Antibiotic Resistance during COVID-19: A Systematic Review

One of the public health issues faced worldwide is antibiotic resistance (AR). During the novel coronavirus (COVID-19) pandemic, AR has increased. Since some studies have stated AR has increased during the COVID-19 pandemic, and others have stated otherwise, this study aimed to explore this impact. Seven databases—PubMed, MEDLINE, EMBASE, Scopus, Cochrane, Web of Science, and CINAHL—were searched using related keywords to identify studies relevant to AR during COVID-19 published from December 2019 to May 2022, according to PRISMA guidelines. Twenty-three studies were included in this review, and the evidence showed that AR has increased during the COVID-19 pandemic. The most commonly reported resistant Gram-negative bacteria was Acinetobacter baumannii, followed by Klebsiella pneumonia, Escherichia coli, and Pseudomonas aeruginosa. A. baumannii and K. pneumonia were highly resistant to tested antibiotics compared with E. coli and P. aeruginosa. Moreover, K. pneumonia showed high resistance to colistin. Commonly reported Gram-positive bacteria were Staphylococcus aureus and Enterococcus faecium. The resistance of E. faecium to ampicillin, erythromycin, and Ciprofloxacin was high. Self-antibiotic medication, empirical antibiotic administration, and antibiotics prescribed by general practitioners were the risk factors of high levels of AR during COVID-19. Antibiotics’ prescription should be strictly implemented, relying on the Antimicrobial Stewardship Program (ASP) and guidelines from the World Health Organization (WHO) or Ministry of Health (MOH).


Introduction
On 11 March 2020, the WHO announced the COVID-19 pandemic [1]. The disease known as COVID-19 or SARS-2 spread rapidly from Wuhan City, China, to the rest of the globe [2]. As of early July 2022, oughly 547,901,157 COVID-19 cases and 6,339,899 deaths have been officially reported [3].
During the COVID-19 pandemic, there were improper uses of antibiotics either in healthcare institutions or in communities, which in turn played a role in the increase in AR [4][5][6]. It has been documented that about 72% of COVID-19-admitted patients were treated with antimicrobials, whereas solely 8% of these patients had bacterial or fungal co-infection [4]. Additionally, different antibiotics have been explored or suggested to cure COVID-19 patients, e.g., azithromycin [4,5]. Both the worry and the improper use of antibiotics directly impact access to antibiotics without a prescription, particularly low-and middle-income countries that have a weak system of antibiotic control. In this correlation, Zavala-Flores E et al., 2020, reported that nearly 69% of COVID-19 patients stated that they had used antibiotics (namely, ceftriaxone and azithromycin) before being admitted to the hospital [6].
Furthermore, during the COVID-19 pandemic, there was a huge increase in the use of biocides universally. These biocides probably encouraged more indirect pressure leading to AR [4]. Since early 2020, this situation has expanded globally and might have supported the evolution of extremely resistant microorganisms, which might have played a critical role in worsening the status of some patients, especially those who were admitted to intensive care units (ICUs). It has been reported that there were some deadly co-infections caused by pan-resistant microorganisms among COVID-19 patients. S. aureus and A. baumannii were the major ones that were resistant to extended-spectrum antibiotics, which were mostly used to cure life-threating diseases caused by bacterial infections [7].
Findings from a review stated that despite the bacterial infections associated with COVID-19, patients were less affected than in the influenza pandemic. COVID-19 patients were affected by common types of bacterial co-infection. These included S. aureus, Streptococcus pneumoniae, Klebsiella spp., Mycoplasma pneumonia, Legionella pneumophila, and Haemophilus Sp., Mycobacterium tuberclosis as a co-infection among COVID-19 patients. The study, however, reported that the rates of secondary bacterial co-infection were high among COVID-19 patients admitted to ICU, which could be due to hospital-acquired AR bacteria. Consequently, the study recommended urgently revising the empirical broad-spectrum antibiotics prescribed to COVID-19 patients and considering the importance guidelines of ASP [8].
COVID-19 patients who were admitted to ICU mostly required intubation and were at risk of ventilator-associated pneumonia, especially Gram-negative bacteria (P. aeruginosa, Acinetobacter Sp., and K. pneumoniae) and Gram-positive bacteria, (S. aureus). A study targeting five ICUs in Britain revealed that the prevalent bacteria among COVID-19 patients were Klebsiella aerogenes and K. pneumonia [9], whereas excessive levels of non-fermenters were found in one hospital in France [10]. COVID-19 patients on ventilators often received courses of multiple antibiotics. ASP guidelines were unfortunately overrun during the peak of COVID-19 as the capacities of ICUs increased [11]. In Spain, it was reported that the use of antibiotic increased [12], and as the pressure of COVID-19 increased, the resistance may have increased accordingly.
Another study has reported that the occurrence of multidrug-resistant organisms (MDROs) has increased in the era of COVID-19 compared with three years before the pandemic, and there was a high incidence of extended spectrum beta-lactamase (ESBL) K. pneumonia [13]. Furthermore, recent reports found that AR during the COVID-19 pandemic was higher than in previous periods [14][15][16]. Since there are some studies that have reported that AR increased during COVID-19 and others that have stated otherwise, the aim of this review was to explore the impact of COVID-19 on AR. The specific objectives were to identify the pattern of reported AR during the COVID-19 pandemic, to determine the nature of reported AR during COVID-19, and to report the encountered risk factors of AR during COVID-19.

Materials and Methods
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed for reporting in this systematic review ( Figure 1) [17]. PRISMA is a set of evidence-based items to report systematic reviews and meta-analyses. It concentrates on reporting revisions which assess the impacts of interventions. It could be used for systematic reviewing without assessing the interventions such as the evaluation of the cause or diagnosis, etc. [17]. The protocol of this systematic review was registered on the PROSPERO database (CRD42022326361).

Inclusion Criteria
Studies were included based on the following criteria: 1.
Articles should be original studies.

2.
Studies should report data on at least these two variables: antibiotic resistance, and COVID-19.

3.
Studies should be written in English or at least their abstract should be in English.

4.
Studies should be published between 2019 (since announcing COVID-19 in the country where the included study conducted) and May 2022.

Inclusion Criteria
Studies were included based on the following criteria: 1. Articles should be original studies. 2. Studies should report data on at least these two variables: antibiotic resistance, and COVID-19. 3. Studies should be written in English or at least their abstract should be in English. 4. Studies should be published between 2019 (since announcing COVID-19 in the country where the included study conducted) and May 2022.

Exclusion Criteria
Studies were excluded if they were a case report, letter to the editor, conference articles, commentary, systematic review, or viewpoint. Studies were also excluded if they were written in a non-English language or reported AR in non-human populations.

Search Strategy
An electronic search was employed to find the published articles from December 2019 to 20 May 2022, which reported antibiotic resistance during COVID-19 through the following databases: PubMed, MEDLINE, EMBASE, Scopus, Cochrane, Web of Science, and CINAHL. The medRxiv database was also searched to ensure a comprehensive search for unpublished studies.

Exclusion Criteria
Studies were excluded if they were a case report, letter to the editor, conference articles, commentary, systematic review, or viewpoint. Studies were also excluded if they were written in a non-English language or reported AR in non-human populations.

Search Strategy
An electronic search was employed to find the published articles from December 2019 to 20 May 2022, which reported antibiotic resistance during COVID-19 through the following databases: PubMed, MEDLINE, EMBASE, Scopus, Cochrane, Web of Science, and CINAHL. The medRxiv database was also searched to ensure a comprehensive search for unpublished studies.

Study Selection and Data Extraction
The initial screening for the title and abstract was performed by A.H. and A.A., and the full text screening for the eligible studies was performed by all authors. Data extraction was carried out by all authors, using a detailed extraction sheet including the following data: first author, country, year, setting, study design, duration, sample size, age, antibioticresistance-related data, and causative bacteria. Disagreement between authors was resolved by a joint discussion.

Quality Assessment
The quality assessment of the included studies was assessed based on the Joanna Briggs Institute critical appraisal tool [18]. The tool has 8 items to assess cross-sectional studies and 11 items to assess cohort studies. Each cross-sectional study was scored from 0 to 8, and the cohort study was scored from 0 to 11. Subsequently, the quality of the included studies ranked as high (for score ≥70%), medium (for score 50-69%) and low (for score <50%) [19]. All authors performed the assessment, and the issues encountered during the assessment were resolved by discussion among the authors.

Data Analysis
Data analysis was carried out using Microsoft Excel 2016. Median and IQR were used to present the resistance of each bacterium against various antibiotics. In this systematic review, the resistance of bacteria to the tested antibiotics, which was reported in more than three studies, was combined to identify the median and IQR of AR. In addition, an analysis for each study was performed narratively for the relevant data (AR findings, nature of AR bacteria, and potential risk factors).

Study Characteristics
The search strategy yielded 7189 studies: PubMed (875), Scopus (1367), Medline (217), Embase (2086), Web of Science (2325), CINAHL (24), and Cochrane (24), and an additional 271 studies were retrieved from medRxiv ( Figure 1). After removing the duplicates, 7121 studies remained for title and abstract screening. In total, 148 studies were eligible for screening the full text, of which 125 did not meet the inclusion criteria for the following reasons: short communication (4), brief report (1), and no relevant AR data (120). Thus, 23 studies met the inclusion criteria and were included in this review.
The summary of the characteristics and findings of the 23 included studies are presented in Table 1. The majority of the studies were from Iran (4) and India (4), followed by 2 studies from each the following countries: China, Italy, Turkey, and Saudi Arabia. Only 1 study was from each of the following countries: New York, Serbia, Egypt, Pakistan, Indonesia, Switzerland, and Greece. The majority of the studies employed a retrospective study design (10), followed by a retrospective observational (6), retrospective record review (3), and cross-sectional study (2). One study was a retrospective cohort study, and another was a retrospective follow-up study. Out of the 23 studies, 17 studies reported that AR emerged from ICUs, whereas only 6 studies reported some patient care areas in addition to ICUs (Table 1).   The hitherto observed resistances were as follows: amoxicillin/clavulanic acid = 84%, levofloxacin = 83%, ciprofloxacin = 79%, piperacillin/tazobactam = 77%, and trimethoprim/sulfamethoxazole = 75%. Generally, resistance to third-generation cephalosporins and carbapenems was (64%-69%). Notably, all isolates were found to be sensitive to colistin.
The majority of the studies (17) reported their number of samples as ranging between 13 and 856 patients, whereas only one study had 3532 patients. In total, 5 studies reported their samples as isolates ranging from 168 to 286 isolates, and only 1 study had 17,837 isolates. Only 1 study included 7309 samples pre-pandemic and 4968 samples during the pandemic phase in 2020, such as blood and urine samples. The majority (8) of studies reported the patient' ages as a median ranging from 56 to 67 years, whereas 6 studies reported ages as a mean ranging from 46 to 71 years, 3 studies classified ages as groups, and only 1 study presented the age as a range from 40 to 83 (Table 1).
Three studies were interested in identifying the rate of AR during or after COVID-19 compared with the era before the pandemic [23,33,36]. One study found that AR to imipenem, meropenem, and ciprofloxacin was significantly higher than the era before COVID-19 [23]. The prevalence of the resistance of S. aureus to oxacilin and Conrynebacterium striatum to vancomycin and linezolid during COVID-19 was higher than during the prepandemic era (Table 1) [33]. On the other hand, the rate of ESBL-producing Enterobacterales (MDROs bacteria) was similar in the era before and during COVID-19 [32]. The AR rates were similar before and during the COVID-19 pandemic (Table 1) [37].

Nature of AR during COVID-19
In total, four Gram-negative bacteria and two Gram-positive bacteria were commonly reported. Of the 23 included studies, 16 studies reported A. baumannii as one of the most common resistant bacteria, followed by K. pneumonia (15 studies), E. coli (10 studies), and P. aeruginosa (9 studies). However, among Gram-positive bacteria, S. aureus was mentioned in 3 studies as one of the most frequently resistant bacteria, followed by E. faecalis in 1 study, and E. faecium in another study (Table 1).

Potential Risk Factors
The risk factors of AR during COVID-19 were explored in only three studies. Selfantibiotic medication and antibiotics prescribed by general practitioners were significant risk factors for high levels of AR among the COVID-19 group compared with the non-COVID-19 group [24]. Another study reported that the administration of empirical antibiotics prior to ICU admission resulted in a high prevalence of MDRO [26]. In a study conducted in Iran, it was observed that 100% of patients who had MDR superinfection were imposed to empirical antibiotics, namely, meropenem and levofloxacin, with a median of duration of 12 and 9 days, respectively [28].

Discussion
In this systematic review, we examined the findings of 23 included studies that reported AR during COVID-19, and in three of them the reported risk factors were summarized. AR levels during COVID-19 were high, and the most commonly reported antibioticresistant Gram-negative bacteria were A. baumannii, K. pneumonia. Despite all Gramnegative bacteria in this study showing no resistance to colistin, K. pneumonia was high. Commonly reported Gram-positive bacteria were S. aureus and E. faecium, and a high resistance of E. faecium to ampicillin, erythromycin, and ciprofloxacin was observed. Selfantibiotic medication, empirical antibiotic administration, and antibiotics prescribed by general practitioners were the risk factors for high levels of AR during COVID-19.
Regarding the most commonly reported AR bacteria, a systematic review in 2019 reported E. coli as a common AR bacteria [41]. Another study reported the common Gramnegative bacteria as follows: P. aeruginosa, Klebsiella spp., A. baumannii, E. coli. Coagulasenegative Staphylococcus, Enterococcus spp., and S. aureus were the common Gram-positive bacteria [42]. Additionally, E. coli was a previously common resistant bacteria, followed by S. aureus [43,44]. Our findings are congruent with previous studies; however, among Gramnegative bacteria, A. baumannii and K. pneumonia were the most commonly reported ones.
In the present review, the resistance of A. baumannii to amikacin, cefepime, ceftazidime, gentamicin, meropenem, imipenem, ciprofloxacin, and piperacillin/tazobactam was higher than previously published studies before COVID-19. In a study carried out to report AR over five years before COVID-19, findings reported that A. baumannii was resistant to amikacin (49%), cefepime (78.6%), ceftazidime (73.8%), ciprofloxacin (46.7%), and piperacillin/tazobactam (62.2%). However, the resistance to meropenem was similar to the levels during COVID-19, and for imipenem this was 82.7%, which was still not higher than the levels observed during COVID-19 in the present study [42]. Another study conducted in 2019 mentioned that A. baumannii isolates were 66% resistant to the tested antibiotics, except colistin, which showed no resistance [45]. The WHO issued a report (2014-2019) illustrating the pattern of carbapenem resistance in A. baumannii, which was much lower than in the present study.
This review found that the resistance of E. coli isolates to amoxicillin clavulanate, cefuroxime, ceftriaxone, levofloxacin, and ciprofloxacin was increased during COVID-19.
In comparison, the resistance of E. coli before COVID-19 to ciprofloxacin and levofloxacin was 46% and 43%, respectively [42]. In China, in a study conducted to monitor AR for about 12 years until 2019, the resistance of E. coli to piperacillin/tazobactam did not exceed 8%, and resistance to ciprofloxacin did not exceed 60% [46]. Similarly, the AR of E. coli before COVID-19 to piperacillin/tazobactam had a median of 12%, as well as ciprofloxacin (65%), ceftriaxone (59%), and levofloxacin (62%) [47].
A study concerning AR in P. aeruginosa including 18 countries worldwide showed a low resistance to amikacin, gentamicin, ceftazidime, imipenem, ciprofloxacin, and levofloxacin [49]. A systematic review reported the resistance to imipenem in 2006, which was 42% and dropped gradually to 23% in 2017; moreover, the resistance to ciprofloxacin ranged from 32% to 14% over 11 years prior to COVID-19 [46]. The percentage of AR to carbapenem-resistant bacteria between 2014 and 2019 was 5% [43]. The findings of our review were not in line with previously published articles; moreover, the resistance to imipenem and ciprofloxacin was almost two times higher.
Colistin is an important antibiotic for various types of Gram-negative bacteria, and is the last resort for physicians to treat bacterial infections involving E. coli [50]. Previous studies reported very low resistance to colistin [47,48]. Notably, in our review, the resistance of K. pneumonia to colistin increased during COVID-19, with a median of 21.1% (IQR 12.42-69.82%).
Regarding the resistant Gram-positive bacteria, in the period from 2015 to 2019, the resistance of S. aureus to clindamycin lay between 17 and 15%, and the pattern of Enterococcus species resistance to ampicillin was 5-35%, erythromycin was 65-85%, ciprofloxacin was 60-80%, vancomycin was 10-50%, and tetracycline was 40-80% [44]. In other systematic reviews, S. aureus resistance to clindamycin was 11.7% and oxacillin was 34.5-46% [41,47]. In contrast, in the present review, the resistance of E. faecium to ampicillin, erythromycin, and ciprofloxacin was higher during COVID-19 than before. The resistance of S. aureus to clindamycin was 33.3% and oxacillin was 48.5%, which was still higher than before COVID-19.
Regarding the risk factors, it is important to note that sometimes antibiotics are selfadministered by individuals or prescribed by physicians to avoid bacterial colonization, even with no specific bacterial infection or laboratory-based confirmation. However, antibiotic treatment should be used based on accurate diagnosis [51]. About 72% of COVID-19-admitted patients in hospitals were treated with antimicrobials, whereas about 8% of these patients had bacterial or fungal co-infection [4]. Nearly 69% of COVID-19 patients stated that they had used antibiotics (namely ceftriaxone and azithromycin) before being admitted to the hospital [6]. In the present review, self-antibiotic medication, antibiotics prescribed by general practitioners, and empirical antibiotics prior to ICU admission were the reported risk factors of AR during COVID-19. A recent systematic review assessing the risk factors of AR from 2013 to 2019 reported a similar risk factor, which was current or previous exposure to antibiotics; however, other factors included sociodemographics and admission to hospital [52].

Strengths and Limitations
This review has many strengths as the first systematic review addressing the impact of COVID-19 on AR and the relevant risk factors, based on the analysis of the retrieved evidence from thirteen countries worldwide. Additionally, the data from recent studies conducted during COVID-19 were included in this review, which in turn provided up-todate data on the impact of the COVID-19 pandemic on AR, as well as relevant risk factors.
On the other hand, there are some limitations. First, a potential limitation was our approach of incorporating AR from various groups of patients from various countries to measure the resistance percentages of bacteria to different antibiotics. In this approach, high resistance in various healthcare settings may have balanced out. Nevertheless, given the observed patterns, it is highly probable that the accuracy of the gathered data was enough to display the overall situation. Second, the extracted and related AR data were measured by different laboratory procedures. However, although the guidelines of the Clinical Laboratory Standards Institute and standard disc diffusion were mostly employed in the studies, it is believed that the validity of the outcomes would not be affected. Third, the global generalization of the study findings was another potential limitation. Although there were 23 retrieved studies across 13 countries worldwide, the studies were from different settings, populations, and healthcare systems, which provided an overall view regarding the impact of COVID-19 on AR and related risk factors.

Conclusions
AR during COVID-19 was high, and the most common Gram-negative AR bacteria were A. baumannii, K. pneumonia; the most common Gram-positive AR bacteria were S. aureus and E. faecium. Although the colistin indicated a highly sensitive antibiotic, resistance of K. pneumonia had a median of 21%. Self-antibiotic medication, empirical antibiotic administration, and antibiotics prescribed by general practitioners were the risk factors of high levels of AR during COVID-19. Those prescribing antibiotics should strictly abide by the ASP and guidelines from the WHO and MOH, particularly during pandemics. Healthcare providers and people in the community need more awareness with respect to the proper uses of antibiotics, both during pandemics and in normal situations. Urgent support from policymakers and authorities is needed to issue more restrictions on the uses of antibiotics, more so than in the current situation.

Acknowledgments:
We would like to acknowledge everyone involved directly or indirectly in the preparation of this manuscript. Our gratitude goes to the librarian who helped in facilitating the online search strategy.

Conflicts of Interest:
The authors declare no conflict of interest.