Article Text

Download PDFPDF

Activity of antibiotics against Fusarium and Aspergillus
  1. S Day1,
  2. P Lalitha2,
  3. S Haug1,
  4. A W Fothergill3,
  5. V Cevallos1,
  6. R Vijayakumar2,
  7. N V Prajna2,
  8. N R Acharya1,4,
  9. S D McLeod1,4,
  10. T M Lietman1,4
  1. 1
    Department of Ophthalmology, University of California San Francisco, San Francisco, California, USA
  2. 2
    Aravind Eye Hospital, Madurai, India
  3. 3
    University of Texas Health Sciences Center at San Antonio, Fungal Testing Laboratory, Department of Pathology, San Antonio, Texas, USA
  4. 4
    Francis I Proctor Foundation, University of California San Francisco, San Francisco, California, USA
  1. Dr T M Lietman, Francis I Proctor Foundation, Room S309, 513 Parnassus Avenue, San Francisco, CA 94143-0944, USA; tom.lietman{at}


Background/aims: To study the susceptibility of Fusarium and Aspergillus isolated from keratitis to amoxicillin, cefazolin, chloramphenicol, moxifloxacin, tobramycin and benzalkonium chloride (BAK).

Methods: 10 isolates of Fusarium and 10 isolates of Aspergillus from cases of fungal keratitis at Aravind Eye Hospital in South India were tested using microbroth dilution for susceptibility to amoxicillin, cefazolin, chloramphenicol, moxifloxacin, tobramycin and BAK. The minimum inhibitory concentration (MIC) median and 90th percentile were determined.

Results: BAK had the lowest MIC for both Fusarium and Aspergillus. Chloramphenicol had activity against both Fusarium and Aspergillus, while moxifloxacin and tobramycin had activity against Fusarium but not Aspergillus.

Conclusions: The susceptibility of Fusarium to tobramycin, moxifloxacin, chloramphenicol and BAK and of Aspergillus to chloramphenicol and BAK may explain anecdotal reports of fungal ulcers that improved with antibiotic treatment alone. While some of the MICs of antibiotics and BAK are lower than the typically prescribed concentrations, they are not in the range of antifungal agents such as voriconazole, natamycin and amphotericin B. Antibiotics may, however, have a modest effect on Fusarium and Aspergillus when used as initial treatment prior to identification of the pathological organism.

Statistics from

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.

Fungal keratitis is a leading cause of ocular morbidity throughout the world, particularly in warm climates and developing countries. It is estimated that corneal ulceration and ocular trauma result in 1.5 to 2 million new cases of corneal blindness annually.1 In certain parts of the world, such as South India, nearly half of these corneal ulcers may be secondary to fungal infection.2 Epidemiological studies in India have identified filamentous fungi, Fusarium and Aspergillus, as the two most common causes of fungal keratitis.2 3 While the incidence of fungal ulcers in the USA and Europe has historically been much lower, the 2006 outbreak of Fusarium keratitis related to contact lens wear and the use of specific contact lens solutions has highlighted the challenges of effectively treating fungal keratitis.

Traditionally, fungal keratitis is treated medically with antifungal agents such as natamycin, amphotericin B and voriconazole.4 5 With failure of medical therapy, surgical interventions such as penetrating keratoplasty may be required. In the last few years, however, there have been several anecdotal cases of fungal keratitis that were initially treated with antibiotics and improved with antibiotic therapy alone. Eleven cases of the 66 patients diagnosed as having Fusarium keratitis during the outbreak in Singapore improved with antibiotic treatment alone.6 An additional report of five cases of culture-proven fungal keratitis in Florida showed significant improvement with topical fluoroquinolone therapy as initial treatment, although one case did eventually require topical natamycin treatment.7 Furthermore, a randomised trial of corneal ulcer prevention in south India where 374 patients with corneal abrasions were randomised to either 1% chloramphenicol plus 1% clotrimazole or 1% chloramphenicol plus placebo found equivalent reduction in the development of both fungal and bacterial keratitis compared with historical controls, suggesting that antibiotic prophylaxis of corneal abrasions alone may be sufficient in preventing both bacterial and fungal infection.8

In vitro studies have shown activity of tobramycin against Fusarium oxysporum, possible activity of fourth-generation fluoroquinolones against Fusarium and Candida species, and activity of amoxicillin and erythromycin against the pectolytic and cellulolytic enzymes in Fusarium oxysporum.911 Chodosh et al added tobramycin solution to brain heart infusion agar prior to inoculation with one strain of Fusarium oxysporum isolated from a case of fungal keratitis and found that tobramycin completely inhibited growth at 450 µg/ml.9 Ozdek et al found inhibition of Candida species by moxifloxacin and gatifloxacin, and Alfonso and Miller also reported in an ARVO abstract that commercial preparations of both gatifloxacin and moxifloxacin reduced colony counts of Fusarium oxysporum and Candida albicans.10 12 However, Chai et al found that moxifloxacin only inhibited one of 10 Fusarium corneal isolates.13 These case series and in vitro studies have raised the question of whether or not topical antibiotics alone may offer clinically significant antifungal activity in early fungal keratitis. In this study, we investigate the efficacy of antibiotics (amoxicillin, cefazolin, chloramphenicol, moxifloxacin and tobramycin) and a biocidal agent (benzalkonium chloride) against Fusarium and Aspergillus species isolated from cases of fungal keratitis at a tertiary eye centre in South India.


This prospective, non-randomised interventional study was performed in the Microbiology Department of the Aravind Eye Hospital, Madurai, India after approval by the institutional review boards of the Aravind Medical Research Foundation and the University of California San Francisco.

As previously described by Lalitha et al,14 each patient with keratitis underwent a detailed history and clinical examination consistent with the standard of care at the Aravind Eye Hospital. Scraping of the ulcer was performed with a Kimura spatula and evaluated with Gram stain and 10% potassium hydroxide mount; then cultures were plated onto blood agar and potato dextrose agar. Fungal cultures were permitted to grow for at least 7 days at room temperature, and fungal isolates were identified to the genus level, with Aspergillus isolates identified to the species level.

Ten fungal cultures each were randomly selected from consecutive cases of culture-proven Fusarium keratitis and culture-proven Aspergillus flavus keratitis between 1 January 2006 and 31 December 2006. Of the 10 Fusarium isolates, all of the samples were processed. One of the Aspergillus isolates was contaminated by bacteria and discarded. The remaining 10 Fusarium isolates and nine Aspergillus isolates were tested for susceptibility to five antibiotic agents (amoxicillin, cefazolin, chloramphenicol, moxifloxacin and tobramycin) and one biocidal agent (benzalkonium chloride). For each antibiotic and biocidal agent, a range of concentrations from 1 to 4000 μg/ml were tested. Prior to susceptibility testing, each antibiotic agent was tested against a standard Staphylococcus aureus strain to ensure that the MIC was within the published range. One randomly selected Fusarium strain and one randomly selected Aspergillus strain were tested against amphotericin B to check for antifungal resistance.

Susceptibility was determined according to methods outlined in Clinical and Laboratory Standards Institute document M38-A.15 The inoculum was prepared by overlaying mature slants with sterile distilled water and gently scraping the surface with a plastic pipette tip. The suspension was permitted to sit for 5 min to allow large particles to settle out. The inoculum was then adjusted with Mueller–Hinton broth to a 0.5 McFarland standard, providing an inoculum concentration of 0.4 to 5×104 CFU/ml, which was verified by colony count. Mueller–Hinton broth was used for the inoculum, antibiotics and benzalkonium chloride instead of RPMI-1640, since it is a standard antibiotic test medium. Broth microdilutions were created by adding 0.1 ml of 2× concentrated drug to 0.1 ml of inoculum. Tests were incubated at 35°C for 48 h. Tests on each fungal culture were repeated on a different date by a masked operator to reduce the effect of random variation in test results.

Since there is no standardised endpoint for the minimum inhibitory concentration (MIC) for antibiotic and biocidal agents against fungus, the MIC was defined as the lowest concentration of an antibiotic or biocidal agent that substantially inhibited growth (approximately 50% of the drug-free growth control) of the organism as detected visually at 48 h. Given that the tests for each fungal strain were repeated twice, the geometric mean of the MIC for each antibiotic or biocide was obtained. The MIC median (MIC50) and 90th percentile (MIC90) were determined for Fusarium and Aspergillus species (PERCENTILE function in Microsoft Excel; Microsoft, Redmond, Washington).

After the MIC was determined at 48 h, the minimum lethal concentration (MLC) was determined by plating 100 μl of the drug-free control tube, the MIC tube concentration and each concentration above the MIC tube to a potato dextrose agar quadrant. Plates were incubated at 35°C and read at the first 24 h interval where growth was detected in the control tube quadrant. The MLC was determined to be the lowest concentration which exhibited five colonies or less, equating to a 95% killing.


The MIC50 and MIC90 for each antibiotic and biocidal agent against Fusarium and Aspergillus are displayed in table 1. BAK had the lowest MIC for both Fusarium and Aspergillus. Chloramphenicol had activity against both Fusarium and Aspergillus at relatively high concentrations, while moxifloxacin and tobramycin had activity against Fusarium but not Aspergillus. For Fusarium, moxifloxacin had an MIC50 = 1000 µg/ml and MIC90 = 2000 µg/ml, while tobramycin had an MIC50 = 500 µg/ml and MIC90 = 700 µg/ml. Neither amoxicillin nor cefazolin had any detectable activity against Fusarium or Aspergillus at concentrations of 4000 μg/ml and below. One hundred per cent of masked replications of MICs were within one fourfold dilution.

Table 1 Minimum inhibitory concentration median (MIC50) and 90th percentile (MIC90) results against Aspergillus and Fusarium

For agent–organism combinations which had an MIC less than 4000 μg/ml, the MLC50 and MLC90 were also calculated, and these are displayed in table 2. For Fusarium, the MLC50 and MLC90 of tobramycin, moxifloxacin and chloramphenicol were all >4000 µg/ml, and for Aspergillus, the MLC50 and MLC90 of chloramphenicol were also >4000 μg/ml. The MLC50 and MLC90 for BAK against Fusarium and Aspergillus were very close to the MIC50 and MIC90, indicating that BAK had a fungicidal rather than static effect against both Fusarium and Aspergillus.

Table 2 Minimum lethal concentration median (MLC50) and 90th percentile (MIC90) results against Aspergillus and Fusarium

When considering the clinical significance of the MIC50 and MIC90 for each of these antibiotic and biocidal agents, it is important to recognise that their prescription doses can vary widely for the same medication and between medications. For example, tobramycin is available at a 0.3% concentration (equivalent to 3000 µg/ml) in eye-drop form but is also given in a fortified concentration of 14 mg/ml (equivalent to 14 000 µg/ml). Commonly prescribed dosages of these medications are given in table 3.

Table 3 Commonly prescribed dosages of antibiotics and benzalkonium chloride


Of the antibiotic agents, chloramphenicol had an effect against both Fusarium and Aspergillus, while tobramycin and moxifloxacin had an effect only against Fusarium. Though a few of the antibiotics tested in this study showed efficacy against Fusarium and Aspergillus, these MICs are much higher than comparable values for antifungal agents such as amphotericin B, natamycin and voriconazole. For example, a recent report from our group that examined similar isolates found that the MIC50 for amphotericin B against Fusarium was 2 µg/ml, and the MIC50 for voriconazole against Fusarium was 0.25 µg/ml.14 In the current study, quality control tests with amphotericin B found an MIC of 0.5 µg/ml against two randomly selected Fusarium and Aspergillus strains. Therefore, antifungal agents remain imperative for first-line treatment of fungal keratitis. The results of this study, however, may explain anecdotal reports of fungal keratitis that improved after initial treatment with antibiotics. In particular, tobramycin is often given in a fortified concentration of 14 mg/ml (14 000 µg/ml), which is 30-fold higher than the MIC50 for tobramycin against Fusarium (500 µg/ml). This MIC value is also consistent with Chodosh et al’s report that tobramycin completely inhibited growth of one strain of Fusarium oxysporum at 450 µg/ml.9 Even considering the effect of tear film dilution, tobramycin would likely have some measurable activity against Fusarium keratitis. Of note, any synergy offered by an antibiotic in conjunction with an antifungal agent would not have been identified in these experiments.

The molecular basis for the efficacy of antibiotics against fungus is not completely understood. Fluoroquinolones like moxifloxacin and gatifloxacin work against type II topoisomerase DNA gyrase and topoisomerase IV, and so the presence of DNA topoisomerases I and II in yeast cells has been proposed as a mechanism for these agents’ antifungal activity.7 Aminoglycosides are thought to make fungal cell walls and membranes more permeable to ionic shifts, perhaps explaining the antifungal effect of tobramycin in this study.16 Chloramphenicol has been shown to have some in vitro effect against the polygalacturonase and pectinmethylgalacturonase enzymes of Fusarium oxysporum.11

The mechanisms of action of biocidal agents like benzalkonium chloride are arguably less well elucidated than those of antibiotic agents. Benzalkonium chloride is one of the quarternary ammonium compounds (QACs), membrane-active agents which primarily target the cytoplasmic (inner) membrane of bacteria of the plasma membrane of yeasts.17 The following sequence of events has been proposed in micro-organisms exposed to QACs: adsorption and penetration of the agent into the cell wall, reaction with the cytoplasmic membrane followed by membrane disorganisation, leakage of low-molecular-weight material, degradation of proteins and nucleic acids, and wall lysis caused by autolytic enzymes.17 In vitro testing of the fungicidal activity of benzalkonium chloride against Aspergillus fumigatus showed biocidal activity in less than 5 min of contact time, defined as a 104 or more reduction in viability of A fumigatus strains.18

The finding of relatively low MIC and MLC values for BAK against both Fusarium and Aspergillus raises the question of whether the addition of benzalkonium chloride to antifungal agents for treatment or prophylaxis may be beneficial for its fungicidal activity in addition to its preservative qualities. Ozdek et al found that Zymar (0.3% gatifloxacin + 0.005% BAK, Allergan, Irvine, California) had greater in vitro activity than Vigamox (0.5% moxifloxacin, Alcon Laboratories, Fort Worth, Texas) against Candida isolated from ocular infections, and they postulated that this effect was due to the presence of BAK in Zymar.10 Similarly, Alfonso and Miller found that Zymar (0.3% gatifloxacin + 0.005% BAK, Allergan) reduced the colony count of both Fusarium oxysporum and Candida albicans by 99.9% at 1, 4 and 24 h, while Vigamox (0.5% moxifloxacin, Alcon Laboratories) had similar results for Fusarium oxysporum but not Candida albicans.12 However, it should be noted that the MIC50 of BAK against filamentous fungus is only one to sixfold lower than the concentration typically found in ophthalmic preparations, and the use of higher concentrations or more frequent dosing of BAK is limited by toxicity.

Although several authors have investigated the in vitro effect of fluoroquinolones and tobramycin on Fusarium, this is the first study using the CLSI broth microdilution antifungal susceptibility testing to determine the MICs of antibiotics versus Fusarium and Aspergillus.15 While in vitro testing must eventually be studied in conjunction with in vivo clinical results, the correlation of in vitro fungal drug-susceptibility testing with in vivo therapy and clinical outcomes has gained credibility in the non-ophthalmic literature.19 Similar to susceptibility testing for bacterial disease, approximately 90% of infections with susceptible isolates will respond to therapy, while only 60% of infections due to resistant isolates respond to therapy.19 Further studies would need to be conducted to assess whether tobramycin, moxifloxacin, chloramphenicol and BAK might have a clinically relevant additive, antagonistic or synergistic in vivo effect with antifungal agents.



  • Competing interests: None.

  • Funding: That Man May See, NEI U10-EY015114 and K23EY017897, Research to Prevent Blindness. The funding organisations had no role in the design or conduct of this research.

  • Ethics approval: Ethics approval was provided by the institutional review boards of the Aravind Medical Research Foundation and the University of California San Francisco.