When is vitrectomy for endophthalmitis needed?

In December 1995, a prospective, randomized study (EVS, 1995) was published on the treatment for acute postoperative endophthalmitis (POE). This was a landmark article since it appeared definitely to answer the dilemma whether POE was an entity to be appreciated as a primarily surgical condition or (statistically selected) intravitreal antibiotics would suffice. The study led the vast majority of ophthalmologists to give up surgery for all but the worst POE eyes: It is always easier not to do something (vitrectomy) than to do something. It is certainly simpler to substitute an intraocular injection for an often difficult surgery, especially since the former can be performed by any ophthalmologist, while the latter requires a trained specialist with special equipment. On the contrary, there is certainly a need to rationalize medicine and use expensive and complicated procedures (vitrectomy) only when it is justified and brings benefits (Brody, 2012). The question how a ‘level1 evidence’ study such as the EVS could not prove the superiority of surgical pus removal over the simple killing of the infecting organism by antibiotics has since been answered: The EVS did not require that the vitrectomy be thorough (removing all vitreous, including the creation of a posterior vitreous detachment [PVD]). The death of the bacteria still leaves a toxic brew sloshing against the retina, causing more damage than the live bacteria would (Callegan et al., 1999), showing that it is not the bacteria themselves but their toxins and the body's immune reactions that destroy the intraocular tissues (Callegan et al., 2011). If proper vitrectomy is performed (creating a PVD and removing all, but especially the posterior vitreous) and is performed early (before major retinal damage occurs and when visibility is still good, thereby increasing the safety margin), the proportion of eyes with the excellent functional outcome almost doubles compared with the findings in the EVS. Complete and early vitrectomy for endophthalmitis (CEVE; Kuhn & Gini, 2006) has therefore been increasingly accepted as the preferred treatment for POE. The study by Blom et al. (2022) is another remarkable milestone on the road to optimal treatment of eyes with POE: It no longer compares surgery with antibiotics but eyes that underwent emergency surgery versus those that had to use antibiotics as a bridge to (typically still) early vitrectomy. The findings in both arms of the study showed encouraging functional and anatomical results, confirming the benefits of CEVE. It is time for all ophthalmologists to change their treatment paradigm for POE. Instead of ‘vitrectomy is only for the worst eyes’, the principle should be ‘primary vitrectomy unless circumstances prevent it’. The vitrectomy should be performed as soon as possible after admission and include the creation of a PVD; intravitreal antibiotics and corticosteroids typically represent but an adjuvant therapy. If the postoperative course requires it, the vitrectomy should be repeated (CEVE+; Dib et al., 2020).