key: cord-0834063-6zrn426f
authors: St. Sauver, Jennifer L.; Lopes, Guilherme S.; Rocca, Walter A.; Prasad, Kavita; Majerus, Michelle R.; Limper, Andrew H.; Jacobson, MD Debra J.; Fan, Chun; Jacobson, Robert M.; Rutten, Lila J.; Norman, Aaron D.; Vachon, Celine M.
title: Factors Associated With Severe COVID-19 Infection Among Persons of Different Ages Living in a Defined Midwestern US Population
date: 2021-07-05
journal: Mayo Clin Proc
DOI: 10.1016/j.mayocp.2021.06.023
sha: bec6145f82984b4f4978feecff68719a7a1699f3
doc_id: 834063
cord_uid: 6zrn426f

Objective To identify risk factors associated with severe COVID-19 infection in a defined, Midwestern, U.S. population overall and within different age groups. Patients and Methods We used the Rochester Epidemiology Project research infrastructure to identify persons residing in a defined 27-county Midwestern region who had a positive PCR test for COVID-19 between 3/1/2020 and 09/30/2020 (N=9,928). Age, sex, race, ethnicity, body mass index, smoking status, and 44 chronic disease categories were considered as possible risk factors for severe infection. Severe infection was defined as hospitalization or death due to COVID-19. Associations between risk factors and severe infection were estimated using Cox proportional hazard models overall and within 3 age groups (0-44, 45-64, and 65+ years). Results Overall, 474 (4.8%) persons developed severe COVID-19 infection. Older age, male sex, non-white race, Hispanic ethnicity, obesity, and a higher number of chronic conditions were associated with increased risk of severe infection. After adjustment, 36 chronic disease categories were significantly associated with severe infection. The risk of severe infection varied significantly across age groups. In particular, persons 0-44 years with cancer, chronic neurologic disorders, hematologic disorders, ischemic heart disease, and other endocrine disorders had a greater than 3-fold increased risk of severe infection compared to persons of the same age without those conditions. Associations were attenuated in older age groups. Conclusion Older persons are more likely to experience severe infections; however, severe cases occur in younger persons as well. Our data provide insight regarding younger persons at especially high risk of severe COVID-19 infection.

Since Coronavirus Disease 2019 (COVID-19) was first recognized in December, 2019, 1 investigators have attempted to identify characteristics associated with higher risk of adverse health outcomes following infection. Several studies have shown that older age, male sex, higher body mass index (BMI), and the presence of diabetes, hypertension, cardiovascular disease, and other co-morbidities are associated with more severe outcomes among hospitalized patients. [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] Other studies have reported similar characteristics associated with risk of hospitalization and adverse outcomes among persons presenting to the emergency department. 14, 15 However, many persons with COVID-19 develop mild or no symptoms and do not require an emergency department visit or hospitalization. Failure to identify and include persons who develop milder cases of COVID-19 may overestimate the risk of severe disease, and may result in a failure to recognize risk characteristics for severe COVID-19 infection. 16 Capturing less severe COVID-19 cases has been especially difficult in the United States because access to testing has been extremely limited in some populations. In addition, analyses of cases reported to state and territorial health departments may miss potential risk factors that are not routinely included as part of the case report forms. Finally, disease severity has been strongly associated with increasing age. However, severe cases do occur in younger persons, and there is limited data on whether known risk factors may differentially affect different age groups.

To address this gap, we studied persons residing in a 27-county region of south-eastern Minnesota and west-central Wisconsin, using the resources of the Rochester Epidemiology Project (REP) medical records-linkage system. 17 The REP links together J o u r n a l P r e -p r o o f the medical records of persons residing in this region from multiple health care providers for research studies, and provides comprehensive health information for this population. One of the health care providers in the REP (Mayo Clinic) rapidly made COVID-19 testing available early in the pandemic, and provided testing services to other health care providers in the region. For this reason, COVID-19 testing in this population was more common than in some other areas of the U.S., and tests did not need to be strictly reserved for patients who required emergency department or hospital care. Therefore, the REP provides a unique environment to study the epidemiology of COVID-19 in 27 counties of a defined, U.S. population. In this study, we report hospitalization and death rates due to COVID-19 infection, and we provide detailed information on the demographic and clinical characteristics associated with increased risk for hospitalization and death. Finally, we assess whether the impact of potential risk factors differs across three age groups.

The REP has been previously described. 17 

We searched the REP electronic indexes to identify all persons who resided in the 27county region and who had a positive nasopharyngeal polymerase chain reaction (PCR) test for COVID-19 between March 1, 2020 and September 30, 2020 (7 months).

Persons who did not provide authorization to use their medical records for research were excluded.

We extracted BMI and smoking status information (never, former, or current) closest to the date of the first positive test result. For persons 0-20 years of age, we categorized BMI into 4 categories using percentiles for height and weight compared to persons of the same age and gender using criteria defined by the Centers for Disease Control and Prevention (under/normal weight: <85 th percentile, overweight: ≥85 th percentile to <95 th percentile, obese:≥95 th percentile, and unknown). 19 For persons 21 years or older, BMI was categorized into 4 categories: under/normal weight (<25 kg/m 2 ), overweight (25-<30 kg/m 2 ), and obese (≥30 kg/m 2 ). Missing data are shown in Table 1 and described as "unknown". We then extracted all International Classification of Diseases diagnosis codes (ICD-9 and ICD-10) for the 5 years prior to the positive test result to capture patient comorbidities. However, we excluded all diagnosis codes that occurred within 14 days before the first positive test to exclude diagnoses that could have been due to COVID-19 infection. ICD diagnostic codes were then grouped into categories using the Clinical Classifications Software. 20 For this study, we focused on only the chronic J o u r n a l P r e -p r o o f disease categories, and further combined the categories into 48 broader disease groups, as previously defined by Mukherjee and colleagues. 21 These categories were used in a previous REP study. 22 The final Clinical Classification Codes (CCC), the 48 chronic condition groups, and the number of persons with a diagnosis in each group are shown in Supplemental Table 1 .

We defined severe COVID-19 infection as either hospitalization or death due to the infection. We used the REP resources to electronically extract all hospitalizations and deaths that occurred in the 3 months following the first positive test result or through 10/31/2020, whichever came first. The full text of the medical records was then reviewed by JLS and GSL to determine whether the hospitalization or death was due to the COVID-19 infection.

Follow-up began on the date of COVID-19 diagnosis (positive PCR test) and continued for 3 months, until death, or October 31, 2020, whichever came first. Cumulative incidence plots were constructed to visualize the probability of hospitalization and death conditions that were present in at least 50 persons in the study population. All models were adjusted for continuous age, sex, race, ethnicity, BMI, and smoking status. P values were adjusted for multiple comparisons using the false discovery rate, and an adjusted p value of <0.05 was considered statistically significant. 23 Stratified Cox proportional hazards regression models were used to estimate the risk for hospitalization or death among persons with different chronic condition categories by 3 age categories: 0-44 years, 45-64 years, and 65+ years. All models were adjusted for continuous age within each age group, sex, race, ethnicity, BMI group, and smoking status. Interactions between age group and each chronic condition were assessed by including an interaction term in overall models. Interaction p values<0.05 were considered statistically significant. All analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC) and R, version 3.2.3.

We identified 10,231 persons in the 27-county region with a positive COVID-19 PCR test result between March 1, 2020 and September 30, 2020; 9,928 (97%) had provided authorization to use their medical records for research, and were included in the study. Table 1 . After adjusting for all of the characteristics in Table 1 , older age, male sex, Hispanic ethnicity, Black, Asian, and Other/unknown race, obesity, and having 7 or more chronic conditions were associated with an increased risk of severe infection.

We examined the risk of severe COVID-19 infection (hospitalization or death) among persons with different types of chronic conditions considered one at a time (Table 2 ). In unadjusted analyses, nearly all of the comorbidities were strongly associated with an increased risk of severe infection. After adjustment for demographic characteristics, smoking, and BMI category, associations were attenuated. However, after adjusting for multiple comparisons, most of the chronic condition categories remained significantly associated with risk of severe infection at an adjusted statistical significance threshold of p<0.05 (Table 2 ). Developmental disorders, personality disorders, and affective disorders, schizophrenia, and other psychoses were most strongly associated with risk of severe infection (Table 2 ; all hazard ratios>3.0).

We next stratified our analyses by age, and assessed interactions between the chronic condition categories and age groups to determine if associations between the chronic condition categories and risk of severe COVID-19 infection differed by age. Because severe infections were uncommon in persons 0-19 years (0.7%), and in persons 20-44 years (1.5%), these 2 age groups were combined in a single group for the stratified analyses. Therefore, we considered 3 age groups: persons 0-44 years, 45-64 years, and 65+ years. Associations between 10 of the chronic disease categories and risk of severe COVID-19 infection differed significantly across the three age groups (interaction p values<0.05; Figure 3 ; Supplemental Table 2 ). The risk of severe COVID-19 for persons with these conditions compared to person without these conditions was highest in persons 0-44 years of age compared to the two older age groups. In particular, persons 0-44 years with cancer, chronic neurologic disorders, hematologic disorders, other endocrine disorders, and ischemic heart disease had a 3-fold increased risk of J o u r n a l P r e -p r o o f developing severe infection compared to persons of the same age without these conditions (Figure 3 ; Supplemental Table 2 ).

Finally, most of the persons in our cohort who had developmental disorders, epilepsy or convulsions, and personality disorders were between 0 and 44 years of age (Supplemental Table 2 

In a large, Midwestern, US population, we found that the risk of severe COVID-19 disease was associated with older age, male sex, Hispanic ethnicity, Asian, Black, and Other/Mixed race, obesity, and with an increasing number of chronic conditions. Most of the chronic disease categories we studied were significantly associated with an increased risk of severe COVID-19 disease even after adjusting for patient characteristics and accounting for multiple comparisons. We also found differences in associations between 10 chronic disease categories and severe COVID-19 infection across 3 age groups. Associations were strongest in persons 0-44 years of age, suggesting that younger COVID-19 patients with these conditions may be at especially high risk of severe disease compared to persons 0-44 years who do not have these conditions relative to the other age groups.

Our study results are in agreement with previous studies which have shown that older age is the primary risk factor for more severe COVID-19 disease. 2, 10, 15, 24, 25 In addition, J o u r n a l P r e -p r o o f our findings are consistent with previous studies that have showed that persons with multiple chronic conditions are at higher risk of severe infection. 8, 25-30 It is not clear why older age is such a strong risk factor for severe disease, but the number of chronic conditions increases with increasing age. However, in our study, older age remained significantly associated with severe infection even after adjustment for the number of comorbidities, suggesting that other factors related to older age (e.g. immunosenescence) may also play a role in disease severity. 31 We also found that minority race and ethnicity, obesity, and male sex were associated with an increased risk of severe COVID-19 infection, and these results are consistent with previous studies. 13, 14, 24, 25, 32-37 By contrast, we did not confirm findings from previous studies that have reported smoking as a risk factor for severe COVID-19 outcomes. 37, 38 Before adjustment, former smokers were at an increased risk of severe infection, but after adjustment, that association was significantly attenuated. The associations previously reported may have been caused by confounding by older age and by the presence of multiple chronic conditions. We also found that virtually all of the chronic condition categories we examined were significantly associated with increased risk of severe COVID-19 infection, even after adjusting for patient characteristics and multiple comparisons. Our comprehensive assessment of comorbidities demonstrates that an increased risk of severe infection extends well beyond those already established with cardiovascular diseases, their risk factors, chronic lung disease, and cancer. We found, however, that the risk of severe infection in persons with 10 of these conditions differed significantly by age. Overall, J o u r n a l P r e -p r o o f persons 0-44 years of age were at the lowest risk of severe infection. However, severe cases occurred in 90 persons (1.5%) in this age group. Persons 0-44 years of age who also had cancer, chronic neurologic disorders, hematologic disorders, ischemic heart disease, and other endocrine disorders had greater than 3-fold increased risks of severe COVID-19 infection compared to persons of the same age, but who did not have these conditions. These risks were significantly different across the three age groups we studied, suggesting that such conditions are especially concerning in the younger population. Our results were particularly striking for cancer. Cancer was a strong risk factor for severe COVID-19 disease, but only in this age group (HR: 3.21; 95% CI: 1.66, 6.21). Cancer was not a significant risk factor for severe COVID-19 disease in persons 45-64 or 65+ years.

Finally, we found that some conditions were over-represented in younger persons in our cohort (0-44 years), including developmental disorders, epilepsy and convulsions, and personality disorders. There were not enough persons with these conditions to study the association between the conditions and severe COVID-19 infection in the age groups 45-64 and 65+ years. However, in analyses restricted to persons 0-44 years of age, each of these chronic condition categories was strongly associated with risk of severe COVID-19 infection.

Strengths of our study include access to historically collected data on all chronic conditions in a defined Midwestern population, across age, sex, and racial and ethnic groups. We note that the REP captures approximately 61% of the population residing in J o u r n a l P r e -p r o o f this region, and that the characteristics of persons in this region are similar to those of persons residing in the upper Midwest. 17 In addition, testing for COVID-19 has been readily available to this population since early in the pandemic for persons with at least one symptom of COVID-19, exposure to a known positive case, or for screening prior to a medical or surgical procedure. These less stringent testing criteria make it more likely that less severe cases were captured. However, we expect to still miss persons who were asymptomatic, sought testing outside of the REP health care partners, or did not seek testing for other reasons. Therefore, hospitalization and death rates may still represent an overestimate of the risk of severe infection in the entire population.

An additional limitation of our study is the inability to verify the ICD codes for this study.

ICD codes may be assigned in error (overdiagnosis), and manual review of the medical records is often needed to determine whether an individual truly has the disease or condition of interest. We also may have missed people who should have been assigned a code of interest, but were not (underdiagnosis). However, because we limited our diagnosis categories to those that are considered chronic conditions, most patients would be seen at least once within the 5-year period we considered for this study, making it unlikely that we missed conditions that require medical treatment. This approach may have caused us to include persons in the chronic disease categories who have recovered from the disease, or who have a well-controlled chronic condition. Table 3 ). Finally, there were 1,391 persons with less than 6 months of information between the date of first visit in the REP system and their COVID-19 diagnosis. We may underestimate chronic conditions in this population, and inclusion of these persons in our unexposed group may have biased our study results. We conducted a sensitivity analysis excluding these persons, and found that results were again largely unchanged (Supplemental Table 3 ).

The chronic condition categories that we examined were broad, and some categories included disparate conditions (e.g. "Cancer" included all cancers). In addition, the conditions we considered are not all equivalent in severity, and using a simple count of conditions to identify persons with a high disease burden may overestimate disease severity in persons with a high number of less serious conditions. However, our approach is useful as a way to examine a wide range of conditions, and to screen for groups of conditions that may increase the risk of severe infection. However, this approach will miss conditions that are important risk factors for severe infection if there were not enough persons with those conditions within the categories that we studied.

Similarly, there were too few severe cases in persons 0-19 years to study this population as a separate group, and a larger sample size is needed to identify all risk factors in this age group.

Finally, PCR tests for COVID-19 sometimes produce inaccurate results, and we may have missed some infections (false negatives), and we may have incorrectly identified some persons as infected who were not truly infected (false positives). Poor sensitivity will result in an underestimation of all cases, and poor specificity will result in an J o u r n a l P r e -p r o o f overestimation of all cases. If demographic and clinical characteristics routinely affect sensitivity and specificity our results may be biased. A range of PCR tests has been used in this community, and accurate sensitivity and specificity data are not currently available for all tests. Therefore, it is difficult to determine in which direction testing variability may have biased our study results.

In a large, Midwestern, U.S. population, we found that older age, male sex, minority ethnicity or race, obesity, and a wide range of chronic conditions were significantly associated with the risk of severe COVID-19 infection. We also found that the risk of severe COVID-19 infection associated with many of the chronic conditions differed by age group, and younger persons in our cohort with these conditions were at especially high risk for severe infection compared to persons of the same age. Older persons are much more likely to experience severe infections; however, severe cases do occur in younger persons as well. Our data provide insight regarding younger persons at especially high risk of severe COVID-19 infection. 

Timeline of WHO's response to COVID-19

Clinical Characteristics and Outcomes of Older Patients with Coronavirus Disease 2019 (COVID-19) in Wuhan, China: A Single-Centered, Retrospective Study

Diabetes is a risk factor for the progression and prognosis of COVID-19

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Risk factors for severity and mortality in adult COVID-19 inpatients in Wuhan

Risk Factors Associated with Clinical Outcomes in 323

COVID-19 Hospitalized Patients in Wuhan, China. Clin Infect Dis

A clinical risk score to identify patients with COVID-19 at high risk of critical care admission or death: An observational cohort study

Coronavirus disease 2019 (COVID-19) and cardiovascular risk: A meta-analysis

Risk Factors of Fatal Outcome in Hospitalized Subjects With Coronavirus Disease 2019 From a Nationwide Analysis in China

Severe Obesity as an Independent Risk Factor for COVID-19 Mortality in Hospitalized Patients Younger than 50

Years Is a Risk Factor for COVID-19 Hospital Admission

Factors associated with hospital admission and critical illness among 5279 people with coronavirus disease 2019 in New York City: prospective cohort study

Defining the Epidemiology of Covid-19 -Studies Needed

Characteristics of Adult Outpatients and Inpatients with COVID-19 -11 Academic Medical Centers

Multimorbidity and SARS-CoV-2 infection in UK Biobank

Beyond chronological age: Frailty and multimorbidity predict in-hospital mortality in patients with coronavirus disease 2019

Cardiometabolic multimorbidity is associated with a worse Covid-19 prognosis than individual cardiometabolic risk factors: a multicentre retrospective study (CoViDiab II)

Age and Multimorbidity Predict Death Among COVID-19 Patients: Results of the SARS-RAS Study of the Italian Society of Hypertension

Association of Comorbidities With Pneumonia and Death Among

Race/Ethnicity, Underlying Medical Conditions, Homelessness, and Hospitalization Status of Adult Patients with COVID-19 at an Urban Safety-Net Medical Center

COVIDView: A Weekly Surveillance Summary of U.S. COVID-19 Activity (Week 34)

Overweight, obesity, and risk of hospitalization for COVID-19: A community-based cohort study of adults in the United Kingdom

Lifestyle risk factors, inflammatory mechanisms, and COVID-19 hospitalization: A community-based cohort study of 387,109 adults in UK

Obesity predisposes to the risk of higher mortality in young COVID-19 patients