key: cord-0797418-mjxnd3az authors: Kurstjens, Steef; van der Horst, Armando; Herpers, Robert; Geerits, Mick W.L.; Kluiters-de Hingh, Yvette C.M.; Göttgens, Eva-Leonne; Blaauw, Martinus J.T.; Thelen, Marc H.M.; Elisen, Marc G.L.M.; Kusters, Ron title: Rapid identification of SARS-CoV-2-infected patients at the emergency department using routine testing date: 2020-04-24 journal: nan DOI: 10.1101/2020.04.20.20067512 sha: c73ed64f506b8e073aeb06e076b6dd95cef83afa doc_id: 797418 cord_uid: mjxnd3az Background: The novel coronavirus disease 19 (COVID-19), caused by SARS-CoV-2, spreads rapidly across the world. The exponential increase in the number of cases has resulted in overcrowding of emergency departments (ED). Detection of SARS-CoV-2 is based on an RT-PCR of nasopharyngeal swab material. However, RT-PCR testing is time-consuming and many hospitals deal with a shortage of testing materials. Therefore, we aimed to develop an algorithm to rapidly evaluate an individual′s risk of SARS-CoV-2 infection at the ED. Methods: In this multicenter retrospective study, routine laboratory parameters (C-reactive protein, lactate dehydrogenase, ferritin, absolute neutrophil and lymphocyte counts), demographic data and the chest X-ray/CT result from 967 patients entering the ED with respiratory symptoms were gathered. Using these parameters, an easy-to-use point-based algorithm, called the corona-score, was developed to discriminate between patients that tested positive for SARS-CoV-2 by RT-PCR and those testing negative. Computational sampling was used to optimize the corona-score. Validation of the model was performed using data from 592 patients. Results: The corona-score model yielded an area under the receiver operating characteristic curve of 0.91 in the validation population. Patients testing negative for SARS-CoV-2 showed a median corona-score of 3 versus 11 (scale 0-14) in patients testing positive for SARS-CoV-2 (p<0.001). Using cut-off values of 4 and 11 the model has a sensitivity and specificity of 96% and 95%, respectively. Conclusion: The corona-score effectively predicts SARS-CoV-2 RT-PCR outcome based on routine parameters. This algorithm provides the means for medical professionals to rapidly evaluate SARS-CoV-2 infection status of patients presenting at the ED with respiratory symptoms. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 The subsequent exponential increase in prevalence has resulted in 100 overcrowding of emergency departments (ED) and has led to a shortage of 101 isolation rooms (4). For correct triaging of patients diagnostic testing is of key 102 importance. The leading standard test for detecting SARS-CoV-2 is an RT-103 PCR of nasopharyngeal swab material (5). However, RT-PCR testing is time-104 consuming and shortage of testing materials and capacity imposes a serious 105 threat (6). 106 107 Doctors at the ED are required to assess the probability of SARS-CoV-2 108 infection in each patient entering the ED. To accelerate the triage process at 109 the ED, we integrated routine demographic, laboratory and imaging data of 110 patients presenting at the ED with COVID-19-like symptoms to develop a 111 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 point-based algorithm. This algorithm can assess whether a person, 112 presenting at the ED with respiratory symptoms, is likely to have In case of a shortage of testing capacity, adoption of this algorithm could 114 reduce the number of patients for whom RT-PCR testing is required. 115 Moreover, implementation of the corona-score enables rapid decision making 116 at the ED, lowering pressure on isolation rooms. 117 In this retrospective multicenter study, 375 patients from three different 121 hospitals presenting at the ED with respiratory symptoms and subsequent 122 SARS-CoV-2 RT-PCR testing were included ( Figure 1 and Table 2 ). Patients 123 from other departments and patients without any respiratory symptoms or 124 suspicion of COVID-19 were excluded. An independent cohort of 592 patients 125 from four hospitals was used to validate the model ( Figure 1 and (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. score.nvkc.nl) for more information). The corona-score is obtained by the 157 summation of the score for each parameter. The final score is clamped from a 158 minimum of 0 to a maximum of 14 points. Cut-off points and weights of 159 demographic, laboratory and imaging parameters were computationally 160 sampled using Python (v3.7.0, Python Software Foundation, USA) to optimize 161 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. Using a cohort of 375 ED patients with respiratory symptoms a point-based 177 algorithm was created and subsequently validated using a separate 178 independent cohort of 592 patients (Table 2 ). At the time of presentation at 179 the ED the parameters sex, age, CRP, ferritin, LDH, ALC, ANC and CXR 180 were significantly different between the COVID-19 positive and negative 181 patients ( Figure 2 and Table 2 ). Together, these parameters were used to 182 develop an algorithm, named 'corona-score'. Inclusion of albumin, 183 procalcitonin or clinical parameters such as fever, cough and dyspnea did not 184 sufficiently improve the performance of the algorithm (data not shown). The 185 corona-score resulted in a model with an AUROC of 0.94 ( Figure 3A , 95% CI 186 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. . https://doi.org/10. 1101 /2020 0.91 -0.96). Patients with a negative RT-PCR test had a median of 4 187 compared to a median of 11 for SARS-CoV-2 positive patients ( Figure 3B and 188 Table 2 ). The corona-score algorithm was validated with data from 592 189 patients, yielding an AUROC of 0.91 ( Figure 3C , 95% CI 0.89-0.94). In the 190 validation population SARS-CoV-2 negative patients had a median of 3 191 versus a median of 11 for SARS-CoV-2 positive patients ( Figure 3D and 192 (Table 3 ). Using corona-score cut-offs of 4 (96% sensitivity) 196 and 11 (95% specificity) at a 70% prevalence, this model showed negative 197 and positive predictive values of 88% and 96% ( Figure 3E ). The total false 198 rate given these conditions is 4% ( Figure 3E ). 199 2 after repeated RT-PCR testing (n=13) had an initial median corona-score of 206 12, while patients that remained negative (n=12) had an initial median corona-207 score of 4 ( Figure 3F ). This shows that the corona-score is able to distinguish 208 between true and false negatives. 209 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. without respiratory symptoms. Therefore, this algorithm should only be used 235 for patients at the ED with respiratory symptoms. Secondly, patients that only 236 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. . https://doi.org/10. 1101 /2020 have mild respiratory symptoms, and therefore do not have large alterations in 237 their laboratory parameters, generally have a low corona-score. However, in 238 most cases the patients with a mild presentation were not hospitalized. 239 Therefore, we consider that the low corona-score corresponds with the clinical 240 findings. On the other hand, some negatively-tested patients received a high 241 corona-score. This could be due to false-negative RT-PCR testing or possibly 242 other viral infections. Interestingly, four patients that were positive for 243 influenza and negative for SARS-CoV-2 had a low corona-score (2 -6). 244 During this COVID-19 pandemic, the prevalence of other respiratory viruses 245 appears very low; hence, the discriminative potential of the corona-score in 246 patients infected by such viruses could not be systematically established. 247 Notably, in case of any viral outbreak, a similar modelling approach could be 248 considered to develop an algorithm as described here. 249 The four laboratories involved in this study deploy different instruments from 251 the major in-vitro diagnostic device providers. Most measurands that were 252 included in the algorithm have an identical metrological traceability and hence 253 comparable results in the commutable EQA scheme of the SKML (8). 254 However, there is no reference method for ferritin (9). The different 255 calibrations lead to approximately 20% difference in ferritin results between 256 the methods employed by the laboratories in this study. Therefore, a 1.2 257 harmonization factor was applied to the ferritin values obtained from Siemens 258 instruments, before calculating corona-scores, correcting the lack of 259 standardization. Generally, methodological harmonization between 260 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. Hospital. This study had no effect on the behaviour of patients or medical 297 decision-making. 298 299 The data that support the findings of this study are available from the 301 corresponding author upon reasonable request. More information can be 302 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020 . . https://doi.org/10.1101 /2020 Table 1 . The point-based scoring system for the calculation of the corona-score. The final score is clamped from a minimum of 0 to a maximum of 14 points. More information can also be found at www.corona-score.com. Bilateral infiltrate 4 CRP, C-reactive protein; LDH, lactate dehydrogenase; ALC, absolute lymphocyte count; ANC, absolute neutrophil count; CXR, chest X-ray. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2020. . Table 3 . Sensitivity and specificity at different lower and upper cut-off values for the corona-score (value included, ≤ for 2 to 5 and ≥ for 9 to 12) determined using the validation population (n = 592). The right column depicts the number of true and false negative and positive patients. True | false negative (n) 2 3 4 5 AMP (98) Chi square test †: JBZ = Jeroen Bosch Hospital Sensitivity (95% CI) Specificity (95% CI)