Clinical Characteristics, Hospitalization, and Mortality Rates of Coronavirus Disease 2019 Among Liver Transplant Patients in the United States: A Multicenter Research Network Study

Abstract

Liver transplant (LT) patients represent one of the largest immunosuppressed cohorts. However, outcomes of coronavirus disease (COVID-19) in this population remain poorly defined although liver injury has been reported in patients with COVID-19.1Zhang C. et al.Lancet Gastroenterol Hepatol. 2020; 5: 428-430Abstract Full Text Full Text PDF PubMed Scopus (1151) Google Scholar We sought to examine the characteristics of LT patients infected with COVID-19 and study the rates of hospitalization, mortality, thrombosis, or intensive care unit (ICU) requirement in LT with COVID-19 in the United States. We used a large health research network (TriNetX) to compile electronic medical records (EMRs) of adult (aged ≥18 years) LT recipients with confirmed severe acute respiratory syndrome coronavirus 2 infection (LT group) from 35 health care organizations in the United States, from January 1, 2020, to June 23, 2020. Within this same time period, we also identified COVID-19–positive patients with no history of LT (non-LT group). For both cohorts, we collected demographics, comorbidities, clinical symptoms, and laboratory findings at COVID-19 diagnosis and presentation. To address confounders, cohorts were balanced using 1:1 greedy nearest neighbor propensity score matching (PSM) based on age, race, and key comorbidities (Table 1).2Guan W.J. et al.Eur Respir J. 2020; 55: 2000547Crossref PubMed Scopus (8) Google Scholar The 4 outcomes of interest were risk of hospitalization (defined as composite outcome of inpatient or critical care services), mortality, thrombosis (defined as composite outcome of deep vein thrombosis, acute pulmonary embolism, stroke, or myocardial infarction), and ICU requirement (requiring mechanical ventilation or extracorporeal membrane oxygenation) after a diagnosis of COVID-19. Further details on methodology are provided in the Supplementary Material.Table 1Comparison of Medication Use Among Non-LT Patients With COVID-19 and LT Patients With COVID-19, Before 6 Months of Diagnosis of COVID-19CharacteristicsBefore matchingAfter matchingLTNon-LTRR (95% CI)P valueLTNon-LTRR (95% CI)P valuen12643,508125125Demographics Age, y, mean ± SD57.08 ± 13.2850.06 ± 18.66<.000157.03 ± 13.3259.83 ± 14.71.116 Female, n (%)43 (34)23,844 (55)<.000143 (34)40 (32).687 Male, n (%)83 (66)19,576 (45)<.000182 (66)85 (68).687 Unknown sex, n (%)088 (<1).613300 White, n (%)73 (58)19,901 (46).006173 (58)71 (57).798 Black or African American, n (%)34 (27)11,157 (26).730834 (27)40 (32).4058 American Indian or Alaska Native, n (%)10 (8)a142 (<1)<.000110 (8)10 (8)1 Asian, n (%)10 (8)a1207 (3).000410 (8)10 (8)1 Native Hawaiian or Other Pacific Islander, n (%)10 (8)a88 (<1)<.000100 Unknown race, n (%)13 (10)11,013 (25).000113 (10)10 (8).5115 Hispanic or Latino, n (%)14 (11)6064 (14).360214 (11)13 (10).8385 Not Hispanic or Latino, n (%)78 (62)18,279 (42)<.000177 (62)54 (43).0036 Unknown ethnicity, n (%)34 (27)19,165 (44).000134 (27)58 (46).0016Comorbid conditions, n (%) Essential (primary) hypertension29 (23)4180 (10)<.000129 (23)23 (18).3498 Chronic kidney disease25 (20)1411 (3)<.000124 (19)26 (21).7518 Diabetes mellitus20 (16)3106 (7).000120 (16)20 (16)1 Nicotine dependence10 (8)a673 (2)<.000110 (8)0.0012 Chronic lower respiratory diseases10 (8)a2346 (5).20710 (8)10 (8)1 Heart failure10 (8)a1131 (3).000210 (8)10 (8)1 Cerebrovascular diseases10 (8)a485 (1)<.000110 (8)10 (8%)1 Alcohol-related disorders10 (8)a207 (<1)<.000110 (8)a0.0012 Dyspnea10 (8)a4632 (11).324610 (8)a11 (9).8196 Ischemic heart diseases10 (8)a1438 (3).003810 (8)a10 (8)a1Presenting symptoms, n (%) Fever of other and unknown origin12 (10)4664 (11).664712 (10)18 (14).2429 Cough10 (8)a6863 (16).015910 (8)a17 (14).1538 Nausea and vomiting10 (8)a1122 (3).000210 (8)a10 (8)a1 Malaise and fatigue10 (8)a1750 (4).025710 (8)a10 (8)a1 Diarrhea, unspecified10 (8)a1493 (3).005610 (8)a10 (8)a1 Abdominal and pelvic pain10 (8)a695 (2)<.000110 (8)a10 (8)a1 Acute pharyngitis10 (8)a1045 (2).000110 (8)a10 (8)a1 Hypoxemia10 (8)a2139 (5).117710 (8)a10 (8)a1Laboratory test results, mean (SD) Sodium, mEq/L135.55 ± 5.03136.58 ± 5.1.1262135.67 ± 5137.06 ± 6.09.2015 Creatinine [mass/volume] in serum, plasma, or blood,2.03 ± 2.071.36 ± 1.65.00212.04 ± 2.093.14 ± 4.45.103 Hemoglobin, g/dL10.88 ± 2.4112.52 ± 2.48010.88 ± 2.4112.42 ± 2.38.0018 Platelets, n/μL167.11 ± 103.94220.64 ± 94.550167.11 ± 103.94199.18 ± 69.46.0797 Leukocytes, n/μL6.56 ± 5.427.74 ± 5.55.12196.56 ± 5.426.76 ± 2.7.8378 Alanine aminotransferase, U/L57.6 ± 154.5641.59 ± 99.82.250757.6 ± 154.5635.87 ± 37.27.3931 Aspartate aminotransferase, U/L74.06 ± 245.7955.22 ± 214.63.528474.06 ± 245.7949 ± 54.5338 Alkaline phosphatase, U/L152.15 ± 14388.51 ± 58.230152.15 ± 14388 ± 74.0123 Potassium, mEq/L4.28 ± 0.653.9 ± 0.5804.29 ± 0.664.06 ± 0.8.1173 Total bilirubin, mg/dL1.61 ± 3.390.63 ± 0.8601.61 ± 3.390.59 ± 0.27.0655 Albumin, g/dL3.29 ± 0.823.43 ± 0.76.2053.29 ± 0.823.35 ± 0.68.7183 Neutrophils, n/μL4.53 ± 5.17.51 ± 98.46.86424.53 ± 5.15.19 ± 1.86.512 Body mass index, kg/m227.74 ± 5.4230.14 ± 8.11.11827.74 ± 5.4230.45 ± 9.03.2246 Prothrombin time, s14.94 ± 4.1714.25 ± 6.76.612314.94 ± 4.1714.3 ± 4.68.656 C-reactive protein, mg/dL48.74 ± 6376.44 ± 86.42.13348.17 ± 64.49102.94 ± 92.72.0242 Lactate dehydrogenase, mmol/L244.9 ± 90.61400.01 ± 315.65.0281244.9 ± 90.61394.88 ± 225.85.008 Ferritin, ng/mL9333.93 ± 35,380.0222,453.26 ± 76,506.63.49393,33.93 ± 35,380.0223,411.78 ± 89,271.63.5533 Activated partial thromboplastin time, s34.31 ± 11.831.38 ± 10.61.301434.31 ± 11.832.9 ± 10.39.7445 Creatine kinase, mg/dL105.92 ± 63.84399.89 ± 2840.48.7091105.92 ± 63.84280.08 ± 288.57.0446 Fibrin D-dimer FEU2.81 ± 5.97196.24 ± 949.4.49952.81 ± 5.973.31 ± 2.22.8075 Gamma glutamyl transferase, U/L42 ± 31.11139.51 ± 174.46.081442 ± 31.1144 ± 0.8412 Erythrocyte sedimentation rate44 ± 31.5745.44 ± 29.36.877244 ± 31.5732.6 ± 17.26.3296 Interleukin 6, pg/mL31.6 ± 0120.58 ± 389.98.471131.6 ± 0224.37 ± 277.5.0414 Procalcitonin, ng/mL2.33 ± 6.3922.07 ± 485.75.89782.33 ± 6.391.21 ± 1.08.5728 Lymphocytes, n/μL0.79 ± 0.280.9 ± 3.14.90820.79 ± 0.280.41 ± 0.27.0058Outcomes Hospitalization50 (40)5510 (13)3.13 (2.52–3.89)<.000150 (40)29 (23%)1.72 (1.17–2.53).0043 Mortality10 (8)a1523 (4)2.27 (1.24–4.12).006910 (8)a10 (8)a1 (0.43–2.32)1 Thrombosis10 (8)a972 (2)3.55 (1.95–6.46)<.000110 (8)a10 (8)a1 (0.43–2.32)1 Intensive care10 (8)a1310 (3)2.64 (1.45–4.79).001310 (8)a11 (9)0.91 (0.40–2.06).8196NOTE. Comparison shown both before and after propensity score matching.SD, standard deviation. Open table in a new tab NOTE. Comparison shown both before and after propensity score matching. SD, standard deviation. Between January and June 2020, there were a total of 43,508 non-LT patients with COVID-19 and 126 LT patients with COVID-19 in the database (Table 1). LT patients were significantly older and predominately male and white, and they had a higher prevalence of comorbidities (Table 1). Thus, we performed (1:1) PSM for age, race, and comorbidities. The LT and non-LT groups were relatively balanced after PSM (n = 125 each group) (Table 1). LT patients were more likely to have nausea and vomiting, malaise and fatigue, diarrhea, and abdominal and pelvic pain. LT patients were more likely to have higher mean levels of creatinine (Cr), total bilirubin, and alkaline phosphatase (Table 1). Within 6 months before diagnosis of COVID-19, 39% of LT patients were receiving prednisone, 9% hydrocortisone, 61% tacrolimus, 37% mycophenolate mofetil, and 8% each azathioprine, cyclosporine, sirolimus, everolimus, and basiliximab (Supplementary Table 1). Patients in the LT group had a significantly higher risk of hospitalization compared to the non-LT group, both before and after PSM (Table 1). After PSM, in adjusted analysis, 40% of patients in the LT group required hospitalization compared to 23% of patients in the non-LT group (risk ratio [RR], 1.72; P < 0.0043). In unadjusted analyses, the risk of mortality (RR, 2.27; P = .0069), thrombosis (RR, 3.55; P < .0001), and ICU requirement (RR, 2.64; P = .0013) was higher in the LT group; however, after PSM, there was no difference in risk of mortality, thrombosis, and ICU requirement between LT and non-LT patients with COVID-19 (Table 1). We found LT patients with COVID-19 to have significantly higher risk of hospitalization but not a higher risk of mortality, thrombosis, or ICU requirement compared to patients without LT and COVID-19 upon adjusted analyses. This is the largest study of LT patients with COVID-19 in the United States to date, to our knowledge. Yi et al3Yi SG et al. Transplantation. https://doi.org/10.1097/TP.0000000000003339.Google Scholar reported 21 solid organ transplant recipients diagnosed with COVID-19, including 3 LT patients, at a US high-volume transplant center. In this study, 33% (1/3) of LT patients required hospitalization compared to 40% in our study, and 33% (7/21) of solid organ transplant patients required ICU care compared to 8% in our study. Belli et al4Belli L.S. et al.Lancet Gastroenterol Hepatol. 2020; 5: 724-725Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar reported the European experience in 103 LT patients with COVID-19 from centers located in Italy, Spain, and France. Although they found fever, cough, and shortness of breath to be the most common presenting symptoms, we found LT patients to have a predominance of nausea and vomiting, malaise and fatigue, diarrhea, abdominal and pelvic pain. These differences in presenting symptoms might be due to differences in study design, methods of data collection, data analyses, and the size of source population. Although 40% of patients in our study were admitted to the hospital and 8% required ICU care, 81% of patients in their study required hospitalization, and 15% were admitted to the ICU. Importantly, 16% of LT patients died in their study compared to a mortality rate of 8% in our study. Our lower rate of hospitalization and ICU care requirements compared to the European experience likely suggests earlier presentation and/or diagnosis in our patients. Furthermore, ICU requirement in our study was defined as requiring mechanical ventilation or extracorporeal membrane oxygenation, whereas in other studies, the definitions were more liberal, thereby leading to a lower estimate of ICU requirement in our study. Other factors such as increased accessibility to a multidisciplinary post-LT team and decreased threshold of admission for LT patients may also have played a role. LT patients with COVID-19 had higher mean levels of Cr (2.03), suggestive of acute kidney injury compared to non-LT patients with COVID-19. Approximately 15% to 29%5Yang X. et al.Lancet Respir Med. 2020; 8: 475-481Abstract Full Text Full Text PDF PubMed Scopus (6085) Google Scholar of patients with COVID-19 have been reported to have elevated Cr. Although a significant proportion of LT patients were on calcineurin inhibitors, ACE2 expression in kidney is known to be nearly100-fold higher than in respiratory organs and may increase the risk of acute kidney injury in patients with COVID-19.6Cheng Y. et al.Kidney Int. 2020; 97: 829-838Abstract Full Text Full Text PDF PubMed Scopus (1592) Google Scholar In a recent Dutch study7Klok F.A. et al.Thromb Res. 2020; 191: 145-147Abstract Full Text Full Text PDF PubMed Scopus (2866) Google Scholar of patients with COVID-19 pneumonia admitted to the ICU, 27% developed venous thromboembolism, and 3.7% developed arterial thrombotic events. In our study, the overall rate of thrombosis was 8%. The decreased rates of thrombosis in our study might be due to the differences in the study population because the Dutch study included only patients with COVID-19 pneumonia admitted to the ICU who had severe disease leading to increased inflammatory burden. Ours is one of the first studies in LT patients with COVID-19 to provide data on thrombosis, which appear to be reassuring. This study is limited by its retrospective nature; the inability to access treatment regimens, if any, for patients with COVID-19; and other information unavailable in the TriNetX database, such as information about socioeconomic status, exposure history, and geographic data of the patient population. In addition, data from EMR-based databases is susceptible to coding errors during the translation of patient information into International Classification of Diseases, 10th Revision, codes. However, TriNetX aggregates data from EMRs in real time, which minimizes errors in data collection and analysis. Furthermore, patients with mild disease who were undiagnosed and did not present to health care organizations were not captured in our study, and thus, our cohort likely represents a relatively severe spectrum of COVID-19. However, compared to prior studies on LT and COVID-19, our estimate of hospitalization and mortality rates in LT might be more precise given our higher sample size in both the LT and non-LT groups with COVID-19. In conclusion, in one of the largest multicenter network studies on LT and COVID-19 to date, we found LT patients with COVID-19 to have a significantly higher risk of hospitalization but not mortality, thrombosis, or ICU requirement compared to patients without LT and COVID-19 when matched for severity of illness. Given the limitations and retrospective nature of this study, further prospective studies are needed to evaluate the burden of care in LT patients and the long-term outcomes of LT patients with COVID-19. Collaborators: Alexandra Mills, MBA,1 Kayla Schlick, MS,1 and Ahmad Khan, MD2 1Clinical Research Center, University Hospitals Cleveland Medical Center, Cleveland, Ohio; and 2Department of Internal Medicine, West Virginia University–Charleston Division, West Virginia. Emad Mansoor, MD (Conceptualization: Lead; Investigation: Supporting; Methodology: Supporting; Project administration: Equal; Writing – original draft: Lead; Interpretation: Equal); Abe Perez, PhD (Data curation: Supporting; Formal analysis: Lead; Methodology: Lead); Mohannad Abou-Saleh, MD (Writing – review & editing: Supporting); Seth N. Sclair, MD (Writing – review & editing: Supporting); Stanley Cohen, MD (Writing – review & editing: Supporting); Gregory S. Cooper, MD (Conceptualization: Supporting; Investigation: Equal; Methodology: Supporting; Supervision: Lead; Writing – review & editing: Lead). Download .pdf (.12 MB) Help with pdf files Characteristics of Liver Transplant Patients Infected with COVID-19GastroenterologyVol. 161Issue 3PreviewWith great interest, we have read this multicenter research network study from Mansoor and colleagues entitled “Clinical Characteristics, Hospitalization and Mortality Rates of COVID-19 Among Liver Transplant Patients in the United States.”1The study had described the characteristics of liver transplant (LT) patients infected with coronavirus disease 2019 (COVID-19) and the rates of hospitalization, mortality, thrombosis, and intensive care unit (ICU) requirements. Within their study, fever (10%) was the most common presenting symptom, followed by cough (8%), malaise and fatigue (8%), dyspnea (10%), and gastrointestinal symptoms (8%). Full-Text PDF COVID-19 Among Liver Transplant RecipientsGastroenterologyVol. 161Issue 3PreviewMansoor et al1 sought to examine the rates of hospitalization, mortality, thrombosis, or intensive care (ICU) requirement of liver transplant (LT) patients infected with coronavirus disease 2019 (COVID-19) in the United States compared with patients without LT and COVID-19. The authors found that upon adjusted analyses LT patients with COVID-19 had a significantly higher risk of hospitalization, but not a higher risk of mortality, thrombosis, or ICU requirement compared with patients without LT and COVID-19. Full-Text PDF