The clinical impact of bacterial co-infection among moderate, severe and critically ill COVID-19 patients in the second referral hospital in Surabaya

Background: Data on the prevalence of bacterial co-infections among COVID-19 patients are limited, especially in our country, Indonesia. We aimed to assess the rate of bacterial co-infections in hospitalized COVID-19 patients and report the most common microorganisms involved and the antibiotic use in these patients. Methods: This study is a retrospective cohort study,among COVID-19 adult patients admitted to Universitas Airlangga Hospital Surabaya from 14 March-30 September 2020. The bacterial infection is defined based on clinical assessment, laboratory parameters, and microbiology results. Results: A total of 218 patients with moderate to critical illness and confirmed COVID-19 were included in this study. Bacterial infection was confirmed in 43 patients (19.7%). COVID-19 patients with bacterial infections had longer hospital length of stay (17.6 6.62 vs 13.317.12), a higher proportion of respiratory failure, intensive care treatment, and ventilator use. COVID-19 patients with bacterial infection had a worse prognosis than those without bacterial infection (p<0.04). The empirical antibiotic was given to 75.2% of the patients. Gram-negative bacteria were commonly found as causative agents in this study (n = 39; 70.37%). Conclusion: COVID-19 patients with bacterial infection have a longer length of stay and worse outcomes. Healthcare-associated infections during intensive care treatment for COVID-19 patients must be carefully prevented.


Introduction
Coronavirus disease (COVID)-19 has experienced an increase in 2,995,758 positive cases and 204,987 deaths in distribution areas of more than 213 countries 1 . In Indonesia, until November 2020, there were 522,581 confirmed cases of COVID-19 with 68,604 active cases, 437,456 recovered cases, and 16,521 (3%) deaths. According to national data, among the total number of cases, East Java is the second-highest prevalence of 60,190 (11.6%). As of November 26, 2020, the cumulative data on confirmed COVID-19 patients in Surabaya were 16,763 with 1204 (7.03%) deaths 1 .
Data regarding secondary respiratory infection in COVID-19 patients are still limited in Indonesia even though the cases distribution is still worldwide. Several reports suggest that secondary infection will impact the patient's survival and increase intensive care unit (ICU) treatment 2 . COVID-19 pneumonia itself is also related to increasing ICU care, secondary infection rate, and higher invasive treatment 3 . Co-infection of SARS-CoV-2 and other microorganisms such as viruses, bacteria, and fungi is an essential factor in COVID-19 treatment since this condition may raise the difficulty in diagnosis, treatment, and the prognosis, also increasing the mortality 4,5 . Upper respiratory tract infection, a typical manifestation of COVID-19, is challenging to differentiate from other causes of pneumonia.
There are various types of co-infection in COVID-19; such as: 1) secondary SARS-CoV-2 following bacterial infection or colonization; 2) mixed infection between viral and bacterial pneumonia infection; 3) secondary bacterial superinfection following SARS-CoV-2 infection. The mechanisms underlying those types of infection are very dependent on the onset and involving complex interactions between three different agents (virus, host, and bacteria). Immune response towards SARS-CoV-2 infection only differs with mixed infection to bacteria or viral pneumonia. Therefore, it could be hypothesized that any co-infection will worsen the outcome and severity of COVID-19 6 .
Although several studies have investigated the epidemiological and clinical characteristic of COVID-19, information regarding SARS-CoV-2 infection with secondary infection are limited 7 . This study aims to describe bacterial co-infection and antibiotic use among patients who confirmed SARS-CoV-2 infection from moderate to critically ill manifestation in Universitas Airlangga hospital Surabaya.

Study design and population
This study is a retrospective cohort study, total sampling, among COVID-19 adult patients admitted to Universitas Airlangga Hospital Surabaya. This hospital is an academic hospital and also a referral hospital for COVID-19 management in Surabaya, East Java Region. We included cases of moderate to critically ill COVID-19 patients between 14 March and 30 September 2020 that were admitted in the intensive care unit or highcare unit. COVID-19 diagnosis was made based on World Health Organization (WHO) guidelines 8 and the Indonesian Ministry of Health guidelines 9 . Confirmed COVID-19 patients were proven by oropharyngeal and nasopharyngeal swabs SARS-CoV-2 PCR (polymerase chain reaction). Bacterial infection was defined based on clinical assessment, laboratory parameters, and inflammatory parameters (C-reactive protein (CRP) and procalcitonin). Bacterial co-infection of SARS-CoV-2 defines if the culture samples were taken at patient presentation to the hospital or < 48 hours admission, while secondary bacterial infection defined if the culture samples were taken > 48 hours of admission. The bacterial causative agents were extracted from data of microbiology, that were identified by Microbiology automated machine Vitek-2 compact, as a routine procedure in this hospital.
This study was approved by the Ethical Committee of Universitas Airlangga Hospital (171/KEP/2020). Written informed consent was obtained from all participants prior to the start of the study.

Data collection
Data were taken from medical records and microbiology reports from the laboratory. Incomplete medical records were excluded. Clinical characteristics were divided according to the severity of the disease. Moderate case definitions are: 1) clinically sign of pneumonia (fever, cough, dyspnea, tachypnea); 2) Oxygen saturation ≥93% free air. Severe case definitions are if there were clinically sign of pneumonia, and one of the following: 1) respiration rate >30 times per minute, or 2) severe respiratory distress, or 3) oxygen saturation < 93% free air. Critically ill cases defined if there were acute respiratory distress syndrome (ARDS), sepsis, and septic shock 10 . Culture examination was performed when there was suspicion of bacterial infection or sepsis. The sample for culture was taken from blood, urine, and respiratory tract.

Statistical analysis
Data were analyzed with SPSS version 24.0 (Chicago, IL, USA) 11 . Descriptive statistics included categorical variables reported as number (percentage) and continuous variables as mean (standard deviation). For missing data, we used listwise deletion or univariable and multivariable analysis. Chi-square test and Mann-Whitney test were used accordingly to the type of variable. Categorical variables were shown as number (%) and continuous variables as mean (standard deviation) or median (range) depending on whether the data are normally distributed

Amendments from Version 1
In the introduction, we have emphasized the fact that the research on secondary infection accompanying COVID-19 in Indonesia is still limited and there is no prior investigation linked to this topic ever done before, by adding "Indonesia" in our first sentence of paragraph 2.
We also made minor revisions by correcting the "leukocyte" to "leucocytes" and "leukocytosis" to "leucocytosis" We also insert citation for SPSS version 24.0 Any further responses from the reviewers can be found at the end of the article From the laboratory findings, the mean leucocyte count (p = 0.014), neutrophil count (p = 0.001), procalcitonin (p = 0.033), basal urea nitrogen (BUN) (p = 0.007), alanine aminotransferase (ALT) levels (p = 0.01) were significantly higher in either severe or critically ill COVID-19 patients than moderate cases. We also found a lower lymphocyte count (p = 0.001) and PaO 2 : FiO 2 ratio (p = 0.005) in either severe or critically ill COVID-19 patients than moderate cases. The procalcitonin level significantly increased in severe and critical illness conditions. The CRP level also increased but was not statistically significant. The majority of patients (82.9%) were recovered and discharged from the hospital, while 9.6% of the patient died.
Most of the patients who died were in critically ill condition at presentation to hospital.

Bacterial infections
Bacterial infection was confirmed in 43 patients (19.7%) (see Table 2). There were no sex differences between bacterial infection and no bacterial infection patients. Patients with a bacterial infection have an older mean age than no bacterial infection, although, among elderly patients, there were no differences in bacterial infection rate. We documented one patient with chronic obstructive pulmonary disease (COPD) and one with liver disease as a comorbidity, and both of them suffered from bacterial infection. Other comorbidities such as diabetes,  hypertension, heart disease, stroke, chronic kidney disease did not differ between the two categories. COVID-19 patients with bacterial infections had a longer hospital length of stay, a higher proportion of respiratory failure, ICU treatment, ventilator use.
Mortality occurs in 16.28% of COVID-19 patients with bacterial infections, higher than those without bacterial infection (8%). Overall, COVID-19 patients with bacterial infection had a worse prognosis than those without bacterial infection (p<0.04).

Our study focuses on bacterial infection results in COVID
In this study, inflammation signs such as leucocyte count, neutrophil count, procalcitonin in the critically ill group were higher than the others. Lymphopenia in COVID-19 patients occurs through various mechanisms such as direct virus invasion lymphocytes, lymphatic organ destruction, altered inflammatory cytokines production leading to lymphocyte apoptosis, and inhibiting lymphocytes function by metabolic molecules produced by metabolic disorders, such as hyperlactic academia 16 . Neutrophil and lymphocyte ratio (NLR) has been proposed as a prognostic marker of severity in various chronic inflammatory diseases, including cardiovascular diseases and oncological processes 17 .
Our result showed a significant increase of procalcitonin level and a higher proportion of bacterial infection in severe and critically ill COVID-19 patients. This finding is consistent with the study result from Wang et al. in elderly COVID-19 patients. They also concluded that bacterial infection was also considered as a predictor of mortality in these patients 18 . Increased procalcitonin level in COVID-19 patients may be associated with the release of some cytokines, especially IL-6. It is established that procalcitonin is a better marker to predict severity, prognosis, or the sepsis course and is also helpful to guide antibiotic usage. Increased procalcitonin in critically ill patients with COVID-19 can represent a bacterial co-infection, and blood cultures for bacteria detection are needed to a prompt response 19 .
In this study, we found that bacterial infection was confirmed in 43 patients (19,7%  The most common bacterial infection in this study was from respiratory tract infection, followed by bloodstream infection and urinary tract infection. The culture result was dominated by gram-negative bacteria (75.68%). This finding is similar to the result from Zhang et al. (50%), although they also included virus and fungal cultures 3 . Gram-negative bacteria were reported responsible for more than 30% of healthcareassociated infections, and these bacteria predominate in cases of ventilator-associated pneumonia (47%) and urinary tract infections (45%) 24 . In intensive care units (ICUs), gram-negative bacteria account for about 70% of these types of infections 26 . Gram-negative bacteria are highly efficient at up-regulating or acquiring genes that code for antimicrobial resistance mechanisms, especially in the presence of antibiotic selection pressure. Among gram-negative bacteria, the Enterobacteriaceae family being the most commonly found and multidrug-resistant organisms, including Pseudomonas aeruginosa, Acinetobacter baumannii, and extended-spectrum β-lactamase (ESBL)producing or carbapenemase-producing Enterobacteriaceae, are increasingly being reported worldwide 24 . The identification of bacterial infection with gram-negative organisms is more reflected as a complication of ICU care and is not suggested as a specific predilection for co-infections in COVID-19 23 .
Acinetobacter baumannii is the most common bacteria found in this study. Several reports regarding bacterial infection in COVID-19 also found a similar result 3,7,9,19,24 . Acinetobacter baumannii is the common cause of respiratory tract infection, especially in ICU settings where patients often received mechanical ventilation. Environmental contamination also plays a role in this phenomenon. Our ICUs for COVID-19 patients were set as a large ward consist of 16 beds. Although we already managed the distance between beds for more than 1.5 meters, the inter-patient transmission is hard to avoid. It is crucial to maintain hand hygiene compliance of the staff, strict cleanliness, and disinfection of the hospital environment in the hospital, especially in high-care or ICU setting.
Antibiotic is often given to COVID-19 patients for various reasons. The clinical manifestation on presentation was sometimes challenging to distinguish with bacterial infection 2 . Secondary infection and nosocomial infection are also other considerations in the use of antibiotics. In this study, 75.3% of patients included were given antibiotics. A meta-analysis conducted about bacterial co-infection and secondary infection in COVID-19 patients reported that the majority of patients with COVID-19 received antibiotics (71.9%, 95%CI 56.1 to 87.7%) 27 . Another meta-analysis reported that >90% of COVID-19 patients were given an empirical antibiotic, while the bacterial infection was detected only in 7% of hospitalized patients and 14% of ICU patients 28 . Quinolones (levofloxacin and moxifloxacin) were frequently prescribed, followed by cephalosporins and carbapenems. Indonesian national guidelines for COVID-19 treatment recommend using antibiotics in severe and critically ill patients, especially if there was suspicion of bacterial infection. The empirical antibiotic choice was intravenous azithromycin or levofloxacin 10 . In our hospital, levofloxacin is more feasible than azithromycin, so clinicians prefer to prescribe this drug. In our hospital, antibiotic use usually was decided based on clinical signs and laboratory parameters (leukocytosis, increased CRP, or procalcitonin). Culture examinations were not routinely performed, preferably in ICU/high-care settings and in patients showing bacterial infection.
The widespread use of antibiotics in this pandemic era raises concern about antimicrobial resistance 4,6 . Although bacterial infection in COVID-19 patients has already been reported in various results, evidence supports the restrictive use of antibiotics. Clinical guidelines and standard testing to diagnose bacterial infection in COVID-19 are not clearly available. The microbiological examination is an important strategy to confirm bacterial infection and decide the antibiotic choice 4 . Sputum, blood culture samples, and also pneumococcal urinary antigen testing should also be performed.

Erni J. Nelwan
Division of Tropical and Infectious Disease, Universitas Indonesia, Jakarta, Indonesia A well written article and proper data collection to support the conclusion. From a LMIC like Indonesia, this article shares important findings, since the co-bacterial infection is not more than 20%, higher than reported in IDSA data but lower compared to what had been expected in clinical practice among Indonesian doctors.
Authors might add value by putting a highlight on what is the clinical appearance of the cobacterial infected vs. the non co-bacterial infected for readers to apply in daily practice.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes © 2021 Ekawati F. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Fitriana Murriya Ekawati
Department of Family and Community Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia Thank you very much for the opportunity for reviewing this manuscript. This manuscript has presented interesting findings of bacterial coinfection in COVID-19 patients. Data collected for the research were from patients' medical records in Univ Airlangga hospital in Indonesia. The presentations of Methods, Results, and Discussion in the manuscript are appropriate. However, there are some minor revisions needed, which I believe are beneficial for improving the manuscript: Introduction ○ -Line 1. I would recommend the authors to improve the first sentence of the first and second paragraph to improve clarity. -I would recommend a language proofreading for this manuscript.
Best wishes to the authors for the manuscript revision.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com