Modulation of the Association Between Age and Death by Risk Factor Burden in Critically Ill Patients With COVID-19

Abstract

Older age is a key risk factor for adverse outcomes in critically ill patients with COVID-19. However, few studies have investigated whether preexisting comorbidities and acute physiologic ICU factors modify the association between age and death.

Design: Multicenter cohort study.

Setting: ICUs at 68 hospitals across the United States.

Patients: A total of 5,037 critically ill adults with COVID-19 admitted to ICUs between March 1, 2020, and July 1, 2020.

Interventions: None.

Measurements and main results: The primary exposure was age, modeled as a continuous variable. The primary outcome was 28-day inhospital mortality. Multivariable logistic regression tested the association between age and death. Effect modification by the number of risk factors was assessed through a multiplicative interaction term in the logistic regression model. Among the 5,037 patients included (mean age, 60.9 yr [± 14.7], 3,179 [63.1%] male), 1,786 (35.4%) died within 28 days. Age had a nonlinear association with 28-day mortality (p for nonlinearity <0.001) after adjustment for covariates that included demographics, preexisting comorbidities, acute physiologic ICU factors, number of ICU beds, and treatments for COVID-19. The number of preexisting comorbidities and acute physiologic ICU factors modified the association between age and 28-day mortality (p for interaction <0.001), but this effect modification was modest as age still had an exponential relationship with death in subgroups stratified by the number of risk factors.

Conclusions: In a large population of critically ill patients with COVID-19, age had an independent exponential association with death. The number of preexisting comorbidities and acute physiologic ICU factors modified the association between age and death, but age still had an exponential association with death in subgroups according to the number of risk factors present. Additional studies are needed to identify the mechanisms underpinning why older age confers an increased risk of death in critically ill patients with COVID-19.

Keywords: COVID-19; age; critical care; death; risk factors.

Figure 1.

Nonlinear association between age and 28-day mortality. A, The model is unadjusted. Age-linear akaike information criterion (AIC): 30,713 versus age-spline AIC: 30,686, likelihood ratio test p value for AIC difference: < 0.001. B, This model is further adjusted for demographic characteristics, including male sex, and presence of hypertension, diabetes mellitus, coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease, and active cancer. Age-linear AIC: 30,218 versus Age-spline AIC: 30,160, likelihood ratio test p value for AIC difference: < 0.001. C, This model is further adjusted for acute ICU physiologic factors, including symptom onset less than or equal to 3 d prior to ICU admission, lymphocyte count less than 1,000/uL, degree of hypoxemia and respiratory support, shock, SOFA coagulation greater than 0, SOFA liver greater than 0, and SOFA renal greater than 0, and the number of ICU beds. Age-linear AIC 26,913 versus Age-spline AIC: 26,783, likelihood ratio test p value for AIC difference: < 0.001. D, This model is further adjusted for COVID-19 treatments, including: remdesivir, tocilizumab, and corticosteroids. Age-linear AIC: 26,778, Age-spline AIC 26,643, likelihood ratio test p value for AIC difference: < 0.001. SOFA = Sequential Organ Failure Assessment.

Figure 2.

Nonlinear association between age and 28-day mortality stratified by number of risk factors. Nonlinear association of age (continuous) with 28-day mortality stratified by categories of the number of significant preexisting comorbidities and acute physiologic ICU factors. p for nonlinearity in each group less than 0.001. Risk factors: male sex, hypertension, diabetes mellitus, coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease, active cancer, symptom onset less than or equal to 3 d, lymphocyte count less than 1,000/µL, invasive mechanical ventilation, shock, SOFA Coagulation Score greater than 0, SOFA Renal Score greater than 0, SOFA Liver Score greater than 0, and number of ICU beds less than 100. Models are further adjusted for body mass index (in categories), White versus non-White, current smoker, remdesivir, tocilizumab, and corticosteroids. SOFA = Sequential Organ Failure Assessment.