Characteristics of participants
76 eligible AILD patients and 136 HC were included in the study. The characteristics of participants are shown in Table 1. Briefly, the median age was 54.0 years (IQR: 48.8-60.2 years) in AILD patients and 52.0 years (IQR:33.0-62.2 years) in the HC group. The majority of participants were female (85.5% [65/76] in AILD patients and 55.1% [75/136] in HC). The median BMI and proportion of vaccine types in AILD patients and HC were similar (22.5 kg/m2 [IQR: 21.2-23.9 kg/m2] vs. 22.9 kg/m2 [IQR: 21.1-25.3 kg/m2]). The median post-vaccination time was 41.5 days (IQR: 29.0-66.2 days) and 55.0 (IQR: 33.0-86.5 days) for the AILD patients and HC. Additionally, among these AILD patients, 20 (26.3%) had cirrhosis and almost half (46.1%, 35/76) received one or more immunosuppressant medications.
COVID‑19 vaccination safety
The overall incidence of AEs within 7 days (25.0% [19/76] vs 17.6% [24/136]) and 30 days (25.0% [19/76] vs 17.6% [24/136]) after COVID-19 vaccination were slightly higher in AILD patients than in HC group (p = 0.20) (Table 2). All AEs were mild and none of them had any serious AEs. The common AEs in AILD patients in local AEs were pain at the injection site (4.0%, 3/76); in systemic AEs were fatigue (5.3%, 4/76) and headache (5.3%, 4/76). The most common AEs in HC was local pain at the injection site (8.1%, 11/136) (Table 2). Notably, one patient increased serum gamma-glutamyl-transpeptidase (GGT) level from 20 U/L to 90 U/L (upper limit of normal: 45 U/L) after vaccination and returned to normal after continuing the original treatment strategy during the following up. Another patient's antinuclear antibody (ANA) was positive at a titer of 1:320 before vaccination. It increased to 1:1000 after vaccination, but the liver function test, serum IgG, anti-liver-kidney microsomal, anti-smooth muscle, anti-mitochondrial antibodies, and anti-soluble liver antigen were normal, and the patient had no symptoms of discomfort.
Antibody responses after COVID-19 vaccination
The seropositivity for anti-RBD-IgG was 97.4% (74/76) in AILD patients, which was similar with HC group (100%) (p = 0.13) (Figure 1A). However, anti-RBD-IgG levels were significantly lower in AILD patients than in HC (mean: 49.1 AU/mL vs 71.9 AU/mL, p = 0.02) (Figure 1B). Regarding NAbs, seropositivity (63.2% [48/76] vs 84.6% [115/136]) and antibody titers were both significantly lower in AILD patients than in HC group (p < 0.001) (Figure 1D, 1E). Compared with controls, anti-RBD-IgG and NAbs levels seem to decrease slightly over time since the second dose vaccination in AILD patients (Figure 1C, 1F).
Effect of immunosuppressants on antibody responses
Analysis of clinical features showed no significant correlations between poor antibody responses and parameters such as age, sex, BMI, cirrhosis, and comorbidities (all p > 0.05) (Table 3). However, low-level antibodies were significantly related to the types of vaccine (p = 0.01) (Table 3, Supplementary figure 1A) and the use of immunosuppressants (p = 0.04) (Table 3, Supplementary figure 1C). Furthermore, after adjusting for potential confounding factors (age, BMI, gender, cirrhosis, comorbidities, types of vaccine, and days between final dose and antibody test) in multiple logistic regression analysis, the use of immunosuppressants remained significantly related to low-level antibody levels (Table 4). Compared with patients with no immunosuppression medication, the crude odds ratio (OR) of low-level antibody response risk among patients who used immunosuppressants was 3.3 (95% CI, 1.3-8.5; p = 0.01), and their adjusted OR (AOR) increased to 4.9 (95% CI, 1.5-15.6; p = 0.01). Notably, the risk trend does not seem to increase with the number of immunosuppressants. The AOR of use of one and more immunosuppressive medications were 5.0 (95% CI, 1.1-23.1; p = 0.04) and 4.9 (95% CI, 1.1-19.2; p = 0.03), respectively.
Similar results were also observed for NAbs responses in the AILD patients. Negative NAbs were significantly associated with the types of vaccine (p = 0.01) (Supplementary figure 1B) and immunosuppressants (p = 0.02) (Supplementary figure 1D), except for age, gender, BMI, cirrhosis, and comorbidities (Table 3). After adjusting for confounding factors, the immunosuppressant was associated with a reduced probability of NAbs seropositivity (AOR, 3.0; 95% CI, 1.0-8.9; p = 0.04), especially when the use of ≥2 immunosuppressive medications (AOR, 4.4; 95% CI, 1.3-15.3; p = 0.02) (Table 4).
Our study suggested that immunosuppressive therapy was an independent risk factor for the poor antibody responses to COVID-19 vaccination in patients with AILD.
Specific B cells responses after COVID-19 vaccination
To furtherly investigate the humoral immune response to the SARS-CoV-2 vaccine, the frequency and phenotype of specific B cells were also detected. As expected, the percentage of specific B cells was very low in the peripheral blood of AILD patients and the HC group. No significant difference was found in the frequency of RBD-specific B cells (CD3-CD19+RBD+) and IgG RBD-specific memory B cells (MBCs) (IgG+CD3-CD19+RBD+CD27+) between AILD and HC groups, regardless of immunosuppressants (Figure 3A, 3B). However, the frequency of IgM RBD-specific MBCs (IgM+CD3-CD19+RBD+CD27+) was significantly lower in AILD patients with (17.2% vs 25.3%, adjusted p < 0.01) or without immunosuppressant (19.4% vs 25.3%, adjusted p = 0.03) than HC group (Figure 3C). To better understand the functional phenotype of the RBD-specific MBCs, we further compared RBD-specific resting MBCs (CD3-CD19+RBD+CD21+CD27+), RBD-specific activated MBCs (CD3-CD19+RBD+CD21-CD27+), RBD-specific atypical MBCs (CD3-CD19+RBD+CD21-CD27-), and RBD-specific intermediate MBCs (CD3-CD19+RBD+CD21+CD27-) in between AILD patients and HC groups. Compared with HC, AILD patients without immunosuppressants had a lower frequency of RBD-specific activated MBCs (13.0% vs 16.9%, adjusted p = 0.03) and a higher frequency of RBD-specific intermediate MBCs (47.1% vs 39.9%, adjusted p = 0.02), but not in patients with immunosuppressant. Moreover, there was no significant difference in RBD-specific resting MBCs and RBD-specific atypical MBCs between AILD patients and the HC groups (Figure 3). These results indicate that patients with AILD may develop the humoral immunity as robust as in a healthy population when receiving a booster dose or against SARS-CoV-2 infection despite ongoing immunosuppression.