In most of the cases published so far, PACK-CXL was performed according to the Dresden CXL protocol for KC, with the following modifications: First, the beam is either focused on the lesion or includes the lesion. In the case of a peripheral lesion, the surgeon might need to expose the limbus. Few publications reported on the effect of CXL on the limbus with clear consensus on the potential risk [57, 58]. Second, surgeons manually remove the epithelium surrounding the infiltrate over a few millimeters to allow complete penetration of the riboflavin around the lesion. Usually, this is combined with the corneal scraping for laboratory analysis. Third, a recent report enlightened the consequences of fluorescein staining immediately prior to PACK-CXL. Fluorescein is competing with riboflavin for UVA absorption thus reducing the overall antiseptic effect of the surgery [59] and should be avoided.

The microbicidal effect of PACK-CXL is obtained through several molecular mechanisms during the surgery and explains further the postoperative effects on the cornea that contribute to healing the infection (Table 9.1).

We assume that PACK-CXL may directly damage pathogens via the following pathways: on one hand, UV light may directly interact with microbial DNA and RNA. When penetrating a cell and its nucleus, UV light alone creates cross-links between nucleic acids as for example between two thymines generating a dimer [60]. This alters the genomic structure and compromises pathogen replication. UV light inactivation depends on a number of factors: type of pathogen (specific wavelength is more efficient for some types of pathogens), type of genome (DNA, RNA, single or double stranded), exposure time, and the growth medium [61].

On the other hand, direct damage is also produced by photo-activated riboflavin: reactive oxygen species (ROS) are created in a type II photochemical reaction and will interact with the nucleic acids and cell membranes [62–64]. Wollensak and colleagues have compared the cytotoxicity of UVA and riboflavin separately and combined on corneal keratocytes: the combination of UVA and riboflavin showed a 10-fold increased cytotoxicity when compared to UVA alone [65].

An indirect effect of PACK-CXL is the increased postoperative resistance of the cornea against enzymatic digestion [66]. In their experiment, Spoerl and colleagues tested the effect of various enzymes on two groups of porcine corneas ex vivo and showed that corneas pretreated with CXL presented a 2-fold increased resistance against enzymatic digestion [66].

Several clinical trials reported in some patients a temporary increase of the hypopion after PACK-CXL that was resolving in a few days [25, 67]. The transitory progress of the hypopyon in the anterior chamber could be explained by a local Jarisch-Herxheimer reaction to released bacterial endotoxins or attributed to the penetration of riboflavin in the anterior chamber that intensifies the local immune response [67, 68].

In human corneas, several clinical trials reported a reduction of corneal sensation and improvement of the pain perception after PACK-CXL [27, 67]. This could be explained by the diminished number of corneal nerves that can be observed immediately after CXL for KC: confocal bio-microscopy shows disappearance of the subepithelial plexus and anterior-midstromal nerves that completely heals 6 months after surgery with a corneal sensitivity back to baseline [69].

Schnitzler and colleagues first used CXL to treat advanced noninfectious melting corneal ulcers before using alternative surgery such as a chaud penetrating keratoplasty (PKP) [70]. Four patients were selected with progressing ulcers of various origins: rosacea, trophic ulceration, transplant melting, and one unclear keratitis. They performed CXL using a 30 min treatment time with energy of 2.5 mW/cm² at a wavelength of 370 nm with a 6–7 mm diameter beam and 0.1% riboflavin solution. The treatment successfully completely halted the progression of the melting process in three patients and allowed to delay the surgical procedure without any emergent intervention. The CXL did not halt the progression of the melting process in the fourth patient who underwent a chaud keratoplasty 3 weeks after the procedure.

It is only 8 years later that the first clinical observation on infectious melting keratitis took place. Iseli and colleagues treated five patients that presented with therapy-resistant progressing infectious corneal melting: four patients had a history of LASIK and one presented with a contact lens-induced infectious keratitis [71]. Various pathogens were responsible for the infection; three patients presented ulcers from bacterial origin (mycobacterium species) and two patients presented filamentous fungi infections (fusarium and acremonium spp). All patients were treated with PACK-CXL at an irradiance of 3 mW/cm² for 30 min at 365 nm wavelength with 0.1% riboflavin solution similar to the Dresden protocol. In all but one case, patients presented reduced infiltrate size and arrested melting process after the treatment. The remaining patient showed progressing melting caused by a persistent immune reaction without any sign of a residual fungal infection. This study introduced PACK-CXL as a mean to successfully delay a chaud keratoplasty in infectious keratitis. Other case series confirmed the initial success of halting melting progression in infectious keratitis [72–74]. Makdoumi and colleagues treated seven eyes with bacterial keratitis and PACK-CXL stopped the melting process as well as allowed for a complete re-epithelialization in all eyes [27]. Larger studies were also organized to evaluate the efficacy of PACK-CXL in bacterial keratitis. Price and colleagues published a prospective dual-center case series on 40 eyes in 2012 [75]. They treated patients with PACK-CXL concomitantly to antibiotics treatment with excellent outcomes for bacterial keratitis. Said and colleagues set up a study comparing patients receiving antibiotics and PACK-CXL with a control group of patients receiving the standard antibiotic regimen [25]. They concluded that the smaller the infiltrate, the faster the re-epithelialization was completed. In advanced cases, however, the additional PACK-CXL did not significantly shorten the time to complete epithelialization. They concluded that PACK-CXL might be suited to treat early ulcers rather than advanced cases [25].

Following these successes in treating therapy-resistant bacterial diseases, Makdoumi and colleagues published in 2011 a prospective nonrandomized clinical study to evaluate the potential efficiency of PACK-CXL alone as a first-line treatment in early infectious keratitis [76]. They treated 16 patients diagnosed with suspected bacterial keratitis that did not receive any prior antibiotics by PACK-CXL using the Dresden protocol. All eyes presented a reduction of infection and inflammation with complete epithelial healing. Two patients needed an additional antibiotic therapy during follow-up controls: the first patient was diabetic and showed a slow re-epithelialization rate. The second patient presented with a deep infection that required additional treatment to contain the inflammatory response and allow for epithelial healing. This clinical trial demonstrated the proof of principle that PACK-CXL can also be efficient as a first-line treatment for bacterial keratitis without any adjunct of antibiotics.

In their initial study in 2008, Iseli and colleagues already included two fungal keratitis caused by filamentous fungi: Acremonium and Fusarium spp. [71]. The first patient presented an infection related to a LASIK surgery and the second one to contact lens use. Both patients were treated with antimycotic therapy before receiving PACK-CXL. The surgery successfully delayed emergency keratoplasty in both cases, but the second patient showed further progression at 3 weeks after treatment and received a PKP with good clinical outcome. Other clinical trials reported also mitigated results in fungal keratitis treatment with PACK-CXL when concomitant to standard antimycotic therapy [67, 75, 77, 78]. In a large retrospective study involving 41 cases of mycotic keratitis, Vajpayee and colleagues compared standard antifungal therapy with or without additional PACK-CXL [78]. They did not find any statistically significant difference in any of the postoperative parameters (infiltrate size, CDVA, time to healing, complications, vascularization) although they recognized the study design as a limiting factor [78]. On the other hand, Li and colleagues treated eight eyes with either Fusarium or Aspergillus keratitis with adjuvant PACK-CXL and reported good clinical outcome after the surgery for all cases [79]. Muller and colleagues reported also a patient successfully treated with PACK-CXL while suffering from Candida Albicans keratitis [73]. Some more clinical trials also reported effective results with PACK-CXL in combination to the standard therapy [25, 80, 81].

In another noteworthy in vitro experiment, Arboleda and colleagues tested PACK-CXL with a modified chromophore on three fungi: Fusarium, Aspergillus, and Candida [82]. As in other studies, they demonstrated that riboflavin 0.1% combined with UVA was inefficient in treating alone those fungi. But on the contrary when using RB as an alternative chromophore and irradiating the fungal isolates with a 518 nm wavelength, they were able to inhibit the growth of all three fungi. These promising results need to be confirmed in vivo and safety issues should also be assessed [82].

In vivo, Galperin and colleagues tested PACK-CXL to treat Fusarium keratitis. They used 24 New Zealand White rabbits, infected them with Fusarium solani isolate, and separated them in two groups: one serving as control did not receive any treatment and the other group was treated with PACK-CXL 5.4J/cm² procedure (30 min, 3 mw/cm², 370 nm wavelength) [83]. At 1 week the PACK-CXL group showed a significantly reduced clinical gravity score. Ex vivo pathological studies showed lower Fusarium hyphae and reduced inflammation compared to the control group. Table 9.2 gives an overview of the various studies related to fungal infections.

PACK-CXL has also been successfully used in animals. Hellander and colleagues treated nine horses presenting melting corneal keratitis from fungal and bacterial origin with PACK-CXL concomitant to regular antimicrobial treatment [84]. In this study, a modified Dresden protocol was applied. Eight horses out of nine improved their infection and closed their epithelium. One horse with a deep fungal keratitis did not improve after PACK-CXL and medical treatment. Enucleation was performed to treat a consequent endophthalmitis. As in human, deep infections do not seem to respond well to standard PACK-CXL protocol. A prospective interventional study by Pot and colleagues recruited 49 cats and dogs with melting keratitis and reported usefulness of PACK-CXL as an adjunctive therapy in case of melting corneal infective keratitis [24]. Another study on cats and dogs showed similar results [26]. In two studies, Famose tested the efficiency of accelerated PACK-CXL on cats and dogs [85, 86]. He used a modified Dresden protocol where he soaked the cornea with a 0.1% riboflavin solution for 30 min and then irradiates the lesion with UVA at 370 nm wavelength and 30 mW/cm² irradiance for 3 min. In both studies all cases showed good results with clinical improvement and epithelial closure. Thus, PACK-CXL seems also to be an efficient therapy to treat infectious keratitis in various types of animals. The limitations seem to be the same as in human regarding depth of treatment for profound stromal infections. Furthermore, as corneal thickness is variable between species, modifications of protocol parameters could improve or accelerate treatment efficiency [85, 86].

PACK-CXL is a recent and promising technology in treating infectious keratitis. It demonstrates efficiency in stabilizing melting processes and early ulcers [71, 76].

One potential complication of UV light treatment could be reactivation of herpes simplex virus. Price and colleagues treated a patient with negative culture and melting cornea that developed a dendritic lesion after the treatment [75]. Another case report presented a patient treated with standard CXL for KC that developed herpes simplex keratitis after the treatment although she had negative history for herpes infections [87]. They concluded that PACK-CXL might activate herpes simplex viruses and thus recommend avoiding any UV light treatment in patients with active herpes or positive previous history. Depth of infection is also an important factor to take in account when deciding to treat an ulcer. It seems to be even more relevant in fungal and amibian infections as those ones are usually more profound than bacterial infections. Depth of infection could thus represent a major risk factor for treatment failure 68. Protocol parameters could also be adapted to the type of pathogen and the depth of infection. New chromophores such as RB could be an alternative tool when treating resisting pathogens to standard PACK-CXL protocol. Accelerated treatment shows also interesting perspectives. Recent data from our group studied the effect of accelerated protocols on several different pathogens [88]. We found that PACK-CXL maintains a high killing rate even when performed with accelerated settings respecting the Bunsen-Roscoe law of reciprocity. Thus, protocol parameters modifications could shorten the time of treatment while keeping the same microbicidal effect.