key: cord-0851678-8b19y0hk
authors: Ryu, Sukhyun; Hwang, Youngsik; Ali, Sheikh Taslim; Kim, Dong-Sook; Klein, Eili Y; Lau, Eric H Y; Cowling, Benjamin J
title: Decreased use of broad-spectrum antibiotics during COVID-19 epidemic in South Korea
date: 2021-04-15
journal: J Infect Dis
DOI: 10.1093/infdis/jiab208
sha: 6a36928a86794830087d60fddef558660381883e
doc_id: 851678
cord_uid: 8b19y0hk

BACKGROUND: Early in the COVID-19 pandemic, there was a concern over possible increase in antibiotic use due to co-infections among COVID-19 patients in the community. Here, we evaluate the changes in nationwide use of broad-spectrum antibiotics during the COVID-19 epidemic in South Korea. METHODS: We obtained national reimbursement data on the prescription of antibiotics, including penicillin with beta-lactamase inhibitors, cephalosporins, fluoroquinolones, and macrolides. We examined the number of antibiotic prescriptions compared with the previous three years in the same period from August to July. To quantify the impact of the COVID-19 epidemic on antibiotic use, we developed a regression model adjusting for changes of viral acute respiratory tract infections (ARTIs) which are an important factor driving antibiotic use. RESULTS: During the COVID-19 epidemic in South Korea, the broad-spectrum antibiotic use dropped by 15%–55% compared to the previous 3 years. Overall reduction in antibiotic use adjusting for ARTIs was estimated to be 14%–30%, with a larger impact in children. CONCLUSIONS: Our study found that broad-spectrum antibiotic use was substantially reduced during the COVID-19 epidemic in South Korea. This reduction can be in part due to reduced ARTIs as a result of stringent public health interventions including social distancing measures.

M a n u s c r i p t 3

The spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus, the cause of coronavirus disease 2019 (COVID- 19) , has had an enormous health impact worldwide. Beyond the direct health impacts of infection and hospitalization with COVID-19, there is also the possibility of impact on other health conditions, and on changes in the use of medications such as antibiotics in the community [1] . Widespread changes in antibiotic use would have knock-on effects on antimicrobial resistance (AMR) which is a global health problem [2] [3] [4] . However, to date, there has been limited assessment of the impact of the COVID-19 pandemic on antibiotic use.

In South Korea, the first COVID-19 case was identified on 20 January 2020, and the national public health alert was raised to the highest level on 23 February 2020 shortly after a localized outbreak in Daegu city [5] . Combined public health measures were implemented to control the spread of COVID-19 [5] and strict social distancing measures were implemented on 22 March and relaxed on 20 April 2020 [6] .

As public health measures were implemented, reduced respiratory virus activity was observed globally which could lead to a slowdown in antibiotic use [7] and potentially AMR in the community [8] . Monitoring nationwide patterns of antibiotic use is thus crucial to evaluate antimicrobial stewardship during the COVID-19 pandemic [9] . Here, we aimed to quantify the impact of the COVID-19 epidemic on the antibiotic use in South Korea in 2020. To achieve this, we analyzed the data on usage of broad-spectrum antibiotics that are commonly used to treat acute respiratory tract infections (ARTIs), which include bronchitis, rhinitis, and sinusitis, and their sequelae.

A c c e p t e d M a n u s c r i p t 4

We obtained nationwide antibiotic prescribing data between August 2016 and July 2020 from the Korean Health Insurance Review and Assessment Service based on reimbursement data from over 80,000 health care providers in South Korea, covering around 95% of the South Korean population.

The monthly antibiotic prescription data were classified by the Anatomic Therapeutic Chemical (ATC)

Classification System for penicillin with beta-lactamase inhibitors (J01CR), cephalosporins (J01D), fluoroquinolones (J01MA), and macrolides (J01FA). Data were obtained for different age groups (<5, 5-19, 20-49, 50-65, and ≥65-year-old) and types of medical institutions (primary clinics, secondary and tertiary hospitals). Primary clinics were defined as institutions with fewer than 30 beds that mainly provided outpatient care, and secondary hospitals were defined as medical institutions with more than 30 beds. Tertiary hospitals were defined as hospitals commonly conducting complicated treatment for severe illnesses with more than 100 beds designated by the Korean Ministry of Health and Welfare [10] .

The number of patients hospitalized for common viral ARTIs was used as a proxy measure for the overall activity of ARTIs, which is known as a major driver of antibiotic use in the community [11] .

Data were collected from weekly sentinel surveillance reports by the Korea Disease Control and Prevention Agency obtained from 196 sentinel hospitals [12] . The clinical samples acquired from hospitalized patients were tested according to a consistent protocol, using multiplex polymerase chain reaction (PCR) or real-time reverse transcription PCR. Respiratory viruses including human adenovirus, bocavirus, parainfluenza virus, respiratory syncytial virus, rhinovirus, metapneumovirus, A c c e p t e d M a n u s c r i p t 5 coronavirus, and influenza virus were tested. Using the extracted hospitalization data, we estimated the nationwide rate of patients hospitalized for viral ARTIs per 100,000 individuals.

To examine the effect of the COVID-19 epidemic on antibiotic use, we first compared the pattern of antibiotic use during 2019/20 with the preceding 3 years (2016/17, 2017/18, and 2018/19). We computed weekly antibiotic use from monthly data by interpolating the splines functions keeping the monthly counts the same [13] . We then quantified the change in the weekly use of antibiotics comparing the pattern during pre-and post-timing of the highest public health alert (23 February 2020) for COVID-19 in South Korea [5, 14] . To compare the changes in antibiotic use, we used Welch two-sample t-test on the weekly antibiotic use for pre-and post-periods of the highest public health alert. Furthermore, to identify the excess number of reductions on weekly antibiotic use, we measured the weekly difference of the use between 2016/17 -2018/19 and 2019/20, and compared the difference for the pre-and post-periods of the highest public health alert. We also used a paired t-test on the weekly antibiotic use during the months March through July in 2020 and the preceding 3 years.

To quantify the impact of the COVID-19 epidemic on antibiotic use accounting for potential change in ARTIs, we estimated the overall reduction in antibiotic use during the COVID-19 epidemic by estimating the difference of each fitted regression model for the weekly antibiotic use, adjusted for viral ARTI hospitalizations.

As antibiotic use is highly correlated with the activity of viral ARTIs (i.e. , the number of hospitalizations for viral ARTIs at time t) [15] , and affected by other extrinsic factors such as the COVID-19 epidemic (i.e. , an indicator variable defined as for the weeks after declaration of the national public health alert (23 February -31 July 2020) and 0 otherwise), the regression A c c e p t e d M a n u s c r i p t 6 model was defined as . The decreases (/increases) over time as the decreases (/decreases) and further modified by accounting for the effects of the public health measures on COVID-19. and are the regression coefficients that quantify the effect of COVID-19 epidemic on antibiotic use ( ). Simplifying, we constructed a log-linear multivariable regression model as , where and were estimated, and the overall reduction in was estimated using these parameters.

We conducted similar analyses for different age groups and types of medical institutions. All analyses were done in R version 3.6.1 software (R Foundation for Statistical Computing, Vienna, Austria).

We identified a clear seasonal pattern in broad-spectrum antibiotic use in South Korea from 2016-2019, with a large peak around December and a second peak around April ( Figure 1 ). In 2019/20 the pattern was similar until emergence of the COVID-19 epidemic, after which the use of all broadspectrum antibiotics dropped substantially compared with the previous 3 years (all p-values < 0.001, Table 1 ). The estimated reduction on weekly antibiotic use in 2019/20 was 178 per 1000 persons for penicillin with beta-lactamase inhibitors, 106 per 1000 persons for cephalosporins, 72 per 1000 persons for macrolides, and 35 per 1000 persons for fluoroquinolones (Table 1 ). Compared to March-July in the previous three years, the use of penicillin with beta-lactamase inhibitors, cephalosporins, macrolides, and fluoroquinolones in 2020 were reduced by 54%, 32%, 55%, and 15%, respectively ( Figure 1 and Table 1 ). Antibiotic use rebounded after strict social distancing measures were relaxed in April, but still remained below mean use in the prior years for all broadspectrum antibiotics except fluoroquinolones (Figure 1) . (95% CI, 4%-25%) for fluoroquinolones use. For macrolides, the reduction was 24% (95% CI, -3%-59%), but this was not statistically significant.

For the use of penicillin with beta-lactamase inhibitors, cephalosporins, and macrolides, similar effects were estimated across different age groups, including pre-schoolers (0-5 years; 31%-51%), school-aged children (6-19 years: 23%-51%), young adults (20-49 years: 16%-36%), and the elderly (above 65 years: a 15-30%) ( Table 2) . However, the effects on the use of fluoroquinolone were not statistically significant for children and adults. For the medical institutions, primary clinics and the secondary hospital similarly had reduced antibiotic use of around 16%-46% and 10%-45%, respectively. However, the decline was lower at tertiary hospitals (7%-16%) and the changes in fluoroquinolones use were not statistically significant at tertiary hospitals (Table 3) .

After declaration of the highest public health alert on COVID-19 in South Korea, use of penicillin with beta-lactamase inhibitors, cephalosporins, macrolides, and fluoroquinolones dropped by 54%, 32%, 55%, and 15% compared to prior years. Accounting for reduced ARTI hospitalizations during the COVID-19 epidemic in South Korea, the estimated overall reduction on antibiotic use was 14%-30%.

This result was significant across all broad-spectrum antibiotics except macrolides. However, in the different age groups and different types of medical institutions, the overall reduction on macrolides A c c e p t e d M a n u s c r i p t 8 was significantly associated with the COVID-19 epidemic, when adjusted for viral ARTI hospitalizations. This suggests that macrolides may be used in a more targeted manner and thus were less affected by broad reductions in overall use.

For the overall reduction adjusting for viral ARTI hospitalizations by the different age-groups, the changes in the use of fluoroquinolones were not significantly associated with the COVID-19

epidemic. This is likely to be affected by the narrow indication (e.g. genitourinary infections or hospital-acquired infections) on the fluoroquinolone prescription for the children as well as adults.

This also explains to the non-significance of changes in fluoroquinolone use at the tertiary hospitals.

We also observed a clear rebound in antibiotic use after relaxing strict social distancing measures in each of the antibiotics analyzed. This may have been because strict social distancing measures reduced use of medical services resulting in fewer prescriptions for viral ARTIs, which are typically unnecessary. However, as we estimated hospitalizations with the ARTIs, and these are unlikely to be affected by medical seeking behavior, it is more likely that social distancing measures reduced the spread of respiratory viruses, and that relaxing restrictions led to an increase in viral transmission,

ARTIs and subsequently antibiotic use [16] .

Consistent with the latter theory, in which tertiary hospitals generally have a more stringent prescription on broad-spectrum antibiotics, the level of reduction of antibiotic use was lower in tertiary hospitals than in other types of medical institutions.

Our findings are similar to a previous report that found reductions in the risk of acquiring and spreading respiratory infections during the COVID-19 pandemic through implementation of nonpharmaceutical measures including hand-washing, face masks, physical distancing, and travel restrictions [17] . Our findings also suggested that the implementation of public health measures A c c e p t e d M a n u s c r i p t 9 during the COVID-19 pandemic could lead to reductions in AMR in the community [18] . Because patients were less likely to seek care in the outpatient setting (e.g., primary physicians), there were fewer opportunities for antibiotics to be prescribed, which should reduce the selection pressure that drives AMR. Furthermore, there was no observed increase in the prescribing rate per ARTI patient visit suggesting that the pandemic did not erode antimicrobial stewardship efforts for ARTIs (Supplementary Figure 1 ).

Our findings have several limitations. First, since February 2020, the Korean government has recommended that all ARTI patients be screened for COVID-19 without cost [19] . Furthermore, prescribing requirements were loosened to allow prescriptions to be written through telephone A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t 15 A c c e p t e d M a n u s c r i p t 17 

Antibiotic use in the COVID-19 crisis in Spain

COVID-19, superinfections and antimicrobial development: What can we expect?

Will coronavirus disease (COVID-19) have an impact on antimicrobial resistance? Euro surveillance : bulletin Europeen sur les maladies transmissibles

PRO: The COVID-19 pandemic will result in increased antimicrobial resistance rates

Effect of Nonpharmaceutical Interventions on Transmission of Severe Acute Respiratory Syndrome Coronavirus 2, South Korea

Stepwise Strategy of Social Distancing in Korea

The Impact of Influenza Vaccination on Antibiotic Use in the United States

COVID-19 and the potential long-term impact on antimicrobial resistance

Changes in Antimicrobial Utilization During the COVID-19 Pandemic after Implementation of a Multispecialty Clinical Guidance Team

Factors predicting antibiotic prescription and referral to hospital for children with respiratory symptoms: secondary analysis of a randomised controlled study at out-of-hours services in primary care

Korea Disease Control and Prevention Agency

Mitigation of Influenza B Epidemic with School Closures

KCDC Risk Assessments on the Initial Phase of the COVID-19 Outbreak in Korea

Temporal relationship between antibiotic use and respiratory virus activities in the Republic of Korea: a time-series analysis

Effects of School Holidays on Seasonal Influenza in South Korea

COVID-19 and Antimicrobial Resistance: Parallel and Interacting Health Emergencies

Antimicrobial resistance and COVID-19: Intersections and implications

Korean Ministry of Health and Welfare

Estimated overall reduction in the weekly number of antibiotic prescriptions per 1000 antibiotic use per 1000 persons, % (95% CI)

A c c e p t e d M a n u s c r i p t 12 A c c e p t e d M a n u s c r i p t