Risk factors for hospitalization and death due to COVID-19 among frail community-dwelling elderly people: a retrospective cohort study

ABSTRACT BACKGROUND: Advanced age, multiple chronic diseases and frailty have been correlated with worse prognosis among coronavirus disease 2019 (COVID-19) inpatients. OBJECTIVE: To investigate potential risk factors for hospitalization and death due to COVID-19 among frail community-dwelling elderly people. DESIGN AND SETTING: Retrospective cohort study of patients followed up at a geriatric outpatient clinic in Belo Horizonte, Minas Gerais, Brazil. METHODS: The associations of demographic characteristics (age and sex) and clinical characteristics (frailty, multimorbidity, number of medications with long-term use, obesity, smoking, diabetes mellitus, pulmonary diseases, cardiovascular diseases, cerebrovascular disease, and chronic kidney disease) with the risk of hospitalization and death due to COVID-19 were explored using a multivariable logistic regression model. RESULTS: 5,295 patients (mean age 78.6 ± 9.4 years; 72.6% females) were included. After adjustments, the number of medications with long-term use was found to increase the odds of hospitalization due to COVID-19 (odds ratio, OR: 1.13; 95% confidence interval, CI: 1.06-1.22). Frailty, multimorbidity and diabetes mellitus also increased the odds of hospitalization (OR: 1.06, 95% CI: 1.02-1.09; OR: 1.17, 95% CI: 1.09-1.26; and OR: 2.27, 95% CI: 1.45-3.54, respectively) and the odds of death due to COVID-19 (OR: 1.07, 95% CI: 1.00-1.14; OR: 1.16, 95% CI: 1.03-1.32; and OR: 2.69, 95% CI: 1.79-6.14, respectively). CONCLUSIONS: Multimorbidity, frailty and diabetes mellitus increased the odds of hospitalization and death due to COVID-19 and the number of medications with long-term use increased the odds of hospitalization due to COVID-19 among frail community-dwelling elderly people.


INTRODUCTION
The coronavirus disease 2019 (COVID-19) pandemic remains a major global public health problem. The World Health Organization (WHO) registered over 153 million cases and three million deaths up to the beginning of May 2021. 1 In Brazil, there have been more than 500,000 deaths due to this disease and, recently, there has been a sharp increase in the number of cases (a "second wave"), which has put a lot of pressure on the healthcare system, with occupation rates of more than 90% in intensive care units in many Brazilian states. 2 Advanced age and several diseases, such as diabetes mellitus, chronic kidney disease, cardiovascular disease and chronic respiratory disease have been correlated with worse outcomes, such as hospitalization, need for invasive ventilation and mortality due to COVID-19. [3][4][5][6] Furthermore, in addition to older age and presence of chronic diseases, it is also important to consider frailty and multimorbidity, particularly among the elderly. 4 Frailty is characterized by decreased strength, resistance and physiological response, which translates into faulty reestablishment of homeostasis after a stressing event, thus leading to a high risk of incapacity. 7 Studies on patients older than 60 years of age have demonstrated that frailty is associated with a higher risk of death 4 and with greater disease severity among patients hospitalized due to COVID-19. 8 Similarly, multimorbidity, which consists of co-occurrence of multiple diseases or clinical conditions in one person, 9 is a factor to be considered among older adults with COVID- 19

Design and sample
A retrospective cohort study was conducted using data from an electronic In the present study, a diagnosis of COVID-19 registered in the medical records was considered to be an incident case of the disease.
Similarly, hospitalization and death due to COVID-19 were considered from the registrations of these events in the medical records.

Assessment of demographic and clinical characteristics
Sex, age and clinical characteristics were evaluated regarding the risks of hospitalization and death due to COVID-19. Among the clinical characteristics, frailty, multimorbidity, long-term use of medications, obesity, smoking, diabetes mellitus (DM), respiratory diseases, cardiovascular diseases, cerebrovascular disease and chronic kidney disease (CKD) were considered. 5,14 Frailty was evaluated using the Clinical-Functional Vulnerability  15 Each section has a specific score and these add up to a maximum of 40 points. The higher the resulting value is, the higher the clinical-functional vulnerability risk of the older adult also is.
Total scores higher than 7 characterize older adults as frail. 15 This instrument has been found to be a valid and reliable measurement for assessing frailty among community-dwelling patients ≥ 60 years of age. 15 Multimorbidity was measured through the number of medical diagnoses. The diagnosis count has been commonly used to assess multimorbidity within primary healthcare. 16 The number of medications with long-term use was evaluated as listed in the medical record. Weight and height were taken from the mobility section of IVCF-20. 15 In this section, these measurements are recorded and the body mass index (BMI, in kg/m 2 ) is calculated. Patients were considered to be obese if they had BMI ≥ 30. 17 Smoking, DM and CKD were evaluated from their diagnoses in the medical records. Respiratory diseases were assessed through diagnoses of asthma and/or chronic obstructive lung disease.
Cardiovascular disease was assessed through any of the following diagnoses: cardiomyopathy, coronary disease with or without previous acute myocardial infarction, heart failure with preserved ejection fraction and congestive heart failure. Cerebrovascular disease was assessed through a diagnosis of previous stroke.

Statistical analysis
Descriptive analyses were conducted using frequencies and percentages (%) for categorical variables and using means and stan- The results were presented as odds ratios (OR) and 95% confidence intervals (CI). All analyses were performed using STATA (version 14.1, StataCorp LP, College Station, Texas, United States).

RESULTS
A total of 5,295 patients were included, of mean age 78.6 ± 9.4 years, and most of them were women (72.6%). The mean IVCF-20 score was 16.8 ± 6.9), the mean number of diagnosis was 5.4 ± 3.0) and the mean of number of medications with long-term use was 5.2 ± 3.1).
Out of these, 82 (24.2%) needed to be hospitalized and 23 died due to complications from the infection (case-fatality rate of 6.8%).
The monthly numbers of diagnoses and deaths due to COVID-19 during the 12-month period are shown in Figure 1.
The demographic and clinical characteristics of the sample are described in Table 1. The patients hospitalized due to COVID-19 and those who died from COVID-19 were older and frailer, with more morbidities and more long-term use of medications.
Furthermore, DM and cardiovascular diseases also were more prevalent in this group.

DISCUSSION
In this retrospective cohort study among frail community-dwelling elderly people, older age and frailty increased the odds of hospitalization and death due to COVID-19. Our results also showed that, independently of the effects of age and frailty, multimorbidity and DM increased the odds of hospitalization and death. Furthermore, the number of medications with long-term use increased the odds of hospitalization due to COVID-19.
Advanced age has already been consolidated as a risk factor for death due to COVID-19, 18 with a case-fatality rate of 19.3% among individuals over 65 years old, 19 and a risk of death three times greater for those between 60 and 69 years old, compared    years, also showed that higher multimorbidity was associated with higher risk of death. 25 While a growing number of studies has supported the notion that frailty and morbidity, including diabetes mellitus, are risk factors for adverse prognosis and death among patients hospitalized due to COVID-19, our study is one of the few to explore these risk factors among elderly people who were followed up on an outpatient basis. Therefore, our study provides an important contribution to investigation of prognostic factors relating to COVID-19.
Moreover, it was possible to demonstrate that multimorbidity contributed to the COVID-19 prognosis regardless of frailty and age.
Nonetheless, the present study had some limitations relating to its observational design. Despite adjustments for potential confounding variables, the presence of residual confounders cannot be ruled out. Our use of a retrospective source of data might have compromised the quality of information, but we minimized this problem through the specificity of the electronic medical records used in our clinic. In the same way, even though the data used came from required fields in the electronic medical records, we cannot rule out the possibility that some information may have been lost due to inadequate and incomplete filling out of the electronic medical records.

CONCLUSIONS
Our results help to identify independent risk factors for hospitalization and death due to COVID-19 among frail communitydwelling elderly people. This is of fundamental importance for planning healthcare actions in this population, which remains vulnerable to complications from SARS-CoV-2 infection.