Clinical predictors of late SARS‐CoV‐2 positivity in Italian internal medicine wards

Le résumé fourni par la source

SARS-CoV-2 represents a "Pandora's box" highlighting critical issues for in-hospital management of infectious patients. Over the past fifty years, the organization of hospital departments in Western countries has been driven by the need to face off noncommunicable diseases, in accordance with epidemiological data. Transmission-based precautions have been instead limited to dedicated departments. During the last year, strong efforts have been then required to improve infection prevention strategies, a process involving countless aspects of patient management. In forerunners countries (e.g., Italy), a first step was to identify different in-hospital paths for positive and nonpositive patients.1, 2 Due to the air contagiousness of disease, the identification and segregation of SARS-CoV-2-positive patients from negative ones represented a critical step. As weeks went by, it became evident that the case definition of SARS-CoV-2 infection was not straightforward. Clinical features of SARS-CoV-2 infection (different from those observed in previous SARS and MERS pandemics),3 patients overcrowding and pitfalls in molecular diagnosis4 contributed together to overwhelm the traditional pre-SARS-CoV-2 hospital organization. Clinicians had to lower the threshold of suspicion for deciding who needed to be tested/isolated and how to do that.5 When possible, separate pavilions were committed to SARS-CoV-2-positive patients and somewhere for those presenting with high clinical suspicion of disease patients but testing negative for SARS-CoV-2 infection. Our Institution–IRCCS Policlinico San Martino–in Genoa (Italy) early approached such categorization of patients, being aware of the dramatic impact that an intrahospital spread would have. Even hospital facilities were then reorganised to segregate positive/suspicious patients to prevent the dissemination of the infection among SARS-CoV-2-negative patients. By focussing on this critical point, we here retrospectively analyzed cases of late SARS-CoV-2 positivity occurring within internal medicine wards, formally not SARS-CoV-2-dedicated. Clinical and management variables were investigated, aiming at identifying potential predictors of late SARS-CoV-2 positivity and their impact on patient outcome.6, 7 Alongside, potential value of common laboratory test at hospital admission was explored. This is a retrospective analysis of patients admitted to internal medicine wards not dedicated to SARS-CoV-2 between 24 February 2020 and 24 May 2020 at IRCCS Ospedale Policlinico San Martino in Genoa (Italy). The screened population was composed of 690 patients admitted during the enrolment period. Only those admitted from the ED and tested with at least one nasal-pharyngeal swab for SARS-CoV-2 were included in the analysis (n = 668). Then, we excluded patients with positive for SARS-CoV-2 infection at ED reaching a final cohort of n = 478 (Figure S1). Clinical and biochemical data performed at ED admission have been collected from hospital records, including the provenance of patients (home, nursing home), any history of contact with SARS-CoV-2 cases and common symptoms of SARS-CoV-2 infection (i.e., fever, cough, dyspnoea, asthenia, anosmia, dysgeusia and diarrhoea). Comorbidities were then stratified according to the Charlson Comorbidity Index. Routinely, biochemical profile performed at ED admission and later during internal medicine ward hospitalization–including arterial blood gas analysis (BGA) and inflammatory biomarkers (i.e., fibrinogen, lactate dehydrogenase [LDH], ferritin and C-reactive protein [CRP])–was retrieved from hospital records performed at ED admission–and later during internal medicine ward hospitalization–was retrieved from hospital records and included arterial blood gas analysis (BGA) and inflammatory biomarkers (i.e., fibrinogen, lactate dehydrogenase [LDH], ferritin and C-reactive protein [CRP]). SARS-CoV-2 positivity of nasal-pharyngeal swabs were performed by routine real-time polymerase chain reaction. The present study was approved by the local ethics board of IRCCS Ospedale Policlinico San Martino (200/2020-DB id 10515). The study was carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and adhered to the principles of the transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD), the TRIPOD statement. Cases of late positivity were defined as patients testing positive for SARS-CoV-2 in internal medicine wards, not SARS-CoV-2-dedicated, after a previous negative molecular test performed in the ED (nasopharyngeal swab). The aim of the study was then to identify whether ED length of stay may have a role as predictor of late positivity to SARS-CoV-2 infection independently of other clinical variables, as emerging in literature.7, 8 They were selected among those significantly different at descriptive analyses: contact with suspected cases, fever and dyspnoea. Alongside, the same criteria were used to identify potential biochemical biomarkers: pCO2, fibrinogen, LDH, ferritin and CRP. Sample size was calculated according to latest guidance for developing clinical prediction models.9 More specifically, by considering this study as a model with binary outcome, our sample size (n = 478) satisfied the minimum sample size required for (i) a 95% confidence interval for the overall outcome proportion of 0.5 (n ≥ 385); (ii) a mean absolute precision error (MAPE) <0.05 (n ≥ 252) and (iii) an expected uniform shrinkage factor <10% (n ≥ 340). Furthermore, we have preset a proportion of overall variance explained (R2cs) = 0.1. Analyses were performed with GraphPad Prism version 9.0.0 for Windows (GraphPad Software, San Diego, CA) and R environment for statistical computing (URL http://www. R- project. org/). Categorical data were presented as absolute and relative frequencies, whereas continuous ones as median and interquartile range [IQR] since the normality assumption was not demonstrated. Unpaired intergroup comparisons were drawn by Fisher's exact test and Mann-Whitney U-test, as appropriate. Logistic regression analyses were then used to estimate the predictive role toward a late nasal-pharyngeal swab positivity to SARS-CoV-2 infection of clinical variables and biochemical parameters separately. Results were expressed as odds ratio (OR) with 95% confidence interval (CI). Model calibration performances have been evaluated via the Hosmer-Lemeshow goodness of fit test with 10 groups. Model performance was assessed through receiver-operator characteristic (ROC) curve analysis by reporting the area under the curve, sensitivity, specificity and +/− likelihood ratio. Predictive ability toward overall and SARS-CoV-2-related mortality was tested through Cox hazard regression models presented as hazard ratio (HR) with 95% CI. Survival rate was also estimated with Kaplan-Meier curve and log-rank test. For all statistical analyses, a 2-sided p-value <0.05 was considered as statistically significant. Based on a minimum measurable difference of 1 day between the groups in our primary endpoint, a late positivity group >25 patients implies a statistical power >95%, as calculated with the two-tailed Mann-Whitney U-test. Clinical characteristics of the study cohort are reported in Table S1. Patients were almost all Caucasian (99.0%), elderly (median age of 80 years [70–86]) and equally distributed across sexes (231 males, 48.3%). When study cohort was categorised in persistent negative vs. late SARS-CoV-2-positive, the latter group was associated with history of previous contact with suspicious or ascertained cases (24.5% vs. 5.8%, p = 0.003) and clinical presentation with fever (65.0% vs. 25.6%, p < 0.001) and dyspnoea (62.5% vs. 25.3%, p < 0.001) at admission. Except for the history of ischemic stroke and peptic ulcer disease (Table S2), the impact of comorbidity burden–as assessed by Charlson C

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.