Overuse of empirical antibiotics in a COVID-19 intensive care unit led to the spread of carbapenem-resistant Gram-negative bacteria in a teaching hospital.

Sir,

Bacterial resistance is a global concern. The alarming annual rate estimated for 2050 of 10 million deaths has been increased with the global COVID-19 pandemic [1]. Treatment of multidrug-resistant Gram-negative bacilli (MDR-BGN) infections, particularly carbapenemase-producing Klebsiella pneumoniae (CP-Kp) carrying either bla NDM or bla KPC genes, or both, has become a major challenge for physicians.

Since 2011, our hospital has maintained a surveillance program that reduced the incidence of nosocomial infections from about 5.3% to about 3%. This lower rate remained until the beginning of the SARS-CoV-2 pandemic [2]. In May 2020, a new intensive care unit (ICU) was opened for COVID-19 patients. Toward the end of the year, an unexpected increase of CP-Kp occurred, mainly in this unit. A gradual increase in CP-Kp isolation rates was observed in 2019, 2020 and 2021; 55, 79 and 291 isolates were identified, respectively (an increase of more than 400%).

In the present study, we assessed the possible causes and consequences of rapid increase in CP-Kp carrying either bla NDM or bla KPC, or both, and the dissemination of these genes among MDR-BGN in a COVID-19 ICU. All SARS-CoV-2 positive patients (RT-qPCR-method) included in this study were admitted to the COVID-19 ICU and had a positive culture for CP-Kp in a rectal swab, clinical sample or both.

Bacterial identification and antimicrobial susceptibility testing were performed using the BD-Phoenix automated system (Becton-Dickinson, Franklin Lakes, NJ). All CP-Kp isolates were tested by the polymerase chain reaction (PCR) method for carbapenemase genes (bla KPC, bla NDM, bla OXA-48) and typified by the Enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) method. The band profile generated was analysed with BioNumerics version 6.5 software (Applied Maths, Sint-Martens-Latem, Belgium). Dice similarity coefficient ≥0.93 classified isolates as belonging to the same cluster.

From September 2020 to April 2021, a surge of CP-Kp was detected. Fifty-two CP-Kp isolates were recovered from 41 patients in the COVID-19 ICU. All isolates showed 100% genetic similarity by ERIC-PCR and were found to carry at least one carbapenemase (39 isolates carried bla KPC, 10 carried bla NDM, and 3 carried both genes).

Among 41 patients evaluated, 19 had CP-Kp infection (bloodstream, urinary tract and ventilator-associated pneumonia), and 13 of them had CP-Kp with other microorganism infections, which were considered superinfections (Fig. 1 ) [3].

Fig. 1

Schematic of cases of a dissemination of carbapenemase-producing Klebsiella pneumoniae (CP-Kp) in a COVID-19 unit. aTemporal representation of cases (weekly). ICU, intensive care unit; PU, SARS-CoV-2-positive patient with positive culture for CP-Kp in COVID-19 ICU; P, SARS-CoV-2-positive patient in COVID-19 medical clinic; EO, patient who was transferred from the COVID-19 ICU to another unit when tested negative for SARS-CoV-2; c, patients tested positive for SARS-CoV-2; d, died; CTR, ceftriaxone; CLI, clindamycin; AZI, azithromycin; PIT, piperacillin + tazobactam; FLC, fluconazole; CTZ, ceftazidime; CIP, ciprofloxacin; MER, meropenem; VAN, vancomycin; LEV, levofloxacin; POL, polymyxin B; TRS, trimethoprim-sulfamethoazole; MCF, micafungin; COL, colistin; TIG, tigecycline; AMI, amikacin; LIN, linezolid; CZO, cefazolin; ACY, acyclovir; IMI, imipenem; AMP, ampicillin. OMO, other microorganisms: Candida sp. (PU1, PU6, PU9, PU14, PU20, PU28, PU40); Stenotrophomonas sp. (PU2, PU12, PU14, PU22); Pseudomonas aeruginosa (PU4, PU13, PU15, PU31); Acinetobacter baumannii (PU4, PU7, PU14, PU20, PU22, PU23, PU28, PU29, PU32, PU35, PU36, PU38, PU39, PU40); A. baumannii carrying blaKPC (PU20) and blaNDM (PU29); Burkholderia sp. (PU7, PU9); Staphylococcus sp. (PU9, PU10, PU18, PU19, PU35); Enterococcus sp. (PU11, PU24, PU33, PU39); RS, rectal swab; ETA, endotracheal aspirate (>106 CFU); urine (>104 UFC - urinary catheter).

Among the 19 infected patients, the evaluation of treatment could be performed in 10 patients (47% died before the start of treatment for CP-Kp infection) and revealed that antibiotic combination was used in 9/10 with 56% therapeutic success (five patients survived). Although new antimicrobials with activity against CP-Kp have been approved by the European and North American international agencies [4], our hospital is a Brazilian public hospital where these new drugs are not standardized.

Among the five patients who survived, 80% used polymyxin B plus meropenem associated or not with amikacin, being the best combinations for this study. In a recent guideline from the European Society of Clinical Microbiology and Infections Diseases for the treatment of infections caused by MDR-BGN, Paul et al. (2022) concluded with low-certainty evidence an advantage of the combination of meropenem-polymyxin therapy over polymyxin monotherapy in the treatment of infections caused by CP-Kp. Our data, despite being from a low number of patients, allowed us to agree with them [4].

All COVID-19 patients received antibiotic therapy prior to colonization or infection by CP-Kp, with a mean of five different antimicrobials per patient (Fig. 1). Antimicrobial empiric overuse during the SARS-CoV-2 pandemic favoured a significant increase of the carbapenemase-producing BGN. Furthermore, there was a spread of these genes to endemic microorganisms in our hospital such as A. baumannii (PU20) and (PU29).

In our study, in addition to the use of antimicrobials, all patients used glucocorticoids, possibly linked to a high frequency of superinfections, which may have been associated with high mortality. Patients who require broad-spectrum empirical antibiotics and immunomodulatory drugs such as COVID-19 patients are at greater risk of acquiring superinfections [3,5].

With the advance of the pandemic, in our institution as well as in others, the addition of hospital beds and the hiring of new health workers were required, which may have influenced the rapid dissemination of MDR-BGN in this COVID-19 ICU [5]. The experience described here has shown that the darkest side of this pandemic was the high mortality associated with empirical overuse of antimicrobials and the lack of a trained infection-control team to care for COVID-19 patients. We conclude that we must increasingly invest in training, both in management of antimicrobials and in healthcare infection control, so that in the post-pandemic period, antimicrobial stewardship and surveillance programs remain active, minimizing the emergence and spread of MDR-microorganisms.

Declaration of Competing Interest

None declared.