Assessing the impact of COVID-19 on kidney diseases: insights from real-world data
Restricted (Penn State Only)
- Author:
- Zhang, Yue
- Graduate Program:
- Epidemiology (PhD)
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 10, 2024
- Committee Members:
- Kathryn Risher, Major Field Member
Vernon Chinchilli, Chair & Dissertation Advisor
Leslie Parent, Outside Unit & Field Member
Nasrollah Ghahramani, Special Member
Djibril M Ba, Special Member
Duanping Liao, Major Field Member
Kathryn Risher, Professor in Charge/Director of Graduate Studies - Keywords:
- COVID-19
Kidney Diseases
Long COVID
Real-World Evidence - Abstract:
- Coronavirus disease 2019 (COVID-19) has been reported to be associated with multiple short-term and long-term conditions following infections. Kidney diseases, including short-term conditions like acute kidney injury (AKI) and long-term conditions like chronic kidney disease (CKD) have attracted significant research attention, particularly regarding the long-term and short-term effects of COVID-19. The biological mechanism underlying the association between COVID-19 and kidney diseases suggests that SARS-CoV-2 may enter cells via Angiotensin-Converting Enzyme 2 (ACE2), which is highly expressed in renal cells. Meanwhile, in vitro studies suggest that the use of Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers (ACEI/ARB) may increase the risk of SARS-CoV-2 entering cells, particularly in the kidney, thereby raising safety concerns regarding the use of ACEI/ARB. Current studies present controversial conclusions regarding whether ACEI/ARB use increases the risk of kidney diseases in patients without COVID-19. While evidence suggests that ACEI/ARB use may be safe for COVID-19 patients in the short term, there remains insufficient real-world evidence on long-term outcomes, especially in patients with Long COVID symptoms. In addition to studying the effects of COVID-19 compared to non-infected individuals, there is growing interest in comparing the impact of COVID-19 with influenza, one of the most common viral respiratory illnesses in the U.S. This comparison can provide more comprehensive information to the public. Research indicates that COVID-19 is associated with higher risks of mortality and several comorbidities, including cardiovascular diseases, mental health disorders, and neurological diseases; however, comparative results on kidney diseases remains limited. Moreover, CKD is highly prevalent but underdiagnosed, with up to 90% of individuals unaware of their condition. With the rapid advancement of machine learning (ML) algorithms, these techniques have been increasingly applied to predict the risk of kidney diseases. However, in the post-pandemic period, COVID-19 has emerged as a significant risk factor for kidney diseases, but most ML algorithms do not account for this factor since their training data predates 2020. Additionally, the use of electronic health record data, the application of advanced ML models, and the inclusion of COVID-19 as a predictor still need assessment. The overall objective of this dissertation is to assess the impact of COVID-19 infections on kidney diseases using real-world data. Based on the findings, we aim to develop ML models to predict individual risk of kidney disease during the post-pandemic period. In Aim 1, we conducted a retrospective cohort study using national electronic health record (EHR) data to explore the association between ACEI/ARB therapy and the risks of AKI, CKD, and all-cause mortality in patients with and without Long COVID. This study was based on the biological mechanism suggesting that ACEI/ARB use may increase the risk of kidney diseases by facilitating SARS-CoV-2 entry into cells. Our findings indicate that ACEI/ARB treatment does not increase the incidence of CKD or AKI, regardless of Long COVID status. However, Long COVID itself is associated with an increased risk of kidney diseases and all-cause mortality. From Aim 1, we concluded that it is not ACEI/ARB therapy but Long COVID that increases the risk of kidney diseases. Building on this, we sought to explore the direct effects of COVID-19 on kidney diseases in Aim 2. We conducted a large retrospective cohort study using national claims data to assess the association between COVID-19 infections and subsequent kidney diseases, using influenza as a positive control and incorporating a negative control to establish clearer associations. Our findings revealed that COVID-19 infections were associated with a 2.3-fold increased risk of developing AKI, a 1.4-fold increased risk of CKD, and a 4.7-fold increased risk of end-stage renal disease compared to influenza. Based on the conclusions from Aim 1 and Aim 2, we leveraged national EHR data and ML to predict the risk of incident AKI and CKD in both the short and long term during the post-pandemic period in Aim 3. We emulated a prospective cohort study to define the cohorts and assess variables, trained and tested eight different ML models using 69 variables. Through a model selection and feature selection process, we finally identified the best-performing models with 9 key variables to predict the risk of AKI/CKD in 1 month/year. A user-friendly application was then developed to support clinicians in conducting timely risk assessments and ongoing surveillance of kidney diseases. In summary, our real-world evidence research demonstrated that COVID-19 and Long COVID are associated with both short- and long-term effects on kidney diseases. Machine learning models utilizing EHR data, including COVID-19 infections, can effectively predict the incidence of kidney diseases. Ultimately, we developed an application to assist physicians in predicting individual risk in clinical settings.