Most immune deficiencies are acquired, developing because of chronic disease, or immune suppressant therapies. A small proportion (~4%) have genetic origins, including genetic susceptibility to infectious triggers of autoimmune diseases. The vast majority of these inborn errors of immunity result in defects in antibody production (14). An extreme of this diverse group are agammaglobulinemic persons, who are otherwise healthy but lack all antibody production. Nonetheless, they can successfully mount T cell-based immune responses against many microbial infections (15).

Abnormalities in antibody production make standard measures of antibody titer difficult to interpret and of limited usefulness in this population. For example, analysis of antibody and cellular immunity responses after a single Pfizer-BioNTech SARS-CoV-2 vaccination in patients with multiple kinds of inborn errors of immunity revealed highly variable antibody responses that did not reach protective levels in a large fraction of subjects (16). In contrast, T cell responses were comparable between those with inborn immune errors, the healthy vaccinated, and healthy convalescent control populations. The best predictor of immune responsiveness was the cellular immune status prior to vaccination (16). Deficient antibody production after vaccination in subjects with primary antibody deficiencies was also reported by Selinas et al. (17), who further described atypical memory B cells but mostly intact T cell responses in these subjects. Increased memory B cell frequency was also associated with response to vaccination (18). Furthermore, a study of subjects with B cell compartments compromised either by inherited immune deficiencies, blood cancers, or therapy determined that the number of naïve B cells available was an independent predictor of successful vaccination (19). Taken together these findings suggest that exclusive reliance on antibody testing would likely underestimate vaccine responsiveness for a large segment of individuals with inherited immune deficiencies. A better characterization of the deficient B cells would improve our understanding of the impact of those deficiencies on vaccine provided protection while helping to develop new vaccine candidates for particularly susceptible populations.

Individuals with cancer generally have suppressed immune function as a result of the disease (20–24) and of the treatment (25–27). Unsurprisingly, cancer patients with SARS-CoV-2 infections have substantially reduced antibody responses, the vast majority of which did not generate sufficient humoral immunity (27–33).

Persons with hematologic cancers, however, that maintained CD8 (cytotoxic) T cell numbers above a defined threshold had better survival rates than those who could not maintain such levels (33). Thus, simply counting the CD8 T cells may identify most of the hematological cancer patients with the highest risk of severe COVID-19 infection and death. Including these measures as part of regular assessment may allow for preventative measures that improve survival. The authors concluded that boosting T cell immunity in cancer patients would substantially aid their ability to weather an infection (33). This interpretation was further supported by other studies of SARS-CoV-2 vaccine responses in patients with cancer (29, 30, 34). Most patients with impaired antibody responses generated protective levels of cellular immunity, suggesting that cellular immunity may be able to overcome the absence of an antibody response and clear the infection. For patients immunocompromised by cancer, relying on antibody testing alone will be unable to distinguish those who can respond to vaccination from those who cannot. Measurements of cellular immunity can provide those distinctions and help evaluate novel strategies to boost immunity in cancer patients.

Intentional or collateral immunosuppression may result from therapeutic treatments targeting a wide variety of conditions. Immunosuppressive therapies are a staple for transplant recipients and for those with autoimmune morbidities. Among all immune supressed individuals, transplant recipients produced the weakest response to vaccination (35, 36), consistent with the severe immune suppression intentionally induced to prevent transplant rejection. Researchers observed extremely low antibody responses following both SARS-CoV-2 infection (35) and vaccination (36–38). Cellular immune responses, however, impacted outcome severity (35), again illustrating the importance of cellular immunity in resolving SARS-CoV-2 infection in the immunocompromised. Testing for cellular immune responses in transplant recipients may enable clinicians to identify which subjects are at higher risk of not controlling an infection.

Many people receive immunosuppressive treatment for a large variety of autoimmune conditions. These include arthritis, inflammatory bowel disease, and multiple sclerosis, among many others (39, 40). In addition, systemic inflammation is known to suppress immune function (41, 42) and is associated with the development of heart disease, obesity, metabolic syndrome, and type 2 diabetes (41, 42). In total, there are potentially millions of people living with varying degrees of immune deficiency caused by these conditions, and these immune deficiencies may significantly affect the ability to produce humoral or cellular immune responses to vaccination. Despite being high risk for severe disease, these groups were underrepresented in first-generation COVID-19 vaccine trials. Conducting more comprehensive assessments of immune responsiveness that include assessments of cellular immunity would enable a more complete understanding of immune status and of the interventions necessary to establish immunity in these individuals.

For example, a study of COVID-19 vaccinated patients with psoriasis receiving immunosuppressant drugs (43) demonstrated variable, but overall poorer antibody responses compared to a control population. Their cellular immune responses, however, were at least as robust as the responses of the controls (43). This differential response to vaccination would not have been observed if only antibody titer measurements were used. Additionally, some treatment types appeared comparatively more immune sparing than others, which may become an important consideration when selecting therapeutic options. Thus, manufacturers of immune suppressants and treating physicians would benefit from a more complete evaluation of the impact immunosuppressant therapies have on SARS-CoV-2 vaccine efficacy. Including cellular immunity measurements along with antibody titer assessments will begin to fill those gaps.

Aging in healthy individuals is independently associated with a decline of immune responsiveness due to a reduction in the total number of immune cells available as well as a specific reduction in the ability to respond to novel antigens (44–46). Even healthy elderly have immune systems that differ significantly from younger adults (44–48) and typically require higher vaccine doses to elicit a sufficient immune response (44, 49–51). Chronic diseases, infections, and other underlying health conditions can combine with the effects of aging to further weaken the immune systems of elderly individuals. As a result, there is an increased vulnerability to external stressors, including infections (52) and a significantly higher risk of vaccine failure (53).

The effects of aging are not uniform. Recent studies define approximately half the over-55 population as immunologically frail (54, 55). Greater immunological frailty was associated with a poorer response to vaccination (55). Subsequent studies have further supported the association of frailty with poor immune responses (56), and diminished vaccination efficacy (56–58). In SARS-CoV-2 infection, frailty correlated with severe disease and mortality more than chronological age (59–61). However, the SARS-CoV-2 vaccine trials to date have excluded volunteers with severe comorbidities and conditions impacting immune functioning (62–64), even in vaccine efficacy studies in the elderly (49). Thus, vaccine efficacy in the frail elderly remains unknown. The elderly share with the other immunocompromised populations described here a significant unmet need for comprehensive evaluation of their immune systems to inform strategies to risk stratify patients and to increase vaccine response in those identified as high risk.

There are potentially many millions of individuals with at least partial immune deficiency. Approximately 4% of US residents have inherited or acquired immune deficiencies (53) and approximately 16% of the US population is over 65 years old. About 2.5% of the US population lived with cancer in 2020 (65), and millions receive immune suppressants for cancer or autoimmune diseases. Even with a significant overlap between these populations, millions of Americans are affected, rendering them at least partially immunocompromised. In persons younger than 65, at least several million may also be indirectly immunocompromised due to obesity, diabetes, and heart disease. Comparable populations are also expected in other developed societies. The size of these populations and the lack of understanding of cellular immune response to vaccination, emphasizes the urgency of the unmet need.

Immunocompromised persons are a significant segment of the population. They are also the most at risk for severe SARS-CoV-2 infection. By excluding these persons from clinical trials, particularly larger-scale phase 2 and 3 trials represent a potential source of observational bias and has led to large gaps in understanding of the true efficacy of vaccination. By including T cell evaluations in clinical development programs, immunocompromised persons no longer need to be excluded for fear of confounding study data. To the contrary, optimal vaccination strategies can be designed specifically to target these groups.

The generation of new viral variants of concern has been discussed extensively elsewhere (66). Viral variants have been repeatedly shown to reduce the effectiveness of neutralizing antibodies developed through vaccination with the prototype strain (67). The corresponding T cell responses, however, remained robust against the variants (67, 68), largely because T cell responses were less sensitive to the single amino acid mutations that define emerging variants. Moreover, T cells have been identified that target sections of the SARS-CoV-2 proteins which cannot mutate without compromising viral fitness (69–71). These T cells, if induced by vaccination or infection by the ancestral strain of SARS-CoV-2, will provide cross-protective benefits against many mutant variants. Indeed, similar cross-protection has been demonstrated with several other viral infections including influenza (72–74), as well as other anti-viral vaccines (75, 76), re-enforcing that this is by no means a COVID-19 specific effect.

T cell response to vaccination may provide the missing link between experimental testing and what is observed empirically in the real world. As mentioned above, highly infectious variants exhibit mutations that were demonstrated to provide immune escape from neutralizing antibodies induced by either infection with the ancestral strain or vaccination (77). Therefore, quantitative, and qualitative analysis of T cell responses induced by the different vaccine candidates should be considered when evaluating the utility of any vaccination strategy or regime, including the need of a third dose booster shoot.

By focusing on antibody testing in vaccine clinical trials, these robust cross-reactive T cell populations remain understudied, limiting our true understanding of vaccine efficacy. As an example, using flow cytometry several recent independent studies have confirmed that the vast majority (>90%) of the T cell response induced by two doses of mRNA vaccines were directed against conserved, “mutation-resistant” sections of the virus (78, 79). When single mutations do occur in these regions, they had limited-to-no impact on T cell activation. However, these studies were limited in size relative to large phase-III clinical trials and lack the regulatory rigor required to submit for vaccine licensing. T cell evaluations of this type must be taken from the academic realm and applied into the world of vaccine clinical trials.

Long COVID-19 is defined as sequelae that extend beyond four weeks after initial infection (80). Statistically significant risk factors included age, with the ≥50 age group having the greatest risk, and number of pre-existing medical conditions, in particular hypertension, obesity, psychiatric or immunosuppressive conditions (81). Chronic fatigue was the most frequently reported symptom following recovery from acute infection (82–85). Other frequently reported symptoms included dyspnea, cardiovascular abnormalities, cognitive impairment (“brain fog”), smell and taste dysfunction, and other less common symptoms (82, 86–93).

The mechanisms leading to long COVID-19 are not fully understood but may involve deficits in cellular immune function. A study that showed no difference in severity of initial infection between individuals with and without post-acute infection sequelae identified lower and more rapidly waning specific T cell responses in the latter population (94). In subjects experiencing neurological post-acute sequelae, no correlation between anti-SARS-CoV-2 T cell responses and IgG production could be found at 155-315 days post-infection, indicating that antibody responses cannot serve as a surrogate indicator of cell-mediated immunity in this group (95). This study also described aberrant T cell memory responses to vaccination in subjects with neurological long COVID-19 (95). The finding that some people experiencing long COVID symptoms notice an improvement after vaccination, and others a worsening of their symptoms supports the potential role of a heterogeneous immune response in the pathogenesis of long COVID (96). Taken together these observations underscore the need for independent monitoring of these two arms of the immune system when following long COVID-19 patients. While our understanding of long COVID-19 is still emerging, currently it appears that the elderly and immunocompromised individuals are most at risk, and that the cellular immune response may be disproportionally affected. Thus, like the other immunocompromised groups described above, long COVID-19 sufferers would also benefit from a comprehensive evaluation of their immune responses to vaccination.

Concurrent with the rise of more infectious viral variants, recommendations in favor of inoculations with more than one type of SARS-CoV-2 vaccine, and clinical studies evaluating the effectiveness of vaccine mixing and matching have been presented (97, 98), and many governments are considering mix-and-matched 3^rd boosters shots for deployment amongst their populations. The argument for comprehensive immune monitoring applies when considering mix-and-matching vaccination are similar to the arguments presented for viral variants or immunocompromised persons. Different SARS-CoV-2 vaccines have been shown to induce T cell responses that can differ in their magnitude, as well as the composition of the T cell response (99). T cell immunity may not be qualitatively the same in all mix-and-match vaccination situations. It cannot be predicted how different vaccines will complement or conflict with each other when used in any given mix-and-match vaccination strategy, nor can it be taken for granted that the quality or quantity of a T cell immune response can be accurately inferred by evaluating antibody responses alone. Therefore, when assessing which mix-and-match vaccination strategies are the best choice against all viral variants, T cell immune monitoring must be included in these clinical trials. This is particularly important when considering immunocompromised groups of persons, as discussed above.

Such an update to our research paradigm will lead to better-informed health policy decisions and establish more robust clinical decision criteria not only for this current pandemic, but for other contemporary vaccination campaigns, as well as future pandemic diseases. These new vaccination strategies could be tailored at the population level to optimize vaccine choices for particular age groups and/or disease states. At the most extreme, vaccination strategies could be personalized based on particular patient idiosyncrasies.