Vets Study Casts Doubt on Accepted COVID-19 Risk Factors

Among more than 10,000 U.S. veterans testing positive for SARS-CoV-2, older age, high regional COVID-19 disease burden, Charlson comorbidity index (CCI) score, and abnormalities in certain blood tests were significantly associated with mortality after adjustment.

However, many previously reported risk factors, including obesity, Black race, Hispanic ethnicity, chronic obstructive pulmonary disease (COPD), hypertension, and smoking, were not significantly associated with mortality, George Ioannou, BMBCh, of Veterans Affairs Puget Sound Health System in Seattle, and colleagues reported in JAMA Network Open.

“Other risk factors for mortality included select preexisting comorbid conditions (i.e., heart failure, chronic kidney disease, and cirrhosis) and presenting symptoms (i.e., fever and dyspnea),” they added.

The authors hypothesized that overall disease burden, rather than individual risk factors, may be more helpful, as they observed “strong associations” with CCI score and all measured outcomes.

The authors noted that while multiple risk factors have been reported with both severe illness and death, most prior studies earlier in the pandemic “did not include multivariable adjustment to identify independent risk factors,” were local or regional rather than national populations, and did not compare patients who tested positive with those who tested negative to determine “excess risk” tied to SARS-CoV-2 infection itself versus underlying conditions in those who have SARS-CoV-2 infection, they said.

Researchers examined data from over 88,000 patients in the Veterans Affairs (VA) healthcare system who were tested for SARS-CoV-2 by polymerase chain reaction (PCR) from Feb. 28 to May 14. Overall, 11.4% of patients tested positive.

Those who tested positive were mostly men and almost half were white, while 42% were Black and about 9% were Hispanic. Compared with those who tested negative, those who tested positive were older, more likely to be Black, more likely to have obesity, and more likely to live in states with a high COVID-19 burden, though they had similar distribution of comorbid conditions and CCI scores, the authors noted.

About two-thirds of patients who died were associated with older age groups versus adults 18-49, with 29% of deaths associated with ages 65-79, and 28% of deaths associated with ages 80 and older. For population-attributable fractions of major risk factors for 30-day mortality, male sex contributed 12.3%, comorbidity burden with a CCI score of at least 1 contributed 11%, fever contributed 5.0%, and dyspnea 4.0%, with “negligible contributions from other risk factors.”

Among those who tested positive, older age (older than age 80 vs younger than age 50, adjusted hazard ratio [aHR] 60.80, 95% CI 29.67-124.61) was significantly associated with mortality, as was high regional COVID-19 disease burden (more than 700 vs less than 130 deaths per 1 million residents, aHR 1.21, 95% CI 1.02-1.45) and CCI score (>5 vs 0, aHR 1.93, 95% CI 1.54-2.42). Interestingly, fever (aHR 1.51, 95% CI 1.32-1.72) and dyspnea (aHR 1.78, 95% CI 1.53-2.07) were also significantly associated with death in COVID-19 patients.

Certain lab abnormalities exhibited a dose-response association with mortality, including aspartate aminotransferase, creatinine, and neutrophil:lymphocyte ratio.

As to why certain risk factors for mortality reported by earlier studies did not reach statistical significance, such as race/ethnicity, BMI, hypertension, and COPD, the authors suggested that this “may reflect differences in the study population … differences in the confounders that were adjusted for, or attenuation of racial/ethnic disparities in access to care in the VA system relative to the private sector.”

One notable limitation to the study was its population of veterans, which may not be generalizable to other populations, especially women. They also noted ICD-10 codes were used to determine comorbid conditions and results were limited to patients who were tested within the VA system.


This study was supported using data from the Veterans Affairs COVID-19 Shared Data Resource.

Ioannou disclosed support from the U.S. Department of Veterans Affairs, Office of Research.

Other co-authors disclosed support from Veterans Affairs Health Services Research and Development, the National Institute of Diabetes and Digestive and Kidney Diseases, the CDC, the VA National Center for Ethics in Health Care, UpToDate, Kaiser Permanente Southern California, University of California San Francisco, University of Pennsylvania, University of Alabama, the Denevir Foundation, Hammersmith Hospital, Dialysis Clinics, Inc, Fresenius Medical Care, Chugai Pharmaceutical Co, the Japanese Society of Dialysis Therapy, the New York Society of Nephrology, the Department of Veterans Affairs, the Firland Foundation, and the Patient-Centered Outcomes Research Institute.

Source Article