If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author. Daisuke Miyamori, Department of General Internal Medicine, Hiroshima University Hospital, 1-2-3 Kasumi, Minamiku, Hiroshima 734-8551, Japan, Tel: +81-82-257-5460
A total of 369,210 bacterial isolates were obtained from in-patients in the J-SIPHE database.
•
The bacterial species before/after the early COVID-19 phase in Japan were compared.
•
The ratio of number of bacteria isolated by droplet infection decreased.
•
Droplet transmission decreased/remained constant after the COVID-19 initial phase.
•
February-May 2020 showed increased behavioral change for COVID-19 prevention.
Abstract
Objectives
During the coronavirus disease 2019 (COVID-19) pandemic, a change in the trend of infections was observed. However, there are few reports comprehensively assessing the impact of the early phase of COVID-19 on the trend of bacteria isolated.
Methods
We extracted the number of positive cultures of hospitalized patients for approximately 200 institutions using the Japanese national database. The outcome was the ratio of 10 species isolated in comparison to the total isolates for each month. Interrupted time-series analyses were conducted between 13 (from Jan-2019 to Jan-2020) and 8 (from May-2020 to Dec-2020) monthly data points.
Results
A total of 369,210 isolates were involved. Differences in the level change for Streptococcus pneumoniae, Haemophilus influenzae, and Streptococcus pyogenes decreased significantly by 0.272 (95% confidence interval [CI]:0.192-0.352), 0.244 (95%CI:0.174-0.314), and 0.324 (95%CI:0.06-0.589), respectively. Bacteria transmitted by contact infection, such as Staphylococcus aureus, did not decrease. Differences in slope change were not significant in all species.
Conclusions
The ratios of isolated bacteria transmitted by droplet infection decreased immediately after the early phase of COVID-19 and maintained the same level. The awareness and behavioral changes toward increased COVID-19 prevention might have a substantial impact on the prevention of bacterial infections, especially droplet infections.
Since the early phase of the coronavirus disease (2019) COVID-19 in Japan, there has been a change in awareness and behavior among the public, including “keeping social distance,” “washing hands,” and “wearing masks,” which have improved public health by preventing infections. The most impactful trigger for the public in Japan for these changes in awareness and behavior was the widespread infection of COVID-19 on cruise ships in early February 2020. At the same time, the number of opportunities for people to come into contact was reduced by the declaration of a state of emergency and the closure of schools
The COVID-19 pandemic brought about a significant change in the trends of bacteria, and no increase in the prevalence of community-acquired pneumonia along with the COVID-19 pandemic during the winter of 2020-2021 was observed although there were concerns. Instead, it has been reported that the number of patients with community-acquired pneumonia admitted to hospitals and the number of patients with invasive bacterial infections that are sequels to respiratory tract infections has decreased since the COVID-19 pandemic
Changes in the incidence of invasive disease due to Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitidis during the COVID-19 pandemic in 26 countries and territories in the Invasive Respiratory Infection Surveillance Initiative: a prospective analysis of surveillance data.
The indirect impact of COVID-19 large-scale containment measures on the incidence of community-acquired pneumonia in older people: a region-wide population-based study in Tuscany, Italy.
However, statistics based on the number of cases, such as those in previous reports 3, only measure the statistical results diagnosed from various symptoms caused by bacterial infections. It is difficult to accurately determine bacterial trends, including subclinical infections and asymptomatic carriers, for these cases. It is unknown whether the decrease in the number of patients with bacterial infections caused a relative increase in the number of subclinical infections or a decrease in the number of bacteria. While it is possible that the interruption of the infection route directly affected the reduction in the number of patients, it is also possible that changes in lifestyle, such as diet, affected the human immune system, which in turn affected the onset of infectious diseases even after infection. Second, for bacteria mainly transmitted by droplet infection, a decrease in the number of patients has been reported, but no reports have compared multiple bacteria simultaneously. It is difficult to refer to a decrease in the number of cases due to the different routes of infection because bacteria from multiple routes of infection have not been comprehensively analyzed.
A comprehensive comparison and understanding of the changes in the trends of detection ratios for each bacterial isolates before and after the COVID-19 pandemic from based on the characteristics of infection routes are essential for the general prevention of infectious diseases. Analyzing the changes in the detection ratios of isolated bacteria from the culture and considering the route of infection mainly affected will help determine the bacterial groups affected by the infection prevention measures against COVID-19 as a secondary effect.
In this study, we investigated the effect of COVID-19 on bacteria using an interrupted time-series analysis (ITSA) of changes in the number of bacteria isolated from the cultures of in-patients during the early phase of COVID-19 using a large Japanese database.
Materials and Methods
Study design
We conducted ITSA of the detection ratios of bacteria isolated during the early phase of the COVID-19 pandemic, one of the most valuable methods for comparing the effects before and after an event using observational data
. ITSA is used to evaluate long-term trends at the population level under the assumption that the influences of confounding, such as contamination, are maintained at a constant rate
. This study determined whether the detection ratios of bacteria isolated in any cultures was associated with the early phase of COVID-19 using a large-scale database in Japan. The conduct of this study complied with the aims of Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Data source
We used data from the Japan Surveillance for Infection Prevention and Healthcare Epidemiology (J-SIPHE) database, which is collected and managed by the Antimicrobial Resistance (AMR) Clinical Reference Center of the National Center for Global Health and Medicine. The J-SIPHE database has been aimed at improving regional and national infection control and AMR measures since 2017. Information is regularly collected from participating registered facilities throughout Japan. As of July 2020, 518 facilities were participating. Among them, 418 facilities had entered information related to microbes and resistant bacteria, including the following: the amount and rate of antimicrobial use at each facility; the number of bacteria isolated and their antimicrobial susceptibilities; the number of multidrug-resistant bacteria isolated; and Antimicrobial Stewardship Program by consultants specializing in infectious diseases and infection control team activities in each hospital. In the section on microorganisms, the 21 major bacteria to be surveyed are voluntarily registered once detected by bacterial culture.
Hospitals that provide information, such as multicenter comparisons, are given feedback to promote AMR activities. The details of J-SIPHE are provided elsewhere
AMR Clinical Reference Center, Japan. Japan Surveillance for Infection Prevention and Healthcare Epidemiology: J-SIPHE. Available at https://j-siphe.ncgm.go.jp/. Accessed 8 May 2021.
This study involved facilities with antimicrobial testing systems contributing to the above registries. The facilities that contribute to these databases account for 75,000 acute care beds and approximately 10% of Japan's total acute care beds.
Study Population
Data on bacterial culture tests performed at the hospital for in-patients from approximately 200 facilities were extracted from an open cohort for which information on microorganisms, resistant bacteria, and all detected bacteria was registered in the database. The database used in this study is in an open cohort format, so the number of participating institutions may vary from year to year. Furthermore, to reduce errors in the species and number of bacteria isolated due to differences in healthcare delivery systems and patient backgrounds during infectious disease treatment, the selected data for in-patients satisfied the criteria for additional reimbursement for infection prevention type 1. The reasons for the errors include the size and function of the facilities. The Ministry of Health, Labour, and Welfare in Japan classified the additional reimbursement for infection prevention in medical institutions into several types according to the size and function of the hospital. Specific facility criteria for infection prevention must be met for additional reimbursement for an infection prevention type
Impact of health policy on structural requisites for antimicrobial stewardship: A nationwide survey conducted in Japanese hospitals after enforcing the revised reimbursement system for antimicrobial stewardship programs.
. The number of bacteria isolated from all cultures submitted to the hospital during the study period, including blood, sputum, urine, spinal fluid, ascites/pleural fluid, open pus, non-open pus, bile, endotracheal aspirate, various drain effluents, joint fluid, and stool, were extracted every month. In cases of multiple bacterial culture tests on the same patient, only the first detected bacteria within 90 days were counted, even if they were from different specimens. Among the bacteria registered in the database, those that could be pathogens in community-acquired infections were included in the study in order to assess whether COVID-19 affected the transmission of bacterial infections. The observation period was from January 2019 until December 2020, before and after the early stage of COVID-19 pandemic, respectively.
Outcome
The primary endpoint of this study was the detection ratios of isolated target bacteria. Since the number of participating institutions could vary by year, it was assumed that basing the outcome on the number of institutions would affect the outcome. Therefore, the number of each bacterium isolated per institution and for each month was calculated, with January 2019 set at “1”, and the ratios for the other months compared to the set January 2019 date. The same ratios as above for all bacterial species were used as the control because the present outcome could be affected by fluctuations in the total number of bacterial isolates due to medical care pressure caused by COVID-19 and refusal to see patients. In this study, We analyzed all the bacterial species registered in the database (Supplemental Table. 1).
The protocol for bacterial culture was based on the guidelines of the Clinical and Laboratory Standards Institute. There were no changes in the guidelines regarding the method of bacterial isolation during the observation period.
The early phase of the COVID-19 pandemic
After the first outbreak of COVID-19 was reported in Wuhan, China, in December 2019, the first case of severe acute respiratory syndrome coronavirus 2 infections was confirmed in Japan on January 15, 2020. In early February 2020, 712 infected persons with 14 deaths were recorded on the Diamond Princess cruise ship that was having 3,700 onboard. This sensitive event was reported on TV every day, and it was one of the most significant events that increased awareness and behavioral change toward COVID-19 infection prevention measures
. At the same time, the Ministry of Health, Labor, and Welfare requested pharmacies and supermarkets in Japan to restrict purchases of masks and disinfectants due to shortages
. The number of newly infected people continued to increase in March, and the Japanese government declared a state of emergency in all prefectures on April 16. After that, the number of infected people started to decrease, and the emergency declaration was lifted in all prefectures on May 25. Therefore, in this study, we set the period from February 2020, when awareness and behavioral changes started, to May 2020, when the emergency ban was lifted, as the early phase of COVID-19, based on previous reports.
Statistical Methods: Interrupted time series analysis
We used the ITSA to estimate changes in the level and the pre-existing trends of the study outcomes during the early phase of the COVID-19 pandemic while controlling for pre-pandemic levels and trends. We estimated the following linear segmented regression model:
where Yt is the outcome at time point t, time indicates the duration from the start of the study, exposure is a dummy variable representing the periods before and after the early phase of the COVID-19 pandemic, and εt is the error term. In this model, β0 estimates the baseline level of each bacterium at the beginning of the study, β1 represents the underlying pre-early phase of the COVID-19 trend, and β2 and β3 estimate the immediate level change and slope change following the early phase of COVID-19, respectively. To account for the lag between the initial effect of COVID-19, we considered a 3-month “phase-in” period, and three monthly data points for this period were excluded from the analysis. Therefore, after the early phase of COVID-19, we had 13 monthly data points from January 2019 to January 2020 and eight monthly data points from May 2020 to December 2020.
We used the “ITSA” command in STATA (Version 16SE; Stata Corporation, College Station, TX, USA), which estimates the model using ordinary least squares (OLS) with Newey-West methods. We set the lag period as 12 months because of the seasonality.
A significant threat to the validity of the ITSA is the history of events that occurred concurrently with the exposure, which may have caused changes in the observed outcomes. To account for this threat, we assessed changes in all the bacteria and compared them to the total number so that the phase-in periods was not expected to be influenced.
Ethics
Since this was an observational study using the AMR control database, which has already been compiled, and not involving human participants, we determined that ethical review was unnecessary according to the Ethical Guidelines for Medical and Health Research Involving Human Participants in Japan.
Results
The analysis involved 369,210 isolates over 24 months from January 2019 to December 2020 for the 21 strains. The number of isolates per month is presented in Supplemental Table 1. Figs. 1 and 2 show plots of the ratios of the bacteria isolates included. From the 21 species registered in the database, we selected 10 species that could cause community-acquired infections in healthy individuals.
The ratios of the isolates detected in January 2020, just before the early phase of COVID-19; May 2020, just after the early phase of COVID-19; and December 2020, at the end of the observation period, were as follows: Streptococcus pneumoniae (S. pneumoniae): 0.633, 0.218, and 0.224; Haemophilus influenzae (H. influenzae): 0.763, 0.253, and 0.252; Streptococcus pyogenes (S. pyogenes): 0.856, 0.337, and 0.27; Staphylococcus aureus (S.aureus): 1.12, 0.79, and 0.872, and Escherichia coli (E.coli): 0.897, 0.718, and 0.758, respectively. Trends in other species of bacteria are described in Supplemental Table 2.
Differences in the level and slope change
Fig. 3 shows the differences in the level and slope change for each bacteria isolate relative to those of the control group. The level changes of S. pneumoniae, H. influenzae, and S. pyogenes, relative to that of the control group, significantly decreased after the early phase of COVID-19, with the differences in level change being -0.272 (95% confidence interval [CI]; -0.352 to -0.192), -0.244 (95% CI; -0.314 to -0.174), and -0.324 (95% CI; -0.589 to -0.06), respectively. On the other hand, the level changes of Escherichia coli (E. coli) and Enterococcus faecalis (E. faecalis), relative to that of the control group, significantly increased after the early phase of COVID-19, with the differences in level change being 0.077(95% CI; 0.029 to 0.125) and 0.084 (95% CI; 0.031 to 0.136), respectively. Meanwhile, no significant difference in slope change compared to the control group was observed for any bacteria. The differences in the level and slope change for other species of bacteria are described in Supplemental Fig. 1.
Fig. 3Forest plots showing differences in the level and slope change between each bacteria isolate and control group. Bars in the left forest plot indicates difference in the level change with 95% confidence interval between each bacteria isolate and control group. The level for S. pneumoniae, H. influenzae, and S. pyogenes after the phase-in periods were significantly lower than those of the control group. Bars in the right forest plot indicates the slope change with 95% confidence interval between each bacteria isolate and control group.
S. pyogenes: Streptococcus pyogenes, S. pneuimoniae: Streptococcus pneumoniae, H. influenzae: Haemophilus influenzae, P. aeruginosa: Pseudomonas aeruginosa, K. pneumoniae: Klebsiella pneumoniae, CNS: coagulase-negative staphylococci, S. agalactiae: Streptococcus agalactiae, S. aureus: Staphylococcus aureus, E. coli: Escherichia coli, E. faecalis: Enterococcus faecalis, CI: Confidence interval.
Fig. 1The ratios of Streptococcus pneumoniae (S. pneumoniae), Haemophilus influenzae (H. influenzae), Streptococcus pyogenes (S. pyogenes), and Streptococcus agalactiae (S. agalactiae) detected were compared with those of the control group for the period between January 2019 and December 2020.
Fig. 2The ratios of intestinal microflora and indigenous skin bacteria detected were compared with those of the control group between January 2019 and December 2020.
This study analyzed the impact of the early phase of the COVID-19 pandemic on bacteria isolates using the large-scale database of J-SIPHE from January 2019 to December 2020. The four months, from February 2020 to May 2020, were designated as the early phase of the COVID-19 pandemic.
The detection ratios of S. pneumoniae and H. influenzae, causing community-acquired pneumonia, and S. pyogenes, causing tonsillitis, showed significant differences in level changes compared to that of the control group of -0.272, -0.324, and -0.244, respectively, and these bacteria were mainly transmitted by droplet infection. We could not identify any differences in level changes compared to that of the control group in the detection ratios of intestinal microflora, indigenous skin bacteria, or S. agalactiae.
S. pneumoniae, H. influenzae, and S. pyogenes are causative bacteria of respiratory tract infections, but they are also causative bacteria of invasive bacterial infections with high mortality rates in people of all ages. In this study, the number of causative bacteria isolates decreased, as well as the number of cases, which has been pointed out in previous reports
Changes in the incidence of invasive disease due to Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitidis during the COVID-19 pandemic in 26 countries and territories in the Invasive Respiratory Infection Surveillance Initiative: a prospective analysis of surveillance data.
The indirect impact of COVID-19 large-scale containment measures on the incidence of community-acquired pneumonia in older people: a region-wide population-based study in Tuscany, Italy.
. Therefore, the decrease in the number of bacteria isolates was behind the decrease in the number of cases of these infections.
Among these bacteria, S. pneumoniae and H. influenzae are the first and second most common causative bacteria of community-acquired pneumonia, accounting for 25% and 10%, respectively, in a survey conducted before the COVID-19 pandemic in Japan
Individual risk factors associated with nasopharyngeal colonization with Streptococcus pneumoniae and Haemophilus influenzae: a Japanese birth cohort study.
. S. pneumoniae causes fatal bacterial infection in older and immunocompromised adults when it causes community-acquired pneumonia; meanwhile, the target populations for vaccination in Japan are infants, older adults, and high-risk patients with underlying diseases. It is also known to be carried in the nasopharynx of 40-60% of infants and is responsible for several subclinical infections
During the period of awareness and behavioral changes, which coincides with the early phase of the COVID-19 pandemic, containment was not enforced as a policy, and voluntary home confinement and the closure of public schools nationwide may have had an impact
. Other specific examples of personal infection prevention measures practiced were “keeping social distance,” “washing hands,” and “wearing a mask.” A Japanese Internet survey on infection prevention measures showed that more than three-quarters of the respondents took preventive measures
, suggesting that the prevalence of respiratory viral infections other than COVID-19, as well as secondary bacterial pneumonia, decreased.
Meanwhile, a little difference in the level change in the contact route of infection compared to all bacterial species was observed. These may be partially due to endogenous transmission by bacterial species that are primarily contact-transmitted with little impact on behavioral change. However, as for endogenous infections, the impact of such infections cannot be assessed because the proportion of endogenous infections and other transmission route among each species of bacteria may varies. Thus, it is possible that there was a certain impact for each species of bacteria.
On the other hand, there was a decrease in the ratios of total isolated cultures of bacterial species (Supplemental Fig.2). The level change was significantly decreased before and after the phase-in periods at -0.2 (95% CI; -0.27 to -0.13). This may be due to the combination of various effects, such as infection control, behavioral restrictions, and refraining from medical visits, therefore, the ratio of total isolates was used as control group in this study, and we compared the differences in levels and slopes between the control and each bacterium to identify the bacteria that were significantly affected.
A snap poll conducted in 28 countries showed that 71% of the respondents frequently washed their hands, as one of the infection prevention measures for COVID-19 in Japan
. However, the following considerations are important for the results that did not show a relative decrease: hand washing may have been inadequate; hand washing alone did not directly affect the reduction of bacteria, and long-term evaluation may be necessary. Therefore, it is impossible to make a definitive statement about the specific changes in awareness and behavior that directly affected the bacteria, as a combination of factors may be suggested. However, there was no relative decrease in the number of bacteria isolates mainly due to contact infection, and the relative decrease in the number of bacteria mainly transmitted through droplet infection suggested that “keeping social distance” and “wearing masks” have a specific impact at the individual level. Therefore, awareness and behavioral changes may be leveraged to prevent respiratory bacterial infections in people who cannot take vaccines for various reasons or in older adults at risk of severe infections.
This study included patients with all diseases for which hospitalization was indicated for bacterial culture testing at a group of major acute care hospitals in Japan. The validity and reliability of the test results were considered high because these facilities performed the test according to the guidelines of the Clinical and Laboratory Standards Institute, which are used globally.
There are several limitations in this study. The limitations of this study include the fact that we were unable to analyze each culture specimen by the type of bacteria because we did not know the number of bacteria isolates for each type of culture specimen. In addition, the absolute number of hospital admissions and culture tests, including sputum tests, may have decreased due to the COVID-19 pandemic because patients could not receive appropriate medical care and undergo appropriate tests due to the reduced workforce and vacant beds. Moreover, there might be an bias due to the increase of false negative in culture tests due to the change of the community-prescribing oral antimicrobial agents.
However, it is unlikely that the participating facilities did not perform culture tests because they are hospitals actively implementing measures against antimicrobial resistance and have infectious disease physicians who monitor the administration of unnecessary antimicrobial agents and conduct appropriate bacterial culture tests. In addition, a decrease in the total number of inpatients has been ruled out in previous reports
. Regarding the possibility of a decrease in the number of respiratory culture specimens submitted, no decrease in the number of specimens submitted were found between 2019 and 2020 in the J-SIPHE database (Supplemental Table 3). Therefore, the impact of changes in the number of submitted specimens after the COVID-19 pandemic was considered negligible. Although culture-positive specimens were analyzed by ITSA, the total number of cultures and the number of culture-negative specimens were unknown, and some effect on the results of the analysis that this may have had, cannot be ruled out. Regarding concerns about bias that false negative culture test may increase due to changes in oral antimicrobial dosing, the total antimicrobial dosing in Japan during the year 2020 decreased compared to 2019
, which would contribute toward underestimating the results. Finally, the data were compiled by open cohort, and because the data on the number of beds per institution were not available, we used the average number of bacteria isolates by participating institutions in each month. However, it is possible that the number of beds per institution may differ, and that the number of beds at participating institutions may be non-uniform.
In conclusion, we conducted a longitudinal retrospective cohort study of bacterial culture test isolates from inpatients using a large-scale nationwide database and compared the bacterial pre- and post-early phase of the COVID-19 pandemic. During the COVID-19 pandemic, the detection ratio of bacteria, transmitted through droplet infections, decreased, and maintained the same level. Our results indicate the changes in public health awareness and the behavior of the public and social communities have a significant impact on bacterial trends.
Declaration of Competing Interest
The authors declare no conflicts of interest associated with this manuscript.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author contributions
MK and DM developed the study design and analyzed the data. SK extracted the data. MK. and DM wrote the manuscript. KI, TK, KO, HO, and MI provided advice on the interpretation of data and revised the manuscript. All authors read and approved the manuscript.
Acknowledgements
we thank the amr clinical reference center, national center for global health and medicine for access to the database used for this study
Changes in the incidence of invasive disease due to Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitidis during the COVID-19 pandemic in 26 countries and territories in the Invasive Respiratory Infection Surveillance Initiative: a prospective analysis of surveillance data.
The indirect impact of COVID-19 large-scale containment measures on the incidence of community-acquired pneumonia in older people: a region-wide population-based study in Tuscany, Italy.
AMR Clinical Reference Center, Japan. Japan Surveillance for Infection Prevention and Healthcare Epidemiology: J-SIPHE. Available at https://j-siphe.ncgm.go.jp/. Accessed 8 May 2021.
Impact of health policy on structural requisites for antimicrobial stewardship: A nationwide survey conducted in Japanese hospitals after enforcing the revised reimbursement system for antimicrobial stewardship programs.
Individual risk factors associated with nasopharyngeal colonization with Streptococcus pneumoniae and Haemophilus influenzae: a Japanese birth cohort study.
00686-7&id=gr3.jpg){kind=link}
00686-7&id=gr1.jpg){kind=link}
00686-7&id=gr2.jpg){kind=link}