COVID-19 mRNA Vaccine in Patients With Lymphoid Malignancy or Anti-CD20 Antibody Therapy: A Systematic Review and Meta-Analysis

Abstract

Background

The humoral response to vaccination in individuals with lymphoid malignancies or those undergoing anti-CD20 antibody therapy is impaired, but details of the response to mRNA vaccines to protect against COVID-19 remain unclear. This systematic review and meta-analysis aimed to characterize the response to COVID-19 mRNA vaccines in patients with lymphoid malignancies or those undergoing anti-CD20 antibody therapy.

Materials and Methods

A literature search retrieved 52 relevant articles, and random-effect models were used to analyze humoral and cellular responses.

Results

Lymphoid malignancies and anti-CD20 antibody therapy for non-malignancies were significantly associated with lower seropositivity rates (risk ratio 0.60 [95% CI 0.53-0.69]; risk ratio 0.45 [95% CI 0.39-0.52], respectively). Some subtypes (chronic lymphocytic leukemia, treatment-naïve chronic lymphocytic leukemia, myeloma, and non-Hodgkin's lymphoma) exhibited impaired humoral response. Anti-CD20 antibody therapy within 6 months of vaccination decreased humoral response; moreover, therapy > 12 months before vaccination still impaired the humoral response. However, anti-CD20 antibody therapy in non-malignant patients did not attenuate T cell responses.

Conclusion

These data suggest that patients with lymphoid malignancies or those undergoing anti-CD20 antibody therapy experience an impaired humoral response, but cellular response can be detected independent of anti-CD20 antibody therapy. Studies with long-term follow-up of vaccine effectiveness are warranted (PROSPERO registration number: CRD42021265780).

Microabstract

Messenger RNA (mRNA) vaccines have been widely used for the prevention of coronavirus disease 2019 (COVID-19). This meta-analysis of 52 articles demonstrated that lymphoid malignancies and anti-CD20 antibody therapy impaired humoral response.

Introduction

Individuals with hematological malignancies are highly susceptible to severe coronavirus disease 2019 (COVID-19).^{1 , 2} The risk for death predominantly among the hospitalized adult population has been reported to be as high as 34%.³ Patients with chronic lymphocytic leukemia (CLL),^{4 , 5} myeloma,^{6 , 7} or lymphoma⁸ are likely to develop serious symptoms due to immunological abnormalities caused by the disease itself and the corresponding immunosuppressive therapies.^{9 , 10} Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) binds to angiotensin-converting enzyme 2 receptors expressed on oral mucosa epithelial cells and lung alveolar type II cells through the receptor binding domain in its spike protein.11, 12, 13 Messenger RNA (mRNA) vaccines targeting the spike protein have been rapidly developed,^{14 , 15} and two, BNT162b2¹⁶ and mRNA-1273,¹⁷ conferred approximately 95% protection against COVID-19 in clinical trials.

However, immunocompromised subjects were excluded from these trials; thus, the efficacy of mRNA vaccines in those with hematological malignancies remains under investigation. Lymphoid malignancies and B cell depletion agents, such as anti-CD20 antibody, have been shown to attenuate conventional vaccine effectiveness.^{18 , 19} Anti-CD20 antibody therapy is efficacious against lymphoma, as well as multiple sclerosis and rheumatic diseases, and some guidelines for rheumatic diseases recommend delaying the administration of vaccines for 5 to 6 months after anti-CD20 therapy to maximize humoral response^{20 , 21}; nevertheless, evidence for mRNA vaccines remains insufficient in this regard.

Neutralizing antibodies generated by the humoral response exert immune protection,²² while the SARS-CoV-2-specific cellular response is also essential for viral elimination and prevention of disease aggravation.23, 24, 25 It remains controversial whether lymphoid malignancies or B cell depletion therapies attenuate cellular responses to the vaccine, and the interaction between T cells and B cells is indispensable for infection control.26, 27, 28, 29

Two mRNA vaccines have been approved for use against COVID-19, and real-world data regarding the response to these vaccines in various patient types have accumulated rapidly. The present systematic review and meta-analysis investigated humoral and cellular immune responses to COVID-19 mRNA vaccines in patients with lymphoid malignancies and those who underwent treatment with anti-CD20 antibody.

Materials and Methods

Literature Search Strategy

This study was registered with PROSPERO (CRD42021265780) and performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines.³⁰ The PubMed and World Health Organization (WHO) COVID-19 database were searched for articles published up to October 2, 2021, without language restriction using the following terms: (“lymphoid malignancy” OR “lymphoid neoplasm” OR lymphoma OR myeloma OR MM OR MGUS OR CLL OR anti-CD20 OR CD20 OR rituximab OR obinutuzumab OR ofatumumab OR ocrelizumab OR veltuzumab OR ocaratuzumab OR ublituximab OR tositumomab OR ibritumomab) AND vaccin* AND (COVID-19 OR SARS-CoV-2).

Study Selection and Quality Assessment

Two authors (YI and AH) independently assessed the titles and abstracts of all articles retrieved in the electronic literature search. Subsequently, the full texts of potentially eligible articles were screened. Studies that lacked sufficient information needed to evaluate outcomes, those that analyzed data only after the first dose of mRNA vaccines, duplicate publications using overlapping patient cohorts, and case series or cohorts with < 10 patients were excluded. Any discrepancies between the authors were resolved through discussion until consensus was reached. A flow diagram of the data extraction process is presented in Figure 1 . The Newcastle-Ottawa scale was used to assess the quality of non-randomized trials.

Figure 1.

PRISMA flow diagram of study selection. After the screening of titles and abstracts of 493 articles, 80 articles were considered to be relevant. Among them, 28 articles were excluded due to several reasons, and 52 articles were included for the analysis.

Endpoints

The primary outcome in the present review was the risk ratio (RR) of the seropositive rates of SARS-CoV-2-specific antibody after the second dose of mRNA vaccine. The secondary outcome was the RR of SARS-CoV-2-specific T cell-positive rates after vaccination. Regarding the interval from anti-CD20 antibody therapy to the first vaccine dose, data were extracted from figures whenever possible. The focus was on mRNA vaccines (BNT162b2 and mRNA-1273); however, some articles included adenoviral vaccines: AZD1222 (ChAdOx1 nCov-19)³¹ and Ad26.CoV2.S.³²

Statistical Analysis

Data were analyzed using EZR (Easy R) statistical software.³³ For each trial, the vaccine response in patients and controls was calculated using RRs. Data were entered into the EZR software for statistical analysis. An RR < 1 indicated an impaired response in the patient group. The random effect model was used in accordance with the method described by Der Simonian-Laird.³⁴ Trial results were assessed using the chi-square test of heterogeneity and the I² measure of inconsistency. Heterogeneity was considered to be statistically significant at P < .10 or an I² statistic > 50%. Publication bias was examined using funnel plots coupled with the Egger's test. Pooled estimates were calculated using the MetaXL add-in for Excel (Microsoft Corporation, Redmond, WA).³⁵

Results

Study Selection

The literature search of the PubMed and WHO COVID-19 database retrieved 493 articles after removal of duplicates, of which 80 were considered to be relevant through evaluation of titles and abstracts. Among them, 52 studies fulfilled the criteria for the present meta-analysis: 30 investigated lymphoid malignancies36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65; and 22 investigated anti-CD20 antibody therapy for non-malignant diseases, such as multiple sclerosis and rheumatic diseases.66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87 28 articles were excluded with the following reasons: (1) insufficient data of outcomes,88, 89, 90, 91, 92, 93, 94, 95, 96 (2) duplicate publications from an overlapping cohort,97, 98, 99, 100, 101, 102, 103, 104, 105, 106 and (3) case series or cohorts with < 10 patients.107, 108, 109, 110, 111, 112, 113, 114, 115 A flow diagram of the article selection process is shown in Figure 1, and the characteristics of each study are summarized in Table 1 .

Table 1.

Characteristics of Studies Included in the Meta-Analysis

Author	Ref	Location	Disease	Total	Pos	Age	Control	Total	Pos	Age	Vaccine	Interval	Antibody	Measurement assay	Cut-off	NOS
- Lymphoid malignancies -
Chiarucci M	36	Italy	Lymphoma, Myeloma after auto-HSCT	38	32	60	healthy	45	NR	NR	BNT162b2	30 d	Spike	LIAISON SARS-CoV-2 Trimeric S IgG assay (CLIA)	15 AU/mL	6
Gavriatopoulou M	37	Greece	WM	74	31	73	healthy	212	181	66	BNT162b2, AZD1222	4 wk	NAb	cPASS SARS-CoV-2 Nabs Detection Kit (ELISA)	50%	9
Shapiro LC	48	US	Lymphoid malignancy	86	71	70.5	-				BNT162b2, mRNA-1273, Ad26.COV2.S	> 2 wk	RBD	AdviseDx SARS-CoV-2 IgG II assay (CLIA)	50 AU/mL	6
Peeters M	59	Belgium	Lymphoid malignancy with RTX	29	2	63	healthy	40	40	48	BNT162b2	28 d	RBD	Wantai SARS-CoV-2 IgG (ELISA)	200 IU/mL	8
Bergman P	60	Sweden	CLL	79	50	NR	healthy	78	78	NR	BNT162b2	14 d	RBD	Elecsys Anti-SARS-CoV-2 S	0.80 U/mL	8
Terpos E	61	Greece	Lymphoid malignancy	132	58	64.6	healthy	214	204	69.8	BNT162b2	4 wk	NAb	cPASS SARS-CoV-2 Nabs Detection Kit (ELISA)	50%	9
Lim SH	62	UK	Lymphoma	55	39	69	healthy	65	65	45	BNT162b2	2-4 wk	Spike	Meso Scale Discovery (ECLIA)	0.55 BAU/mL	8
Perry C	63	Israel	B-NHL	149	73	64	healthy	65	64	66	BNT162b2	2-3 wk	RBD	Elecsys Anti-SARS-CoV-2 S	0.80 U/mL	9
Jurgens EM	64	US	Lymphoma, CLL	67	41	71	healthy	35	35	NR	BNT162b2, mRNA-1273	24.5 d	spike	ELISA	10,000	8
Thakkar A	65	US	Lymphoma with anti-CD20 Ab	23	16	67	healthy	26	26	64	BNT162b2, mRNA-1273, Ad26.COV2.S	> 7 d	RBD	AdviseDx SARS-CoV-2 IgG II assay (CLIA)	50 AU/mL	9
Benda M	38	Austria	Lymphoid malignancy	89	57	65.1	-				BNT162b2	4-5 wk	RBD	Elecsys Anti-SARS-CoV-2 S	0.82 BAU/mL	6
Henriquez S	39	France	Myeloma	60	51	69.86	healthy	20	20	NR	BNT162b2	1-2 mo	spike	S-flow (SARS-Cov-2 IgG)	40%	8
Terpos E	40	Greece	Myeloma	276	158	74	healthy	226	183	NR	BNT162b2, AZD1222	4 wk	NAb	cPASS SARS-CoV-2 Nabs Detection Kit (ELISA)	50%	9
Maneikis K	41	Lithuania	Lymphoid malignancy	163	97	65	healthy	67	67	40	BNT162b2	7-21 d	spike (S1)	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	50 AU/mL	8
Parry H	42	UK	CLL	55	39	69	healthy	37	36	NR	BNT162b2, ChAdOx1	18 d	spike	Dried blood spot ELISA	ratio 1	9
Stampfer SD	43	US	Myeloma	103	50	68	healthy	31	29	61	BNT162b2, mRNA-1273	14-21 d	spike	ELISA	250 IU/mL	9
Gurion R	44	Israel	Lymphoma	162	83	65	-				BNT162b2	2-6 wk	spike	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	50 AU/mL	6
Benjamini O	45	Israel	CLL	373	160	70	-				BNT162b2	2-3 wk	spike	Liaison SARS-CoV-2 S1/S2 IgG or Architect AdciseDx SARS-CoV-2 IgG II or RBD-IgG ELISA	15 U/mL or 50 U/mL or 1.1	6
Avivi I	46	Israel	Myeloma	171	133	70	healthy	64	63	67	BNT162b2	14-21 d	RBD	Elecsys Anti-SARS-CoV-2 S	0.80 U/mL	9
Ghione P	47	US	Lymphoma, Myeloma	86	36	70	healthy	201	197	NR	BNT162b2, mRNA-1273, Ad26.COV2.S	2-8 wk	spike (S1)	KSL chemiluminescence immunoassay (CLIA)	1.0 COI	8
Tzarfati KH	49	Israel	Lymphoid malignancy	194	131	71	healthy	108	107	69	BNT162b2	32 d	spike	Liaison SARS-CoV-2 S1/S2 IgG (CLIA)	12 AU/mL	9
Oekelen OV	50	US	Myeloma	260	219	68	healthy	67	67	NR	BNT162b2, mRNA-1273	> 10 d	spike	Kantrao COVID-SeroKlir IgG Ab kit (ELISA)	5 AU/mL	8
Diefenbach C	51	US	Lymphoma, CLL	18	4	63	healthy	3	3	NR	BNT162b2, mRNA-1273	4-8 wk	RBD	multiplex bead-binding assay	mean +3 x s.d.	8
Pimpinelli F	52	Italy	Myeloma	42	33	73	healthy	36	36	81	BNT162b2	2 wk	spike	Liaison SARS-CoV-2 S1/S2 IgG (CLIA)	15 AU/mL	9
Roeker LE	53	US	CLL	44	23	71	-				BNT162b2, mRNA-1273	21 d	spike	Liaison SARS-CoV-2 S1/S2 IgG (CLIA)	15 AU/mL	6
Herishanu Y	54	Israel	CLL	167	66	71	healthy	52	52	68	BNT162b2	2-3 wk	RBD	Elecsys Anti-SARS-CoV-2 S	0.80 U/mL	9
Agha M	55	US	Lymphoid malignancy	63	33	71	-				BNT162b2, mRNA-1273	23 d	RBD	semi-quantitative Beckman Coulter SARS-CoV-2 platform	1.0 S/CO	6
Dhakal B	56	US	Lymphoma, Myeloma after auto-HSCT	45	27	65	-				BNT162b2, mRNA-1273, Ad26.COV2.S	>2 wk	spike (S1)	EUROIMMUN (ELISA)	NR	6
Greenberger LM	57	US	Lymphoid malignancy	1311	969	66	-				BNT162b2, mRNA-1273	14 d	RBD	Elecsys Anti-SARS-CoV-2 S	0.8 U/mL	6
Re D	58	France	Lymphoid malignancy	79	45	75.5	-				BNT162b2, mRNA-1273	3-5 wk	spike	anti-spike IgG	-	6
- Non-malignant diseases treated with anti-CD20 Ab -
Sormani MP	66	Italy	MS	179	83	45.8	untreated MS	87	87	45.8	BNT162b2, mRNA-1273	4 wk	RBD	Elecsys Anti-SARS-CoV-2 S	0.80 U/mL	8
Disanto G	67	Switzerland	MS	56	29	56	untreated MS	13	13	51.8	BNT162b2, mRNA-1273	26 d	RBD	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	50 AU/mL	9
Brill L	78	Israel	MS	49	20	47.9	healthy	35	35	45.3	BNT162b2	2-4 wk	RBD	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	50 AU/mL	9
Apostolidis SA	81	US	MS	20	10	40	healthy	10	10	35	BNT162b2, mRNA-1273	25-30 d	RBD	ELISA	NR	9
Sabatino JJ	82	US	MS	35	9	46	healthy	13	13	35	BNT162b2, mRNA-1273	2 wk	RBD	Luminex assay	MFI 5.0	8
Novak F	83	Denmark, US	MS	60	22	47	-				BNT162b2	2-4 wk	RBD	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	7.1 BAU/mL	6
Moor MB	84	Switzerland	Autoimmunity/Cancer /Transplantation	96	47	67	healthy	29	29	54	BNT162b2, mRNA-1273	1.8 mo	spike (S1)	EUROIMMUN (ELISA)	1.1 index	8
Mrak D	85	Austria	Immune-mediated inflammatory disease	74	29	61.7	healthy	10	10	NR	BNT162b2, mRNA-1273	21.9 d	RBD	Elecsys Anti-SARS-CoV-2 S	NR	8
Ali A	86	US	MS, NMO	22	8	43.5	healthy	7	7	41.6	BNT162b2, mRNA-1273	3 wk	RBD	Siemens SARS-CoV-2 spike RBD total antibody assay (CLIA)	index value 1	9
Benucci M	87	Italy	RA	14	10	58	-				BNT162b2	3 wk	RBD	ThermoFisher (FEIA)	NR	6
Gadani SP	68	US	MS	39	22	47.78	untreated MS	14	14	57.42	BNT162b2, mRNA-1273, Ad26.COV2.S	4-8 wk	spike (S1)	EUROIMMUN (ELISA)	1.24	9
Prendecki M	69	UK	Autoimmune disease	75	40	53.7	healthy	70	70	41.4	BNT162b2, ChAdOx1	21 d	spike	Abbott Architect SARS-CoV-2 IgG Quant II (CMIA)	7.1 BAU/mL	8
Connolly CM	70	US	AAV	44	17	69	-				BNT162b2, mRNA-1273, Ad26.COV2.S	NR	spike	Elecsys or Liaison or EUROIMMUN	NR	6
Tallantyre EC	71	UK	MS	134	33	50.2	MS without DMT	92	85	50.2	BNT162b2, ChAdOx1	4.6 wk	RBD	Dried blood spot ELISA	0.56	8
Madelon N	72	Switzerland	MS, RD	37	24	45.6/
58.0	healthy	22	22	54.5	BNT162b2, mRNA-1273	30 d	RBD	Elecsys Anti-SARS-CoV-2 S	0.8 IU/mL	9
Stefanski AL	73	Germany	RA, AAV	19	13	58	healthy	30	30	57	BNT162b2, mRNA-1273, ChAdOx1	3-4 wk	spike (S1)	EUROIMMUN (ELISA)	NR	9
Ammitzbøll C	74	Denmark	SLE, RA	17	4	70	-				BNT162b2	1 wk	spike (S1)	VITROS SARS-CoV-2 total antibody (CLIA)	1 S/CO	6
Guerrieri S	75	Italy	MS	16	6	43.3	-				BNT162b2, mRNA-1273	> 2 wk	spike	ECLIA/CMIA/CLIA/DELFIA	NR	6
Bigaut K	76	France	MS	11	3	53.5	MS without DMT	2	2	53.5	BNT162b2, mRNA-1273	18 d	spike	Abbott/Elecsys	0.72-1.54 U/mL	8
Spiera R	77	US	RD	30	10	61.3	-				BNT162b2, mRNA-1273	NR	spike	Elecsys/Siemens healthineers SARS-CoV-2 Total Assay/ADVIA Centaur XP/XPT	NR	6
Achiron A	79	Israel	MS	44	10	53.2	healthy	47	46	54.3	BNT162b2	1 mo	spike (S1)	EUROIMMUN (ELISA)	index value 1.1	9
Deepak P	80	US	chronic inflammatory disease	10	5	45.5	healthy	53	52	43.4	BNT162b2, mRNA-1273	1-2 wk	spike	ELISA	NR	9

Abbreviations: AAV = ANCA-associated vasculitis; Anti-CD20 Ab = anti-CD20 antibody; auto-HSCT = autologous hematopoietic stem cell transplantation; B-NHL = B-cell non-Hodgkin Lymphoma; CLIA = chemiluminescence immunoassay; CLL = chronic lymphocytic leukemia; CMIA = chemiluminescent microparticle immunoassay; DMT = disease modifying therapy; ECLIA = electrochemiluminescence immunoassay; ELISA = enzyme-linked immunosorbent assay; FEIA = fluorimetric enzyme-linked immunoassay; Interval = interval from second vaccination to antibody test; MS = multiple sclerosis; Nab = neutralizing antibody; NMO = neuromyelitis optica; NOS = Newcastle-Ottawa scale; NR = not reported; Pos = positive number; RA = rheumatoid arthritis; RBD = receptor binding domain; RD = rheumatic disease; RTX = rituximab; SLE = systemic lupus erythematosus; Total = total number; UK = United Kingdom; US = United States; WM = Waldenström macroglobulinemia.

Humoral Response in Lymphoid Malignancies

Data regarding humoral response in lymphoid malignancies compared with healthy controls were reported in 20 articles^{37 ,} 39, 40, 41, 42, 43 ^{, 46 , 47 ,} 49, 50, 51, 52 ^{, 54 ,} 59, 60, 61, 62, 63, 64, 65 that included 2203 patients with CLL, myeloma, non-Hodgkin lymphoma (NHL), and Hodgkin lymphoma (HL). Patients with lymphoid malignancies exhibited significantly lower seropositivity rates than healthy controls (RR 0.60 [95% confidence interval (CI) 0.53-0.69]), with high heterogeneity (I² = 94%, P < .01) (Figure 2 A). The funnel plot suggested a publication bias (P < .05, Figure 2B).

Figure 2.

Humoral response in lymphoid malignancies. (A) Risk ratios for seropositivity rates of patients with lymphoid malignancies compared with healthy controls, and (B) funnel plot.

Humoral Response in Individual Subtypes of Lymphoid Malignancies

Individual subtypes of lymphoid malignancies were analyzed. First, for CLL, a positive humoral response was observed in 52% (95% CI 43%-62%)^{42 , 45 , 49 ,} 53, 54, 55 ^{, 57 , 60 , 64} (Figure 3 A). Data for 356 patients in five articles were eligible for the analysis of RR.^{42 , 49 , 54 , 60 , 64} Patients with CLL exhibited significantly lower seropositive rates than healthy controls (RR 0.55 [95% CI 0.43-0.71]) (Figure 3B). Second, for myeloma, a positive humoral response was observed in 78% (95% CI 69%-86%)^{36 ,} 38, 39, 40, 41 ^{, 43 , 46 ,} 48, 49, 50 ^{, 52 ,} 55, 56, 57, 58 (Figure 3C). Data regarding 1041 patients from 8 cohorts were eligible for the analysis of RR,39, 40, 41 ^{, 43 , 46 , 49 , 50 , 52} and myeloma significantly reduced seropositive rates (RR 0.76 [95% CI 0.69-0.83]) (Figure 3D). Third, for NHL, a positive humoral response was observed in 61% (95% CI 50%-71%) (Figure 3E).^{36 , 44 , 48 , 49 , 57 , 63 , 64} Data for 282 patients from 3 articles were eligible for the analysis of RR,^{49 , 63 , 64} which revealed a low seropositivity rate in patients with NHL (RR 0.58 [95% CI 0.48-0.71]) (Figure 3F). When NHL was subdivided into aggressive and indolent NHL, both subgroups exhibited lower seropositivity rates than control (aggressive NHL, RR 0.60 [95% CI 0.42-0.86]; indolent NHL, RR 0.54 [95% CI 0.43-0.67])^{49 , 63} (Figure 3G and H). With regard to T-cell NHL, one article reported that the seropositivity rate was 84.6% (11 out of 13 patients).⁵⁷ Fourth, for HL, a positive humoral response was observed in 95% (95% CI 89%-99%)^{36 , 44 , 48 , 49 , 57 , 64} (Figure 3I). Data that could be compared with healthy controls were available from only 2 articles (20 patients),^{49 , 64} which revealed no significant difference from control (RR 0.95 [95% CI 0.85-1.07]) (Figure 3J).

Figure 3.

Humoral response in each subtype of lymphoid malignancies. Pooled estimates of seropositivity rates for patients with (A) chronic lymphocytic leukemia (CLL), (C) myeloma, (E) non-Hodgkin lymphoma (NHL), and (I) Hodgkin lymphoma (HL). Risk ratios (RRs) for seropositivity rates of patients with (B) CLL, (D) myeloma, (F) NHL, (G) aggressive NHL, (H) indolent NHL, and (J) HL compared with healthy controls.

Humoral Response in Treatment-Naïve Patients

Low-risk patients with CLL, smoldering multiple myeloma (SMM), and indolent NHL are often offered “watchful waiting” until disease progression, and data regarding treatment-naïve patients can be used to estimate the extent to which lymphoid malignancy itself impairs immune function. First, for CLL, positive humoral response was observed in 77% (95% CI 63%-88%) of treatment-naïve patients,^{42 , 45 , 53 , 54 , 60 , 64} which was significantly lower than control (RR 0.79 [95% CI 0.63-1.00], P = .047)^{42 , 54 , 60 , 64} (Figure 4 A and 4B). On the other hand, SMM and treatment-naïve indolent NHL did not exhibit a significant difference from healthy controls, although patient numbers were relatively small (SMM, seropositivity rate, 94% [95% CI 76%-100%],^{40 , 43 , 46 , 50 , 57} RR 0.96 [95% CI 0.75-1.24]^{40 , 43 , 46 , 50}; treatment-naïve indolent NHL, seropositivity rate, 84% [95% CI 75%-92%],^{44 , 63 , 64} RR 0.90 [95% CI 0.81-1.01]^{63 , 64}) (Figure 4C-4F).

Figure 4.

Humoral response in treatment naïve lymphoid malignancies. Pooled estimates of seropositivity rates for treatment naïve patients with (A) chronic lymphocytic leukemia (CLL), (C) smoldering multiple myeloma (SMM), and (E) indolent non-Hodgkin lymphoma (NHL). Risk ratios for seropositivity rates of treatment naïve patients with (B) CLL, (D) SMM, and (F) indolent NHL compared with healthy controls.

Humoral Response in Lymphoid Malignancies With B-Cell Target Therapy

Next, the impact of B-cell target therapy on humoral response was analyzed, first focusing on anti-CD20 antibody. Patients treated with anti-CD20 antibody exhibited a lower seropositivity rate than healthy controls (RR 0.37 [95% CI 0.24-0.57])^{41 , 47 , 49 , 54 ,} 59, 60, 61 ^, 63, 64, 65 (Figure 5 A). When divided according to the interval from the last infusion with anti-CD20 antibody to the first vaccine dose, treatment within the past 6 months was significantly associated with decreased rates of seropositivity compared to treatment > 6 months before vaccination (RR 0.21 [95% CI 0.09-0.46])^{41 , 44 , 48 , 63 , 64} (Figure 5B). Treatment within the past 12 months also decreased the rates of seropositivity (RR 0.23 [95% CI 0.10-0.57])^{36 , 41 , 44 , 45 , 54 , 58 , 64 , 65} (Figure 5C). Moreover, treatment > 12 months before vaccination resulted in a lower seropositivity rate than healthy controls (RR 0.61 [95% CI 0.51-0.73])^{41 , 54 , 64 , 65} (Figure 5D). With regard to other B-cell target therapies, myeloma patients undergoing anti-CD38 therapy exhibited decreased seropositivity rates compared to patients without anti-CD38 therapy (RR 0.86 [95% CI 0.76-0.96])^{39 , 40 , 46 , 48 , 50 , 52} (Figure 5E). Bruton's tyrosine kinase (BTK) inhibitor also reduced seropositivity rates (RR 0.49 [95% CI 0.37-0.64]) (Figure 5F).^{42 , 45 ,} 47, 48, 49 ^{, 53 , 54 , 57 , 60 , 64}

Figure 5.

Humoral response in lymphoid malignancies with B-cell target therapy. (A) Risk ratios (RRs) for seropositivity rates of patients treated with anti-CD20 antibody compared with healthy controls. (B-D) RRs for seropositivity rates of patients with (B) < 6 months from therapy vs. > 6 months from therapy, and (C) < 12 months from therapy vs. > 12 months from therapy, and (D) > 12 months from therapy vs. healthy controls. (E and F) RRs for seropositivity rates of patients (E) with anti-CD38 therapy vs. without anti-CD38 therapy, and (F) with BTK inhibitor vs. without BTK inhibitor. BTK = Bruton's tyrosine kinase.

Humoral Response in Non-Malignant Diseases With Anti-CD20 Antibody

The impact of anti-CD20 antibody on immune response, including cellular immunity, was further analyzed by focusing on non-malignant patients treated with anti-CD20 antibody. First, the relationship between humoral response and anti-CD20 antibody was investigated. Data from 16 articles (900 patients) revealed that anti-CD20 antibody treatment significantly decreased seropositivity rates compared with the control group (RR 0.45 [95% CI 0.39-0.52]) (Figure 6 A).66, 67, 68, 69 ^, 71, 72, 73 ^{, 76 ,} 78, 79, 80, 81, 82 ^, 84, 85, 86 The funnel plot did not reveal any publication bias (P = .12, Figure 6B). Treatments within the past 6 months,67, 68, 69, 70 ^{, 74 ,} 77, 78, 79, 80 ^{, 83 , 84 , 86} 9 months,^{67 , 68 , 74 , 78 , 87} and 12 months^{68 , 77 , 84 , 85 , 87} were all associated with significantly decreased seropositivity rates compared with treatment > 6, 9, and 12 months before vaccination, respectively (within 6 months, RR 0.45 [95% CI 0.35-0.57]; within 9 months, RR 0.54 [95% CI 0.34-0.84]; within 12 months, RR 0.49 [95% CI 0.33-0.73]) (Figure 6C-6E). Patients treated > 12 months before vaccination still had decreased seropositivity rates compared with the control group (RR 0.70 [95% CI 0.55-0.88])^{68 , 84 , 85} (Figure 6F).

Figure 6.

Humoral response in non-malignant diseases with anti-CD20 antibody. (A) Risk ratios (RRs) for seropositivity rates of patients treated with anti-CD20 antibody compared with controls, and (B) funnel plot. (C-F) RRs for seropositivity rates of patients with (C) < 6 months from therapy vs. > 6 months from therapy, (D) < 9 months from therapy vs. > 9 months from therapy, (E) < 12 months from therapy vs. > 12 months from therapy, and (F) > 12 months from therapy vs. controls.

Cellular Response Among Individuals Undergoing Anti-CD20 Antibody Treatment

The influence of anti-CD20 antibody therapy on cellular response was investigated. Six studies examined SARS-CoV-2-specific T cell responses using the interferon gamma (IFN-γ) assay,^{68 , 69 , 78 , 84 , 85 , 87} and revealed that 78% of patients treated with anti-CD20 antibody elicited a positive cellular response (95% CI 45%-99%) (Figure 7 A) that was comparable with the control group (RR 0.77 [95% CI 0.55-1.08])^{68 , 69 , 78 , 84 , 85} (Figure 7B). In addition, four studies examined cellular responses using an activation-induced marker (AIM) assay, two of which revealed no significant difference without available quantitative data.^{81 , 82} The meta-analysis of other two articles exhibited no difference between patients treated with anti-CD20 antibody and controls (AIM-positive CD4 cell, RR 0.98 [95% CI 0.82-1.18]; AIM-positive CD8 cell, RR 0.93 [95% CI 0.43-2.04])^{72 , 73} (Figure 7C and 7D).

Figure 7.

Cellular response in non-malignant diseases with anti-CD20 antibody. (A) Pooled estimates of positive T cell response rates for patients treated with anti-CD20 antibody evaluated by IFN-γ assay. (B-D) Risk ratios for T cell response rates of patients treated with anti-CD20 antibody compared with controls, evaluated by (B) IFN-γ assay, (C) activation induced marker (AIM) assay for CD4-positive cells, and (D) AIM assay for CD8-positive cells. IFN-γ = interferon gamma.

Discussion

Lymphoid malignancies attenuated the humoral response to COVID-19 mRNA vaccines. Moreover, subgroup analysis further revealed that CLL, NHL, and myeloma patients exhibited decreased seropositivity rates. Patients with lymphoid malignancies are immunocompromised due to the disease itself.^{9 , 10} In particular, treatment-naïve CLL patients exhibited lower seropositivity rates than healthy controls. In CLL patients, the humoral response was also impaired after contracting COVID-19,¹¹⁶ and the effectiveness of other conventional vaccines was attenuated.¹⁸ These data reflect the substantial immune abnormalities associated with CLL itself.^{117 , 118} On the other hand, patients with HL and treatment-naïve SMM exhibited high seropositivity rates that were equivalent to healthy controls, suggesting a difference in the influence of disease subtype on the immune system.

With regard to the influence of treatment on humoral response, anti-CD20 antibody, anti-CD38 therapy, and BTK inhibitor significantly decreased seropositivity rates. These agents target B cell function, and many studies have demonstrated a correlation between B cell counts in peripheral blood and seropositivity rates.^{36 , 38 , 40 , 43 , 63} These agents have also been reported to decrease the effectiveness of several other conventional vaccines.119, 120, 121, 122, 123 Regarding anti-CD20 antibody therapy, we analyzed vaccine immunogenicity divided by the interval from the last infusion to the first vaccine dose. An interval of < 6 months and 12 months significantly attenuated seropositivity rates. These results were confirmed in a meta-analysis of non-malignant patients treated with anti-CD20 antibody. B cell reconstitution occurs > 6 months after anti-CD20 antibody therapy,^{124 , 125} and some guidelines for rheumatic diseases recommend delaying administration of the vaccine for 5 to 6 months after anti-CD20 therapy.^{20 , 21} A recent meta-analysis of influenza vaccine also demonstrated that anti-CD20 antibody treatment within at least the past 6 months abrogated the humoral response,¹⁹ which is consistent with the results of our meta-analysis. Moreover, some articles reported that recovery of the memory B-cell pool after anti-CD20 antibody therapy in the lymphoma population is delayed compared with normal B-cell ontogeny and remains impaired after 12 months.^{47 , 124} Our meta-analysis revealed that anti-CD20 antibody therapy >12 months before vaccination still attenuated the humoral response compared with healthy controls, which suggests a prolonged immunosuppressive state caused by B cell depletion.

Humoral and cellular responses work closely together against viral infection and vaccines¹²⁶; however, it is controversial whether B cell activity is essential for T cell priming, activation, and expansion.26, 27, 28, 29 We analyzed T cell responses against mRNA vaccines under conditions of B-cell depletion caused by anti-CD20 antibody therapy, which demonstrated that mRNA vaccines elicited SARS-CoV-2-specific T cell response without adequate B cell function, and there was no correlation between antibody formation capacity and T cell response. A previous study about influenza vaccine was consistent with our data,¹²⁷ suggesting that patients treated with anti-CD20 antibody should not avoid vaccination. On the other hand, several large-scale studies have reported that anti-CD20 therapy increases the exacerbation risk from COVID-19 in patients with multiple sclerosis and rheumatic diseases128, 129, 130, 131, 132; thus, highlighting the importance of B cell response during infection. Our analysis of T cell response was only from studies performed in vitro; thus, whether immune memory by cellular response can prevent infection and disease aggravation in the absence of immune protection by humoral response²² should be further evaluated in clinical long-term follow-up studies. However, CD8-positive T cells can positively influence recovery^{133 , 134}; thus, activation of cellular immunity without humoral response can provide a level of efficacy.

This meta-analysis had several limitations. First, we evaluated antibody formation several weeks after vaccination. In addition to immunogenicity, long-term vaccine effectiveness is an important parameter in evaluating vaccine function. Several studies have shown that antibody titers tend to be lower than in healthy controls, even in seropositive patients with lymphoid malignancies; thus, long-term follow-up is further warranted. Second, the heterogeneity in disease status and treatment history in each study will affect the seroresponse to vaccines. For example, MM patients with a complete response (CR) achieved higher antibody levels than non-CR patients,⁴³ and exposure to > 3 novel anti-myeloma dugs were associated with lower response rates.⁴⁶ Therefore, humoral response in each status should be further analyzed individually. Also, anti-CD20 antibody is often used with cytotoxic agents or other immunosuppressive agents as combination therapy, and these agents will affect the humoral response to vaccines. Third, the measurement method for SARS-CoV-2 antibodies and cut-off values of seropositivity differed among the selected studies as summarized in Table 1. Most studies evaluated antibodies against the receptor binding domain or total spike protein instead of neutralizing antibody using different assays, such as chemiluminescent immunoassay, enzyme-linked immunosorbent assay, and flow-cytometry analysis. The receptor binding domain is poorly conserved among SARS coronaviruses and relatively specific to SARS-CoV-2, whereas the antibody against the entire spike protein can be elevated after other coronavirus infections.^{135 , 136} However, the results from most of the assays were correlated with one another,¹³⁷ and also with neutralizing antibodies.^{81 , 138} Fourth, regarding the measurement method for the SARS-CoV-2-specific T cell response, several studies evaluated the response using SARS-CoV-2 spike peptides through IFN-γ production or AIM assay.¹³⁹ Whether these in vitro data can predict clinical outcomes of COVID-19 remains unknown, and some argue that these available evidences remain insufficient for clear guidance.¹⁴⁰ Additionally, all data used for our meta-analysis were from non-malignant patients, and there is only one study that evaluated the T cell response against patients with lymphoid malignancies to date.³⁹ This article revealed that T cell response was attenuated in myeloma patients compared with healthy controls, and also showed the discrepancy between T cell response and antibody formation capacity. In comparison with multiple sclerosis,¹⁴¹ lymphoid malignancies cause a high disorder of lymphoid systems^{9 , 10}; thus, cellular response in these groups should be further analyzed.

Conclusion

This meta-analysis demonstrated that lymphoid malignancies, as well as some subtypes, including CLL, NHL, and myeloma, attenuated humoral response. Treatment-naïve CLL and B-cell target therapy including anti-CD20 antibody, anti-CD38 therapy, and BTK inhibitor also demonstrated decreased humoral response. Regarding the interval from last anti-CD20 antibody therapy to the first vaccine dose, an interval < 6 months significantly attenuated seropositivity rates, and anti-CD20 antibody therapy > 12 months before vaccination still impaired humoral response. Cellular response was detected independent of anti-CD20 antibody therapy or antibody formation capacity. Further studies focusing on long-term follow-up and T cell responses in lymphoid malignancies are warranted.

Clinical Practice Points

Patients with hematological malignancies are highly susceptible to severe COVID-19. mRNA vaccines have been widely used against COVID-19.

Patients with lymphoid malignancies or those undergoing anti-CD20 antibody therapy experience an impaired humoral response.

Cellular response can be detected independent of anti-CD20 antibody therapy in non-malignant patients.

Authorship statement

Y.I. conceptualized and designed the research, performed literature search, analyzed data, and wrote the manuscript. A.H. performed literature search and analyzed data. M.K. supervised the research.

Disclosure

Y.I. declares no competing financial interests. A.H. reports honoraria from Janssen Pharmaceutical. M.K. reports honoraria from AstraZeneca, Chugai Pharmaceutical, Janssen Pharmaceutical, Sanofi, and Pfizer, and research funding from Chugai Pharmaceutical and Pfizer.