Introduction

Fungal keratitis is a serious ocular infection associated with severe visual loss. It is typically a slowly progressing disease; however, it has poor outcomes owing to delay in diagnosis and difficulty in treatment [1]. For a definitive diagnosis, corneal scrapings from the fungal lesion should be obtained. However, the culture is often negative for fungus owing to the deep location of the lesion, and it takes up to three weeks to culture and identify the organism. Medications available for ocular therapy, including fungal eye drops and ointments, are limited because these medications fail to penetrate deep into the cornea. It is reported that intrastromal injection of voriconazole was effective; however, the dose concentration and dosage frequency still need to be considered [2, 3].

Fungal keratitis is sometimes caused by rapidly proliferating fungi, resulting in tissue necrosis and corneal perforation. In addition, if opacity persists in the center of the cornea without perforation or when the fungal keratitis has low response to treatment, therapeutic corneal transplantation is sometimes performed [2, 3] and long-term treatment is required.

We reported that a case of fungal keratitis with large retrocorneal plaques was successfully treated surgically by retrocorneal plaque aspiration and anterior chamber irrigation [4]. The present study aimed to retrospectively investigate the safety of retrocorneal plaque aspiration for the diagnosis and treatment of fungal keratitis.

Patients and methods

Subjects

This retrospective study was conducted at the Kyoto Prefectural University of Medicine, Kyoto, Japan, and the Baptist Eye Institute, Kyoto, Japan. The study was approved by the institutional review board (#ERB-C-1006) and was conducted in adherence with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all subjects.

The study included five eyes of five patients who visited the Kyoto Prefectural University of Medicine Hospital and received retrocorneal plaque aspiration for the diagnosis and treatment of fungal keratitis. Table 1 shows the subjects’ demographic data.

Table 1 Demographic data
Full size table

The mean age (mean ± standard deviation [SD]) of the patients was 68.4 ± 13.0 years (range: 45–81 years). Corneal scrapings from all patients were examined at the initial visit. Ocular examinations, including best-corrected visual acuity (BCVA) and lens status (i.e., phakic/pseudophakic), previous ocular history, and systemic diseases, were recorded. An anterior segment photograph was obtained before surgical retrocorneal plaque aspiration. Anterior segment optical coherence tomography (AS-OCT) was taken just before retrocorneal plaque aspiration.

Surgical procedure

A two-sided port at 10 o’clock and 2 o’clock was set up. The retrocorneal plaques were aspirated using a tuberculin syringe with a 27-gauge blunt needle or directly grasped using disposable forceps and removed from the eye while maintaining the anterior chamber depth. The retrocorneal plaques and anterior hypopyon left in the anterior chamber were irrigated, aspirated, and sufficiently washed out using bimanual irrigation and aspiration (I/A). The depth of the anterior chamber was carefully maintained by the minimum height of the irrigation bottle. The irrigation bottle did not contain an antifungal drip to avoid endothelial cell damage; instead, intracameral or intrastromal antifungal injections were administered as needed (i.e., intracameral voriconazole: 20–40 μg/ml).

Postoperative medication

Patients were required to continue the intensive fungal treatment prescribed preoperatively after retrocorneal plaque aspiration. In patients with therapeutic keratoplasty, the postkeratoplasty medication was routinely administered as in our previous report [5] (with minor modifications), including 0.3% gatifloxacin eye drops and 0.1% betamethasone eye drops administered four times daily and the initiation of a systemic dose of 4 mg/day of betamethasone. Systemic betamethasone was tapered to 1 mg/day three days postkeratoplasty and discontinued seven days postoperatively.

Examination

Samples were collected and cultured on Sabouraud agar, specific to detect fungi. Fungal detection on the smeared samples was performed using Fungiflora Y® staining (Trust Medical Co., Ltd.), a ready-to-use staining kit for fungi [6]. The corneal button after therapeutic keratoplasty was examined for the presence or absence of fungi.

Results

Three phakic and two pseudophakic eyes from five fungal keratitis patients were included. Table 1 summarizes details of the study subjects. Of the five patients, three (60%) were diagnosed with fungal keratitis at the initial visit according to the culture results taken from the corneal scrapings. The other patients were not definitely diagnosed with fungal infection because they developed a focal lesion into the deep cornea (Table 2). Because fungal keratitis was suspected from the clinical findings, all patients were administered intensive fungal treatment, including 0.1% miconazole eye drops and 1% natamycin ointment frequently throughout the day. However, the response to the intensive fungal treatment was poor and large retrocorneal plaques with hypopyon still remained; three patients had hypopyon height over 2 mm (Fig. 1). AS-OCT was taken in only three of five patients (Cases 2, 3, and 5) just before aspiration; it revealed an unclear boundary line between Descemet’s membrane and corneal plaques, indicating the disruption of the former (Fig. 2). In Cases 1 and 4 a slit-lamp examination found a vague Descemet’s membrane line just before the aspiration indicating a possible disruption. Therefore, we decided to perform retrocorneal plaque aspiration for a precise diagnosis and appropriate treatment. Anterior chamber aspiration improved the accuracy of diagnosis to five out of five patients (100%), including three patients that were positive for yeast-type fungi, one case that was positive for filamentous fungi, and one case that was positive to Fungiflora Y® staining. Two patients with diabetes mellitus and ocular problems who needed topical steroid treatment required therapeutic keratoplasty, and on top of intensive fungal treatment underwent retrocorneal plaques aspiration, as fungi, stained black by Grocott’s methenamine silver staining, was observed in the corneal button (Fig. 3). Severe complications from retrocorneal plaque aspiration included endopthalmitis and severe endothelial cell damage, but the development of cataracts was not observed. All the eyes finally became stabilized (Table 3).

Table 2 Summary of retrocorneal plaques aspiration
Full size table
Fig. 1
figure 1

Preoperative anterior segment photographs of five cases. All of the cases had large plaques on the posterior cornea

Full size image
Fig. 2
figure 2

Preoperative anterior segment optical coherence tomography (AS-OCT) in Cases 2, 3, and 5

Full size image
Fig. 3
figure 3

Grocott’s methenamine silver staining after therapeutic keratoplasty in Cases 4 and 5. Fungi was stained in black. Scale bar indicates 20 μm

Full size image
Table 3 Final outcome of retrocorneal plaques aspiration
Full size table

Case 1

A 68-year-old man struck by a foreign body OS. His ocular infection did not improve after the administration of antibacterial eye drops, and he was referred to the Kyoto Prefectural University of Medicine Hospital, Kyoto, Japan. Neither bacteria nor fungi were detected in corneal scrapings. The following treatment for fungal keratitis was administered: 0.1% miconazole eye drops six times daily, 1% natamycin ointment six times daily, 1.5% levofloxacin eye drops four times daily, and oral voriconazole 400 mg daily. At the initial visit, his visual acuity was hand motion and his intraocular pressure was 27 mmHg. Notwithstanding intensive fungal treatment, corneal infiltration failed to clear up and worsened, and Descemet’s membrane line looked unclear in slit-lamp examination, interpreted as a disruption of Descemet’s membrane, retrocorneal plaque aspiration was performed for large retrocorneal plaques on the 15th day following administration of intensive treatment. Filamentous fungus positive to Fungiflora Y® staining was identified from the samples on the retrocorneal plaques; however, no species were detected from the cultured samples. On the next day, the hypopyon height decreased but remained, and the keratitis became one size smaller. One week after the aspiration treatment, the hypopyon disappeared, and the infectious keratitis gradually improved. One month later, no hypopyon was found, and the fungal keratitis became convalescent, resulting in corneal scaring.

Case 3

A 45-year-old woman noticed visual loss and ocular pain OD. On examination Candida albicans was detected from corneal scrapings. Although intensive fungal treatment, including 0.1% miconazole eye drops every hour and 1% natamycin ointment five times daily combined with 0.5% moxifloxacin eye drops eight times daily, was administered, her keratitis did not improve, and she was referred to the Kyoto Prefectural University of Medicine Hospital five months after the administration of fungal treatment. Her visual acuity was hand motion, and her intraocular pressure was 15 mmHg. AS-OCT just before retrocorneal plaque aspiration revealed stromal infiltration and retrocorneal plaques with indistinct boundaries. Retrocorneal plaque aspiration was performed two days in a row owing to the mild recurrence of retrocorneal plaque. The obtained plaques were grabbed by forceps and were found to be positive for Candida albicans. At the time of aspiration, intracameral voriconazole (20 μg/ml on the first aspiration and 40 μg/ml on the second aspiration) injection was performed. After the aspiration, the fungal focus gradually decreased and finally became corneal scaring.

Discussion

We demonstrated the safety of retrocorneal plaque aspiration in removing retrocorneal plaques and hypopyon in patients with fungal keratitis developing into the deep cornea. In this study, the diagnosis accuracy improved from three (60%) to five out of five patients (100%), and three out of five patients developed stable corneal scarring without the need for therapeutic keratoplasty. There were no adverse events, including endopthalmitis.

The diagnostic accuracy of fungal keratitis is relatively low owing to the limited sensitivity of the culture, the need for prolonged culture time, and the limited number of samples from the cornea. For deep stromal fungal keratitis penetrating into the anterior chamber through Descemet’s membrane, the severity of fungal keratitis was evaluated using AS-OCT; this enables providing detailed findings on the posterior cornea [7, 8]. As Takezawa et al. demonstrate, an unclear boundary exists between the cornea and the plaque in patients with a break in Descemet’s membrane [7]. Also, AS-OCT findings suggest that Descemet’s membrane in three patients who were examined by AS-OCT seemed to be disrupted, possibly leading to higher accuracy of final diagnosis by the direct capture of retrocorneal plaques and suggesting that a sign of Descemet’s membrane disruption may be useful to determine the indication of retrocorneal plaque aspiration.

In cases in which the infection progresses through Descemet’s membrane into the anterior chamber, as occurs with retrocorneal plaques and hypopyon, antifungal medications often fail to eradicate the active infection. We believe that, compared with anterior chamber aspiration, this approach may improve the diagnostic accuracy and reduce secondary inflammation through the capture of plaque itself. It is reported that the outcomes of early deep anterior lamellar keratoplasty (DALK) for fungal keratitis that is poorly responsive to medical treatment could represent a beneficial approach to effectively eradicate active corneal infections [9]. In addition, studies suggest that it is important to protect a clear optical zone before it is affected by fungal keratitis. The concept of our approach seems to be similar; the earlier removal of fungi itself entwined with fibrins minimizes the secondary inflammation and corneal scarring, thereby preserving the optical zone.

Fungi can colonize the deep stroma and penetrate an intact Descemet’s membrane into the anterior chamber, thereby exposing the eye globe to high risk and finally leading to endophthalmitis. Patients ultimately need to undergo therapeutic keratoplasty for the impending corneal perforation. The outcome of keratoplasty for the treatment of infectious keratitis is poor, even if successful keratoplasty was performed, because of frequent allograft rejection and postoperative complications [10, 11]. It is reported that fungal keratitis was more likely to require therapeutic keratoplasty compared with bacterial keratitis [12]. Lalitha et al. [13] used multivariate analysis and found that the presence of a hypopyon was the most significant risk factor for primary treatment failure, and the presence of a deep infiltrate and hypopyon height over 2 mm were significant predictors for corneal perforation in fungal keratitis. In fact, our cases had deep fungal keratitis with hypopyon and were poorly responsive to primary fungal treatment. Nevertheless, three of the five patients did not require therapeutic keratoplasty. This suggests that retrocorneal plaque aspiration could present a valid approach to suppress active fungal keratitis and prevent the development of corneal perforation.

It should be noted that this study has several limitations. First, two of the five patients had a recurrence following retrocorneal plaque aspiration because they had poor medical histories; this can be associated with an immunodeficient ocular surface due to local steroid treatment or systemic problems due to diabetes mellitus. It is important to pay attention to the patient background when there is a poor response to retrocorneal plaque aspiration. Second, the anterior chamber approach is an intraocular surgery that is associated with a risk of endophthalmitis, especially in cases with pseudophakia. In cases with phakia, cataract development could occur. Other concerning issues are endothelial cell damage and progression to perforation by this approach. However, the improvement of precise diagnosis provides us with greater benefits in cases with poor response to intensive treatment. Another limitation is that this is a retrospective study with a small case series, and AS-OCT was taken only in three out of five cases to evaluate the presence or absence of Descemet’s membrane disruption. Therefore, a prospective randomized controlled study should be performed to confirm the indication and the efficacy of retrocorneal plaque aspiration.

In conclusion, in advanced cases in which retrocorneal plaques are suspected to be penetrating the anterior chamber, retrocorneal plaque aspiration may be positively selected for both diagnosis and treatment.