Introduction
The novel coronavirus disease (COVID-19) has been spreading rapidly since the first case of infection was confirmed in December 2019 in Wuhan, China, and cases are now being reported worldwide [
1]. The cumulative number of COVID-19 infections has reached 36 million, with more than 10 million deaths reported [
2]. In Japan, the number of infected people has been increasing, and outbreaks, not only in cities but also in medical and nursing care facilities, have become a public health problem [
3]. The spread of infection in healthcare facilities can lead to facility disruption and the disintegration of healthcare systems in the community. Therefore, clinicians are required to properly identify and manage patients with COVID-19.
COVID-19 often causes fever, upper respiratory tract symptoms, cough, malaise, olfactory, and gustatory dysfunction after 7–14 days of the incubation period [
4‐
7]. However, the clinical course and symptoms are nonspecific, making it difficult to distinguish COVID-19 from common cold and other febrile diseases. Although several hematologic and biochemical changes associated with COVID-19 have been reported [
8‐
10], there is no specific blood test for COVID-19. Inflammatory marker levels and lymphocyte counts have been investigated as predictors of severe COVID-19 [
11,
12], but few have been validated as tools for the diagnosis of COVID-19. On the contrary, chest computed tomography (CT) has shown a high frequency of abnormalities, and several CT findings characteristic of COVID-19 have been reported [
13‐
15]. However, there are several pathological conditions such as interstitial pneumonia and pulmonary edema that need to be differentiated.
The definitive diagnosis of COVID-19 was confirmed mainly by polymerase chain reaction (PCR) testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) RNA. The sensitivity of the PCR test for COVID-19 is approximately 70% [
16‐
18]; therefore, there is a risk of producing false-negative results. Thus, it is still challenging to manage suspected COVID-19 cases in the absence of a highly sensitive and specific diagnostic test system. In addition, limited medical and human resources make it difficult to perform PCR testing sufficiently in some facilities and geographical regions. To date, no useful clinical indicators have been established for the management of suspected COVID-19 cases in healthcare settings.
In response to the spread of COVID-19 infection in Hokkaido, Japan, Hokkaido University Hospital has systematically managed suspected COVID-19 cases under the leadership of respiratory physicians and infection control teams. In this study, we retrospectively analyzed the clinical characteristics and management details of suspected COVID-19 patients at Hokkaido University Hospital. We also compared the clinical, laboratory, and radiological characteristics of suspected and confirmed COVID-19 patients to clarify the differences. Based on these findings, we developed a clinical score (COVID-19 Clinical Risk Score) to ensure the accurate management of suspected COVID-19 patients.
Discussion
In the present study, we analyzed the clinical characteristics of the medical management of suspected COVID-19 patients and compared the characteristics of suspected and confirmed COVID-19 patients who visited our medical institution. Patients who did not undergo PCR tended to have no history of high-risk behaviors, had lower CT scores, and had alternative diagnoses. There were significant differences in the COVID-19 group and the non-COVID-19 group with respect to some symptoms, blood test findings, CT score values, and the presence of alternative diagnoses. Based on our findings, we developed the COVID-19 Clinical Risk Score to guide proper management of suspected COVID-19 patients. As a result, the distribution of scores differed significantly among patient groups, with non-COVID-19 patients who did not undergo PCR scoring less than 4 and COVID-19 patients scoring greater than 5.
Although Hokkaido in Japan had been under the COVID-19 emergency alert during most of the study period, only two patients were confirmed to have COVID-19 among more than 100 suspected cases observed at our hospital. Because our facility is a university hospital with a wide range of specialized departments, the patient backgrounds and referral sources were diverse. Therefore, our patients’ backgrounds might be different from those of patients in primary care clinics and city hospitals. Nonetheless, COVID-19 is a disease with a high risk of outbreak and hospital shutdown once it is identified in a medical institution [
22]. Even if the actual number of COVID-19 patients is low, there will be a large number of suspected cases. Thus, proper management of suspected COVID-19 cases is critical.
In this study, we compared the clinical characteristics of confirmed COVID-19 and non-COVID-19 patients. Significant differences between the two groups were found in the rates of sore throat and olfactory and gustatory disturbances. Although the symptoms of COVID-19 have been described as nonspecific, it has been recently found that olfactory and gustatory disturbances are characteristic of patients with COVID-19 [
7,
23,
24]. The present study also confirmed that olfactory and taste disorders are more frequent in patients with COVID-19 than in suspected patients. However, the positivity rate was not high (13.3%); hence, these disorders were not included in the clinical risk score.
Leukopenia and lymphopenia have been reported as hallmarks of COVID-19 blood tests. The neutrophil-to-lymphocyte ratio tends to be higher in severe COVID-19, and lymphopenia is an important indicator of the severity of COVID-19 [
25]. In the present study, COVID-19 patients had significantly lower WBC counts, and leukopenia was added to the risk score, but the rate of lymphocytopenia was not significantly different between COVID-19 and non-COVID-19 patients. Several non-COVID-19 patients had inflammatory diseases in which lymphocytes were depleted due to an elevated WBC count along with elevated neutrophil ratio; therefore, a low lymphocyte count may not be helpful in the diagnosis of COVID-19 in clinical practice. Contrary, eosinophil count has been reported to be decreased in the early stage of the disease and subsequently elevated in COVID-19 [
26]. Several of the confirmed COVID-19 patients in this study also showed a significant decrease in eosinophil count, and eosinopenia was adapted in the risk score. Previous studies on COVID-19 have shown that elevated procalcitonin levels are associated with disease severity [
10,
27]. However, in the present study, which compared COVID-19 cases with other febrile illnesses and cases with abnormal images on chest CT, a low level of procalcitonin was evident in patients with COVID-19, whereas the CRP level was elevated. Procalcitonin level is predominantly elevated in bacterial infections [
28]. A substantial number of cases of bacterial infections may be observed in non-COVID-19 patients, and these patients were more likely to have elevated procalcitonin level at an early stage of the disease. Furthermore, we hypothesize that few cases of bacterial coinfection were observed in the COVID-19 patient group in the early stage of the disease. In the early stage of the disease, elevated procalcitonin appears to help reduce the risk of COVID-19. Thus, we selected leukopenia, eosinopenia, and low procalcitonin levels along with high CRP levels as useful routine blood tests for the diagnosis of COVID-19.
In the present study, we scored the chest CT findings and examined their usefulness in the diagnosis of COVID-19. Some of the CT findings characteristic of COVID-19, such as linear or rounded opacities, may be difficult for non-specialists to read [
13‐
15]. Therefore, we focused on simple chest CT findings that could be read by general physicians and included the absence of atypical findings as an item of CT imaging score. As a result, there was a large difference in CT imaging scores between non-COVID-19 patients without PCR and confirmed COVID-19 patients. Even without professional CT reading, this result suggests that it is possible to distinguish COVID-19 from other diseases to some extent based on scoring chest CT findings.
Our focus in this study was to determine the risk of any disease besides COVID-19. This perspective is critical in the calculation of pretest probability prior to diagnostic testing, as exemplified by Well’s criteria used for the diagnosis of pulmonary embolism [
29,
30]. COVID-19 is a disease with few specific clinical manifestations. Therefore, when only blood tests and CT findings are used as indicators for diagnosis, the score may underestimate or overestimate the likelihood of COVID-19. Based on the clinical course, patient background, and other clinical findings, we aimed to determine whether any other differential diagnoses would be more likely than COVID-19 and added this judgment to the COVID-19 Clinical Risk Score. The results showed a large difference in the risk scores between COVID-19 patients and non-COVID-19 patients. We believe that adding the judgment of clinicians regarding the differential diagnosis to objective indicators made the risk score clinically relevant.
The sensitivity of the PCR test is not sufficiently high with COVID-19, and if the PCR test alone is used as an indicator of judgment, several false-negative results will be missed. In addition, due to the limitations of medical and human resources, it is not always possible to perform PCR testing on all eligible patients. For appropriate medical care and infection control, it is reasonable to estimate the pretest probability based on the clinical score to determine whether PCR should be performed. The identification of patients who have a low risk of COVID-19 and are not eligible for PCR testing by the COVID-19 Clinical Risk Score is important for the proper use of the PCR test and to conserve resources. Patients with high scores should be carefully considered for retesting and continuing infection control, even if PCR results are negative. However, the actual decision to perform PCR will be influenced by a variety of factors, including the prevalence of COVID-19, severity of the patient’s illness, fear of medical staff or patients, and policies of medical institutions or administrative government. The clinical score should be used as a reference criterion, and decision-making should be flexible in the clinical setting. In addition, if sensitive and simple diagnostic tests are developed in the future, the implications of the clinical score will change. Such testing may take some time to develop, and for the time being, the COVID-19 Clinical Risk Score will be an important indicator in clinical practice.
This study had several limitations. First, since this was a retrospective study, a prospective validation is needed to prove the usefulness of the clinical risk score. Second, the backgrounds of the suspected and confirmed COVID-19 patients were different. Because only a low number of true COVID-19 patients were identified among the suspected cases, we used data from patients admitted for the treatment of COVID-19. Third, we cannot rule out the possibility that some of the patients we classified in the non-COVID-19 group might have had COVID-19. However, since only a single case of infection was identified in the subsequent clinical course of the patients without PCR testing and no nosocomial outbreaks occurred, we believe that our management was successful.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.