Introduction
Older age is the strongest risk factor for infection with the severe acute respiratory virus syndrome coronavirus-2 (SARS-CoV-2) and older adults have been especially hard-hit during the pandemic [
1,
2]. Older adults with SARS-CoV-2 are at increased risk for hospitalization [
3‐
5], critical illness [
4,
6‐
13], prolonged hospitalization [
11,
14] and mortality [
8,
9,
14‐
29] compared with their younger counterparts. Because a disproportionate number of hospitalizations and deaths in patients with coronavirus disease 2019 (COVID-19) occur among older adults [
30‐
32], the burden of SARS-CoV-2 has been greatest in countries [
33‐
36] and communities [
37] with older populations, and in healthcare facilities [
38‐
40] and health systems serving older adults [
41].
However, the prevalence of most comorbid conditions increases with age, which can make it challenging to disentangle the separate effects of age and comorbid burden on outcomes among patients infected with SARS-CoV-2. Prior studies have reported an association between Charlson comorbidity score and other measures of comorbid burden with adverse outcomes in SARS-CoV-2 after adjustment for age [
41‐
49] and differences across age groups in clinical presentation and outcomes of SARS-CoV-2 [
11,
14,
27,
31,
50‐
53]. There are also reports of age differences in the association of individual comorbid conditions [
54‐
56] with adverse outcomes in infected patients.
While few studies have examined the relationship between age, comorbid burden and adverse outcomes in SARS-CoV-2, available data suggest that the relationship between these factors may be complex. Among the first 11,122 infected patients in Denmark, mortality rates were noted to be extremely low in those < 80 years old without comorbid conditions but uniformly high in those ≥80 years old regardless of the number of comorbid conditions [
57]. Because both age and comorbidity have figured prominently in guidelines for clinical care, vaccination and social distancing during the pandemic, a more detailed understanding of the relationship of age and comorbid burden with a range of adverse outcomes among adults infected with SARS-CoV-2 could inform clinical care, current and future vaccine prioritization and prognostication [
58,
59] and efforts to estimate disease burden and service needs related to SARS-CoV-2 [
33].
The US Department of Veterans Affairs (VA) supports the largest integrated national health care system in the US and provides care for more than six million veterans annually [
60]. Because the VA serves both younger and older adults and many veterans have multiple comorbid conditions [
61], the system may offer unique insights into the relationship between age, comorbid burden and adverse outcomes in those infected with SARS-CoV-2. We designed a study to evaluate differences across age groups in the association of comorbid burden with a range of adverse outcomes among patients infected with SARS-CoV-2.
Methods
Data source and study population
The VA uses a single comprehensive electronic healthcare information network. Data elements from patients’ electronic medical records at individual VA facilities are stored centrally in the VA’s Corporate Data Warehouse (CDW) which is maintained by the VA Informatics and Computing Infrastructure (VINCI). To support research on the health system impacts of the SARS-CoV-2 pandemic, VA maintains the VA National Surveillance Tool, which includes updated clinical and administrative data extracts for all patients tested for SARS-CoV-2 within the VA, with adjudication of all positive test results [
60]. Using this resource, we assembled a cohort of all Veterans who tested positive for SARS-CoV-2 nucleic acid by polymerase chain reaction (PCR) at least once between February 28, 2020 and December 31, 2020 with complete information on test date (
n = 170,528). The index date for the present study was defined as the date of each patient’s earliest positive test result unless this occurred within the first 15 days of a hospital admission, in which case we used the date of hospital admission as the index date. Follow-up for all study outcomes was available through January 31, 2021, allowing for analysis of all outcomes occurring within 30 days after the index date for all cohort members. This study was approved by the Institutional Review Board of the Veterans Affairs Puget Sound Healthcare System which granted a waiver of informed consent as the study was deemed minimal risk because it involved secondary analyses of existing data and could not otherwise have been conducted due to the large size of the cohort and inclusion deceased patients. Our research was conducted in accordance with the Declaration of Helsinki.
Exposure
The primary exposure was each patient’s Charlson Comorbidity Index (CCI) score (categorized by approximate quartile of the distribution within our cohort as 0, 1, 2–3 or ≥ 4) based on International Classification of Diseases 10th Revision (ICD-10) codes recorded in VA administrative data on or within 2 years before the index data [
62].
Outcomes
The following outcomes were ascertained for the 30-day period following the index data using the VA National Surveillance Tool [
60]: 1) hospitalization; 2) ICU admission; 3) mechanical ventilation; and 4) death.
Covariates
All analyses were stratified by patients’ age group on the index date (categorized as 18–64, 65–79 and ≥ 80 years) and adjusted for sex, race (Black, White, Other), Hispanic ethnicity, body mass index body (BMI) (categorized as underweight (< 18.5 kg/m2), normal weight (18.5–24.9 kg/m2) overweight (25–29.9 kg/m2), stage 1 obesity (30–34.9 kg/m2) and stages 2 & 3 obesity (≥35 kg/m2)) and U.S. Federal Region. To account for potential age differences in outcomes within each age group, multivariate analyses were also adjusted for age as a continuous variable.
Statistical analysis
We used Pearson’s χ2 test to compare patient characteristics across age groups. We plotted 30-day cumulative rates of mortality, hospitalization, ICU admission, and mechanical ventilation by age using a lowess smoother with a bandwidth of 80%. We used a time-to-event analysis accounting for censoring at the time of death to calculate the crude cumulative 30-day incidence of hospitalization, ICU admission, mechanical ventilation, and death within each age group.
We used Cox proportional hazards models stratified by age group to measure the adjusted association of CCI with hospitalization, ICU admission, mechanical ventilation, and death within the first 30 days after the index date after adjustment for age (as a continuous variable), sex, race, Hispanic ethnicity, BMI, and region of residence. We used a two-sided p-value threshold of < 0.05 to assess statistical significance. We confirmed there were no violations of the proportional hazard assumption via visual inspection of log-log graphs. All analyses were stratified by VA medical center. A likelihood ratio test (or χ2 test of predicted hazard ratios) was used to evaluate for interaction between age group and CCI in adjusted analyses.
We conducted a supplementary analysis in which we used competing risk models to estimate the 30-day cumulative incidence of hospitalization, ICU admission and mechanical ventilation and to measure the adjusted association of CCI with each of these non-death outcomes [
63]. We also repeated the primary analyses using approximate quartiles of the Elixhauser Comorbidity Index (ECI).
All analyses were conducted using Stata/MP Version 15 (StataCorp, LLC. College Station, TX).
Discussion
Among VA patients who tested positive for SARS-CoV-2 during the first 10 months of the pandemic, there were marked differences across age groups in the prevalence and prognostic significance of comorbid burden. Associations of comorbid burden with adverse outcomes were generally attenuated at older ages leading to systematic differences across age groups in the threshold level of the CCI (and ECI) associated with an increased risk for adverse outcomes. In cohort members < 65 years and 65–79 years, risk for most outcomes was increased for patients with a CCI in the second or third quartile (CCI of 1 or 2–3, respectively) or higher. On the other hand, among those aged ≥80 years, only a CCI in the 4th quartile (CCI ≥4) was associated with an increased risk for hospitalization, ICU admission and death and risk of mechanical ventilation did not vary greatly with CCI.
The disproportionate number of deaths occurring among older adults is a distinguishing feature of the SARS-CoV-2 pandemic as compared with prior influenza pandemics [
32]. The very steep age gradient in risks of death and hospitalization and strong association of comorbid burden with adverse outcomes among younger and middle aged members of our cohort generally supports the US Centers for Disease Control-recommended approach to vaccine prioritization based on age and the presence of comorbid conditions among younger and middle-aged (i.e. < 65 years) adults [
64]. However, as we and others have argued elsewhere, there may be opportunities to refine ongoing vaccine prioritization strategies by accounting for the presence of sizeable differences within age groups in risk for adverse outcomes [
65‐
67].
Consistent with the crude mortality rates reported by Reilev et al. for the first 11,122 infected patients in Denmark, risk of death was extremely high for the oldest members of our cohort regardless of CCI [
57]. However, there were substantial differences in the relative risk for all adverse outcomes across quartiles of comorbid burden in younger age groups, including those aged 65–79 years. Further, rates of hospitalization, ICU admission and mechanical ventilation for patients with a CCI in the fourth quartile were extremely high regardless of age. Prioritization for vaccination and risk stratification based on comorbid burden may be especially impactful for those aged 65–79 years given that their absolute mortality and hospitalization rates substantially exceed those aged < 65 years. Collectively, these findings suggest there may be opportunities to refine ongoing vaccine allocation strategies to improve risk stratification within this relatively large high-risk group.
While prior studies have described strong associations of CCI and other measures of comorbid burden with adverse outcomes in patients infected with SARS-CoV-2 after adjusting for age [
41‐
48], none to our knowledge have evaluated the adjusted association of comorbid burden with adverse outcomes after stratification by age group. However, our major finding of systematic differences across age groups in the strength and magnitude of the association of comorbid burden with hospitalization, death and ICU admission among patients with SARS-CoV-2 seems consistent with other work describing age differences in the association of individual comorbid conditions [
54‐
56] with adverse outcomes among infected patients. An attenuation of the association of comorbid burden with adverse outcomes in patients with SARS-CoV-2 at older ages also resonates with earlier work demonstrating that in older adults, the presence or absence of particular comorbid conditions and/or abnormalities in specific disease markers often have less prognostic significance than more global measures of functional status and frailty that are not necessarily tied to the presence or severity of specific underlying disease processes [
68‐
70]. Our findings may also suggest the need for alternative approaches to risk stratification within the oldest age group, perhaps based on non-disease-specific measures such as frailty and functional status (71) or different threshold levels of comorbid burden.
Consistent with earlier studies conducted in a range of different populations infected with SARS-CoV-2 [
8,
9,
14‐
29,
71], the age gradient for crude mortality was extremely steep and mortality rates increased exponentially while slopes for non-death outcomes plateaued at older ages. Age-related increases in rates of ICU admission and mechanical ventilation among members of our cohort were much less steep than those for mortality and hospitalization. Of particular note, crude rates of mechanical ventilation peaked among patients in their late 70’s and declined at older ages and the association of CCI with mechanical ventilation was substantially attenuated among those aged 65–79 years and extinguished in the oldest age group. These findings support the possibility that clinical management of hospitalized patients infected with SARS-CoV-2 might vary depending on the patient’s age and comorbid burden [
72]. Collectively, our work argues for studies to understand the impact of age and comorbidity on real-world care practices for patients infected with SARS-CoV-2.
Limitations of our study include first, that our results may not be generalizable to other populations, particularly non-veteran populations and those that include a higher percentage of women. Second, ascertainment of comorbidity was limited to information included in ICD-10 codes which may not fully capture the severity of individual comorbid conditions or comorbid disease burden. Third, we only had information on hospitalizations occurring within the VA or under VA Community Care, thus our results may underestimate rates of non-death outcomes among cohort members. Fourth, due to selective testing for SARS-CoV-2 within our health system, our results may not be generalizable to infected Veterans not tested within the VA and may be subject to bias based on age and comorbidity. Finally, our analyses are adjusted for a limited number of measured patient characteristics and may reflect confounding by unmeasured characteristics.
In conclusion, while prior studies have demonstrated strong associations between older age and comorbid burden with the severe manifestations of COVID-19, none have rigorously evaluated for age differences in the strength and magnitude of the association of comorbid burden with a range of adverse outcomes. Our finding of systematic differences across age groups in the distribution and prognostic significance of comorbid burden among patients infected with SARS-CoV-2 could help to inform clinical care, ongoing vaccine prioritization and prognostication. Age differences in outcomes for hospitalized patients with SARS-CoV-2 also argue for more work to understand the role of age and comorbid burden in shaping clinical care and decision-making.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.