Introduction

Coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and became a pandemic when the first cases were reported in late 2019. During the period of April 2023, more than 760 million confirmed cases of COVID-19 have been reported worldwide. Clinical symptoms of COVID-19 can range from no symptoms to critical illness characterized by acute respiratory failure requiring mechanical ventilation, septic shock and multiple organ failure1. The most common initial symptoms are fever, cough, sore throat, nasal congestion, dyspnea, nausea/vomiting, myalgia and diarrhea2. Several neurological signs/symptoms have been reported, including anosmia, ageusia, stroke, cranial nerve deficits, encephalopathy and seizures3. Recently, a study using digital surveillance platforms in six countries revealed that anosmia and ageusia had strong associations with self-reported positive SARS-CoV-2 tests4.

Although COVID-19 may present primarily as a lower respiratory tract infection transmitted via air droplets, accumulating data suggest multiorgan involvement in patients who are infected5. This systemic involvement is thought to be mainly due to SARS-CoV-2 binding to angiotensin-converting enzyme 2 (ACE2) receptors on several different human cells6. Several clinical features and systemic inflammatory markers could be used as predictors for poor prognosis of COVID-19 infection, such as older age, underlying comorbidities, obesity, higher neutrophil-lymphocyte ratios (NLRs), platelet-lymphocyte ratios (PLRs), D-dimer levels and ANDC scores (which stand for age (A), NLR (N), d-dimer (D) and C-reactive protein (C), an early warning score to predict mortality risk for patients with COVID-19)7,8,9. Some studies have been performed on the importance of anosmia/ageusia as clinical features10,11. Previous studies revealed that anosmia/ageusia may be a predictor of a milder course of COVID rather than a more severe form12,13. Moreover, olfactory disturbances appeared to have less prognostic value in predicting the severity of COVID-19 disease than systemic inflammation markers had been reported14. Recently, new variants of SARS-CoV-2 have different properties than the alpha variant, and omicron which has been dominant for over two years causes much less chemosensory dysfunction15. Because geographic variations in host predisposition to COVID-19 related anosmia/ageusia with high rates in Western countries and lower rates in East Asia had been reported16. In addition, clinical symptoms, comorbidities and inflammatory profile such as levels of C-reactive protein (CRP) showed a significant difference between the China and North America with COVID-19 infection17. It would be of interest to explore the association between the anosmia/ageusia and systemic inflammation in Taiwan and compare with previous literatures.

There are only a few studies in Taiwan that focus on clinical manifestations, including a retrospective study of the clinical features of COVID-19 among Taiwanese pediatric patients18. In this retrospective observational study, we aimed to elucidate the clinical features, anosmia/ageusia and systemic inflammatory markers and prognosis of patients admitted with COVID-19. We hypothesize that hospitalized COVID-19 patients with anosmia/ageusia may have different clinical presentations and prognosis.

Results

Clinical and laboratory characteristics of hospitalized COVID-19 patients

A total of 231 patients with a confirmed diagnosis of COVID-19 based on nasopharyngeal screening and/or detection of SARS-CoV-2 by RT‒PCR were recruited. The median age was 62 years (IQR: 49–71 years), and 51.9% of patients were male. The median total score for initial symptoms was 3 (IQR: 2–4), and the first three symptoms were cough (64.1%), fever (58.9%), and dyspnea (56.3%). Anosmia/ageusia affected 9.1% of patients. The median severity of dyspnea was grade 1 (IQR: 0–2), and the median severity of CXR was 2 (IQR: 2–4). Laboratory tests revealed increased systemic inflammation, such as elevated median values of NLR, CRP, D-dimer and ferritin tests. The median value of the lowest RT‒PCR Ct value was 22.3 (IQR: 18.5–28.9). The median total length of stay was 14 days (IQR: 10–21), and approximately 10% of patients experienced in-hospital mortality. For 63.6% of patients, an antiviral agent was used, and for 33.8% of patients, a monoclonal antibody drug was used as an emergency treatment for COVID-19 infection. Table 1 depicts the detailed characteristics of all hospitalized COVID-19 patients.

Table 1 Demographic data and laboratory values of in-hospital patients with COVID-19 infection in two different hospitals.
Full size table

To compare the clinical characteristics of the patients from the tertiary medical center and those from the district hospital, we used Wilcoxon rank-sum tests. As shown in Table 1, there was no significant difference in median age, but there was a higher proportion of male patients in the tertiary medical center (p = 0.53 and 0.02, respectively). There was a significantly higher median total score of initial symptoms and a higher severity of dyspnea among the patients admitted to the tertiary medical center than among those admitted to the district hospital (p < 0.01 and p = 0.02, respectively). Laboratory tests showed significantly higher WBC counts, higher median NLR, ANDC and inflammatory markers such as CRP, D-dimer and ferritin in the patients admitted to the tertiary medical center than in those admitted to the district hospital (all p < 0.05). No significant difference was found in the lowest RT‒PCR Ct values between the different hospital settings. Regarding admission outcomes, the total length of stay was significantly longer in the patients admitted to the tertiary medical center than in those admitted to the district hospital (p < 0.01), but there was no significant difference in in-hospital mortality (p = 0.60). A significantly higher proportion of patients used antiviral agents (p = 0.02) and monoclonal antibodies (p < 0.01) in the tertiary medical centers than in the district hospitals. All these results showed that COVID-19 patients have a higher clinical severity in tertiary medical centers than in district hospitals.

Significant features in COVID-19 patients with anosmia /ageusia and gender differences

To investigate the clinical significance of anosmia/ageusia symptoms in patients with COVID-19, we compared the clinical and laboratory findings in both groups (Table 2). Among patients with anosmia/ageusia symptoms, there was a significant difference in median age, BMI and higher median total score of initial symptoms (p < 0.01, p < 0.01 and p = 0.01, respectively). However, COVID-19 patients with anosmia/ageusia had no significant differences of initial endotracheal intubation, severity of dyspnea, total length of stay and in-hospital mortality than those without anosmia/ageusia (all p > 0.05). Laboratory tests showed a significantly lower values of ESR, CRP, LDH and ferritin levels in patient with anosmia/ageusia group (all p < 0.05) but there were no significant differences in the NLR or PLR values between the two groups. A significant lower ANDC total scores were found in patients with anosmia/ageusia (p < 0.01). There was no significant difference in the lowest RT‒PCR Ct values between the two groups (p = 0.74). In terms of admission outcomes, there were no significant differences in in-hospital mortality or total length of stay. When compared the clinical features and laboratory tests between the gender difference in patients with COVID-19, no significant differences were found in age, BMI, anosmia/ageusia, total scores of initial symptoms and other clinical features (all p > 0.05)(Table 3). However, significant lower values of NLR, ANDC, CRP and ferritin were found in female group (all p < 0.01).

Table 2 Comparison of clinical features and laboratory values in patients with and without anosmia/ageusia symptom.
Full size table
Table 3 Comparison of clinical features and laboratory values in patients with COVID-19 infection by gender.
Full size table

In addition, we examined the associations between the severity of dyspnea and clinical characteristics and the interaction between the laboratory tests. We graded the severity of dyspnea from 0 to 2 as low grade and 3 to 4 as high grade. Approximately 23.8% of patients were classified as having a high grade of dyspnea during the initial admission. The median age was 62 years in the low dyspnea severity group and 59 years in the high dyspnea severity group, with borderline significance (p = 0.05). The median total score of initial symptoms was significantly higher in the high severity dyspnea group than in the low severity dyspnea group, but there was no significant difference in the presence or absence of anosmia/ageusia symptoms. In the high dyspnea group, a significantly higher proportion of patients underwent endotracheal intubation during the hospital course (p = 0.03), but there were no associations with in-hospital mortality or total length of stay. Regarding laboratory tests, there were no significant differences in terms of the NLR or PLR, CRP, ferritin or the lowest RT‒PCR Ct values between the two groups.

Associations between systemic inflammatory markers, ANDC scores and clinical features

Finally, the NLR and PLR values for patients with COVID-19 were investigated in the current study. The NLR, PLR, ESR, CRP, D-dimer and ferritin variables were logarithmically transformed because these variables did not fit normal distributions. Regression analysis showed that the NLR and PLR were significantly associated with age, ESR, CRP, D-dimer and ferritin levels (all p < 0.05) (Fig. 1). Using multivariable regression analysis, the NLR showed significant associations with CRP and ferritin levels (all p < 0.01). The PLR had significant associations with ESR and ferritin levels (p < 0.01 and 0.01, respectively). The lowest RT‒PCR Ct value was not associated with the NLR or PLR (p = 0.08 and 0.10, respectively). The NLR and PLR did not show significant differences in relation to the Taylor scores from CXR, in-hospital mortality or in-hospital endotracheal intubation (all p > 0.05).

Fig. 1
figure 1

The log-transformed neutrophil-lymphocyte ratio (NLR) showed significant associations with the levels of ferritin (a) and the levels of CRP (b). The platelet-lymphocyte ratio (PLR) showed significant associations with the levels of ferritin (c) and ESR (d) in patients with COVID-19.

Full size image

To explore the associations between the ANDC score and clinical outcomes, Wilcoxon rank-sum tests were performed among lower, median and higher ANDC scores group and the total length of stay. Significant between group differences were found in low score group (median:11, IQR:10–14 days), moderate score group (median:14.5, IQR:11–21 days) and high score group (median:24.5, IQR:13.3–44.3) (all p < 0.01 by post-hoc analysis). However, no significant group difference was found in in-hospital mortality (p = 0.73).

Discussion

In this study, we presented the clinical and laboratory characteristics of hospitalized patients with COVID-19 in Taiwan from April 2021 to July 2021. Our results showed that admitted patients presented with an average of 3 initial clinical symptoms, of which 23.8% had high-grade dyspnea. 9% of COVID-19 patients presented with anosmia/ageusia. Patients admitted to the tertiary medical center had a more severe degree of COVID-19 signs and symptoms, as evidenced by a higher severity of dyspnea, higher inflammatory markers, a higher proportion of use of monoclonal antibodies as rescue therapy for COVID-19, and a longer length of hospital stay. The presence of anosmia/ageusia symptoms was associated with younger age, lower BMI and lower systemic inflammatory markers. Lower ANDC scores indicate less disease severity in patients with the presence of anosmia/ageusia symptoms. In addition, female patients with COVID-19 had less systemic inflammation as evidenced by lower levels of CRP, ferritin, NLR, and better prognostic markers as evidenced by lower ANDC scores. Finally, laboratory tests revealed that NLR and PLR had significant associations with inflammatory markers and different severity of ANDC scores revealed the difference in total length of stay.

Anosmia/ageusia is now generally recognized as a relatively early symptom and/or complication in patients with COVID-19, with highly variable outcomes between studies and wide ranges of prevalence for olfactory dysfunction (0-98%) and gustatory dysfunction (0–89%)19,20. After 4 months, 52.9% of patients reported partial recovery, and 2.0% reported no recovery in one study21. However, the geographic variations were highly significant: Caucasians had a three times higher probability of chemosensory dysfunctions (54.8%) than Asians (17.7%)22. The psychological impact of anosmia/ageusia on activities of daily living can be very distressing, and it is important for physicians to be aware of it, as many patients reported that their symptoms were trivialized by health care providers23. Psychological distress, depression and anxiety might be related to anosmia/ageusia after COVID-19 infection24. Currently, there is no effective treatment, but recovery is generally possible over time25.

In our study, we demonstrated that patients with anosmia/ageusia had a younger age, lower BMI and higher number of initial clinical symptoms, which is compatible with previous findings10,22. Besides, lower inflammatory response was found in patient with anosmia/ageusia than those without had been documented, which is compatible with our study26. Furthermore, a mild disease course in patients with anosmia/ageusia was also demonstrated, as revealed in those patients with lower ANDC scores. From our study, no significant differences were found in the lowest RT‒PCR Ct value between patients with anosmia/ageusia and those without, suggesting that the viral load in the blood may not represent the viral load in the nasopharyngeal mucosa. The lower inflammatory response may reveal local rather than systemic inflammation in these patients as the previous report27. Gender difference in COVID-19 infection had also been found as male had high risk of critical illness and mortality rate28,29. Several factors may contribute to these differences such as lifestyle, genetic factor, the role of sex hormone, comorbidities and inflammation28. In our study, although no significant differences were found in clinical outcomes such as total length of stay or in-hospital mortality, however, a significantly better prognosis revealed by low ANDC score was found in female. Because the ANDC score was calculated by NLR, CRP and D-dimer, lower systemic inflammatory response in female may contribute to the good outcome.

The NLR and PLR are indicators of a systematic inflammatory response and have been used in several COVID-19 studies30,31. In one study, the NLR was used as an excellent predictor of disease severity of COVID-19 infection, in-hospital mortality, risk of death or clinical course deterioration32. The PLR is another parameter calculated from the complete blood count and is referred to as a nonspecific marker of inflammation that is associated with increased morbidity and mortality in COVID-1933. In our study, both the NLR and PLR showed significant associations with other inflammatory markers, even after adjustment for multiple covariates. Moreover, an association between CRP and the NLR has been found in previous studies34. The combined use of CRP with NLR may lower the CRP cutoff point in distinguishing between infectious and noninfectious inflammation in hemodialysis patients34. However, NLR and PLR markers were not associated with in-hospital mortality, in-hospital endotracheal intubation, or total length of stay in our study. These findings may be due to the relatively small sample size in our study. In addition, the NLR and PLR were significantly associated with the severity scores of chest computed tomography in a previous study35. In contrast, in this study, neither the NLR nor the PLR showed significant associations with the CXR Taylor score (all p > 0.05), which may relate to our sample size. The ANDC score had been used as prognostic marker for COVID-19 mortality. In our study, we also found that categorized the ANDC as low, moderate and high score could differentiate total length of stay in hospital.

Limitations.

The current study has several limitations. First, this study only recruited a relatively small sample of patients with COVID-19 in one tertiary medical center and one district hospital, which cannot represent all hospitalized patients in Taiwan. Nevertheless, our findings revealed the detailed clinical characteristics of admitted patients in the first surge wave of COVID-19 in 2021. Due to the early and appropriate measures conducted by the Taiwanese medical authority, the incidence of newly confirmed COVID-19 cases was lower during that time36. Second, we recorded initial clinical symptoms, including anosmia/ageusia symptoms, based on patients’ self-reports during history taking, which may have the potential for recall bias in this study. In addition, mild to moderate patients who might not have been admitted to the hospital, and very severe patients who initially had impaired consciousness, could not be recruited in the current study. The anomia/ageusia symptoms in these non-admitted patients could not be adequately assessed. Third, we did not perform chest computed tomography for all admitted patients. Furthermore, no vaccinations were performed on our medical staff, and therefore routine ear-nose-throat examinations or chest computed tomography imaging might have been difficult to perform for all patients at that time. In addition, no objective tests for anosmia were performed due to the lack of vaccination for medical personnel at that time, which may affect the accuracy of the reported prevalence of anosmia based solely on self-reports. Although we performed severity grading scores based on CXR and in the high dyspnea group, a significantly higher proportion of patients underwent endotracheal intubation during the hospital course. We acknowledge that this score could not totally represent the severity score from chest computed tomography. We were limited by the availability of the facility during the COVID-19 pandemic periods. However, the severity grading scores based on CXR were carefully evaluated by radiologists who were blinded to the patients’ clinical status. Nevertheless, our clinical grading system for dyspnea provided a feasible tool to assess the severity of dyspnea and showed an association with future in-hospital endotracheal intubation. These findings need to be verified by a prospective cohort study in the future.

Conclusions

We presented the clinical manifestations of hospitalized patients with COVID-19 in Taiwan during the first wave of the COVID-19 pandemic period in 2021. Patients with anosmia/ageusia had younger age, lower BMI, and lower inflammatory markers than those without anosmia/ageusia. Female patients had less severe inflammation and a better prognostic score. The prevalence of COVID-19 patients with olfactory and gustatory dysfunction showed geographic differences, but our study elucidated the clinical features and inflammatory profiles were similar. The ANDC score could be used as a prognostic marker in patients with COVID-19.

Methods

Subjects: From April 2021 to July 2021, patients admitted to two hospital wards (one tertiary medical center and one district hospital) with a diagnosis of COVID-19 based on nasopharyngeal screening and/or detection of SARS-CoV-2 by real-time reverse transcription polymerase chain reaction (RT‒PCR) were recruited37. At that time, more than 95% of COVID-19 infections were caused by B.1.1.1.7 (Alpha variant), which was identified by next-generation sequencing by the laboratory of the Taiwan Centre of Disease Control (https://www.cdc.gov.tw/EpidemicTheme/Detail/zKNFqsVWxUoUqE6fyhmBNA?archiveId=gkvCMsVNSYDHOByZjP2jkA). Using the chart review method, we recorded demographic data, the date of onset of symptoms and clinical features. Fourteen initial clinical symptoms, such as fever (defined as a body temperature > = 37.5 °C), dry cough, sore throat, sputum production, dyspnea, myalgia, diarrhea, nausea/vomiting, nasal congestion, conjunctivitis, anosmia, ageusia, headache and impaired consciousness, were documented as present or absent and summed as the total score of initial symptoms for further analysis. Major comorbidities such as hypertension, diabetes mellitus, dyslipidemia, obesity (defined as body mass index (BMI) > = 25 kg/m2), chronic obstructive pulmonary disease, chronic kidney disease, smoking, heart disease and active cancer treatment were also documented as present or absent and summed as total scores of comorbidities.

Dyspnea was classified as present or absent by the treating physician based on the patient’s subjective feeling of shortness of breath. Although the severity of dyspnea could be measured using the Borg scale38, it was difficult to perform this measurement for all patients with COVID-19. Rather, we graded the severity of dyspnea according to the nursing care record into 5 grades: grade 0: no dyspnea symptoms; grade 1: only one record of dyspnea during one-third of the day; grade 2: at least two records of dyspnea during two-thirds of the day; grade 3: dyspnea while performing basic daily activities all day; and grade 4: dyspnea while lying down and during preparation for intubation. We recorded the first available measurement of oxygen saturation by pulse oximetry together with the conditions under which it was measured, either room air or supplemental inspired oxygen.

Laboratory data, including total white blood cell (WBC) and differential counts, erythrocyte sedimentation rate (ESR), inflammatory markers (such as the NLR and PLR, CRP, D-dimer, ferritin), aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatine, lactate dehydrogenase (LDH), chest X-ray (CXR) and other imaging data, were documented at the most severe stage of the hospital course. The NLR, PLR and ANDC [Total points of ANDC = 1.14×age − 20 (years) + 1.63 ×NLR + 5.00×D − dimer(mg/L) + 0.14×CRP(mg/L)] were calculated by previous literatures9,39. The severity of ANDC was classified as lows score group (ANDC < 59), moderate score group (59 ≤ ANDC ≤ 101) and high score group (ANDC > 101) as previous literature9. The lowest RT‒PCR cycle threshold (RT‒PCR Ct) values were also collected and analyzed. The severity of the CXR was measured using the Taylor score; in brief, CXR findings were categorized as follows: 1 = normal; 2 = patchy atelectasis and/or hyperinflation and/or bronchial wall thickening; 3 = focal consolidation; 4 = multifocal consolidation; and 5 = diffuse alveolar changes40. Finally, the admission outcome was measured by length of stay, in-hospital mortality, in-hospital endotracheal intubation and emergency use of antiviral drugs such as remdesivir and/or monoclonal antibodies (such as tocilizumab, bamlanivimab, etesevimab, casirivimab and imdevimab) and a novel traditional Chinese medicine formula, Taiwan Chingguan Yihau (NRICM101) for treating COVID-19, and the medications were documented as used or not used for each patient41.

Institutional review board

This study was approved by the ethics review board of Chang Gung Memorial Hospital (approval number 202101086B0). The Chang Gung Memorial Hospital IRB has approved the waiver of informed consent. All methods were performed in accordance with the relevant guidelines and regulations.

Statistical analysis

Normally distributed continuous variables were assessed with the Shapiro‒Wilk test. Continuous variables were log-transformed in the statistical analysis whenever appropriate. A linear regression model was used to model the effect of the studied variables on continuous data. Data are presented as the medians and interquartile ranges (IQRs), and nonparametric tests such as the Wilcoxon rank-sum test were performed where appropriate. Pearson’s χ2 test was used for cross-tabulations. The significance level was set at 0.05. All data analyses were performed using SPSS (version 21.0, Chicago, IL).