Comparison of different disinfection protocols against contamination of ceramic surfaces with Klebsiella pneumoniae biofilm

For the purposes of this study we compared ceramic tile disinfection with gaseous ozone (O₃), citric acid (CA), and two marketed professional disinfecting products (DP). Gaseous ozone was produced in the laboratory with a mobile ozone generator (Mozon GPF 8008, Mozon d.o.o., Sisak, Croatia) and used in the concentration of 49.914 mg/m³. Citric acid was purchased from manufacturer (Kemig d.o.o., Zagreb, Croatia) and diluted in the laboratory to the working concentration of 15 %. The first disinfecting product (DP1) contains 1 % BAC as the only active substance, while the second (DP2) contains 4.8 % BAC with 0.1 % 2-phenoxyethanol, 0.098 % ethanol, 0.05 % glycolic acid, and 0.02 % N-(3-aminopropyl)-N-dodecylpropane-1,3-diamine. Both products were obtained from retail and were used in the following working concentrations: 5 % and 20 % for DP1 and 1 % for DP2.

Anti-biofilm efficacy of biocidal active substances alone or in combination was tested against the standard K. pneumoniae ATCC 700603 strain and the clinical K. pneumoniae 14 strain. The standard strain was obtained from the collection of the University of Rijeka Faculty of Medicine’s Department of Microbiology and Parasitology, while the clinical isolate was obtained from a urine sample provided by the Dr. Ivo Pedišić General Hospital in Sisak, Croatia. Both strains were stored in 10 % glycerol broth at -80 °C.

Biofilm was let to form on small ceramic tiles (2.5 × 2.5 cm), which were previously brushed and washed thoroughly and then sterilised in autoclave. The biofilm formation method has been described in detail earlier (8, 32). Briefly, to 250 mL of distilled water we added 5 g of 2 % agar, which was then melted and poured around three ceramic tiles placed in a Petri dish. The upper tile surface was not covered in agar but was layered with diluted overnight bacterial suspension (around 10⁵ CFU/mL) and then incubated in the Petri dish placed on an orbital shaker (Unimax model 1010, Heidolph Scientific Products GmbH, Schwabach, Germany) at 30–50 rpm and 25±2 °C for 24 h.

We employed disinfection protocols divided into four groups as follows (Table 1): disinfection with O₃, CA, DP1, or DP2 alone (group A), combined disinfection with O₃ followed by CA, DP1, or DP2 (group B), combined disinfection with CA, DP1, or DP2, followed by O₃ (group C), and combined treatment with CA, O₃, and DP1 (group D). All protocols involved one-hour exposure to O in the concentration of 49.914 mg/m³. All experiments were done in triplicate. Controls (untreated tiles) were provided for all disinfection protocols.

Petri dishes with ceramic tiles with formed K. pneumoniae biofilm were placed in a sealed experimental chamber (V=0.125 L) and O₃ inserted into the chamber with a silicon tube until it reached the concentration of 49.914 mg/m³. Exposure lasted 1 h, during which time we monitored the temperature (23.4 °C), relative humidity (56 %), and O₃ concentration with ozone detector Keernuo GT-901 (Keernuo, Shenzhen, China) and Auriol 4-LD5531 weather station (OWIM GmbH, Neckarsulm, Germany). After one hour, the tiles were removed from the agar with sterile pincers, rinsed with 10 mL saline, placed in a Falcon tube (one tile per tube) containing 10 mL sterile saline, and sonicated in an ultrasound bath (BactoSonic, Bandelin, Berlin, Germany) at 40 kHz for 1 min. Falcon tubes with tiles were then vortexed to enhance biofilm detachment from the tiles.

CA was poured over ceramic tiles with formed K. pneumoniae biofilm previously removed from agar, washed with sterile saline, and dried in a laminar flow chamber for 1 min. Exposure time to CA was 10 min. After that, the tiles were washed with sterile saline, transferred into a Falcon tube, and prepared for the determination of culturable bacterial count.

DP1 (in either 5 % or 20 % working concentration) or DP2 was poured over the ceramic tiles with formed biofilm and left for 10 min. After exposure, each tile was transferred into a new Petri dish containing a 10 % sodium thiosulphate solution (Kemika d.o.o., Zagreb, Croatia) for 10 min to neutralise BAC. Culturable bacterial count was determined immediately as described below.

K. pneumoniae biofilm on ceramic tiles was first exposed to O₃ as previously described and then to either CA, DP1, or DP2 as follows: O₃⁺CA; O₃⁺5 % DP1; O₃⁺20 % DP1, and O₃⁺1 % DP2.

Ceramic tiles with K. pneumoniae biofilm were first treated with either CA, DP1, or DP2 as described above, and then with O₃ as follows: CA + O₃; 5 % DP1 + O₃; 20 % DP1 + O₃, and 1 % DP2 + O. After the pre-treatment with CA, DP 1 and DP2, the tiles were neutralised, rinsed with sterile saline, dried off in laminar flow chamber, and then exposed to O₃ in a sealed chamber for 1 h.

K. pneumoniae biofilm on ceramic tiles was first treated with 15 % CA for 10 min as described above, then with O₃ for 1 h, and finally with 20 % DP1 for 10 min.

After all disinfection protocols, culturable bacterial count was determined using ten-fold serial dilutions prepared and inoculated on Muller Hinton agar. After incubation at 35±2 °C for 24–48 h, culturable bacteria were counted and are expressed as CFU/cm².

For imaging, the ceramic tiles with representative strain K. pneumoniae ATCC 700603 biofilm were rinsed with sterile saline to remove excess material, fixated in a dry heat steriliser (ST-01/02, Instrumentaria, Zagreb, Croatia) at 80 °C for 30 min, and stained with 0.1 % crystal violet (CV) dye for 30 min. Images were taken with a DSX 1000 digital microscope (Olympus, Tokyo, Japan) at 20× magnification and the stained tiles are presented as 3D images.

For statistical analysis we used the TIBCO Statistica 14.0.1 (TIBCO Software Inc., Palo Alto, CA, USA). The normality of data distribution was tested with the Shapiro-Wilk test. Statistical differences in bacterial counts between control and treated samples were tested using the non-parametric Wilcoxon signed-rank test for paired samples. Differences in bacterial counts between treatments were tested with the non-parametric Mann-Whitney U test, and the average of rank was determined with Friedman’s ANOVA and Kendall’s coefficient of concordance.