Bacterial coinfection and antibiotic resistance in hospitalized COVID-19 patients: a systematic review and meta-analysis

Introduction

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2, 2019-nCoV) is a novel coronavirus that causes the disease known as coronavirus disease 2019 (COVID-19), which primarily affects the respiratory system (Yuki, Fujiogi & Koutsogiannaki, 2020), and the majority of viral respiratory tract infections are linked with microbial coinfection (Langford et al., 2020). Coinfection with a virus, bacteria, or fungus significantly impacts how COVID-19 infection develops and spreads, complicating COVID-19 diagnosis, treatment, and prognosis and even worsening disease symptoms and mortality (Chen et al., 2020).

A previous meta-analysis involving 20 published studies from December 2019 to June 2021 estimated the pooled prevalence of bacterial coinfection in patients with COVID-19 was 5.62% (95% CI [2.26–10.31]) (Alshaikh et al., 2022). The coinfections frequently occurred in patients hospitalized for an extended period, aggravated respiratory viral infections, and frequently resulted in fatalities during influenza pandemics (Westblade, Simon & Satlin, 2021). It was similar for COVID-19 patients in a multicenter retrospective cohort study in the United States, where the coinfections were more common in critically ill COVID-19 patients (Langford et al., 2020), particularly those who required invasive mechanical ventilation (Petty et al., 2022).

The most common bacteria isolated from co-infected COVID-19 patients were Streptococcus pneumoniae, Staphylococcus aureus, and Haemophilus influenzae (Moreno-García et al., 2022). Furthermore, a review reported that the most frequent bacteria were coagulase-negative Staphylococcus isolated from blood, Streptococcus pneumoniae from the respiratory tract, and Escherichia coli from urinary tract samples (Santos et al., 2022). Another review (Abu-Rub et al., 2021) evaluated the presence of bacteria such as Klebsiella pneumonia, Pseudomonas aeruginosa, and Acinetobacter baumannii co-infected in an intensive care unit (ICU) COVID-19 patients.

According to literature analysis, patients with COVID-19 admitted to the ICU had an unexpectedly high frequency of infections caused by methicillin-resistant Staphylococcus aureus (MRSA), carbapenem-resistant Acinetobacter baumannii (CRAB), carbapenem-resistant Enterobacteriaceae (CRE), and Candida auris (Segala et al., 2021). Meanwhile, a review discovered that over half of ICU patients with SARS-CoV-2 infection were prescribed antibiotics, significantly higher than the estimated frequency of identified bacterial co-infection (Abu-Rub et al., 2021). The review also stated that the WHO’s guidance on the clinical management of COVID-19 alarmed that antibiotic overuse increases the risk of multidrug-resistant organisms (MDROs) emerging and spreading.

This review aimed to reveal the pooled prevalence of bacterial coinfection and antibiotic resistance in hospitalized COVID-19 patients. This information integrates existing information to determine if scientific conclusions are consistent and generalizable across populations and situations or whether findings significantly differ by subgroup. It also aids healthcare shareholders in making decisions because the pooled prevalence increases the power and precision of effect estimations. It also included the prevalence of isolated bacteria related to bacterial coinfection and the prevalence of antibiotic resistance related to bacterial coinfection. However, only certain bacteria and selected antibiotic resistance bacteria were involved in this review.

Materials and Methods

Results

Study selection

Databases and secondary citations yielded a total of 2,451 studies. Titles and abstracts screened were 1,857 after removing duplicate studies. Then, 265 studies were retrieved from databases and secondary citations and evaluated for eligibility. In this review, only 49 studies that matched the inclusion and exclusion criteria were considered (Fig. 1).

Publication bias

The funnel plots showed that the distributions were asymmetry for all plots (Figs. 2 and 3). The distributions were concentrated on the right side of the plots. Egger’s test proved that there were small studies effects by significant p-values for all variables (Table 4).

Table 4:

Egger’s test.

Egger test	Coefficient	(95% CI)	p-value
Bacterial coinfection	11.21	[8.08–14.24]	<0.001
Subgroup year
2020	5.57	[2.92–8.21]	0.007
2021	9.76	[7.02–12.50]	<0.001
2022	14.81	[5.76–23.86]	0.003
Subgroup wards
ICU	10.89	[3.40–18.39]	0.007
Unspecific wards	10.67	[7.29–14.05]	<0.001
Subgroup regions
Africa	NA		NA
America	12.37	[−16.19 to 40.94]	0.203
Asia	12.66	[6.89–18.42]	<0.001
Europe	10.74	[7.04–14.44]	<0.001
Type of bacteria
Acinetobacter baumannii	6.11	[4.32–7.90]	<0.001
Escherichia coli	2.85	[1.83–3.86]	<0.001
Klebsiella pneumoniae	5.54	[4.18–6.89]	<0.001
Pseudomonas aeruginosa	3.75	[2.61–4.90]	<0.001
Staphylococcus aureus	3.40	[1.69–5.11]	<0.001
Antibiotic resistance bacteria
Methicillin-resistant Staphylococcus aureus	2.49	[1.33–3.64]	<0.001
ESBL-producing Enterobacterales	4.20	[0.55–7.86]	0.028
Carbapenem-resistant Acinetobacter baumannii	4.50	[2.25–6.74]	0.001
Carbapenem-resistant Enterobacteriaceae	2.89	[2.00–3.79]	<0.001
Carbapenem-resistant Pseudomonas aeruginosa	2.49	[1.21–3.77]	0.003
Vancomycin-resistant Enterococcus	1.83	[0.67–2.99]	0.008

DOI: 10.7717/peerj.15265/table-4

Discussion

In this review of 49 studies, the pooled prevalence of bacterial coinfection in hospitalized COVID-19 patients was 26.84% (95% CI [23.85–29.83]). It was the highest compared to review studies published in 2020 (prevalence = 6.9% (95% CI [4.3–9.5])) (Langford et al., 2020) involving COVID-19 patients from 24 studies. Another review showed a prevalence of 7% (95% CI [3–12]) involving hospitalized COVID-19 patients from 30 studies (Lansbury et al., 2020). Meanwhile, a meta-analysis published in 2021 reported the prevalence of bacterial coinfection increased to 19% (95% CI [14–25]), and the prevalence of bacterial superinfection was 24% (95% CI [19–30]) (Musuuza et al., 2021) which pooled from 118 studies. Another meta-analysis in 2021 also reported the pooled estimated prevalence of bacterial coinfection in COVID-19 patients from 42 studies was 20.97% (95% CI [15.95–26.46]) (Soltani et al., 2021). The pooled prevalence of bacterial coinfection in COVID-19 patients for review of 25 studies published in 2022 was 24% (95% CI [8–40]) (Kariyawasam et al., 2022). There was an increasing pattern of prevalence from the year 2020 to 2022. Another factor that may influence the high prevalence of bacteria coinfection in this review was the consideration of bloodstream infection, secondary infection, and bacterial superinfection as bacterial coinfection.

A significantly increasing trend of bacterial coinfection prevalence was also shown by year subgroup analysis from 13% in 2020 to 24% in 2021 and higher increased in 2022 to 34% in this review. However, this situation might be influenced by the under-reporting of the incidence due to limitations of laboratory resources, especially during the early phase of the COVID-19 pandemic (Petty et al., 2022) or a limited number of studies investigating the coinfection (Jeong et al., 2022). The coinfection rate also differs due to geographical regions, study duration (Wee et al., 2020), variation in case definition (Baskaran et al., 2021), or differences in diagnostic approaches used for coinfection detection (Moreno-García et al., 2022).

Bacterial coinfection was higher in ICU COVID-19 patients compared to COVID-19 patients from unspecific wards. This result was similarly reported by a previous meta-analysis that compared ICU and non-ICU COVID-19 patients (Kariyawasam et al., 2022). According to a study from Saudi Arabia, ICU COVID-19 patients had higher levels of bacterial coinfection found in their blood and sputum, which may be related to their more extended stays in the ICU and increased use of catheters like endotracheal, arteriovenous, and urinary tubes (Bazaid et al., 2022). Another study (Soriano et al., 2021) discovered that ICU-acquired infections were high (51.2%), with the respiratory tract, bloodstream, urinary tract, soft tissue, and abdominal focus being the most common sites of infection.

Based on regions subgroup analysis, most of the studies were from Asia, with America having the largest number of participants, and the region’s highest prevalence of bacterial coinfection was Europe. In contrast, another review Alshaikh et al. (2022) found that America had the highest prevalence of bacterial coinfection, followed by Asia and Europe. According to the report, the prevalence may be influenced by potentially higher rates of microbial lab investigation and excessive interpretation of laboratory data. Additionally, variations in population, regions, and environment, access to care, and infection control and prevention practices may impact the frequency of bacterial coinfections (Langford et al., 2020).

In hospitalized COVID-19 patients, this review pooled the prevalence of isolated bacteria for Acinetobacter baumannii (23%), Escherichia coli (11%), Klebsiella pneumoniae (19%), Pseudomonas aeruginosa (11%) and Staphylococcus aureus (12%). All these bacteria were associated with nosocomial infection (Bazaid et al., 2022). Gram-negative was more prominent isolated in COVID-19 patients, and Acinetobacter baumannii has been identified as the primary pathogen in the patient’s respiratory tracts (Bazaid et al., 2022). Acinetobacter baumannii coinfection secondary to SARS-CoV-2 infections has been the subject of numerous reports, with an incidence rate of up to 1% of hospitalized COVID-19 patients from various regions (Rangel, Chagas & De-Simone, 2021). It has also been reported that Acinetobacter baumannii was frequently found in ICU COVID-19 patients (Sharifipour et al., 2020) with a high mortality rate and a more extended stay in the ICU (Ceparano et al., 2022). Meanwhile, there were high rates of Staphylococcus aureus isolated from respiratory tract samples in COVID-19 patients with ventilator-associated pneumonia and COVID-19 patients with community-acquired bacterial coinfections (Wu et al., 2022).

High prescription of antibiotics was a risk factor for antibiotic resistance. The prevalence of antibiotic prescription in COVID-19 patients was estimated as high as 74.6% (95% CI [68.3–80.0]) with an odd ratio increased in age years and mechanical ventilation used, but the estimated prevalence of bacterial coinfection in the review was low (8.6% (95% CI [4.7–15.2])) involving 31 studies (Langford et al., 2021). A review (Abu-Rub et al., 2021) also discovered a significant pattern of antibiotic prescription in the ICU settings for COVID-19 patients, along with much lower detection rates for bacterial identification. Acinetobacter baumannii, Klebsiella pneumonia, Escherichia coli, and Pseudomonas aeruginosa were the most frequently reported antibiotic-resistant Gram-negative bacteria during the COVID-19 pandemic. At the same time, Staphylococcus aureus and Enterococcus faecium were the most commonly encountered Gram-positive bacteria, indicating that antibiotic resistance was widespread during the pandemic (Sulayyim et al., 2022). The study also found that resistance of ESBL-producing Escherichia coli was 75%, 100% of Staphylococcus aureus and Coagulase-negative Staphylococci were resistant to methicillin, and, Acinetobacter baumannii and Klebsiella pneumonia were 91.7% and 76.6% respectively, carbapenem-resistant.

Particularly in COVID-19 patients with high rates of antimicrobial use and low rates of secondary or coinfection, the prevalence of multidrug-resistant organisms has rapidly risen (Lai et al., 2021). This review pooled the prevalence of antibiotic-resistant bacteria in ESBL (15.2%), CRAB (14.6%), CRPA (7.0%), MRSA (5.1%), CRE (5.0%), and VRE (1.3%). The difficulty differentiating between respiratory viral infection and bacterial coinfection led to the widespread use of broad-spectrum antibiotics. It has also been suggested that the severity of the clinical conditions in an environment with a high prevalence of bacteria that produce ESBL may have contributed to the widespread use of carbapenems in empirical therapy of COVID-19 infection (Rusic et al., 2021).

A review reported that the rate of Acinetobacter baumannii and Klebsiella pneumoniae carbapenem resistance significantly increased in 2020 compared to isolates from the pre-COVID-19 era, also a significant rise in polymyxin B resistance, particularly for Klebsiella pneumoniae isolates (Wu et al., 2022). According to several studies (Abdollahi et al., 2021; Eckardt et al., 2022; Shinohara et al., 2022), a CRAB strain had broken out in ICU COVID-19 wards due to cross-contamination between ICU equipment like ventilators, infusion pumps, and hemodialysis machines with the ICU equipment and central lines most likely contributed to the hospitalized COVID-19 patients’ five times higher prevalence of MRSA bacteremia (Aslam et al., 2022). MRSA and VRE were the most commonest detected among the Gram-positive antimicrobial resistance identified in Covid-19 patients (Kariyawasam et al., 2022). Therefore, coinfections with multidrug-resistant bacteria may result in COVID-19 patients having more extended hospital stays, extra costly treatment, and less desirable clinical outcomes (Senok et al., 2021).

This review provided the most recent pooled prevalence of bacterial coinfection and antibiotic resistance in hospitalized COVID-19 patients by including published publications from 2021 to 2022, as opposed to previous reviews that only had articles published in 2021. These findings should increase the consistency and accuracy of policymakers’ conclusions. However, the validity of these findings depended on the validity of the results of included studies.

As a limitation of this review, the regions subgroup may not represent the region since not all geographical parts were included in this study. For example, the African region was only represented by a study in Egypt and had the lowest prevalence of bacteria coinfection compared to other regions, with a greater number of included studies. Next, asymmetry funnel plots and significant Egger’s tests detected publication bias, indicating small-study effects. Therefore, the prevalence in this review may overestimate compared to the true value (van Aert, Wicherts & van Assen, 2019).

Conclusions

Bacterial coinfection that includes bloodstream infection, secondary infection, and bacterial superinfection from published studies between January 2020 and September 2022 showed that the pooled prevalence was 27%. The subgroup analysis showed that the prevalence was highest in 2022, in the ICU and Europe, with the largest study participants. Meanwhile, Acinetobacter baumannii had the highest isolated bacteria prevalence compared to Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Staphylococcus aureus. The prevalence of antibiotic-resistant bacteria in hospitalized COVID-19 patients was 1.2%, and ESBL-producing Enterobacteriaceae contributed the highest prevalence compared to MRSA, VRE, CRE, CRAB, and CRPA. The findings of this review should be interpreted with caution due to their limitations. However, this review provided some evidence of bacterial coinfection in COVID-19 patients. This information might be valuable in developing an infection management and prevention strategy for COVID-19 patients.

Supplemental Information