The course of COVID-19 in patients with IEI and CVID has been reported in several studies (17–27). Study designs vary and overall mortality of COVID-19 in CVID is difficult to assess, but a recent review by Tangye et al. estimated the case fatality rate in CVID to 7% as compared to 1-4% in the general population (28). Notably, known risk factors for severe COVID-19 as age, obesity and chronic lung disease are of significant importance also in CVID (19, 22–24). Lymphopenia is another known risk factor for severe COVID-19 in the general population and has been associated with a worse outcome also in CVID (19, 24, 29). Patients with Goods syndrome and untreated Severe Combined Immunodeficiency (SCID) seem particularly vulnerable to COVID-19 suggesting a particularly important role for T-cells when the humoral immune response fails. On the other hand, patients with pure B-cell deficiency in the form of X-linked agammaglobulinemia (XLA) have a risk of severe Covid-19 comparable to CVID (28, 30, 31). Clinical data thus present conflicting evidence regarding the potential role of cellular immunity in controlling COVID-19 in CVID.

The role of T-cells in COVID-19 in CVID patients has been studied in several articles (1, 3, 5, 6). Kinoshita et al. report findings in five unvaccinated COVID-19 patients with primary antibody deficiencies (PAD) (6). Three of these patients had CVID with mild COVID-19 and were sampled 30-80 days after a positive SARS-CoV-2 test. The patients had CD4+ T-cell responses similar to healthy COVID-19 convalescents as assessed by cytokine production in spike, membrane and nucleocapsid protein stimulated peripheral blood mononuclear cells (PBMC) cultures.

The Hanitsch group has presented their work on T-cell responses to COVID-19 in CVID in two articles. In the first one, analyzing activation markers and cytokine production in stimulated PBMC cultures, Steiner et al. show that of 11 uninfected CVID-patients with no lymphopenia, seven had reactive CD4+ T-cells against SARS-CoV-2 spike proteins but none against nucleocapsid protein (3). Notably, four of these seven patients also had reactive CD4+ T-cells against spike protein from endemic coronaviruses. Healthy controls, both exposed and unexposed to COVID-19, had a higher frequency of all these reactive T-cell populations. The authors found a significant correlation between reactivity to SARS-CoV-2 and endemic coronaviruses in unexposed healthy controls, suggesting the possibility of cross-reactive T-cells.

In a second paper, Steiner et al. demonstrate strong T-cell responses in three unvaccinated CVID-patients with severe COVID-19 (1). The CVID patients were sampled 24-40 days after the debut of symptoms. All CVID-patients had CD4+ T-cells reactive to spike and nucleocapsid protein. Notably, frequencies of reactive T-cells were significantly higher in CVID-patients than convalescent healthy controls, but these controls had mild disease. Similar to their previous report, reactive CD4+ T-cells produced cytokines IFN-γ, TNF and IL-2 but here the strongest response was seen in CVID-patients. None of these CVID-patients with severe COVID-19 showed a serologic response. The relationship between serologic^nbsp;and cellular responses has also been studied in a different publication from the same group (2). In a group of immunocompetent convalescents of mild COVID-19 infection, CD4+ and CD8+ T-cells reactive to spike and nucleocapsid proteins of SARS-CoV-2 were found both in seropositive and seronegative patients. While similar reactivity also was found in healthy unexposed controls there was a significantly increased cytokine production in both convalescent groups. The authors conclude that there is evidence for specific T-cell responses in COVID-19.

In a study by Oksenhendler et al. of infections in 252 patients with CVID, 16 (6%) patients presented with opportunistic infections like CMV or Pneumocystis jirovecci pneumonia (PCP) but we do not know the CD4 levels of the afflicted patients (33). In a seminal paper from the same group, Malphettes et al, reported that 28 of 313 CVID patients presented with opportunistic infections or CD4 <200 and termed these patients late combined immunodeficiency (LOCID) (36). A similar finding has been done by Cunningham-Rundles et al. where 5% out of 248 patients studied with CVID had CD4 cell counts lower than 200×10⁶ cells/L (32). Five of these lymphopenic patients presented with opportunistic infections, and 55 patients in the same study had mild or moderate infections that also pointed towards a T cell defect. Resnick et al. reported a higher number of opportunistic infections (15,4%) in a study of 473 CVID patients followed over four decades, but lymphopenic patients could have been included in this cohort (37).

CMV related disease occurs almost exclusively in patients with impaired cellular immunity. Kralickova et al. studied 32 patients to determine the frequency of CMV-related disease in CVID patients (38). Symptomatic CMV infection was documented in three CVID patients, where two patients were diagnosed with CMV pneumonia, and two patients suffered from CMV enteritis (one patient had both). None of these patients had severe abnormalities in T-cell subpopulations and all had CD4 ≥200 at the time of CVID diagnosis. However, all three patients had various other comorbidities, and these findings are in line with the studies by Cunningham-Rundles et al. and Resnick et al.

The first study presenting data on vaccine efficacy in IEI was authored by Hagin et al. using a spike protein stimulated IL-2/IFN-γ ELISPOT assay to evaluate cellular responses at week five after start of vaccination (47). In a group of 26 patients with IEI that included 13 patients with CVID, 19 patients had a positive cellular response similar to healthy controls. Notably, all controls had a cellular response to vaccination whereas four of the patients with CVID were non-responders. These patients did have a humoral response to vaccination with anti-S antibody levels that were variable but blunted. Interestingly, patients with XLA had a stronger cellular response than both healthy controls and the rest of the IEI group.

In 41 patients with CVID that received basal mRNA vaccination, Salinas et al. found no increase in spike protein induced IFN-γ producing cells in patients as assessed by an ELISPOT assay four weeks after the start of vaccination, and found defective cellular response in 30% of patients (40). In contrast, all six XLA-patients did respond. The CVID-patients again had a significant but blunted humoral response to vaccination as compared to healthy controls.

Antoli et al. demonstrated an impaired cellular response to basal vaccination in a group of 28 CVID patients (27 mRNA, 1 viral vector vaccine) sampled four weeks after the second dose (43). Patients had significantly lower IFN-γ levels in a SARS-CoV-2 specific IGRA-test compared to healthy controls, and 8 patients had no response. The authors identified several predictors for poor humoral responses, including an inverted CD4/CD8 ratio, but these predictors were not associated with an impaired cellular response.

In a study looking at the effect of basal vaccination in 18 patients with CVID (17 mRNA, 1 viral vector vaccine), Arroyo-Sánchez et al. report 83% positive response four weeks after the second dose using a spike protein stimulated IFN-g ELISPOT assay (42). Again, this was significantly lower than the 98% positive cellular response observed among healthy controls, and likewise 15 of 18 CVID-patients were anti-S1 IgG positive compared to 50 of 50 healthy controls. Three patients had no cellular response to vaccination but two of them did mount an antibody response. Notably, CVID patients with an infection only phenotype seemed to have a better humoral and cellular response to basal COVID-19 vaccination than patients with inflammatory complications.

Bergman and colleagues present their findings of COVID-19 vaccine responses in a large cohort of patients with IEI in two articles (41, 44). In their second article they show that in 11 patients with CVID and a group of healthy controls, there was a significant and comparable cellular response at week five after the start of basal vaccination as assessed by an IFN-γ ELISPOT assay. There was also an increase in spike protein activated CD4+CD69+CD154+ T-cells in CVID-patients and controls after vaccination, and here response was impaired in patients as compared to controls. The antibody response in CVID was also impaired with anti-spike IgG detected in 68 vs 100% of patients and controls, respectively. Interestingly, they found that low CD4+ T-cells and high CD21^Low B-cells were associated with a poor humoral response to vaccines, but there were no data on any relationship to the cellular response.

In a large cohort of over 500 patients with IEI that included 212 patients with CVID, van Leeuwen et al. looked at the effect of basal vaccination 8 weeks after the start of vaccination (49). Cellular responses were evaluated in a subgroup of the cohort (56 CVID patients) using spike protein IGRA and finding a significantly reduced proportion of responders among CVID patients (67%) versus healthy controls (88%). Notably, while CVID patients had impaired cellular responses compared to healthy controls this was not found in other groups of IEI like XLA and IgG specific antibody deficiency. Seroconversion among CVID patients were significantly reduced compared to healthy controls (81% vs 100%, patients and controls, respectively).

Looking at the long-term effect of basal vaccination, Kralickova et al. sampled a group of 46 patients with IEI that included 34 patients with CVID 6 and 16 weeks after the start of vaccination (51). Cellular responses were analyzed using spike protein IGRA-assays that stimulated CD4+ and CD4+/CD8+ T-cells, respectively. A positive CD4+ T-cell response was noted in 41% and 50% of CVID patients at week 6 and 16, respectively, while the combined CD4+/CD8+ assay showed a negative trend over time with 50% and 41% positive at week 6 and 16, respectively. Seven CVID patients had previously been treated with rituximab and notably only two of these patients were able to mount a cellular response at week six.

Milota et al. have also looked at the long-term cellular effect of basal vaccination in CVID following a group of 12 CVID patients over 6 months (52). Cellular responses were evaluated by flowcytometry of intracellular IL-2/IFN-γ/TNF in spike/nucleoprotein stimulated PBMCs. CVID patients had a similar proportion of spike protein reactive CD4+ T-cells as controls at one and six months after vaccination (46 vs 73%, and 50% vs 60%, respectively). Notably, anti-receptor binding domain (RBD) antibody were significantly and incrementally lower in CVID patients compared to controls at 1, 3 and 6 months.

In a study that included 9 CVID basal vaccine recipients and 15 CVID COVID-19 convalescents, Pulvirenti et al. showed that both groups had impaired T-cell responses as assessed by spike protein IGRA compared to healthy controls (53). Immunized patients were sampled at week 4 after the start of vaccination whereas convalescents were sampled at week 12 after first positive SARS-CoV-2 PCR. The same group further explored the role of T-cells in a different study that included 47 CVID patients with at third (booster) vaccine dose (54). Patients were sampled two weeks after the third dose and cellular response were measured in spike protein stimulated cultures of PBMC by staining for IFN-γ, TNF and CD40L. Notably, the spike protein induced expression of CD40L and TNF in CD4+ T-cells were significantly reduced in CVID patients compared to controls, while the expression of IFN-γ was similar. Likewise, even if CVID patients had increased levels of anti-S IgG after the third dose, they were still significantly lower than what was seen in healthy controls.

Ainsua-Enrich et al. report findings in a study that included 12 CVID-patients receiving three mRNA vaccine doses (55). All healthy controls had positive spike protein IFN-γ ELISPOT at week 8 and 24 after basal vaccination, while this was found in 67% and 33% of patients with CVID, respectively. After receiving a third mRNA vaccine dose, the proportion of positive CVID patients increased to approximately 70%. CVID-patients also had a significant but still impaired response in activated CD8+ T-cells compared to controls at week 8 and 28.

In an early study of 21 patients with IEI that included 14 CVID-patients, Amodio et al. found a significant increase in spike protein activated CD4+CD40L+ T-cells at week four after the start of basal vaccination (48). This increase was significantly lower than what was seen in healthy controls, and four of the 14 CVID patients did not respond. There were no significant differences in pre-vaccination T-cell subsets between cellular responders and non-responders in the CVID group. Three of the four cellular non-responders had a serologic response comparable to healthy controls.

As noted, Bergman and colleagues found an increase in spike protein activated CD4+CD69+CD154+ T-cells after basal vaccination in CVID-patients that was significantly lower than what was seen in healthy controls (41). In an in-depth article from the same vaccine study, Gao et al. present T-cell profiles of 48 patients with IEI, including CVID and XLA (56). Overall, they found that patients with IEI had a relatively high frequency of spike-specific polyfunctional CD4+ T-cells, while the frequency of spike-specific CD8+ T-cells were lower but comparable to healthy controls. Interestingly, they found that XLA patients had a strong cellular response despite their lack of humoral response, but similar data on CVID humoral non-responders are not presented.

Sauerwein et al. studied the response of CD4+ T-cells to basal mRNA vaccination in 31 CVID patients a median of 6 weeks after the second dose (46). Using a spike protein induced activation marker assay they found reduced activation of CD4+ memory T-cells and CD4+ circulating follicular T-cells in CVID patients vs healthy controls. Anti-spike IgG levels were also significantly lower in CVID patients compared to controls with 16 of 31 patients defined as non-responders. These non-responders were characterized by reduced proportion of CD4+ vaccine specific memory T-cells as compared to responders.

In an article on predictors of poor immune responses after basal mRNA vaccination, Shin et al. found that a group of 12 CVID patients had a delayed cellular response to vaccination as compared to healthy controls (50). The cellular response was evaluated using a spike protein induced activation marker assay identifying CD4+OX40 + 4-1BB+ effector memory T-cells.

The above-mentioned study by Sauerwein et al. included a spike-protein induced 3H-thymidine proliferation assay. The assay showed significantly reduced proliferation in a group of 16 vaccinated CVID patients as compared to 14 healthy vaccinated controls (46).

Summarizing the studies performed 4-8 weeks after the start of vaccination assays using ELISPOT, IGRA and flowcytometry of activation markers, all detect a significantly impaired response in CVID patients compared to healthy controls ( Figure 1 ).

CVID patients will typically have an impaired antibody response to vaccines containing T cell independent antigens like polysaccharides, but interestingly, pneumococcal vaccine induction of cytokines not related to B-cells are also found to be impaired (57–59). This is demonstrated in a study by Hong et al, where Pneumovax-23-induced secretion of IL-6 and TNF-alpha by monocytes was significantly lower in 14 patients with CVID as compared to controls (59).

CVID patients have an impaired humoral response to influenza vaccine, while data on cellular responses are variable. In their study of 8 CVID patients, Hanitsch et al. found that the influenza-specific antibody and T-cell cytokine responses (vaccine specific IFN-γ, TNF-α and IL-2) in the CVID group were similar to that of the healthy controls (60). A similar finding is described in case reports from Pedersen et al, while Friedmann et al. found that cellular immunity is preserved in most CVID patients after vaccination with the seasonal flu vaccine (61, 62). In contrast, van Assen et al. demonstrated a reduction in vaccine specific cytokine-producing CD4⁺ T cells in 15 CVID patients compared to healthy controls (63).