Clinical isolates from patients were collected and culture-confirmed as either Staphylococcus aureus or Pseudomonas aeruginosa (PA) at our institution. Six methicillin-sensitive (MSSA), six methicillin-resistant (MRSA) isolates, and six PA isolates were obtained from common ocular sources and were observed to have different virulence factors ( Table 1 ) and minimum inhibitory concentrations (Supplemental Tables 1 and 2). Additionally, two reference strains were also included: Staphylococcus aureus ATCC 29 213 and Pseudomonas aeruginosa ATCC 27 583. All isolates were cultured on 5% sheep blood agar and incubated at 37 °C. Following two growth cycles, isolates were transferred to tryptic soy broth (TSB) slants (Remel, Lenexa, Kansas, United States) and stored at room temperature. A combination of standard microbiological techniques and DNA microarrays was used to characterize the molecular profiles of the isolates. Phenotypic identification was confirmed using the VITEK 2 automated system (bioMérieux, Durham, NC, USA).

Ten commercially available ATs were selected to represent a diverse range of formulations and were categorized into three groups: preservative-containing MDBs (n = 5), preservative-free MDBs (n = 3), and preservative-free SDVs (n = 2). Among the preservative-containing products, six different preservative agents were represented, including benzalkonium chloride (BAK), EDTA, hexadecenoic acid, polyquaternium-1, stabilized oxychloro complex, and sodium perborate ( Figure 1 ). A summary of all AT products and their formulations is provided in Table 2 .

Overview of methods.

All ATs were stored at room temperature and protected from light, heat, and humidity. To minimize contamination, opened bottles were kept in separate containers and handled using personal protective equipment. Expiration dates were recorded, and all bottles were disposed of upon reaching expiration or if open for more than 3 months.

On the day of experimentation, isolates were suspended in TSB and standardized to a 0.5 McFarland turbidity standard, corresponding to 1.5 × 10⁸ colony-forming units per milliliter (CFU/mL). The inoculum was then diluted to a concentration of 1.5 × 10 ⁷ CFU/mL. All 18 clinical strains and the two reference strains were tested.

For each isolate, the following five experimental conditions were tested by adding 200 µL of the respective media (ATs, saline, or TSB) and 20 µL of organism: (1) ATs + saline (no organism, to confirm sterility the ATs), (2) ATs + organism (test condition), (3) saline only (no organism, negative control), (4) saline + organism (negative growth control of the organism), and (5) TSB + organism (positive growth control of the organism). The TSB provides nutrients for the organisms that allow for maximum growth, while the saline solution lacks any nutrients, so the organisms exist in a minimal growth state. Each condition was prepared in quadruplicate in a 96-well plate (SPL Life Sciences, Pocheon-Si, Gyeonggi-do, South Korea).

After plating, the 96-well plates were placed in a multimode microplate reader (Varioskan Lux, Thermo Fisher Scientific, Waltham, MA) and incubated at 35 °C for 24 hours. Turbidimetric measurements were recorded (λ = 620 nm) at 15-minute intervals. Between measurements, the plate underwent low intensity shaking for 10 seconds.

After 24 hours of incubation, the microplate was removed, and wells containing organisms were serially diluted in saline. Five microliters of each dilution were plated onto labeled Mueller-Hinton agar (Remel, Lenexa, Kansas, United States) plates and incubated at 37 °C for 48 hours. Bacterial colonies were counted and multiplied by the corresponding dilution factor to determine the final CFU/mL. Bacterial growth was quantified by calculating the logarithmic difference (log ₁₀ ) between the initial inoculum (1.5 × 10⁶ CFU/mL) and the final CFU/mL count.

Turbidimetric growth curves were log-transformed. The curves were averaged for each condition. Growth was qualitatively assessed based on the presence or absence of a logistic growth pattern. Curves exhibiting a sigmoidal increase were interpreted as evidence of growth, while flat or irregular curves were considered indicative of no detectable growth. For quantitative characterization, growth curves were fitted using the Gompertz model ^, in MATLAB (R2024a, MathWorks, Natick, MA). Fits with R² values greater than 0.95 were considered acceptable for further analysis of growth kinetics. Parameters derived from the model included the lag time (λ), representing the delay before exponential growth; the specific growth rate (µ), corresponding to the slope of the exponential phase. Additionally, the area under the curve (AUC) was calculated as an integrated measure of total microbial proliferation.