Evolution of anti-SARS-CoV-2 spike protein titers after two-dose of COVID-19 vaccination among people living with HIV

Abstract

Background

A community COVID-19 outbreak caused by the B.1.1.7 SARS-CoV-2 variant occurred in Taiwan in May 2021. High-risk populations such as people living with HIV (PLWH) were recommended to receive two doses of COVID-19 vaccines. While SARS-CoV-2 vaccines have demonstrated promising results in general population, real-world information on the serological responses remains limited among PLWH.

Methods

PLWH receiving the first dose of SARS-CoV-2 vaccine from 2020 to 2021 were enrolled. Determinations of anti-SARS-CoV-2 spike IgG titers were performed every one to three months, the third dose of the SARS-CoV-2 vaccine or confirmed SARS-CoV-2 infection. All serum samples were tested for anti-nucleocapsid antibody and those tested positive were excluded from analysis.

Results

A total of 1189 PLWH were enrolled: 829 (69.7%) receiving two doses of the AZD1222 vaccine, 232 (19.5%) of the mRNA-1273 vaccine, and 128 (10.8%) of the BNT162b2 vaccine. At all time-points, PLWH receiving two doses of mRNA vaccines had consistently higher antibody levels than those receiving the AZD1222 vaccine (p <0.001 for all time-point comparisons). Factors associated with failure to achieve an anti-spike IgG titer >141 BAU/mL within 12 weeks, included type 2 diabetes mellitus (DM) (adjusted odds ratio [aOR], 2.24; 95% CI, 1.25–4), a CD4 T cell count <200 cells/mm³ upon receipt of the first dose of vaccination (aOR, 3.43; 95% CI, 1.31–9) and two homologous AZD1222 vaccinations (aOR, 16.85; 95%CI, 10.13–28). For those receiving two doses of mRNA vaccines, factors associated with failure to achieve an anti-spike IgG titer >899 BAU/mL within 12 weeks were a CD4 T cell count <200 cells/mm³ on first-dose vaccination (aOR, 3.95; 95% CI, 1.08–14.42) and dual BNT162b2 vaccination (aOR, 4.21; 95% CI, 2.57–6.89).

Conclusions

Two doses of homologous mRNA vaccination achieved significantly higher serological responses than vaccination with AZD1222 among PLWH. Those with CD4 T cell counts <200 cells/mm³ and DM had consistently lower serological responses.

1. Introduction

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (COVID-19) had become pandemic since 2020 and caused a high morbidity and mortality, especially among immunocompromised patients.¹ Previous studies had demonstrated that people living with HIV (PLWH) of older age or with comorbidities such as diabetes mellitus (DM) or cardiovascular diseases might have poor outcomes when infected with SARS-CoV-2 compared with those without HIV infection.^2,3 Further studies indicated that a worse outcome was associated with a lower CD4 T cell count, even in PLWH receiving stable antiretroviral therapy (ART) with virological suppression.3, 4, 5, 6, 7 Therefore, current guidelines suggest that PWLH who have mild to moderate COVID-19 are eligible to receive antiviral agents or monoclonal antibody treatment to prevent severe or critical disease, especially those with advanced HIV infection, which is defined as having CD4 T cell counts below 200 cells/mm³ or previous AIDS-defining illnesses.⁸

Compared with other countries, the epidemic of COVID-19 in Taiwan was relatively mild before 2022. There were only a few indigenous cases of COVID-19 before May 2021 when under strict non-pharmaceutical interventions such as border control, contact tracing and isolation, social distancing and a nationwide campaign for using personal protective equipments.⁹ A community outbreak caused by the B.1.1.7 (Alpha) variant took place in May 2021, just after the implementation of the SARS-CoV-2 vaccination program. Overall, a total of 14,308 indigenous cases of COVID-19 were reported from May to August 2021, of those, only 192 (1.3%) PLWH were found to be infected by SARS-CoV-2.¹⁰ There was a consensus that all PLWH should receive two doses of COVID-19 vaccine, regardless of CD4 T cell counts or plasma HIV RNA load,⁸ though the safety and efficacy data of COVID-19 vaccine for PLWH was limited. Previous studies have demonstrated that certain vaccines, such as the influenza or hepatitis B (HBV) vaccine, induce suboptimal serological responses in PLWH.^11,12 Data from most clinical trials of COVID-19 vaccines conducted in PLWH was scarce, and safety and short-term immunogenicity remained limited^9,13, 14, 15, 16, 17, 18; moreover, the durability of antibody responses and the correlation between antibody titers and the risk of SARS-CoV-2 infection remained unclear. In this prospective, observational study, we aimed to investigate the serological responses in PLWH who had been on stable ART and had received two homologous doses of COVID-19 vaccines, including the AZD1222, mRNA-1273 and BNT162b2 vaccines.

2. Methods

2.1. Study population and setting

This observational study was conducted at the National Taiwan University Hospital (NTUH) to include PLWH aged 20 years or older who had been receiving HIV care as outpatients. Those receiving the first dose of SARS-CoV-2 vaccine between January 2020 and December 2021 were enrolled. A second dose of homologous vaccine, including the AZD1222, mRNA-1273 and BNT162b2 vaccine, was given according to the guidelines by Taiwan Centers of Disease Control, with at least an interval of 4–10 weeks between the two doses. Determinations of anti-SARS-CoV-2 spike IgG titers were performed every one to three months, depending on the clinical appointments made for HIV care, until the participants had received the third dose of SARS-CoV-2 vaccine, the diagnosis of COVID-19, loss to follow-up or death, whichever took place first. Participants who had a history of confirmed SARS-CoV-2 infection were excluded. All serum samples were also tested for anti-SARS-CoV-2 nucleocapsid antibody and those tested positive at baseline and during follow-up were excluded from the analysis. Medical records of the included PLWH were reviewed and the information on the demographic and clinical characteristics was reviewed, including age, body-mass index (BMI), date of vaccination, CD4 T cell count and plasma HIV RNA load (PVL) at vaccination, ART, and underlying diseases that might influence immune responses such as type 2 diabetes (DM), chronic kidney disease (CKD) of stage 3–5 (defined as an estimated glomerular filtration rate less than 60 mL/min/1.73m²), malignancies, autoimmune disease and viral hepatitis. The study was approved by the Research Ethics Committee of NTUH (NTUH 202106149RIND) and all participating PLWH gave written informed consent.

2.2. Laboratory investigations

All serum samples were stored and centrifuged at -20°C before testing. Anti-SARS-CoV-2 spike IgG in serum samples was determined using SARS-CoV-2 IgG II Quant assay (Abbott, Abbott Park, Illinois, U.S.A.) according to the manufacturer's instructions. An anti-SARS-CoV-2 spike IgG level higher than 50 arbitrary units per milliliter (AU/mL) was considered positive. The mathematical relationship of the Abbott AU/mL unit to WHO unit (binding antibody unit per mL [BAU/mL]) was as follows: BAU/mL = 0.142∗AU/mL.

In addition, anti-SARS-CoV-2 nucleocapsid antibody IgG was determined using Elecsys® Anti-SARS-CoV-2 assay (Roche, U.S.A), while an anti-SARS-CoV-2 nucleocapsid IgG level higher than 1.0 cutoff index (COI) was considered reactive.

2.3. Outcome assessment

Primary endpoints included serologic responses at weeks 1–24 after the second dose of SARS-CoV-2 vaccination. Acquisition of COVID-19 included symptomatic infection and asymptomatic infection; the history of symptomatic infection was retrieved from the National Notification System for Infectious Diseases, while asymptomatic infection was defined as a positive result of anti-SARS-CoV-2 nucleocapsid antibody in the absence of clinical symptoms. In order to speculate the potential vaccine effectiveness through antibody measurements, two cut-off values including 141 and 899 BAU/mL of anti-spike IgG were used. Dimeglio C et al. recently demonstrated that an anti-spike antibody titer greater than 141 BAU/mL correlated with the presence of neutralizing antibodies through the evaluation of 8758 vaccinated and unvaccinated healthcare workers.¹⁹ Prediction of clinical efficacy performed by Feng S et al. suggested that an anti-spike antibody titer of 899 BAU/mL predicted a 90% vaccine efficacy of AZD1222 vaccine against symptomatic infection by the B.1.1.7 variant.²⁰

2.4. Statistical analysis

Categorical variables, such as gender, underlying diseases and ART, were compared between different vaccination groups using Fisher's exact test and Pearson's chi-squared test. Continuous variables, such as age and laboratory results at vaccination were analyzed using Mann-Whitney U test. The geometric mean titers (GMTs) of SARS-CoV-2 anti-spike IgG were calculated in Ln-transformed data for statistics. Antibody titers in Ln form after receiving different vaccines were compared using Student's t-tests of all time-points. We applied the logistic regression model to estimate the adjusted odds ratios (aORs) for those with relatively low anti-spike IgG (<141 or <899 BAU/mL) within 12 weeks, and for those whose anti-spike IgG titers rapidly declined within 12–24 weeks after receiving the second dose of COVID-19 vaccine. A backward stepwise regression with removal threshold of p = 0.2 was used to select among covariates to be included into the multivariable model. A two-tailed p value less than 0.05 was considered statistically significant. All analyses were performed using Stata/SE software, Version 17.0 (https://www.stata.com).

3. Results

Between January 2020 and December 2021, 1252 PLWH who were followed at the HIV clinics of the NTUH and had received the first dose of COVID-19 vaccine were included into this study. After excluding those with confirmed COVID-19, positive anti-SARS-CoV-2 nucleocapsid antibody at baseline or during follow-up, and loss to follow-up, 1189 PLWH who completed the 2-dose vaccination schedule were included for further analysis, including 829 (69.7%) who had received two homologous doses of the AZD1222 vaccine, 232 (19.5%) of the mRNA-1273 vaccine and 128 (10.8%) of the BNT162b2 vaccine (Fig. 1).

Fig. 1.

Study population.

Table 1 shows the characteristics of included PLWH. There were mostly male (97.6%) with a median age of 40 years and 98.6% were virologically suppressed with ART with a median baseline CD4 counts of 632 cells/mm³. Overall, 1145 (96.3%) PLWH were receiving integrase strand-transfer inhibitor-based antiretroviral regimens before vaccination. Those who received two doses of mRNA vaccines tended to be older compared with those who received two doses of AZD1222 vaccines (41 vs 39 years, p < 0.001), while PLWH who received mRNA vaccines were more likely to have type 2 DM, CKD stage 3–5 and a previous history of lymphoma. The intervals between the two homologous doses of AZD1222 vaccination and mRNA vaccination were 13 (interquartile range, 13–14) and 13 (6–15) weeks, respectively.

Table 1.

Baseline characteristics of the included participants. Group 1 represents the participants receiving two doses of AZD1222 vaccines, while Group 2 were those who had two doses of mRNA vaccines, including mRNA-1273 and BNT162b2 vaccines.

	Total (N = 1189)	Group 1 (N = 829)	Group 2 (N = 360)	p value
Age, median (IQR), years	40 (33–48)	39 (33–46)	41 (35–50)	<0.001
Male gender, n (%)	1160 (97.6)	815 (98.3)	345 (95.8)	0.01
BMI >30 kg/m2	110 (9.3)	76 (9.2)	34 (9.4)	0.88
Interval between the two doses, median (IQR) weeks	13 (12–14)	13 (13–14)	13 (6–15)	0.002
HIV status
CD4, median (IQR), cells/mm3	632 (474–812)	641 (496–810)	612 (436–830)	0.11
CD8, median (IQR), cells/mm3	809 (624–1062)	798 (624–1057)	845 (618–1076)	0.58
CD4 <200 cells/mm3, n (%)	25 (2.1)	14 (1.7)	11 (3.1)	0.13
CD4 <350 cells/mm3, n (%)	125 (10.5)	74 (8.9)	51 (14.2)	0.007
CD4 <500 cells/mm3, n (%)	336 (28.3)	211 (25.5)	125 (34.7)	0.001
CD4/CD8 ratio, median (IQR)	0.77 (0.57–1.03)	0.77 (0.59–1.03)	0.78 (0.51–1.03)	0.18
Median PVL, (IQR), log copies/mL	0 (0–0)	0 (0–0)	0 (0–0)	0.6
PVL >20 copies/mL, n (%)	192 (16.2)	131 (15.8)	61 (16.9)	0.62
PVL >200 copies/mL, n (%)	17 (1.43)	11 (1.3)	6 (1.7)	0.65
cART, n (%)
INSTI-based	1145 (96.3)	794 (95.8)	351 (97.5)	0.15
NNRTI-based	42 (3.5)	35 (4.2)	7 (1.9)	0.051
PI-based	6 (0.5)	3 (0.4)	3 (0.8)	0.38
Comorbidities, n (%)
Type 2 DM	74 (6.22)	42 (5.1)	32 (8.9)	0.01
CKD stage III–V	49 (4.12)	24 (2.9)	25 (6.9)	0.001
Solid-organ cancer	25 (2.1)	14 (1.7)	11 (3.1)	0.13
Lymphoma	24 (2)	13 (1.6)	11 (3.1)	0.09
Autoimmune disease	15 (1.3)	11 (1.3)	4 (1.1)	1
Concurrent systemic steroid use	7 (0.6)	4 (0.5)	3 (0.8)	0.44
Chronic hepatitis B	138 (11.7)	97 (11.8)	41 (11.5)	0.88
Anti-HCV positivity	170 (14.3)	116 (14)	54 (15)	0.65

Abbreviations-: BMI, bodymass index; cART, combination antiretroviral therapy; CKD, chronic kidney disease; DM, diabetes mellitus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; PVL, plasma HIV RNA load.

Among the PLWH who had received two homologous doses of the AZD1222 vaccine and mRNA vaccines, anti-spike IgG titer reached their peak on the third and fourth weeks and subsequently declined. At all time-points, PLWH receiving two homologous doses of the mRNA vaccines had consistently higher antibody levels than those who had received two homologous doses of the AZD1222 vaccine (p < 0.001 for all time-point comparisons).

Of 1136 PLWH who underwent anti-spike IgG testing within 12 weeks after receiving the second dose of vaccines, 385 (33.9%) failed to achieve anti-spike IgG titers >141 BAU/mL, including 372 (46.7%) receiving the AZD1222 vaccine, 8 (3.6%) receiving the mRNA-1273 and 10 (8.5%) receiving the BNT162b2 vaccine (Fig. 2). Associated factors with failure to achieve anti-spike IgG titers >141 BAU/mL in the multivariate logistic regression model including PLWH with CD4 T cell counts <200 cells/mm³ before vaccination (aOR, 3.43; 95% confidence interval [CI], 1.31–9), type 2 DM (aOR, 2.24; 95% CI, 1.25–4), and those who had received two homologous doses of the AZD1222 vaccine (aOR, 16.85; 95% CI, 10.13–28) (Table 2) (Fig. 3). Moreover, 733 (72.9%) of 1005 PLWH who underwent anti-spike antibody testing 4–12 weeks after the second dose of vaccination failed to achieve anti-spike IgG titers >899 BAU/mL. PLWH receiving 2 consecutive doses of the AZD1222 vaccine were more likely to fail to achieve anti-spike IgG titers >899 BAU/mL compared with those receiving two homologous doses of the mRNA vaccines (OR, 45.91; 95% CI, 30.7–68.65) (data not shown).

Fig. 2.

Serologic responses after the second dose of COVID-19 vaccination at different follow-up intervals. (A). Evolution of anti-spike IgG of PWLH receiving two AZD1222 or mRNA vaccines. (B). Evolution of anti-spike IgG of PWLH receiving two AZD1222 vaccines. (C). Evolution of anti-spike IgG of PWLH receiving two mRNA vaccines. (D). Evolution of anti-spike IgG of PWLH receiving two mRNA-1273 (M) or BNT162b2 (B) vaccines. The number in the table below the X-axis in each figure represented the number of PLWH undergoing antibody testing and the GMT of anti-spike IgG in each period.

Table 2.

Factors associated with low titers of anti-spike antibody (<141 BAU/mL) in PLWH within 12 weeks after receiving the second dose of COVID-19 vaccine.

	Univariable		Multivariable
	HR (95% CI)	p value	aOR (95% CI)	p value
Age, per 1-year increase	1 (0.99–1.01)	0.71	–	–
Male gender	3.14 (1.08–9.13)	0.04	2.41 (0.77–7.59)	0.13
BMI >30 kg/m2	1 (0.66–1.53)	0.99	–	–
Type 2 DM	1.43 (0.88–2.32)	0.15	2.24 (1.25–4)	0.006
CKD stage III–V	0.94 (0.51–1.74)	0.85	–	–
Solid-organ cancer	0.64 (0.25–1.64)	0.36	–	–
Lymphoma	0.91 (0.37–2.25)	0.84	–	–
Autoimmune disease	0.71 (0.22–2.23)	0.55	–	–
Concurrent systemic steroid use	0.39 (0.05–3.34)	0.39	–	–
Chronic hepatitis B	1.39 (0.96–2.01)	0.08	1.4 (0.92–2.12)	0.12
Anti-HCV positivity	1.03 (0.73–1.46)	0.85	–	–
CD4 <200 cells/mm3	1.84 (0.86–3.95)	0.12	3.43 (1.31–9)	0.01
CD4 <350 cells/mm3	1.12 (0.75–1.66)	0.59	–	–
CD4 <500 cells/mm3	0.9 (0.69–1.19)	0.48	–	–
PVL >20 copies/mL	0.91 (0.64–1.27)	0.57	–	–
PVL >200 copies/mL	1.72 (0.62–4.78)	0.3	–	–
PVL >100,000 copies/mL	1.96 (0.27–13.94)	0.5	–	–
Two doses of AZD1222 vaccine	15.3 (9.33–25.1)	<0.001	16.85 (10.13–28)	<0.001
Interval between the two doses >8 weeks	4.89 (2.76–8.63)	<0.001	–	–

Abbreviations-: BMI, bodymass index; CKD, chronic kidney disease; DM, diabetes mellitus; HCV, hepatitis C virus; PVL, plasma HIV RNA load.

Fig. 3.

Factors associated with low anti-spike IgG response within the first 12 weeks after PLWH received two COVID-19 vaccines in the multivariate logistic regression model. (A) All PLWH with antibody titers failing to reach 141 BAU/mL. (B) PLWH with two AZD1222 vaccines, with antibody titers failing to reach 141 BAU/mL. (C) PLWH with two mRNA vaccines, with antibody titers failing to reach 899 BAU/mL.

Of the 796 participants who had anti-spike IgG titers determined within the first 12 weeks after receiving the second dose of the AZD1222 vaccine, 367 (46.1%) and 759 (95.4%) failed to achieve anti-spike IgG titers >141 and 899 BAU/mL, respectively. The major independent factor associated with failure to achieve anti-spike IgG titers >141 BAU/mL was type 2 DM (aOR, 2.66; 95% CI, 1.35–5.23) (Supplementary Table 1). In addition, among 340 PLWH receiving two homologous mRNA vaccines, 18 (5.3%) and 105 (30.9%) failed to achieve anti-spike IgG titers >141 and > 899 BAU/mL, respectively. In the multivariate analysis, factors associated with an anti-spike IgG <899 BAU/mL were CD4 T cell counts <200 cells/mm³ (aOR, 3.95; 95% CI, 1.08–14.42) and two doses of the BNT162b2 vaccine (aOR, 4.2; 95% CI, 2.57–6.89) (Supplementary Table 2) (Fig. 3).

Among the 228 PLWH who underwent anti-spike IgG testing within 12–24 weeks after receiving the second dose of vaccine, 111 (48.7%) had rapid declines of anti-spike IgG titer, which was defined as a decline of anti-spike IgG titers to <141 BAU/mL, including 108 (70.6%) PLWH who had received the AZD1222 vaccine, 1 (1.8%) who had received the mRNA-1273, and 2 (10%) who had received the BNT162b2 vaccine (Fig. 2). Receipt of two homologous doses of the AZD1222 vaccine was the major factor associated with the rapid declines of the anti-spike antibody (aOR, 57.62; 95% CI, 16.9–196.6) (Supplementary Table 3). Furthermore, 151 (98.7%) PLWH who had received two doses of the AZD-122 vaccine, 42 (76.4%) the mRNA-1273 and 19 (95%) the BNT162b2 vaccine had a decline of anti-spike IgG titer to <899 BAU/mL 12–24 weeks after the second dose of vaccination.

4. Discussion

In this study investigating the serologic responses of COVID-19 vaccination among PLWH, we have shown the longitudinal follow-up of antibody responses in PLWH who had been mostly well-controlled with ART and had received different types of a 2-dose homologous vaccine in the real-word setting. PLWH who was immunized with two homologous mRNA vaccines were more likely to achieve higher and more sustained antibody responses compared with those who had received a 2-dose homologous AZD1222 vaccine.

To date, several studies have demonstrated that well-controlled PLWH have similar serological responses to those without HIV infection within 2–4 weeks after the second dose of AZD1222 or BNT162b2 vaccine.^12,16 However, the titers of anti-spike IgG in our cohort were slightly lower when compared with those observed in the above mentionned studies .^12,16 In a retrospective study of 100 PLWH and 152 matched HIV-negative control participants, HIV was not significantly associated with the magnitude of any humoral response. However, two homologous doses of AZD1222 vaccination was significantly associated with lower antibody responses.²¹ When compared to the antibody responses in a study that assessed the evolution of antibody titers in young, HIV-negative participants of similar age who had received homologous or heterologous SARS-CoV-2 vaccines,²² we found lower antibody responses in our study consisting of only PLWH. In addition to HIV infection, the types of vaccines used, and vaccination strategies (homologous vs heterologous), the discrepancy may be attributed to the intervals between the two doses of vaccines. The PLWH in our study mostly received the second doses of vaccines after a longer interval (>12 weeks) due to a limited availability of vaccines in Taiwan in 2021 compared with those of other studies in which participants had received the second dose after 4–8 weeks.^12,16,22 In our study, we found that an interval over 8 weeks between the two doses of vaccination had a 4.9-fold higher risk associated with lower antibody responses. However, such correlation was not found in the subgroup analysis in PLWH receiving the AZD1222 or mRNA vaccine, respectively. Previous meta-analysis by Voysey M et al. has demonstrated that an interval of 8–12 weeks for AZD1222 vaccine led to a higher immunogenicity while an extended interval of 14–17 weeks showed a similar trend in antibody responses.^23,24 Nevertheless, there was scarce data to support this finding among PLWH. In addition, the storage of serum samples at -20°C before testing might also have contributed to the lower antibody responses observed in our study, although previous studies have demonstrated that multiple freeze-thaw cycles had no effect on the ability of the ELISA assay to detect SARS-CoV-2 IgG antibodies.^25,26

Similarly to the findings in studies among healthy adults, we have demonstrated a waning of anti-spike IgG.^27,28 Lapointe HR et al. pointed out that virologically suppressed PLWH showed typical antibody durability after two doses of COVID-19 vaccination, regardless of the types of vaccine administered.¹⁴ However, in the context of new variants and evidence of correlation between anti-spike IgG and neutralizing ability, investigations to examine the clinical efficacy of a higher standard of antibody titer is warranted. Seventy percent of PLWH who had received the AZD1222 vaccine in this study failed to achieve sustained antibody responses by maintaining anti-spike IgG titers >141 BAU/mL, an indicator of presence of neutralizing antibodies. Given that the peaks of anti-spike IgG titers might be lower in PLWH compared with those without HIV infection, the durability of maintaining neutralizing antibodies might be overestimated. Therefore, our findings support that additional doses of booster vaccination are needed in PLWH during the ongoing SARS-CoV pandemic.

Even with only a small number of them in our cohort, PLWH with a CD4 T cell count <200 cells/mm³ were shown to be associated with poor serological responses to COVID-19 vaccination. PLWH with CD4 T cell counts <200 cells/mm³ in this study continued to achieve inferior antibody responses, even if they had received two homologous doses of the mRNA vaccines. Such findings were consistent with those of previous studies.29, 30, 31, 32, 33 Moreover, our study illustrated that, in addition to CD4 T cell counts and vaccine type, traditional risk factors including CKD stage 3–5 and type 2 DM were associated with poor serological responses.34, 35, 36 As a result, PLWH with comorbidities that may affect immune responses should be prioritized for a second booster dose of SARS-CoV-2 vaccine.

The strengths of this study include a relatively large number of PLWH and a longer observational period for the evaluation of durability of antibody responses. However, the observational nature of this study results in several limitations. First, the anti-spike IgG measurements were performed at intervals according to clinical care practices or convenient time-points for PLWH but not at fixed time-points; while trends were similar to those observed in other studies, serological responses should be interpreted with caution. Second, only a few PLWH underwent antibody testing 12–24 weeks after the second dose of vaccination, so the power to understand the extent to which the serological responses waned might be dampened. Third, the low-level viral transmission rate in Taiwan at the time of the study was conducted hindered the evaluation of vaccine effectiveness. Nevertheless, our previous study during the early phase of implementation of the COVID-19 vaccination program in Taiwan had demonstrated that vaccination was still clinically effective among PLWH during the outbreak setting where non-pharmaceutical interventions were strictly implemented.^9,10 Fourth, no HIV-negative participants were included as a control group, which might preclude us from understanding if virologically-suppressed PLWH with improved immunity might mount similar immune responses to COVID-19 vaccination to HIV-negative people of the same age groups. Last, we did not assess the neutralizing antibody or T cell-related immunity. Instead, two cut-off values derived from an immune-bridging model were used in our study. The performance of using these two cut-off values in predicting clinical vaccine effectiveness warrants more investigations among PLWH.

In conclusion, we found that PLWH receiving two homologous SARS-CoV-2 vaccine mounted antibody responses with similar trends of antibody responses over time to those observed in clinical trial settings. Two doses of homologous mRNA vaccination had significantly higher immunogenicity than AZD1222 among PLWH. PLWH with CD4 T cell counts less than 200 cells/mm³ had consistently lower antibody responses to either mRNA or non-mRNA vaccination. Traditional risk factor such as DM also contributed to lower anti-spike IgG titers.

Author contributions

Wang-Da Liu – manuscript drafting, data analysis.

Man Wai Pang – laboratory assistant for antibody test.

Sui-Yuan Chang – laboratory chief, handling the antibody measurement.

Jann-Tay Wang – statistics, data analysis.

Hsin-Yun Sun – participants enrolling.

Yu-Shan Huang - participants enrolling.

Kuan-Yin Lin - participants enrolling;

Un-In Wu - participants enrolling;

Guei-Chi Li – data collection.

Wen-Chun Liu - participants enrolling, laboratory assistant.

Yi-Ching Su - laboratory assistant.

Pu-Chi He - participants enrolling, data collection.

Chia-Yi Lin – data collection.

Chih-Yu Yeh – data collection.

Yu-Chen Cheng - laboratory assistant for antibody test.

Yi Yao - laboratory assistant for antibody test.

Yi-Ting Chen - participants enrolling.

Yu-Zhen Luo - participants enrolling.

Pei-Ying Wu - participants enrolling.

Ling-Ya Chen - participants enrolling.

Hsi-Yen Chang - participants enrolling.

Wang-Huei Sheng - participants enrolling;

Szu-Min Hsieh - participants enrolling.

Chien-Ching Hung – concept of the study, manuscript drafting.

Shan-Chwen Chang – concept of the study.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: