Patients with a vitamin D deficiency were four times more likely to be COVID-19 positive than those with a sufficient amount of the crucial vitamin, a University of Florida research team has found. The study also found African American patients with insufficient vitamin D levels were three times more likely to be infected by the coronavirus.
Insufficient vitamin D has been linked with respiratory infections, and research suggests vitamin D treatment may help lessen symptoms of COVID-19. African Americans are disproportionally affected by COVID-19 and vitamin D deficiency compared to other races for a variety of reasons including skin pigmentation, which reduces vitamin D production.
An interdisciplinary team of investigators took a unique approach for the cross-sectional association study, which was published in the December 2020 edition of the medical journal Nutrition.
Led by Dr. Joseph Katz, a UF College of Dentistry oral medicine professor, the team used the university’s powerful patient database and its Informatics for Integrating Biology and Bedside, or i2b2, patient registry platform to review aggregate data of more than 980,000 patients seen at UF Health medical practices between 2015 and 2020.
The National Institutes of Health-funded i2b2 is a self-service tool accessible to researchers across the university.
“Research can start out ‘on the bench,’ where you move from small details and work to create a big picture and discover conclusions about a condition or ailment,” Katz said. “In our case, however, we approached the challenge like a jigsaw puzzle in reverse. We were hypothesis-driven: We started with the big picture, that oral diseases are linked to COVID-19. But we didn’t know exactly why. We had to break down the puzzle to learn the reasons.”
Vitamin D deficiency is a key factor in oral health challenges, such as periodontal disease, and its effects on the immune system are well-documented. The receptor that acts as a host to the COVID-19 virus, allowing it to enter and infect human cells, is abundant not only in the lungs and other major organs, but also in the mouth and oral tissues.
Now that vitamin D deficiency has been linked to COVID-19 susceptibility, the next step, Katz said, is for nutrition experts and others to continue that line of research, looking at individualized patient details for more specific causes and results. For example, studies could be conducted to validate the hypothesis that vitamin D supplementation can help prevent or treat COVID-19.
Katz collaborated with Dr. Wei Xue, a UF research assistant professor of biostatistics in the UF College of Public Health and Health Professions and the UF College of Medicine, and her colleague Sijia Yue, a UF biostatistics data management analyst. They used the patient data registry to define cohorts of interest and to drill down to identify commonalities and significant risk factors.
They sorted the exhaustive patient dataset by diagnoses codes, risk factors and demographic factors, and made statistical adjustments for comorbidities and demographic variables. For Katz, the ability to collaborate with Xue enabled the team to go deep into the data.
For her part, Xue said, “I came to UF last year and was not aware of the i2b2 until September, when Dr. Katz suggested we use the aggregated data to study these associations. Without the right data, investigators cannot test and verify their theories. Because of this tool, we could conduct more complicated analyses and not just a summary of statistics.”
The i2b2 provides de-identified patient clinical information (without individual personal/protected details), which is pre-approved for use and bypasses the standard Institutional Review Board, or IRB, process. Use of the i2b2 tool can replace requests for tailored data from the Integrated Data Repository Research Services team.
“Traditionally, researchers use i2b2 to determine if a study premise is feasible by asking if enough patients exist within certain criteria for an in-depth study,” said UF Health Chief Research Information Officer Dr. Chris Harle, who leads the IDR as a joint program of the UF Clinical and Translational Science Institute and UF Health Information Technology.
“This research team went beyond the norm,” Harle said. “They demonstrated how the i2b2 can be used to conduct cross-sectional analyses and to identify hypothesis-generating associations that stimulate future studies. Using i2b2 in this creative way can lead to more rapid research production by eliminating the need for IRB approval and customized data extractions.”
Katz has published other papers based on i2b2 data and encourages other research teams with strong biostatistics abilities to maximize use of the platform.