Grading of Recommendations, Assessment, Development, and Evaluation (GRADE): Protein subunit RSV vaccines (GSK Arexvy and Pfizer Abrysvo) in older adults

Table 1a: Policy Question and PICO for Adults Aged ≥75 Years

Table 1b: Policy Question and PICO for Adults Aged 60–74 Years at Increased Risk of Severe RSV Disease

a Three options: 1. Critical; 2.  Important but not critical; 3. Not important for decision-making
b No events were recorded in included studies, so the outcome could not be evaluated.

Table 3a: Summary of Studies Reporting RSV LRTD

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a ≥2 lower respiratory trace disease symptoms or signs for ≥24 hours including ≥1 lower respiratory sign, or ≥3 lower respiratory symptoms for ≥24 hours. Lower respiratory symptoms included new or increased sputum, new or increased cough, and new or increased dyspnea. Lower respiratory signs included new or increased wheezing, new or increased crackles or rhonchi based on chest auscultation, respiratory rate ≥20 respirations per minute, low or decreased oxygen saturation (<95% or ≤90% if baseline was <95%), and need for oxygen supplementation.

b ≥3 lower respiratory trace disease signs or symptoms. Lower respiratory signs and symptoms included new or worsened cough, new or worsened sputum production, new or worsened wheezing, new or worsened shortness of breath, and new or worsened tachypnea.

c Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 3b: Summary of Studies Reporting Medically Attended RSV LRTD

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Medically attended RSV LRTD was defined as RSV LRTD with attention at ≥1 inpatient or outpatient health care service.

b Medically attended RSV lower respiratory tract illness (LRTI) included LRTI prompting any health care visit (including any outpatient or inpatient visit such as hospitalization, ER visit, urgent care visit, home healthcare services, primary care physician office visit, pulmonologist office visit, specialist office visit, other visit, or telehealth contact).

c Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 3c: Summary of Studies Reporting Hospitalization for RSV Respiratory Illness

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 3d: Summary of Studies Reporting Severe RSV Respiratory Illness Requiring O2/Respiratory Support

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 3e: Summary of Studies Reporting Serous Adverse Events

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Serious adverse events were defined as any untoward medical occurrence (during six months after injection in the phase 3 trial and 60 days after injection in the phase 1/2 trial) that resulted in death, was life threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent disability or incapacity, or was a congenital anomaly or birth defect.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c Includes all events for 6 months after injection.

d Includes all events after dose 1, but before dose 2 (day 61) in a placebo-controlled phase 1/2 dosing selection study.

e Serious adverse events were defined as any untoward medical occurrence (through day 182 after injection in the phase 3 trial and 60 days for the phase 1/2 trial) that resulted in death, was life threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent disability or incapacity, or was a congenital anomaly or birth defect.

f Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

g All events reported from vaccination through day 182.

h Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random-effect model.

Table 3f: Summary of Studies Reporting Inflammatory Neurologic Events^a

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Guillain-Barre syndrome (including variants thereof), chronic inflammatory demyelinating polyneuropathy, and acute central nervous system inflammation (e.g., transverse myelitis, acute disseminated encephalomyelitis) occurring within 42 days after injection.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c Calculated using 0.5 correction factor to account for zero events.

d Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

e In the Pfizer phase 3 RCT there were a total of three cases of inflammatory neurologic events initially identified as occurring within 42 days after injection reported by investigators. All occurred in the intervention arm. These included GBS (14 days post-vaccination), Miller Fisher syndrome (a variant of GBS; 10 days post-vaccination), and undifferentiated motor-sensory axonal polyneuropathy (site investigator reported this case as not associated with vaccination; participant had worsening of pre-existing symptoms 21 days post-vaccination). After further investigation, the case of undifferentiated motor-sensory axonal polyneuropathy was excluded. This is a revision since the June 2023 GRADE.

f Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random-effect model.

Table 3g: Summary of Studies Reporting Severe Reactogenicity

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Severe reactogenicity events were defined as grade 3 solicited local reactions (injection site pain, redness and swelling) or systemic reactions (fatigue, fever, headache, gastrointestinal symptoms [nausea, vomiting, diarrhea or abdominal pain], arthralgia, myalgia and shivering) recorded during days 0–4 after vaccination in the phase 3 trial and days 0–7 after vaccination in the phase 1/2 trial. For injection site erythema and swelling, grade 3 corresponded to a diameter >100 mm. For fever, grade 3 corresponded to a temperature >39°C. For all other reactions, grade 3 corresponded to reactions that prevented normal, everyday activities. Grade 4 events were not defined in these trials.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c All events within 4 days of injection.

d All events within 7 days of injection.

e Severe reactogenicity events were defined as grade 3 or higher injection site reaction (pain at injection site, redness and swelling) or systemic reaction (fever, fatigue/tiredness, headache, nausea, muscle pain, joint pain, vomiting, diarrhea, and other systemic event) during days 0–7 after vaccination. For injection site redness and swelling, grade 3 corresponded to a diameter >10.0 cm from e-diary or severe grade from adverse event case report form. For fever, grade 3 corresponded to a temperature >38.9°C from e-diary or severe grade from adverse event case report form. For all other reactions, grade 3 corresponded to reactions that prevented normal, everyday activities. Grade 4 event is only applicable to fever >40℃.

f Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

g Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random-effect model.

Table 4a Summary of Studies Reporting RSV LRTD

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pre-existing comorbidities of interest (“specific medical conditions”) includes adults with chronic obstructive pulmonary disease (COPD), asthma, any chronic respiratory/pulmonary disease, chronic heart failure, diabetes mellitus Type 1 or Type 2, or advanced liver or renal disease.

b Lower respiratory tract disease (LRTD) was defined as ≥2 lower respiratory symptoms or signs for ≥24 hours including ≥1 lower respiratory sign, or ≥3 lower respiratory symptoms ≥24 hours. Lower respiratory symptoms included new or increased sputum, new or increased cough, and new or increased dyspnea. Lower respiratory signs included new or increased wheezing, new or increased crackles or rhonchi based on chest auscultation, respiratory rate ≥20 respirations per minute, low or decreased oxygen saturation (<95% or ≤90% if baseline was <95%), and need for oxygen supplementation

c Pre-specified significant condition includes adults with current tobacco use, diabetes, lung disease [including COPD], heart disease [including congestive heart failure], liver disease, or renal disease.

d LRTD was defined using the trial out of lower respiratory tract illness (LRTI) with ≥3 lower respiratory signs or symptoms. Lower respiratory signs and symptoms included new or worsened cough, new or worsened sputum production, new or worsened wheezing, new or worsened shortness of breath, and new or worsened tachypnea.

e Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 4b Summary of Studies Reporting Medically Attended RSV LRTD

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pre-existing comorbidities of interest (“specific medical conditions”) includes adults with chronic obstructive pulmonary disease (COPD), asthma, any chronic respiratory/pulmonary disease, chronic heart failure, diabetes mellitus Type 1 or Type 2, or advanced liver or renal disease.

b Pre-specified significant condition includes adults with current tobacco use, diabetes, lung disease [including COPD], heart disease [including congestive heart failure], liver disease, or renal disease.

c Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 4c: Summary of Studies Reporting Hospitalization for RSV Respiratory Illness

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 4d: Summary of Studies Reporting Severe RSV Respiratory Illness Requiring O2/Respiratory Support

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel fixed-effect model.

Table 4e: Summary of Studies Reporting Serous Adverse Events

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Serious adverse events were defined as any untoward medical occurrence (during six months after injection in the phase 3 trial and 60 days after injection in the phase 1/2 trial) that resulted in death, was life threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent disability or incapacity, or was a congenital anomaly or birth defect.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c Includes all events for 6 months after injection.

d Includes all events after dose 1, but before dose 2 (day 61) in a placebo-controlled phase 1/2 dosing selection study.

e Serious adverse events were defined as any untoward medical occurrence (through day 182 after injection in the phase 3 trial and 60 days for the phase 1/2 trial) that resulted in death, was life threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent disability or incapacity, or was a congenital anomaly or birth defect.

f Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

g All events reported from vaccination through day 182.

h Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random effect model.

Table 4f: Summary of Studies Reporting Inflammatory Neurologic Events^a

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Guillain-Barre syndrome (including variants thereof), chronic inflammatory demyelinating polyneuropathy, and acute central nervous system inflammation (e.g., transverse myelitis, acute disseminated encephalomyelitis) occurring within 42 days after injection.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c Calculated using 0.5 correction factor to account for zero events.

d Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

e In the Pfizer phase 3 RCT there were a total of three cases of inflammatory neurologic events initially identified as occurring within 42 days after injection reported by investigators. All occurred in the intervention arm. These included GBS (14 days post-vaccination), Miller Fisher syndrome (a variant of GBS; 10 days post-vaccination), and undifferentiated motor-sensory axonal polyneuropathy (site investigator reported this case as not associated with vaccination; participant had worsening of pre-existing symptoms 21 days post-vaccination). After further investigation, the case of undifferentiated motor-sensory axonal polyneuropathy was excluded. This is a revision since the June 2023 GRADE.

f Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random-effect model.

Table 4g: Summary of Studies Reporting Severe Reactogenicity

Abbreviations: CI = confidence interval; RCT = randomized controlled trial.

a Severe reactogenicity events were defined as grade 3 solicited local reactions (injection site pain, redness and swelling) or systemic reactions (fatigue, fever, headache, gastrointestinal symptoms [nausea, vomiting, diarrhea or abdominal pain], arthralgia, myalgia and shivering) recorded during days 0–4 after vaccination in the phase 3 trial and days 0–7 after vaccination in the phase 1/2 trial. For injection site erythema and swelling, grade 3 corresponded to a diameter >100 mm. For fever, grade 3 corresponded to a temperature >39°C. For all other reactions, grade 3 corresponded to reactions that prevented normal, everyday activities. Grade 4 events were not defined in these trials.

b Intervention was a single dose of GSK Arexvy (120µg antigen + AS01_E adjuvant, 1 dose IM).

c All events within 4 days of injection.

d All events within 7 days of injection.

e Severe reactogenicity events were defined as grade 3 or higher injection site reaction (pain at injection site, redness and swelling) or systemic reaction (fever, fatigue/tiredness, headache, nausea, muscle pain, joint pain, vomiting, diarrhea, and other systemic event) during days 0–7 after vaccination. For injection site redness and swelling, grade 3 corresponded to a diameter >10.0 cm from e-diary or severe grade from adverse event case report form. For fever, grade 3 corresponded to a temperature >38.9°C from e-diary or severe grade from adverse event case report form. For all other reactions, grade 3 corresponded to reactions that prevented normal, everyday activities. Grade 4 event is only applicable to fever >40℃.

f Intervention was a single dose of Pfizer Abrysvo (120µg antigen, 1 dose IM).

g Pooled estimate using data from all studies for this outcome generated using Mantel-Haenszel random-effect model.

Table 5a Summary of Evidence for Outcomes of Interest in Adults Aged ≥75 Years

Table 5b: Summary of Evidence for Outcomes of Interest in Adults Aged 60–74 Years at Increased Risk of Severe RSV Disease

Table 6a: Grade Summary of Findings Table in Adults aged ≥75 years

1. 7,358 person-years of observation.
2. 8,416 person-years of observation.
3. Pooled rate ratio and the corresponding 95% CIs were calculated as 1- vaccine efficacy observed in the manufacturers’ phase 3 clinical trials for the specified outcome using all available follow-up time. The input vaccine efficacy was calculated by CDC using reported case counts and person-time without adjustment for covariates.
4. An absolute effect could not be calculated for this outcome because there were no reliable data to estimate the true baseline rate of this outcome in the general population. The baseline rate captured in the placebo arm of the trial was felt to underestimate the true population-based incidence due to atypical RSV burden and seasonality during the COVID-19 pandemic. There were also no estimates of rate of ≥3 symptom RSV LRTD available in the literature.
5. Serious concern for indirectness due to inclusion of adults aged 60–74 years.
6. 20,774 person-years of observation.
7. 25,706 person-years of observation.
8. This represents a control rate of 476 cases of medically attended RSV LRTI with ≥3 symptoms per 100,000 adults aged ≥60 years per RSV season (196 cases of medically attended RSV illness over one RSV season per 10,000 adults aged ≥60 years with chronic cardiopulmonary conditions [as a proxy for those 75 years and older], of which 24.3% are RSV LRTI with ≥3 symptoms). Control rate data are from Belongia EA, et al. Clinical Features, Severity, and Incidence of RSV Illness During 12 Consecutive Seasons in a Community Cohort of Adults ≥60 Years Old. Open Forum Infect Dis. 2018 Nov 27;5(12):ofy316).
9. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote h.
10. Serious concern for imprecision due to the confidence intervals containing measures of absolute risk for which different policy decisions might be considered.
11. 39,323 person-years of observation.
12. 41,962 person-years of observation.
13. This represents a control rate of 314 cases of RSV-associated hospitalization over one RSV season per 100,000 adults aged 75 and older. Control rate data are unpublished data from CDC’s RSV-NET. Available here: https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2024-06-26-28/08-RSV-Adult-Hutton-508.pdf.
14. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote m.
15. Serious concern for inconsistency because the point estimates between the studies differed substantially, although the confidence intervals overlapped.
16. 39,322 person-years of observation.
17. This represents a control rate of 290 cases of severe RSV illness requiring supplemental oxygen or other respiratory support per 100,000 adults aged ≥60 years per RSV season (196 cases of medically attended RSV illness over one RSV season per 10,000 adults aged ≥60 years with chronic cardiopulmonary conditions [as a proxy for those 75 years and older], of which 14.8% require supplemental oxygen in the 28 days after medical attention). Control rate data are from Belongia EA, et al. Clinical Features, Severity, and Incidence of RSV Illness During 12 Consecutive Seasons in a Community Cohort of Adults ≥60 Years Old. Open Forum Infect Dis. 2018 Nov 27;5(12):ofy316
18. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote q.
19. 34,167 person-years of observation.
20. 43,696 person-years of observation.
21. Measures of relative and absolute effect could not be estimated due to zero events.
22. Pooled relative risk estimates were independently calculated using counts of events and participants in the pivotal phase 3 trials, as well as from earlier the phase 1/2 trials.
23. For calculation of absolute effect, the baseline number of events was taken from the rates observed in the placebo arm.
24. Very serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered and fragility in the estimate.
25. Absolute risk was unable to be calculated due to zero events in the placebo arm and no reliable data to estimate the true baseline rate of this outcome in the general population.
26. Differences in the risk of grade ≥3 reactogenicity between GSK and Pfizer vaccines are not unexpected given that one of the products is adjuvanted and one is not; therefore inconsistency was felt to be not serious, as we would expect that the risk of grade ≥3 reactogenicity is different between the vaccine products.

Table 6b: Grade Summary of Findings Table in Adults aged 60–74 Years at Increased Risk of Severe RSV Disease

1. 19,516 person-years of observation.
2. 21,340 person-years of observation.
3. Pooled rate ratio and the corresponding 95% CIs were calculated as 1- vaccine efficacy observed in the manufacturers’ phase 3 clinical trials for the specified outcome using all available follow-up time. The input vaccine efficacy was calculated by CDC using reported case counts and person-time without adjustment for covariates.
4. An absolute effect could not be calculated for this outcome because there were no reliable data to estimate the true baseline rate of this outcome in the general population. The baseline rate captured in the placebo arm of the trial was felt to underestimate the true population-based incidence due to atypical RSV burden and seasonality during the COVID-19 pandemic. There were also no estimates of rate of ≥3 symptom RSV LRTD available in the literature.
5. 19,524 person-years of observation.
6. 21,372 person-years of observation.
7. This represents a control rate of 476 cases of medically attended RSV LRTI with ≥3 symptoms per 100,000 adults aged ≥60 years per RSV season (196 cases of medically attended RSV illness over one RSV season per 10,000 adults aged ≥60 years with chronic cardiopulmonary conditions, of which 24.3% are RSV LRTI with ≥3 symptoms). Control rate data are from Belongia EA, et al. Clinical Features, Severity, and Incidence of RSV Illness During 12 Consecutive Seasons in a Community Cohort of Adults ≥60 Years Old. Open Forum Infect Dis. 2018 Nov 27;5(12):ofy316).
8. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote g.
9. Serious concern for indirectness as this outcome was assessed for all adults 60-74 years, not only those at increased risk of severe RSV disease.
10. Serious concern for imprecision due to the confidence intervals containing measures of absolute risk estimates for which different policy decisions might be considered.
11. 39,323 person-years of observation.
12. 41,962 person-years of observation.
13. This represents a control rate of 198 cases of RSV-associated hospitalization over one RSV season per 100,000 adults aged 60-74 at increased risk of severe RSV disease. Control rate data are unpublished data from CDC’s RSV-NET with at least one condition among: COPD, asthma, CAD, diabetes, obesity (>40 kg/m2), or CKD. Available here: https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2024-06-26-28/08-RSV-Adult-Hutton-508.pdf.
14. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote m.
15. Serious concern for inconsistency because the point estimates between the studies differed substantially, although the confidence intervals overlapped.
16. 39,322 person-years of observation
17. This represents a control rate of 290 cases of severe RSV illness requiring supplemental oxygen or other respiratory support per 100,000 adults aged ≥60 years per RSV season (196 cases of medically attended RSV illness over one RSV season per 10,000 adults aged ≥60 years, of which 14.8% require supplemental oxygen in the 28 days after medical attention). Control rate data are from Belongia EA, et al. Clinical Features, Severity, and Incidence of RSV Illness During 12 Consecutive Seasons in a Community Cohort of Adults ≥60 Years Old. Open Forum Infect Dis. 2018 Nov 27;5(12):ofy316).
18. For calculation of absolute effect, the baseline number of events was not taken from the placebo arm of the trial given concerns that the baseline captured in the trial underestimated the true population-based incidence. Instead the calculation uses the control rate described in footnote q.
19. 34,167 person-years of observation.
20. 43,696 person-years of observation.
21. Measures of relative and absolute effect could not be estimated due to zero events.
22. Pooled relative risk estimates were independently calculated using counts of events and participants in the pivotal phase 3 trials, as well as from earlier the phase 1/2 trials.
23. For calculation of absolute effect, the baseline number of events was taken from the rates observed in the placebo arm.
24. Very serious concern for imprecision due to fragility of the estimate, adjustment of relative risk for zero events observed, and confidence interval containing estimates for which different policy decisions might be considered.
25. Absolute risk was unable to be calculated due to zero events in the placebo arm and no reliable data to estimate the true baseline rate of this outcome in the general population.
26. Differences in the risk of grade ≥3 reactogenicity between GSK and Pfizer vaccines are not unexpected given that one of the products is adjuvanted and one is not; therefore inconsistency was felt to be not serious, as we would expect that the risk of grade ≥3 reactogenicity is different between the vaccine products.