Discussion

In this study, we investigated the molecular epidemiological transition of A. baumannii at a single hospital over >10 years. We observed the loss of clonal diversity due to the introduction of drug-resistant epidemic clones into our hospital and the occurrence of outbreaks, ^{Reference Baang, Axelrod and Decker5} as well as the recovery of clonal diversity after the outbreaks, along with changes in the in-hospital antibiogram. Interestingly, the epidemic clone was not eliminated and remained in the facility, suggesting that outbreaks may occur in the future and that continued surveillance is therefore warranted. Furthermore, POT was an effective surveillance method in terms of convenience and strain preservation.

The epidemiology of antimicrobial resistance of Acinetobacter differs among countries. ^{Reference Xie, Zhang, Zhao, Peng and Zheng7} Although the drug-resistance rate of Acinetobacter in Japan is low, ⁸ MDRA outbreaks have been reported. ^{Reference Ushizawa, Yahata and Endo15} Molecular epidemiology has shown that specific epidemic clones are prevalent within facilities during MDRA outbreaks, ^{Reference Corbella, Montero and Pujol4,Reference Baang, Axelrod and Decker5} and the effect of continuously monitoring epidemic clones as part of a bundled approach to prevent outbreaks has also been noted. ^{Reference Nguyen and Joshi2,Reference Rodríguez-Baño, García and Ramírez16} POT not only can determine a specific epidemic clone but also can evaluate strain homology in detail. POT can be used to observe the trends of specific strains and the diversity of clones in a hospital. The results of this study highlight the utility of the method. ^{Reference Fujikura, Yuki and Hamamoto12–Reference El Far, El-Mahallawy and Attia14}

In molecular epidemiological analysis, MLST has been a common approach for determining epidemic clones. In addition, PFGE has been used to evaluate the association between strains detected during outbreaks. ^{Reference Villalon, Valdezate and Cabezas10,Reference Peleg, Seifert and Paterson17} However, PFGE is less convenient than POT in that it requires simultaneous analysis of all potentially related strains, which is time-consuming and expensive. POT can be used for clone determination and to assign a strain-specific number by a simple multiplex PCR reaction, providing quantitative and objective unique data even if the strains are not preserved. The data can also be stored easily. POT is suitable for long-term monitoring because it allows easy comparisons of previous and new clinical specimens. ^{Reference Fujikura, Yuki and Hamamoto12}

The POT#1=122 clone that entered our facility had a distinctly different drug-resistance profile from the other clones identified and was the primary clone responsible for outbreaks at our hospital. POT#1=122 belongs to International Clone-2, ^{Reference Suzuki, Hosoba, Matsui and Arakawa11,Reference El Far, El-Mahallawy and Attia14} which tends to become persistent in the hospital environment ^{Reference Giannouli, Antunes, Marchetti, Triassi, Visca and Zarrilli18} and to acquire drug resistance, ^{Reference Karah, Sundsfjord, Towner and Samuelsen19} which also makes it an outbreak-causing clone in an institution. At the time of both outbreaks, our hospital took a bundled approach, ^{Reference Rodriguez-Bano, Garcia and Ramirez20} including thorough cleaning of the environment. However, the persistent detection of this clone means that it has not been eliminated. It is presumed to have persistently colonized the environment and to be carried in a certain number of patients attending our hospital. ^{Reference Baang, Axelrod and Decker5,Reference Karah, Sundsfjord, Towner and Samuelsen19}

In-hospital antibiograms of A. baumannii showed that its resistance rate increased prior to the outbreaks. Furthermore, there was a temporary deterioration in resistance during these outbreaks, which improved afterward. When the frequency of drug-resistant epidemic clones is socially low, it is possible that the in-hospital antibiogram will improve when the outbreak subsides and the diversity of clones that remain susceptible will be restored. Conversely, in our hospital, the epidemic clones from 2019 that acquired antimicrobial resistance were detected more frequently and increased the rate of antimicrobial resistance, indicating the need for careful monitoring to prevent future outbreaks of MDRA.

This study had several limitations. First, epidemiological data are only partially reflected, especially because our analysis was limited to blood cultures. However, the use of only blood-culture strains did provide information on molecular epidemiological trends at our hospital. Second, because this was a retrospective study and there were limitations in the collection of specimens, the data are similarly partial. These limitations could be resolved in the future by performing POT in all cases, regardless of specimen type, when A. baumannii is detected.

In this study, POT was able to show the introduction, outbreak, and subsequent transition of an epidemic clone of A. baumannii, which can readily acquire drug resistance and cause an outbreak within a single facility. POT analysis showed that during an outbreak, the overall diversity of A. baumannii in the facility disappeared and a specific epidemic clone emerged, and that the diversity of the clones recovered again after a certain period. However, specific epidemic clones are not eliminated from a hospital, and it is highly likely that clones with characteristics that make them more likely to spread in an institution cannot be eliminated. POT is a simple and useful method to monitor the presence of such clones in an institution prior to and during an MDRA outbreak.