Evaluation of the Prognosis of COVID-19 Patients According to the Presence of Underlying Diseases and Drug Treatment

Certain underlying diseases such as diabetic mellitus and hypertension are a risk factor for the severity and mortality of coronavirus disease (COVID-19) patients. Furthermore, both angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) are controversial at role in the process of COVID-19 cases. The aim of the study was to investigate whether underlying diseases and taking ACEi/ARBs, affect the duration of hospitalization and mortality in patients with confirmed COVID-19. Medical usage claims data for the past three years until 15 May 2020, from the “CORONA-19 International Cooperation Research” project was used. We analyzed the medical insurance claims data for all 7590 coronavirus (COVID-19) patients confirmed by RT-PCR tests nationwide up to 15 May 2020. Among the comorbidities, a history of hypertension (hazard ratio [HR], 1.51; 95% confidence interval [CI], 1.056–2.158) and diabetes (HR, 1.867; 95% CI, 1.408–2.475) were associated significantly with mortality. Furthermore, heart failure (HR, 1.391; 95% CI, 1.027–1.884), chronic obstructive pulmonary disease (HR, 1.615; 95% CI, 1.185–2.202), chronic kidney disease (HR, 1.451; 95% CI, 1.018–2.069), mental disorder (HR, 1.61; 95% CI, 1.106–2.343), end stage renal disease (HR, 5.353; 95% CI, 2.185–13.12) were also associated significantly with mortality. The underlying disease has increased the risk of mortality in patients with COVID-19. Diabetes, hypertension, cancer, chronic kidney disease, heart failure, and mental disorders increased mortality. Controversial whether taking ACEi/ARBs would benefit COVID-19 patients, in our study, patients taking ACEi/ARBs had a higher risk of mortality.


Introduction
After its emergence in China in December 2019, the coronavirus disease  caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spread to South Korea in January 2020. As of 18 April 2021, one year after the first outbreak of COVID-19 in Korea, the cumulative number of cases was 104,006, with a total of 1797 deaths (http://ncov.mohw.go.kr/, accessed on 18 April 2021), and the country was experiencing the fourth wave of the epidemic. The outbreak of the infection in specific groups, such as some religious groups, nursing homes, mental hospitals, and occupational groups (e.g., call centers and others) had epidemiological implications. The characteristics of these groups affected the prognosis of diseases according to their demographics and the presence or absence of underlying diseases [1][2][3]. Older age and the presence of underlying diseases have been shown to be factors leading to the severe stage of COVID-19 [4] and death of these patients [5,6]. In particular, the COVID-19 pandemic has shown that the presence of certain underlying diseases such as cancer, hypertension, stroke, and DM, are risk factors for severe disease and mortality of COVID-19 patients in various countries [7][8][9].
Furthermore, the angiotensin converting enzyme is used by SARS-CoV-2 to enter the target cell [10]. It has been reported that both angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) can significantly increase the mRNA expression of cardiac angiotensin-converting enzyme 2 [11]. Based on this evidence, it has been hypothesized that these drugs may play a role in COVID-19 disease severity [12]; however, there is a lack of evidence on the potential negative effects of these drugs on COVID-19 [13]. ACEi/ARBs are renin-angiotensin-aldosterone system (RAS) inhibiting agents and is commonly used anti-hypertensive agents to protect cardiovascular and cerebrovascular diseases, also in patients with chronic kidney disease for anti-proteinuric effects. Therefore, this concern remains to be elucidated.
We explored the risk factors related to the mortality and hospitalization of COVID-19 patients, and investigated their impact. In addition, we investigated the effects of underlying diseases (such as chronic kidney disease, hypertension, diabetes mellitus, ischemic heart disease, heart failure, cancer, chronic obstructive pulmonary disease, mental disorder, and cerebrovascular disease [14][15][16][17][18]) and a history of ACEi/ARBs treatment on mortality and length of hospital stay in these patients.

Patients and Data
Medical usage claims data for the past three years until 15 May 2020 from the "CORONA-19 International Cooperation Research" project hosted by the Ministry of Health and Welfare of the Republic of Korea and Health Insurance Intensive Assessment Service were used. In the Republic of Korea, under the entire national health insurance system, claims data that include all the costs incurred when using medical services at medical institutions nationwide, are routinely collected. This study used the national health insurance system claims database from three years prior to the date each patient was diagnosed with COVID-19 to 23 May 2020. Patients were defined as those diagnosed with COVID-19 based on by real-time reverse transcription polymerase chain reaction (RT-PCR) testing at the Korea Centers for Disease Control and Prevention by 15 May 2020. The mortality cases were included only during the COVID-19 infection status and it was excluded when patient died after COVID-19 infection was cleared up.
Using the specific International Classification of Diseases Tenth Revision codes, patients with the following underlying diseases for three years before the diagnosis of COVID-19 were selected: chronic kidney disease, hypertension, diabetes mellitus, ischemic heart disease, heart failure, cancer, chronic obstructive pulmonary disease, mental disorder, and cerebrovascular disease, as well as those with a history of ACEi/ARBs treatment. Patients who received an ACEi/ARBs prescription at least 2 times, and the next prescription interval after the first prescription was within 30 days were defined as those who received an ACEi/ARBs prescription by tracking the drug prescription history back from the patients claims history.
The Charlson Comorbidity Index (CCI) is the most widely used method among comorbidity index adjustment methods. It is a method to adjust the sum of these weights after assigning a constant weight from 1 to 6 points to 19 diseases defined through a medical records survey. In this study, we used the Quan algorithm [19] that combines the advantages of the existing algorithm. Additionally, the Quan algorithm is more code defined than other algorithms, which increases the chance to observe comorbid diseases. Therefore, it is thought that this study will be sensitively applied to define patients who have various diseases as underlying diseases at the same time [20] (Supplementary Table S1).

Statistical Analyses
Categorical variables were reported as frequencies (percentages with 95% confidence intervals (CIs)) and continuous variables as means (with standard deviations (SDs)) according to the type of distribution. We calculated Kaplan-Meier survival estimates and used the log-rank test to compare groups in terms of survival. The association of risk factors with time to death was assessed using univariable and multivariable Cox proportional hazards regression models. A generalized linear model was used to investigate the association of the risk factors with hospitalization. In order to investigate the factors affecting the length of hospital stay on of the prognostic indicators, a subgroup analysis was performed according to socioeconomic status that excluded patients who died during the study period.
Multivariable models were developed for demographics and diabetic mellitus and hypertension as the underlying disease comorbidities in the basic model, with drug treatment and additional underlying diseases added one by one, to measure their effect on mortality and hospitalization. In addition, the use of ACEi/ARBs, or history, and the use of a ventilation during the treatment were also applied to the model.
Cox proportional hazards regression was used to analyze whether the effect of the drug, which was discussed as a COVID-19 drug in the absence of a COVID-19 treatment, affects mortality, the prognosis of COVID-19 patients. The effect of the drug was investigated by adjusting for underlying diseases that increase the risk of death, and whether ACEi/ARBs was prescribed (Model 2), or whether the patient was treated with a ventilator (Model 3). In addition, to consider the importance of each underlying disease, we performed an analysis of the scores of the CCI. To reflect the level of management of the underlying disease, we conducted subgroup analysis for each group according to the disease severity; groups were also classified based on specific drug treatment. We used insurance types as surrogate variables to obtain economic information. To assess the degree of underlying disease management, the socioeconomic conditions that affect chronic disease management were divided into medical aid and National Health Insurance Service beneficiaries and then analyzed.
We calculated Kaplan-Meier survival estimates and used log-rank tests to compare groups for each variable by sex, socioeconomic status, and underlying disease in terms of survival. It showed that there were differences in the estimated survival curve in the groups according to the sex, socioeconomic status, and underlying disease ( Figure 1).
We calculated Kaplan-Meier survival estimates and used log-rank tests to compare groups for each variable by sex, socioeconomic status, and underlying disease in terms of survival. It showed that there were differences in the estimated survival curve in the groups according to the sex, socioeconomic status, and underlying disease ( Figure 1).

Observation Time and Main Outcomes
For the 7590 patients, the cumulative observation time was 164,329 patient-days from admission to the end of the follow-up (median observation time, (range, 1-128) days; IQR, 12-29 days) and there were 225 deaths (overall mortality, 2.96%). After a median follow-up of two days from admission to ICU transfer (IQR, 0-6; range, 0-47), there were 30 ICU deaths (13.3%). The distribution of the ICU length of stay is presented in Figure S1.

Observation Time and Main Outcomes
For the 7590 patients, the cumulative observation time was 164,329 patient-days from admission to the end of the follow-up (median observation time, (range, 1-128) days; IQR, 12-29 days) and there were 225 deaths (overall mortality, 2.96%). After a median followup of two days from admission to ICU transfer (IQR, 0-6; range, 0-47), there were 30 ICU deaths (13.3%). The distribution of the ICU length of stay is presented in Figure S1.

Subgroup Analyses
In the model based on age, sex, diabetes mellitus, and hypertension, in the group with high economic status, diabetes mellitus and hypertension had a statistically significant effect on the increase in hospital stay. In particular, hypertension was significant in all groups regardless of economic status. The size of the coefficient according to the economic status of hypertension was high in the group with low economic status. When each underlying disease was added to the basic model, hypertension (coefficient: 0.159, 95% CI: 0.008-0.310), cardiovascular disease (coefficient: 0.203, 95% CI: 0.038-0.372), a high socioeconomic status, and chronic obstructive pulmonary disease (coefficient: 0.1924, 95% CI: 0.0433-0.3431) were associated with a significantly longer duration of hospitalization ( Figure 3).   When the CCI was as a continuous variable, CCI were statistically significant at high socioeconomic status in Model 1 (coefficient: 0.0303, 95% CI: 0.0168-0.0441), Model 2 (coefficient: 0.029, 95% CI: 0.0154-0.0429 and Model 3 (coefficient: 0.0284, 95% CI: 0.0148-0.0423). When CCI was categorized by level, a higher CCI score was associated with a longer hospital stay. These results were statistically significant in patients with a high socioeconomic status. In patients with low socioeconomic status, the trend was similar, but was not statistically significant ( Figure 4). The effect of the underlying disease and whether or not taking ACEi/ARBs in the past or present on hospital stay according to the subgroup analysis based on the variable for classifying insurance types as a substitute variable for the socioeconomic status. Note: Model0: adjusted covariates age, sex and insurance type, Model1: adjusted covariates age, sex, insurance type and ischemic heart disease, Model2: adjusted covariates age, sex, insurance type and hear failure, Model3: adjusted covariates age, sex, insurance type and chronic obstructive pulmonary disease, Model4: adjusted covariates age, sex, insurance type and cancer, Model5: adjusted covariates age, sex, insurance type and chronic kidney disease, Model6: adjusted covariates age, sex, insurance type and end stage renal disease, Model7: adjusted covariates age, sex, insurance type and angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs), Model8: adjusted covariates age, sex, insurance type and mental disorder. Model9: adjusted covariates age, sex, insurance type and cerebrovascular disease. DM, diabetic mellitus; HTN hypertension; IHD ischemic heart disease; HF heart failure; COPD chronic obstructive pulmonary disease; CAN cancer; CKD chronic kidney disease; ESRD end stage renal disease; ACEi/ARBs angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs); MEN mental disorder; CVD cerebrovascular disease.
When the CCI was as a continuous variable, CCI were statistically significant at high socioeconomic status in Model 1 (coefficient: 0.0303, 95% CI: 0.0168-0.0441), Model 2 (coefficient: 0.029, 95% CI: 0.0154-0.0429 and Model 3 (coefficient: 0.0284, 95% CI: 0.0148-0.0423). When CCI was categorized by level, a higher CCI score was associated with a longer hospital stay. These results were statistically significant in patients with a high socioeconomic status. In patients with low socioeconomic status, the trend was similar, but was not statistically significant ( Figure 4).

Figure 4.
The effect of the Charlson Comorbidity Index and whether or not taking ACEi/ARBs in the past or present on hospital stay according to the subgroup analysis based on the variable for classifying insurance types as a substitute variable for the socioeconomic status. Note: Model1: adjusted covariates age, sex, insurance type and Charlson Comorbidity Index as continuous variables, Model2: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) and Charlson Comorbidity Index as continuous variables, Model3: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs), ventilator and Charlson Comorbidity Index as continuous variables, Model4: adjusted covariates age, sex, insurance type and Charlson Comorbidity Index as categorical variables with reference CCI score0 (without underlying disease), Model5: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) and Charlson Comorbidity Index as categorical variables with reference CCI score0 (without underlying disease), Model6: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs), ventilator and Charlson Comorbidity Index as categorical variables with reference CCI score is 0 (without underlying disease). CCI Charlson Comorbidity Index score; CCI1 Charlson Comorbidity Index score=1; CCI2 Charlson Comorbidity Index score=2; CCI3 Charlson Comorbidity Index score ≥ 3.

Discussion
The prediction of the prognosis of patients confirmed with COVID-19 is generally based on age and the presence or absence of an underlying disease. Men had a higher mortality rate than women, and older adults had a higher mortality rate than younger adults, which is consistent with the results of previous studies. [21][22][23][24]. In univariable analysis, the underlying disease had a significant effect on increasing the mortality rate [1,23,25]. To investigate the effect of patient mortality, not only sex and age but also diabetes, hypertension with high incidence rate and consistent management [25] and whether to take ACEi/ARBs [13], were applied as adjustment variables together to affect the COVID-19 fatality rate. In addition, cancer, chronic kidney disease, heart failure, and mental disorder were added to the model. Patients with these underlying diseases and patients with a history of ACEi/ARBs treatment showed a higher mortality rate. As a result of performing a subgroup analysis according to economic conditions, hypertension showed a significant effect on mortality at both economic levels, in all the models. This result was consistent with that of a previous study [26]. In a study conducted in Lombardy, Italy [27], inpatients receiving ICU care for COVID-19 had an increased risk of DM (HR, 1.18; 95% CI, 1.01-1.39), and in a study of patients hospitalized for COVID-19 in Spain [22], patients with cancer and DM had an increased mortality rate. In our study, cancer (HR: 1.615, 95% CI: 1.185-2.202) and DM (HR: 1.867, 95% CI: 1.408-2.475) in the base model and in the model that included other underlying diseases were associated with a significantly increased risk of mortality in patients with COVID-19 (Supplementary Table S2), which is consistent with the results of previous studies. Based on these results, there is a need to . The effect of the Charlson Comorbidity Index and whether or not taking ACEi/ARBs in the past or present on hospital stay according to the subgroup analysis based on the variable for classifying insurance types as a substitute variable for the socioeconomic status. Note: Model1: adjusted covariates age, sex, insurance type and Charlson Comorbidity Index as continuous variables, Model2: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) and Charlson Comorbidity Index as continuous variables, Model3: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs), ventilator and Charlson Comorbidity Index as continuous variables, Model4: adjusted covariates age, sex, insurance type and Charlson Comorbidity Index as categorical variables with reference CCI score0 (without underlying disease), Model5: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs) and Charlson Comorbidity Index as categorical variables with reference CCI score0 (without underlying disease), Model6: adjusted covariates age, sex, insurance type, angiotensin converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARBs), ventilator and Charlson Comorbidity Index as categorical variables with reference CCI score is 0 (without underlying disease). CCI Charlson Comorbidity Index score; CCI1 Charlson Comorbidity Index score = 1; CCI2 Charlson Comorbidity Index score = 2; CCI3 Charlson Comorbidity Index score ≥ 3.

Discussion
The prediction of the prognosis of patients confirmed with COVID-19 is generally based on age and the presence or absence of an underlying disease. Men had a higher mortality rate than women, and older adults had a higher mortality rate than younger adults, which is consistent with the results of previous studies. [21][22][23][24]. In univariable analysis, the underlying disease had a significant effect on increasing the mortality rate [1,23,25]. To investigate the effect of patient mortality, not only sex and age but also diabetes, hypertension with high incidence rate and consistent management [25] and whether to take ACEi/ARBs [13], were applied as adjustment variables together to affect the COVID-19 fatality rate. In addition, cancer, chronic kidney disease, heart failure, and mental disorder were added to the model. Patients with these underlying diseases and patients with a history of ACEi/ARBs treatment showed a higher mortality rate. As a result of performing a subgroup analysis according to economic conditions, hypertension showed a significant effect on mortality at both economic levels, in all the models. This result was consistent with that of a previous study [26]. In a study conducted in Lombardy, Italy [27], inpatients receiving ICU care for COVID-19 had an increased risk of DM (HR, 1.18; 95% CI, 1.01-1.39), and in a study of patients hospitalized for COVID-19 in Spain [22], patients with cancer and DM had an increased mortality rate. In our study, cancer (HR: 1.615, 95% CI: 1.185-2.202) and DM (HR: 1.867, 95% CI: 1.408-2.475) in the base model and in the model that included other underlying diseases were associated with a significantly increased risk of mortality in patients with COVID-19 (Supplementary Table S2), which is consistent with the results of previous studies. Based on these results, there is a need to develop more elaborate treatment guidelines and strategies for managing common underlying diseases that are associated with increased mortality, in order to improve the prognosis of COVID-19 patients with underlying conditions. Among COVID-19 patients, patients with an underlying disease had a higher mortality risk, and hypertension and CVD patients at all levels of socioeconomic status had longer hospital stays; however, patients with CVD as an underlying disease had longer hospital stays at a lower socioeconomic condition. Despite the government's full support for the cost of COVID-19 treatment, these results suggest that the response to treatment of COVID-19 is affected by the degree of chronic disease control among vulnerable groups. Therefore, it is important to improve the management of chronic diseases that increase vulnerability to severe COVID-19. Due to the limitations in the billing data, the definitions of the ACEi/ARBs-prescribed patient group and operational definition of the level of economy were unclear. Despite the fact that the number of patients in the low-economic group occupied a small proportion of the total number of patients, it had a significant effect. In addition, data on clinical test results at an individual patient level were not available, so it was not possible to adjust for individual variations in response to treatment. However, despite these limitations, with the inclusion of all confirmed cases of COVID-19 in Korea, evaluating effects of the underlying diseases and drugs on the prognosis, provides evidence that could contribute to the prevention and treatment of COVID-19. In our study, the hospitalization of patients with low economic status, that is, patients with medical aid, was higher than those who were National Health Insurance Service beneficiaries. Unlike previous research findings that the type of insurance did not significantly affect the mortality rate, [28], in our study, when investigating the length of stay for alleviating diseases, excluding the deceased, the type of insurance had an effect on the length of hospital stay. This suggested that the management of the underlying disease may have had an effect on recovery; however, depending on the underlying disease, this effect was significant in both insurance types.
Our data showed that most of the deaths occurred within seven days of the average hospital stay; thus, the duration of hospitalization was an outcome, and when the subgroup analysis was performed by economic level, the effect of the underlying diseases was investigated after excluding deceased patients. In the low-economic group, even if the patient with underlying disease received the same quality of management as the patient in the high-economic group, the effect of the underlying disease may have indirectly influenced the patient's prognosis.
For the same underlying disease, the low-economic group had a greater effect on hospital stay than the high-level group with hypertension and cerebrovascular bleeding being significant, and in particular, the effect of hypertension was significantly higher in all the models. The underlying diseases affecting the length of stay were hypertension and cerebrovascular disease in all the groups, regardless of the economic level. In the higheconomic group, patients with diabetes, cardiovascular diseases such as ischemic heart disease, and heart failure, as well as cancer showed an increased probability of an increased length of hospital stay. On the other hand, for chronic obstructive pulmonary disease, this effect was statistically significant in the low-economic group. Chronic obstructive pulmonary disease was not significant in the analysis of the association between the underlying disease and effect of mortality, which was different from the findings of previous studies [29]. However, during the recovery process, it was found that underlying diseases whose management was neglected affected the group. Those with underlying diseases such as hypertension and cerebrovascular disease were more likely to have a longer length of hospital stay regardless of their economic level. Based on the overall results, it appears that patients with hypertension, diabetes, and cerebrovascular disease had a longer recovery time, and those with chronic obstructive pulmonary disease had a longer recovery in a group with low economic status. There is a limit to the adjustment of individual clinical conditions during hospitalization due to the characteristics of the claim data, although the inclusion of all COVID-19 patients nationwide will be meaningful for infectious diseases that originate from a specific group.
The use of hydroxychloroquine, lopinavir/ritonavir (Kaletra), rivabirin, steroids, and IVIG for the treatment of COVID-19 patients were analyzed by gender, age, and socioeconomic status while adjusting the underlying disease model, except in the case of hydroxychloroquine. In this case, it was found that there was an increased mortality rate of patients who were treated with it. Its use can be interpreted as treatment for the purpose of symptom relief in patients with severe symptoms, without a treatment option. After adjusting for whether ACEi/ARBs were prescribed or not, the effect of medicine prescription on mortality showed the same pattern as the above result. When adjusting for ventilator use, the effect of drug prescription on death, except for steroids was not statistically significant. The results of increasing the risk of COVID-19 death with the use of ACEi/ARBs were inconsistent with the results of previous studies, which are still controversial [30]. There is sufficient scientific evidence that ACEi/ARBs upregulates the human ACE2 receptor which has been shown to be a viral entry of SARS-CoV-2 into cells [31][32][33]. Therefore, the use of ACEi/ARBs has been suggested to increase the risk of COVID-19 infection and adverse outcomes with COVID-19 infection which is concordant with our data. Further large-scale epidemiological studies should be elucidated.

Conclusions
Our data includes COVID-19 patients across the country. In this data, underlying diseases such as diabetes, hypertension, cancer, chronic kidney disease, heart failure, and mental disorders increased mortality. It is controversial whether taking ACEi/ARBs would benefit COVID-19 patients but in our study, patients taking ACEi/ARBs had a higher risk of mortality. The underlying diseases affecting hospital stay excluding the deceased were hypertension and cerebrovascular disease regardless of the type of insurance, and COPD in the medical aid group and diabetes and heart failure in the medical insurance group had an effect on increasing the length of stay. In addition, the higher the comorbid disease index, the longer the hospital stay.
Author Contributions: Research idea and study design, E.K., Y.C.K., J.P.L. and H.K.; data analysis/interpretation, E.K., Y.C.K., J.Y.P., J.J., J.P.L. and H.K.; drafted and revised the paper, E.K., Y.C.K. and J.P.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement: Ethical review and approval were waived for this study, due to the use of secondary data anonymized aggregated data to which access was restricted.

Informed Consent Statement: Not applicable.
Data Availability Statement: Under the management and supervision of the Ministry of Health and Welfare and the Health Insurance Review and Assessment Service of South Korea, data cannot be available.

Acknowledgments:
The authors appreciate healthcare professionals dedicated to treating COVID-19 patients in Korea, and the Ministry of Health and Welfare and the Health Insurance Review & Assessment Service of Korea for sharing invaluable national health insurance claims data in a prompt manner.

Conflicts of Interest:
The authors declare no conflict of interest.