Evaluation of liposomal amphotericin B for the treatment of fungal keratitis in a tertiary eye care hospital

Abstract

Purpose:

To evaluate the efficacy of liposomal amphotericin B (L-AMB) for the treatment of fungal keratitis.

Methods:

Patients with fungal keratitis confirmed by potassium hydroxide (KOH) smear and/or confocal microscopy were administered topical L-AMB and randomized into three groups treated with three different formulations. The medication was administered two hourly till clinical improvement was achieved, followed by six hourly till complete resolution. The outcome measures were time to clinical improvement, resolution of epithelial defect, stromal infiltrate, hypopyon, extent and density of corneal opacity, neovascularization, and best corrected visual acuity (BCVA) at 3 months.

Results:

Mean age of the patients was 46.6 ± 14.8 years, and trauma with vegetative matter was the most common predisposing factor. Aspergillus flavus (36%) was the most common fungus cultured, followed by Fusarium (23%). Mean time to clinical improvement, time to resolution of epithelial defect, mean time to resolution of infiltrate, and time to resolution of hypopyon were 3.45 ± 1.38, 25.35 ± 8.46, 37.97 ± 9.94, and 13.33 ± 4.90 days, respectively, and they were comparable among the three groups. There was a significant difference between treatment failure and success cases in terms of days of presentation (P < 0.01), size of the epithelial defect (P-value 0.04), and infiltrate size at presentation (P-value 0.04). At 3 months follow-up, no statistically significant difference was noted in BCVA and mean scar size among groups.

Conclusion:

L-AMB in a gel form is an effective antifungal agent that promotes the healing of fungal ulcers with notably least vascularization and better tolerance.

Trial registration number: CTRI/2020/04/024550

Corneal blindness is responsible for about 1.5–2 million new cases of monocular blindness every year.[1] Infectious keratitis is one of the major causes of corneal blindness, especially in the developing world.[2,3] Fungal corneal ulcer remains the important cause of blindness in India and accounts for nearly 50% of all culture-positive cases based on various studies.[1,4-7]

Fungal keratitis is difficult to treat compared to bacterial keratitis and has been reported to have worse outcomes due to poor ocular penetration and limited bioavailability of the various antifungal agents available.[8-10] Based on the clinical evidence available so far, natamycin is the drug of the first choice for topical use in the treatment of fungal keratitis and is the only Food and Drug Administration (FDA)-approved drug.[11,12] Poor success rates and the problem of recalcitrant cases have driven the need for constant exploration of alternative agents in the treatment of fungal keratitis, of which amphotericin B, a macrocyclic polyene, has proven to be useful in cases with deep mycosis, endothelial plaque, and presence of hypopyon and/or inflammation of the anterior chamber.[13,14] Amphotericin B is now available as a liposomal formulation that improves the solubility and bioavailability of the active pharmacological agent, lowering the minimum inhibitory concentration, to improve the clinical response.[15] Topical amphotericin B ointment has been reported to be useful for controlling keratomycosis caused by antimycotic-resistant fungi in experimental animal models.[16,17] A randomized controlled trial demonstrated topical liposomal amphotericin B (L-AMB) sonicated eye drops to be equally efficacious to natamycin in inducing healing and improving visual acuity in moderate fungal corneal ulcers.[18] A new gel form of L-AMB is now available, which remains stable in unilamellar form and does not require refrigeration. Given the previous success of treatment of fungal keratitis with sonicated L-AMB in comparison to conventional treatment with natamycin, this study was undertaken to compare the different formulations of L-AMB, namely, sonicated eye drop, unsonicated eye drop, and gel, in the treatment of fungal keratitis in terms of efficacy, safety, and tolerability.

Methods

Study design and ethical approval

This was a prospective, comparative, interventional, randomized, single-center pilot study. The study was commenced after obtaining ethical approval from the Institute Ethics Committee and was carried out in accordance with the tenets of the Declaration of Helsinki. This trial is registered under the Clinical Trial Registry, National Institute of Medical Statistics, Indian Council of Medical Research, India, with registration number CTRI/2020/04/024550.

Patient characteristics

Forty-five patients aged 18 years and above with fungal hyphae positive on potassium hydroxide (KOH) smear and/or in vivo confocal microscopy were included in the study. Patients with impending perforation or perforated corneal ulcer, bilateral involvement, no light perception in the affected eye, patients who were allergic to the study medication, and pregnant or lactating females were excluded from the study. At the initial presentation, the history and demographic data of each patient were recorded and a basic ophthalmological workup of all patients was done. A detailed history of any systemic disease and local predisposing factors such as corneal trauma with organic matter, contact lens use, and steroid usage was also taken. The best corrected visual acuity (BCVA) of patients was recorded, and intraocular pressure was assessed through digital palpation. The slit-lamp examination included the size of the epithelial defect, the size of the infiltrate, and the size of the hypopyon. B scan ultrasonography was done to rule out any posterior segment abnormality. The epithelial defect was measured by diaphragm slit on a slit-lamp biomicroscope after staining with fluorescein dye under a cobalt blue filter. The longest perpendicular diameter was measured and noted. The ulcers were divided into mild, moderate, and severe based on the grading system by Harrison [Table 2].[19,20] A Kimura spatula using aseptic techniques was used to obtain a scrape from the leading edge and base of the corneal ulcer. The scrapings were smeared directly onto two separate microbiology glass slides for Gram stain and KOH wet mount. The patients were randomized into three groups using a random number table as group 1, group 2, and group 3 and were started on different formulations of L-AMB in the form of gel, sonicated drop, and unsonicated drop, respectively.

Table 2.

Grading of corneal ulcer

	Mild	Moderate	Severe
Size	<2 mm	2-5 mm	>5 mm
Depth of infiltrate	<20%	20%-50%	>50%
Scleral involvement	Not involved	Not involved	May be involved

Drug formulations

Three formulations of L-AMB were used in this study. First, the gel form of L-AMB (0.1%) is commercially available as Fungisome gel (Lifecare Innovation, Lucknow, India), which was used for group 1. The second group was treated with L-AMB 0.1% (Fungisome, Lifecare Innovation) containing (10 mg/10 ml) injectable unsonicated form, which was further sonicated using a commercially available sonicator for 90 min and used as sonicated eye drops after dispensing it in 5-ml bottles for group 2. Third, the unsonicated form (10 mg/10 ml) was dispensed in 5-ml bottles and used in group 3.

Study parameters and outcomes

The medications were initially started two hourly till clinical improvement was achieved, followed by six hourly till complete resolution. Patients were followed up on days 3, 7, 14, 21, 28, and 90. On each follow-up visit, various parameters were evaluated, which included BCVA, digital tonometry, size of the epithelial defect, size of hypopyon, size of infiltrate, corneal neovascularization, and drug tolerability. Treatment success was defined as resolution of the corneal infiltrate with scarring, disappearance of the hypopyon, and a healed epithelial defect. Treatment failure was considered if the infiltrate/epithelial defect size increased by 2 mm or more and/or there was an increase in the size of hypopyon or the endothelial plaque continued to enlarge for three consecutive days or there was perforation of ulcer.

Outcome measures

Primary outcome measures included time taken for the epithelial defect to heal, time taken for resolution of hypopyon, and time taken for resolution of infiltrate. BCVA, corneal scarring, and corneal neovascularization at 3 months were included as secondary outcome measures.

Statistical analysis

Data were recorded in a Microsoft Excel sheet, and the statistical test was carried out using the Statistical Package for Social Sciences (version 27.0.1). Data were presented as mean ± standard deviation (SD) and range. Baseline parameters were compared by Chi-square or Fisher’s exact test (categorical) and independent t-test or Wilcoxon rank-sum test (continuous variables) as applicable. The Mann–Whitney U test was used for variables that were not normally distributed. P values < 0.05 were considered statistically significant.

Results

The present study was conducted on 45 eyes of 45 patients with a fungal corneal ulcer. The patients were divided into three groups of 15 patients each as group 1, group 2, and group 3 and were treated with gel, sonicated drop, and unsonicated drop of L-AMB (0.1%), respectively.

Demographic and baseline clinical characteristics

The mean ages of patients in group 1, group 2, and group 3 were 42.2 ± 13.2, 45 ± 15.1, and 52.6 ± 15.07 years, respectively (P-value 0.13). The fungal corneal ulcer was more common in males (62.2%). Baseline logMAR BCVA was 1.72 ± 0.38 in group 1, 1.58 ± 0.51 in group 2, and 1.62 ± 0.48 in group 3. There was no significant difference in BCVA among the three groups (P-value 0.70) at presentation. In the present study, 69% of the patients had moderate corneal ulcers, whereas 31% had severe corneal ulcers. The distribution of moderate corneal ulcers (P-value 0.66) and severe corneal ulcers (P-value 0.77) was comparable in the three groups. The most common predisposing factor for fungal ulcers was found to be vegetative matter injury in 51% of patients, which was followed by sand particle injury in 20% of patients. Diabetes mellitus was the only systemic risk factor that was found in 4% of patients. Fungal culture was positive in 26 patients (57%) and the most common fungus isolated was Aspergillus flavus (35%) followed by Fusarium sp. (23%). Other species isolated from the culture in this study included Aspergillus niger (15%), Acremonium (11%), Alternaria (8%), and Curvularia (7%) [Table 1].

Table 1.

Demographic and baseline characteristics of patients in three groups

	Group 1 Sonicated gel	Group 2 Sonicated drop	Group 3 Unsonicated drop	P
Age (years)	42.2±13.2	45±15.1	52.6±15.07	0.13
Gender (male/female)	11/4	8/7	9/6	0.51
Eye (right/left)	8/7	9/6	11/4	0.51
BCVA±SD	1.72±0.38	1.58±0.51	1.62±0.48	0.70
Vegetative matter injury	8	6	9	0.961
Confocal scan positive	13 (86.6%)	13 (86.6%)	12 (80%)	0.73
KOH smear positive	10 (66.67%)	9 (60%)	9 (60%)	0.910
Hypopyon positive	9 (60%)	10 (66.6%)	5 (33.33%)	0.945
Epithelial defect size (mm)	3.69±1.23	3.62±1.64	3.55±1.24	0.962
Infiltrate size (mm)	4.34±1.45	4.67±1.68	4.9±1.31	0.595
Culture positive	8 (53.33%)	9 (60%)	9 (60%)	0.684
Aspergillus flavus	3 (20%)	3 (20%)	3 (20%)
Fusarium species	2 (13.3%)	2 (13.3%)	2 (13.3%)
Aspergillus niger	1 (6.66%)	1 (6.66%)	2 (13.3%)
Acremonium	2 (13.3%)	0	1 (6.66%)
Alternaria	0	2 (13.3%)	0
Curvularia	0	1 (6.66%)	1 (6.66%)

SD=standard deviation

Hypopyon was present in 24 patients (53.3%). In group 1, it was present in nine patients (60%), whereas in group 2 and group 3, it was present in 10 (66.6%) and five (33.3%) patients, respectively. In group 1, the mean hypopyon size at presentation was 1.05 ± 0.58 mm and in group 2 and group 3, it was 1.01 ± 0.42 and 1.2 ± 0.75 mm, respectively. The mean epithelial defect size at presentation in group 1 was 3.69 ± 1.23 mm and in group 2 and group 3, it was 3.62 ± 1.64 and 3.55 ± 1.24 mm, respectively. The epithelial defect size at presentation was comparable in the three groups (P-value 0.96). In group 1, the mean infiltrate size at presentation was 4.34 ± 1.45 mm and in group 2 and group 3, it was 4.67 ± 1.68 and 4.9 ± 1.31 mm, respectively. The mean size of the infiltrate was comparable in all three groups (P-value 0.59).

Stinging sensation on applying medication was complained by all the patients of group 2 and group 3 (n = 30, 100%). It was less commonly observed with group 1 patients. Only five patients of group 1 complained of a stinging sensation on putting gel in the eye. The results were statistically significant on comparison (P < 0.01). The mean time to resolution of epithelial defect, resolution of infiltrate, and resolution of hypopyon was 25.35 ± 8.46, 37.97 ± 9.94, and 13.33 ± 4.90 days, respectively.

The mean time to resolution of the epithelial defect in group 1, group 2, and group 3 was 26 ± 6.96, 24.5 ± 10.55, and 25.45 ± 8.44 days, respectively. The mean time to resolution of infiltrate in group 1, group 2, and group 3 was 36 ± 6.64, 40.33 ± 12.75, and 37.90 ± 10.34 days, respectively. The time to resolution of the epithelial defect (P-value 0.90) and infiltrate (P-value 0.55) was comparable in the three groups. The mean time to the disappearance of hypopyon in group 1, group 2, and group 3 was 14 ± 4.94, 13.12 ± 4.48, and 12.25 ± 6.70 days, respectively. The time to the disappearance of hypopyon (P-value 0.63) was comparable in the three groups The mean BCVA in logMAR at 3 months in group 1, group 2, and group 3 was 0.9 ± 0.53, 0.84 ± 0.83, and 0.9 ± 0.33, respectively. There was no significant difference in visual acuity at 3 months among the three groups (P-value 0.91). All patients healed with corneal opacity at the end of 3 months [Figs. 1-3]. In group 1, five patients (35.71%) healed with macular corneal opacity and nine patients (64.29%) healed with leucomatous corneal opacity. In group 2, five patients (41.67%) healed with macular corneal opacity and seven patients (58.33%) healed with leucomatous corneal opacity. In group 3, five patients (45.45%) healed with macular corneal opacity and six patients (54.55%) healed with leucomatous corneal opacity. There was no significant difference in corneal opacity at the end of 3 months among the groups (P-value 0.882). The mean size of corneal opacity in group 1, group 2, and group 3 was 2.18 ± 0.88, 2.22 ± 1.11, and 2.27 ± 0.9 mm, respectively. There was no significant difference in the size of corneal opacity at 3 months (P-value 0.98). Superficial corneal vascularization and deep corneal vascularization were present in all patients in the present study at the end of 3 months. In group 1, seven patients (58.33%) had deep corneal vascularizations in two quadrants, two patients (16.67%) had deep corneal vascularizations in three quadrants, and three patients (25%) had deep corneal vascularizations in all four quadrants. In group 2, nine patients (90%) had deep corneal vascularizations in more than three quadrants and one patient (10%) had two quadrants of deep vascularization. In group 3, 10 patients (90%) had deep corneal vascularizations in all four quadrants and one patient (10%) had deep vascularization in three quadrants.

Figure 1.

Clinical photographs of fungal corneal ulcer treated with a gel form of amphotericin B 0.1% on day 1 (a) and day 90 (b)

Figure 3.

Clinical photographs of fungal corneal ulcer treated with unsonicated eye drop of amphotericin B 0.1% on day 1 (a) and day 90 (b). SD = standard deviation

Figure 2.

Clinical photographs of fungal corneal ulcer treated with sonicated eye drop of amphotericin B 0.1% on day 1 (a) and day 90 (b)

There were eight treatment failure cases. In group 1, there was only one failure (6%), whereas in group 2 and group 3, there were three (20%) and four (26.6%) patients, respectively. In group 1, one treatment failure patient underwent therapeutic keratoplasty; in group 2 and group 3, the number of patients undergoing keratoplasty was three and two, respectively. Two patients healed with leucomatous corneal opacity after starting natamycin therapy. There was a significant difference between treatment response and treatment failure cases in terms of days of presentation (P < 0.01), infiltrate size (P-value 0.04), and epithelial defect at presentation (P-value 0.04). There was no significant difference between these two groups in BCVA and hypopyon size at presentation and the type of ulcer.

Discussion

The trend of infectious keratitis has changed over the past 10 years in terms of increasing cases of fungal keratitis including its management with topical antifungal agents, whereas surgical management is reserved for nonresponding, recalcitrant, and perforated corneal ulcers.[5,19] The commonly used antifungal agents in the treatment of fungal corneal ulcers are natamycin, amphotericin B, itraconazole, voriconazole, and ketoconazole.[21] These antifungal agents have various limitations such as poor ocular penetration, patient compliance, development of resistance among fungi, and need for proper storage. Currently, the studies done on L-AMB have shown that this formulation has better antifungal action than the conventional form of amphotericin B. Conventional amphotericin B has poor ocular penetration and it is epitheliotoxic.[15,16] Liposomes can increase the affinity between amphotericin B and sterols in the fungal cell and, at the same time, reduce the affinity of the mammalian cell membrane to cholesterol. Therefore, they not only increase antifungal activity, but also reduce damage to the host tissues in experimental rabbit models.[22]

Sonicated formulation of L-AMB is a preparation in which the drug is present in a unilamellar form that is stable for 24 h at room temperature and has better absorption. L-AMB needs refrigeration for its storage to prevent its conversion into multilamellar form. Due to this, other formulations of L-AMB that have good ocular penetration, can be stored at room temperature, and have similar antifungal activity need to be evaluated. The gel form is a new formulation in which amphotericin B is present in the sonicated form, and it does not require refrigeration for its storage. To the best of our knowledge, no clinical data is available in human eyes investigating different formulations of topical L-AMB in fungal keratitis. Hence, this is the first study evaluating the efficacy of three different formulations of L-AMB in the treatment of fungal corneal ulcers.

In the present study, 73% of patients were in the age group of 18–58 years. This age group constitutes the working population of the country and they are involved in maximal outdoor activities. So, fungal corneal ulcer is a major economic burden to any country. The majority of studies done in various parts of India reported that fungal corneal ulcers are more common in males compared to females. This is mainly due to the reason that males are more commonly involved in outdoor activities.[9,11] In our study, 62% of the patients were male, which is similar to that reported in previous studies.

The most common predisposing factors in the present study were found to be vegetative matter injury followed by sand particle injury. Multiple studies done previously have also reported corneal trauma, particularly with vegetative matter, to be the most common predisposing factor.[23-26] Diabetes mellitus was the only systemic risk factor that was found in 4% of our patients. Filamentous fungi, especially Aspergillus sp. and Fusarium sp., are the commonest causative fungal pathogens for corneal ulcers reported from our country.[2,5,8] Our results are similar wherein all the fungi isolated on culture were filamentous, that is, Aspergillus sp. (35%), Fusarium sp. (23%), Curvularia sp. (8%), and Alternaria sp. (8%).

In the present study, the mean visual acuity at presentation was 1.64 ± 0.45 (logMAR). In a study done by Prajna et al.,[27] the mean baseline visual acuity reported among patients was 0.93 (logMAR). The difference in initial visual acuity may be due to the nonuniform recruitment of the patients. The mean duration of the presentation of symptoms in this study was 7 days. Gopinathan et al.[8] reported that 66% of patients presented within 1–2 weeks of the onset of symptoms. Larger-sized ulcers had a longer duration of presentation and also poorer prognosis.

The mean time to resolution of the epithelial defect in our study was 25 days. In a study done by Yoon et al.,[28] they reported the mean time to closure of epithelial defect as 20 days in the intracameral amphotericin B group and 33 days in the topical amphotericin group. The possible reason for this difference in the topical amphotericin B group may be due to the use of conventional amphotericin B. We observed that the mean time to resolution of infiltrate was 38 days. Yoon et al.[28] reported the mean time to resolution of infiltrate in their study as 52 days in the topical amphotericin B group. Parchand et al.[29] reported the mean time to resolution of infiltrate in their study to be 37 days, which is similar to the result of the present study. The mean time to resolution of hypopyon in our study was 13 days. In the study done by Yoon et al.,[28] the mean time to the disappearance of hypopyon was 27 days in the topical amphotericin group, which is higher than the result of the present study. In a study done by Hirose et al.,[17] the time to the disappearance of hypopyon after administration of amphotericin B eye ointment in patient 1 and patient 2 was 10 and 16 days, respectively. The possible reason for the time to the disappearance of hypopyon being less than in the present study may be due to the administration of systemic amphotericin B along with eye ointment.

The mean BCVA at presentation in the present study was 1.64 (logMAR). The final BCVA at 3 months was comparable in the three groups (P-value 0.91). There was a significant improvement in the final BCVA in the three groups. There was an improvement of 48%, 47%, and 44% in group 1, group 2, and group 3, respectively. Yoon et al.[28] reported the final logMAR visual acuity in topical amphotericin B group as 1.36 and the visual acuity gain as 35%, which is lower than that found in the present study. This difference in final visual acuity between different studies may be due to the location of the scar and the difference in initial visual acuity in the study groups. In the present study, corneal ulcers healed as a corneal opacity in all the three groups. The corneal opacity was comparable in the three study groups.

In the present study, superficial and deep neovascularization was present in all patients at 3 months. There was a significant difference in quadrants of deep neovascularization among the three groups. Neovascularization in all four quadrants was more in group 3 patients. A study done by Sharma et al.[30] reported superficial neovascularization in all patients in their study and deep neovascularization in 45% of patients.

In the present study, 17% (n = 8) of cases were nonresponding to treatment and were shifted to topical natamycin eye drops. Out of these nonresponding cases, 13% (n = 6) underwent therapeutic penetrating keratoplasty. In the MUTT 1 trial, Prajna et al.[12] reported that 23% of patients in the voriconazole group and 12% of patients in the natamycin group had corneal perforation or therapeutic keratoplasty. Yoon et al.[28] reported a 17% failure rate in patients treated with topical amphotericin B, which is similar to the result of the present study. In the present study, it was observed that the gel form of L-AMB had a less stinging sensation on putting medication in the eye compared to drop formulations of L-AMB, which led to better patient compliance (P-value 0.01). No drug toxicity has been reported so far and was not found in the present study either.

The study indicates that L-AMB can be considered for the treatment of fungal keratitis when required, though being an expensive drug, cost could be a limitation.

Conclusion

The present study concludes that L-AMB is an effective antifungal agent in promoting the healing of fungal ulcers with notably least vascularization and better tolerance using the gel form.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.