Antimicrobial stewardship challenges during the COVID-19 pandemic: implications of antibiotic overuse and antimicrobial resistance

Patterns of antibiotic prescribing during COVID-19

Since its emergence in Wuhan in late 2019, SARS-CoV-2 rapidly spread worldwide and resulted in substantial global morbidity and mortality. Global excess mortality associated with the COVID-19 pandemic was estimated to about 18.2 million deaths between January 2020 and December 2021, reflecting both direct and indirect effects of the pandemic (11). Despite the viral aetiology of COVID-19, antibiotics were prescribed widely throughout the pandemic. Early COVID-19 recommendations promoted empirical antibiotic therapy because of concerns regarding bacterial co-infection and superinfection (12). Diagnostic uncertainty, overlapping clinical manifestations between viral and bacterial pneumonia, and early hypotheses regarding potential antiviral effects of macrolides further contributed to extensive empirical antibiotic prescribing in both hospital and community settings (13, 14). However, subsequent large, randomised trials demonstrated no clinical benefit of azithromycin-based regimens in COVID-19 patients (15).

Most empirical antibiotic regimens targeted community-acquired pneumonia and primarily consisted of ceftriaxone combined with azithromycin (16). Frequently prescribed antibiotics also included amoxicillin-clavulanate, piperacillin-tazobactam, macrolides, carbapenems, cephalosporins and fluoroquinolones (17). Carbapenem consumption increased markedly in some hospital settings. In one centre, it increased from a baseline of 67.1 to 142.9 defined daily doses (DDD) per 1,000 patient-days during the COVID-19 pre-intervention period, while intensive-care unit (ICU) consumption increased from 125.7 to 240.8 DDD per 1,000 patient-days. Following implementation of targeted stewardship measures, carbapenem consumption decreased by 49.5 % hospital-wide and by 36 % in ICUs (18). High-risk antibiotics associated with Clostridioides difficile infection (CDI), including clindamycin, fluoroquinolones, and third-generation cephalosporins, were increasingly prescribed during the pandemic (12). Broad-spectrum antibiotic prescribing became common despite the viral nature of the COVID-19 (19). Importantly, early antibiotic administration did not slow down disease progression, shorten the length of hospital stays, or lower mortality in most hospitalised COVID-19 patients without confirmed bacterial infection (17).

The substantial increase in antibiotic prescribing during the pandemic challenged antimicrobial stewardship programmes and raised concerns regarding the emergence and spread of multidrug-resistant organisms. Regional differences in prescribing practices were also observed. North America reported the highest prevalence of antibiotic use among COVID-19 patients (68.84 %, 95 % CI 62.27–75.05), followed by Europe (60.01 %, 95 % CI 25.50–89.67), whereas Asia reported the lowest prevalence (40.81 %, 95 % CI 7.75–79.65) (14). Although Europe experienced lower overall antibiotic consumption in 2020, likely due to reduced medical treatment activity and social distancing, antibiotic prescribing among hospitalised COVID-19 patients remained high (14). In the United States, outpatient antibiotic prescription also declined substantially in 2020, particularly for commonly used antibiotics such as amoxicillin, amoxicillin-clavulanate, doxycycline, and azithromycin (7). Antibiotic consumption patterns also changed following the acute pandemic period. After the marked reductions observed in 2020 and 2021, community antibiotic consumption in the EU/EEA rebounded in 2022. Total EU antibacterial consumption reached 19.4 DDD per 1,000 inhabitants per day, representing a significant increase compared to 2020–2021, although remaining approximately 2.5 % below the 2019 baseline (20).

The pandemic also altered broader antibiotic utilisation and resistance patterns. Longer hospital stays were observed during the pandemic (3.86 days in 2019 v 4.29 days in 2020), accompanied by greater use of duplicate anaerobic therapy (4.58 % in 2019 v 5.71 % in 2020) and greater prescribing of antibiotics for viral and fungal infections (17.08 % in 2019 v 22.38 % in 2020) (21). In addition, erroneous empirical antibiotic prescribing increased during the pandemic, while antimicrobial resistance (AMR) challenges persisted despite modest improvements in stewardship practices (21). These findings emphasise the need for strengthened antimicrobial stewardship programmes and continued surveillance to mitigate the long-term consequences of inappropriate antibiotic prescribing and AMR development.

Frequent empirical prescribing of antibiotics stood out against the relatively low prevalence of documented bacterial infections. In a comparative cohort study (22), bacterial infections were identified in only 8 % of hospitalised patients with COVID-19, compared to 13 % with non-COVID-19 pneumonia, and most bacterial infections among COVID-19 patients were nosocomial rather than community-acquired. The overwhelming number of hospitalised COVID-19 patients further amplified antibiotic consumption. One study (23) demonstrated a quadrupling of ceftriaxone use and a doubling of azithromycin use during the initial stages of the pandemic. One systematic review and meta-analysis (24) reported a pooled prevalence of bacterial co-infection of 8 % and bacterial superinfection of 20 %, while antibiotics were administered in 98 % of studies reporting antibiotic use. In a stewardship-focused intervention study including 833 hospitalised patients (25), antibiotics were prescribed to 438 (53 %) patients at enrolment. Among 429 patients allocated to the intervention group, 203 (47 %) received at least one antibiotic at enrolment, while 32 (7 %) more were prescribed antibiotics later during hospitalisation. A total of 301 prospective audit and feedback events were conducted, including 235 initial audits and 66 follow-up audits. Ceftriaxone (216 prescriptions) and azithromycin (167 prescriptions) were the two most audited antibiotics, in line with local recommendations for the management of hospitalised patients with community-acquired pneumonia. These findings are consistent with additional cohort and meta-analytic evidence demonstrating low rates of bacterial co-infection alongside substantial antibiotic over-prescription among hospitalised COVID-19 patients (26, 27). The main patterns of inappropriate antibiotic use, including route of administration, duration-related problems, and key stewardship concerns, are summarised in Table 1.

Bacterial co-infections and mismatch with antibiotic use

Current evidence suggests that bacterial co-infection rates in patients with COVID-19 were substantially lower than those observed during previous viral pandemics. Direct pre-pandemic comparisons support this observation. In a retrospective cohort study (22) comparing patients hospitalised with non-COVID-19 pneumonia in 2019 with patients admitted with COVID-19 and pulmonary infiltrates in 2020, microbiologically confirmed bacterial infections were identified in 13 % and 8 %, respectively (P<0.001). During the 2009 influenza A (H1N1) pandemic, bacterial pneumonia contributed to up to 55 % of fatalities (14). In contrast, one meta-analysis (26) conducted during the COVID-19 pandemic demonstrated that bacterial co-infections at hospital admission were relatively uncommon, with reported prevalence rates ranging between 3 % and 5 %, whereas secondary bacterial infections occurred more frequently among critically ill and mechanically ventilated patients. Despite these low rates of confirmed bacterial co-infection, antibiotics were prescribed to 60–75 % of hospitalised COVID-19 patients, constituting a substantial mismatch between microbiological evidence and prescribing practices. Although recognition gradually emerged that bacterial co-infections were uncommon in COVID-19, antibiotics continued to be prescribed at disproportionately high rates (25). Bacterial co-infection and secondary infection were also associated with adverse clinical outcomes. One systematic review and meta-analysis (24) found that COVID-19 patients with co-infection or superinfection had more than three times higher odds of death than patients with SARS-CoV-2 infection alone (OR 3.31; 95 % CI 1.82–5.99).

Several factors contributed to this discrepancy between confirmed bacterial infection and empirical antibiotic use such as pandemic-related healthcare burden, diagnostic uncertainty, concerns about bacterial superinfection, and disruption of established AMS activities (27). The central stewardship challenge was not the complete absence of bacterial infections, but rather the difficulty in distinguishing early viral pneumonia from true bacterial co-infection, particularly in severely ill patients, leading to extensive empirical antibiotic prescribing without microbiological confirmation (17).

In addition, one systematic review (28) demonstrated inadequate reassessment of empirical antibiotic therapy during hospitalisation and identified opportunities for improvement to address issues such as prolonged antibiotic treatment exceeding seven days in 37 % of patients receiving empirical antibiotics despite negative cultures, intravenous therapy exceeding 48 h, and reluctance to switch to oral therapy, ranging from only 9.9 % to 40.5 %, depending on reports covered by this review. It also revealed that antibiotic prescribing increased from 60.1 % at admission to 72.3 % after seven days of hospitalisation, even though bacterial co-infection was confirmed in only 1.2 % of patients.

Impact on antimicrobial resistance dynamics

The COVID-19 pandemic created a unique and highly complex epidemiological and clinical environment involving extraordinary pressure on healthcare systems, altered population behaviour, and substantial deviations in antibiotic prescribing practices. Under such circumstances, certain aspects of the pandemic response may have accelerated AMR development, while others temporarily exerted protective effects (29). The widespread empirical antibiotic prescribing in the early stages of the pandemic generated considerable antimicrobial selection pressure and raised concerns regarding long-term AMR consequences in hospital settings. In fact, the emerging evidence indicates increasing resistance to critically important antibiotics, particularly carbapenems and colistin (6, 29). Secondary bacterial infections were observed predominantly among critically ill and mechanically ventilated patients and were frequently caused by nosocomial multidrug-resistant pathogens, including Acinetobacter baumannii, Klebsiella pneumoniae, and Pseudomonas aeruginosa (26, 30). In some healthcare settings, increased incidence of carbapenem-resistant Enterobacterales was also reported (31). Collectively, these findings suggest that pandemic-related prescribing practices, prolonged hospitalisation, and increased ICU exposure may have accelerated the emergence and spread of multidrug-resistant organisms.

Besides, the pandemic disrupted infection prevention and control (IPC) activities and AMS programmes worldwide. Healthcare systems faced severe staff shortages, overwhelming workloads, and limited diagnostic capacity, all of which contributed to delayed microbiological confirmation, excessive empirical antibiotic use, and transmission of multidrug-resistant organisms within healthcare facilities (27, 32). In a cohort of hospitalised COVID-19 patients with documented bacterial or fungal co-infection mortality reached 57 %, highlighting the poor prognosis associated with clinically significant secondary infections (33). The pandemic therefore highlighted the critical importance of maintaining diagnostic capacity, microbiological surveillance, and rational antibiotic prescribing, reserved for confirmed or strongly suspected bacterial infections, even during public health emergencies (23, 32).

Structure and role of AMS teams during the pandemic

For multiple reasons, AMS became “a victim” of COVID-19, as major disruptions in clinical service delivery, global antibiotic supply chains, and the tendency toward overtreatment with antibiotics weakened established stewardship practices (19), particularly in regions where stewardship programmes were still developing and resource limitations already existed (14). The pandemic also disrupted ongoing antimicrobial stewardship programmes in high-income healthcare systems, highlighting the vulnerability of stewardship infrastructure (27).

Despite these challenges, AMS programmes played a critical role in optimising antibiotic use during the COVID-19 pandemic. Effective stewardship strategies included infectious disease specialist participation in clinical rounds, biweekly review of ongoing antibiotic therapy, dissemination of COVID-19 treatment guidelines through mobile applications, and prospective weekly audits (17). One carbapenem-focused AMS intervention utilised pharmacy alerts to infectious disease specialists following carbapenem prescription, with specialist consultation occurring within 72 h. This intervention reduced carbapenem use from 4.1 % to 2.3 % and was associated with lower in-hospital mortality and improved treatment success rates (17). Predictive models combining white blood cell count, procalcitonin levels, and comorbidity burden were also proposed as useful tools for identifying patients at low risk of bacterial co-infection and supporting rational antibiotic prescribing (17).

Prospective audit and feedback emerged as one of the most effective stewardship interventions during the pandemic and was shown to be safe and effective in optimising and reducing antibiotic use among adults hospitalised with COVID-19 (25). Chen et al. (25) reported high physician acceptance rate of stewardship recommendations (84 %) including discontinuation of antibiotic therapy when bacterial infection was not suspected or confirmed (57 %), modification of treatment duration (21 %), antibiotic spectrum narrowing (8 %), dose adjustment (7 %), and route modification (3 %).

The pandemic further bolstered preserving previously established stewardship achievements. Several AMS programmes implemented prior to COVID-19, including the Antibiotic Review Kit (ARK) and the Treat Antibiotics Responsibly, Guidance, Education, Tools (TARGET) toolkit, proved safe and effective in discontinuing antibiotics during acute hospital care (28). These findings highlight the need for sustained stewardship and its integration into future pandemic planning.

AMS implementation was strengthened through multidisciplinary collaboration involving pharmacists, infectious disease specialists, and microbiologists. Clinical pharmacists emerged as particularly important members of AMS teams during COVID-19. They ensured continuity of pharmaceutical supply, participated in the development and revision of COVID-19-specific treatment guidelines, and supported pandemic response systems by adapting existing stewardship programmes (19). Pharmacists also contributed through prospective audit and feedback activities, antimicrobial review, dose optimisation, de-escalation support, therapeutic drug monitoring and education of healthcare professionals regarding rational antibiotic use. Their unique expertise enabled tailored recommendations for appropriate antimicrobial prescribing and contributed significantly to minimising AMR development.

Antibiotic prescribing patterns after COVID-19 vaccination

The introduction of large-scale COVID-19 vaccination programmes lowered disease severity, hospitalisation rates, and intensive care unit admissions, which likely contributed to reducing unnecessary antibiotic exposure. Jorgensen et al. (34) report a significant reduction in outpatient antibiotic prescribing, particularly of antibiotics commonly prescribed for respiratory tract infections, while no significant effect was observed for antibiotic treatment of urinary tract infections. One meta-analyses (35) reports 34 % lower odds of antibiotic use (OR: 0.662; 95 % CI: 0.540–0.811) in vaccinated COVID-19 patients.

The available evidence on the association between COVID-19 vaccination and antibiotic prescribing is summarised in Table 2.

One Health implications of antibiotic overuse during the pandemic

Self-medication and uncontrolled access to antibiotics during the pandemic may have intensified global resistance pressure, particularly in low- and middle-income countries, increasing the risk of the rapid global spread of resistant microorganisms (27).

The implications of such antibiotic overuse are particularly evident in the disruption of the gut microbiome and increased incidence of CDI recorded among hospitalised elderly COVID-19 patients (12).

Furthermore, exposure to broad-spectrum antibiotics accelerated the emergence of multidrug-resistant organisms. One study (17) reported carbapenem resistance in hospital and intensive care settings in more than 69 % of Gram-negative isolates – Acinetobacter baumannii, Klebsiella pneumoniae, and Pseudomonas aeruginosa in particular. Widespread use of carbapenems, cephalosporins, and fluoroquinolones during the pandemic has raised serious concerns about long-term environmental effects and the amplification of AMR reservoirs (28).

Yet the consequences of excessive antibiotic use extend beyond antimicrobial resistance, as it has been associated with increased healthcare costs, adverse drug reactions, and prolonged hospitalisation (16). One study (21) reported an increase in healthcare costs from about $18,000 to $45,000, excluding staffing and other direct or indirect costs. Moreover, increased antimicrobial consumption must have strained global pharmaceutical supply chains and contributed to medication shortages worldwide (19).

One Health consequences, however, extend beyond patients and healthcare institutions, as antimicrobial resistance crosses human, animal, and environmental compartments (36). Antibiotics and their active metabolites are frequently excreted in urine and faeces and subsequently enter municipal wastewater and hospital effluents, whose residues cannot be completely eliminated by wastewater treatment and therefore persist in aquatic environments, facilitating interactions between human and environmental microbial communities (37, 38). Although direct evidence linking COVID-19-associated antibiotic consumption to subsequent environmental AMR remains limited, these established mechanisms call for the integration of wastewater surveillance, pharmaceutical waste management, and environmental AMR monitoring into broader One Health stewardship strategies (36,37,38).

Diagnostic stewardship and rapid molecular diagnostics

Rapid molecular diagnostics support antimicrobial stewardship by facilitating earlier differentiation between bacterial and viral pneumonia. One study in critically ill COVID-19 patients (39), reported an 89.3 % sensitivity and 99.1 % specificity of a polymerase chain reaction diagnostic test that prompted initiation or modification of antibiotic therapy in 15 % and discontinuation in 28 % of episodes.

The effects of infection prevention and non-pharmaceutical interventions on antimicrobial resistance

The pandemic produced several epidemiological effects that temporarily reduced antimicrobial demand. Non-pharmaceutical interventions, including social distancing, mask use, and better hand hygiene were associated with marked drops in viral and invasive bacterial infections (40, 41). These measures may also have reduced transmission of some healthcare-associated pathogens and mitigated the expected consequences of increased antibiotic exposure in certain settings (12).

In contrast, the widespread use of disinfectants and biocidal agents may also have contributed to selective pressure and potential cross-resistance among some microorganisms (42). These opposite effects further demonstrate that AMR trends cannot be understood from antibiotic consumption data alone and call for maintaining IPC measures alongside AMS during future health emergencies.

Implications for future research and pandemic preparedness

What we have learned from the COVID-19 pandemic is that future AMS strategies should focus not only on reducing antibiotic prescribing but also on improving its quality and timing. Priority should be given to rapid and accessible microbiological diagnostics, structured reassessment of empirical therapy, or early discontinuation when bacterial infection is not confirmed, optimisation of treatment duration, restriction of unnecessary broad-spectrum agents, and to the application of multidisciplinary AMS infrastructure during public health emergencies (17, 19, 25, 27, 29, 32).

Future research should evaluate which stewardship interventions remain most effective under crisis, how diagnostic tools can best distinguish viral disease from bacterial co-infection in severely ill patients, and whether reductions in antibiotic exposure translate into measurable long-term reductions in AMR. Additional prospective evidence regarding the effect of vaccination on antibiotic prescribing and its environmental consequences would help future pandemic responses to balance the immediate need for safe empirical treatment against the long-term need to preserve antimicrobial effectiveness.

Study limitations

This narrative review is limited by an inherent selection bias. The included studies may reflect early clinical practices that have changed since, whereas their designs, patient populations, and definitions of bacterial co-infection vary to such an extent that makes direct comparisons and synthesis nearly impossible. These limitations highlight the need for more high-quality, systematic studies to better understand the relationship between COVID-19 and antimicrobial resistance.