Grading of Recommendations, Assessment, Development, and Evaluation (GRADE): Pfizer Maternal RSV Vaccine

Overview

A Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) review of the evidence for benefits and harms for Pfizer Respiratory Syncytial Virus (RSV) vaccine (RSVpreF, Abrysvo, Pfizer Inc.) was presented to the Advisory Committee on Immunization Practices (ACIP) on June 22, 2023 and September 22, 2023. GRADE evidence type indicates the certainty in estimates from the available body of evidence. Evidence certainty ranges from high certainty to very low certainty [1].

The policy question was, “Should Pfizer RSVpreF vaccine be recommended for pregnant people to be given during 32 through 36 weeks gestation to prevent RSV lower respiratory tract infection in infants?” The benefits chosen by the ACIP Maternal/Pediatric RSV Work Group (Work Group) as critical or important to policy decisions were prevention of: medically attended RSV-associated lower respiratory tract infection (LRTI) in infants (critical); hospitalization for RSV-associated LRTI in infants (critical); intensive care unit (ICU) admission from RSV hospitalization in infants (important); mechanical ventilation from RSV hospitalization in infants (important); RSV-associated death in infants (important); all-cause medically attended LRTI in infants (important); and all-cause hospitalization for LRTI in infants (important). The harms chosen by the Work Group as critical or important to policy decisions were serious adverse events in pregnant people (critical), reactogenicity (grade ≥3) in pregnant people (important); serious adverse events in infants (critical) and preterm birth (<37 weeks) (important). A systematic review of evidence on the efficacy and safety of Pfizer RSVpreF vaccine among pregnant persons identified one phase 3 randomized controlled trial (RCT) and one phase 2b RCT; these two RCTs, plus additional data provided by the sponsor and FDA, were assessed using the GRADE approach [2-4].

A lower risk of medically attended RSV-associated LRTI* in infants^† was observed with vaccination compared with placebo (relative risk [RR] 0.487, 97.58%^§ confidence interval [CI]: 0.332, 0.706, evidence certainty: high), vaccine efficacy of 51.3% (97.58% CI: 29.4%, 66.8%). A lower risk of hospitalization for RSV-associated LRTI in infants was also observed (RR 0.432; 99.17% CI: 0.193, 0.899; evidence certainty: moderate), vaccine efficacy 56.8% (99.17% CI: 10.1%, 80.7%). A possible lower risk of ICU admission from RSV hospitalization in infants was observed with vaccination compared with placebo (RR: 0.571 95% CI: 0.123, 2.248; evidence certainty low), vaccine efficacy 42.9% (95% CI: -124.8%, 87.7%). A lower risk of mechanical ventilation from RSV hospitalization was also observed (RR: 0.001 95% CI: 0.001, 1.091, evidence certainty low), vaccine efficacy 100% (95% CI: -9.1%, 100%). A lower risk of all-cause medically attended LRTI in infants was not observed (RR 0.975; 99.17% CI 0.806, 1.179; evidence certainty: moderate), vaccine efficacy 2.5% (99.17% CI: -17.9, 19.4). A possible lower risk of all-cause hospitalization for LRTI in infants was observed (RR: 0.711 95% CI: 0.492, 1.020, evidence certainty moderate), vaccine efficacy 28.9% (95% CI: -2.0%, 50.8%). The trial was not powered to detect a lower risk of death due to RSV.^¶

In terms of harms, available data was pooled from the phase 3 and phase 2b RCTs.  The evidence indicated that serious adverse events (SAEs)** in pregnant people were balanced between the vaccine and placebo arms (RR 1.06; 95% CI: 0.95, 1.17; evidence certainty: low). Reactogenicity grade ≥3^†† was similar between the vaccine and placebo arms of the trials (RR 0.97; 95% CI: 0.72, 1.31; evidence certainty: moderate), with 2.3% of vaccine recipients and 2.3% of placebo recipients reporting any systemic reactions grade ≥3 following injection. For infants, SAEs were balanced between the vaccine and placebo arms (RR 1.01; 95% CI: 0.91, 1.11; evidence certainty: low). Preterm births^§§ were unbalanced between the vaccine and placebo arms, with more preterm births among RSVpreF vaccine recipients than placebo recipients (RR 1.20; 95% CI: 0.99, 1.46; evidence certainty: very low).

Introduction

On August 21, 2023, the U.S. Food and Drug Administration (FDA) approved the Biologics License Application (BLA) for a single dose of RSV vaccine (Abrysvo, Pfizer Inc.) for administration in pregnant people at 32 through 36 weeks for prevention of RSV in infants [5]. Pfizer RSV vaccine is a bivalent (RSV A and B) stabilized prefusion F protein subunit vaccine (RSVpreF).  As part of the process employed by the ACIP, a systematic review and GRADE evaluation of the evidence for Pfizer RSV vaccine was conducted and presented to ACIP. ACIP adopted a modified GRADE approach in 2010 as the framework for evaluating the scientific evidence that informs recommendations for vaccine use. Evidence of benefits and harms were reviewed based on the GRADE approach [1]. No conflicts of interest were reported by CDC or Work Group members involved in the GRADE analysis.

In the Pfizer phase 2b and 3 trials, vaccination was given during 24 through 36 weeks gestation. In both the phase 2b and phase 3 trials, a numerical imbalance in preterm births in RSV vaccine recipients was observed compared with placebo recipients, with more preterm births among RSVpreF vaccine recipients than placebo recipients. However, available data were insufficient to establish or exclude a causal relationship between preterm birth and Pfizer RSV vaccine. FDA approved the use of the Pfizer RSV vaccine in pregnant people at 32 through 36 weeks gestation to reduce the potential risk of preterm birth and complications from preterm birth [6]. For the GRADE assessment, data were included from phase 2b and phase 3 trials using the trial dosing interval of 24–36 weeks gestation window, as per the original trial design. Utilizing all available data from the trial dosing interval increased power to allow better detection of potential benefits and harms. The policy question was, “Should Pfizer RSVpreF vaccine be recommended for pregnant people to be given during 32 through 36 weeks gestation to prevent RSV lower respiratory tract infection in infants? (Table 1).

Methods

We conducted a systematic review of evidence on the efficacy and safety of Pfizer bivalent RSVpreF vaccine. We assessed outcomes and evaluated the quality of evidence using the GRADE approach from February–September 2023.

The Work Group was asked to pre-specify and rate the importance of relevant patient-important outcomes before the GRADE assessment. Outcomes of interest included individual benefits and harms (Table 2). The critical benefits of interest for infants selected by the Work Group were medically attended RSV-associated LRTI and hospitalization for RSV-associated LRTI. Important benefits of interest in infants were ICU admissions from RSV hospitalization, mechanical ventilation from RSV hospitalization, RSV-associated death, all-cause medically attended LRTI, and all-cause hospitalization for LRTI. Harm outcomes rated by the workgroup as critical were SAEs in both pregnant people and infants as well as preterm birth. The outcome of reactogenicity (grade ≥3) in pregnant people was rated as important.

A systematic literature search was completed to review all available evidence on the efficacy and safety of Pfizer RSVpreF vaccine. Records of relevant observational studies as well as randomized controlled trials were included if they 1) provided data on pregnant persons vaccinated with RSVpreF; 2) involved human subjects; 3) reported primary data; and 4) included data relevant to the efficacy and safety outcomes being measured. We identified relevant studies through Medline, Embase, Cochrane Library, CINAHL, Scopus, and clinicaltrials.gov. Relevant observational studies were restricted to the defined population, intervention, comparison, and outcome outlined in the policy question, or related outcomes if direct data were not available. In addition, unpublished relevant data were obtained through direct communication with the vaccine manufacturer. The systematic review was conducted through April 2023. Characteristics of all included studies are shown in Appendix 1 and evidence retrieval methods are found in Appendix 2.

The evidence certainty assessment addressed risk of bias, inconsistency, indirectness, imprecision, and other characteristics. The GRADE assessment across the body of evidence for each outcome was presented in an evidence profile; evidence could be assessed as certainty of high, moderate, low, or very low certainty.

For benefit outcomes of medically attended RSV-associated LRTI, hospitalization for RSV-associated LRTI, and all-cause medically attended LRTI, vaccine efficacy was published by Kampmann et al. 2023 and was calculated as 1–(P/[1–P]), where P is the number of cases in the RSVpreF group divided by the total number of cases [2]. Risk ratios for GRADE were estimated by 1-(Vaccine Efficacy/100). 95% confidence intervals were calculated unless different confidence intervals were specifically agreed upon by the sponsor and FDA. Data on the outcomes of ICU admission from RSV hospitalization in infants, mechanical ventilation from RSV hospitalization, and all-cause hospitalization for LRTI in infants were provided directly from the manufacturer as a post-hoc analysis using the same method for calculating vaccine efficacy as the Kampmann publication [2]. For the harm outcomes, risk ratios were calculated using counts of events and total participants available in the body of evidence, including both the phase 3 and phase 2b RCTs (estimates were pooled using R version 4.1.2).

Results

The results of the GRADE assessment were presented to ACIP on June 22, 2023 and September 22, 2023. Data were reviewed from one published phase 3 RCT, one published phase 2b RCT, plus additional data provided by the sponsor and FDA [2-4, 6].

For benefit outcomes, the Pfizer RSV vaccine reduced the risk of medically attended RSV LRTI in infants 0–180 days after birth (vaccine efficacy: 51.3% [97.58% CI: 29.4%, 66.8%]) (Table 3a). Pfizer RSVpreF also reduced the risk of hospitalization for RSV-associated LRTI in infants 0–180 days after birth (vaccine efficacy: 56.8% [99.17% CI: 10.1%, 80.7%]) (Table 3b). Pfizer RSVpreF may be effective in preventing ICU admission for RSV hospitalization in infants 0–180 days after birth (vaccine efficacy: 42.9% [95% CI: -124.8, 87.7]) (Table 3c). Pfizer RSVpreF may be effective in preventing mechanical ventilation for RSV hospitalization in infants 0–180 days after birth (vaccine efficacy: 100% [95% CI: -9.1, 100]) (Table 3d). The vaccine was not effective in preventing all-cause medically attended LRTI in infants 0–180 days after birth (vaccine efficacy: 2.5% [99.17% CI: -17.9%, 19.4%]) (Table 3e). Pfizer RSVpreF may be effective in preventing all-cause hospitalization for LRTI in infants 0–180 days after birth (vaccine efficacy: 28.9% [95% CI: -2.0, 50.8]) (Table 3f). One RSV-associated death occurred in the placebo arm of the phase 3 RCT that was recorded at day 120 after birth. No RSV-associated deaths were recorded in the phase 2b trial.

For evaluation of potential harms, SAEs in pregnant people were similar between the vaccine group and the placebo group (RR: 1.06; 95% CI: 0.95, 1.17) (Table 3g). SAEs in four vaccine recipients (pain in an arm followed by bilateral lower extremity pain, premature labor, systemic lupus erythematosus, and eclampsia) and in one placebo recipient (premature placental separation) were assessed by the investigator as being related to the injection. Based on review of the event narratives and temporal association of these events to vaccination, FDA agreed with the investigator’s assessments that there was a reasonable possibility that these events were related to the study intervention [5].  Grade ≥3 systemic reactogenicity events in pregnant people were reported by 2.3% of vaccine recipients and 2.3% of placebo recipients (RR: 0.97; 95% CI: 0.72, 1.31) (Table 3h). SAEs in infants were similar between the vaccine and placebo groups (RR: 1.01; 95% CI: 0.91, 1.11) (Table 3i). No serious adverse events in infants were considered by the investigators to be related to the vaccine. The FDA agreed with the investigator’s conclusions for 4 out of 5 of the infant deaths in the RSVpreF group. For 1 case of extreme prematurity in an infant born to an 18-year-old mother at 10 days after vaccination who died from prematurity-related complications, FDA was unable to exclude the possibility of the extreme prematurity and subsequent death being related to receipt of the investigational product. No non-fatal SAEs in infant participants were considered related to maternal vaccination by FDA [5]. Preterm births were unbalanced between the vaccine and placebo arms, with more preterm births among RSVpreF vaccine recipients than placebo recipients (RR: 1.20; 95% CI: 0.99, 1.46) (Table 3j).

GRADE Summary

The initial GRADE evidence level was high for each outcome because the body of evidence consisted of randomized controlled trials [1]. Of note, these trials were conducted during the COVID-19 pandemic which did impact RSV incidence, but outcomes were not downgraded for indirectness because this was judged to be not a serious concern of bias.

In terms of critical benefits, the available data indicated that the vaccine was effective at preventing medically attended RSV-associated LRTI with no concerns in the certainty assessment. The available data indicated that the vaccine was effective at preventing hospitalization for RSV-associated LRTI with moderate certainty. Certainty was downgraded once for serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered.

The available data showed that RSVpreF vaccine may be effective at preventing the important outcomes of ICU admission from RSV hospitalization and mechanical ventilation from RSV hospitalization with low certainty. Certainty in estimates of both of these outcomes of ICU admission from RSV hospitalization and mechanical ventilation from RSV hospitalization were downgraded twice for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered as well as fragility of the estimates.

The data available for the important benefit of all-cause medically attended LRTI, showed that the vaccine was not effective at preventing the outcome with moderate certainty. Certainty was downgraded once for serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered.

The data available for the important benefit of all-cause hospitalization for LRTI, showed that the vaccine was effective at preventing the outcome with moderate certainty. Certainty was downgraded once for serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered.

For the critical harm outcomes of SAEs in both pregnant people and infants, the available data showed that SAEs were balanced between the vaccine and placebo groups with low certainty. Certainty was downgraded once for serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered and once for serious concern for indirectness as 55% of the phase 3 RCT and 62% of the phase 2b RCT did not receive vaccine or placebo in the approved dosing interval (32–36 weeks gestation). In the approved dosing interval, there is less opportunity for serious adverse events, including preterm birth/delivery compared with the trial dosing interval (24–36 weeks gestation).

The data available on the critical harm of preterm birth showed that preterm births were unbalanced between the vaccine and placebo groups, with more preterm births among RSVpreF vaccine recipients than placebo recipients, with very low evidence certainty. Certainty was downgraded twice for very serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered and not meeting optimal information size requirements; certainty was additionally downgraded once for serious concern for indirectness as 55% of the phase 3 RCT and 62% of the phase 2b RCT did not receive vaccine or placebo in the approved dosing interval (32–36 weeks gestation). In the approved dosing interval, there is less opportunity for preterm birth compared with the trial dosing interval (24–36 weeks gestation).

The data available on the important harm of reactogenicity (grade ≥3) in pregnant people showed that reactogenic events were balanced between the vaccine and placebo groups with moderate certainty. Certainty was downgraded once for serious concern for indirectness as these data only include systemic reactions. When selecting the a priori harm outcomes, the Work Group defined reactogenicity as both local and systemic events (Table 4).

Footnotes

* Medically-attended LRTI was defined as a medical encounter (inclusive of both outpatient and inpatient settings) with ≥1 symptom of fast breathing (respiratory rate ≥60 bpm [<2 months of age] or ≥50 bpm [≥2 to 12 months of age]), peripheral blood oxygen saturation measured in room air <95%, or chest wall indrawing, plus a positive validated reverse transcription-polymerase chain reaction in central laboratory or by a certified laboratory with nucleic acid amplification test, and confirmation by endpoint adjudication committee

^† 0–180 days after birth

^§ Confidence intervals that are not 95% were adjusted using the Bonferroni procedure and accounting for the primary endpoints results.

^¶ There was 1 RSV-associated death in a placebo recipient in the phase 3 trial. There were no RSV-associated deaths recorded in the Phase 2b study.

** Serious adverse event is defined as any untoward medical occurrence that results in death, is life threatening, requires inpatient hospitalization or prolongation of existing hospitalization, results in persistent disability or incapacity, or is a congenital anomaly or birth defect.

^†† Grade 3: prevents daily routine activity. For redness or swelling is >10 cm. For vomiting, requires intravenous hydration. For diarrhea, includes 6 or more loose stools in 24 hours. Grade 4: requires emergency room visit or hospitalization; for redness included necrosis or exfoliative dermatitis; for swelling included necrosis.

^§§ Footnote: Preterm birth was defined as birth at <37 weeks gestational age

References

1. Ahmed F. U.S. Advisory Committee on Immunization Practices (ACIP) Handbook for Developing Evidence-based Recommendations.
2. Kampmann B, Madhi SA, Munjal I, et al. Bivalent Prefusion F Vaccine in Pregnancy to Prevent RSV Illness in Infants. New England Journal of Medicine. 388(16): 1451-1464. https://doi.org/10.1056/NEJMoa2216480
3. Simões EAF, Center KJ, Tita ATN, et al. Prefusion F Protein–Based Respiratory Syncytial Virus Immunization in Pregnancy. New England Journal of Medicine. 2022. 386(17): 1615-1626. https://doi.org/10.1056/NEJMoa2106062
4. Fleming-Dutra KE. Evidence to Recommendations Updates. Presentation to ACIP September 2023. https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2023-09-22/06-Mat-Peds-Fleming-Dutra-508.pdf
5. Food and Drug Administration (FDA). FDA Approves First Vaccine for Pregnant Individuals to Prevent RSV in Infants. Press briefing. https://www.fda.gov/news-events/press-announcements/fda-approves-first-vaccine-pregnant-individuals-prevent-rsv-infants
6. Food and Drug Administration (FDA). Abrysvo package insert. https://www.fda.gov/media/171482/download?attachment
7. Food and Drug Administration (FDA). FDA Briefing Document Pfizer RSV vaccine. https://www.fda.gov/media/168185/download

Table 1: Policy Question and PICO

Abbreviations: RSV= respiratory syncytial virus; IM = intramuscular; LRTI= lower respiratory tract infection; ICU= intensive care unit

Table 2: Outcomes and Rankings

Abbreviations: RSV= respiratory syncytial virus; LRTI= lower respiratory tract infection; ICU= intensive care unit

^a 1 RSV-associated death occurred in the placebo arm of the phase 3 trial that was recorded at day 120 after birth. No RSV-associated deaths were recorded in the phase 2b trial.

Table 3a. Summary of studies reporting RSV medically attended RSV-associated LRTI in infants^a

Table 3b: Summary of studies reporting hospitalization for RSV-associated LRTI in infants^a

Abbreviations: LRTI= Lower Respiratory Tract Infection; CI = confidence interval
^a 0–180 days after birth
^b Based on data cutoff September 30, 2022
^c Intervention was a single dose of Pfizer bivalent RSVpreF vaccine (120µg antigen, 1 dose IM).
^d Efficacy was calculated 1–(P/[1–P]), where P is the number of cases in the RSVpreF group divided by the total number of cases.
^e The confidence interval was adjusted using the Bonferroni procedure and accounting for the primary endpoints results.

Table 3c: Summary of studies reporting ICU admission from RSV hospitalization in infants^a

Table 3d: Summary of studies reporting mechanical ventilation from RSV hospitalization in infants^a

Table 3e: Summary of studies reporting all-cause medically attended LRTI in infants^a

Table 3f: Summary of studies reporting all-cause hospitalizations for LRTI in infants^a

Table 3g: Summary of studies reporting serious adverse events^a in pregnant people

Table 3h: Summary of studies reporting reactogenicity^a (grade ≥3)^b in pregnant people

Table 3i: Summary of studies reporting serious adverse events^a in infants

Table 3j: Summary of studies reporting preterm births^a

Table 4. Grade Summary of Findings Table

CI: confidence interval; RR: risk ratio

Explanations

1. The effects noted are from a per protocol analysis. In a modified intention to treat analysis there were 60 cases among 3568 persons in the vaccine arm and 118 cases among 3558 persons in the placebo arm (RR = 0.51 [0.37 to 0.69])
2. The clinical trial participants consisted of healthy people ≤49 years of age with an uncomplicated, singleton pregnancy, who were at no known increased risk for complications. The clinical trial population may not be representative of all pregnant people and their infants in the United States.
3. Absolute risk was calculated using the observed outcomes in the placebo arm during the available clinical trial follow-up. Absolute risk estimates should be interpreted in this context.
4. Rate used from Lively JY, Curns AT, Weinberg GA, et al. Respiratory Syncytial Virus–Associated Outpatient Visits Among Children Younger Than 24 Months, Journal of the Pediatric Infectious Diseases Society, Volume 8, Issue 3, July 2019, Pages 284–286, https://doi.org/10.1093/jpids/piz011
5. Rate calculated from Lively et al study, assuming 47.5% of ARI cases are LRTI. Rainisch G, Adhikari B, Meltzer MI, Langley G. Estimating the impact of multiple immunization products on medically attended respiratory syncytial virus (RSV) infections in infants. 2020. Vaccine 38(2): 251-257. ISSN 0264-410X. https://doi.org/10.1016/j.vaccine.2019.10.023.
6. The clinical trial excluded pregnant persons with certain conditions and therefor may not represent all pregnant people and their infants in the United States
7. Serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered.
8. Rate of acute respiratory infection (ARI) hospitalizations for infants 0-5 months (2016-2020 NVSN, unpublished)
9. Very serious concern for imprecision due to the confidence interval containing estimates for which different policy decision might be considered and fragility of the estimate.
10. ICU admission rate of hospitalizations of 35% from Arriola CS, Kim L, Langley G, et al. Estimated Burden of Community-Onset Respiratory Syncytial Virus–Associated Hospitalizations Among Children Aged <2 Years in the United States, 2014–15, Journal of the Pediatric Infectious Diseases Society, Volume 9, Issue 5, November 2020, Pages 587–595, https://doi.org/10.1093/jpids/piz087. Using NVSN hospitalization rate (unpublished)
11. Mechanical ventilation rate of 11% of hospitalizations from Arriola CS, Kim L, Langley G, et al. Estimated Burden of Community-Onset Respiratory Syncytial Virus–Associated Hospitalizations Among Children Aged <2 Years in the United States, 2014–15, Journal of the Pediatric Infectious Diseases Society, Volume 9, Issue 5, November 2020, Pages 587–595, https://doi.org/10.1093/jpids/piz087. Using NVSN hospitalization rate (unpublished)
12. Data were pooled from one phase 2B trial and one phase 3 trial
13. Serious concern for indirectness as 55% of the phase 3 RCT and 62% of the phase 2b RCT did not receive vaccine or placebo in the approved dosing interval (32–36 weeks gestation). In the approved dosing interval, there is less opportunity for serious adverse events, including preterm birth/delivery, compared with the trial dosing interval (24–36 weeks gestation).
14. Serious concern for indirectness as this measure only contains information on systemic reactogenicity. Phase 3 local reactogenicity grade ≥3: 11/3660 vaccine arm;| 0/3639 placebo arm. Phase 2b local reactogenicity grade ≥3: 0/114 vaccine arm; 0/117 placebo arm.
15. Very serious concern for imprecision due to the width of the confidence interval containing estimates for which different policy decisions might be considered and not meeting optimal information size requirements.

Appendices

Appendix 1. Studies Included in the Review of Evidence

Appendix 2. Databases and strategies used for systematic review^a