Article Text

Download PDFPDF

Intrastromal voriconazole for deep recalcitrant fungal keratitis: a case series
  1. Ganapathy Kalaiselvi,
  2. Sivananda Narayana,
  3. Tiruvengada Krishnan,
  4. Sabyasachi Sengupta
  1. Cornea and Refractive Services, Aravind Eye Hospital, Pondicherry, India
  1. Correspondence to Dr Ganapathy Kalaiselvi, Cornea and Refractive Services, Aravind Eye Hospital, Thavalakuppam, Pondicherry 605007, India; kalaisvrain{at}


Aim To evaluate the outcomes of treating deep recalcitrant fungal keratitis with intrastromal voriconazole injection.

Methods Twenty-five patients with culture proven fungal keratitis, not responding to a combination of topical 5% natamycin and 1% voriconazole were treated with intrastromal voriconazole (50 µg/0.1 mL) injected in five divided doses around the infiltrate to form a depot of the drug around the circumference of the lesion.

Results The mean age of the patients was 52.52±12.21 years and mean time to presentation was 17.12±13.75 days from the onset of symptoms. The mean area of the infiltrate was 30.41±17.2 mm2, hypopyon was present in 88% and all cases had infiltrates that extended beyond the mid-stromal level. Intrastromal voriconazole helped to resolve the infection in 18 (72%) patients and about 15% of these needed more than one injection. Smaller ulcers responded better to treatment. Fusarium spp were responsible for six of the seven cases that failed treatment.

Conclusions Targeted delivery of voriconazole by intrastromal injection (50 µg/0.1 mL) is a safe and effective way to treat deep recalcitrant fungal keratitis, though some may need repeated injections. Fusarium keratitis may show suboptimal response but this needs further study.

  • Fungal keratitis
  • voriconazole
  • intrastromal
View Full Text

Statistics from



A large proportion of infectious corneal ulcers that are referred to tertiary care centres in south India are fungal in aetiology.1 Management of fungal keratitis is challenging due to fungistatic effect of most of the topical antifungal agents and their poor penetration to the deeper layers of the cornea resulting in suboptimal therapeutic levels at the site of infection.2 ,3 Therefore, treatment of fungal keratitis generally results in protracted therapy with poor final outcomes.4 Newer antifungal agents, such as voriconazole, posaconazole, ravuconazole and caspofungin are fungicidal and have been shown in laboratory studies to have a very good safety profile and better corneal penetration than the conventional topical antifungal agents such as natamycin, thereby increasing the potential for better treatment outcomes.5–7

Therapeutic keratoplasty (TPK) is the last choice for eradicating fungi from the corneal surface and preserve its structural integrity. However, TPK has a lower success rate in terms of graft reinfection and graft failure compared with that performed for bacterial keratitis.8 An increase in the referral of fungal keratitis to tertiary care centres coupled with a decreased availability of donor corneas in the developing world merit the study of alternative and lesser invasive modalities for treating eyes with aggressive fungal keratitis.

Targeted drug delivery has the potential to achieve adequate drug concentrations at the site of infection and serve as an alternative modality of treatment in eyes with recalcitrant fungal keratitis.9–18 Among the targeted drug delivery modalities, intrastromal injection of antifungal agents has shown promising results.12–18 This study was done to evaluate the structural and functional outcomes of intrastromal voriconazole in deep recalcitrant fungal keratitis showing poor response to topical antifungals.

Materials and methods

The study was approved by the ethics committee and the institutional review board of the Aravind eye care systems. Eyes with culture proven fungal keratitis, not responding to topical antifungal medications, including natamycin and voriconazole, were enrolled in this prospective interventional study after obtaining an informed consent, and eligible eyes underwent intrastromal injections of voriconazole (50 µg/0.1 mL).

Corneal infiltrates negative for fungus by 10% potassium hydroxide mount, Gram staining and culture, mixed corneal infections, perforated corneal ulcers or those with impending perforation, presence of descemetocele, involvement of the adjacent sclera, ulcers with clinical features of non-infective and autoimmune conditions, fungal ulcer associated with endophthalmitis, patients <16 years and beyond 80 years and patients with one eye were excluded from the study.

At initial presentation, each patient underwent a detailed clinical evaluation including the mode of injury, duration of symptoms, occupation and best corrected Snellen visual acuity testing. Slit-lamp biomicroscopy was performed and the size of ulcer, depth of the infiltrate, presence of an endothelial plaque and height of the hypopyon were recorded. The area of the ulcer was calculated from its maximum diameter and the dimension perpendicular to the maximum diameter. Corneal scrapings were obtained under topical anaesthesia in a standardised manner as described previously, and were sent for microbiological investigations including potassium hydroxide wet mount preparation, Gram stained smear, and cultures on blood agar, chocolate agar and potato dextrose agar.19

Topical natamycin sulfate (5%) monotherapy was instilled once every hour for the first 2 weeks after the initial diagnosis. The response to therapy was noted on slit lamp examination and defined as ‘not improved’ if there was no change in the size and area of the ulcer or infiltrates and defined as ‘worsened’ if there was an increase in area, size or depth of the ulcer or infiltrate by 20%, or perforation. The ulcer was defined as ‘healing’ if the area and size of the epithelial defect and the infiltrate reduced by more than 20% from that at presentation. If there was no improvement or response to therapy after 2 weeks or if the infection showed signs of worsening, topical voriconazole 1% (VOZOLE; Aurolab, India) eye drops were added once every hour. If no response to this combined therapy was observed after two more weeks, the patients received intrastromal injection of voriconazole (50 µg/0.1 mL) around the fungal infiltrate and were included in the study. Oral antifungal medications were not administered to any of our patients during the course of treatment.

Technique of intrastromal injection

Injection voriconazole (VOZOLE PF; Aurolab, India) is available as 1 mg of white lyophilised powder in a glass vial. The powder was reconstituted with 2 mL of lactated Ringer's solution to a concentration of 0.5 mg/mL (50 µg/0.1 mL). The reconstituted solution was loaded in a 1 ml tuberculin syringe with a 30-gauge needle.

After administration of peribulbar anaesthesia, the patient was shifted to the operating table. Under full aseptic conditions, the preloaded drug was administered under the operating microscope. A dose of 50 µg/0.1 mL reconstituted intrastromal voriconazole was prepared fresh each time. With the bevel down, the needle was inserted obliquely from the uninvolved, clear area to reach the infiltrate at the mid-stromal level (as the intended level for drug deposit) in each case. The drug was then injected and the amount of hydration of the cornea was used as a guide to assess the area covered. Once the desired amount of hydration was achieved, the plunger was withdrawn slightly to ensure discontinuation of the capillary column and thus prevent back-leakage of the drug. Five divided doses were given around the infiltrate to form a deposit of the drug around the circumference of the lesion. Circumferential injection ensured the formation of a barrage of intrastromal voriconazole around the entire infiltrate. Intraoperative complications, if any, were recorded.


After the intrastromal injection, all patients were continued on the previously mentioned topical antifungal regimen. Patients were examined every day and the response to therapy was recorded on the slit lamp. Once the infiltrate showed signs of healing, the patients were reviewed after 1 week, then once every 2 weeks for 3 months or until the ulcer had healed completely. On each follow-up the patient’s best corrected visual acuity was recorded and size of the infiltrate, height of the hypopyon and occurrence of any complications were noted by slit-lamp biomicroscopy. The infection was considered resolved when there was complete healing of epithelial defect with resolution of corneal infiltrate and scar formation. The patients were continued on topical antifungal therapy for at least 2 weeks after the complete resolution of infection.

In case of worsening or no response to the previous injection within 1 week, the intrastromal injection of voriconazole was repeated. Patients with impending perforation underwent application of cyanoacrylate glue with bandage contact lens along with continuation of the topical antifungal regimen as mentioned previously. Patients with perforations and progressive worsening of infiltrates despite the intrastromal voriconazole were taken up for TPK.

The final outcome was assessed in terms of treatment success, which included patients in whom the ulcer healed with scar formation and treatment failure which included non-healing, progressive infiltrate and perforation.

Statistical analysis

The data were entered in an Excel data sheet and analysed using STATA statistical software package (V.11, Texas, USA). The continuous variables are presented as mean±SD and the categorical variables are presented as percentages. The differences in the demographics and clinical parameters between those that had treatment success and those whose treatment failed were compared using the Wilcoxon rank-sum test (for continuous variables) and the Fisher's exact test (for categorical variables).


The study included 25 eyes of 25 patients. All patients had deep-seated corneal infiltrates with or without hypopyon and endothelial plaque. The depth of corneal involvement in all these cases extended up to or deeper than the mid-stromal level (figure 1A, B). The age of our patients ranged from 20 years to 80 years with the mean being 52.5±12.2 years. There were 14 men (56%) and 11 women (44%). A history of trauma with agricultural agents was elicited from 18 patients (72%) out of which five patients gave a history of taking some form of native treatment like breast milk and plant juice. The mean time to presentation at our centre was 17.1±13.8 days from the onset of symptoms. Systemic disease was present in four (20%) patients including uncontrolled diabetes in two, asthma in one and leprosy in one.

Figure 1

(A and B) Show eyes with deep stromal keratitis that were included in the study, (C) shows an eye that required a second injection of intrastromal voriconazole and (D) shows an eye that showed good resolution of keratitis with scarring.

The mean initial visual acuity at presentation was 2.3±0.5 logMAR units. The mean infiltrate area was 30.41±17.2 mm2. Surface pigmentation was present in 4 patients (16%) and hypopyon was present in 22 (88%). The mean height of the hypopyon was 1.76±0.50 mm. Nineteen eyes (76%) showed moniliaceous (non-pigmented) fungi and six (24%) eyes showed dematiaceous (pigmented) fungi in culture. Table 1 shows a list of the fungal species isolated in culture. The predominant fungal species isolated were Fusarium spp in 13 patients (52%).

Table 1

Fungal pathogens isolated

Of the 25 patients enrolled in the study, intrastromal injections of voriconazole helped resolve the infection in 18 (72%) patients, of which 16 cases resolved without any additional intervention while two patients had impending perforation during follow-up and were managed successfully using cyanoacrylate glue with bandage contact lens. Out of the 16 successfully treated patients, 12 patients received a single injection, 2 received the injection twice and 2 had three injections. Figures 1C and D show an example of a patient who required a repeat injection and showed successful outcome. Of the two patients who developed impending perforation during follow-up, one had received a single injection and the other received two injections. All patients presented with slight pain immediately after the injection despite anaesthesia.

In seven (28%) patients, treatment with intrastromal injections failed. Six out of these had progression of infection despite the injections and finally received therapeutic penetrating keratoplasty to eradicate infection, while one patient developed corneal perforation requiring an emergency tectonic penetrating keratoplasty to maintain structural integrity. Among these seven eyes that failed, four received two injections and three received one injection. The organism identified was Fusarium in six of the seven patients that failed and in one (14%) it was an unidentified hyaline fungus. Of these seven patients, two were patients with diabetes, with inadequate control of blood sugar levels, and two patients had used topical steroid drops prior to presenting to us.

Out of 18 successfully treated patients, 16 patients had improvement in best corrected visual acuity and two patients had no change in visual acuity. The mean healing time was 45.68±11.49 days.

The infiltrate area and height of the hypopyon were important factors associated with treatment success (table 2).

Table 2

Significant predictors of outcome of intrastromal voriconazole


Voriconazole has the best in vitro efficacy against pathogenic fungi compared with other agents.5–7It has low minimum inhibitory concentration (MIC) values against Candida and Aspergillus spp known to be resistant to amphotericin B, fluconazole and itraconazole.6,20,21 In order to achieve adequately deep intracorneal concentration of antifungals, intrastromal injections of voriconazole have been described in the past.12–18 In our study, voriconazole 50 µg/0.1 mL was injected around the infiltrate to form a drug deposit at the circumference of the lesion, thus forming a barrage of intrastromal voriconazole around the entire infiltrate. We believe that this technique of injection, as described previously by other authors, has the best chance of achieving MIC of the drug at the tissue level.

Prakash et al13 reported successful use of intrastromal voriconazole in three patients with fungal keratitis. Similarly, Tu16 reported success with intrastromal voriconazole in cases of Alternaria infection, and Jain et al18 reported good outcome in a case of fungal infection of the phacoemulsification site tunnel with intrastromal voriconazole. In our study, all infiltrates involved more than half the thickness of the stroma and were unresponsive to 4 weeks of topical natamycin and voriconazole, and about two-thirds of our patients responded to intrastromal voriconazole.

Need for repeat injections have been reported by Tu16 previously for Alternaria keratitis with good results. Sharma et al,14 also reported that 10 of 12 eyes required two or more injections to attain optimal response. In our series, 15% of eyes that showed treatment success required reinjections.14 ,16 As there is little knowledge of the pharmacokinetics of voriconazole when injected into the corneal stroma, need for reinjection and the interval between injections needs to be determined on clinical grounds and may differ from case to case. However, it appears from our own results and also from the literature, that a considerable number of such patients with deep keratomycosis require repeated injections with good results.

Treatment success was achieved in 72% of our patients; similarly Sharma et al14 reported a success rate of 83% (10/12 resolved). We suspect that if not for intrastromal voriconazole, many more would have required keratoplasty. However, to establish this beneficial treatment effect, an adequately powered randomised clinical trial is required for subjects with deep keratomycosis.

We found that the infiltrate area and height of the hypopyon were the most important determinants of treatment success. Lalitha et al22 have also showed that an ulcer area of more than 14 mm2, presence of a hypopyon and infection with Aspergillus spp are important risk factors for treatment failure. In our study, out of seven failures, Fusarium spp was responsible for six cases (86%). Overall, we found that only half the eyes (7/13) with Fusarium keratitis responded to our treatment regimen, illustrating the fact that Fusarium spp may be resistant to many of the antifungal drugs that are currently used to treat fungal infections.23 Some Fusarium spp, such as F. verticillioides, are resistant to most antifungal drops. Among these resistant fungi, F. solani is the most resistant and, along with F. oxysporum, is the most common aetiological agent in fungal infections.23 Though our results are inferior to those reported by Siatiri et al15 for treating Fusarium keratitis, the mycotic ulcer treatment trial, a large randomised clinical trial comparing topical natamycin with voriconazole for the treatment of fungal keratitis also found unfavourable response to Fusarium keratitis in the voriconazole arm.24

Sharma et al25 found no significant benefit of adding intrastromal voriconazole to topical natamycin for treating recalcitrant fungal keratitis in 40 eyes randomly assigned to receive topical or intrastromal voriconazole in addition to natamycin. However, we injected intrastromal voriconazole in eyes that did not respond to combination of topical voriconazole and natamycin. Sharma et al did not have a similar interventional arm in their study. It may be worthwhile to attempt intrastromal voriconazole in eyes with deep fungal keratitis that appear to be unresponsive to combination of topical antifungal medication before considering TPK.

In conclusion, the judicious and timely use of intrastromal voriconazole may be undertaken in selected patients with deep keratomycosis who are unresponsive to topical antifungal therapy. A majority of eyes show a good response to this targeted treatment, though a few patients may require repeated injections. The area of the infiltrate and the height of the hypopyon are important determinants of treatment success in such eyes. Fusarium spp appear to show suboptimal response but this phenomenon needs to be investigated in a better manner in future studies.


View Abstract


  • Contributors KG and SN were involved in manuscript preparation, design and execution of the study. SS was involved in data collection, statistical analysis and manuscript writing. TK was involved in all the steps and mentored the rest of the coauthors.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics Committee and the Institutional Review Board of the Aravind Eye Care Systems.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost Jonas