Population Risk Factors for Severe Disease and Mortality in COVID-19 in the United States during the Pre-Vaccine Era: A Retrospective Cohort Study of National Inpatient Sample

Background-Previous studies on coronavirus disease 2019 (COVID-19) were limited to specific geographical locations and small sample sizes. Therefore, we used the National Inpatient Sample (NIS) 2020 database to determine the risk factors for severe outcomes and mortality in COVID-19. Methods-We included adult patients with COVID-19. Univariate and multivariate logistic regression was performed to determine the predictors of severe outcomes and mortality in COVID-19. Results-1,608,980 (95% CI 1,570,803–1,647,156) hospitalizations with COVID-19 were included. Severe complications occurred in 78.3% of COVID-19 acute respiratory distress syndrome (ARDS) and 25% of COVID-19 pneumonia patients. The mortality rate for COVID-19 ARDS was 54% and for COVID-19 pneumonia was 16.6%. On multivariate analysis, age > 65 years, male sex, government insurance or no insurance, residence in low-income areas, non-white races, stroke, chronic kidney disease, heart failure, malnutrition, primary immunodeficiency, long-term steroid/immunomodulatory use, complicated diabetes mellitus, and liver disease were associated with COVID-19 related complications and mortality. Cardiac arrest, septic shock, and intubation had the highest odds of mortality. Conclusions-Socioeconomic disparities and medical comorbidities were significant determinants of mortality in the US in the pre-vaccine era. Therefore, aggressive vaccination of high-risk patients and healthcare policies to address socioeconomic disparities are necessary to reduce death rates in future pandemics.


Introduction
Coronavirus disease 2019 (COVID- 19) is an ongoing pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. There is a broad spectrum of disease chronic pulmonary disorders, liver disease, and rheumatoid arthritis/collagen vascular disorders were part of the exhauster comorbidity index and were used directly. The remaining variables were extracted using the International Classification of Diseases, Tenth Revision, and Clinical Modification (ICD-10-CM) diagnosis codes. Comorbidities were selected based on the CDC guidelines (Table S1/Table S2).

Statistical Analysis
Descriptive analysis was conducted to determine the mean for continuous variables and the proportion for categorical variables. As NIS is a survey dataset, the proportion and the total number of events are reported with confidence intervals. First, we studied the distribution of socioeconomic variables, comorbidities, and complications among different severities of COVID-19. Univariate logistic regression was performed to determine the predictors of severe complications and mortality associated with COVID-19. Variables with significant p values (p < 0.05) were used in the multivariate logistic regression to determine the independent predictors of mortality and complications in COVID- 19. Analysis was carried out using StataCorp. 2021. Stata Statistical Software: Release 17. StataCorp LLC (College Station, TX, USA), BE version. Stata's every command and appropriate weights were used in all estimations. The overall fit was assessed using Receiver Operative Curves (ROC), and sensitivity analysis was performed using the value package. The study was exempt from institutional review board approval as the database uses previously collected de-identified data.
Compared with asymptomatic COVID-19 and COVID-19 bronchitis/LRTI, patients with COVID-19 pneumonia and COVID-19 ARDS were older and had fewer female patients. The COVID-19 ARDS group had a higher number of Hispanics compared with other groups. There were no significant differences in primary payer, zip code income quartiles, and hospital ownership. Large hospitals and urban teaching hospitals had more COVID-19 ARDS when compared with small and medium hospitals. Hospitals in the southern US had more COVID-19 cases in all subtypes (Table 1).
Stroke, chronic kidney disease (CKD), heart failure, cardiomyopathy, and ischemic heart disease were more prevalent in COVID-19 ARDS. Diabetes, dyslipidemia, malnutrition, OSA, overweight/obesity, long-term steroid/immunomodulator use, and hypertension showed an increasing prevalence with worsening severity. Substance abuse and nicotine use showed decreasing prevalence with worsening severity. COVID-19 LRTI and bronchitis were much more prevalent among patients with chronic pulmonary disorders. COVID-19 ARDS showed a higher proportion of severe liver disease when compared to other subtypes. The remaining variables did not show any clear trend with worsening severity (Table 2).  (Table 3).

Risk Factors for Severe Complications in COVID-19
On univariate analysis, ages 40-65 and age > 64 were associated with more complications when compared to the age group of 18 to 40 years. On multivariate analysis, the female sex was associated with lower odds of COVID-19 complications (aOR 0.72, 95% CI 0.70-0.73). Medicare (aOR 1.25, 95% CI 1.21-1.29) and Medicaid (aOR 1.10, 95% CI 1.06-1.14) were associated with higher complications when compared with private insurance. When compared with whites, all races were associated with higher odds of COVID-19 complications. The native American race had the highest risk for complications (aOR 1.81, 95% CI 1.63-2.02). Geographical locations with higher income (>25th percentile) were associated with lower odds of COVID-19 complications when compared with low income (<25th percentile). Medium and large hospitals were associated with higher odds of COVID-19 complications when compared with small hospitals. Similarly, urban nonteaching and urban teaching hospitals were associated with higher odds of COVID-19 complications when compared with rural hospitals. Private-not-profit hospitals were associated with lower odds for COVID-19 complications when compared with government hospitals (Table 4).
Stroke, malnutrition, and liver disease had an adjusted odds ratio (aOR) of more than 2. CKD, heart failure, primary immunodeficiency, long-term steroid/immunomodulatory use, COPD, complicated DM, obstructive sleep apnea, and elixhauser comorbidity index had an aOR of more than 1.24 for COVID-19 complications. Cardiomyopathy, uncomplicated DM, complicated hypertension, solid cancer, overweight and obesity, other chronic pulmonary disorders, and rheumatoid arthritis/collagen vascular disorders had an aOR from 1.01 to 1.24 (Table 5).

Risk Factors for Mortality in COVID-19
Elderly patients (age 65 and above) had the highest odds for mortality, followed by the age group 40 to 64 years. The female sex was associated with lower mortality (aOR 0.69, 95% CI 0.68-0.71). Medicaid and self-pay were associated with slightly higher odds of mortality when compared with private insurance. When compared with whites, blacks had similar mortality (aOR 1.03, 95% CI 0.99-1.07). Hispanics, Asian/pacific islanders, and other races had higher odds of mortality. Native Americans had the higher odds of mortality among all races (aOR 1.98, 95% CI 1.73-2.27). Geographical locations with higher income (>25th percentile) were associated with lower odds of mortality when compared with low income (<25th percentile).
Medium and large hospitals were associated with higher mortality when compared with small hospitals. Similarly, urban non-teaching and urban teaching hospitals were associated with lower mortality when compared with rural hospitals. Hospitals in all geographical locations had lower mortality when compared with hospitals in the northeast region. Private-owned hospitals were associated with lower mortality when compared with government hospitals (Table 6).  On multivariate analysis, stroke and liver disease had an aOR of more than 2 for mortality. Cerebral palsy, heart failure, complicated DM, malnutrition, primary immunodeficiency, long-term steroid/immunomodulatory use, COPD, and an aOR of more than 1.24 for mortality. CKD, cardiomyopathy, uncomplicated DM, solid cancer, overweight and obesity, rheumatoid/collagen vascular disorders, and other chronic pulmonary disorders had an aOR of 1.01 to 1.24 for mortality. Cardiac arrest, septic shock, and intubation had the highest odds for mortality (Table 7).

Discussion
Our study is one of the largest studies on the epidemiology of COVID-19. While it has strengthened some of the results from the prior smaller studies, it has also shown some key differences. In our study of 1.6 million COVID-19 hospitalizations, we stratified the mortality depending on the COVID-19 severity, which is unique. Interestingly, we noticed a 16.6% mortality rate in COVID-19 pneumonia patients and a heightened mortality rate of 54% in COVID-19-associated ARDS patients. This is consistent with prior literature, which showed an in-hospital mortality rate of 11.4% in the US during the pre-vaccine era [10]. Consistent with the prior literature, age, male sex [11], non-Caucasian race, and certain comorbidities [12] were associated with higher odds of experiencing complicated COVID-19 and death.
Our study showed that the elderly population (age above 65) has higher complications and mortality when compared to middle-aged adults. Extremes of age are known to have a weaker immune system predisposing to COVID-19 complications [13]. Anyway, elderly patients have more comorbidities that predispose them to COVID-19 complications. Similar to the multiple prior works of literature [14,15], the male sex is associated with increased complications and mortality associated with COVID-19 despite adjusting for other risk factors. Although several mechanisms have been proposed, estrogen is shown to suppress pro-inflammatory cytokines such as IL-6, which is implicated in pulmonary vessel leakage and lung damage seen with COVID-19 [16].
Some studies have shown minority races to have a worse prognosis [17,18], whereas other studies have shown no such association [19,20]. Although these studies may be limited to their respective geographical locations, our study is much more generalizable and shows that all non-white races have higher odds of COVID-19 complications. However, when compared with whites, blacks had similar mortality. Hispanics, Asian/pacific islanders, native Americans, and other races had higher mortality when compared with whites. Native Americans had the highest complications and COVID-19-related mortality, consistent with the current literature [21].
Although the exact cause is unclear, several reasons have been proposed, such as high poverty rates, high prevalence of risk factors, and lack of access to healthcare among minority races. However, our analysis showed that despite adjustment for these factors, there were higher odds for complications and mortality. This suggests other factors, such as genetic predisposition, may be seen in non-white races.
Larger and urban hospitals likely have higher mortality because they were referral centers catering to the sickest patients. It is unclear why hospitals in the northeast had higher complications and mortality. One possible explanation was that the northeastern US is one of the most densely populated areas and was the pandemic's epicenter, which may have led to resource overutilization [22]. The association between low-income zip code locations, government insurance or no insurance, and government hospitals with increased morbidity and mortality points to the deficiency in the US healthcare system to effectively care for the economically underprivileged during the pandemic [23].
Similar to the prior literature, stroke, CKD, heart failure, malnutrition, primary immunodeficiency, long-term steroid/immunomodulatory use, complicated DM, and liver disease were strong predictors of mortality [11,24,25]. Anyway, uncomplicated DM, cardiomyopathy, solid cancer, COPD, and other chronic pulmonary disorders were also predictors of COVID-19 mortality in our study. The poor cardiopulmonary health and impaired immune function associated with these diseases are likely responsible for this increased risk [26].
Diabetes is a well-established risk factor for severe COVID-19 [27,28]. Our study is consistent with a prior study which shows that complicated diabetes mellitus is a more decisive risk factor for COVID-19 mortality than uncomplicated diabetes mellitus [11]. Insulin resistance enhances pro-inflammatory cytokine production and a worse prognosis [28]. Hypertension is the most common comorbidity and is not associated with increased mortality risk, as seen in our study and prior studies [11]. With regards to stroke, COVID-19 can increase the risk of stroke, and stroke can increase the risk of severe COVID-19, as seen in our study [29].
Chronic liver disease was the most significant risk factor for COVID-19 severity. A meta-analysis by CDC showed a similar pooled odds ratio for mortality [30]. Immune dysregulation, coagulopathy, and intestinal dysbiosis have been postulated as potential reasons. Cancer is associated with worse adverse outcomes in patients with COVID-19 [31]. The pathophysiology stems from the immunosuppressive state as a result of malignancy and anti-cancer treatment modalities putting such a cohort at higher risk for adverse prognosis in COVID-19 infection [32,33] Our study shows that being overweight and obese is weakly associated with increased COVID-19 mortality (aOR 1.07, 95% CI 1.03-1.11). Prior literature showed that obesity is a strong risk factor for COVID-19 mortality, and the risk increases with increasing BMI [11,24]. One reason for the weaker association in our study could be the adjustment for complications of obesity, such as diabetes and chronic pulmonary disorders, which were significant for mortality. This may be the reason for the lack of association of smoking and hypertension for mortality as we adjusted for cardiovascular diseases such as heart failure and ischemic heart disease. This was shown in this study [11]. Initially, hypertension showed a negative association with mortality. However, hypertension was significantly associated with mortality after removing obesity and diabetes from the multivariable model. The lack of association of transplant status with mortality could also be due to the adjustment for immunosuppressants/steroid use.
The prior literature has shown intellectual disability, paralysis, mood or psychotic disorders, asthma, HIV/AIDS, and cystic fibrosis to be associated with increased mortality [26]. However, our study failed to demonstrate this association. Cerebral palsy was the only chronic neurological disease associated with increased mortality in our study. Although OSA was associated with higher complications, it was not associated with higher mortality. Consistent with the literature, the number of comorbidities predicted severe outcomes, as shown by the strong association of the elixhauser comorbidity index on a categorical and numerical scale. Cardiac arrest, septic shock, and intubation had very high odds of mortality. Therefore, it is important to aggressively manage these patients and consider hospice in case of poor functional outcomes as appropriate.

Limitations
Our study has some limitations. First, NIS is an inpatient database and does not track patients post-discharge. This is not captured if a patient was discharged alive but died at home or at a rehab facility. Second, NIS, being an administrative database, relies on ICD-10 codes, which is inferior to manual chart review. There is a lack of information on imaging studies, laboratory investigations, treatment regimens, and vaccination status, which could affect outcomes. As with any observational study, association does not mean causation, and conclusions should be drawn cautiously. Nevertheless, NIS is an extensive database that can reveal socioeconomic and medical factors in healthcare utilization and patient outcomes.

Conclusions
COVID-19 is associated with high inpatient mortality in the US during the prevaccination era. Racial and Socioeconomic predictors of health appear to play a crucial role in this pandemic. It is imperative that we develop healthcare policies to address these gaps to alter the course of future pandemics. The number of medical comorbidities and certain comorbidities such as stroke, complicated DM, CKD, Chronic liver disease, and malnutrition appear to be associated with the highest risk for COVID-19 complications and mortality. These patients need aggressive preventive care such as vaccination and close monitoring to prevent complications.
Funding: This research received no external funding.

Institutional Review Board Statement:
The study was exempt from institutional review board approval as the database uses previously collected de-identified data.
Informed Consent Statement: Consent was not needed as this was a retrospective chart review study of previously collected de-identified data.