Intended for healthcare professionals

Clinical Review State of the Art Review

Vaccines for older adults

BMJ 2021; 372 doi: (Published 22 February 2021) Cite this as: BMJ 2021;372:n188
  1. Anthony L Cunningham, director1,
  2. Peter McIntyre, professor2,
  3. Kanta Subbarao, director3,
  4. Robert Booy, professorial fellow4 5,
  5. Myron J Levin, professor6
  1. 1Centre for Virus Research, The Westmead Institute for Medical Research, Faculty of Medicine and Health, University of Sydney, Australia
  2. 2Women’s and Children’s Health, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand
  3. 3WHO CollaboratingCentre for Reference and Research on Influenza and Department of Microbiology and Immunology, University of Melbourne at the Peter Doherty Institute for Infection and Immunity, Melbourne, Australia
  4. 4National Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases, The Children’s Hospital at Westmead, New South Wales, Australia
  5. 5Marie Bashir Institute for Infectious Diseases and Biosecurity, School of Biological Sciences and Sydney Medical School, University of Sydney, Australia
  6. 6Departments of Pediatrics and Medicine, University of Colorado School of Medicine Anschutz Medical Campus, Aurora, Colorado, USA
  1. Correspondence to A Cunningham tony.cunningham{at}


The proportion of the global population aged 65 and older is rapidly increasing. Infections in this age group, most recently with SARS-CoV-2, cause substantial morbidity and mortality. Major improvements have been made in vaccines for older people, either through the addition of novel adjuvants—as in the new recombinant zoster vaccine and an adjuvanted influenza vaccine—or by increasing antigen concentration, as in influenza vaccines. In this article we review improvements in immunization for the three most important vaccine preventable diseases of aging. The recombinant zoster vaccine has an efficacy of 90% that is minimally affected by the age of the person being vaccinated and persists for more than four years. Increasing antigen dose or inclusion of adjuvant has improved the immunogenicity of influenza vaccines in older adults, although the relative effectiveness of the enhanced influenza vaccines and the durability of the immune response are the focus of ongoing clinical trials. Conjugate and polysaccharide pneumococcal vaccines have similar efficacy against invasive pneumococcal disease and pneumococcal pneumonia caused by vaccine serotypes in older adults. Their relative value varies by setting, depending on the prevalence of vaccine serotypes, largely related to conjugate vaccine coverage in children. Improved efficacy will increase public confidence and uptake of these vaccines. Co-administration of these vaccines is feasible and important for maximal uptake in older people. Development of new vaccine platforms has accelerated following the arrival of SARS-CoV-2, and will likely result in new vaccines against other pathogens in the future.


Influenza, herpes zoster, and pneumococcal infections cause significant morbidity and—for influenza and pneumococcal disease—mortality, in aging people (fig 1, 2). Estimates of the burden of vaccine preventable diseases are similar in Australia,1 Europe,2 and Canada.3 Effective vaccines against influenza, pneumococcal infection, and herpes zoster have the greatest potential for preventing morbidity and mortality among people aged ≥65. Estimates from Australia of disability-adjusted life years lost annually are shown in figure 1, and all-age incidence and mortality in figure 2.

Fig 1
Fig 1

The effect of influenza, herpes zoster, and invasive pneumococcal disease (IPD) on disability-adjusted life years per 100 000 by age group. Data from Australian Burden of Disease study 2015

Fig 2
Fig 2

Cases, deaths, and population burden of vaccine preventable disease. Size of bubbles indicates population-level burden (DALY per 100 000 population). Data from Australian Burden of Disease study 2015

Vaccines protecting against these diseases have long been available, however, efficacy and persistence of immunity are suboptimal, especially in people of 70 and older. This can be improved, as was demonstrated by the efficacy (~90%) of the recombinant zoster vaccine (RZV). Improved influenza and pneumococcal vaccines are needed, since efficacy of seasonal influenza vaccine in people aged 65 and older is low. This review examines the evidence for current best use of vaccines against influenza, pneumococcal disease, and herpes zoster in adults of 60 and older, and identifies evidence gaps and potential directions for future research. Functional status (frailty) predisposes to infectious risk and vaccine responsiveness and may be more important than chronological age.4


Herpes zoster

Varicella zoster virus-specific cell mediated immunity declines with increasing age (immunosenescence)56 and immunecompromise,7 resulting in increased frequency and severity of herpes zoster infection.8910 People aged 85 have a 50% lifetime risk of infection. In the USA and Australia, two thirds of cases occur in people aged 50 or older.10


Influenza A and B viruses cause seasonal epidemics of influenza in humans. Currently, two influenza A virus subtypes (A/H1N1pdm09 and A/H3N2) and two influenza B lineage viruses (B/Victoria and B/Yamagata), are responsible individually or in combination for seasonal influenza epidemics. Seasonal influenza A and B viruses are constantly evolving in nature, often resulting in antigenic change or “drift.” The composition of influenza vaccines is updated annually to keep pace with antigenic drift.11 Adults aged 65 or older have the highest risk for hospitalization, complications, and death resulting from influenza,1213 and this risk increases markedly in people over 85.13 The gradual accumulation of underlying health conditions contributes to this risk. The estimated global mean annual influenza associated respiratory excess mortality is 0.1-6.4 per 100 000 people under 65 and 17.9 to 223.5 for people over 75. This wide variability is attributed to annual differences in circulating virus strains and the severity of associated illness.14 However, non-respiratory deaths are often not included.15

Pneumococcal disease

The age specific incidence of invasive pneumococcal disease (IPD) is U shaped, with peaks at <2 years and >85 years. In the UK16 and USA17 incidence increases with chronic medical conditions, especially immunocompromise. Among placebo recipients in the Netherlands’ CAPiTA vaccine trial, community acquired non-bacteremic pneumococcal pneumonia (NBPP) was identified five times more commonly than IPD in people aged 65 years or older.18 In a single UK region, NBPP progressively increased from 65 years of age, such that in people >85 years, incidence per 100 000 people was fivefold higher (349) compared with 65-74 years (60)19 and, similar to CAPiTA, was much higher than concurrent UK incidence estimates for IPD of 27.6 per 100 000 people of >65 years.20

Sources and selection criteria

References for this review were identified by searching Pubmed and OVID Medline, all for items in English published in peer reviewed journals between 2010 and July 2020. We prioritized randomized controlled trials, systematic reviews with meta-analyses, and large case series. The Medical Subject Heading (MeSH) terms used in the search included:

For herpes zoster: “Immunization”, “Immunization Programs”, “Vaccines”, “Herpes Zoster”, “Herpes Zoster Vaccine”, “Immunogenicity, Vaccine”, “Treatment Outcome”, “Safety-Based Drug Withdrawals”, “Product Surveillance”, “Drug Evaluation”, “Postmarketing Population Surveillance”, “Adverse Drug Reaction Reporting Systems”, and “Aged”.

For influenza: “Immunization”, “Immunization Programs”, “Vaccines”, “Influenza, Human”, “Influenza Vaccines”, “Immunogenicity, Vaccine”, “Treatment Outcome”, “Safety-Based Drug Withdrawals”, “Product Surveillance”, “Drug Evaluation”, “Postmarketing Population Surveillance”,“Adverse Drug Reaction Reporting Systems” , and “Aged”.

For pneumococcal disease:“Immunization”, “Immunization Programs”, “Vaccines”, “Pneumococcal Infections”, “Pneumococcal Vaccines”, “Immunogenicity, Vaccine”, “Treatment Outcome”, “Safety-Based Drug Withdrawals”, “Product Surveillance”, “Drug Evaluation”, “Postmarketing Population Surveillance”, “Adverse Drug Reaction Reporting Systems,” and “Aged”.

Equivalent text words were used to supplement the MeSH terms. The search was limited to results in high impact Abridged Index Medicus journals.

Prioritisation and inclusion/exclusion criteria: we focused on selection of moderate to high impact publications in the field. All searches were limited to moderate- high impact AIM (Abridged Index Medicus) journals. The pneumococcal and influenza searches also had review limits applied owing to concerns about the higher levels of retrieval for those topics.


Immunosenescence is the progressive decline in immunity with age, especially past the age of 60, that results in increased frequency and severity of common infectious diseases, including herpes zoster, influenza, and pneumococcal disease, increased prevalence of cancer and autoimmune diseases, and poor responses to immunization. Several immune modalities decline, creating deficits in innate immunity, such as in the function of natural killer cell, (monocyte derived) dendritic cells, and polymorphonuclear leukocytes. Adaptive immunity also declines, with reduced diversity in naïve T and B cells, reduced T cell memory, reduced B cell memory, and reduced differentiation of B cells to plasma cells. As aging progresses, lymph nodes shrink and fibrosis occurs, limiting the ability of dendritic cells and T and B cells to migrate and encounter each other.21222324

Vaccines for herpes zoster

Pathogenesis and significance of herpes zoster

Herpes zoster results from reactivation of varicella zoster virus from a latent state in dorsal root and cranial nerve sensory ganglia following varicella.25 Reactivation of latent varicella zoster virus is prevented, or controlled as a sub-clinical infection, by varicella zoster virus-specific T cell-mediated immunity.2526 When varicella zoster virus-specific T cell-mediated immunity falls below some critical threshold, the reactivated varicella zoster virus damages neurons and causes a ganglionitis, resulting in neuropathic pain in the affected dermatome. Reactivated varicella zoster virus also is transported anterogradely in the affected sensory nerve to the skin, causing the pathognomonic dermatomal rash and nociceptive pain.

The typical dermatomal herpes zoster rash evolves from erythematous papules to vesicles and crusts, over 7-10 days, with healing occurring within four weeks.27 Pain, which may precede rash, is often severe. Complications of herpes zoster increase with age and immune compromise. The decline of varicella zoster virus-specific T cell-mediated immunity with age is strongly correlated with the incidence of herpes zoster and severity of herpes zoster pain.2829 Significant pain persists for at least three months in 15% of people over 50.30 This post-herpetic neuralgia is the most frequent complication of herpes zoster, and increases in frequency, severity, and duration with advancing age, usually after the age of 50. Post-herpetic neuralgia is not prevented by antiviral therapy and the pain is difficult to treat, especially in older patients. Other complications of herpes zoster include ocular infection in ~10% (uveitis in ~1.5%), bacterial superinfection (2-5%), and disseminated disease in immune compromised patients. Neurological complications such as cerebral arteritis and stroke,31 Ramsay Hunt syndrome, myelitis, motor neuropathy, and encephalitis32 are uncommon complications.

Live attenuated herpes zoster vaccine (ZVL)

The recognized relationship between declining varicella zoster virus-specific T cell-mediated immunity and increasing incidence and severity of herpes zoster led to experiments showing that varicella zoster virus-specific T cell-mediated immunity could be significantly and safely increased in older people by administration of the Oka (Merck) strain of live attenuated varicella zoster virus. An investigational herpes zoster vaccine containing 19 400 pfu of this virus (14-fold greater than in varicella vaccine) was administered33 to 38 500 immune competent individuals aged ≥60. This live attenuated varicella zoster virus vaccine (ZVL; Zostavax) was 51.3% effective in preventing herpes zoster and 66.5% effective in preventing post-herpetic neuralgia (table 1a). However, a statistically significant age effect was evident, in that efficacy was 63.9% for people who were 60-69 years at the time of vaccination, 41.0% for those 70-79 years, and even lower for older vaccinees.33 Furthermore, long term follow-up of the efficacy trial indicated that efficacy declined significantly at 4-8 years after vaccination.34 This was replicated by one effectiveness study,35 but not by another large one where protection against herpes zoster remained at 47.2-41.0% for at least four years and against post-herpetic neuralgia at 60-70%, regardless of age at the time of vaccination.36 An additional limitation of ZVL is that, as a live attenuated vaccine, it is contraindicated for use in moderate-to-severely immunocompromised patients for fear of disseminated and/or fatal disease.37

Table 1a

Efficacy of ZVL in the pivotal clinical trial

View this table:

Despite these limitations, ZVL was an important advance. It is the only herpes zoster vaccine licensed for immunocompetent adults aged ≥50 in many countries, although national recommendations and program funding often limit its use to older age groups. Furthermore, ZVL has an important attenuating effect (even when herpes zoster occurs) such that post-herpetic neuralgia is prevented in 67% of vaccinees aged >703336 and largely preserves the quality of life of older vaccinees who develop herpes zoster.38 In the UK, where ZVL is recommended for adults aged 70, with a catchup program of 71-79 years, coverage after three years was 72% for the first routine cohort and 58% for the catchup cohorts after three years.39 In this case, herpes zoster vaccine was estimated to have reduced the incidence of herpes zoster by 35% and of post-herpetic neuralgia by 50%.39 Another study found similar levels of vaccine effectiveness against herpes zoster.40

Attempts to improve the clinical characteristics of ZVL

Measures to overcome the limitations of ZVL were complicated by the absence of a specific immune surrogate for protection. Levels of antibody and varicella zoster virus-T cell-mediated immunity before and after herpes zoster vaccination correlated with protection on a population basis. However, none predicted protection for an individual vaccine recipient28 with the exception that that a fivefold rise in varicella zoster virus antibody titer at six weeks post vaccination was predictive of 90% protection in individuals aged 50-59.41 Vaccine immunogenicity was not increased by increasing the concentration of vaccine virus or giving two doses.4243 Re-immunization of some individuals a decade after first immunization resulted in significantly higher levels of varicella zoster virus-specific T cell-mediated immunity than obtained in age matched first time vaccinees. Thus the declining immunity was reversed, and was even additive to the varicella zoster virus-specific T cell-mediated immunity increase from the initial vaccination,44 persisting at least partially for up to three years.45

Long term vaccine safety

A 10 year review of vaccine safety46 after administration of 34 million doses reported 23 556 adverse events, of which 93% were not serious. Seven cases of vaccine strain related severe disease have occurred, including disseminated disease, and three deaths, all in immunocompromised patients, emphasizing the importance of avoiding administration of ZVL to severely immunocompromised patients.4647

Recombinant subunit herpes zoster vaccine (RZV)

The more effective subunit herpes zoster vaccine (Shingrix; RZV) consists of recombinant varicella zoster virus glycoprotein E (gE) and the AS01B adjuvant system. The glycoprotein induces gE-specific immune responses and the AS01B shapes the characteristics of these responses. AS01B consists of QS21, a saponin in a purified extract from Quillaja Saponaria, and the toll-like receptor 4 (TLR4) agonist monophosphoryl lipid A (MPL); both contained within liposomes. ASO1B stimulates gE-specific CD4+T cell and antibody responses in animal models.4849 Similar adjuvants have been used for numerous pathogens, including herpes simplex virus, hepatitis B, and malaria51525354 with partial success. Varicella zoster virus gE was selected because it is the most abundant glycoprotein on varicella zoster virus infected cells and is a major target for varicella zoster virus specific antibody and T cells.505152 RZV also induces virus neutralizing antibodies.53

In phase I/II clinical trials, combinations and concentrations of the antigen and adjuvant tested in adults aged ≥50 indicated that the adjuvant system added to recombinant gE induced much higher gE-specific CD4+T cell and antibody responses than did the glycoprotein alone. This was apparent regardless of age, including participants aged ≥70.54 Administration of a second dose of gE/AS01B two months after a first dose increased VZV-specific CD4+T cell frequencies and antibody levels by approximately 30%. A two dose schedule was therefore selected for phase III trials.

RZV efficacy against the incidence of herpes zoster (tables 1b, 1c)

Table 1b

Age specific efficacy of RZV in pivotal clinical trials

View this table:
Table 1c

Duration of RZV efficacy over the four years of the phase III trials

View this table:

Two randomized, blinded, placebo controlled phase III trials of RZV were conducted concurrently in 18 countries (in five continents) in adults aged ≥50 (Zoster Efficacy Study in Adults ≥50 Years of Age [ZOE-50]) and in adults ≥70 Years old [ZOE-70]).5556 In ZOE-50, participants were aged 70-79 and ≥80. In both trials, efficacy for prevention of herpes zoster infection and vaccine safety and reactogenicity were determined in the overall study populations, and in each age group. The ZOE-70 trial was designed to better define efficacy against herpes zoster and post-herpetic neuralgia in people aged ≥70, since both are much more common in this age group.5556 Herpes zoster was confirmed by polymerase chain reaction (PCR) of lesion swabs for varicella zoster virus DNA (90%) or, if diagnosis was equivocal, by examination of clinical reports and photos. In ZOE-50, 210 people in the placebo group (9.1/1000 person years) and six people in the vaccine group (0.3/1000 person years) had herpes zoster (vaccine efficacy of 97.2%). Efficacy was similar in all three age groups (table 1a) and persisted over the median of 3.2 years of the study and then over seven years in an extension of these trials.57 Efficacy was similar in different regions of the world. In ZOE-70, 223 placebo and 23 vaccine recipients developed herpes zoster (vaccine efficacy 89.8%). Efficacy was similar in the 70-79 year age group (90.0%) and those aged ≥80 (89.1%; P<0.001) for people aged ≥70 old pooled from both trials (91.3%) unlike the age effect with ZVL.333435

RZV efficacy against post-herpetic neuralgia

Pooled data from ZOE-50/70 showed that incidence of post-herpetic neuralgia was 0.9 and 0.1 cases per 1000 person years for placebo and vaccine recipients, respectively, for an efficacy against post-herpetic neuralgia of 91.2% (95% confidence interval, 75.9 to 97.7%). No cases of post-herpetic neuralgia were recorded in people aged <70. The incidence of post-herpetic neuralgia in cases of herpes zoster was similar in placebo and vaccine recipients (12.5% versus 9.6%, P=0.54) indicating that RZV prevents post-herpetic neuralgia mainly through preventing herpes zoster rather than attenuating post-herpetic neuralgia, as seen with ZVL.333656 The lack of an age effect with RZV on herpes zoster and post-herpetic neuralgia is important, as the incidence and severity of both increases markedly with age in unvaccinated populations. Although the incidence of herpes zoster has been reported to differ across gender and racial groups,58 no such differences were observed in the efficacy of RZV.59

Safety, reactogenicity, and tolerability of RZV

RZV was associated with more injection site and systemic reactions within seven days of administration than in most other vaccines, and can be disabling (table 2). Myalgia and fatigue, and pain at the injection site, were the most frequent reactions after either dose.546061 Grade 3 injection site reactions (impairing everyday activities, similar to that defined in the Common Terminology Criteria for Adverse Events scale) were reported in 8.5-9.5% of vaccine recipients; 6-11% reported grade 3 systemic reactions.

Table 2

Local and systemic reactogenicity to RZV

View this table:

Both systemic and local reactions were short lived (1-3 days), and did not vary significantly between the first and second dose.62 Most (66%) subjects with a grade 3 reaction after dose one had a lower grade reaction after dose two. Both local and systemic reactions were less frequent in people aged ≥80 and those who were frail. Vaccine reactions did not prevent 96% of participants from receiving a second dose.5556 The adjuvant system, AS01B, was shown in the phase II RZV trials to be mainly responsible for reactogenicity.5461 No new immune mediated diseases or exacerbations of old ones were detected in RZV recipients.6364 No differences were recorded in serious adverse events or deaths between vaccine and placebo groups. The safety and immunogenicity of RZV was similar in vaccinees who had herpes zoster in the five years before RZV.65

Immunogenicity of RZV

In the ZOE trials (and previous phase II) trials, RZV induced marked increases in both cell mediated and humoral immunity with no age effect (table 3).

Table 3

RZV induced antibody and CD4 T cell responses

View this table:

Almost all vaccinees showed notable antibody responses above their baseline. The geometric mean concentration of anti-gE antibodies peaked at one month after dose two and plateaued at 12-36 months. Statistically significant increases in gE-activated CD4 T cell frequency were observed in >90% of RZV recipients. Peak CD4+T cell responses in RZV recipients also occurred at one month after dose two and declined at 12 months in all age groups, but plateaued and remained significantly above baseline at 36 months. An early decline was noted in CD4+T cells with single functions but persistence of those with multiple functions at 36 months in all age groups.5366 In particular, interleukin-2 responses were retained. An immunological correlate of protection could not be determined in the ZOE-50/70 trials because there were few cases of herpes zoster in recipients of RZV.

When immune responses to ZVL and RZV were compared in participants aged 50-85, both VZV and gE-specific T cells were more frequent after RZV administration; gE-specific CD4 and CD8 effector and memory T cell responses were >10-fold higher after RZV,6768 and the peak responses in RZV recipients approximately predicted their subsequent persistence. gE-specific T cell responses to RZV have remained above baseline levels for at least nine years.54697071 The most important immunologic principle emerging from the phase I/II and III trials is that a single viral protein combined with an appropriate adjuvant combination provided strong protection against herpes zoster in both older and immunocompromised populations, with persistence for many years. This is probably because varicella zoster virus gE is an excellent immune target to stimulate CD4+T cell responses, and because of enhancement of gE-specific memory immune responses by AS01B.4872

Mechanism of action of AS01B in RZV

In mouse models, AS01B stimulates a local innate immune response in injected muscle and draining lymph nodes, leading to recruitment and activation of antigen-presenting dendritic cells in the lymph nodes.727374 These take up and present gE to CD4 and CD8 T cells and B cells (fig 3). MPL synergizes with QS-21 to enhance the antibody and T cell responses to gE.66

Fig 3
Fig 3

Mechanism of AS01B action in mouse lymph node. Initially QS21 stimulates peripheral (sinus lining) macrophages, to release interleukins 12 and 18, which stimulate natural killer cells to release interferon gamma. This attracts monocyte derived dendritic cells into the nodes and with MPL activates them and resident dendritic cells. The dendritic cells then present the gE antigen to both T and B cells.7274 The encapsulation of MPL and QS21 in liposomes enhances cellular uptake and reduces toxicity. Components of this cascade, especially interferon gamma, have been identified in primate lymph nodes and/or human blood after immunization with RZV6673

Efficacy, reactogenicity, and safety of RZV in people with comorbidities

Apart from immunosuppressive diseases and drugs, several conditions may increase the risk of herpes zoster, for example, rheumatoid arthritis and inflammatory bowel disease.5875 Many of these are markedly increased in aging populations. In the ZOE trials ~82% of participants reported at least one and most had multiple comorbidities. No safety concerns were identified by number or type of comorbidities. Vaccine efficacy did not vary with the presence of any pre-morbid condition or when multiple conditions, up to six, were present.75 Frailty, measured directly using the Cumulative Deficits approach, or by surrogate methods, did not diminish efficacy against herpes zoster in the ZOE participants.6476

Immunization of immunocompromised patients

Immunocompromised patients have an increased incidence of herpes zoster and, on average, increased severity, both in proportion to the degree of immunocompromise. Patients receiving allogeneic or autologous hematopoietic stem cell transplantation (HSCT) are particularly susceptible, with an incidence of herpes zoster of 15-30% in their first year after transplantation.77 The risk of herpes zoster in untreated HIV infected people is 10- to 20-fold higher than in the age matched population, and remains two- to threefold higher with antiretroviral therapy. Complications of herpes zoster, including recurrent episodes, are about threefold higher in people who are infected with HIV.78 Live attenuated vaccines such as ZVL are contraindicated in conditions that result in severe immunocompromise, such as HIV with CD4 counts <200/µL, administration of high dose corticosteroids, and other immunosuppressive therapies, and after HSCT or solid organ transplantation.46 Disseminated vaccine strain infections and deaths have occurred in these settings.47 The safety and efficacy of ZVL in less immunocompromising conditions, such as autoimmune diseases treated with biologics (including inhibitors of tumor necrosis factor) is being determined.79 ZVL can also be administered before chemotherapy or transplantation, when this is practical.37

Five clinical studies of RZV have been reported in immunocompromised patients (table 4). In each study safety was demonstrated. Efficacy was also shown in trials undertaken in patients receiving autologous HSCT transplants83 and in patients receiving chemotherapy for hematologic malignancy.84 In three smaller trials, substantial immunogenicity was seen in patients with HIV (mostly with substantial reconstitution),78 after renal transplantation,85 or after chemotherapy for solid tumors.86 In the renal transplantation study no vaccine effect was noted on allograft rejection. The solid tumor study determined that immune responses were optimal if the vaccine was not administered during a course of chemotherapy. Additional trials are under way for many autoimmune diseases (such as multiple sclerosis, inflammatory bowel disease, and rheumatoid arthritis) in which biologic modifiers are being administered.

Table 4

Immunogenicity and efficacy of RZV in immunocompromised patients

View this table:

Comprehensive comparisons of the safety and efficacy of RZV and ZVL in diseases or therapeutic settings resulting in mild to moderate immune suppression, such as for autoimmune diseases, would be informative. In particular the effect of additional therapies suppressing CD4 T cell responses to RZV needs testing.


Current research aimed at understanding the exact mechanism of the potent AS01B adjuvant in humans, building on past research in animal models, will provide information on how immunogenicity can be retained while reducing reactogenicity. This information will be relevant to other vaccines for older people.

Vaccines for influenza

Influenza viruses are enveloped viruses with an RNA genome comprised of eight gene segments that encode one or more protein(s). Influenza A and B viruses can co-circulate and cause influenza in all age groups. Humans are exposed to influenza viruses in early childhood. This first experience with influenza viruses can affect how the immune system responds to influenza vaccination later in life, with an anamnestic recall of antibody to influenza virus strains experienced early in life—so called “original antigenic sin.”87

Influenza A viruses are classified into subtypes based on their major surface glycoproteins, hemagglutinin and neuraminidase. Influenza A viruses that include 18 HA and 11 NA subtypes infect animals including birds, pigs, horses, and bats. Only three hemagglutinin subtypes (H1, H2, and H3) and two neuraminidase subtypes (N1 and N2) have caused sustained epidemics of influenza in humans. Influenza B viruses are more limited in their diversity and are classified into two lineages.

Antigenic drift and antigenic shift

Antigenic change in influenza viruses occurs by two mechanisms: antigenic drift and antigenic shift. Antigenic drift is a continuous process in both influenza A and B viruses, where point mutations in the hemagglutinin and neuraminidase proteins arise, allowing the virus to evade neutralization by antibodies induced by previous infection or vaccination. In contrast, antigenic shift is a major antigenic change that occurs when a virus bearing a novel hemagglutinin—with or without a novel neuraminidase and/or other accompanying gene segments derived from an influenza A virus from another species, typically birds or pigs—is introduced and becomes established in humans.

A pandemic occurs when a novel influenza A virus, to which a large proportion of the population lacks immunity, crosses the species barrier, causes disease, and spreads from person to person through sustained chains of community wide transmission. Influenza pandemics occur rarely at irregular intervals, recent examples being in 1918, 1957, 1968, and 2009. They can result from direct infection of humans by an animal influenza virus or by reassortment between an animal influenza virus and the previously circulating human influenza virus. An influenza pandemic triggers the need for a new influenza vaccine. This review focuses on seasonal influenza vaccines.

Influenza infection induces a protective systemic and mucosal antibody response mainly to the hemagglutinin and neuraminidase surface glycoproteins. “Follicular helper” CD4+ T cells are also stimulated to maintain antibody production. CD8+ T cells that recognize the highly conserved internal virion proteins clear virus and reduce disease severity by lysing infected cells and secreting cytokines, but do not prevent infection. Therefore, the principle underlying currently licensed influenza vaccines to prevent infection is the induction of protective antibody against the hemagglutinin; hemagglutination inhibition (HAI) titers are measured as a correlate of protection.

Effectiveness and composition of licensed vaccines

Currently licensed seasonal influenza vaccines are multivalent. They contain antigens representing two circulating influenza A subtypes and either one or both of the circulating influenza B virus lineages.11 The standard influenza vaccine is a trivalent or quadrivalent preparation comprised of inactivated split virions, enriched for hemagglutinin and neuraminidase and formulated to contain 15 μg of hemagglutinin of each virus component. Influenza vaccine effectiveness depends on many factors,88 including the antigenic match between the circulating strain and the strain included in the vaccine, the presence of egg adaptation mutations resulting from growth of the vaccine virus in eggs (discussed below), and age. Influenza vaccine effectiveness has been low for the last several years, as summarized (for the US) in table 5.

Table 5

Adjusted vaccine effectiveness estimates in the US (2015-2019) from the US Vaccine Effectiveness Network

View this table:

The composition of the seasonal influenza vaccines is updated annually, based on epidemiologic and virologic surveillance conducted by the World Health Organization’s Global Influenza Surveillance and Response System, a network of national influenza centers and collaborating centers that monitors influenza virus activity around the world, including antigenic and genetic characterization, and makes predictions based on models from hemagglutinin sequences.

Three key challenges exist in selecting viruses for inclusion in seasonal influenza vaccines. First, there is considerable genetic heterogeneity among the influenza A viruses that are circulating globally. Currently licensed influenza vaccines induce strain specific immunity, and viruses from different genetic clades do not elicit cross reactive antibodies. Increasing the breadth of vaccine induced immunity is a priority research area that includes engineering broadly cross reactive hemagglutinins, standardizing the amount of neuraminidase, incorporating adjuvants, and ultimately, the development of universal influenza vaccines. Second, vaccine strains are selected about six months ahead of implementation because vaccine manufacture and delivery takes several months. In the interim, influenza viruses continue to evolve, so better prediction methods are needed and are being explored.93 Third, most of the global supply of influenza vaccine is manufactured in embryonated chicken eggs. Egg adaptation mutations resulting from growth of the vaccine virus in eggs can alter the antigenic characteristics and glycosylation of the hemagglutinin protein.88 A vaccine manufactured in a qualified cell line and a recombinant hemagglutinin protein vaccine—both of which avoid egg adaptation mutations—are available in some countries.

Standard influenza vaccines have consistently shown lower immunogenicity and efficacy in older adults than in young people,94 irrespective of their composition. Vaccine developers have exploited two strategies to develop “enhanced” vaccines for older people, based on the principle that enhanced immunogenicity of the vaccine will improve vaccine efficacy. The first is to increase the dose of vaccine, as a high dose vaccine contains four times the dose of hemagglutinin in the standard dose vaccine or as a recombinant hemagglutinin (rHA) (tri-)vaccine expressed from a baculovirus vector in insect cell culture. The second strategy includes an adjuvant (MF59 adjuvanted vaccine).

Until enhanced quadrivalent preparations are available, older adults should receive enhanced trivalent instead of standard dose quadrivalent vaccines because the increase in immunogenicity is more important than the inclusion of two strains of influenza B in this age group. Unlike young children, older adults have cross reactive influenza B antibody from a lifetime of exposure to influenza B viruses.

High dose vaccine

Fluzone high dose TIV contains 60 μg of each hemagglutinin. Men and women both experienced a greater response to high dose than to standard dose vaccine. In post hoc analyses, the improved immunogenicity of high dose vaccine was maintained in participants aged ≥75 and in those with chronic heart or lung conditions (table 6).102

Table 6

High dose influenza vaccine

View this table:

The US Food and Drug Administration approved high dose trivalent vaccines for use in older adults in 2009, with a requirement for post-licensure data on the effectiveness of the vaccine in preventing influenza in older people. A phase IIIb/IV study in nearly 32 000 older adults over two influenza seasons suggested that using high instead of standard dose vaccine could have prevented about one quarter of breakthrough influenza illnesses and one third of breakthrough illnesses caused by strains similar to those in the vaccine.96 The number needed to vaccinate with high dose instead of standard dose vaccine to prevent one all-cause hospital admission ranged from 52.6 to 71.4 in a meta-analysis of randomized studies alone or both randomized and observational studies103; 68.7 to 83.7 nursing home residents would need to receive high dose instead of standard dose vaccine to avoid one hospitalization.101

High dose vaccine is cost effective and likely cost saving in older adults.102 Modeling showed that a shift in the US from standard dose to high dose trivalent vaccine would prevent 195 958 cases of influenza, 22 567 influenza related hospitalizations, and 5423 influenza related deaths in older adults, and a shift from standard dose quadrivalent to high dose trivalent vaccine would prevent 169 257 cases of influenza, 21 222 influenza related hospitalizations, and 5212 influenza related deaths. Factoring in the costs of vaccine and reduced healthcare utilization, high dose trivalent vaccine was more cost effective than standard dose trivalent or quadrivalent.104 Analysis of data from a phase IIIb/IV trial estimated that the mean per-participant medical costs were approximately US$116 lower in the high dose trivalent group than the standard dose trivalent group. High dose trivalent vaccine was also cost saving compared with standard dose trivalent in participants with one or more comorbid conditions ($106) and in those aged ≥75 ($12).105

Adjuvanted vaccines

MF59 is an oil-in-water adjuvant that contains squalene.106 Its mechanism of adjuvanticity is only partially understood.107 MF59 triggers immunostimulation via cytokines that indirectly activate dendritic cells. MF59 enhances recruitment and activation of immune cells at the site of injection and enhances uptake of antigen by antigen-presenting cells and their transport to the draining lymph nodes.108 MF59 induces the release of extracellular ATP in muscle, which is the “danger signal” that is able to enhance the antibody response to the co-administered antigens.109

Magnitude, breadth, diversity, and avidity of hemagglutinin antibodies were greater in older adults given the MF59 adjuvanted influenza vaccine than in those given the standard dose. Higher geometric mean titer of hemagglutinin inhibiting antibodies and percentage of people with a ≥fourfold rise in antibody titer were observed following the MF59 adjuvanted vaccine; these differences lasted six months post-vaccination.110111 The antibody response to MF59 adjuvanted vaccine is broadened, with higher antibody titers to antigenic drift strains. The adjuvanted antibody repertoire showed an increased proportion of anti-hemagglutinin antibodies directed to the HA1 receptor binding domain than to the antigenically less important stem region.112

A systematic review of 11 observational studies that enrolled >546 000 person-seasons showed that MF59 adjuvanted vaccine is effective in preventing hospitalizations for various influenza complications and is superior to standard vaccines. MF59 adjuvanted vaccine is also effective against a number of influenza related outcomes in older people living in communities and long term care facilities (table 7).107115

Table 7

MF59 adjuvanted vaccine

View this table:

Recombinant hemagglutinin vaccine

The recombinant hemagglutinin (rHA) vaccine is a highly purified product formulated to contain 45 μg of recombinant hemagglutinin per strain, three times more than the amount of hemagglutinin in standard vaccines. Notably, the rHA vaccine does not contain neuraminidase and avoids the presence of egg protein.

In a randomized controlled trial of trivalent rHA compared with standard dose trivalent vaccine, geometric mean titers of serum hemagglutinin inhibiting antibody and seroconversion rates against influenza A antigens were significantly higher in the group given rHA. These differences were even more pronounced in people who were aged ≥75.116 In a multicenter randomized controlled trial of a quadrivalent, recombinant influenza vaccine (RIV4) compared with standard dose quadrivalent vaccine, RIV4 provided better protection against laboratory confirmed influenza-like illness among adults who were aged ≥50.117 The probability of influenza-like illness was 30% lower with RIV4 than with quadrivalent vaccine (95% confidence interval, 10 to 47%; P=0.006). However, the relative efficacy of the vaccine was lower in the subgroup of participants aged >64 (17%, 95% confidence interval, -20 to 43%) than in participants in the 50-64 year age group (42%, 95% confidence interval, 15 to 61%). The safety profiles of the vaccines were similar.117

Comparison of enhanced vaccines

Several studies compare enhanced vaccines with standard dose vaccines in older people, but few with each other. In a systematic review of 39 randomized controlled trials comparing high dose or adjuvanted vaccines with standard dose in persons aged ≥60 at low risk of bias (in selection, reporting, and other sources), and with sufficiently complete outcome data, enhanced vaccines induced higher antibody responses than standard dose vaccine.118 Increases in post-vaccination geometric mean titer against A(H3N2) in trials comparing enhanced and standard dose vaccines were substantially greater for high dose (82%, 95% confidence interval, 73 to 91%) than for MF59 adjuvanted vaccine (52%, 95% confidence interval, 35 to 72%). Post-vaccination titer ratios following high dose vaccines were significantly higher than for MF59 adjuvanted vaccine (P=0.04). High dose vaccine also elicited higher post-vaccination geometric mean titers than MF59 adjuvanted vaccines for A(H1N1) and B/Victoria viruses.118

Vaccine safety: A randomized controlled trial comparing three enhanced vaccines (MF59 adjuvanted TIV, high dose trivalent, or quadrivalent rHA) with standard dose quadrivalent vaccines in community dwelling older adults aged 65-82 in Hong Kong119 showed that mild and local adverse events were common with all the injected influenza vaccines. Some acute local reactions were more frequent with MF59 adjuvanted and high dose influenza vaccines, compared with standard dose vaccine, whereas systemic symptoms occurred at similar frequencies in all groups.119

Vaccine Immunogenicity: Recipients of enhanced vaccines showed improved humoral and cell mediated immune responses, compared with standard dose vaccine recipients.120 The proportion of participants with ≥fourfold rises to titers ≥40 was significantly higher for all enhanced vaccines (range 59-60%) compared with standard dose quadrivalent (42%). The proportions of participants achieving titers of ≥40 were higher for the MF59 adjuvanted (82%) and high dose (83%) groups, compared with the standard dose group (72%). Furthermore, the proportions achieving very high titers of ≥160 were significantly higher for all enhanced vaccines (45-55%) when compared with the standard dose vaccine (35%). Recipients of enhanced vaccines achieved higher post-vaccination geometric mean titers and greater mean fold rises to A(H3N2) antigens than recipients of standard dose vaccines. Among the enhanced vaccines, antibody responses to A/H3N2 strains were significantly higher among quadrivalent rHA recipients, compared with standard dose vaccine, MF59-adjuvant, and high dose recipients.120


All three types of enhanced vaccines were significantly more immunogenic than the standard dose vaccines. Thus, they could improve vaccine induced immune responses in older adults. Antibody responses were significantly enhanced for higher dose (rHA and high dose) than MF59 adjuvanted vaccines 30 days after vaccination, but determining their duration requires further study. Relative effectiveness of high dose vaccines in prevention of influenza illness will require large efficacy trials to guide future recommendations.

Pneumococcal vaccines

Streptococcus pneumoniae is a gram positive bacterium commonly carried in the nasopharynx that causes disease either by extension from the nasopharynx (pneumonia, otitis media, and sinusitis) or by entering the bloodstream, with spread to normally sterile sites such as the alveoli, meningeal space, or joint fluid (invasive pneumococcal disease [IPD]). The polysaccharide capsule on the surface of pneumococci is the major factor responsible for virulence by preventing opsonization by complement and subsequent phagocytosis.121 More than 90 capsular serotypes of S pneumoniae exist, each characterized by the molecular structure of its specific polysaccharide. Serotypes differ in colonization and invasion capacity such that each could be seen as a unique pathogen.122 This is emphasized by the specificity of antibody protection by serotype. Although antibody to some serotypes provides cross protection (6B against 6A and 6A against 6C), this is not the case for others (such as for 19F and 19A).123


Pneumococcal polysaccharide vaccines (PPVs)

Trials in adults during the early-mid 20th century showed good efficacy of pneumococcal capsular polysaccharides vaccines, but were superseded by penicillin.124 However in infants, antibody responses to polysaccharide antigens were short lived and variable.124 In the 1970s a 23-valent polysaccharide vaccine (23vPPV) was developed and has remained unchanged since licensure in the US in 1989. Despite consistent evidence of protection against IPD, especially in vulnerable adults, protection against pneumonia without IPD has been uncertain, owing to lack of sensitive and specific diagnostic tests.1819124

Pneumococcal conjugate vaccines (PCVs)

The first randomized controlled trial of a pneumococcal conjugate vaccine, conjugating the seven most common US serotypes to the mutant diphtheria toxin (CRM157), was completed in the late 1990s. It showed significant efficacy against all seven serotypes in children under 2.125 Following widespread use of PCVs in infants in the US, demonstration of concomitant “indirect” reductions in IPD in adults was instrumental in establishing the cost effectiveness of PCV. Rapidly accelerating use of pediatric PCVs in other high income countries—first 7-valent and then 13-valent—resulted in dramatic reductions in pneumococcal disease caused by vaccine serotypes in both children and adults with the exception of serotype 3.126 However, this indirect reduction in adult disease has rendered uncertain the incremental benefit of direct immunization of adults, as vaccine serotypes have been replaced with other serotypes, especially in adults aged >65. The degree of replacement varies between countries for reasons which are incompletely understood.127

Table 8 shows serotypes included in the two pneumococcal vaccines licensed in adults: 13-valent conjugate and 23-valent polysaccharide. Conjugate vaccines currently in clinical trials contain 15 and 20 serotypes.

Table 8

Licensed and pipeline pneumococcal vaccines for use in adults

View this table:

Immune responses important for protection

The importance of measuring functional antibody as a correlate of pneumococcal vaccine protection was first shown with the 7-valent PCV where, although IgG antibody measured by enzyme linked immunosorbent assay (ELISA) against serotype 19F cross reacted in vitro with antibody against 19A, it was not associated with clinical protection.123 In contrast with ELISA, opsonophagocytic assays (OPA) for 19A and 19F did correlate with clinical protection and now are the gold standard for measuring seroprotection. However for many serotypes, ELISAs are valid and better standardized.128.

A systematic review of five randomized trials comparing OPA antibody geometric mean titers by serotype for 23vPPV and 13vPCV included 4561 adult vaccinees aged ≥50.129 Among people aged >65, OPA results were significantly higher one month post 13vPCV than 23vPPV for all 13v serotypes except 3, 7F, and 14. Serotype specific responses to the two vaccines varied even more among persons aged 50-64.129 Local and systemic reactions did not differ between PCV13 and 23vPPV.129

Cell mediated immunity, particularly CD4 T cells, may also be important for protection against pneumococcal infection with aging in human and mouse models. Studies in aging mice indicate that substantially more pneumococcal antigen may be needed to induce CD4 T cell responses.21 More studies are needed. Mucosal immunity may be important in protection against non-bacteraemic pneumococcal pneumonia, especially in older people, but has been poorly studied for pneumococcal vaccines.22

Vaccine safety

In randomized controlled trials comparing polysaccharide and conjugate vaccines, no significant difference was noted in local or systemic reactions.129 Fewer severe local reactions were seen following PCV13 than PPV23 (relative risk 0.51; 95% confidence interval, 0.29 to 0.90), but among people who had never received a pneumococcal vaccine, local reactions were significantly more common in the PCV13 arm (relative risk 1.15; 95% confidence interval 1.05 to1.26). Frequency of local reactions was correlated with higher antibody titers, consistent with a more vigorous immune response to vaccine.129

Vaccine efficacy and effectiveness

Pneumococcal polysaccharide vaccines

For IPD, 17 studies and six meta-analyses found significant effectiveness of 23vPPV, although some studies were limited by small case numbers and inclusion criteria varied.130 Effectiveness decreased as age increased above 65, especially >85, which is the age group with the highest incidence. In the largest observational study of vaccine effectiveness against IPD, including almost 10000 cases in the UK, vaccine effectiveness was low and non-significant among immunocompromised persons >65 years; 53-63% among high risk immunocompetent people aged 65-84; and in those without risk factors vaccine effectiveness was significant even in those aged >85. Importantly, although vaccine effectiveness declined if 23vPPV had been given >5 years previously, it remained statistically significant, and significant vaccine effectiveness was demonstrated against IPD due to serotypes included in 23vPPV but not in PCV13.131 A smaller Korean hospital based case-control study, conducted in the setting of high childhood PCV13 coverage, also found high vaccine effectiveness against IPD due to 23vPPV-unique serotypes (78%; 95% confidence interval, 34.6 to 92.6).132

Whether use of 23vPPV should be restricted to a single dose or repeated has been controversial. A meta-analysis of studies of re-vaccination with 23vPPV found that long term antibody responses and safety of a second dose were comparable to a first dose, but clinical endpoints were lacking in the studies, therefore effectiveness could not be determined.133

For non-bacteraemic pneumonia, a Japanese study was the first to evaluate effectiveness of 23vPPV using sensitive and specific diagnostic methods able to identify all 23 vaccine serotypes.134 The study, conducted in the context of an existing childhood PCV7, but not PCV13 program, used the test negative method, whereby cases were identified by serotype-specific PCR of respiratory specimens or a urinary pneumococcal assay identifying all serotypes combined (Binax Now); controls were negative on both assays (table 8). Although vaccine effectiveness was significant against all pneumococcal pneumonia, serotype-specific data showed effectiveness was higher for serotypes included in the 13vPCV (27.4%; 3.2 to 45.6%), than for serotypes in 23vPPV but not included in 13vPCV (12.0%; -62.8 to 52.4%). Duration of protection was estimated at 1-2 years post-PPV. The previously mentioned Korean study identified non-bacteraemic pneumococcal pneumonia (NBPP) by clinical and radiological criteria plus identification of pneumococci in sputum or urine, but lacked serotype data. Among people aged 65-74, vaccine efficacy was 35% (2.3 to 56.7), but was low and non-significant (13%; 5 to 18%) for people aged >75.132

Pneumococcal conjugate vaccines

The large CAPiTA randomized controlled trial, with almost 85 000 participants, provided high quality evidence of vaccine effectiveness.18 The study used a novel type-specific urinary assay to identify non-bacteraemic community-acquired pneumonia (CAP) caused by PCV13 serotypes, as well as the non-specific urinary pneumococcal assay (Binax Now). As summarized in table 9, 1843 cases of CAP were identified, of which 309 (20.1%) had evidence of pneumococcal cause, 172 were due to vaccine types, and 70 also had IPD. Vaccine efficacy increased from 5.1% (95% confidence interval, -5 to 14%) for all CAP to 30.6% (95% confidence interval, 9.8 to 46.7%) for pneumococcal CAP, and 45.6% (95% confidence interval, 21.8 to 62.5%) for vaccine type pneumococcal CAP. Subsequent ad hoc analyses reported that vaccine efficacy varied by comorbidity status (reduced in lung disease and increased in diabetes mellitus),135 was maintained for all five years of the study,136 and when CAP was defined solely on laboratory criteria (with no requirement for specific radiologic findings) vaccine efficacy was significant for serotypes 3 and 7F but not for serotype 1.137

Table 9

Studies of serotype specific effectiveness of PCV 13 and PPV 23 against pneumonia (CAP) in patients aged >65

View this table:

However, the CAPiTA study was conducted in the absence of PCV13 use in infants (2008-2010), limiting conclusions about the incremental direct protection by PCV on the residual burden of PCV13 serotype disease in older people. In 2014, the US became the only country to recommend direct immunization of people aged ≥65,138 although PCV13 was funded by insurance in Saxony, Germany.139 The only published data on pneumococcal CAP after the US 2014 recommendation come from a single hospital in 2015-16.140 The proportion of CAP resulting from 13v serotypes in this study, using diagnostic assays identical to CAPiTA, was 8.1% (0.3% IPD), which was substantially lower than the 20% (4% IPD) documented in CAPiTA and 21% (1% IPD) in Japan. Of the 164 pneumococcal CAP cases identified, the proportion owing to 13v serotypes was lower than in CAPiTA or Japan, but still greater than 40% (table 8), with efficacy against vaccine-type CAP estimated at 72% (95% confidence interval, 8.7 to 91.4), higher than point estimates from CAPiTA or Japan (40 to 46%), but with overlapping confidence intervals.140 Thus, when considering the overall population impact of PCV13 in the US, the relative importance of direct effect immunization of older adults versus indirect effects of the infant program must be considered. National surveillance through the Centers for Disease Control and Prevention’s ABCs network suggested that, although substantial reductions in pneumococcal CAP have been seen post infant PCV13, there was a non-significant additional reduction of 5% post adult PCV13 vaccination.141

Is vaccine efficacy for pneumococcal vaccines sex specific?

Three studies of 23vPPV for pneumococcal pneumonia from Japan,134 Germany,142 and multi-country143 found substantially lower vaccine efficacy in men than in women, with similar findings from Germany for the efficacy of PCV13.139

Long term trends in indirect effects of pneumococcal conjugate vaccines

In the UK144 and the Netherlands145 reductions in IPD in people over 65 have been less than 20% compared with the pre-PCV7 baseline. In contrast, in the US reductions in IPD versus pre-PCV7 are two to threefold greater (74%), albeit from a high baseline,17 and in Australia146 were 32%. However only the US has had continued decreases in IPD following the introduction of PCV10/13 as measured by incidence rate ratios (table 10). IPD incidence following high childhood coverage of PCV10 (Netherlands) and PCV13 (Australia, UK, and US) varies substantially from 51 per 100 000 in the Netherlands to around 15 per 100 000 in Australia and the US, but the proportion owing to PCV13 serotypes is lowest in the UK (19%), related to more serotype replacement disease.

Table 10

Incidence of IPD per 100 000 and indirect impact of PCV7 and PCV10/13 in adults aged >65

View this table:

Impact of PCV13 in older adults in the US and relative cost effectiveness of PCV13 and PPV23

The US Advisory Committee on Immunization Practices reviewed the five year experience following the 2014 recommendation for PCV13 followed by PPV23 for adults aged ≥65. It concluded that no evidence existed of incremental reductions in vaccine type IPD attributable to direct PCV13 use in older adults and found scant evidence of reductions in non-IPD CAP.138 An updated economic analysis found PCV13 immunization of older adults to be not cost effective.148 Where indirect reductions in IPD in older adults have not been as great, recent cost effectiveness evaluations conducted in the Netherlands,147 Australia,149 and the UK150 concluded that 23vPPV—but not PCV13—met cost effectiveness criteria, especially in adults aged over 70. In Australia, PCV13 was recently deemed cost effective for all adults aged >70, but sequential use of 23vPPV following PCV13 was recommended only for persons with significant comorbidities and for indigenous populations.151


Similar to high dose influenza vaccine, double dose PCV13 in persons aged over 55 showed superior immunogenicity to single dose, but higher or more long lasting clinical efficacy has not yet been demonstrated.152 Higher valency conjugate vaccines, 15-valent, or 20-valent serotype vaccines are in clinical trials, with commercial availability anticipated in 2-5 years. The 20-valent vaccine includes the serotypes most prominent in serotype replacement in countries where this has been a problem.153 Development of universal pneumococcal vaccines to avoid serotype replacement is being attempted.154

Co-administration of vaccines for older people

In studies that compare vaccines containing pneumococcal polysaccharide plus influenza significant reductions in pneumonia were documented compared with either influenza vaccine alone (15% greater reduction; 95% confidence interval, 4 to 24%) or with pneumococcal vaccine alone (24%; 95% confidence interval, 16 to 31%).159. No studies have evaluated effectiveness of concomitant influenza and PCVs but lower OPA antibody responses to some serotypes suggested that co-administration should be avoided where practicable.155

Co-administration of PCV13 with the first herpes zoster vaccine, ZVL, was initially reported to reduce the antibody response to ZVL, but this result has been re-examined and the US Centers for Disease Control and Prevention now recommends co-administration to ensure maximum protection of older individuals. The reactogenicity, efficacy, or immunogenicity of RZV administered with pneumococcal or influenza vaccine is similar for dual vaccination to each administered alone.156

Guidelines (where relevant)8081

View this table:


RZV is the benchmark for efficacious vaccines in older people because it greatly reduces the risk of herpes zoster and post-herpetic neuralgia in adults aged ≥50, independent of age and frailty status. RZV will substantially reduce the societal costs of herpes zoster in older adults, including the use of antivirals and pain medications, and hospitalization. It is also likely that RZV will be a partial solution to the problem of frequent and severe herpes zoster occurring in immune compromised patients. ZVL remains useful in preventing or attenuating herpes zoster in countries where RZV is not yet available.

During the covid-19 pandemic, where influenza infection may mimic or synergize with SARS-CoV-2 infection, optimal coverage of older populations with enhanced influenza vaccines in autumn and winter is essential. Optimizing prevention of influenza and pneumococcal diseases will require still more effective vaccines. Lessons learned from RZV and its development suggest that using one or a few target antigens combined with appropriate adjuvant combinations may be optimal for vaccines for other diseases of older adults, perhaps including influenza, respiratory syncytial virus, and covid-19.

Questions for future research

Herpes zoster

In what other immunosuppressive diseases and/or treatments will RZV be effective?

What is the real world effectiveness of RZV in clinical practice and the duration of protection?353671?

What is the safety of adjuvanted vaccines in patients with autoimmune diseases over long term follow-up?


Is one of the enhanced influenza vaccines more effective than the others?

How many months does protection against influenza last? What is the optimal timing of vaccinating in relation to onset of the influenza season? Does data support a second dose of vaccine in the same influenza season?

Will enhanced influenza vaccines reduce the risk of severe influenza related complications, such as bacterial pneumonia, myocardial infarction, congestive heart failure, stroke, and death in older people, and what are the economic benefits?

Pneumococcal disease

Could new adjuvants provide higher or more long lasting protection against pneumococcal infection in adults >80?

Are current pneumococcal vaccines effective in severe immunocompromise?

Do pneumococcal vaccines reduce incidence or severity of lung disease from covid-19?


  • State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors

  • Patient involvement: no patients were asked for input in the creation of this article.

  • Competing interests: ALC has consulted on vaccines for Merck, BioCSL/Sequirus, and GlaxoSmithKline, and his institution has received resulting honoraria. MJL received fees for serving on advisory boards from Merck, GlaxoSmithKline, and Curevo, and grant support from Merck and GlaxoSmithKline.

  • Acknowledgments: Thanks to Kerry-Anne Baxter, community representative to the Australian National Centre for immunisation Research and Surveillance, for reviewing the manuscript, Catherine King also in NCIRS for the Pubmed and Medline searches and to Joanne Camilleri for processing it.

  • Contributors: AC and ML planned the scope of the review and wrote the sections on immunosenescence and herpes zoster vaccine. KS and PM wrote the sections on influenza and pneumococcal immunization sections. RB contributed to the influenza vaccine section and the literature searches. All authors contributed to the introduction, conclusions, abstract, and other general sections.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract