Dear Editor,
In a representative cohort of immunocompromised young people, we observed marked differences in humoral and cellular responses to a standard two‐dose mRNA vaccination regime between disease categories. These differences are only partially explained by prevailing immunosuppression, which highlights the limitations of the current approach in selecting young patients for additional vaccine doses based primarily on the use of immunosuppressive medication.
Immunocompromised adults are recommended to receive a third primary dose of SARS‐CoV‐2 mRNA vaccines to promote seroconversion, and more recently a second booster dose. 1 , 2 There is, however, a paucity of vaccine response data from immunocompromised young people to justify this approach. We characterised the humoral and cellular responses to a standard two‐dose regime of SARS‐CoV‐2 mRNA vaccines in a single‐centre cohort of immunocompromised young people with a representative range of medical conditions and subsequently identified risk factors for an attenuated response.
A total of 97 immunocompromised patients and 22 controls were prospectively recruited. Six participants did not complete the study (Figure S1) and were excluded. The immunocompromised cohort included patients with inflammatory bowel disease (IBD), kidney transplant (KTX), liver transplant (LTX), idiopathic nephrotic syndrome (INS), kidney failure (ESKD), IgA nephropathy/vasculitis (IGA), juvenile idiopathic arthritis (JIA), systemic lupus erythematosus (SLE) and idiopathic uveitis (UVI).
Baseline characteristics (Table 1, Figure 1A) were comparable between controls and the immunocompromised cohort, and disease groups were homogenous in terms of age and dose interval (p > .05) which have been reported to affect antibody titres.
TABLE 1.
Baseline characteristics, immunosuppressive medication use, humoral and cellular vaccine responses in controls and immunocompromised young people
| Controls | All immunocompromised | IBD | LTX | KTX | INS | ESKD | IGA | JIA | SLE | UVI | Others | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N | 20 | 93 | 12 | 8 | 12 | 12 | 11 | 7 | 8 | 12 | 4 | 7 |
| Baseline characteristicsa | ||||||||||||
| Age (years) | 17.10 ± .66 | 17.58 ± .35 | 16.71 ± .89 | 17.28 ± 1.35 | 18.55 ± 1.03 | 17.43 ± .99 | 17.82 ± 1.12 | 18.88 ± 1.51 | 15.72 ± .9 | 18.50 ± .93 | 16.91 ± 1.43 | 17.29 ± 1.29 |
| Female | 7/20 (35) | 47/93 (51) | 8/12 (67) | 5/8 (63) | 6/12 (50) | 4/12 (33) | 1/11 (9) | 5/7 (71) | 3/8 (38) | 11/12 (92) | 2/4 (50) | 2/7 (29) |
| Pfizer | 18/20 (90) | 91/93 (98) | 11/12 (92) | 7/8 (89) | 12/12 (100) | 12/12 (100) | 11/11 (100) | 7/7 (100) | 8/8 (100) | 12/12 (100) | 4/4 (100) | 7/7 (100) |
| Days between doses 1 and 2 | 31.70 ± 1.75 | 35.28 ± .93 | 32.67 ± 1.51 | 41.63 ± 4.30 | 39.33 ± 1.48 | 33.25 ± 2.07 | 41.09 ± 2.00 | 30.71 ± 4.3 | 38.63 ± 3.22 | 27.33 ± 1.17 | 32.50 ± 4.33 | 35.86 ± 4.69 |
| Days after dose 2 | 38.75 ± 3.22 | 39.42 ± 1.36 | 40.92 ± 3.69 | 46.13 ± 3.36 | 37.67 ± 3.19 | 36.08 ± 3.38 | 38.82 ± 4.12 | 34.86 ± 3.53 | 37.63 ± 5.92 | 38.00 ± 3.75 | 42.50 ± 8.26 | 46.14 ± 7.40 |
| Immunosuppressive medications | ||||||||||||
| Steroids | 0/20 (0) | 26/93 (28)** | 0/12 (0) | 2/8 (25) | 12/12 (100) | 1/12 (8) | 0/11 (0) | 4/7 (58) | 0/8 (0) | 4/12 (33) | 0/4 (0) | 3/7 (43) |
| Anti‐metabolites | 0/20 (0) | 64/93 (69)*** | 11/12 (92) | 3/8 (38) | 10/12 (83) | 11/12 (92) | 0/11 (0) | 7/7 (100) | 4/8 (50) | 12/12 (100) | 2/4 (50) | 4/7 (57) |
| Calcineurin inhibitors | 0/20 (0) | 28/93 (30)** | 0/12 (0) | 7/8 (88) | 12/12 (100) | 5/12 (42) | 0/11 (0) | 1/7 (14) | 0/8 (0) | 0/12 (0) | 0/4 (0) | 3/7 (43) |
| Biologics | 0/20 (0) | 16/93 (17) | 4/12 (33) | 0/8 (0) | 0/12 (0) | 0/12 (0) | 0/11 (0) | 0/7 (0) | 6/8 (75) | 0/12 (0) | 3/4 (75) | 3/7 (43) |
| Humoral response to vaccination | ||||||||||||
| Anti‐S > 250 U/ml | 20/20 (100) | 68/93 (73)** | 10/12 (83) | 8/8 (100) | 6/12 (50) | 5/12 (42) | 11/11 (100) | 2/7 (29) | 7/8 (88) | 8/12 (67) | 4/4 (100) | 7/7 (100) |
| Anti‐S > 100 U/ml | 20/20 (100) | 74/93 (80)* | 12/12 (100) | 8/8 (100) | 9/12 (75) | 5/12 (42) | 11/11 (100) | 2/7 (29) | 7/8 (88) | 9/12 (75) | 4/4 (100) | 7/7 (100) |
| Anti‐S > .8 U/ml | 20/20 (100) | 86/93 (92) | 12/12 (100) | 8/8 (100) | 10/12 (83) | 10/12 (83) | 11/11 (100) | 5/7 (71) | 8/8 (100) | 11/12 (92) | 4/4 (100) | 7/7 (100) |
| Cellular response to vaccination | ||||||||||||
| Pan‐T‐cell IFNγ response (IU/ml) | 1.75 (1.43–2.12) | .71 (.59–.87)* | .55 (.29–1.06) | .45 (.17–1.24) | .17 (.09–.33) | .53 (.34–.82) | 1.54 (.99–2.38) | .75 (.46–1.23) | 1.72 (1.20–2.46) | 1.99 (1.30–3.05) | .58 (.32–1.03) | .70 (.41–1.21) |
| Pan‐T‐cell responder | 19/20 (95) | 66/93 (71)* | 10/12 (83) | 6/8 (75) | 5/12 (42) | 8/12 (67) | 8/11 (73) | 5/7 (71) | 7/8 (88) | 11/12 (92) | 2/4 (50) | 4/7 (57) |
| CD4+ T‐cell IFNγ response (IU/ml) | .99 (.81–1.22) | .32 (.25–.42)* | .2 (.08–.48) | .28 (.11–.73) | .04 (.02–.10) | .17 (.08–.40) | .89 (.57–1.37) | .4 (.24–.66) | 1.3 (.86–1.98) | .84 (.42–1.70) | .35 (.23–.53) | .44 (.21–.92) |
| CD4+ T‐cell responder | 19/20 (95) | 72/93 (77) | 9/12 (75) | 6/8 (75) | 4/12 (33) | 9/12 (75) | 11/11 (100) | 6/7 (86) | 7/8 (88) | 11/12 (92) | 4/4 (100) | 5/7 (71) |
Note: ‘Others’ consist of participants with atopic dermatitis (N = 2), corneal transplants (N = 2) as well as juvenile spondyloarthropathy, juvenile dermatomyositis and membranous glomerulopathy. Biologics used include Adalimumab (N = 10), Infliximab (N = 2), Tocilizumab (N = 2) and Dupilumab (N = 2).
Abbreviations: ESKD, end‐stage kidney disease; IBD, inflammatory bowel disease; IGA, IgA nephropathy and IgA vasculitis; INS, idiopathic nephrotic syndrome; JIA, juvenile idiopathic arthritis; KTX, kidney transplant; LTX, liver transplant; SLE, systemic lupus erythematosus; UVI, idiopathic uveitis.
Frequency data are given as N/N (%). Summary data are given as mean ± SEM, except for IFNγ responses for which the geometric mean (SEM) is provided.
*, **, ***Refer to significant differences between controls and the immunocompromised cohort with p < .05, .01 and .001, respectively.
FIGURE 1.
(A) Immunosuppression of immunocompromised participants at time of vaccination. Humoral and cellular responses to vaccination comparing (B) between controls and immunocompromised participants, and (C) within immunocompromised participants across disease categories. *, ** Refer to significant differences between controls and the immunocompromised cohort with p < .05 and .01, respectively. †, ‡ Refer to BH p = .0009 and p = .042, respectively, comparing between immunocompromised disease groups. Error bars are geometric mean (SEM).
Following vaccination, only 68 (73%) patients had anti‐spike (Anti‐S) protein titres >250 U/ml, compared to 20 (100%) controls (p = .006) (Table 1, Figure 1B). Similarly, patients had reduced pan‐T‐cell (.71 [.59–.87] vs. 1.75 [1.43–2.12] IU/ml, p = .041) and CD4+ T‐cell responses (.32 [.25–.42] vs. .99 [.81–1.22] IU/ml, p = .048) to SARS‐CoV‐2 spike protein (Table 1, Figure 1B).
Using a definition of non‐responders as those with T‐cell responses two standard deviations below the mean for controls, patients had a lower proportion of pan‐T‐cell responders compared to controls (71% vs. 95%, p = .023), but no significant difference in CD4+ T‐cell responders (77% vs. 95%, p = .12).
Within immunocompromised patients, there were significant differences between disease groups in the proportion of patients with titers >250 U/ml (p < .001) (Table 1, Figures 1C and S2). Compared to immunocompromised patients as a whole, ESKD patients had a superior humoral response with all having titers >250 U/ml (Benjamini–Hochberg [BH] p‐value = .042), whereas IGA patients had a poorer humoral response with 29% having titers >250 U/ml (BH p‐value = .042). There were also significant differences between disease categories in terms of CD4+ T‐cell responders (p = .017) but not pan‐T‐cell responders (p = .266). Specifically, the proportion of CD4+ T‐cell responders in KTX patients was reduced compared to immunocompromised patients taken together (BH p‐value = .0009).
To examine if the differences in vaccine responses could be explained by the different immunosuppressive medications, we performed a multivariable analysis (Table 2). Patients using corticosteroids (OR = .142 [95% CI: .033–.620], p = .009) and anti‐metabolites (OR = .083 [95% CI: .009–.781], p = .030) were less likely to have anti‐S titers >250 U/ml, whereas patients using calcineurin inhibitors had reduced pan‐T‐cell (−.641 [95% CI: −1.058 to −.223] log IU/ml, p = .003) and CD4+ T‐cell responses (‐.742 [95% CI: −1.290 to −.194] log IU/ml, p = .009).
TABLE 2.
Multivariable analysis of clinical risk factors affecting vaccine responses
| Anti‐S > 250 U/ml | Pan‐T‐cell IFNγ response (log IU/ml) | CD4+ T‐cell IFNγ response (log IU/ml) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Odds ratio | 95% CI | p‐Value | Regression coefficient | 95% CI | p‐Value | Regression coefficient | 95% CI | p‐Value | |
| Female | 1.878 | .528–6.678 | .330 | −.183 | −.537 to .171 | .306 | −.265 | −.73 to .199 | .259 |
| Pfizer | 0 | .999 | −.025 | −1.263 to 1.213 | .968 | .133 | −1.492 to 1.759 | .871 | |
| Age (years) | .944 | .800–1.114 | .497 | −.042 | −.091 to .008 | .101 | −.047 | −.112 to .019 | .158 |
| Days between doses | 1.035 | .957–1.119 | .389 | −.003 | −.025 to .018 | .762 | .011 | −.017 to .04 | .440 |
| Days from dose 2 | 1.020 | .978–1.064 | .354 | −.005 | −.018 to .007 | .393 | −.007 | −.023 to .009 | .377 |
| Corticosteroids | .142 | .033–.620 | .009** | −.148 | −.601 to .306 | .519 | −.385 | −.98 to .211 | .203 |
| Anti‐metabolites | .083 | .009–.781 | .030* | .010 | −.42 to .441 | .962 | .185 | −.381 to .75 | .518 |
| Calcineurin inhibitors | .861 | .204–3.634 | .839 | −.641 | −1.058 to −.223 | .003** | −.742 | −1.29 to −.194 | .009** |
| Biologics | .718 | .108–4.786 | .732 | −.367 | −.836 to .103 | .124 | −.284 | −.9 to .333 | .362 |
*, **Refer to significant differences between controls and the immunocompromised cohort with p < .05 and .01, respectively.
To determine if there was a residual effect of disease category, as disease category and immunosuppressive medications could not be included in the same statistical model due to collinearity (Table S1), we performed a stratified analysis on 28 participants only on anti‐metabolite monotherapy (Table 3). There remained significant differences in pan‐T‐cell responses between categories (p = .036), with post hoc testing revealing lower responses in IBD patients compared to SLE patients (.767 [.516, 1.141] vs. 3.97 [2.878–5.478] IU/ml, p = .039). There were no significant differences in CD4+ T‐cell (p = .093) or antibody (p = .214) responses between categories.
TABLE 3.
Humoral and cellular responses to vaccination amongst participants on anti‐metabolite monotherapy
| IBD | INS | IGA | JIA | SLE | |
|---|---|---|---|---|---|
| N | 8 | 7 | 3 | 2 | 8 |
| Humoral response to vaccination a | |||||
| Anti‐S > 250 U/ml | 7 (88) | 4 (57) | 2 (67) | 2 (100) | 8 (100) |
| Anti‐S > 100 U/ml | 8 (100) | 4 (57) | 2 (67) | 2 (100) | 8 (100) |
| Anti‐S > .8 U/ml | 8 (100) | 7 (100) | 2 (67) | 2 (100) | 8 (100) |
| Cellular response to vaccination | |||||
| Pan‐T‐cell IFNγ response (IU/ml) | .767 (.516, 1.141) | .892 (.606, 1.314) | 1.189 (.48, 2.947) | 3.002 (1.749, 5.152) | 3.97 (2.878, 5.478) |
| Pan‐T‐cell responder | 7 (88) | 6 (86) | 3 (100) | 2 (100) | 8 (100) |
| CD4+ T‐cell IFNγ response (IU/ml) | .239 (.098, .584) | .499 (.327, .761) | .676 (.263, 1.735) | 2.579 (1.572, 4.231) | 2.38 (1.699, 3.334) |
| CD4+ T‐cell responder | 6 (75) | 6 (86) | 3 (100) | 2 (100) | 8 (100) |
Frequency data are given as N (%). IFNγ responses are given as the geometric mean (SEM).
Finally, we examined the relationship between vaccine adverse effects and immunogenicity (Table S2). Amongst immunocompromised patients, an adequate pan‐T‐cell response was associated with a decreased prevalence of local side effects (55% vs. 82%, p = .018).
The key finding of this study is that within our immunocompromised cohort of young people, there are marked differences in vaccine responses between patients with different underlying conditions. For instance, all LTX patients had anti‐S titers >250 U/ml, compared to only 29% of IGA patients. This was partly associated with the use of immunosuppressive medication, with corticosteroid and anti‐metabolite use being associated with reduced humoral responses, and calcineurin inhibitor use with reduced T‐cell responses, consistent with existing adult data. 3 , 4 , 5 , 6
However, differing patterns of immunosuppression use do not fully explain the different vaccine responses between disease categories, as there remained significant differences in T‐cell responses even after restricting the analysis to patients on anti‐metabolite monotherapy. This may be due to the immune dysregulation intrinsic to certain conditions. For instance, in IBD, peripheral blood mononuclear cell hyporesponsiveness has been described 7 as well as therapy‐independent attenuation of responses to various vaccines. 8
Currently, additional SARS‐CoV‐2 mRNA vaccine doses are prioritised for young patients deemed to have an equivalent level of immunosuppression to solid organ recipients, and various public health bodies have published criteria to define this operationally. 1 , 2 In the non‐oncological setting, this would include many young people on various immunosuppressive medications, such as those in our cohort. In this context, our data revealed two main limitations of such an approach. First, given the heterogeneity in vaccine responses, some patient groups who qualify for an additional primary vaccination dose, for example JIA patients, may not benefit significantly, resulting in suboptimal use of vaccines and vaccination capacity. Second, the finding of disease‐specific, immunosuppression‐independent effects on vaccine responses suggest that young patients with mild disease on minimal immunosuppression, for example ulcerative colitis on rectal mesalazine, may also benefit from additional doses of vaccines which they currently do not qualify for. Therefore, further studies defining baseline immunological predictors of vaccine response are urgently required to further inform and optimise vaccination strategies, particularly in this vulnerable patient population.
In conclusion, although immunocompromised young people display attenuated humoral and cellular responses to vaccination in general, there are marked differences in vaccine responses between disease subgroups, not completely explained by immunosuppression used. A more differentiated, evidence‐based framework in determining patient subgroups which will benefit most from additional vaccine doses is required.
CONFLICTS OF INTEREST
There were no potential conflicts of interest to disclose.
Supporting information
Supporting Information
ACKNOWLEDGEMENTS
This study was funded by Paediatrics Renal Division Fund and Biomedical/Immunology Division Fund. The funding sources played no role in the study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.
Liangjian Lu, Chang Yien Chan and Pauline PL Chan‐Ng contributed equally to this manuscript.
REFERENCES
- 1. Mbaeyi S, Oliver SE, Collins JP, et al. The advisory committee on immunization practices’ interim recommendations for additional primary and booster doses of Covid‐19 vaccines — United States, 2021. Morb Mortal Wkly Rep. 2021;70(2021):1545‐1552. doi: 10.15585/mmwr.mm7044e2external. November 5, 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. David Salisbury MR, Karen N, eds. Immunisation Against Infectious Disease (The Green Book). The Stationery Office; 2021. https://www.gov.uk/government/collections/immunisation‐against‐infectious‐disease‐the‐green‐book [Google Scholar]
- 3. Crespo M, Barrilado‐Jackson A, Padilla E, et al. Negative immune responses to two‐dose mRNA COVID‐19 vaccines in renal allograft recipients assessed with simple antibody and interferon gamma release assay cellular monitoring. Am J Transplant. 2021;22:786‐800. doi: 10.1111/ajt.16854. Sep 22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Deepak P, Kim W, Paley MA, et al. Effect of immunosuppression on the immunogenicity of mRNA vaccines to SARS‐CoV‐2: a prospective cohort study. Ann Intern Med. 2021;174(11):1572‐1585. doi: 10.7326/M21-1757. Nov. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Moyon Q, Sterlin D, Miyara M, et al. BNT162b2 vaccine‐induced humoral and cellular responses against SARS‐CoV‐2 variants in systemic lupus erythematosus. Ann Rheum Dis. 2022;81(4):575‐583. doi: 10.1136/annrheumdis-2021-221097. Oct 4. [CrossRef] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Rabinowich L, Grupper A, Baruch R, et al. Low immunogenicity to SARS‐CoV‐2 vaccination among liver transplant recipients. J Hepatol. 2021;75(2):435‐438. doi: 10.1016/j.jhep.2021.04.020. Aug. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. MacDermott RP, Bragdon MJ, Kodner IJ, Bertovich MJ. Deficient cell‐mediated cytotoxicity and hyporesponsiveness to interferon and mitogenic lectin activation by inflammatory bowel disease peripheral blood and intestinal mononuclear cells. Gastroenterology. 1986;90(1):6‐11. doi: 10.1016/0016-5085(86)90067-3. Jan. [DOI] [PubMed] [Google Scholar]
- 8. Wellens J, Colombel JF, Satsangi JJ, Wong SY. SARS‐CoV‐2 vaccination in IBD: past lessons, current evidence, and future challenges. J Crohns Colitis. 2021;15(8):1376‐1386. doi: 10.1093/ecco-jcc/jjab046. Aug 2. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information