Reply.

Abstract

We thank Dr. Duan for the interest in our paper that reported the association between postvaccination and overall and coronavirus disease 2019 (COVID‐19)–related mortality among participants with cirrhosis. We adjusted for the etiology of alcohol‐associated liver disease versus others based on studies that showed a worse prognosis of COVID‐19 associated with alcohol.1–4 However, as suggested, we now include the various etiologies of liver disease in the baseline characteristics (Table S1). The most common cause of cirrhosis in both the postvaccination and unvaccinated COVID‐19 cohorts was NAFLD (28.7% and 28.2%, respectively), and the two groups were well matched with respect to liver disease etiology. We agree that exposure to proton pump inhibitors (PPIs) may be a potential confounder. Participants with postvaccination COVID‐19 were more likely to be exposed to PPIs than those with unvaccinated COVID‐19 (80.3 vs. 65.8%; p < 0.0001). Third, Dr. Duan suggests socioeconomic status as a potential confounder. Although we did not have data on individual income levels, we examined the socioeconomic status by identifying participant locations using residential zip codes and the median household income associated with these locations. Median household incomes were similar between the two groups ($47,400 vs. 47,100; p = 0.64). We repeated the analysis by including these three variables in the multivariable model (Table 1). Compared with NAFLD cirrhosis, alcohol and HCV cirrhosis were not associated with an increase in overall or COVID‐19‐related death. We observed no association between the median household income and overall (per $1000 change in household income; adjusted HR [aHR] 0.98, 95% CI 0.93–1.07; p = 0.16) or COVID‐19‐related death (aHR 0.99, 95% CI 0.97–1.03; p = 0.11). However, PPI exposure was associated with an increase in overall mortality (aHR 1.61, 95% CI 1.06–2.15; p = 0.001), but not COVID‐19‐related death (aHR 1.08, 95% CI 0.55–1.54; p = 0.75). After inclusion of these variables, postvaccination COVID‐19 continued to be associated with a decrease in overall (aHR 0.25, 95% CI 0.12–0.49; p < 0.0001) and COVID‐19‐related death (aHR 0.27, 95% CI 0.13–0.60; p = 0.001). TABLE 1 - Multivariable HRs for the risk of overall death or COVID‐19‐related death in patients with postvaccination COVID‐19 versus unvaccinated COVID‐19 Variable Overall death COVID‐19‐related death aHR (95% CI) p‐Value aHR (95% CI) p‐Value Number of patients 762 – 762 – Number of events 87 – 64 – Group Control REF REF Vaccine 0.25 (0.12, 0.49) <0.0001 0.27 (0.13, 0.60) 0.0011 Location, n (%) Northeast REF REF Southeast 1.13 (0.51, 2.53) 0.7606 1.42 (0.55, 3.65) 0.4726 Midwest 1.30 (0.64, 2.64) 0.4759 1.46 (0.61, 3.48) 0.3914 South 0.61 (0.26, 1.42) 0.2528 0.87 (0.32, 2.38) 0.7897 Northwest 2.47 (0.98, 6.23) 0.0557 2.53 (0.82, 7.81) 0.1067 Southwest 1.54 (0.68, 3.49) 0.2992 2.29 (0.91, 5.80) 0.0801 Age 1.05 (1.02, 1.08) 0.0015 1.06 (1.02, 1.10) 0.0012 BMI 1.00 (0.97, 1.02) 0.6657 1.01 (0.98, 1.03) 0.6598 Diabetes No REF REF Yes 0.92 (0.56, 1.53) 0.9894 0.78 (0.45, 1.34) 0.3673 Etiology at cirrhosis NAFLD REF REF Alcohol 1.01 (0.49, 2.08) 0.9894 1.41 (0.64, 3.12) 0.3942 HCV+Alcohol 0.68 (0.32, 1.43) 0.3043 0.51 (0.20, 1.29) 0.1544 HCV 0.85 (0.45, 1.59) 0.6024 0.93 (0.52, 2.17) 0.8666 Others 0.29 (0.07, 1.18) 0.0841 NA NA AUDIT‐C score Low REF REF High 1.17 (0.53, 2.57) 0.6902 1.21 (0.47, 3.06) 0.6959 eCTP A REF REF B 1.06 (0.60, 1.89) 0.8335 0.83 (0.42, 1.62) 0.5747 C 1.28 (0.11, 4.17) 0.6659 N/A N/A Dexamethasone No REF REF Yes 4.25 (2.08, 8.69) <0.0001 3.78 (1.55, 9.22) 0.0035 Remdesivir No REF REF Yes 0.75 (0.33, 1.70) 0.4926 1.41 (0.55, 3.62) 0.4778 MELD‐Na 1.03 (0.99, 1.08) 0.1183 1.04 (0.99, 1.09) 0.1661 PPI exposure No REF REF Yes 1.61 (1.06, 2.15) 0.0007 1.08 (0.55, 1.54) 0.7511 Median household income per $1000 0.98 (0.93, 1.07) 0.1567 0.99 (0.97, 1.03) 0.1078 Abbreviations: aHR, adjusted HR; AUDIT‐C, Alcohol Use Disorders Identification Test–Concise; BMI, body mass index; COVID‐19, coronavirus disease 2019; eCTP, electronic Child Turcotte Pugh; MELD‐Na, Model for End‐Stage Liver Disease–Sodium; NA, not available; PPI, proton pump inhibitor.Bold indicates p < 0.05. These analyses reveal similar associations described in our original estimates, indicating that postvaccination COVID‐19 is associated with consistent reductions in overall and COVID‐19‐related death.