key: cord-0769308-p4emro9e
authors: Rakiro, Joe; Shah, Jasmit; Waweru-Siika, Wangari; Wanyoike, Ivy; Riunga, Felix
title: Microbial co-infections and super-infections in critical COVID-19: A Kenyan retrospective cohort analysis
date: 2021-10-04
journal: IJID Regions
DOI: 10.1016/j.ijregi.2021.09.008
sha: 6c18aa36fa401d5b7e68db6258c40e91a2568c84
doc_id: 769308
cord_uid: p4emro9e

Objectives To outline the burden, risk factors and outcomes for critical COVID-19 patients with co-infections or superinfections. Methods This was a retrospective descriptive study of adults who were admitted with critical COVID-19 for ≥ 24 hours. Data collected included demographic profiles and other baseline characteristics, laboratory and radiological investigations, medical interventions, and clinical outcomes. Outcomes of interest included presence or absence of co-infections and superinfections and in-hospital mortality. Differences between those with and without co-infections and superinfections were compared for statistical significance. Results We reviewed 321 patient records. Baseline characteristics included a median age (IQR) of 61.4 (51.4-72.9) years, male (71.3%) and African/black predominance (66.4%). Death occurred in 132 (44.1%) patients with significant difference noted between those with added infections (58.2%) compared to those with none (36.6%) (p = 0.002, odds ratio (OR) = 2.41). One patient had co-infection with pulmonary tuberculosis. Approximately two-thirds of patients received broad-spectrum antimicrobial therapy. Conclusion Added infections in critically ill COVID-19 patients were relatively uncommon but where present, were associated with higher mortality. Empiric use of broad spectrum antimicrobials was common and may have led to selection of multidrug resistant organisms. More robust local data on antimicrobial susceptibility patterns may help in appropriate antibiotic selection to improve outcomes without driving up rates of drug resistant pathogens.

The COVID-19 pandemic has resulted in more than 220 million infections and 4.5 million deaths, posing unique challenges to healthcare systems globally (Ouma, Masai et al. 2020 , Vaillancourt and Jorth 2020 , Moynihan, Sanders et al. 2021 . Determining the optimal therapeutic approach for COVID-19, including the role of antibiotics, continues to be an area of active research. Use of antibiotics such as Azithromycin as therapy for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection or as empiric treatment for suspected co-infections and superinfections has been extensive (Cavalcanti, Zampieri et al. 2020 , Furtado, Berwanger et al. 2020 . However, there is limited data to support this practice (Ginsburg and Klugman 2020, Hsu 2020) . A recent review of COVID-19 management in 10 African countries, including Kenya, was undertaken by Adebisi and others. They found that use of empiric or prophylactic antibiotics was generally recommended by local guidelines, contrary to World Health Organization (WHO) guidelines (Adebisi, Jimoh et al. 2021) . In another review, upto 75% of patients with COVID-19 received empiric antimicrobial medications (Langford, So et al. 2021) . This widespread empiric antimicrobial use has raised concerns regarding increased antibiotic resistance and an upsurge in the rates of multidrug resistant organisms.

Antimicrobial use has partly been driven up by the concern that co-infections and superinfections in patients with COVID-19 leads to poor outcomes (Musuuza, Watson et al. 2021) . A recent study by Lansbury and others demonstrated that the rate of co-infection in patients hospitalized with COVID-19 pneumonia was low (7%) except for those admitted to the intensive care unit (ICU) (51%). Superinfection was associated with a higher mortality (75%) in comparison to noninfected patients (44%) (Lansbury, Lim et al. 2020) . This is comparable to higher mortality rates seen with viral-bacterial co-infections in community acquired pneumonia (Voiriot, Visseaux et al. 2016) . Other studies also point to the low incidence of co-infections in patients admitted with COVID-19 (Garcia-Vidal, Sanjuan et al. 2020 , Hughes, Troise et al. 2020 . This is in contrast to observations made during outbreaks of influenza, where rates of bacterial co-infections of upto 65% were reported (Klein, Monteforte et al. 2016) . The combined effects of reported higher rates of added microbial infections and increased risk of death in patients with critical COVID-19 warrants local characterization of the rates, risk factors and likely pathogens among these patients.

In Kenya, 246,643 confirmed cases and 4,995 deaths have been reported since March 2020 (WHO dashboard, accessed September 22, 2021: https://covid19.who.int/). Our experience in managing COVID-19 points to widespread empiric use of antibiotics, often with little evidence of co-infections or superinfections. This has the potential to worsen antibiotic resistance patterns and facilitate emergence of multidrug resistant organisms which is already a major healthcare challenge in our region (Tadesse, Ashley et al. 2017) . To the best of our knowledge, no studies thus far have examined co-infections and superinfections in those hospitalized with critical COVID-19 in Kenya and the wider sub-Saharan Africa. We therefore performed a study of the demographics, incidence, risk factors and outcomes of co-infections and superinfections in critically ill COVID-19 patients in a tertiary health facility in Nairobi, which has been the epicentre of the outbreak in our country.

This was a retrospective cohort study involving patients admitted with critical COVID-19 for a period of at least 24 hours and who subsequently died or were discharged. The study was carried out at the Aga Khan University Hospital, Nairobi (AKUHN), a 258-bed, private, not-for-profit, tertiary level teaching and referral hospital. The institution has 15 critical care beds available for the management of critical COVID-19 patients.

All critical COVID-19 patients aged ≥ 18 years with a confirmed diagnosis of COVID-19 pneumonia were enrolled. Confirmation was based on a positive real-time reverse transcriptionpolymerase chain reaction (RT-PCR) assay for SARS-CoV-2 in nasopharyngeal swabs or samples from lower respiratory tract obtained via tracheal aspirate or broncho-alveolar lavage. In addition, radiological evidence of pneumonia based on chest radiographs or chest computed tomography (CT) scans was required. We excluded patients with a clinical presentation that was suggestive of COVID-19 but who had negative RT-PCR tests, and any re-admissions for post COVID-19 complications.

COVID-19 pneumonia: Clinical and radiologic features consistent with pneumonia and a positive SARS-CoV-2 PCR from a respiratory specimen. The extent of lung parenchymal involvement on baseline chest CT scans was determined using the CT severity score (Saeed, Gaba et al. 2021) . We defined Critical COVID-19 based on the WHO COVID-19 disease severity classification (Organization 2021) to include those with acute respiratory distress syndrome (ARDS), sepsis or septic shock. ARDS was characterized using the 2012 Berlin definition (Ferguson, Fan et al. 2012) whereas sepsis or septic shock were defined as per the third international consensus definitions (Singer, Deutschman et al. 2016) .

Bloodstream infection (BSI): Isolation of a pathogenic microbe from at least one blood culture or isolation of an organism considered to be a skin commensal from at least two sets of blood cultures in the presence of a compatible syndrome as ascertained by an infectious disease specialist in the study team (JR or FR).

Bacterial/Viral/Mycobacterial/Fungal Pneumonia: Growth of a bacterial or fungal isolate from a respiratory sample (sputum/tracheal aspirate/ broncho alveolar lavage fluid) or a positive PCR from a respiratory sample with a compatible clinical syndrome as ascertained by an infectious disease specialist in the study team.

All of these clinically significant added infections were categorized either as co-infections (diagnosis made at the time of or within the first 48 hours of COVID-19 hospital admission) or superinfections (diagnosis made after 48 hours of hospitalization). Multi-drug resistant organisms (MDROs) including extended spectrum beta-lactamase producing enterobacterales, carbapenem-resistant enterobacterales, carbapenem-resistant Acinetobacter baumannii, multidrug-resistant Pseudomonas aeruginosa and methicillin-resistant Staphylococcus aureus were classified using the Centers for Disease Control and Prevention (CDC) definitions for nosocomial infections (Garner, Jarvis et al. 1988 .

The manual and electronic health records for all the patients admitted to the critical care units with a diagnosis of critical COVID-19 prior to the study period were reviewed and data of interest was extracted and entered into a REDCap © database.

Data obtained included: clinical demographic profiles (age, sex, race); comorbid conditions (hypertension, diabetes, cancer, rheumatologic diseases, HIV, chronic heart/lung/kidney disease); days since onset of symptoms; source of admission to critical care (direct, medical ward, inter-hospital transfer); baseline inflammatory markers at admission to critical care ward (c-reactive protein, lymphocyte counts, procalcitonin, ferritin, d-dimer); microbiologic data on record (blood, urine, respiratory tract cultures; viral multiplex PCRs from respiratory specimens; mycobacterial nucleic acid amplification tests; mycobacterial cultures); radiologic tests (CT scans of the chest, chest radiographs); specific therapies offered to the patients for COVID-19 (remdesivir, dexamethasone, tocilizumab); antimicrobial therapy given before/during hospitalization; presence and duration of indwelling catheters (central venous catheters, urethral catheters); duration of hospitalisation; duration of critical care stay; duration of mechanical ventilation and discharge status (alive at discharge/transfer to another hospital or dead). After data collection and coding, the data collected was exported to SPSS for analysis (IBM Statistical Package for the Social Sciences version 22.00).

Continuous variables such as baseline inflammatory markers, duration of hospital stay and duration of mechanical ventilation were expressed as medians with interquartile ranges (IQRs)

whereas categorical variables such as the outcome variable, sex, co-morbidities as well as the interventions used were expressed as frequencies and percentages. Normality of the data was analyzed using the Shapiro Wilk test for continuous data. Differences between groups with no infections and those with added infections were compared using Student's t-test or the Wilcoxon-Mann-Whitney test (depending on the normality of the data) for continuous variables and using Chi square (χ^2) or Fisher's exact test for categorical variables. A p-value < 0.05 was considered significant.

We reviewed data collected between March 2020 and May 2021. During this period, a total of 321 patients were admitted to the AKUHN critical care units with respiratory failure secondary to critical COVID-19 pneumonia. Of these, 229 (71.3%), were male, 213 (66.4%) African/black and 257 (80.1%) had at least one co-morbid condition. The median (IQR) age of this cohort was 61.4 (51.4-72.9) years. Systemic arterial hypertension, diabetes mellitus and pre-existing renal disease were the most frequently reported co-morbidities at 184/321 (57.3%), 154/321 (48%) and 41/321 (12.8%) respectively. Thirteen (4%) of the patients had a co-existing malignancy while 8

(2.5%) had HIV co-infection. Oral antibiotic use prior to admission was reported in 40 patients (12.5%). These were predominantly penicillins, cephalosporins and macrolides at 16/40 (40%), 14/40 (35%) and 14/40 (35%) respectively. The median (IQR) time to presentation to hospital from onset of symptoms was 6 (3-7) days, and 287 (90%) of the patients had crepitations on chest examination at arrival. Baseline chest radiographs showed severe (˃50%) lung parenchymal involvement in 199/293 (67.9%) patients. Table 1 provides a summary of these clinical demographics by different categories.

The patients were admitted to critical care units from the medical isolation wards (51.4%), directly through the Accidents and Emergency (A&E) department (44.2%) or as transfers in from other facilities (4.4%). 143 patients (44.5%) eventually required invasive mechanical ventilation.

Other interventions offered included dexamethasone (87.5%), tocilizumab (62%) and methylprednisone (14%). Antimicrobial drugs were prescribed to 224 (69.8%) patients, including 57.1% who had no microbiologically confirmed concomitant infections. The antibiotics prescribed in the 224 patients included: penicillins (52.2%), carbapenems (46.8%), cephalosporins (31.7%), macrolides (31.2%), vancomycin (21.5%), echinocandins (10.2%), voriconazole (3.9%), polymyxins (2.9%) and trimethoprim/sulfamethoxazole (2.9%). 99 patients received more than one antimicrobial agent.

The overall mortality in our study population was 41.1% (132/321). All except five were 

uncertainty in the management of COVID-19 infection which led us to investigate their frequency and impact in critically ill patients in our institution. Low rates of co-infections and superinfections in COVID-19 patients have been reported in previous studies (Hughes, Troise et al. 2020) . However, this is disproportionately higher (up to 56%) in critically ill COVID-19 patients (Fattorini, Creti et al. 2020 , Kim, Quinn et al. 2020 , Lv, Cheng et al. 2020 ) and has been associated with increased mortality (Shafran, Shafran et al. 2021 , Silva, Lima et al. 2021 ). Most guidelines currently discourage empiric use of antibiotics in COVID-19 except for those with clinical suspicion for co-infections or those requiring direct ICU admission due to critical illness (Adebisi, Jimoh et al. 2021) . In our study of critically ill COVID-19 patients, coinfections and superinfections were documented in 29% of the patients with microbiologic samples, which is substantially lower than the rates reported in other studies. Previous authors have suggested that use of prophylactic antibiotics may be a factor in the low rates of added infections (Garcia-Vidal, Sanjuan et al. 2020). However, our study did not demonstrate a statistically significant difference in the rates of added infections between those with or without prior exposure to outpatient antibiotics.

Almost two-thirds of the patients in this study received in-patient antimicrobials (antibiotics and/or antifungals) which is consistent with those reported in a review by Rawson et al. (Rawson, Moore et al. 2020) . The discrepancy between the frequency of antimicrobial use and actual confirmed infections as demonstrated in our study and others (Chedid, Waked et al. 2021) is concerning for the potential of selecting for multidrug resistant organisms (Clancy and Nguyen 2020) . Broad-spectrum antibiotics were administered in a large proportion of our study population, with 46.8% of those given antibiotics receiving carbapenems. Use of broad-spectrum antibiotics is a documented risk for driving up antimicrobial resistance rates (Sonmezer, Ertem et al. 2016) , and discretion is necessary in ensuring their utilization is appropriate. Our study found a high percentage of multi-drug resistant organisms (MDROs) among the commonly isolated pathogens: 21 of the 22 (95.5%) A. baumannii, 10 of 46 (22.7%) K. pneumoniae, 2 of 32 (6.3%) E. Coli and 5 of 36 (13.9%) P. aeruginosa isolates were carbapenem resistant. This was a higher percentage of MDROs compared to our institution's 2020 antibiotic susceptibility report in which carbapenem resistance was documented in 65% of A. baumannii, 7% of K. pneumoniae, 1% of E. Coli and 15% of P. aeruginosa bloodstream isolates (report in the supplementary appendix). A recent study on Acinetobacter infections in our institution also revealed that 86% of the isolates were multi-drug resistant (Patel, Shah et al. 2019) , which is still a lower frequency than that noted in our study. This aberration in the rates of MDROs between our study findings and the institutional antibiotic susceptibility report may be in part related to increased empiric use of broad spectrum antimicrobial agents.

Factors significantly associated with high rates of added infections were need for and duration of mechanical ventilation, presence and duration of invasive urethral or central venous catheters, tocilizumab administration and duration of ICU stay, consistent with the findings of other authors (Garcia-Vidal, Sanjuan et al. 2020 . The necessity for these interventions should be carefully considered daily and discontinued whenever indicated in order to mitigate against the risk of infection that they pose. The presence of added bacterial or fungal infections was associated with higher mortality in our study population (58% compared to 36% for those with no added infections). Due to the association of added microbial infections with increased mortality especially in critically ill patients, we advocate for characterization of the local patterns of susceptibility of these concomitant infections. This would facilitate appropriate utilization of antimicrobial agents to improve outcomes without unnecessarily driving up the rates of drug resistant pathogens.

PTB-COVID-19 co-infection was only demonstrated in one of our patients, who was HIV negative. This patient had an excellent response to concurrent therapy for both infections, did not require invasive mechanical ventilation and was discharged after a brief stay in the ICU. Given that our country is in a TB-endemic zone, this was a reassuringly low rate of co-infection given the risk of increased COVID-19 mortality in the presence of TB co-infection (Sarkar, Khanna et al. 2021) . Despite the low incidence in our study, vigilance is still required because therapies like steroids and tocilizumab may exacerbate or unmask undiagnosed TB infection (Bandyopadhyay, Palepu et al. 2020 , Tadolini, García-García et al. 2020 . HIV in COVID-19 pneumonia has also been cited as a risk factor for increased mortality. Though our numbers were limited, this pattern was not demonstrated in our study. Out of 8 HIV positive patients, 6 survived to discharge. All 8 patients had been on antiretroviral drugs with documented viral suppression at admission.

We appreciate some important limitations relevant to our study. As this was a single-centre study, our findings may not be generalizable given that there may exist variation in terms of frequently isolated nosocomial pathogens, availability of adjunctive laboratory tests or interventions such as invasive mechanical ventilation and vasopressors between different institutions in our region. Secondly, the absence of a microbiologic isolate does not rule out the possibility of an intercurrent infection and reliability of microbiology test results vary between laboratories. In addition, antibiotics may have been prescribed to some of the patients before microbiologic samples were obtained. Lastly, as this was a retrospective cohort analysis, it is conceivable that the attending physicians might not have requested microbiologic samples in patients who could have had co-infections and super-infections.

Our study reports relatively low rates of microbial infections in our cohort of critically ill COVID-19 patients. However, where present, these were associated with higher death rates, greater need for mechanical ventilation and prolonged hospitalization. Multi-drug resistant pathogens were frequent among the commonly isolated bacteria such as K. pneumoniae, P.

aeruginosa, E. coli and A. baumannii. Given the high rates of empiric broad-spectrum antimicrobial use in our study cohort, broader regional epidemiology, microbiology and susceptibility patterns of likely added infections in critical COVID-19 patients is emergently necessary in order to develop local guidance on approach to rational empiric antimicrobial usage.

The authors of this paper report no conflicts of interest with regards to this study.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Ethical approval for this study was obtained from the Institutional Ethics Review Committee (IERC) of the Aga Khan University, East Africa Medical College (reference number: 2021/IERC-16 (v1)). A waiver of consent was granted by the IERC since this was a retrospective analysis of data obtained as part of routine care, with no collection of patient identifiers such as name or hospital number. A research license was also secured from the National Commission for Science, Technology and Innovation (NACOSTI) (license number: NACOSTI/P/21/9035). Aspergillus flavus Candida parapsilosis 1 t 8 patients had pathogens isolated from bloodstream, respiratory tract and urine; 24 patients had blood and tracheal aspirate isolates; 11 patients had blood and urine isolates; 11 patients had urine and tracheal aspirate isolates.

10 patients had similar pathogens isolated from blood and tracheal aspirate cultures,2 patients had similar pathogens isolated from blood and urine cultures whereas 5 patients had similar pathogens isolated from urine and tracheal aspirate cultures. 

Antibiotic resistance threats in the United States

The use of antibiotics in COVID-19 management: a rapid review of national treatment guidelines in 10 African countries

COVID-19 and tuberculosis co-infection: a neglected paradigm

Hydroxychloroquine with or without Azithromycin in Mild-to-Moderate Covid-19

Antibiotics in treatment of COVID-19 complications: a review of frequency, indications, and efficacy

Coronavirus Disease 2019, Superinfections, and Antimicrobial Development: What Can We Expect?

Bacterial coinfections in COVID-19: an underestimated adversary

The Berlin definition of ARDS: an expanded rationale, justification, and supplementary material

Azithromycin in addition to standard of care versus standard of care alone in the treatment of patients admitted to the hospital with severe COVID-19 in Brazil (COALITION II): a randomised clinical trial

Incidence of co-infections and superinfections in hospitalized patients with COVID-19: a retrospective cohort study

CDC definitions for nosocomial infections

COVID-19 pneumonia and the appropriate use of antibiotics

How covid-19 is accelerating the threat of antimicrobial resistance

Bacterial and fungal coinfection among hospitalized patients with COVID-19: a retrospective cohort study in a UK secondary-care setting

Rates of Co-infection Between SARS-CoV-2 and Other Respiratory Pathogens

The frequency of influenza and bacterial coinfection: a systematic review and meta-analysis

Antibiotic prescribing in patients with COVID-19: rapid review and meta-analysis

Co-infections in people with COVID-19: a systematic review and meta-analysis

Clinical characteristics and co-infections of 354 hospitalized patients with COVID-19 in Wuhan, China: a retrospective cohort study

Impact of COVID-19 pandemic on utilisation of healthcare services: a systematic review

Prevalence and outcomes of co-infection and superinfection with SARS-CoV-2 and other pathogens: A systematic review and meta-analysis

COVID-19 clinical management: living guidance

Health coverage and what Kenya can learn from the COVID-19 pandemic

Acinetobacter Infections: A Retrospective Study To Determine In -Hospital Mortality Rate And The Clinical Factors Associated With Mortality

Bacterial and Fungal Coinfection in Individuals With Coronavirus: A Rapid Review To Support COVID-19 Antimicrobial Prescribing

Correlation between Chest CT Severity Scores and the Clinical Parameters of Adult Patients with COVID-19 Pneumonia

Impact of COVID-19 in patients with concurrent co-infections: A systematic review and meta-analyses

Secondary bacterial infection in COVID-19 patients is a stronger predictor for death compared to influenza patients

Fungal and bacterial coinfections increase mortality of severely ill COVID-19 patients

The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3)

Evaluation of Risk Factors for Antibiotic Resistance in Patients with Nosocomial Infections Caused by Pseudomonas aeruginosa

Antimicrobial resistance in Africa: a systematic review

On tuberculosis and COVID-19 co-infection

The Unrecognized Threat of Secondary Bacterial Infections with COVID-19

Viral-bacterial coinfection affects the presentation and alters the prognosis of severe community-acquired pneumonia

Risks and features of secondary infections in severe and critical ill COVID-19 patients