key: cord-322868-o3r0y6bz authors: Williams, E. J.; Mair, L.; de Silva, T. I.; Green, D. J.; House, P.; Cawthron, K.; Gillies, C.; Wigfull, J.; Parsons, H.; Partridge, D. G. title: Routine measurement of serum procalcitonin allows antibiotics to be safely withheld in patients admitted to hospital with SARS-CoV-2 infection. date: 2020-07-02 journal: nan DOI: 10.1101/2020.06.29.20136572 sha: doc_id: 322868 cord_uid: o3r0y6bz Background. It can be a diagnostic challenge to identify COVID-19 patients without bacterial co-infection in whom antibiotics can be safely stopped. We sought to evaluate the validity of a guideline that recommends withholding antibiotics in patients with a low serum procalcitonin (PCT). Methods. We retrospectively collected 28-day outcome data on patients admitted to Sheffield Teaching Hospitals NHS Foundation Trust, UK, between 5 March and 15 April 2020, with a positive SARS-CoV-2 polymerase chain reaction (PCR) and PCT within 48 hours of diagnosis. PCT was considered negative if [≤]0.25ng/ml and positive if >0.25ng/ml. Primary outcomes included antibiotic consumption, mortality, intensive care admission and length of hospital stay. Results. 368 patients met the inclusion criteria; 218 (59%) had a negative PCT and 150 (41%) positive. At 48 hours post-diagnosis, 73 (33%) of those with a negative PCT were receiving antimicrobials compared to 126 (84%) with a positive PCT (p<0.001), with a corresponding reduction in antimicrobial usage over 28 days (median DDD of 3.0 vs 6.8 (p<0.001); median DOT 2 vs 5 days (p<0.001) between the negative and positive PCT groups.) In the negative PCT group, there were fewer deaths (62 (28%) vs. 54 (36%), (p=0.021)) and critical care admissions (19 (9%) vs. 28 (19%), (p=0.007)) than in the positive PCT group. Median length of hospital stay was 8.7 and 9 days in the negative and positive PCT groups respectively. Conclusions. Procalcitonin is a valuable tool in the assessment of patients with SARS-CoV-2 infection, safely reducing the potential burden of unnecessary antibiotic usage. In patients with COVID-19, the presentation of fever, tachypnoea and hypoxia, together with lung infiltrates on chest imaging and a frequent rise in biomarkers such as C-reactive protein (CRP) [1] presents a challenge to rational use of antimicrobials, as it is difficult to confidently identify or exclude bacterial co-infection. Rates Furthermore, the co-chairs of the American Thoracic Society and Infectious Diseases Society of America Guideline for Treatment of Adults with Community Acquired Pneumonia (ATS) have endorsed the use of the assay to guide antibiotic usage [11] . Unlike other biomarkers such as CRP, PCT is not thought to be routinely raised in COVID-19 [12] [13] [14] [15] , although data on its utility in this context are scarce. A recent meta-analysis showed that COVID-19 patients with increased PCT values had a nearly 5-fold higher risk of severe infection, thought to be due to the presence of bacterial co-infection [16] . PCT has been shown to have a greater impact on antibiotic exposure than standard stewardship strategies alone [17] , however this hypothesis remains untested in the setting of COVID-19. We set out to evaluate whether PCT use had an impact on i) antibiotic usage and ii) outcomes in patients with confirmed COVID-19 at a large NHS Foundation Trust Hospital in the United Kingdom (UK). This retrospective observational study was undertaken at two sites of Sheffield Teaching Hospitals NHS Foundation Trust (STHNHFT; Royal Hallamshire Hospital and Northern General Hospital) with a combined total of 1700 beds. Patients diagnosed between 5 March and 15 April 2020 were included in the study, allowing assessment of 28-day outcome in all patients by the time of analysis. Those diagnosed prior to 5th March were excluded as at that stage COVID-19 was managed as a high consequence infectious disease and patients were admitted regardless of symptom severity, making the group unrepresentative of those admitted later in the epidemic. The enrolment end date of 15th April was before mandatory SARS-CoV-2 screening of all patients admitted to hospital was introduced. Eligible patients were aged ≥ 18 years old, admitted to STHNFT, with both a positive SARS-CoV-2 reverse-transcriptase polymerase chain reaction (RT-PCR) result on nose and/or throat swabs and/or deep respiratory samples, and a PCT assay undertaken within 48 hours of collection of the SARS-CoV-2 sample. RT-PCR was performed using an in-house assay targeting E and RdRP genes [18] . Patients with both community and nosocomial acquisition of COVID-19 were included. STHNFT guidelines recommended that antibiotics could be withheld in COVID-19 patients with a PCT value of ≤ 0.25ng/ml unless felt necessary by a senior clinician, as concomitant bacterial infection is unlikely in such patients [19] . The study was granted approval by the STHNFT Clinical Effectiveness Unit (Ref: 9863) Demographic and clinical characteristics of patients were drawn from existing laboratory, pharmacy and clinical databases and from examination of physical and electronic patient notes. Data were entered into an electronic case report form (Access 2010, Microsoft, Redmond, WA, USA) including age, sex, ethnicity, height, weight, comorbidities, and antibiotic allergies. Antibiotic usage was recorded for the 28-day follow up period including antibiotic agent, duration, and WHO defined daily doses (DDD). Where different DDDs are defined for oral and parenteral preparations, the parenteral figure was used for calculations to avoid inappropriate weighting by route of administration for some commonly used agents (e.g. clarithromycin) which would not be relevant from a stewardship perspective [21] . Prescribing rates were calculated using DDD/evaluable day for which data of patients who died during the 28-day follow up period were censored from the date of death. Days of treatment (DOT) were defined as the number of days in the 28-day period for which antibiotics were prescribed. All values from patients meeting eligibility criteria were summarised using the most appropriate form, either using frequency/percentages, or medians with IQR (Inter-Quartile Range). In the scenario where there are multiple zeros values, and the IQR becomes uninformative with values of 0.00 to 0.00, the 10 th and 90 th centile are reported instead. Differences between demographics were analysed with the suitable significance test, depending on whether parametric assumptions were met as is detailed in the tables. To investigate the relationship between PCT positivity and total DDD, antibiotic receipt at 48 hours post-diagnosis and meropenem prescription, linear and logistic regression models were explored adjusting for demographic confounders (age, sex, ethnicity and comorbidities). Regression model assumptions were examined, and appropriate adjustments made. All p-values presented were two-sided, and estimates provided with a 95% confidence interval (CI). All statistical analyses were performed in Stata version 16 A total of 368 patients met the eligibility criteria and were included in the analysis; overall 60% were male, with a median age of 75. Of these, 218 (59%) had a PCT level of Table 1 . There was no significant difference in demographics between the two groups in terms of age, sex, BMI or ethnicity. Comorbidities between the two groups were also similarly distributed with the exception of malignancy, which was more common in the PCT negative group. There were no pregnant women in the cohort. Of those patients with a negative PCT, 73 (33%) were on antibiotics 48 hours after their COVID-19 diagnosis compared to 126 (84%) with a positive PCT (p<0.001) suggesting good compliance with the guideline (Table 2) . Data on total DDD of antibiotics received in the 28-day follow-up period and DDD per evaluable day are presented in Table 2 and Figure 2 . Patients with a negative PCT received significantly fewer DDDs of antibiotics (both total and per evaluable day) than those with positive PCT with a median DDD of 3.0 vs 6.8 (p<0.001). Due to the log-normal distribution of total DDD, a log transformation was performed on the values (after adding the smallest non-zero value of 0.17 to ensure patients with a total DDD of zero were included). Therefore, a log-linear model was computed in order to explore the relationship with PCT positivity after adjusting for demographic confounders (comorbidities, age, sex, ethnicity) to ensure regression assumptions were met. A statistically significant relationship between PCT and total DDD remained after accounting for these confounders; on average a person with PCT>0.25 had almost three times as many DDDs of antibiotics compared to those Table 1 ). Over the 28-day follow-up period, 116 (32%) of the included patients died, 229 (62%) were discharged and 23 (6%) were still in hospital. Median length of stay was 8.35 days. 47 (13%) were admitted to intensive care, and of these, 32 (68%) were intubated and ventilated. The PCT, age and 28-day mortality distribution of the patients are illustrated graphically in Figure 1 . In the PCT negative group, 62 (28%) patients died compared to 54 (36%) of those with a positive PCT (p=0.021), and 19 (9%) were admitted to ITU, compared with 28 (19%) of the positive PCT group (p=0.007); Table 2 . Meropenem was the only carbapenem used in the study population. With specific reference to meropenem consumption, positive PCT was associated with a 3-fold increase in the odds of receiving any meropenem during the course of the admission (OR= 3.16, 95% CI: 1.50, 6.65, p=0.002) after adjusting for demographic confounders (Supplementary Table 2 ). There was also a statistically significant association between receipt of antibiotics at 48 hours after COVID-19 diagnosis and receipt of meropenem (OR=3.63, 95% CI: 1.53,8.63, p=0.003). Table 3 ). Data on the potential adverse consequences of antibiotic usage are presented in Table 2 . With the exception of hospital or ventilator associated pneumonia, adverse events were all uncommon and none were significantly different between PCT positive and negative groups. This observational study supports the hypothesis that implementation of a local guideline advising against the use of antibiotics for patients with confirmed COVID-19 and a PCT level Previous studies have demonstrated that procalcitonin-guided therapy in lower respiratory tract infections substantially reduces antibiotic use without compromising outcome [22] [23] [24] , including in the critically ill [25] . Although some studies have excluded immunocompromised patients when evaluating PCT, others have shown that a rise in the biomarker correlates with bacterial infection [26, 27] . Drawbacks of PCT-guided therapy include the potential for false negative results in localised infection such as empyema, in atypical infection -and in the context of renal replacement therapy [28, 29] . As it may take 24-48 hours for PCT to reach its peak, false negatives may be seen if samples are taken early in the course of infection [30, 31] . PCT levels also fall by 50% every 24-36 hours [31] and false negatives may occur in patients due to resolution of infection. It is important, therefore, that interpretation of PCT is made in the context of other laboratory and clinical findings [17] . PCT results can be obtained rapidly from a readily obtained sample. This presents a major advantage over respiratory tract culture results, which are challenged by the inability to obtain adequate sample, slower turnaround times and insensitivity. Sputum culture has been shown to reveal a definitive diagnosis in less than 20% of cases of community-acquired pneumonia [32] . Clinical uncertainty, combined with this diagnostic insensitivity, further compromised following initiation of antibiotics, has the potential to provoke unnecessary antibiotic usage and consequent short-and long-term morbidity in patients without bacterial infection. The adopted PCT threshold of 0.25 was intentionally conservative and it may be that a higher threshold can be adopted safely, which would lead to a further reduction in antibiotic usage in patients with COVID-19. This higher mortality seen in the PCT >0.25 group supports those of other authors, demonstrating an association between higher PCT values and severe disease or death [33, 34] . It is possible that higher PCT in these patients reflects bacterial superinfection, which increases the production of PCT from non-thyroidal sources through the interplay of interleukin (IL)-1β, tumor necrosis factor (TNF)-α and IL-6, whereas viral infections result in a rise in interferon (INF)-γ which inhibits PCT synthesis [16] . It is also possible that PCT is Reducing the unnecessary use of antibiotics is a key component to mitigating this risk. The risk of severe COVID-19 disease increases with age and the elderly are also at greatest risk of the adverse consequences of excessive antibiotic use [35] . Numerous metrics exist for the measurement of antibiotic consumption, each with their own strengths and weaknesses [21] . World Health Organisation Defined Daily Doses have the advantage that they are standardised and, at institution level, can be assessed from procurement records without recourse to individual prescribing data. They do, however, have several drawbacks. They do not factor in exposure equivalent dose adjustments such as those that occur in renal or liver dysfunction, and combination therapy results in elevation of the DDDs received which may correlate poorly with relative effects on host ecology and adverse consequences of therapy. Finally, they are based on usual prescribing patterns and therefore markedly different DDD may apply to different settings or routes of administration of the same drug [36] . This latter weakness was addressed in our study through the use of the parenteral DDD for those drugs where there was a difference. We also assessed days of treatment (DOT), which may be a more relevant metric from a stewardship perspective and again this was found to be significantly lower in the negative PCT group. We also demonstrated a 3-fold increase in the odds of carbapenem prescription in those with a positive PCT. This is important in the context of the increasing global incidence of carbapenemase-producing Enterobacteriales and is likely to be a direct consequence of increased initial antibiotic usage and a concern that this had induced resistance leading to prescription of broader spectrum agents for any subsequent indication and despite the fact that isolation of ESBL or AmpC positive organisms occurred no more frequently in this group. The impact of early antimicrobial therapy on later prescription of broad spectrum agents is supported by the association between antibiotic use at 48 hours and subsequent meropenem use, which was also statistically significant. The limitations of our study include the fact that it is from a single centre and retrospective in design. However, patients were recruited systematically from the beginning of the epidemic locally until a date which was defined in advance and which preceded the introduction of routine COVID-19 testing of all admissions regardless of symptoms. It is possible that some patients may have been readmitted to other regional hospitals, but this is felt unlikely as the Treatment of Community-Acquired Pneumonia During the Coronavirus Disease 2019 (COVID-19) Pandemic Prognostic value of interleukin-6, C-reactive protein, and procalcitonin in patients with COVID-19 Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China Procalcitonin in patients with severe coronavirus disease 2019 (COVID-19): A meta-analysis Impact of Procalcitonin-Guided Antibiotic Management on Antibiotic Exposure and Outcomes: Real-world Evidence Improved sensitivity using a dual target, E and RdRp assay for the diagnosis of SARS-CoV-2 infection: Experience at a large NHS Foundation Trust in the UK Effectiveness of a procalcitonin algorithm to guide antibiotic therapy in respiratory tract infections outside of study conditions: a post-study survey Metrics for quantifying antibiotic use in the hospital setting: results from a systematic review and international multidisciplinary consensus procedure Effect of procalcitonin-based guidelines vs standard guidelines on antibiotic use in lower respiratory tract infections: the ProHOSP randomized controlled trial Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections Effect of procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: clusterrandomised, single-blinded intervention trial Efficacy and safety of procalcitonin guidance in reducing the duration of antibiotic treatment in critically ill patients: a randomised, controlled, open-label trial Assessment of Procalcitonin as a Diagnostic Marker of Underlying Infection in Patients with Febrile Neutropenia Markers of bacteremia in febrile neutropenic patients with hematological malignancies: procalcitonin and IL-6 are more reliable than C-reactive protein Procalcitonin and Proinflammatory Cytokine Clearance during Continuous Venovenous Haemofiltration in Septic Patients The challenge of removal of sepsis markers by continuous hemofiltration Procalcitonin in bacterial infections--hype, hope, more or less Serum Procalcitonin in Lower Respiratory Tract Infections in Adult Patients Assessment of the usefulness of sputum culture for diagnosis of community-acquired pneumonia using the PORT predictive scoring system Risk factors of critical & mortal COVID-19 cases: A systematic literature review and meta-analysis Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study Special considerations of antibiotic prescription in the geriatric population Measurement of adult antibacterial drug use in 130 US hospitals: comparison of defined daily dose and days of therapy VAP= Ventilator-Associated Pneumonia; ESBL= Extended Spectrum Beta Lactamase; MRSA= Methicillin-Resistant Staphylococcus aureus. CDAD= Clostridioides difficile associated disease The authors would like to thank Dr Eirini Koutoumanou of UCL for discussion around the transformation analysis.