Skip to main content

COVID-19 and dementia: Analysis of risk, disparity, and outcomes from electronic health records in the US

Our take —

This retrospective case-control study used electronic health records from 61.9 million adults in the United States from January to August 21, 2020. Patients with dementia had twice the odds of COVID-19 compared to patients without dementia, after adjusting for several risk factors. Patients with both dementia and COVID-19 had greater odds of hospitalization and death than patients with either disease individually. The odds of COVID-19 were highest among patients with vascular dementia (adjusted OR = 3.17), followed by pre-senile (AOR=2.62), senile (AOR=1.99), Alzheimer’s disease (AOR=1.86), and posttraumatic (AOR=1.67) dementia. The study also examined racial disparities and found that Black (vs. white) patients with dementia were both more likely to develop COVID-19 and to be hospitalized. This study was limited by a small number of subjects in certain dementia subtypes (e.g., Lewy body and frontotemporal dementia), excluding patients with missing data for certain analyses, potential residual confounding, and lack of adjustment for COVID-19 risk factors when examining the mortality and hospitalization outcomes.

Study design

Case Control

Study population and setting

The study used data from an electronic health record (EHR) database, which was collected from 360 hospitals across 50 states in the United States and comprised 20% of the United States population. On Aug 21, 2020, 61,916,260 adult subjects were identified through the database and formed the study population. The EHR database deidentified the data and tabulated it into different categories at the population level. Therefore, for the current study, the adjusted odds ratio (aOR) was calculated between subjects in different categories. A dementia diagnosis was categorized into Alzheimer dementia, vascular dementia, senile dementia, pre-senile dementia, and post-traumatic dementia, with separate analyses conducted for each. Lewy body, frontotemporal, and mixed type dementia were excluded due to the small number of subjects. The outcome variables were positive COVID-19 infection, hospitalization and death (within 6 months post-hospitalization). The following factors known to be associated with COVID-19 were included as potential confounders: asthma, cardiovascular disease, chronic obstructive pulmonary disease, cancer, type II diabetes, obesity, chronic kidney disease, substance use disorders (including alcohol use disorder, cannabis use disorder and cocaine use), history of transplant (bone marrow or solid organ transplant), and nursing home stay status. Also, subjects were categorized according to their age (18-65 years and more than 65 years), race, and sex. The study compared the odds of COVID-19 infection, hospitalization, or death between subjects with and without dementia and across different dementia categories, with and without adjustment for COVID-19 risk factors. Odds ratios adjusted for potential confounders were calculated using the Cochran-Mantel-Hanzel method. A Benferroni correction was used to account for the multiple comparisons.

Summary of Main Findings

In total, 1,064,960 (1.75% of the study population) had: dementia, 351,590 (0.57%); Alzheimer dementia, 126,450 (0.2%); post-traumatic dementia, 31,960 (0.05%); pre-senile dementia, 17,2630 (0.28%); senile dementia, and vascular dementia, 117,860 (0.19%). 15,770 (0.025%) subjects had SARS-COV-2 infection and 810 subjects had dementia and SARS-COV-2 infection. Alzheimer dementia was the most common type of dementia among subjects with COVID-19, followed by presenile dementia, vascular dementia, senile dementia and post-traumatic dementia. After adjusting for COVID-19 risk factors, patients with dementia were more likely to develop COVID-19 compared to non-dementia patients (adjusted OR 2, 95 CI:1.94-2.06). After stratifying for the dementia types, patients with vascular dementia were more likely to develop COVID-19 (OR 3.17, 95CI: 2.97-3.37), followed by presenile dementia, Alzheimer dementia, and senile dementia with (aOR 2.26,95 CI: 2.28-3.0), (aOR 1.86, 95 CI:1.77-1.96) and (aOR1.99, 95 CI:1.86-2.13) respectively. Patients with post-traumatic dementia had the lowest odds for COVID-19 infection compared to other dementia types with (aOR 1.67,95CI:1.51-1.86). When the authors repeated the analysis without adjusting for COVID-19 risk factors, the strength of association between all dementia types and the COVID-19 infection increased. This suggests that patients with all types of dementia have many risk factors that predispose them to COVID-19 infection. However, absence of lack of association after adjusting for the COVID-19 risk factors, suggests that dementia is independently associated with risk of COVID-19 infection. African American people with dementia (Alzheimer dementia and vascular dementia only, due to sample size limitations) were more likely to have COVID-19 infection compared to white people with dementia after adjusting for age, sex and COVID-19 risk factors (aOR 2.86,95 CI: 2.67-3.06). This relationship did not change after excluding COVID-19 risk factors, suggesting factors other than the COVID-19 risk factors contributed to the observed disparity. Among subjects with vascular dementia, senior patients (>65 years) were more likely to develop COVID-19 compared to younger subjects. Of those with COVID-19, 25.17% (n=3969) were hospitalized. African Americans with COVID-19 and dementia (73.08% VS 53.85%), Alzheimer dementia (75.00% VS 58.82%), and without dementia (31.99% VS 18.57%) were more likely to be hospitalized than white participants. The 6 month hospitalization risk was statistically significantly higher among African American people compared to white people in the group of adults and senior with COVID-19 but without dementia (31.99% VS 18.57%), among adults and seniors with dementia but without COVID-19 (16.92% VS 12.24%) and among adults and seniors with Alzheimer dementia but without COVID-19 (19.9% VS 13.37%). Patients with COVID-19 and dementia were more likely to be hospitalized or die compared to the additive effect of each group, suggesting a synergistic effect of dementia on COVID-19 hospitalization and death. African American people with dementia but without COVID-19, COVID-19 without dementia and those with Alzheimer dementia but without COVID-19 were more likely to die at 6 months compared to white people. No disparities were found among patients with dementia or Alzheimer dementia and COVID-19.

Study Strengths

The study used a large nationwide database of electronic health records in the United States comprising 20% of its population, which decreases the chance of selection bias. The authors adjusted the analysis for multiple COVID-19 risk factors in their analysis for COVID-19 infection. The study stratified the analysis for several racial groups and for multiple outcomes which strengthen the study conclusion.

Limitations

The authors mentioned that missing data pertaining to SARS-CoV-2 infection and dementia were excluded from the analysis. This could bias the results if participants missing data had different relationships between dementia, COVID-19, and hospitalization or death. Also, because electronic health records are not designed specifically for research, data may have been misclassified, which could have biased the findings. The authors did not indicate how they categorized the COVID-19 risk factors; findings may be distorted if these were not adequately handled in the analysis. The mortality and hospitalization outcomes were measured without adjustment for the COVID-19 risk factors due to the small number of patients in these cohorts, which may limit study inferences. The study lacked data pertaining to socioeconomic and lifestyle factors, which highly correlate with the risk of SARS-CoV-2 infection.

Value added

This is the largest study that looks at the association between different types of dementia and the risk for COVID-19 infection, hospitalization and death, and the first study that looked at the disparity related to COVID-19 infection among dementia patients.

This review was posted on: 26 February 2021