, including disturbance to lens, retinal detachment, vitreous incarceration, bleeding, and infection. It can be argued that those with a known retinal detachment should not have video endoscopy unless it is planned for endoscopic repair. It can be postulated that the risks will be higher in eyes being considered for the OOKP because of difficulty of identifying landmarks, dry eye, and more challenging wound closure. It will also be more difficult to inject a subconjunctival antibiotic, although a quinolone ointment can be used instead. That said, we are among proponents to precede OOKP surgery with a formal endoscopic vitrectomy, which could help with improving the success of glaucoma drainage devices and with reducing the risk of late rhegmatogenous retinal detachment. Any significant diabetic retinopathy can be treated with endolaser (panretinal, focal, or grid) or anti–vascular endothelial growth factor injections. Furthermore, ciliary body ablation could be carried out as necessary. Finally, depending on the type of anesthesia used, it may be possible to project a fine grid pattern onto the macula to test potential visual acuity, or to do the same for pattern electroretinogram.

In summary, preoperative intraocular endoscopic evaluation of the fundus is applicable for select patients with unclear visual potential. Its value may be enhanced if treatment could also be offered in addition to diagnostic use. Possible treatment includes endolaser for diabetic retinopathy and pars plana insertion of a glaucoma drainage device. Routine video endoscopic testing of every patient considered for KPro is not advocated.