Prevalence and predictors of antibiotic prescription among patients hospitalized with viral lower respiratory tract infections in Southern Province, Sri Lanka

Abstract

Background

Antimicrobial overprescription is common for lower respiratory tract infections (LRTI), as viral and bacterial infections generally present with similar clinical features. Overprescription is associated with downstream antimicrobial resistance. This study aims to identify the prevalence and predictors of antibiotic prescription among patients hospitalized with viral LRTI.

Methods

A prospective cohort study was conducted among patients aged ≥1 year hospitalized with viral LRTI in a tertiary care hospital in Southern Province, Sri Lanka from 2018–2021. Demographic, clinical, and laboratory data were recorded. Nasopharyngeal and blood samples were collected for multiplex polymerase chain reaction testing for 21 respiratory pathogens and procalcitonin (PCT) detection, respectively. Demographic and clinical features associated with antibiotic prescription were identified using Chi Square and t-tests; significant variables (p<0.05) were further included in multivariable logistic regression models. The potential impact of biomarker testing on antibiotic prescription was simulated using standard c-reactive protein (CRP) and PCT cut-offs.

Results

Of 1217 patients enrolled, 438 (36.0%) had ≥1 respiratory virus detected, with 48.4% of these patients being male and 30.8% children. Influenza A (39.3%) and human rhinovirus/ enterovirus (28.3%) were most commonly detected. A total of 114 (84.4%) children and 266 (87.8%) adults with respiratory viruses were treated with antibiotics. Among children, neutrophil percentage (median 63.6% vs 47.6%, p = 0.04) was positively associated with antibiotic prescription. Among adults, headache (60.6% vs 35.1%, p = 0.003), crepitations/crackles (55.3% vs 21.6%, p<0.001), rhonchi/wheezing (42.9% vs 18.9%, p = 0.005), and chest x-ray opacities (27.4% vs 8.1%, p = 0.01) were associated with antibiotic prescription. Access to CRP and procalcitonin test results could have potentially decreased inappropriate antibiotic prescription in this study by 89.5% and 83.3%, respectively.

Conclusions

High proportions of viral detection and antibiotic prescription were observed among a large inpatient cohort with LRTI. Increased access to point-of-care biomarker testing may improve antimicrobial prescription.

Introduction

Lower respiratory tract infection (LRTI) remains the world’s deadliest communicable disease and a top five leading cause of infectious admissions and disability-adjusted life years, leading to a total of 489 million cases and 2.5 million deaths in 2019 [1, 2]. LRTI is comprised of conditions including pneumonia, bronchitis, and infective exacerbations of chronic obstructive pulmonary disease (COPD) and asthma [3]. LRTI is caused primarily by bacterial and viral pathogens, and less commonly by fungal organisms; however, viral LRTI is often more commonly detected than bacterial LRTI among both children and adults [4–7]. Both bacterial and viral LRTI generally present with similar clinical features, such as fever, cough, chest pain, and sputum production [3]. The similarity of these clinical manifestations has posed a great challenge for physicians worldwide in managing and distinguishing bacterial and viral LRTI. For this reason, physicians often overprescribe antibiotics for viral LRTI [8–10]. Alternatively, failure to effectively treat LRTI conditions can result in a higher risk for severe complications, prolonged course of symptoms, secondary infections, and mortality, especially among vulnerable populations such as the immunosuppressed, elderly, or those with pre-existing health conditions [3, 4].

Antibiotic overprescription and misprescription contribute to an emerging top ten global health threat: antimicrobial resistance (AMR) [11, 12]. AMR occurs when bacteria, viruses, and other pathogens adapt to antimicrobial drugs and evolve into strong, drug-resistant organisms that decrease antimicrobial drug effectiveness and impede the treatment of infectious diseases [11]. While AMR caused a global burden of approximately 4.95 million deaths in 2019, it has been shown to have a disproportionate impact in low- and middle-income countries (LMICs), causing some of the highest rates of death in South Asia [8].

LRTI may be diagnosed through a combination of clinical assessment, including the evaluation of clinical features, medical history, and physical exam findings, as well as through chest radiography. Specific diagnostic tests to differentiate between bacterial and viral etiologies include pathogen-based diagnostics, such as sputum Gram stain and culture, blood cultures, serologic studies, antigen detection tests, and nucleic acid amplification tests [13, 14]. However, the sensitivity and specificity of these pathogen-based diagnostics may be limited. Invasive procedures–such as bronchoscopic bronchoalveolar lavage, transthoracic needle aspiration, or protected specimen brush–may provide a more accurate etiology but are infrequently performed due to their invasive nature [10]. Biomarkers, such as C-reactive protein (CRP) and procalcitonin (PCT), may also be used as an adjunct to identify the class of infection (bacterial versus viral) and to improve antimicrobial use [10].

In Sri Lanka, a LMIC in South Asia, LRTI is a common reason for hospitalization among both children and adults. Distinguishing viral and bacterial LRTI can be difficult due to insufficient resources and diagnostics [9, 15, 16]. In 2018, a Sri Lankan study reported that 30.1% of hospitalized patients were potentially inappropriately prescribed antibiotics, for which LRTI was the most common indication for prescription [9]. Further, a 2021 study found high proportions of antibiotic prescription among outpatients in Sri Lanka, of whom only 4.1% of patients presenting with respiratory symptoms underwent diagnostic testing [17]. Despite concerns related to the high prevalence of LRTI and inappropriate antibiotic prescription in Sri Lanka, there is limited information on predictors of antibiotic prescription among patients with viral LRTI.

To improve antibiotic prescribing practices and appropriately inform future antibiotic stewardship interventions, it is critical to understand antibiotic prescription patterns and features associated with antibiotic prescription among patients with viral LRTI. The aim of this study was to identify the prevalence and predictors of antibiotic prescription among children and adults hospitalized with viral LRTI in Sri Lanka.

Materials and methods

Setting

The Respiratory Infection Severity and Etiology in Sri Lanka (RISE-SL) study was conducted in a large, 1,800-bed, public tertiary care hospital in Southern Province, Sri Lanka from April 2018 to October 2021. The hospital is administered by the Ministry of Health and offers free outpatient and inpatient services, including medical, pediatric, and surgical care, to a catchment population of 3.5 million.

Participants

Consecutive patients ≥1 year old admitted with acute LRTI were identified for enrollment within 48 hours of admission. Patients were eligible if they met an age-specific case definition for LRTI and had an acute illness for a duration of <14 days. For patients ≥5 years, the case definition included 1) evidence of acute respiratory illness, as indicated by at least one sign or symptom (new cough, sputum production, chest pain, dyspnea, tachypnea [>25 breaths/minute], abnormal lung examination, or need for mechanical ventilation) and 2) evidence of acute infection, as indicated by at least one sign or symptom (reported fever or chills, documented fever or hypothermia, leukocytosis, leukopenia, or new altered mental status). Chest x-ray imaging within 48 hours of admission was required for eligibility for patients ≥5 years. For patients <5 years of age, the case definition for LRTI included 1) evidence of acute respiratory illness, as indicated by cough or difficulty breathing and 2) evidence of acute infection, as indicated by at least one of the following: lower chest wall indrawing, tachypnea (≥40 breaths per minute), oxygen saturation <90%, central cyanosis, severe respiratory distress, inability to drink or breastfeed, vomiting everything, altered consciousness, or convulsions. Chest x-rays were not required in patients <5 years of age. Patients were not eligible to participate in this study if they were outpatients, hospitalized within the past 28 days, unable or unwilling to provide consent or biological samples, had known or suspected infections at other anatomic sites requiring antibacterial therapy, or had any specific condition that could have precluded participation and affected subject safety according to study physicians or clinical providers. The study recruitment period spanned from April 24, 2018, to January 19, 2021.

Data collection

A standardized questionnaire on demographic and clinical information was administered to all patients by trained research assistants. A nasopharyngeal sample and blood sample were collected at enrollment and stored at -80°C until used for testing. Patients were followed longitudinally during hospitalization and information relating to laboratory and radiographic results, treatments administered, and clinical outcomes were obtained from the medical record. Clinical decision-making was conducted by the routine care providers and was not influenced by research personnel.

Laboratory testing

Nasopharyngeal samples were tested using the Luminex NxTAG Respiratory Pathogen Panel (Luminex Corporation, Austin, TX, USA) [18]. This panel detects 3 bacterial pathogens–Chlamydophila pneumoniae, Mycoplasma pneumoniae, and Legionella pneumophila–and 18 viral pathogens, including adenovirus, bocavirus, coronavirus types 229E, HKU1, NL63, and OC43, influenza types A, A H1, A H3, and B, human metapneumovirus, parainfluenza types 1–4, respiratory syncytial virus types A and B, and human rhinovirus/enterovirus; human rhinovirus and enterovirus are closely related members of the Picornaviridae family, therefore, the assay cannot reliably differentiate them due to their genetic similarity [19]. This panel was chosen for its diagnostic capabilities across a wide array of targets, capacity to process up to 96 samples per run, enhanced sensitivity for certain viruses and detection of coinfections, and benefits such as reduced manual labor and lower consumption of consumable and reagents when compared with alternative methods [20–22]. Testing for SARS-Coronavirus-2 was conducted using the Centers for Disease Control and Prevention (CDC) SARS-CoV-2 assay on an AB7500 Fast DX (Applied Biosystems). PCT testing was performed on serum samples using the VIDAS BRAHMS PCT kit using the Enzyme-Linked Fluorescent Assay technique via the Mini-VIDAS platform (bioMérieux, Marcy-l’Étoile, France).

Statistical analysis

Data analyses were conducted using R Statistical Software (R Core Team 2022). Descriptive statistical analysis was conducted to determine the prevalence of respiratory viral positivity and antibiotic prescription. Bivariable analyses were conducted using Chi Square and t-tests to identify the association between sociodemographic characteristics, clinical features, and physical, laboratory, and radiographic examination results and antibiotic prescriptions among patients with viral detection. Variables which were statistically significant (p<0.05) on bivariable analysis were checked for collinearity and included in multivariable logistic regression models. Two separate multivariable models were created for children and adults. Backward elimination methods were implemented, whereby full models were initially fitted with all predictor variables, and variables with the greatest p-values were gradually removed at each step until all variables met the significance threshold (p<0.05). Excluded variables were added back to the models one at a time to verify that they were not significant.

Chest x-ray result (normal/ abnormal) was excluded as a variable from the multivariable logistic regression model for adults, since presence of opacity/ consolidation on x-ray was included as a variable and was thought to be more specific for bacterial infection. Abnormal x-rays could include findings such as interstitial pattern, which is seen with viruses. To prevent model instability, diarrhea and x-ray opacity variables were also excluded from the multivariable model for children, as both had frequency percentages of less than 5% among those with no antibiotic prescription.

The potential impact of PCT and CRP testing on antibiotic prescription, assuming clinical guidelines were followed and PCT test results were available to the physicians at the time of clinical decision-making, were simulated individually using standard PCT and CRP cut-offs of >0.25 ng/mL and ≥ 20 mg/L, respectively, to indicate the likelihood of clinically relevant bacterial infection.

Ethical procedures

Ethical approval for this research study was acquired from the Duke University Institutional Review Board (USA) and Ethical Review Committee of the Faculty of Medicine, University of Ruhuna (Sri Lanka). Written informed consent was obtained from patients ≥18 years of age and the parents or guardians of patients 1–17 years of age. Written assent was obtained from pediatric patients aged 12–17 years.

Results

A total of 1217 patients were enrolled during the study period, including 190 (15.6%) children under 18 years of age, 614 (50.6%) male patients, and 600 (49.4%) female patients. Of the 1217 patients, 438 (36.0%) had at least one respiratory virus detected. The most commonly detected viruses were influenza A (39.3%), human rhinovirus/ enterovirus (HRV/HEV; 28.3%), and respiratory syncytial virus A (RSV A; 10.3%) (Table 1). Among children, the most commonly detected viruses were influenza A (7.5%), HRV/HEV (7.3%), and bocavirus (5.9%); and those among adults were influenza A (31.7%), HRV/HEV (21.0%), and human metapneumovirus (5.3%). Bocavirus (5.9% vs 3.7%, p<0.001) and RSV A (5.3% vs 5.0%, p = 0.002) were detected more commonly in children than adults. Influenza A was detected less commonly in children than adults (7.5% vs 31.7%, p<0.001). None of the patients tested positive for COVID-19 in our study.

[Figure omitted. See PDF.]

The results presented in the remainder of the manuscript focus on the 438 patients with viral detection.

Antibiotic prescription among patients with viral detection

Of the 438 patients with viral detection, 135 (30.8%) were children and 212 (48.4%) were male; the median age was 4 years (IQR 2–7) for children and 61 years (IQR 48–70) for adults (Table 2). Overall, 380 (86.8%) patients with respiratory viruses were treated with antibiotics during hospitalization, including 114 (84.4%) children and 266 (87.8%) adults (Table 3). There was no difference in proportion receiving antibiotic prescription between children and adults (p = 0.34).

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

Among the 380 (86.8%) patients with viral detection who received antibiotics, the most commonly prescribed antibiotics were amoxicillin and clavulanic acid (44.5%), third-generation cephalosporins (30.8%), and clarithromycin (30.4%) (Table 3). These three antibiotics were also the most commonly prescribed among pediatric and adult patients, when analyzed individually. First-generation cephalosporins (0.5% vs 0.0%, p = 0.03), second-generation cephalosporins (4.1% vs 3.4%, p = 0.002), vancomycin (1.8% vs 0.7%, p = 0.002), erythromycin/ azithromycin (5.5% vs 4.1%, p<0.001), and fluoroquinolones (3.0% vs 2.3%, p = 0.006) were prescribed more commonly in children than adults, respectively. Amoxicillin and clavulanic acid (6.6% vs 37.9%, p<0.001) and third-generation cephalosporins (12.1% vs 18.7%, p = 0.01) were prescribed less commonly among children than adults.

Demographic and clinical features associated with antibiotic prescription

On bivariable analysis, there was a non-significant trend related to the likelihood of antibiotic prescription based on age (p = 0.05), with patients aged <65 years being less likely to be prescribed antibiotics than other patients (68.9% vs 84.5%, Table 2). Patients with diabetes were less likely to be prescribed antibiotics than other patients (10.0% vs 19.0%, p = 0.04).

The most common clinical symptom upon admission was cough (411, 10.0%), followed by fever (380, 9.0%) and chills (374, 9.0%) (Table 4). Pediatric patients presenting with diarrhea were significantly more likely to be prescribed antibiotics than other patients (22.4% vs 0.0%, p = 0.02); however, those presenting with wheezing were significantly less likely to be prescribed antibiotics than other patients (19.6% vs 45.0%, p = 0.01). Adult patients presenting with cough (p = 0.03) or headache (p = 0.003) were significantly more likely to be prescribed antibiotics than other patients. Fever was not significantly associated with antibiotic prescription among either children or adults; however, there was a non-significant trend related to the likelihood of antibiotic prescription among adult patients presenting with fever (p = 0.07).

[Figure omitted. See PDF.]

The most common physical examination finding was crepitations/ crackles (199, 43.6%), followed by rhonchi/ wheezing (170, 37.3%) and drowsiness/ confusion (17, 3.7%) (Table 4). Pediatric patients presenting with free fluid were significantly more likely to be prescribed antibiotics than other patients (p = 0.02); however only 1 (0.8%) patient was found to present with free fluid. Adult patients presenting with crepitations/crackles (55.3% vs 21.6%, p<0.001) or rhonchi/ wheezing (42.9% vs 18.9%, p = 0.005) were significantly more likely to be prescribed antibiotics than other patients.

The most common clinical diagnosis upon admission was LRTI (289, 66.0%), followed by asthma exacerbation (55, 12.6%) and COPD exacerbation (25, 5.7%) (Table 4). Both pediatric (p = 0.005) and adult (p = 0.004) patients with an admission clinical diagnosis of LRTI were significantly more likely to be prescribed antibiotics than other patients.

Routine laboratory testing results associated with antibiotic prescription

We looked at laboratory results within the first 48 hours of hospitalization as this is a critical period for antibiotic prescription decision-making. Neutrophil percentage (p = 0.04) and ESR (p<0.001) were significantly associated with antibiotic prescription (Table 5); patients who were prescribed with antibiotics had median values of 71.8 (IQR 60.1–81.4) for neutrophil percentage and 40 (IQR 21.0–64.0) for ESR. Among pediatric patients, laboratory test findings of neutrophil percentage (p = 0.04) were significantly associated with antibiotic prescription; children who were prescribed antibiotics had a median neutrophil percentage value of 63.6 (IQR 46.0–78.1) versus 47.6 (IQR 31.0–62.6) in those who were not. Pediatric patients presenting with leukocytosis (64, 60.4%) were significantly more likely to be prescribed with antibiotics than other patients (3, 23.1%, p = 0.01). Among adult patients, laboratory test results of CRP (p = 0.001) and ESR (p<0.001) were significantly associated with antibiotic prescription; adults who were prescribed antibiotics had CRP and ESR median values of 40mg/L (IQR 17.0–74.0) and 40mm/ hour (IQR 22.0–66.0), respectively.

[Figure omitted. See PDF.]

Chest radiograph features associated with antibiotic prescription

Overall, all 438 patients with viral detection (135 children and 303 adults) received chest radiographs. A total of 179 (47.4%) chest radiographs showed abnormal results, including those of 48 (35.6%) children and 131 (43.2%) adults (Table 5). Among patients with parenchymal abnormalities, 105 (24.0%) had opacities, 110 (25.1%) had effusions, and 24 (5.5%) had interstitial patterns. On bivariable analysis, pediatric patients with opacity/ consolidation on chest x-ray were significantly more likely to be prescribed antibiotics than patients without this finding (24.6% vs 4.8%, p = 0.04). Adult patients presenting with abnormal readings (46.9% vs 28.1%, p = 0.04) or opacity/ consolidation (27.4% vs 8.1%, p = 0.01) were significantly more likely to be prescribed antibiotics than other patients.

Multivariable logistic regression of features associated with antibiotic prescription

In the final multivariable model, neutrophil percentage was associated with higher odds of antibiotic prescription among pediatric patients (OR 1.04, 95% CI 1.01–1.07) (Table 6). Adult patients presenting with rhonchi/ wheezing on exam (OR 4.08, 95% CI 1.76–10.70), opacities on chest x-ray (OR 4.15, 95% CI 1.25–21.85), crepitations/ crackles on exam (OR 4.94, 95% CI 2.15–12.74), or headache (OR 1.39, 95% CI 0.97–1.91) were more likely to be prescribed antibiotics.

[Figure omitted. See PDF.]

C-reactive protein (CRP) testing

Patients with viral detection had a median CRP level of 30.5 (IQR 11.2–66.3), with 17 (IQR 7.0–41.3) for children and 37 (IQR 15.0–72.8) for adults. On simulation using the standard CRP cut-off of ≥ 20 mg/L to identify bacterial LRTI, having adhered to CRP test results may have reduced antibiotic prescription among physicians (Table 7). A level of 60% adherence to CRP-guided antibiotic prescription using CRP ≥ 20 mg/L may have potentially reduced inappropriate antibiotic prescription by 33.6% in children and 16.7% in adults; 80% adherence may have potentially reduced inappropriate antibiotic prescription by 44.8% in children and 22.3% in adults; and 100% adherence may have potentially reduced inappropriate antibiotic prescription by 56.0% in children and 27.9% in adults. The potential reduction in antibiotic prescription was significantly greater among pediatric patients than adult patients at 60%, 80%, and 100% adherence levels.

[Figure omitted. See PDF.]

Procalcitonin (PCT) testing

Patients with viral detection had a median PCT level of 0.2 (IQR 0.1–0.7) (Table 8), with 0.4 (IQR 0.2–1.2) for children and 0.1 (IQR 0.1–0.5) for adults. On simulation using the standard PCT cut-off of >0.25 ng/mL to identify bacterial LRTI, having access to PCT test results may have reduced antibiotic prescription among physicians (Table 7). A level of 60% adherence to PCT-guided antibiotic prescription using PCT >0.25 ng/mL may have potentially reduced inappropriate antibiotic prescription by 23.4% in children and 37.0% in adults; 80% adherence may have potentially reduced inappropriate antibiotic prescription by 31.3% in children and 49.4% in adults; and 100% adherence may have potentially reduced inappropriate antibiotic prescription by 39.1% in children and 61.7% in adults. The potential reduction in antibiotic prescription was significantly greater among adult patients than pediatric patients at 60%, 80%, and 100% adherence levels.

[Figure omitted. See PDF.]

Discussion

We conducted a prospective cohort study in a large tertiary medical center in Southern Province, Sri Lanka to identify the prevalence and predictors of antibiotic prescription among patients hospitalized with viral LRTI. The goal of this study was to inform future antibiotic stewardship interventions and improve antibiotic prescribing practices. We found a high prevalence of antibiotic prescription among patients with viral detection. Features significantly associated with antibiotic prescription on multivariable analysis included neutrophil percentage among children, and headache, crepitations/ crackles on exam, rhonchi/ wheezing on exam, and x-ray opacity among adults. We showed that access to and adherence to CRP and PCT test results could potentially reduce antibiotic prescriptions by up to 56.0% and 39.1% in children and 27.9% and 61.7% in adults, respectively.

We showed a high percentage of viral detection among a cohort of children and adults hospitalized with LRTI in Sri Lanka, with over one third (36.0%) of enrolled patients having at least one virus detected. A Chinese study found a similar viral detection level of 36.6% among acute LRTI inpatients across 108 hospitals [23]. In contrast, the percentage of viral detection in our study was twice as great as the viral detection level of 15.4% found by a Canadian study among patients hospitalized with respiratory symptoms [24]. Among the viruses detected in the Canadian study were influenza, RSV, parainfluenza, and coronaviruses, all of which are common causes of viral pneumonia [25–27]. While pneumonia was formerly thought to be primarily caused by bacteria, recent studies have increasingly recognized that a high proportion of pneumonia conditions are caused by viruses [4, 28]. No patients were found to test positive for COVID-19 in our study, which could be attributed to the early pandemic practices in Sri Lanka, when patients with COVID-19 were being admitted to specialized isolation hospitals.

Despite finding a high level of viral detection in this study, the results also showed a high proportion of antibiotic prescription (86.8%) among patients with viral detection; therefore, a majority of the antibiotic prescriptions were likely unnecessary. This high prevalence of potential overprescription was found to be nearly equivalent among children and adults (84.4% vs 87.8%), respectively. Similarly, the aforementioned Canadian study found a high prevalence of antibiotic prescription (94.5%) among adult inpatients with viral detection, and thus suggested a positive viral nasopharyngeal test result was not associated with decreased antibiotic prescription [24]. Further, a study in the Netherlands and Israel found that children were significantly less likely to receive an antibiotic prescription than adults (37% vs 83%, p<0.001) among patients at emergency departments or hospitalized with viral RTI [29].

We examined associations between 89 distinct patient features and antibiotic prescription among pediatric and adult patients with viral detection. We found high antibiotic prescription prevalence among children and adults presenting with fever and cough, which are common LRTI symptoms [4]; yet, neither symptom was significantly associated with antibiotic prescription in our study. In contrast, a 2019 study from Malta found acute RTI patients presenting with fever (OR 2.6, 95% CI 2.08–3.26, p<0.001) or productive cough (OR 1.3, 95% CI 1.03–1.61, p = 0.03) were significantly more likely to be prescribed antibiotics than other patients [30]. Headache, also a common LRTI symptom, was found to be significantly associated with antibiotic prescription among adults in the reduced multivariable model. A 2016 Polish study reported an association of headaches while bending forward with antibiotic prescription among pediatric and adult outpatients with RTI (OR 1.93, 95% CI 1.26–2.95, p = 0.003) [31]. However, a 2006 Netherlands study reported headaches were not significantly associated with antibiotic prescription among adult outpatients with LRTIs [32]. Crepitations/ crackles and rhonchi/ wheezing on exam were found to be significantly associated with antibiotic prescription in the reduced adult multivariable model. The previously mentioned 2016 Polish study also reported similar significant associations of rales/ crepitations (OR 26.90, 95% CI 9.00–80.35, p<0.001) and rhonchi/ wheezing (OR 13.03, 95% CI 7.14–23.80, p<0.001) with antibiotic prescription [31]. However, the 2006 Netherlands study found wheezing was not significantly associated with antibiotic prescription among adult outpatients with LRTIs [32]. Neutrophil percentage was the only patient feature that was significantly associated with antibiotic prescription in the children’s multivariable model; no associations with neutrophil percentage were found among similar studies. We also found a significant association between x-ray opacity/ consolidation and antibiotic prescription in the reduced model for adults. Similarly, a 2021 Scottish study found abnormal x-ray results were significantly associated with antibiotic prescription (OR 1.88, 95% CI 1.22–2.90, p = 0.005) among adult outpatients with RTI; however, x-ray abnormality types, such as opacity, were not specified [33]. Although clinical guidelines recommend x-ray testing for LRTI patients [4], there is limited literature that report on the associations of x-ray results and antibiotic prescription.

Despite finding five features that were significantly associated with antibiotic prescription among children and adults, patients without these features also had high rates of antibiotic prescription. For example, patients who presented with cough were significantly more likely to be prescribed antibiotics than those who were not; yet 70.6% of those who did not present with cough were also prescribed antibiotics. Therefore, the patient features significantly associated with antibiotic prescription in this study did not aid in distinguishing whether patients had bacterial or viral LRTI infections. However, our team also conducted PCT tests for research purposes, which could have potentially helped decrease inappropriate antibiotic prescription in this study by 83.3%.

Past studies have suggested biomarker tests such as CRP and PCT may help physicians distinguish bacterial from viral LRTI [10]. When physicians are uncertain about the cause of LRTI, clinical guidelines discourage antibiotic prescription for LRTI patients if their CRP levels are < 20 mg/L or PCT levels are ≤0.25 ng/mL, which show a bacterial infection is unlikely; antibiotic prescription is encouraged if CRP ≥20 mg/L or PCT >0.25 ng/mL, as this points to a high probability of bacterial infection [10]. CRP data were collected as part of clinical testing for 68.5% of the patients with viral detection and were made available to physicians; 38.0% of the patients had low CRP results (0.20 mg/L), and 89.5% of these patients with low CRP results were prescribed antibiotics. We also performed PCT testing for 79.2% of the patients with viral detection; over half of the patients (58.8%) had low PCT results (≤0.25 ng/mL), and 83.3% of these patients with low PCT results were prescribed antibiotics. In contrast, 90.2% of patients with high PCT results (>0.25 ng/mL) received antibiotic prescriptions, as would be recommended by clinical guidelines. The physicians in this study did not have access to PCT test results prior to or during the time of clinical decision making; however, the high level of potential inappropriate antibiotic prescription may have been potentially reduced by more than 80% if the physicians had had access to PCT testing results and followed clinical guidelines. Past studies have found PCT-guided care may lead to physicians to shorten or discontinue antibiotic therapy; it has been associated with a reduction in total antibiotic exposure of 2.9 days among patients with pneumonia [34] and 2.4 days among patients with general LRTI [35], and a reduction in overall antibiotic prescription of 32.4% among patients with pneumonia [36] and between 25% [37] and 47% [38] among patients with general LRTI.

While biomarker testing, such as that of PCT, may be used as tools to guide decision-making on antibiotic therapy for patients with LRTI, physicians are not recommended to use this data as the sole evidence to initiate or end antibiotic therapy [39]. Biomarker testing does not replace clinical judgement or other laboratory tests, and is recommended to supplement other available evidence, such as microbiological data [39]. For this reason, adequate training, such as that offered through antibiotic stewardship, is needed for the optimal use of PCT and other biomarker data [40]. However, it is important to consider that adherence to biomarker-guided antibiotic stewardship training is usually left to the voluntary discretion of physicians [41]; despite having access to CRP results, physicians still overprescribed antibiotics in this study.

Biomarker-guided antibiotic stewardship in LMICs may lead to long-term benefits, such as increased overall cost-savings of hospitals and decreased antimicrobial prescription [40, 42]; however, there are existing barriers to implementing biomarker-guided antibiotic stewardship in LMICs. A 2021 Sri Lankan study reported that despite most physicians stating PCT testing could be helpful, 41% had limited or no prior clinical experience with PCT [43]. In another study, PCT use experts agreed that PCT-guided antibiotic stewardship could help fight acute RTIs in Asia-Pacific countries; however, they expressed concern for a lack of PCT education among physicians [42]. Thus, increased physician training is required to successfully expand PCT testing in LMICs. Moreover, increased training is required for healthcare authorities and administration to adequately assess the long-term cost-effectiveness of PCT-guided antibiotic stewardship in LMIC hospitals [40].

Strengths of this study include the large sample size and unbiased, consecutive sampling over a long duration, allowing for more precise estimates of the associations between patient features and antibiotic prescription. Further, the study population included pediatric and adult patients, which improved representativeness of the results. Our study had a few limitations. Data regarding ESR and CRP biomarkers were only collected when conducted as part of clinical testing. Biomarkers, such as CRP, have been used as tools to diagnose infections and guide antibiotic prescription [40, 44]; however, past studies have shown PCT levels have greater specificity and sensitivity in detecting bacterial infections than CRP. Due to the limited availability of CRP data in our study, we were unable to determine whether CRP was significantly associated with antibiotic prescription. In this study, we also assumed detection of a respiratory viral pathogen was equivalent to infection, however asymptomatic colonization can be seen with viral organisms. Following standard procedure, we tested nasopharyngeal swabs to diagnose etiology instead of utilizing invasive methods such as bronchoalveolar lavage (BAL); however, past studies have shown that there is substantial concordance in the detection of respiratory viruses between nasopharyngeal swabs and BAL when using multiplex polymerase chain reaction tests [45, 46]. Furthermore, we did not account for the presence of bacterial co-infection. Moreover, this study was conducted at a single center in Southern Province, Sri Lanka. While the study population includes pediatric and adult patients, the generalizability of the results is limited to this center and local area, therefore limiting our ability to expand future interventions to other settings. Further, this study was conducted at a tertiary care hospital and may not reflect care at other lower-level hospitals in Sri Lanka and other LMICs.

Conclusions

Among a large inpatient cohort with LRTI, viral infection was most common and antibiotic prescription was high. We identified neutrophil percentage, headache, crepitations/ crackles, rhonchi/ wheezing, and x-ray opacity to be associated with antibiotic prescription. CRP and PCT testing, if available, has potential to reduce inappropriate antibiotic prescription by up to 89.5% and 83.3% among inpatients with viral LRTI, respectively. Further research needs to be conducted on the impact of CRP and PCT biomarker testing on antibiotic prescription in low-resource settings to properly inform future antibiotic stewardship interventions.

Acknowledgments

We would like to acknowledge the great contributions and support from the researchers, physicians, collaborators, study staff, and laboratory personnel of RISE-SL and Ruhuna-Duke Centre for Infectious Diseases. We would also like to acknowledge the support of bioMérieux for providing the procalcitonin kits used in this study.

Citation: Medrano PG, Weerasinghe N, Nagahawatte A, Vanderburg S, Park LP, Wijayaratne GB, et al. (2024) Prevalence and predictors of antibiotic prescription among patients hospitalized with viral lower respiratory tract infections in Southern Province, Sri Lanka. PLoS ONE 19(6): e0304690. https://doi.org/10.1371/journal.pone.0304690

About the Authors:

Perla G. Medrano

Contributed equally to this work with: Perla G. Medrano, Nayani Weerasinghe

Roles: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliations: Duke University, Durham, North Carolina, United States of America, Duke Global Health Institute, Durham, North Carolina, United States of America

ORICD: https://orcid.org/0000-0001-5292-5265

Nayani Weerasinghe

Contributed equally to this work with: Perla G. Medrano, Nayani Weerasinghe

Roles: Data curation, Investigation, Methodology, Resources, Writing – review & editing

Affiliation: Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Ajith Nagahawatte

Roles: Investigation, Methodology, Resources, Writing – review & editing

Affiliations: Duke Global Health Institute, Durham, North Carolina, United States of America, Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Sky Vanderburg

Roles: Data curation, Funding acquisition, Investigation, Methodology, Supervision, Writing – review & editing

Affiliation: University of California, San Francisco, California, United States of America

Lawrence P. Park

Roles: Formal analysis, Methodology, Supervision, Writing – review & editing

Affiliations: Duke University, Durham, North Carolina, United States of America, Duke Global Health Institute, Durham, North Carolina, United States of America

Gaya B. Wijayaratne

Roles: Data curation, Investigation, Methodology, Writing – review & editing

Affiliation: Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Vasantha Devasiri

Roles: Investigation, Methodology, Writing – review & editing

Affiliation: Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Buddhika Dilshan

Roles: Data curation, Methodology, Project administration, Writing – review & editing

Affiliation: Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Tianchen Sheng

Roles: Formal analysis, Investigation, Methodology, Writing – review & editing

Affiliations: Duke University, Durham, North Carolina, United States of America, Duke Global Health Institute, Durham, North Carolina, United States of America

Ruvini Kurukulasooriya

Roles: Methodology, Project administration, Writing – review & editing

Affiliation: Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

Jack Anderson

Roles: Data curation, Methodology, Writing – review & editing

Affiliation: Duke University, Durham, North Carolina, United States of America

Bradly P. Nicholson

Roles: Data curation, Methodology, Supervision, Writing – review & editing

Affiliation: Institute for Medical Research, Durham, North Carolina, United States of America

Christopher W. Woods

Roles: Investigation, Methodology, Resources, Supervision, Writing – review & editing

Affiliations: Duke University, Durham, North Carolina, United States of America, Duke Global Health Institute, Durham, North Carolina, United States of America

Champica K. Bodinayake

Roles: Conceptualization, Investigation, Methodology, Resources, Supervision, Writing – review & editing

¶‡ CKB and LGT also contributed equally to this work.

Affiliations: Duke Global Health Institute, Durham, North Carolina, United States of America, Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

ORICD: https://orcid.org/0000-0003-1339-1883

L. Gayani Tillekeratne

Roles: Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing

¶‡ CKB and LGT also contributed equally to this work.

Affiliations: Duke University, Durham, North Carolina, United States of America, Duke Global Health Institute, Durham, North Carolina, United States of America, Faculty of Medicine, University of Ruhuna, Galle, Sri Lanka

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

References

1. Lower respiratory infections—Level 3 cause. Institute for Health Metrics and Evaluation. The Lancet; 2020. Available from: https://www.healthdata.org/results/gbd_summaries/2019/lower-respiratory-infections-level-3-cause

2. Vos T, Lim SS, Abbafati C, Abbas KM, Abbasi M, Abbasifard M, et al. Global Burden of 369 Diseases and Injuries in 204 Countries and territories, 1990–2019: a Systematic Analysis for the Global Burden of Disease Study 2019. The Lancet. 2020 Oct 17;396(10258):1204–22. Available from: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30925-9/fulltext pmid:33069326

3. Dasaraju PV, Liu C. Infections of the Respiratory System. Nih.gov. University of Texas Medical Branch at Galveston; 2014. Available from: https://www.ncbi.nlm.nih.gov/books/NBK8142/

4. Woodhead M, Blasi F, Ewig S, Garau J, Huchon G, Ieven M, et al. Guidelines for the management of adult lower respiratory tract infections—Full version. Clinical Microbiology and Infection [Internet]. 2011 Nov 26;17:E1–59. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7128977/ pmid:21951385

5. O’Brien KL, Baggett HC, Brooks WA, Feikin DR, Hammitt LL, Higdon MM, et al. Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia: the PERCH multi-country case-control study. The Lancet. 2019 Aug;394(10200):757–79. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6727070/#!po=15.0000 pmid:31257127

6. Jain S, Self WH, Wunderink RG, Fakhran S, Balk R, Bramley AM, et al. Community-Acquired Pneumonia Requiring Hospitalization among U.S. Adults. New England Journal of Medicine. 2015 Jul 30;373(5):415–27. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4728150/ pmid:26172429

7. Jain S, Williams DJ, Arnold SR, Ampofo K, Bramley AM, Reed C, et al. Community-Acquired Pneumonia Requiring Hospitalization among U.S. Children. New England Journal of Medicine. 2015 Feb 26;372(9):835–45. Available from: https://www.nejm.org/doi/full/ pmid:25714161

8. Murray CJ. Global Burden of Bacterial Antimicrobial Resistance in 2019: A Systematic Analysis. The Lancet. 2022 Jan 19;399(10325):629–55. Available from: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)02724-0/fulltext

9. Sheng T, Wijayaratne GB, Dabrera TM, Drew RJ, Nagahawatte A, Bodinayake CK, et al. Point-prevalence study of antimicrobial use in public hospitals in southern Sri Lanka identifies opportunities for improving prescribing practices. Infection Control & Hospital Epidemiology. 2018 Dec 7;40(2):224–7. pmid:30522538

10. Noviello S, Huang D. The Basics and the Advancements in Diagnosis of Bacterial Lower Respiratory Tract Infections. Diagnostics [Internet]. 2019 Apr 3;9(2):37. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6627325/ pmid:30987144

11. World Health Organization. Antimicrobial resistance. Who.int. World Health Organization: WHO; 2021. Available from: https://www.who.int/news-room/fact-sheets/detail/antimicrobial-resistance

12. Ramachandran P, Rachuri N, Martha S, Shakthivel R, Gundala A, Battu T. Implications of Overprescription of Antibiotics: A Cross-Sectional Study. Journal of Pharmacy And Bioallied Sciences. 2019;11(6):434. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6555336/ pmid:31198382

13. Wijnands GJA. Diagnosis and interventions in lower respiratory tract infections. The American Journal of Medicine [Internet]. 1992 Apr 6;92(4):S91–7. Available from: https://www.amjmed.com/article/0002-9343(92)90317-5/abstract pmid:1580272

14. Carroll KC. Laboratory Diagnosis of Lower Respiratory Tract Infections: Controversy and Conundrums. Journal of Clinical Microbiology [Internet]. 2002 Sep 1;40(9):3115–20. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC130746/ pmid:12202540

15. Vanderburg S, Wijayaratne G, Danthanarayana N, Jayamaha J, Piyasiri B, Halloluwa C, et al. Outbreak of severe acute respiratory infection in Southern Province, Sri Lanka in 2018: a cross-sectional study. BMJ Open. 2020 Nov;10(11):e040612. pmid:33158834

16. Nirubaa U. Burden of respiratory disease on hospital admission of children: Experience from a tertiary care centre. Sri Lanka Journal of Child Health. 2021 Dec 5;50(4):661.

17. Zhang HL, Bodinayake C, Wijayaratne GB, Jayatissa P, Piyasiri DLB, Kurukulasooriya R, et al. Point-prevalence survey of outpatient antibiotic prescription at a tertiary medical center in Sri Lanka: opportunities to improve prescribing practices for respiratory illnesses. BMC Infectious Diseases. 2021 Jan 21 [cited 2023 Aug 25];21:97. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7818802/ pmid:33478430

18. NxTAG® Respiratory Pathogen Panel Test | Diasorin [Internet]. int.diasorin.com. DiaSorin; Available from: https://int.diasorin.com/en/molecular-diagnostics/kits-reagents/nxtag-respiratory-pathogen-panel

19. NxTAG ® Respiratory Pathogen Panel + SARS-CoV-2 Package Insert [Internet]. Luminex Corporation. U.S. Food and Drug Administration; 2022 Nov. Available from: https://www.fda.gov/media/146495/download

20. Lee CK, Lee HK, Ng CWS, Chiu L, Tang JWT, Loh TP, et al. Comparison of Luminex NxTAG Respiratory Pathogen Panel and xTAG Respiratory Viral Panel FAST Version 2 for the Detection of Respiratory Viruses. Annals of Laboratory Medicine [Internet]. 2017;37(3):267. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5339100/

21. Gonsalves S, Mahony J, Rao A, Dunbar S, Juretschko S. Multiplexed detection and identification of respiratory pathogens using the NxTAG® respiratory pathogen panel. Methods [Internet]. 2019 Apr 1;158:61–8. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7128802/

22. Popowitch EB, Kaplan S, Wu Z, Tang YW, Miller MB. Comparative Performance of the Luminex NxTAG Respiratory Pathogen Panel, GenMark eSensor Respiratory Viral Panel, and BioFire FilmArray Respiratory Panel. Mostafa HH, editor. Microbiology Spectrum [Internet]. 2022 Jun 29; Available from: https://journals.asm.org/doi/ pmid:35766513

23. Feng L, Li Z, Zhao S, Nair H, Flaxman S, Xu W, et al. Viral Etiologies of Hospitalized Acute Lower Respiratory Infection Patients in China, 2009–2013. PLOS ONE. 2014 Jun 19;9(6):e99419–9.

24. Mulpuru S, Aaron SD, Ronksley PE, Lawrence N, Forster AJ. Hospital Resource Utilization and Patient Outcomes Associated with Respiratory Viral Testing in Hospitalized Patients. Emerging Infectious Diseases. 2015 Aug;21(8):1366–71. pmid:26197268

25. CDC. Causes of Pneumonia. CDC. 2019. Available from: https://www.cdc.gov/pneumonia/causes.html

26. Viral pneumonia Information | Mount Sinai—New York. Mount Sinai Health System. Available from: https://www.mountsinai.org/health-library/diseases-conditions/viral-pneumonia

27. Ruuskanen O, Lahti E, Jennings LC, Murdoch DR. Viral pneumonia. Lancet (London, England). 2011;377(9773):1264–75. Available from: https://www.ncbi.nlm.nih.gov/pubmed/21435708/ pmid:21435708

28. Freeman AM, Soman-Faulkner K, L Jr. Viral Pneumonia. Nih.gov. StatPearls Publishing; 2019. Available from: https://www.ncbi.nlm.nih.gov/books/NBK513286/

29. van Houten CB, Cohen A, Engelhard D, Hays JP, Karlsson R, Moore E, et al. Antibiotic misuse in respiratory tract infections in children and adults—a prospective, multicentre study (TAILORED Treatment). European Journal of Clinical Microbiology & Infectious Diseases. 2019 Feb 1;38(3):505–14. pmid:30707378

30. Saliba-Gustafsson EA, Dunberger Hampton A, Zarb P, Orsini N, Borg MA, Stålsby Lundborg C. Factors associated with antibiotic prescribing in patients with acute respiratory tract complaints in Malta: a 1-year repeated cross-sectional surveillance study. BMJ Open. 2019 Dec;9(12):e032704. pmid:31857311

31. Strumiło J, Chlabicz S, Pytel-Krolczuk B, Marcinowicz L, Rogowska-Szadkowska D, Milewska AJ. Combined assessment of clinical and patient factors on doctors’ decisions to prescribe antibiotics. BMC family practice. 2016 Jun 3 [cited 2020 Sep 7];17:63. Available from: https://pubmed.ncbi.nlm.nih.gov/27255505/ pmid:27255505

32. Hopstaken RM, Butler CC, Muris JWM, Knottnerus JA, Kester ADM, Rinkens PELM, et al. Do clinical findings in lower respiratory tract infection help general practitioners prescribe antibiotics appropriately? An observational cohort study in general practice. Family Practice. 2005 Dec 2;23(2):180–7. Available from: https://academic.oup.com/fampra/article/23/2/180/527307 pmid:16326800

33. Seaton RA, Cooper L, Gibbons CL, Malcolm W, Choo-Kang B, Griffith D, et al. Antibiotic prescribing for respiratory tract infection in patients with suspected and proven COVID-19: results from an antibiotic point prevalence survey in Scottish hospitals. JAC Antimicrob Resist. 2021 Jun 18;3(2):dlab078. pmid:34223137; PMCID: PMC8211567.

34. Rhee C, Mansour MK. UpToDate. UpToDate. 2022 Nov 3; Available from: https://www.uptodate.com/contents/procalcitonin-use-in-lower-respiratory-tract-infections

35. Schuetz P, Wirz Y, Sager R, Christ-Crain M, Stolz D, Tamm M, et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections. Cochrane Database of Systematic Reviews. 2017 Oct 12;(10).

36. Schuetz P, Christ-Crain M, Thomann R, Falconnier C, Wolbers M, Widmer I, et al. Effect of procalcitonin-based guidelines vs standard guidelines on antibiotic use in lower respiratory tract infections: the ProHOSP randomized controlled trial. JAMA. 2009 Sep 9;302(10):1059–66. Available from: https://www.ncbi.nlm.nih.gov/pubmed/19738090 pmid:19738090

37. Kristoffersen KB, Søgaard OS, Wejse C, Black FT, Greve T, Tarp B, et al. Antibiotic treatment interruption of suspected lower respiratory tract infections based on a single procalcitonin measurement at hospital admission—a randomized trial. Clinical Microbiology and Infection. 2009 May;15(5):481–7. pmid:19416298

38. Christ-Crain M, Jaccard-Stolz D, Bingisser R, Gencay MM, Huber PR, Tamm M, et al. Effect of procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: cluster-randomised, single-blinded intervention trial. The Lancet. 2004 Feb;363(9409):600–7. pmid:14987884

39. Nora D, Salluh J, Martin-Loeches I, Póvoa P. Biomarker-guided antibiotic therapy—strengths and limitations. Annals of Translational Medicine. 2017 May;5(10):208–8. pmid:28603723

40. Lee C, Kwa A, Apisarnthanarak A, Feng J, Gluck E, Ito A, et al. Procalcitonin (PCT)-guided antibiotic stewardship in Asia-Pacific countries: adaptation based on an expert consensus meeting. Clinical Chemistry and Laboratory Medicine (CCLM). 2020;58(12): 1983–1991. pmid:31926074

41. Chung GW, Wu JE, Yeo CL, Chan D, Hsu LY. Antimicrobial stewardship. Virulence. 2013 Feb 15;4(2):151–7. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3654615/

42. de Carvalho FRT, Telles JP, Tuon FFB, Rabello Filho R, Caruso P, Correa TD. Antimicrobial Stewardship Programs: A Review of Strategies to Avoid Polymyxins and Carbapenems Misuse in Low Middle-Income Countries. Antibiotics. 2022 Mar 12;11(3):378. pmid:35326841

43. Rolfe R, Kwobah C, Muro F, Ruwanpathirana A, Lyamuya F, Bodinayake C, et al. Barriers to implementing antimicrobial stewardship programs in three low- and middle-income country tertiary care settings: findings from a multi-site qualitative study. Antimicrobial Resistance & Infection Control. 2021 Mar 25;10(1). pmid:33766135

44. van Griensven J, Cnops L, De Weggheleire A, Declercq S, Bottieau E. Point-of-Care Biomarkers to Guide Antibiotic Prescription for Acute Febrile Illness in Sub-Saharan Africa: Promises and Caveats. Open Forum Infect Dis. 2020 Jun 30;7(8):ofaa260. pmid:32818139; PMCID: PMC7423291.

45. Bouzid D, Le Hingrat Q, Salipante F, Ferré V, Chevallier T, Tubiana S, et al. Agreement of respiratory viruses’ detection between nasopharyngeal swab and bronchoalveolar lavage in adults admitted for pneumonia: a retrospective study. Clinical Microbiology and Infection [Internet]. 2023 Jul 1;29(7):942.e1–6. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9873593/ pmid:36708772

46. Ulrich P, Chen D. 1332. Clinical Indicators for When Bronchoalveolar Lavage (BAL) Is Needed Beyond Nasopharyngeal Swab (NP) Testing for Viral Respiratory Infections. Open Forum Infectious Diseases [Internet]. 2021 Nov 1;8(Supplement_1):S754–4. Available from: https://academic.oup.com/ofid/article/8/Supplement_1/S754/6451045