Methods

We completed a comparative experimental design to evaluate different sampling and detection methods for fungal contamination on common healthcare surface materials, including aluminum, formica, linen, and high efficiency particulate air (HEPA) material, each measuring 10 × 10 cm. We uniformly inoculated approximately 10^4 colony-forming units (CFUs) of Aspergillus fumigatus (ATTC 204305) or Candida parapsilosis (ATCC 22019)) on 10 × 10 cm of each type of surface. Inocula were prepared using standard curves with OD600 and CFU/mL metrics. Standard spore collection was completed for A. fumigatus and liquid culture for C. parapsilosis, both using Sabouraud dextrose (SD) media. OD values were used to generate desired CFU/mL values, and serial dilutions were used to confirm CFU/mL. Following inoculation, the surfaces were allowed to air-dry for 10 minutes at ambient temperature to simulate real-world conditions. Each surface was evaluated 240 times (120 for A. fumigatus and 120 for C. parapsilosis).

Three primary sampling techniques were employed: 1) polyurethane sponge sticks, 2) nylon flocked swabs (both pre-moistened with a neutralizing buffer), and 3) Replicate Organism Detection And Counting (RODAC) plates containing SD agar. Sponges were positioned on the sampling surface ensuring full contact of one side of the sponge and then used to scrub the entire surface. Sponges were then flipped to ensure the opposite side made complete contact, and the entire surface was scrubbed again. Swabs were positioned similarly and scrubbed the entire surface while constantly rotating the swab to ensure usage of the entire swab surface. Replicate Organism Detection And Counting (RODAC) plates were directly placed onto the surface and carefully removed to avoid smearing. The recovery process for sponges involved the stomacher technique: sponges and 45 mL of phosphate-buffered saline with 0.1% Tween 20 were combined and stomached (Seward, Bohemia, NY, USA) at 260 RPM for 60 seconds. Resulting homogenates were centrifuged at 3100 rpm for 15 minutes, and all but 5 mL of the supernatant was decanted. ⁵ Swabs were processed by vortexing in 1 mL of phosphate buffered saline (PBS). The eluents obtained from both sponges and swabs were then subjected to two detection methods: culture-based quantification and quantitative polymerase chain reaction (qPCR) using the FungiQuant primers for the fungal 18S rRNA gene in the QIAquant 96-5plex real-time PCR thermal cycler (Qiagen, Germany). ^{Reference Liu, Kachur and Dwan6,Reference Wagner, Springer, Pires and Keller7} Select PCR products from each experiment were sent for sequencing to rule out contamination. A standard curve was generated for both species between qPCR cycle thresholds and culture-based quantifications. Measured qPCR cycle thresholds were then used to calculate expected culture-based quantification to control for the increased sensitivity of qPCR compared to culture. The primary metric for comparison was the percent recovery of organisms, calculated as the total recovered sample area CFU relative to the known inoculum CFU. CFU were calculated from RODAC plates by counting plate CFUs, converting to CFU/cm², and extrapolating that value to the size of the surface (100 cm²).

On each experiment day, all possible combinations of species, surface, sampling methodology, and detection methodology were run to control temporal variation. The Wilcoxon signed rank test was used to compare all CFU measurements, and t tests were used for qPCR cycle threshold comparisons. P < .05 was considered to be significant. All statistical tests were two-tailed and performed using SAS, version 9.4M7 (SAS Institute Inc).