COVID-19 vaccination in systemic lupus erythematosus: a systematic review of its effectiveness, immunogenicity, flares and acceptance

Abstract

Objectives

COVID-19 infection is associated with significant morbidity in systemic lupus erythematosus but is potentially preventable by vaccination, although the impact of the myriad vaccines among SLE patients is not established. We aimed to assess the effectiveness, efficacy, acceptance and safety of COVID-19 vaccination in SLE.

Methods

We performed a systematic review of PubMed, EMBASE, CENTRAL, and Scopus publications until 8 June 2022 without language, publication year or publication status restrictions. Reports with fewer than 5 patients or incomplete information on study outcomes were excluded. Risk of bias was assessed, and results reported according to the PRISMA 2020 guidelines.

Results

We identified 32 studies (34 reports) comprising 8269 individuals with SLE. Post-vaccine COVID-19 infections ranged from 0 to 17% in 6 studies (5065 patients), while humoral and cellular immunogenicity was evaluated in 17 studies (976 patients) and 5 studies (112 patients), respectively. The pooled seropositivity rate was 81.1% (95% CI: 72.6, 88.5%, I² = 85%, P < 0.01), with significant heterogeneity and higher rates for mRNA vaccines compared with non-mRNA vaccines. Adverse events and specifically lupus flares were examined in 20 studies (3853 patients) and 13 studies (2989 patients), respectively. Severe adverse events and moderate to severe lupus flares were infrequent. The pooled vaccine acceptance rate was 67.0% (95% CI: 45.2, 85.6%, I²=98%, P < 0.01) from 8 studies (1348 patients), with greater acceptance in older patients.

Conclusion

Among SLE patients, post-vaccine COVID-19 infections, severe flares, and adverse events were infrequent, while pooled seropositivity and acceptance were high, with significant heterogeneity. These results may inform shared decision-making on vaccination during the ongoing COVID-19 pandemic.

Trial registration

PROSPERO, https://www.crd.york.ac.uk/PROSPERO/, CRD42021233366.

COVID-19 vaccine, SARS-CoV-2, vaccination, safety, efficacy

Rheumatology key messages

• COVID-19 vaccination was efficacious in SLE patients on various immunosuppressive therapies.

• Severe adverse events and lupus flares occurred infrequently, demonstrating vaccine safety in SLE patients.

• COVID-19 vaccine acceptance was high among SLE patients, with greater acceptance in older patients.

Introduction

SLE is a chronic relapsing disease with potentially debilitating symptoms and end-organ damage and thus often requires immunosuppressive therapy to achieve disease remission. Coronavirus disease (COVID-19) infection was associated with worse prognosis in individuals with active inflammatory disease [1], and in immunosuppression [2], thus the EULAR (in December 2020) and the ACR (in February 2021) recommended COVID vaccination in SLE empirically, while acknowledging the lack of data on its effectiveness [3]. However, concerns of suboptimal efficacy during immunosuppressive treatment and/or possibility of SLE flare after immunization had led to suboptimal coverage of even established vaccines, such as influenza vaccines [4]. The current literature regarding COVID vaccine efficacy and safety in SLE has largely been limited to small or single-centre studies, which has limited its power to detect significant associations with patient, disease or treatment factors. An early review intended to study vaccine efficacy in SLE reported 11 studies of auto-immune inflammatory rheumatic diseases that included SLE, but data specific to SLE patients was limited, and the pooled analyses were based on highly heterogeneous diseases, including non-SLE conditions [5]. Other systematic reviews of COVID-19 vaccination in rheumatic diseases or immune-mediated inflammatory diseases were dominated by inflammatory arthritis or gastrointestinal disease [6–8], included only mRNA vaccines [1], and evaluated booster vaccination [8], or only vaccine immunogenicity and/or safety [1, 7–9]. Yet, data on utilization and vaccine hesitancy can highlight deficiencies in vaccine coverage, and indicate patients' concerns that need to be addressed to improve vaccination rates among at-risk individuals with SLE. Thus, we performed a systematic review on the efficacy, effectiveness, acceptance or utilization, and safety of COVID-19 vaccines in individuals with SLE on various immunosuppressive therapies, to facilitate informed decision-making about vaccination.

Methods

Search strategy

A systematic search was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [10] in PubMed (from 1979), EMBASE (from 1981), Cochrane Central Register of Controlled Trials (CENTRAL), and Scopus. The databases were searched up to 8 June 2022 without language, publication year, or publication status restrictions.

The search strings included Boolean terms ‘AND’, ‘OR’ and ‘NOT’ with the following keywords and their respective variants or derivates in any relevant combination: ‘systemic lupus erythematosus’ (‘SLE’), ‘lupus nephritis’, ‘vaccine’, ‘coronavirus’ and ‘COVID’ (Supplementary Table S1, available at Rheumatology online). Two authors (S.Y.S.T. and A.M.Y.) independently selected full studies for evaluation by analysing titles and abstracts from the databases searched. We retrieved the full text of potentially relevant reports selected for evaluation and linked multiple reports of the same study together. A hand search of the references was performed to identify relevant articles. Disagreement about selection of a study was resolved by discussion. We contacted authors of included papers and abstracts to request additional information where required, such as for abstracts or when SLE patients comprised a subgroup of the study cohort.

Study selection

Articles were selected for systematic review if they included individuals with LN or SLE who were eligible for or received the COVID-19 vaccine, with or without comparison with placebo, active or no treatment and were assessed for the outcomes of interest. We included randomized controlled trials (RCTs) and quasi-RCTs, both prospective and retrospective cohort studies, as well as cross-sectional studies. We included both published and non-published studies. For multiple studies that included overlapping patient data, the largest study was included, and the others were excluded, or data was combined. We excluded case reports and case series with fewer than five patients, as recommended by the Cochrane Statistical Methods Group [11]. Reports available as abstracts and preprints were not excluded. For studies that included individuals with SLE but did not report SLE-subgroup data, the study authors were contacted for SLE-specific data. Reports with incomplete information regarding the patient population, vaccine, and study outcomes and whose authors did not respond were also excluded, since the study eligibility could not be ascertained.

Outcomes and data management

The primary outcomes of interest were the (1) effectiveness and efficacy, and (2) acceptance and/or utilization of COVID-19 vaccination in SLE. Vaccine effectiveness was evaluated according to occurrence of clinical events related to COVID-19 infection [6, 12], while vaccine efficacy was defined by humoral and cellular immunogenicity. The other outcomes of interest were (1) lupus flares and (2) other adverse events.

Relevant data (such as study design and setting; participant, disease and treatment characteristics; vaccination type and doses; detailed definitions; and nature of outcomes) were extracted by two authors (S.Y.S.T. and A.M.Y.) using a structured form. Risk of bias in non-randomized studies was assessed using the Risk Of Bias In Non-randomized Studies—Interventions (ROBINS-I) tool as low, moderate, serious or critical risk in seven domains (bias due to: confounding; participant selection; classification of interventions; deviation from intended interventions; missing data; outcome measurement; and selective reporting) [13]. A study was assessed to have low risk of bias if it achieved low risk in all domains for each outcome; otherwise, it was considered to have high risk of bias.

Statistical analysis

Statistical analyses were performed using R software (version 3.6.3) with the packages meta and dmetar [14]. Where means (s.d.) of continuous variables were not available, they were estimated from the medians and minimum, maximum or interquartile ranges using the equations proposed by Wan et al. [15]. Weighted pooled means and standard deviations were calculated based on the sample size. Pooled proportions were calculated with the inverse variance method using the Freeman–Tukey double arcsine transformation [16]. Meta-analysis for the effect of vaccination on the outcomes of interest was performed if there were at least three studies with similar study design. Unadjusted estimates for dichotomous outcomes were analysed using the Mantel Hanszel method with random effects analysis, as the cohorts were clinically heterogeneous in age, ethnicity, and immunosuppressive therapy [17]. We evaluated heterogeneity by visual inspection of the forest plots, the Chi² test and the I² statistic. I²<30% suggested mild heterogeneity, and I²>50% suggested substantial heterogeneity [18], with a Chi²P-value of <0.1 being considered as significant for heterogeneity. Subgroup difference was statistically significant if the P-value was <0.1 for the test for subgroup difference. Univariable study-level meta-regressions were shown as bubble plots to explore potential sources of heterogeneity or prognostically relevant study-level covariates [19]. Funnel plots and Egger’s test were used to document the potential presence of publication bias, where asymmetrical distribution of studies was considered suggestive of bias [20].

Results

Literature review

Fig. 1 shows the study flow as recommended by the PRISMA 2020 updated guideline [21]. Our search yielded 411 unique articles. After screening the title and abstract, we reviewed the full text of 103 reports. Five studies were associated with multiple reports; hence, reports of different outcomes arising from the same study were combined [22–25], or only the report with the largest sample size and/or more outcomes of interest was included [26–28]. For reports that included SLE among inflammatory or rheumatic conditions but did not report the SLE-subgroup data, the authors of 14 reports provided further information via personal communication [22, 23, 25, 27, 29–38].

Figure 1.

Study flow diagram

Table 1 details the 32 studies (34 reports) comprising 8269 individuals with SLE, while Supplementary Table S2 (available at Rheumatology online) summarizes the study characteristics and patient’ characteristics. Most studies were conducted in Europe, North America, and Asia. The participants were predominantly female. Only 1 study evaluated adolescents younger than 18 years [39]. The pooled mean age was 49.0 (5.6) years (7694 patients, 28 studies), mean disease duration was 11.9 (5.9) years (1530 patients; 13 studies) and mean SLEDAI score was 2.9 (1.7) (721 patients; 10 studies). Supplementary Table S2, available at Rheumatology online, shows that comorbidities such as renal involvement or LN, hypertension, and use of glucocorticoid and other immunosuppressants were prevalent. A total of 7383 (89.3%) patients with SLE received one or more doses of COVID-19 vaccines from Pfizer (2124 patients; 19 studies), Moderna (443 patients; 10 studies), Oxford-AstraZeneca (305 patients; 8 studies), Johnson and Johnson (37 patients; 5 studies), and Sinovac or Sinopharm (734 patients; 8 studies). Vaccine type was not specified for 3728 patients (6 studies) and a minority received Cansino, CureVac, Sputnik V, and Covaxin vaccines.

Table 1.

Characteristics of included studies of COVID-19 vaccination in individuals with SLE

Study	Country	Study design	Study month/year	SLE, N	Age, years		Female, %	Disease			Immunosuppressantse	Vaccine typef	Outcomes evaluated
					Mean	s.d.		Activityc	Duration, years				Efficacy/Effectiveness	Safety	Acceptance/Utilization
									Mean	s.d.
Barbhaiya et al. 2021 [55]	USA	SS CS	Mar 2021	136	NR	NR	NR	NR	NR	NR	NR	mRNA 95% viral vector 3% NR 2%	Yes	Yes	Yes
Moyon et al. 2022 [46]	France	SS PC	NR	126	46.3d	18.6d	90.5	2 (3)d Remission 80.9%	1.2d	1.0d	GC 55% HCQ 84% MMF 19% AZA 4% MTX 16% BeliAb 12%	mRNA 100%	Yes	Yes	Yes
Izmirly et al. 2022 [47]	USA	SS PC	NR	90	45.5	14.2	87.8	3.2 (3.8)	NR	NR	GC 29% HCQ 79% MMF 23% AZA 6% MTX 9% TAC 6% BeliAb 11% RTX 3%	mRNA 94% viral vector 6%	Yes	Yes	No
Ammitzbøll et al. 2021 [24]	Denmark	SS PC	Dec 2020 – Apr 2021	61	57.9d	15.6d	67.2	SLICC/ACR median 1 (IQR 1–2)	NR	NR	GC 47% HCQ 61% MMF 26% AZA 25% MTX 10% BeliAb 5% RTX 3%	mRNA 100%	Yes	No	No
Bartels et al. 2021 [25]				128	52.7	16	88.3	SLICC/ACR median 1 (IQR 0–2)	16.2d	15d	GC 38% HCQ 69% MMF 17% AZA 21% MTX 9% RTX 4% BeliAb 4%	mRNA 100%	No	Yes	No
Felten et al. 2021 [42]	Global	MC CS	Mar – May 2021	696	42.3d	12.6d	96.1	Remission 92.9%	NR	NR	GC 54% HCQ 78% MMF/AZA/MTX/CYC 50% BeliAb 11% RTX 3%	mRNA 66% viral vector 11% inactivated viral 23%	Yes	Yes	No
Boyarsky et al. 2021 [26]	USA	MC PC	Jan – Feb 2021	24	NR	NR	NR	NR	NR	NR	NR	mRNA 100%	Yes	No	No
Spiera et al. 2021 [48]	USA	SS RC	Feb – Apr 2021	9	NR	NR	NR	NR	NR	NR	NR	mRNA 100%	Yes	No	No
Wang et al. 2022 [52]	China	MC PC	Aug –Sep 2021	60	40.1	12.5	96.7	2.3 (3.7)	5.0d	4.6d	GC 92% HCQ 93% MTX 13% CSA 18%	Inactivated viral 100%	Yes	Yes	No
Tang et al. 2022 [53]	China	SS CS	NR	188	39.7	29.1	91.5	Remission 94.7%	7.4d	3.8d	GC 53% HCQ 89% MMF 12% AZA 2% MTX 11% CSA 11%	Inactivated viral 100%	No	Yes	Yes
Heshin-Bekenstein et al. 2022 [39]	Israel, Slovenia	MC PC	Apr – Nov 2021	13	NR	NR	NR	3.3 (NR)	NR	NR	GC 54% HCQ 92% MMF 46% AntiCD20 8%	mRNA 100%	Yes	Yes	No
So et al. 2022 [49]	Hong Kong	SS PC	Jun–Aug 2021	65	46.2	10.4	93.8	2.9 (2.0) Remission 100%	NR	NR	GC 75% MMF 26% AZA 12% CNI 19%	mRNA 59% Inactivated viral 41%	Yes	Yes	No
Tien et al. 2022 [50]	Taiwan	SS PC	NR	36	44.6	14.3	97.2	NR	NR	NR	GC 94% HCQ 94% MMF 42% AZA 39% CSA 14% RTX 3%	mRNA 61% viral vector 39%	Yes	No	No
Yuki et al. 2022 [28]	Brazil	SS NRCT	Feb–Apr 2021	232	43.7d	41d	89.7	7.0 (14.3)d	22.7d	39.5d	GC 54% HCQ 81% MMF 31% AZA 27% MTX 11% CYC 3% CNI 5% BeliAb 14%	Inactivated viral 100%	Yes	Yes	No
Ferri et al. 2021 [51]	Italy	MC PC	Feb – Jul 2021	38	46	17	94.7	NR	14	12	GC 82% HCQ 66% MMF 11% AZA 11% MTX 8% BeliAb 37%	mRNA 100%	Yes	Yes	No
Zavala-Flores et al. 2022 [56]	Peru	SS PC	May – Jun 2021	100	37.2d	16.2d	94.0	NR	8.2d	6.8d	GC 71% HCQ 75% MMF 23% AZA 24% MTX 8% CYC 5%	mRNA 100%	No	Yes	No
Gerosa et al. 2022 [41]	Italy	MC RC	Dec 2020 – Oct 2021	452	46.3d	15.6d	92.3	Remission 88.9%	11.1d	7.7d	GC 71% HCQ 83% CYC 25% RTX 13%	mRNA 98% viral vector 2%	Yes	Yes	No
Larsen et al. 2021a [45]	Denmark	SS PC	NR	90	52.7d	16.6d	87.8	2.7 (4.5)d	14.7d	14.3d	GC 27% HCQ 60% Other IS 43%	mRNA 98% viral vector 2%	Yes	No	No
Petri et al. 2021a [69]	USA	SS RC	May 2021	228	NR	NR	NR	3.2 (2.7)	NR	NR	NR	mRNA 96% viral vector 4%	No	Yes	No
Sadun et al. 2021a [54]	USA	MC CS	Feb 2021	63	52.8	13.4	98.4	NR	17.1	9.6	NR	NR	No	No	Yes
Singh et al. 2021a [43]	USA	MC RC	Jan 2020 – Sep 2021	3601	54.0d	18.5d	87.7	NR	NR	NR	NR	NR	Yes	No	No
Liew et al. 2022 [29]	Singapore	SS CS	Jun–Jul 2021	18	43.7d	20.8d	83.3	NR	4.8d	3.2d	NR	NR	No	No	Yes
Alsaed et al. 2022b[30]	Qatar	MC RC	Jun 2021	5	41.52	5.26	100.0	NR	NR	NR	GC 40% HCQ 40% MMF 20% MTX 20% RTX 80%	mRNA 100%	Yes	No	No
Li et al. 2022 [31]	Hong Kong	MC CS	Jun–Oct 2021	384	NR	NR	NR	NR	NR	NR	NR	NR	No	Yes	Yes
Guaracha et al. 2022b [32]	Mexico	MC CS	Mar–Sep 2021	400	39.7d	14.1d	92.0	RDCI median 0 (IQR 0–1) Remission 64.5%	NR	NR	NR	NR	No	No	No
Rider et al. 2022b [27]	Global	MC CS	Apr–Aug 2021	763	48.1	12.8	96.6	PGA 4.2 (2.4)	NR	NR	GC 40% HCQ 73% MMF 20% MTX 14% RTX 5%	mRNA 74% viral vector 20% inactivated viral 6%	No	Yes	No
Fornaro et al. 2022b [33]	Italy	SS PC	NR	62	46.5	15.5	87.1	2 (3)	NR	NR	GC 77% HCQ 79% MMF 26% AZA 26% MTX 11% RTX 3% BeliAb 15%	mRNA 100%	No	Yes	No
Gaur et al. 2021 [34]	India	SS CS	Mar–Apr 2021	14	32.14	10.4	92.9	Remission 71.4%	5.1	3.3	NR	Viral vector 100%	No	Yes	Yes
Goulenok et al. 2022 [35]	France	SS PC	Dec 2020 – Mar 2021	101	48.7d	42.1d	87.1	NR	NR	NR	NR	mRNA 100%	No	No	Yes
Shenoy et al. 2021b [36]	India	SS PC	NR	49	53.0	11.3	83.7	NR	NR	NR	GC 27% HCQ 63% MMF 20% AZA 6% MTX 41% TAC 12% RTX 20%	Viral vector 80% Inactivated viral 20%	Yes	No	No
Assawasaksakul et al. 2022a [22]	Thailand	SS PC	NR	71	39	11.9	95.8	3.1 (2.7)	12.6	8.6	GC 89% AZA 14% HCQ 70% MMF 69% MTX 6% CNI 7%	Inactivated viral 38% Viral vector 37% Both inactivated viral and viral vector 25%	Yes	Yes	No
Hadjadj et al. 2022b [37]	France	SS PC	Jan–Apr 2021	16	36.2	12.4	100.0	Remission 37.5%	12.5	11.6	MTX 25% Other IS 56%	mRNA 100%	Yes	Yes	No
Szebeni et al. 2022 [38]	Hungary	SS PC	May–Sep 2021	11	52	14.1	90.9	NR	13	9.4	GC 36% HCQ 36% MTX 9% BeliAb 18%	mRNA 64% viral vector 36%	Yes	No	No

a

Report available only as abstract.

b

Data of SLE patients as communicated by study authors via email may differ from original report that combined SLE patients with other immune-mediated or rheumatic diseases.

c

Disease activity at enrolment or before vaccination according to the SLEDAI, where available, and reported as mean (s.d.) unless otherwise stated; remission was defined according to study or in absence of ‘active disease’ or ‘lupus flare’ at time of enrolment or vaccination.

d

Where means (s.d.) of continuous variables were not available, they were estimated from the medians and minimum, maximum or interquartile ranges.

e

Immunosuppressants most frequently used among patients with SLE in the study.

f

Vaccine types: mRNA (Pfizer, Moderna, CureVac), viral vector (Oxford-AstraZeneca, Johnson & Johnson, Sputnik V, CanSino), inactivated viral (Sinovac, Sinopharm, Covaxin). SS: single-centre; MC: multicentre; CS: cross-sectional; PC: prospective cohort; RC: retrospective cohort; NRCT: non-randomized controlled trial; NR: not reported; N: number of patients; RDCI: Rheumatic Disease Comorbidity Index; ACR DI: ACR Damage Index; PGA: Patient Global Assessment of Disease Activity; GC: glucocorticoid; BeliAb: belimumab; CNI: calcineurin inhibitor; HCQ: HCQ or other anti-malarials; IS: immunosuppressants; RTX: rituximab; TAC: tacrolimus.

Clinical effectiveness, defined by COVID-19 infection, was assessed in 6 studies (5065 patients), while vaccine efficacy, defined by humoral and cellular immunogenicity, was evaluated in 17 studies (976 patients) and 5 studies (112 patients), respectively. Adverse events and specifically lupus flares were examined in 20 studies (3853 patients) and 13 studies (2989 patients), respectively. Additionally, 8 studies (1348 patients) evaluated vaccine acceptance and/or utilization. Most studies had moderate risk of bias or no information (Supplementary Fig. S1, available at Rheumatology online) [40], largely due to incomplete information regarding participant selection.

Clinical effectiveness and efficacy

COVID-19 infections occurred in 461 (9%) of 5065 patients who received at least 1 vaccine dose from various countries such as Israel, Slovenia, Brazil, France, Italy, the USA and Thailand (Table 2) [28, 39, 41–43]. Most studies reported low infection rates of 0–6% [28, 39, 42, 43], while only one study reported a much higher infection rate of 17% [41]. The follow-up duration after full vaccination or booster ranged from 14 to 90 days [28, 39, 43, 44], but was not specified in other studies [41, 42]. Data was not pooled due to heterogeneity in the study design.

Table 2.

Clinical infections, immunogenicity, and safety after COVID-19 vaccination

Study/Reports	Vaccine dose or regimen	Follow-up, days	COVID-19 infection, n/N (%)	Immunogenicity		Adverse events		Flares
				Humoral, n/N (%)	Cellular, n/N (%)	All, n/N (%)	Severe, n/N (%)	All, n/N (%)	Severe, n/N (%)
Barbhaiya et al. 2021 [55]	First	7	–	–	–	83/136 (61%)	0/136 (0%)	8/136 (6%)	1/136 (1%)
	Second	7	–			96/136 (71%)	0/136 (0%)	4/136 (3%)	0/136 (0%)
Moyon et al. 2021 [46]	First	14	–	–	–	97/126 (77%)	0/126 (0%)	–	–
		28	–	–	–	46/126 (37%)	0/126 (0%)	–	–
	Second	14–15	–	NR/126 anti-S Ab positive NR/126 anti-RBD Ab positive	18/40 (45%) detectable SARS-CoV-2–specific T cell response	70/126 (57%)	0/126 (0%)	3/126 (2%)	0/126 (0%)
Izmirly et al. 2022 [47]	Fulla	14	–	64/90 (71%) anti-RBD Ab positive (>100 U/ml)	NR/16	–	–	–	–
		90–120	–	–	–	–	–	9/79 (11%)	1/79 (1%)
Ammitzbøll et al. 2021 [24]	Second	7	–	54/61 (89%) anti-S1 Ab positive	–	–	–	–	–
Bartels et al. 2021 [25]	Second	7	–	–	–	125/128 (98%) 85% local 85% systemic	NR	–	–
Felten et al. 2021 [42]	First	NR	0/696 (0%)	–	–	316/696 (45%)	451/696 (17%)	21/696 (3%)	15/696 (2%)
	Second	NR	0/343 (0%)	–	–	181/343 (53%)
Boyarsky et al. 2021 [26]	First	NR	–	16/24 (67%) anti-RBD Ab positive	–	–	–	–	–
Spiera et al. 2021 [48]	NR	NR	–	7/9 (78%) anti-RBD Ab positive	–	–	–	–	–
Wang et al. 2022 [52]	Second	>14	–	30/60 (50%) anti-CoV Ab positive	–	15/60 (25%)	–	–	–
Tang et al. 2022 [53]	First	42	–	–	–	82/188 (44%)	0/188 (0%)	0/188 0)	–
	Second	42	–	–	–	23/116 (20%)	0/116 (0%)	1/116 (1%)	–
Heshin-Bekenstein et al. 2022 [39]	First	14	–	–	–	NR	0/13 (0%)	–	–
	Second	14–28	–	–	–	NR	1/13 (8%)	–	–
		14–63	0/10 (0%)	10/10 (100%) anti-S1/S2 Ab positive (>15 BAU)	–	–	–	–	–
		60–90	0/10 (0%)	–	–	–	–	0/10 (0%)
So et al. 2022 [49]	First	NR	–	–	–	53/65 (82%) 71% local 66% systemic	0/65 (0%)	–	0/65 (0%)
	Second	28	–	60/65 (92%) anti-RBD NAb positive (inhibition >30%)	–	51/65 (79%) 66% local 63% systemic	0/65 (0%)	SLEDAI-2k NSC	0/65 (0%)
Tien et al. 2022 [50]	First	21–28	–	16/36 (44%) anti-RBD Ab positive (>10 AU/ml)	–	–	–
	Second	21–28	–	29/36 (81%) anti-RBD Ab positive	–	NR	1/36 (3%)	NR	NR
Yuki et al. 2022 [28]	First	28	–	57/215 (27%) anti-S1/S2 Ab positive (≥15 AU/ml) 58/213 (27%) anti-RBD NAb positive (inhibition ≥30%)	–	132/223 (59%) 32% local 49% systemic	0/223 (0%)	–	–
	Second	42	–	151/215 (70%) anti-S1/S2 Ab positive 131/213 (62%) anti-RBD NAb positive		115/223 (47%) 28% local 37% systemic	0/223 (0%)	–	–
		52	9/232 (4%)	–	–	–	–	–	–
		90	–	–	–	–	–	11/223 (5%)	NR
Ferri et al. 2021 [51]	Second	14	–	33/38 (87%) anti-RBD Ab positive (>7 BAU/ml)	–	17/38 (45%)	–	2/38 (5%)	0/38 (0%)
Zavala-Flores et al. 2021 [56]	First	10	–	–	–	90/100 (90%)	NR	9/100 (9%)	2/90 (2%)
	Second	10	–	–	–	83/90 (92%)	NR	18/90 (20%)	0/90 (0%)
Gerosa et al. 2022 [41]	First	NR	–	–	–	56/452 (12%)	NR	–	–
	Second	NR	77/452 (17%)	–	–	94/452 (21%)	NR	19/452 (4%)	4/452 (1%)
Larsen et al. 2021b [45]	First	21		31/55 (56%) anti-S Ab positive (>8.7 IU/ml)	–	–	–	–	–
	Second	28–56	–	84/90 (93%) anti-S Ab positive	–	–	–	–	–
Petri et al. 2021b [69]	NR	NR	–	–	–	–	–	SLEDAI NSC	–
Singh et al. 2021b [43]	Partial	≥14	203/3601 (6%)	–	–	–	–	–	–
	Fulla	≥14	172/3224 (5%)	–	–	–	–	–	–
Alsaed et al. 2022 [30]	Second	NR	–	3/5 (60%) anti-RBD Ab positive (≥132 BAU/ml)	–	–	–	–	–
Li et al. 2022 [31]	NR	NR	–	–	–	62/70 (89%)	0/70 (0%)	–	–
Rider et al. 2022 [27]	NR	60	–	–	–	336/763 (44%)	–	53/791 (7%)	–
Fornaro et al. 2022 [33]	First	NR	–	–	–	33/62 (53%)	–	–	–
	Second	28	–	–	–	38/57 (67%)	–	SLEDAI-2K NSC	–
		Mean 49	–	–	–	–	–	4/62 (6%)	0/64 (0%)
Gaur et al. 2021 [34]	NR	NR	–	–	–	6/16 (25%)	–	–	–
Shenoy et al. 2021 [36]	Second	28	–	40/49 (82%) anti-S1 Ab	–	–	–	–	–
Assawasaksakul et al. 2022a [22]	Second	7	–	–	–	55/64 (86%)	–	–	–
		28	–	52/64 (81%) anti-RBD Ab positive (>0.4 AU/ml)	NR	–	–	0/34 (0%) SLEDAI NSC	–
Assawasaksakul et al. 2022b [23]	Third	14	0/71 (0%)	69/71 (97%) anti-RBD positive (>50 AU/ml) 32/71 (45%) anti-RBD high-efficacy (>2360 BAU/ml) 15/36 (42%) anti-RBD NAb positive (inhibition >30%)	31/34 (91%) positive SARS-CoV-2–specific T cell response (>50 spots per 10 PBMCs)	59/71 (83%)	0/71 (0%)	0/71 (0%)	–
		90	0/71 (0%)	–	–	–	–	–	–
	Fourth	14	–	NR/23 anti-RBD high-efficacy	–	21/23 (91%)	0/23 (0%)	2/23 (7%)	–
Hadjadj et al. 2022 [37]	Second	60	NR	13/16 (81%) anti-S Ab positive (BAU >1.1)	NR/11 positive SARS-CoV-2–specific CD3+ T cell response (>11 SFU per 106 cells)	–	–	2/16 (13%)	2/16 (13%)
		150	NR	14/16 (88%) anti-S Ab positive		–	–	–	–
	Third	NR	–	NR	–	–	–	–	–
Szebeni et al. 2022 [38]	Second	30	–	8/11 (73%) anti-RBD Ab positive (21.8 BAU/ml)	–	–	–	–	–
		120	–	8/11 (73%) anti-RBD positive	7/11 (63%) positive SARS-CoV-2–specific CD4+ T cell response (>400 reactive cells per 106 parental cells) 4/11 (36%) positive SARS-CoV-2–specific CD8+ T cell response	–	–	–	–

a

Full vaccine regimen = 2 doses of mRNA vaccines or 1 dose of Johnson and Johnson viral vector vaccine. anti-CoV Ab: anti-SARS-CoV-2 (antigen not specified) IgG; anti-RBD Ab: anti-spike protein receptor-binding domain IgG; anti-S Ab: anti-SARS-CoV-2 trimeric spike glycoprotein IgG; AU/ml: arbitrary unit per millilitre; BAU: binding antibody units; FU: follow-up; n: number of SLE patients with event; N: number of SLE patients evaluated for that event; NAb: neutralizing antibodies; NR: not reported for SLE patients; NSC: no significant change; PBMC: peripheral blood mononuclear cells; S1: spike protein subunit S1; S2: spike protein subunit S2; –: outcome not studied; SFU: spot forming units.

b

Report only available as abstract.

Humoral immunogenicity to COVID vaccination was evaluated in 976 patients (17 studies, 18 reports). A variety of assays (Supplementary Table S3, available at Rheumatology online) were used to detect antibodies against the SARS-CoV-2 spike protein [24, 28, 36, 37, 39, 45, 46], and/or specifically the receptor-binding domain (RBD) of the S1 subunit [22, 23, 26, 28, 30, 38, 46–51], while one study did not specify the target antigen [52]. Results regarding anti-nucleocapsid antibodies were not included, since these antibodies likely reflect COVID-19 infection rather than vaccine immunogenicity. Most studies reported humoral response at 1 week to 6 months after one or two vaccine doses; only 2 studies evaluated humoral response after booster vaccinations [23, 37]. Analyses for factors associated with humoral immunogenicity in SLE patients were mostly univariate or had limited adjustment for possible confounders, such as age, gender, SLE duration and activity, and immunosuppressant use and dose. There was no association with older age [24, 28, 46], gender [28, 46], race [28], SLE disease activity [46], higher dsDNA [46], glucocorticoid use or dose [46, 52], MTX [28, 52], HCQ [46, 52], AZA [28, 46], CYC [28], calcineurin inhibitor [28, 52], or belimumab [28, 46]. However, conflicting data noted that glucocorticoid use and/or higher doses [28, 47, 49] and MTX [46] were associated with reduced humoral response, while higher anti-dsDNA [47] and HCQ monotherapy [28, 47] were associated with seropositivity. In contrast, MMF [28, 46, 47, 49], combination immunosuppressants [47], and non-mRNA vaccines (compared with mRNA vaccines) [47, 49] were consistently reported to be associated with reduced humoral immunogenicity. The use of anti-CD20 therapy was generally low or not reported in the studies in this review. Pooled seropositivity after full vaccination was 81.1% (95% CI: 72.6–88.5%), with significant heterogeneity (Fig. 2). Sensitivity analysis that excluded studies with high risk of bias and fewer than 10 SLE patients did not alter the seropositivity rate or heterogeneity [rate 80.1% (95% CI: 71–87.9%, I² = 87%, P < 0.01]. Pooled seropositive rates were 91.3% (95% CI: 83.0–97.3%), 67.0% (95% CI: 50.8–81.4%) and 52.2% (95% CI: 7.6, 95.1%) for mRNA, inactivated viral and viral vector vaccines, respectively (Supplementary Fig. S2, available at Rheumatology online). Thus, vaccine type, but not type of antibodies (Supplementary Fig. S3, available at Rheumatology online) or follow-up duration (Supplementary Fig. S4, available at Rheumatology online), likely contributed to the heterogeneity in seropositivity. Meta-regression analyses (Supplementary Fig. S5, available at Rheumatology online) showed that seropositivity was not significantly associated with mean age, mean SLE disease duration, or glucocorticoid and mycophenolic acid analogues use among the study cohorts.

Figure 2.

Seropositivity after completing full primary vaccination with COVID-19 vaccines

Cellular immunogenicity was evaluated in 112 patients with SLE (5 studies). T cell response correlated with humoral response to COVID vaccines in studies that focused on patients with SLE [46, 47], and when other immune-mediated or rheumatic conditions were included [22], but not if patients received rituximab or MTX [37].

Acceptance and utilization

Eight cross-sectional studies (Table 3) [29, 31, 32, 34, 35, 46, 53, 54] (one each from China, Hong Kong, Singapore, India, the USA and Mexico, and two from France) reported data on the acceptance of the COVID-19 vaccines among 1168 patients with SLE. Most studies were conducted in 2021 during the worldwide vaccination drive, and 7 studies utilized structured questionnaires administered in-person (n = 4), by electronic mail (n = 1), by telephone (n = 1) or were online web-based (n = 1). Vaccine acceptance was usually expressed as ‘preference’ or ‘willingness’ to be vaccinated, while non-acceptance included study terms such as ‘hesitancy’, ‘no intention’, ‘prefer not’ and ‘refusal’ to be vaccinated. The pooled vaccine acceptance rate was 67.0% (95% CI: 45.2–85.6%). Fig. 3 noted significant heterogeneity but no statistically significant subgroup differences by geographical region. Meta-regression analyses (Supplementary Fig. S6, available at Rheumatology online) showed that vaccine acceptance was significantly associated with the mean age, but not country’s gross domestic product or the sample size of the study cohorts. The lowest rate (18.2%) was reported among 384 patients with SLE in Hong Kong, where safety concerns regarding vaccine side effects (59.6%) and flares (56.5%) dominated. Vaccine hesitancy was associated with immunosuppressant use among patients with SLE [32], and when patients with other immune-mediated or inflammatory rheumatic conditions were included [31, 35].

Figure 3.

Subgroup analysis of COVID-19 vaccine acceptance rates according to study region

Table 3.

COVID-19 vaccine acceptance among patients with SLE

Studya	Centre	Country	Study period	Survey typeb	Patients	Age (years)	Acceptance (%)	Factors associated with VA/VH
Sadun et al. 2021 [54]	Multi	USA	Feb 2021	Electronic mail, structured	63	52.8 (13.4)	95.2	VA not associated with health literacy or numeracy
Liew et al. 2022 [29]	Single	Singapore	Jun–Jul 2021	In-person, structured	18	43.7 (20.8)	88.9	NRc
Moyon et al. 2021 [46]	Single	France	NR	NR	180	46.3 (18.6)	80.6	NR
Guaracha et al. 2022 [32]	Multi	Mexico	Mar–Sep 2021	In-person, structured	400	39.7 (14.1)	69.3	VA associated with older age, longer formal education years; VH associated with immunosuppressant use
Goulenok et al. 2022 [35]	Single	France	Dec 2020 – Mar 2021	Telephone, structured	101	48.7 (42.1)	59.4	NRc
Tang et al. 2022 [53]	Single	China	NR	Web-based, structured	188	39.7 (29.1)	57.5	NRc
Gaur et al. 2021 [34]	Single	India	Mar–Apr 2021	In-person, structured	14	32.14 (10.4)	57.1	NRc
Li et al. 2022 [31]	Multi	Hong Kong	Jun–Oct 2021	In-person, structured	384	NR	18.2	NRc

Categorical variable reported as number or proportion (percentage) and continuous variable reported as mean (s.d.). Where means (s.d.) of continuous variables were not available, they were estimated from the medians and minimum, maximum or interquartile ranges.

a

Studies presented in descending order for acceptance rates.

b

In-person surveys may be conducted in clinics or participants’ homes and self-administered or assisted by interviewers.

c

Descriptive evaluation of factors related to vaccine acceptance, but no comparison between VA group and VH group for SLE patients. VA: vaccine acceptance; VH: vaccine hesitancy; NR: not reported.

Safety

Post-vaccine adverse events occurred in 948 patients (44.8%) after their first vaccination [28, 31, 33, 41, 42, 46, 49, 53, 55, 56], 876 patients (50.8%) after their second dose [22, 25, 28, 33, 41, 42, 49, 52, 53, 55, 56], and 340 patients (43.6%) after an unspecified dose [27, 34]. The occurrence of adverse events after the first dose was not significantly different from that of the second dose [odds ratio (OR) 0.94 (95% CI: 0.62–1.40), I²=82%, P = 0.75; 8 studies]. Table 2 showed that both localized and systemic events were frequent, but severe adverse events, defined as occurrence of anaphylaxis [55], prevented daily activities [56], required hospitalization [25, 31, 39], or according to the Common Terminology Criteria for Adverse Events (CTCAE) [50], or the World Health Organization criteria [28], were infrequent. Among studies that performed analyses to compare SLE patients with and without adverse events, occurrence of adverse events was not associated with age [41, 42, 53], gender [42, 53], disease duration [41, 53], organ involvement [41], glucocorticoid dose [53], or vaccine type [42], but was more likely in SLE patients with pre-existing constitutional symptoms [41], on belimumab therapy [41], or who had received a mRNA vaccine (compared with a non-mRNA vaccine) [49].

Lupus flares were described in 166 patients (5.5%) among SLE patients who received COVID-19 vaccines [22, 23, 27, 28, 33, 37, 41, 42, 46, 47, 51, 53, 55, 56]. Among 6 studies that detailed the organ or system involved, flares were most frequently reported to be musculoskeletal (n = 48), dermatological (n = 22), haematological (n = 12) and kidney (n = 7); no CNS flares were reported. Moderate to severe flares, defined by studies as organ- or life-threatening, required change in SLE treatment [42, 49], i.v. steroid pulses [41], or hospitalization [56], were generally infrequent, at 0–2% (Table 2); only 1 study reported 2 (13%) of 16 SLE patients developed disease flare within 3 months of the first vaccine dose and required modification of their immunosuppressants [37]. Univariate analyses comparing SLE patients with and without flares post-vaccination found that flares were not associated with age [41, 56], SLE disease duration [41, 56], organ involvement [41], glucocorticoid [56], MTX [56], or MMF [56], but were more likely in SLE patients with pre-vaccination positive anti-dsDNA [41], high SLE DAS [41], AZA [56], or belimumab [41]. The risk of flare was reduced in SLE patients with a history of renal involvement [56] and HCQ use [56].

Publication bias

Funnel plots and Egger’s regression tests did not suggest significant publication bias for humoral immunogenicity and vaccine acceptance (Supplementary Fig. S7, available at Rheumatology online).

Discussion

This systematic review included 34 reports from 32 studies that evaluated the emerging data on the efficacy, safety and acceptance of a variety of COVID-19 vaccines among 8269 individuals with SLE worldwide, and it is reported according the PRISMA guidelines (Supplementary Table S4, available at Rheumatology online). It found that post-vaccine COVID-19 infection rates were generally low, and the pooled seropositivity rate was 81.1% (95% CI: 72.6, 88.5%) in individuals with SLE treated with various immunosuppressant regimens. mRNA vaccines were associated with greater humoral immunogenicity compared with non-mRNA vaccines. The pooled vaccine acceptance rate was 67.0% (95% CI: 45.2, 85.6%), with higher rates noted in studies with older patients.

A prior systematic review reported a pooled seropositivity of 90.7% (95% CI: 85.4, 94.2, I² = 0%, P = 0.61) after two doses of mRNA vaccines in 175 SLE patients from three studies [9]. Although data for the viral vector vaccines was limited in this review, we noted that mRNA vaccines had significantly higher seropositive rates than non-mRNA vaccines. The earlier systematic reviews were unable to explore the difference between vaccine types, since those included studies that had limited data on non-mRNA vaccines for SLE patients [5, 7, 9]. The seropositive rates for anti-spike and anti-RBD antibodies were similar, but the heterogeneity in the results may be contributed by the different serological assays. The impact of immunosuppressants was synthesized narratively in this review, since there was likely considerable variability in the immunosuppressant doses and permutations utilized in combination therapy. Glucocorticoid, MMF and MTX were similarly associated with reduced humoral response in reviews in other immune-mediated or rheumatic diseases [1, 5, 6, 9]. The use of anti-CD20 therapy was generally low or not reported in the included studies in this review, but other studies among immune-mediated or rheumatic diseases have noted that anti-CD20 therapies reduce seroconversion after COVID-19 vaccination [6–9].

Information regarding the impact of the different types and intensity of immunosuppressants on the ability of COVID-19 vaccines to prevent severe infections can guide physician and patient shared decision-making. However, real-world data in SLE and other immune-mediated or rheumatic diseases were seldom explored in prior systematic reviews, which largely focused on immunogenicity [1, 5, 7, 9]; one review included only 2 retrospective studies on COVID-19 infection–related hospitalization and mortality rates [6]. Few studies have undertaken the evaluation of clinical effectiveness and long-term adverse events in the short time since vaccine availability (in early 2021 in most countries), possibly because evaluating clinical outcomes such as COVID-19 infection, pneumonia or hospitalizations require significant resources to recruit large numbers of participants for follow-up over months, while results may be confounded by fluctuating risks according to rapidly evolving SARS-CoV-2 viral strains during an extended study. Although there was insufficient data to compare the effectiveness of the different COVID-19 vaccines in SLE, post-vaccination antibody titres in patients with autoimmune diseases were predictive of protection against COVID-19 infection [57], and among the general population, the real-world effectiveness against COVID-19 infection was higher in mRNA vaccines [91.2% (95% CI: 87.9, 94.5%) and 98.1% (95% CI: 96.0, 100%) for Pfizer [23 studies] and Moderna [5 studies], respectively) than the inactivated viral vaccine CoronaVac [65.7% (95% CI: 63.0, 68.5%); 3 studies] [12].

Interestingly, study cohorts with higher mean age had higher vaccine acceptance, echoing COVID-19 vaccine acceptance studies among the general population and health-care workers [58]. Public health policies and campaigns initially prioritized older adults (who were at greater risk of COVID-related hospitalization and death) to receive vaccination [58], so younger individuals may have perceived themselves to be at lower personal risk of severe sequelae from COVID-19 infection. Severe infections were also known to be more frequent among older patients with SLE [59], thus greater physician advocacy or patient acceptance of vaccinations for other preventable infections may have had spill-over effects on reducing COVID-19 vaccine hesitancy in older adults with SLE. Although concerns regarding vaccine safety contributed to vaccine hesitancy, our review found that the rates of lupus flares after COVID-19 vaccines were similar to those after vaccinations against influenza [4], HPV [60], and herpes zoster [61], while severe flares and adverse events were rare. Like influenza, where annual vaccination is recommended due to the antigenic drift and waning of vaccine-induced antibodies over time [62], it is possible that booster or repeated COVID-19 vaccines will be required for at-risk individuals with SLE. Hence contemporaneous data on COVID-19 vaccination rate and acceptance will continue to be relevant in ensuring adequate ongoing seroprotection in patients with SLE and in formulating public health policies to improve vaccination rates.

There are some limitations to this systematic review. The results were based on observational studies with significant heterogeneity, and information may be limited where studies did not include sufficient data regarding potential confounders. While subgroup and meta-regression analyses were performed to evaluate the heterogeneity, these may not comprehensively account for patients’ highly variable immunosuppression doses and permutations, assays, and follow-up time. Most studies were performed in the first half of 2021; hence, the results for the outcomes of interest were likely to be based on the predominant SARS-CoV-2 variants circulating in 2020 and 2021, but may not be generalized to subsequent variants with accumulated mutations on the spike protein that can allow immune escape and vaccine resistance [63]. Notably, only a handful of studies assessed neutralizing antibodies using viral neutralization tests [28, 37, 44, 49], while functional immune response may be considered more clinically relevant than the quantity of total anti-spike- or even anti-RBD-binding antibodies [64]. Follow-up was generally short in most studies, while vaccine-induced immunity may wane over months [65], so the effectiveness and efficacy may be reduced at longer intervals after vaccination. Booster vaccines may prolong host immunity, but this review was unable to evaluate booster effectiveness, since only two studies evaluated booster vaccines in SLE. The paucity of studies of adolescents with SLE in this review may be attributed to health-care policies, since health authorities worldwide initially prioritized older adults for vaccinations and extended vaccinations to adolescents and children in late 2021 [66]. Males and individuals from Africa and Australia were underrepresented in the included studies; hence, the results of COVID-19 vaccine immune responses and acceptance may not be generalizable to people with different intrinsic host factors, such as genetically determined HLA haplotype [67], comorbidities and nutrition [68], and extrinsic factors such as environment and geopolitical circumstances [68], respectively. Despite these limitations, this systematic review of COVID-19 vaccine effectiveness, efficacy, safety, acceptance, and utilization among individuals with SLE provides an update on the published literature to inform patients’ decisions about receiving vaccination to protect themselves during the ongoing COVID-19 pandemic.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability statement

The data underlying this article are available in the article and in its online supplementary material.

Contribution statement

S.Y.S.T. and A.M.Y. contributed equally to this paper and are joint first authors. S.Y.S.T., A.M.Y., J.J.L.S. and C.C.L. conceived and designed the study. S.Y.S.T. and A.M.Y. selected the articles and extracted data. J.J.L.S. and C.C.L. were responsible for statistical analysis. S.Y.S.T. and A.M.Y. wrote the first draft of the manuscript. All authors provided critical conceptual input, interpreted the data analysis, and read and approved the final draft. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Funding

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement: The authors have declared no conflicts of interest.

Acknowledgements

We would like to acknowledge and extend our appreciation to the following authors who have graciously provided data over email correspondence: Dr Lisa Rider, Dr Sebastian Sattui, Dr Jeffrey Sparks, Dr Julia Simard, Dr Kevin Kennedy, Dr Eloisa Bonfa, Dr Ana Medeiros-Ribeiro, Dr Italo Ribeiro Lemes, Dr Omar Alsaed, Dr Chung Ho Yin, Dr Virginia Pascual-Ramos, Dr Christian Ammitzbøll, Dr Marco Fornaro, Dr Florenzo Iannone, Dr Priyanka Gaur, Dr Anuj Shukla, Dr Tiphaine Goulenok, Dr Karim Sacre, Dr Padmanabha Shenoy, Dr Aby Paul, Dr Wonngarm Kittanamongkolchai, Dr Jérome Hadjadj, Dr Benjamin Terrier, Dr Szebeni Gábor János and Dr Attila Balog.

Author notes

S.Y.S.T. and A.M.Y. contributed equally.