Study population and setting
This retrospective cohort study included all 43,634 patients seen at health facilities in the TriNetX Research Network from January 1 to June 23, 2020 in the United States with a confirmed SARS-CoV-2 infection. It compared the risk of hospitalization, mortality, thrombosis, and intensive care among those with a history of liver transplant (n=126) to those without (n=43,508) before and after 1:1 nearest neighbor propensity score matching based on age, race, and ten comorbid conditions defined by diagnostic and procedural codes. Of note, patient counts < 10 for any particular comorbidity or symptom were rounded up to 10. Continuous data were compared with t-tests, categorical data were compared with chi-squared tests, and the authors calculated risk ratios of hospitalization, mortality, thrombosis, and intensive care by liver transplant status.
Summary of Main Findings
The 126 patients with a history of liver transplant were older (mean age 57 vs. 50 years), more likely to be male (66 vs. 45%) and white (58 vs. 46%), and had a higher prevalence of hypertension (23 vs. 10%), chronic kidney disease (20 vs. 3%), and hypertension (16 vs. 7%) than the 43,508 patients without a history of liver transplant. Patients with and without liver transplant were more similar after propensity score matching, which included 125 individuals with and 125 without a history of liver transplant. It is difficult to compare COVID-19 presenting symptoms by liver transplant status due to small numbers of liver transplant recipients with any recorded symptoms. Individuals with a history of liver transplant had higher baseline mean bilirubin, alkaline phosphatase, potassium and lower baseline mean hemoglobin, platelets, and lactate dehydrogenase than individuals without a history of liver transplant, although these differences resolved with propensity score matching. The propensity score adjusted risk ratio of 1.72 (95% CI 1.17, 2.53) suggested an increased risk of hospitalization among individuals with a history of liver transplant, however, there were insufficient events to draw any conclusions about differences in risk of mortality, thrombosis, or intensive care before or after propensity score adjustment.
This is one of the largest available studies on the impact of SARS-CoV-2 on patients with a history of liver transplant. Additionally, the propensity score matching allowed the authors to control for important confounding variables such as age and medical comorbidities.
The prevalence of COVID-19 symptoms was extremely low relative to most other published studies, which may reflect a difference in how these data are collected as compared to other COVID-19 studies. No information is provided about the geographic location of patients, the health facility type (community hospitals, academic medical centers, outpatient clinics, etc.), or to what extent diagnostic and procedural code misclassification exists within and between health centers participating in TriNetX, including missing data. These considerations may bias these results, however, it is difficult to predict the extent of the bias or its direction. Rounding up counts less than ten also made it difficult to compare comorbidities or COVID-19 symptoms by liver transplant status, and prevented interpretation of risk ratios for outcomes with ten or fewer occurrences. Finally, the small sample size makes it difficult to generalize results to all patients with a history of liver transplantation who are infected with SARS-CoV-2.
This is one of the largest studies to date of SARS-CoV-2 infection in individuals with a history of liver transplant.
This review was posted on: 14 January 2021