Study selection
The study selection flowchart is presented in Figure 1. The literature search, after removing duplicates, resulted in 2093 studies, of which 1992 were considered irrelevant following title and abstract screening. Of the remaining 101, a further 75 were removed according to the exclusion criteria. Therefore, in total, 26 studies (16-41) were eligible for the meta-analysis of seroconversion after the second dose of the vaccine.
Characteristics of included studies
Characteristics of the included studies are provided in Table 1. All 26 included studies on 3207 IC patients and 1726 healthy controls showed that 65.8 % IC patients and 99.2 % healthy controls had seropositive IgG test following second dose of COVID-19 mRNA vaccines. All of the studies were conducted in 2021. Sample sizes, from which relevant data were available for extraction, varied from 40 to 807. Participants’ mean age ranged from 42 to 71.4 years. The majority of the studies (18, 20, 23-25, 27, 29-34, 36-41) had a prospective cohort design (n = 18). Five studies (17, 19, 21, 22, 28) had a retrospective cohort design and three (16, 26, 35) were cross-sectional.
Table 1
Details of the data presented by the included studies.
Study (First Author)
|
Country
|
Study design
|
Total sample size
|
Case
|
Control
|
Etiology of IC condition
|
Type of vaccine
|
No of cases
|
Male% of cases
|
Age
|
No of Non-cases (if applicable)
|
Male% of Non-cases
|
Age
|
Sattler A
|
Germany
|
Prospective Cohort
|
78
|
39
|
71.8
|
57.3
|
39
|
51.2
|
53.0
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Rincon-Arevalo H
|
Germany
|
Prospective Cohort
|
75
|
40
|
70
|
62.4 [51.2-69.5]*
|
35
|
57.1
|
51 [34-80]*
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Korth J
|
Germany
|
Prospective Cohort
|
46
|
23
|
48
|
57.7
|
23
|
39
|
44.4
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Rabinowich L
|
Israel
|
Cross-sectional
|
105
|
80
|
70
|
60.1
|
25
|
32
|
52.7
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Schramm R
|
Germany
|
Prospective Cohort
|
100
|
50
|
64
|
55
|
50
|
34
|
47
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Cao J
|
USA
|
Retrospective Cohort
|
47
|
37
|
72.9
|
64 [50-69]*
|
10
|
20
|
66 [57-75]*
|
Transplant
|
mRNA-1273 (Moderna) or BNT162b2 (Pfizer/BionTech)
|
Grupper A
|
Israel
|
Retrospective Cohort
|
151
|
136
|
81.7
|
58.6
|
25
|
32
|
52.7
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Marinaki S
|
Greece
|
Prospective Cohort
|
150
|
34
|
79.4
|
60 [49.1-68.4]*
|
116
|
-
|
-
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Rashidi-Alavijeh J
|
Germany
|
Prospective Cohort
|
63
|
43
|
60.5
|
57 [49–64]*
|
20
|
45
|
43.5 [38-53.5]*
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Hod T
|
Israel
|
Prospective Cohort
|
322
|
120
|
80
|
59.7
|
141
|
30.2
|
57.04
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Stumpf J
|
Germany
|
Prospective Cohort
|
512
|
368
|
65.5
|
57.3
|
144
|
23.6
|
48
|
Transplant
|
(a) mRNA-1273 (Moderna) (n=143);
(b) BNT162b2 (Pfizer/BionTech) (n=369)
|
Firket L
|
USA
|
Retrospective Cohort
|
40
|
20
|
45
|
51.2
|
20
|
65
|
48.3
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Peled Y
|
Israel
|
Prospective Cohort
|
213
|
77
|
64
|
62 [49-68]*
|
136
|
37
|
63
|
Transplant
|
BNT162b2 (Pfizer/BionTech)
|
Monin L
|
UK
|
Prospective Cohort
|
205
|
151
|
52
|
73 [64.5-79.5]*
|
54
|
52
|
40.5 [31.3-50]*
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Pimpinelli F
|
Italy
|
Prospective Cohort
|
128
|
92
|
53/2
|
70*
|
36
|
0
|
81
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Massarweh A
|
Israel
|
Prospective Cohort
|
180
|
102
|
57
|
66 [56-72]*
|
78
|
32
|
62 [49-70]*
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Agbarya A
|
Israel
|
Cross-sectional
|
355
|
140
|
54
|
65.3
|
215
|
37.2
|
62.5
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Herishanu Y
|
Israel
|
Prospective Cohort
|
219
|
167
|
67.1
|
71 [63-76]*
|
52
|
-
|
69 [63-73.7]*
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Iacono D
|
Italy
|
Cross-sectional
|
108
|
36
|
41.6
|
82*
|
72
|
-
|
≥66
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Malard F
|
France
|
Retrospective Cohort
|
225
|
195
|
60
|
68.9*
|
30
|
-
|
-
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Eliakim-Raz N
|
Israel
|
Prospective Cohort
|
161
|
95
|
58
|
65 [56-72]*
|
66
|
32
|
62 [50-70]*
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Herzog Tzarfati K
|
Israel
|
Prospective Cohort
|
423
|
315
|
56
|
71 [61-78]*
|
108
|
44
|
69 [58-74]*
|
Malignancy
|
BNT162b2 (Pfizer/BionTech)
|
Reuken P
|
Germany
|
Prospective Cohort
|
55
|
28
|
46.4
|
42 [36–59]*
|
27
|
-
|
-
|
Autoimmune
|
BNT162b2 (Pfizer/BionTech)
|
Geisen UM
|
Germany
|
Retrospective Cohort
|
68
|
42
|
35.7
|
50.5
|
26
|
30.8
|
37.5
|
Autoimmune
|
mRNA-1273 (Moderna) or BNT162b2 (Pfizer/BionTech)
|
Furer V
|
Israel
|
Prospective Cohort
|
807
|
686
|
30.7
|
59 [19-88]*
|
121
|
35
|
50*
|
Autoimmune
|
BNT162b2 (Pfizer/BionTech)
|
Prendecki M
|
UK
|
Prospective Cohort
|
155
|
85
|
52.1
|
52 [39.9-63.9]*
|
70
|
-
|
41.4*
|
Autoimmune
|
BNT162b2 (Pfizer/BionTech)
|
*: Median [IQR] is reported; otherwise the mean is reported.
|
Quality assessment of included studies
Quality assessment of the included studies is presented in Supplementary Table S1. The majority of the studies (n = 18) (16-18, 21-23, 25, 27, 28, 30, 31, 33, 34, 37-41) were of good quality and 8 (19, 20, 24, 26, 29, 32, 35, 36) had fair quality.
Seroconversion in immunocompromised patients vs. controls
Meta-analysis of 26 studies revealed that the risk of positive seroconversion in IC patients were 48% lower than healthy controls. (RR= 0.52; 95% CI: 0.42, 0.65; P < 0.01). Subgroup meta-analysis based on type of IC (i.e. autoimmune, transplant, and malignancy), revealed a statistically significant between-group difference (P < 0.01) (Figure 2). When comparing each two subtypes of immunodeficiency, the results showed that IC patients due to transplant were less likely to develop positive seroconversion than IC patients due to autoimmune disorder (P < 0.01) as well as IC patients due to malignancy (P < 0.01). There was no statistically significant difference in seroconversion between IC patients with autoimmune disorder and those with malignancy (P= 0.19).
Seroconversion in patients with autoimmune disease vs. controls
Four (20, 21, 34, 37) of the included studies were conducted on IC patients with autoimmune immunodeficiency. Although the proportion of positive seroconversion in these patients was lower than the controls, the pooled analysis showed no statistically significant difference in relative risk of seroconversion between two groups. (RR= 0.87; 95% CI: 0.75, 1.01; P = 0.07) (Figure 2).
Seroconversion in patients with malignancy vs. controls
Meta-analysis of 9 studies (16, 18, 23, 24, 26, 28, 30, 31, 33) revealed IC patients with malignancy were 0.75 times as likely to seroconvert than healthy controls (RR = 0.75; 95% CI: 0.63, 0.89; P < 0.01). Subgroup meta-analysis was conducted based on type of malignancy (hematologic vs. solid organ). Four (23, 24, 28, 33) of the studies were on patients with hematologic malignancy and three (16, 18, 30) were on patients with solid organ malignancy. The relative risk of seroconversion among IC patients with solid organ was significantly higher than those with hematologic malignancies (RR= 0.88; 95% CI: 0.85, 0.92 vs. RR= 0.61; 95% CI: 0.44, 0.86; P = 0.0.03) (Figure 3).
Seroconversion in transplant patients vs. controls
Of the included studies, 13 (17, 19, 22, 25, 27, 29, 32, 35, 36, 38-41) were on IC patients with transplant. The meta-analysis showed transplant patients were 67% less likely to develop seroconversion than controls (RR= 0.33; 95% CI: 0.24, 0.47; P < 0.01). Seven (19, 22, 25, 27, 38, 39, 41) of the included studies were on patients with kidney transplant and the remaining (17, 29, 32, 35, 36, 40) were on patients with different transplants; none of which with more than three studies to be separated in the subgroup analysis. Hence, subgroup meta-analysis was conducted based on type of transplantation (kidney vs. others). The analysis did not reveal any statistically significant difference in relative risk of seroconversion in patients with kidney transplant compared to other types of transplants (RR = 0.30; 95% CI: 0.20, 0.47) vs. RR = 0.38; 95% CI: 0.21, 0.66; P = 0.55) (Figure 4).
Publication bias
Funnel plot for seroconversion was asymmetrical and Egger test showed statistically significant evidence of publication bias (P < 0.01, z = -9.09). Trim and fill method was used to adjust the effect size (Pooled estimate = 0.87; 95% CI: 0.85, 0.88; number of studies = 84) (Figure 5).