2.1 Study design and participants

Between June 1, 2021, and September 30, 2021, we performed a prospective observational study at the Second Affiliated Hospital of Chongqing Medical University, a large tertiary medical center in southwestern China. Participants included PWCD treated in outpatient clinics and healthy volunteers (healthy controls) who underwent annual physical examinations at the health management center. Inclusion criteria were age 18 years or older, no SARS-CoV-2 infection before receipt of the first vaccine dose (determined based on either a negative anti-SARS-CoV-2 immunoglobulin (Ig)-M/IgG test or the absence of a positive polymerase chain reaction assay result for SARS-CoV-2, with no history of suspected clinical SARS-CoV-2 infection), and ability to understand and complete questionnaires in Chongqing. Participants with known pregnancy during study entry, those who did not complete the full course of vaccination, and those who provided incomplete vaccination information (including the date of their first vaccine dose and complete vaccination, and vaccine manufacturer [due to different vaccine types by manufacturers]) were excluded. All participants in this study were required to provide peripheral blood samples for serological assays for SARS-CoV-2 within 21–120 days after full-course vaccination.

2.2 Variables and definitions

Clinical data, including age, sex, body mass index (BMI) (the weight in kilograms divided by the square height in meters), comorbidities, and SARS-CoV-2 vaccination information (including the date of the first vaccine dose and complete vaccination, and vaccine manufacturer), were collected by the investigators using a standardized questionnaire and case record form (CRF). Investigators completed CRFs using data from the electronic patient files. Patient files were used to register the disease diagnosis and laboratory test results.

Serum samples from vaccinated participants were collected by venipuncture and tested for SARS-CoV-2 antibodies, including IgG antibodies against the RBD of the SARS-CoV-2 spike protein (RBD-IgG) and neutralizing antibodies (NAbs) using capture chemiluminescence immunoassays by MAGLUMI X8 (Snibe) according to the manufacturer's instructions. According to the kit labels, RBD-IgG tests have 100% sensitivity and 99.6% specificity for the diagnosis of COVID-19, while NAbs tests have 100% sensitivity and 100% specificity. The cut-off values were 1.0 AU/ml for RBD-IgG and 0.15 µg/ml for NAbs.

Full-course SARS-CoV-2 vaccination was defined as having had two vaccinations of inactivated vaccines (BBIBP-CorV or CoronaVac) or three vaccinations for the RBD-based protein subunit vaccine (ZF2001), regardless of the time interval.^29-33

2.3 Statistical analysis

Data were presented as mean (standard deviation [SD]) or median (interquartile range [IQR]) for continuous variables and as proportions for categorical variables. Descriptive analyses were performed in the PWCD and healthy control groups and included quantification of seroconversion and the antibody titer, as well as a description of their distribution in various disease subgroups. Differences in continuous variables were compared using the Student t-test or Mann–Whitney U-test. Differences in categorical variables (age dichotomized, sex, and vaccine types) were compared using the χ² test. Correlations between NAbs and RBD-IgG levels for each period were assessed using the Spearman rank correlation.

For the primary objective, associations of disease status (binary exposure variable, with or without diseases) with seroconversion (binary outcome variable, positive or negative) were assessed using univariate and multivariate logistic regression models. Associations of disease status with antibody titers (continuous outcome variable) were assessed using linear regression analysis in univariate and multivariate models. In the secondary objective, we also assessed associations between the number of coexisting diseases with seroconversion and the antibody level. Four models were used, with model 1 being unadjusted; model 2 was adjusted for age and sex (binary variable); model 3 was adjusted for model 2 variables plus the BMI, vaccine types, and days between the final dose and antibody test (continuous variables); and model 4 adjusted for model 3 variables plus the white blood cell count, lymphocyte count, hemoglobin level, alanine aminotransferase level, and total bilirubin level (continuous variables). Data are presented as point estimates and corresponding 95% confidence intervals (CIs) of the effect size estimates. Unadjusted and adjusted models were constructed to evaluate the association of diseases and the risk of seroconversion and antibody levels in the primary and secondary objectives. Missing data regarding RBD-IgG or NAbs levels were excluded from the analysis.

Additionally, we performed stratified analyses in prespecified subgroups defined by age, sex, BMI, vaccine type, and time between the final dose and antibody test. Interactions by subgroup characteristics were tested using likelihood ratio tests comparing models with and without multiplicative interaction terms.

All analyses were performed using R (http://www.R-project.org, the R Foundation) and EmpowerStats (http://www.empowerstats.com, X&Y Solutions, Inc.). A p < 0.05 (two-sided) was considered to indicate statistical significance.

3.1 Participant characteristics

In total, 1645 participants were included in the study, including 245 (14.9%) healthy volunteers (healthy control group) and 1400 (85.1%) PWCD (PWCD group). Table 1 presents the demographic characteristics, laboratory test results, comorbidities, and vaccination information of these participants. The mean ages of participants were 46.1 years (SD: 14.8) in the healthy control group and 50.3 years (SD: 13.8) in the PWCD group. The proportions of men were 41.6% (102/245) and 54.1% (757/1400) in the healthy control and PWCD groups, respectively. The PWCD group had patients with 8 common chronic disease types, including 552 (39.4%) with chronic hepatitis B, 289 (20.1%) with cancer (including liver, lung, breast, colon, rectal, cervical, ovarian, endometrial, and gastric cancers), 143 (10.2%) with cardiovascular disease (hypertension and coronary heart disease), 108 (7.7%) with autoimmune disease (autoimmune liver disease, Hashimoto thyroiditis, Sjogren syndrome, and polymyositis), 105 (7.5%) with diabetes mellitus, and 77 (5.5%) with chronic kidney disease.

3.2 Humoral response after SARS-CoV-2 vaccination

The seropositivity rates of NAbs and RBD-IgG were 72.1% (1010/1400) and 84.8% (1182/1400) in the PWCD group, and the median titers of NAbs and RBD-IgG, 0.26 μg/ml (IQR: 0.14–0.50) and 4.22 AU/ml (IQR: 1.58–11.81), respectively, were significantly lower in the PWCD group than in the healthy controls group (Supporting Information: Table 1). As expected, both seropositivity rate and antibody levels tended to decrease within 120 days after completed vaccination. The decrease in NAbs and RBD-IgG titers was rapid within 60 days but slowed thereafter (Supporting Information: Table 1; Figure 1A,B). The seropositivity rate and antibody levels of NAbs and RBD-IgG also varied considerably between the eight predefined disease groups (Figure 1C–F).

3.3 Impaired humoral responses in the PWCD group

In the primary analysis, the PWCD group had significantly lower seroconversion and antibody levels of NAbs and RBD-IgG than the healthy control group in the unadjusted model (model 1) and partially adjusted models (models 2 and 3) (Figure 2B). In the fully adjusted model (model 4), the odds ratios (OR) for NAbs and RBD-IgG seroconversion were 0.28 (95% CI: 0.14, 0.55, p < 0.001) and 0.26 (95% CI: 0.10, 0.68, p = 0.001), respectively, in the PWCD group (Figure 2A). The titers of NAbs (β: −0.76; 95% CI: −1.19, −0.32; p < 0.001) and RBD-IgG (β: −17.01; 95% CI: −23.29, −10.72; p < 0.001) were still significantly lower in the PWCD group than in the healthy control group (Figure 2B). Additionally, chronic hepatitis B (OR: 0.65; 95% CI: 0.46, 0.93, p = 0.02), cancer (OR: 0.65; 95% CI: 0.42, 0.99, p = 0.04), and diabetes mellitus (OR: 0.50; 95% CI: 0.28, 0.89, p = 0.02) were associated with lower seroconversion of NAbs than the other predefined diseases (Supporting Information: Figure 1A). Chronic kidney disease (OR: 0.29; 95% CI: 0.11, 0.76, p = 0.01), cancer (OR: 0.38; 95% CI: 0.23, 0.62, p < 0.001), and diabetes mellitus (OR: 0.37; 95% CI: 0.20, 0.69, p < 0.01) were associated with lower seroconversion of RBD-IgG than the other predefined diseases (Supporting Information: Figure 1B). Regarding the antibody level, only the presence of autoimmune disease showed significantly lower NAbs (β: −0.60; 95% CI: 1.18, −0.03; p = 0.04) and RBD-IgG titers (β: −11.13; 95% CI: −19.47, −2.78; p = 0.01) (Supporting Information: Figure 1C,D).

In the secondary analysis, the seropositivity rates of NAbs and RBD-IgG in healthy controls, participants with 1 disease, and participants with ≥2 coexisting diseases were 84.1% (206/245) and 72.7% (838/1152), 69.4% (172/248) (p < 0.001) and 89.8% (220/245), and 85.7% (987/1152) and 78.6% (195/248) (p = 0.002), respectively. In the fully adjusted model (model 4), significantly lower seroconversion of NAbs (OR: 0.26; 95% CI: 0.12, 0.54, p < 0.001) and RBD-IgG (OR: 0.18; 95% CI: 0.06, 0.48, p < 0.001) were associated with the presence of ≥2 coexisting diseases (Figure 3A–F). Concerning the antibody level, participants with ≥2 coexisting diseases also had a significantly lower NAbs (β: −0.90; 95% CI: −1.44, −0.37; p < 0.001) and RBD-IgG titer (β: −19.37; 95% CI: −27.07, −11.67; p < 0.001) compared with the healthy controls (Figure 3). The negative effects of the coexisting diseases on the humoral response increased with the number of comorbidities (p for trend <0.01) (Figure 3).

3.4 Subgroup analysis

In the subgroup analysis (Table 2), although the estimated effect sizes did not meet statistical significance in some stratifications, a negative association between the presence of disease and seroconversion was consistently observed across the subgroups. The interaction test showed that there was no significant difference between each stratification in the association between disease and humoral responses, indicating that there was no significant dependence of age, sex, BMI, vaccine type, and interval time after vaccination on this negative association (all, p for interaction >0.05).

3.5 Correlation between NAbs and RBD-IgG levels

Correlation analysis revealed a strong correlation between NAbs and RBD-IgG titers in the healthy control and PWCD groups (Spearman correlation coefficients: 0.96 and 0.90, respectively; both, p < 0.001) (Figure 4A). In the PWCD group, the correlation coefficient of NAbs and RBD-IgG titers decreased from 0.92 at 21–30 days to 0.84 at 91–120 days, whereas it was stable in the healthy controls (correlation coefficients: 0.93–0.96). The regression relationship between NAbs and RBD-IgG titers exhibited time dependence (Figure 4B–E), suggesting that NAbs and RBD-IgG levels may have different kinetics in healthy controls and PWCD. These results partially explain differences in the effects of the presence of comorbidities on NAbs and RBD-IgG responses.