key: cord-353116-7t1prfkr
authors: Bhargava, Ashish; Fukushima, Elisa Akagi; Levine, Miriam; Zhao, Wei; Tanveer, Farah; Szpunar, Susanna M; Saravolatz, Louis
title: Predictors for Severe COVID-19 Infection
date: 2020-05-30
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa674
sha: 
doc_id: 353116
cord_uid: 7t1prfkr

BACKGROUND: COVID-19 is a pandemic disease caused by a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Predictors for severe COVID-19 infection have not been well defined. Determination of risk factors for severe infection would enable identifying patients who may benefit from aggressive supportive care and early intervention. METHODS: We conducted a retrospective observational study of 197 patients with confirmed COVID-19 infection admitted to a tertiary academic medical center. RESULTS: Of 197 hospitalized patients, the mean (SD) age of the cohort was 60.6 (16.2) years, 103 (52.3%) were male and 156 (82.1%) were black. Severe COVID-19 infection was noted in 74 (37.6%) patients, requiring intubation. Patients aged above 60 were significantly more likely to have severe infection. Patients with severe infection were significantly more likely to have diabetes, renal disease, chronic pulmonary disease and had significantly higher white blood cell counts, lower lymphocyte counts, and increased C-reactive protein (CRP) compared to patients with non-severe infection. In multivariable logistic regression analysis, risk factors for severe infection included pre-existing renal disease (odds ratio [OR], 7.4; 95% CI 2.5-22.0), oxygen requirement at hospitalization (OR, 2.9; 95% CI, 1.3-6.7), acute renal injury (OR, 2.7; 95% CI 1.3-5.6) and initial CRP (OR,1.006; 95% CI, 1.001-1.01). Race, age and socioeconomic status were not identified as independent predictors. CONCLUSIONS: Acute or pre-existing renal disease, supplemental oxygen at the time of hospitalization and initial CRP were independent predictors for the development of severe COVID-19 infections. Every 1 unit increase in CRP increased the risk of severe disease by 0.06%.

A c c e p t e d M a n u s c r i p t 5 In December 2019, the first pneumonia cases of unknown origins were identified in Wuhan city, Hubei province, China [1] . The pathogen was identified as a novel coronavirus (nCoV), now called severe acute respiratory syndrome coronavirus (SARS-CoV), with the disease termed COVID-19 [2] . Because of its rapid spread, the World Health Organization has declared 2019-nCoV as pandemic [3] . As of April 28, 2020, a total of 3, 090, 844 confirmed cases had been reported in at 184 countries [4] . SARS-CoV-2 infections have been described among asymptomatic (who never developed symptoms) as well as pre-symptomatic patients (who are not yet symptomatic) [5] [6] [7] [8] [9] [10] .

The clinical spectrum from the largest cohort of symptomatic COVID-19 patients from China ranged from mild to critically ill cases [11] . Age was described as a strong risk factor for severe disease, with the highest case fatalities occurring in those 80 years and older [11] [12] [13] .

Preliminary data from the United States (U.S.) also suggested that adverse outcomes were most frequent among persons 85 years of age and older, but it also recognized that severe infections could occur in adults of any age group [14, 15] . Comorbid conditions of hypertension, diabetes, chronic lung and renal disease were also associated with severe infections and adverse outcomes [16] [17] [18] . Medications such as non-steroidal antiinflammatory drugs (NSAIDs), angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) were suggested to increase the severity of infection [19] , but currently there are no data to suggest a link between these medications and adverse outcomes. The determination of risk factors for severity in COVID-19 infection would enable identification of high-risk patients who may benefit from close monitoring, aggressive supportive care and early intervention.

A c c e p t e d M a n u s c r i p t 6 To address this question, we collected clinical data from a cohort of hospitalized patients with the aim of identifying predictors for developing severe COVID-19 infections. Obesity and severe obesity were defined according to CDC definitions [22] . Fever was defined as an axillary temperature of 37.5°C or higher. Lymphocytopenia was defined as a lymphocyte count of less than 1500 cells per cubic millimeter. Thrombocytopenia was defined as a platelet count of less than 150,000 per cubic millimeter. Acute renal injury or elevated creatinine on admission was defined as increase in serum creatinine by ≥0.3mg/dL (≥26.5 micromol/L) within 48 hours or increase in serum creatinine to ≥1.5 times baseline, which is known or presumed to have occurred within the prior seven days [23] . Patients with pre-existing renal disease were on dialysis, had a history of renal transplant, had uremic syndrome, or had a creatinine > 3mg/dL in prior admissions. Using the nine-digit zip code, the area deprivation rank for each individual patient was also obtained. The Area Deprivation Index (ADI) is a measurement of healthcare deprivation based upon where a person lives; it correlates with socioeconomic status [24] . A higher ADI means more deprived. On a national level the ADIs are ranked from 1 to 100 (1 = least deprived, 100 = most deprived) and on a state level it is from 1 to 10 (1 = least deprived, 10 = most deprived).

A c c e p t e d M a n u s c r i p t 8 

Statistical analysis was performed using SPSS v. 26.0 (Armonk, NY). Descriptive statistics were generated to characterize the study population. Continuous variables were described as the mean with standard deviation or median with interquartile range. Univariable analysis was done using Student's t-test, analysis of variance followed by multiple pairwise comparisons using the Bonferroni correction of the p-value, the Mann-Whitney U test and chi-squared analysis. Variables that were found to be significant (p <0.05) predictors of severity were then entered a multivariable logistic regression model using a forward likelihood ratio algorithm. When two variables were measuring the same underlying factor, the variable with the highest univariable measure of association was used in the model.

Results from the regression are reported as odds ratios with 95% confidence intervals. All There was no significant association found between race and severity (p -0.4). To examine this further, we also assessed the association between the median state ADI rank with both race and severity of disease. Although there was a significant association between race and median state ADI rank (White median state rank 6.0 (IQR: 2.8,8) vs. Black median rank 9.0 (IQR: 7,9.75)), (p -<0.0001), there was no association between median state ADI rank and severity of disease.

The most common symptoms at the onset of illness in the studied cohort were cough (141 including higher white blood cell counts, lower lymphocyte and platelet counts, and increased C-reactive protein (CRP) levels compared with those patients with non-severe infection.

Although patients with severe infection had significantly elevated procalcitonin levels, which raises concern for the presence of secondary bacterial infection, these patients also had acute renal injury on admission which may have caused the elevated procalcitonin. has become pandemic. While little has been reported regarding the predictors for severe COVID-19 infections, much is known regarding the risk factors and predictors for mortality [11, 25] . In our study we report pre-existing renal disease, supplemental oxygen requirement at admission, acute renal insufficiency, and initial CRP value as independent predictors of severe COVID-19 infections. It is also interesting that acute renal insufficiency (whether in patients with chronic kidney disease or normal baseline renal function) was associated with adverse outcome. SARS-CoV-2 is strongly suspected to use angiotensin converting enzyme 2 (ACE2) as its receptor, and ACE2 binding affinity has been shown to be one of the most important determinants of SARS-CoV infectivity [28] . Perhaps acute renal insufficiency reflects more efficient binding of SARS-CoV-2 to ACE2 given the location of ACE2 expression. Interestingly, ACE inhibitors and angiotensin II type-1 receptor blockers (ARBs), increase ACE2 expression, yet our analysis did not detect an association between ACE or ARB use, or hypertension or diabetes, and disease severity. Persons taking ACE or ARB at home but presenting with renal insufficiency generally have those medications held upon admission. While SARS-CoV-2 enters cells by binding to ACE2, ACE2 also reduces inflammation [19] . If one hypothesizes that ACE2 expression decreases inflammation, and that withholding drugs that increase its expression was done mainly in patients with acute renal insufficiency, perhaps a proinflammatory reaction to drug withdrawal could contribute to unfavorable outcome in such patients.

The need for supplemental oxygen for baseline hypoxia was an independent factor for severe disease in our study. A recently published study showed oxygen saturation below 90% despite oxygen supplementation was a powerful predictor for fatal outcome [29] . Given that ACE2 is expressed in lung epithelium, hypoxia may represent more avid binding to SARS-CoV-2 in those hosts. Interestingly, ACE2 is also expressed by endothelial cells, which represent one third of lung cells [30] . The endothelium functions to promote vasodilation, fibrinolysis, and anti-aggregation; thus, endothelial damage may lead to a hypercoagulable state [31] . Accumulation of coagulation factors in lungs can drive ARDS through activation of A c c e p t e d M a n u s c r i p t 13 protease activated receptors. Microvascular permeability from endothelial injury can also facilitate viral invasion [31] . Thus, direct effect of viral invasion and indirect effects thorough endothelial damage lead to severe hypoxia.

Initial C-reactive protein (CRP) level also associated with severe infection. Every 1 unit increase in CRP increased the risk of severe disease by 0.06%. CRP is a homopentameric acute-phase inflammatory protein. Baseline CRP values are influenced by age, gender, smoking status, weight, lipid levels, and blood pressure, and by genetics [32] . Recent studies have reported that cases of sever COVID-19 exhibit increased plasma levels of interleukin (IL) 2, IL6, IL7, IL10, granulocyte colony-stimulating factor (GCSF), tumor necrosis factor (TNF) alpha, and others [33] . IL-6 is the main inducer of CRP gene expression, with IL-1 and TNF-alpha also playing a role [32] . Elevated CRP may reflect severe disease as an indirect marker of elevated IL-6 and TNF-alpha. CRP not only reflects inflammation, it also enhances the immune response. CRP can be irreversibly dissociated into monomeric subunits termed monomeric or modified CRP (mCRP) at either high concentrations of urea or elevated temperatures in the absence of calcium, and mCRP promotes monocyte chemotaxis and recruitment of circulating leukocytes to areas of inflammation. Modified CRP also binds immunoglobulin G (IgG) Fc receptors in an interaction leading to release of proinflammatory cytokines [32] . Thus, elevated CRP at admission may both reflect significant inflammation and itself drive further inflammation. And given that elevated levels of urea promote formation of mCRP, it fits that acute and chronic kidney disease may be associated with adverse outcomes.

In our data, having a known sick contact was associated with lower risk of severe disease. Perhaps some of those patients with a sick contact knew they were at risk of A c c e p t e d M a n u s c r i p t 14 exposure and therefore were already attempting to minimize that risk through hand-hygiene, masks, or physical isolation or using separate bathrooms, thus decreasing the potential amount of virus to which they were exposed. This would, however, only account for patients where someone else was symptomatic or exposed soon enough for the patient to be able to take precautions. Patients with known sick contacts may also have presented sooner due to a heightened suspicion of having contracted COVID-19, and thus received care more rapidly [34] . Recall bias may also have contributed. One other potential explanation is that the sickest patients were more confused or were intubated rapidly and therefore could not provide a history of sick contact, which would have impacted the recording of a sick contact in the electronic medical record and thus our results.

Patients over 60 years old were significantly more likely to have severe infection. In multivariable analysis, however, age, was not found to be an independent predicting factor for the severe infection. This may reflect that the occurrence of kidney disease tends to be higher in older people and kidney disease was the stronger predictor in the model. Older age has been significantly associated with death in previous studies [35, 36] . This might be related due to less robust immune responses as older age has been linked with declined of comorbid conditions including significant elevation in creatinine on admission than those not admitted to ICU [13] . Thus, age in our studied cohort might have been a confounding factor to elevated serum creatinine and C-reactive protein.

During the COVID-19 pandemic, racial and ethnic minorities especially blacks, have been reported to be severely or disproportionately impacted. Our study did not show a disparity in severity by race, so we also investigated the relationship by ADI (as a proxy for socioeconomic status). No association was found between ADI and severity or between race and severity after controlling for ADI.

Our study has several limitations. This was a single institution study among all the admitted patients which makes generalization of interpretations difficult. Because of the retrospective nature of the study design, all variables in the studied patients were not available. Therefore, the role of some of these variables in predicting severity of the infection could have been underestimated. Last but not the least, the small sample size of our study and a predominantly black and overweight/obese cohort could have limited the generalizability of interpretation for some of the findings (for eg: race). Nonetheless our study did involve a population of black patients in the Detroit area and can provide valuable information on which factors are most significant predictors of severe disease in that population. 

Pneumonia of unknown cause -China

Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

World Health Organization Home Page

Evidence of SARS-CoV-2 Infection in Returning Travelers from Wuhan, China

Clinical characteristics of 24 asymptomatic infections with COVID-19 screened among close contacts in Nanjing

Clinical outcome of 55 asymptomatic cases at the time of hospital admission infected with SARS-Coronavirus-2 in Shenzhen

Presumed Asymptomatic Carrier Transmission of COVID-19

Asymptomatic and Presymptomatic SARS-CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility -King

Characteristics of and Important Lessons From the

COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention

Clinical Characteristics of Coronavirus Disease 2019 in China

Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China

Severe Outcomes Among Patients with Coronavirus Disease 2019 (COVID-19) -United States

Characteristics and Outcomes of 21 Critically Ill Patients With COVID-19 in Washington State

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Novel Coronavirus Pneumonia Emergency Response Epidemiology T

COVID-19 pathophysiology: A review

Fibrinolytic abnormalities in acute respiratory distress syndrome (ARDS) and versatility of thrombolytic drugs to treat COVID-19

Role of C-Reactive Protein at Sites of Inflammation and Infection

Transcriptomic characteristics of bronchoalveolar lavage fluid and peripheral blood mononuclear cells in COVID-19 patients

Journey of a Thai Taxi Driver and Novel Coronavirus

Risk Factors Associated With Acute Respiratory Distress Syndrome and Death in Patients With Coronavirus Disease

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Abbreviations: n: Number, OR: Odds ratio, CI: Confidence interval, ACEI: Angiotensin converting enzyme inhibitor, ARBs: Angiotensin II receptor blockers

A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t