The analysis built on an earlier experiment in which the Candida strains had been exposed daily to hydrogen peroxide for 30 days. Susceptibility to the disinfectant had already been assessed by broth microdilution, after which strains were maintained in broth containing hydrogen peroxide at 0.5×MIC, 1×MIC or 2×MIC. For the work reported here, antifungal susceptibility was evaluated at baseline and after 10, 20 and 30 days of repeated exposure. Testing followed the European Committee on Antimicrobial Susceptibility Testing broth microdilution protocol. Candida strains were cultivated overnight on Sabouraud dextrose agar at 35 °C, and inocula were adjusted to 0.5 McFarland, reported as 1–5 × 10^6 CFU/mL. Each exposed strain was tested in triplicate for each antifungal and incubated for 24 hours at 35 °C before minimum inhibitory concentrations were determined.

For most antifungals, the MIC was defined as the concentration producing 50% growth inhibition; amphotericin B used a greater-than-90% inhibition endpoint. Data were normalized against changes in the corresponding positive control at each time point and then interpreted against EUCAST antifungal susceptibility breakpoints. The researchers regarded a variation as significant only when it exceeded a two-fold dilution relative to the day-zero positive control. This distinction was consequential because the controls themselves displayed some fluctuations, making it difficult to attribute every change in the exposed cultures to hydrogen peroxide. The results therefore separated control-associated variation from subsequent changes confirmed in particular hydrogen-peroxide-exposed strains.

One C. albicans strain, ATCC 10231, showed several shifts after repeated exposure. Following exposure to 0.5×MIC hydrogen peroxide for 20 days, its micafungin MIC was reported as changing from 0.03 to 0.008 mg/L, a numerical decrease that the paper described as an increase in susceptibility in its results section. By day 30, the amphotericin B MIC rose from 0.125 to 0.5 mg/L after 30 days at 1×MIC, while the caspofungin MIC rose from 0.125 to 0.5 mg/L after exposure at 0.5×MIC and 2×MIC. These changes involved both a polyene antifungal and echinocandins, but the investigators classified them as minor fluctuations because the maximum deviation was two two-fold dilutions. The pattern did not establish a consistent mechanism of adaptation across concentrations, agents or sampling times.

Both C. parapsilosis strains showed exposure-associated changes confined to posaconazole susceptibility. C. parapsilosis ATCC 22019 shifted from an MIC of 0.125 to 0.5 mg/L after 30 days at 1×MIC hydrogen peroxide, indicating increased resistance. At 2×MIC, the same strain changed from a susceptible result of 0.003 mg/L to a resistant result of 0.5 mg/L after both 20 and 30 days. The clinical isolate C. parapsilosis CP 009 changed in the opposite direction, moving from a resistant result of 0.125 mg/L to susceptibility at 0.03 mg/L after 30 days at 0.5×MIC. As with the C. albicans findings, the changes did not follow a single disinfectant concentration or antifungal class, and most appeared late in the 30-day period. The authors therefore regarded them as limited susceptibility fluctuations rather than evidence of a coherent resistance mechanism.

The C. auris results require a more qualified interpretation. Major MIC variations were identified for anidulafungin and caspofungin in C. auris DSMZ 21092 at day 30, but the paper also reported comparable variation in positive controls and did not list C. auris among the strains for which subsequent changes were confirmed as associated with hydrogen peroxide exposure. The three C. auris strains had equivalent initial susceptibility values across the antifungals evaluated. Thus, the study did not demonstrate a stable baseline pattern for every measured value, nor did it confirm an exposure-associated susceptibility shift for C. auris. This distinction prevents the preliminary control variation from being presented either as proof that hydrogen peroxide selected for resistance or as evidence of complete absence of variation during the experiment.

Catalase production remained detectable in every exposed strain. Qualitative testing involved mixing five drops of a 3% hydrogen peroxide working solution with a fresh colony; immediate effervescence, representing peroxide decomposition into water and oxygen, was treated as a positive result. The assay was performed in triplicate at each designated time point. Because all strains were catalase-positive from the outset, the experiment established persistence of the detectable phenotype rather than an increase or decrease in enzyme production. The researchers did not directly quantify catalase, measure gene expression or trace the biochemical pathways involved in peroxide detoxification, so the assay cannot establish whether cells altered the amount or activity of the enzyme while retaining a positive tube-test result.

Biofilm formation was assessed with a microtiter-plate assay in which Candida were grown for 48 hours at 35 °C under static conditions. Suspended cells were removed, attached biomass was washed with phosphate-buffered saline, fixed with methanol and stained with 1% crystal violet. The stain was then resolubilized with 33% acetic acid, and biomass was quantified by absorbance at 570 nanometres. Classification used a cutoff calculated from the negative control and its standard deviation, with C. albicans ATCC 90028 included as a positive biofilm-forming control. Six of the seven strains maintained their original biofilm characteristics. The exception was C. parapsilosis CP 009, which shifted from a weak-producer profile at baseline to a non-biofilm-producer classification on day 10 under all three hydrogen peroxide concentrations, then rose above its baseline value on day 20. The investigators interpreted this pattern as a transient, strain-specific fluctuation rather than a general loss of biofilm-forming capacity.

Overall, the experiment did not reveal a consistent adaptive trajectory across the tested Candida strains. Limited MIC shifts occurred in C. albicans and C. parapsilosis, while major variations in the C. auris positive controls prevented those preliminary differences from being confirmed as exposure-associated changes. Catalase remained detectable in every strain, and biofilm formation was stable in all but one C. parapsilosis isolate, whose response fluctuated over time. The results support the limited conclusion that repeated laboratory exposure did not produce broad or uniform changes across the three measured attributes. They do not establish the safety or effectiveness of hydrogen peroxide as a daily hospital disinfectant, nor do they demonstrate changes in fungal virulence, persistence, transmission or human infection. The authors called for periodic environmental monitoring and further studies, including in vivo work, while emphasizing that the clinical relevance of the observed MIC fluctuations remains unresolved.

Subject of Research: Effects of repeated hydrogen peroxide exposure on antifungal susceptibility, catalase production and biofilm formation in Candida strains.

Article Title: Assessment of Candida attributes under in vitro repeated exposure to hydrogen peroxide

Article References: Fernandes, Â. R., Rodrigues, A. G., & Cobrado, L. (2026). Assessment of Candida attributes under in vitro repeated exposure to hydrogen peroxide. International Microbiology. https://doi.org/10.1007/s10123-026-00886-1

Image Credits: AI Generated

DOI: 10.1007/s10123-026-00886-1

Keywords: Candida, hydrogen peroxide, antifungal susceptibility, biofilm formation, catalase, Candida albicans, Candida parapsilosis, Candida auris, disinfection, in vitro study, Assessment, attributes