Opening methods
Two injection needle product opening methods were employed: the “peel-apart method” (Fig. 1a) where the adhesive surface of the mount for packaging is peeled off, and the “push-off top method” (Fig. 1b) where the needle hub is pressed against the mount to break it [5].
Bacterial strains and preparation of bacterial solutions
The methicillin-susceptible Staphylococcus aureus strain of the American Type Culture Collection (ATCC) 29213 was used [6]. ATCC 29213 was provided by the Kitasato University-Laboratory of Infection Control and Research Center. A bacterial solution was cultured for 10 h with shaking and was diluted with physiological saline to an absorbance of 0.3 using an absorption spectrometer at 578 nm [7]. The concentration of the bacterial suspension was 108 colony-forming units/ml (CFUs/ml) before the experiment. This solution was diluted with physiological saline to six different concentrations (108, 107, 106, 105, 104, and 103 CFUs/ml).
Experimental contamination of mounts
A total of 240 injection needles, including 120 each adherently packaged with a paper-mount and transparent plastic blister (18G: Terumo Co., Ltd., Tokyo, Japan) or with a plastic (combination of polystyrene and polyethylene terephthalate) mount and transparent plastic blister (18G: NIPRO Co., Ltd., Osaka, Japan) were used. The injection needle was taken out of the box just before the initiation of experiments and was stored on a clean bench after disinfection. Injection needles were classified into two groups according to the opening methods: the peel-apart and push-off top methods (60 needles each). Experiments were conducted separately for 10 needles each at six different concentrations of the bacterial suspension. To assess the risk of needle contamination by various quantitative concentrations under clinical settings, 10 μl of each of the bacterial suspensions (108, 107, 106, 105, 104, and 103 CFUs/ml) was applied to the part near the needle hub’s opening at the mount of an unopened injection needle product using a pipette tip on a clean bench (shaded parts in Fig. 1a, b). Using the peel-apart method, the bacterial suspension was applied to the gripped part of the mount (shaded parts in Fig. 1a). Using the push-off method, the bacterial suspensions were applied to the part potentially touching the mount when removing the needle (shaded parts in Fig. 1b).
Injection needles were then opened using the peel-apart or push-off top method with disinfected gloves (Fig. 1a, b). On a clean bench, half of the needles were opened as soon as the bacterial suspensions had been applied (wetness group). The other half were dried using the filtering airflow of the clean bench at room temperature. One hour later, the dry state of suspensions applied was confirmed visually and needles were then opened on the clean bench (dryness group). Injection needles were taken out to examine the degree of contamination in each part of the needle hub.
To examine the degree of contamination on each site of the needle hub (Fig. 2), all lateral surfaces (Fig. 2a) were placed on agar medium and rotated to be brought into contact with the medium. The bottom surface of the needle hub (Fig. 2b) was then pressed against the agar medium. To examine contamination in the inner lumen (Fig. 2c), a 1-ml syringe containing 0.1 ml of saline was connected to the needle to discharge all saline onto the agar medium.
The agar medium was incubated at 37 °C for 30 h for colony counting. Brain heart infusion agar (Becton, Dickinson, and Company, USA) was used as the agar medium.
Emulated clinical contamination (hand and saliva contamination of mounts)
Based on the rare occurrence of needle hub contamination in the previous experiment using the peel-apart method, various conditions (i.e., dry or wet hands, saliva contamination) were examined to evaluate the clinical risk of needle hub contamination using the push-off top method. Five anesthesiologists were included in the present study. This investigation was conducted in accordance with the current Declaration of Helsinki. The authors’ own samples were collected, and patients and volunteers were not included. All samples were anonymized after collection for the impossibility to identify the specific individual. A total of 150 injection needles, including 75 each in the paper-mount group and plastic-mount group, were used.
Anesthesiologists rubbed dry/wet hands on the paper or plastic-mounts without gloves. To simulate wet hands, 10 μl of autoclaved physiological saline was applied to dry hands using a micropipette. To simulate a hand contaminated with a patient’s saliva, gloved fingers licked by anesthesiologists were applied to each paper and plastic-mount.
Five saliva samples were obtained from each of the five anesthesiologists and were quantitatively cultured.
All injection needle products with clinically contaminated mounts were opened on a clean bench in the same manner as described for examination of the experimental contamination of mounts.
Statistical analysis
The numbers of bacteria on the lateral surfaces, bottom surfaces, and total surface (sum of the lateral surface, bottom surface, and inner lumen) of needle hubs were compared between the opening methods, dryness/wetness of bacterial solution, and mount materials using the Cochran-Mantel-Haenszel test considering the concentration of S. aureus as a stratum. By comparing the dryness/wetness of bacterial solution and mount materials, only data obtained using the push-off top method was used because only one needle hub was contaminated in the peel-apart method. The trend test for the concentration of S. aureus-contamination relationship was performed using the push-off top method data with the Jonckheere test.
The number of bacteria in the inner lumen of a needle hub was classified into contaminated (≥ 1 colony) or uncontaminated (no colony), and this binary response was compared between opening methods, the dryness/wetness of the bacterial solution, and mount materials using Fisher’s exact test without considering the concentration of S. aureus. A trend test for the S. aureus concentration-contamination relationship in the number of bacteria in the inner lumen was not performed because of only five hubs were contaminated. Fisher’s exact test was instead applied to compare S. aureus concentrations.
Regarding emulated clinical contamination data, the number of bacteria was compared between mounts using the Cochran-Mantel Haenszel test considering the anesthesiologist and wet/saliva hands as strata, excluding dry hand data because all were zero. The numbers of bacteria on the lateral surfaces, bottom surfaces, inner lumens, and total surface of needle hubs were compared between dry/wet/saliva hands using the Cochran-Mantel-Haenszel test considering anesthesiologist as a stratum. The mount was not included into stratum because a large p value was obtained for the mount comparison. Pairwise comparisons were also performed. The family-wise error rate was controlled using the closed testing procedure; first, data was compared between dry, wet, and saliva hands with a significance level of 5% and the testing procedure was stopped if not significant. Second, pairwise comparisons were performed with a significance level of 5% for each test if the first step was significant. A p value of < 0.05 was considered to be significant. Data were analyzed using SAS version 9.4 (SAS Institute, Cay, North Carolina, USA).