Intended for healthcare professionals

CCBYNC Open access
Research

Waning of vaccine effectiveness against moderate and severe covid-19 among adults in the US from the VISION network: test negative, case-control study

BMJ 2022; 379 doi: https://doi.org/10.1136/bmj-2022-072141 (Published 03 October 2022) Cite this as: BMJ 2022;379:e072141
  1. Jill M Ferdinands, research epidemiologist1,
  2. Suchitra Rao, associate professor of pediatrics2,
  3. Brian E Dixon, director of public health informatics3 4,
  4. Patrick K Mitchell, senior epidemiologist5,
  5. Malini B DeSilva, internal medicine specialist6,
  6. Stephanie A Irving, project director7,
  7. Ned Lewis, data manager8,
  8. Karthik Natarajan, assistant professor of biomedical informatics9 10,
  9. Edward Stenehjem, infectious disease specialist11,
  10. Shaun J Grannis, vice president of data analytics3 12,
  11. Jungmi Han, research analyst9,
  12. Charlene McEvoy, internal medicine specialist6,
  13. Toan C Ong, research instructor2,
  14. Allison L Naleway, senior epidemiologist7,
  15. Sarah E Reese, senior biostatistician5,
  16. Peter J Embi, professor of medicine,
  17. Kristin Dascomb, medical director infection prevention11,
  18. Nicola P Klein, senior research scientist8,
  19. Eric P Griggs, epidemiologist1,
  20. I-Chia Liao, analytics developer13,
  21. Duck-Hye Yang, senior epidemiologist5,
  22. William F Fadel, clinical assistant professor3 4,
  23. Nancy Grisel, analyst11,
  24. Kristin Goddard, senior research manager8,
  25. Palak Patel, epidemiologist1,
  26. Kempapura Murthy, SAS programmer13,
  27. Rebecca Birch, senior epidemiologist5,
  28. Nimish R Valvi, postdoctoral fellow3,
  29. Julie Arndorfer, analyst11,
  30. Ousseny Zerbo, research scientist8,
  31. Monica Dickerson, epidemiologist1,
  32. Chandni Raiyani, biostatistician13,
  33. Jeremiah Williams, surveillance coordinator1,
  34. Catherine H Bozio, epidemiologist1,
  35. Lenee Blanton, research epidemiologist1,
  36. Ruth Link-Gelles, epidemiologist1,
  37. Michelle A Barron, senior medical director2,
  38. Manjusha Gaglani, chief of pediatric infectious diseases13,
  39. Mark G Thompson, epidemiologist1,
  40. Bruce Fireman, biostatistician8
  1. 1Centers for Disease Control and Prevention COVID-19 Response Team, Atlanta, GA, USA
  2. 2Department of Pediatrics, University of Colorado Anschutz Medical Campus, Aurora, CO, USA
  3. 3Center for Biomedical Informatics, Regenstrief Institute, Indianapolis, IN, USA
  4. 4Fairbanks School of Public Health, Indiana University, Indianapolis, IN, USA
  5. 5Westat, Rockville, MD, USA
  6. 6HealthPartners Institute, Minneapolis, MN, USA
  7. 7Center for Health Research, Kaiser Permanente Northwest, Portland, OR, USA
  8. 8Kaiser Permanente Vaccine Study Center, Kaiser Permanente Northern California Division of Research, Oakland, CA, USA
  9. 9Department of Biomedical Informatics, Columbia University Irving Medical Center, New York, NY, USA
  10. 10New York Presbyterian Hospital, New York, NY, USA
  11. 11Division of Infectious Diseases and Clinical Epidemiology, Intermountain Healthcare, Salt Lake City, UT, USA
  12. 12Indiana University School of Medicine, Indianapolis, IN, USA
  13. 13Baylor Scott &White Health, Temple, TX, USA
  1. Correspondence to: J M Ferdinands zdn5{at}cdc.gov
  1. Patients aged <50 years were excluded from estimates of fourth dose effectiveness; thus, column sum might not equal 100% of encounters.

  • Accepted 9 September 2022

Abstract

Objective To estimate the effectiveness of mRNA vaccines against moderate and severe covid-19 in adults by time since second, third, or fourth doses, and by age and immunocompromised status.

Design Test negative case-control study.

Setting Hospitals, emergency departments, and urgent care clinics in 10 US states, 17 January 2021 to 12 July 2022.

Participants 893 461 adults (≥18 years) admitted to one of 261 hospitals or to one of 272 emergency department or 119 urgent care centers for covid-like illness tested for SARS-CoV-2.

Main outcome measures The main outcome was waning of vaccine effectiveness with BNT162b2 (Pfizer-BioNTech) or mRNA-1273 (Moderna) vaccine during the omicron and delta periods, and the period before delta was dominant using logistic regression conditioned on calendar week and geographic area while adjusting for age, race, ethnicity, local virus circulation, immunocompromised status, and likelihood of being vaccinated.

Results 45 903 people admitted to hospital with covid-19 (cases) were compared with 213 103 people with covid-like illness who tested negative for SARS-CoV-2 (controls), and 103 287 people admitted to emergency department or urgent care with covid-19 (cases) were compared with 531 168 people with covid-like illness who tested negative for SARS-CoV-2. In the omicron period, vaccine effectiveness against covid-19 requiring admission to hospital was 89% (95% confidence interval 88% to 90%) within two months after dose 3 but waned to 66% (63% to 68%) by four to five months. Vaccine effectiveness of three doses against emergency department or urgent care visits was 83% (82% to 84%) initially but waned to 46% (44% to 49%) by four to five months. Waning was evident in all subgroups, including young adults and individuals who were not immunocompromised; although waning was morein people who were immunocompromised. Vaccine effectiveness increased among most groups after a fourth dose in whom this booster was recommended.

Conclusions Effectiveness of mRNA vaccines against moderate and severe covid-19 waned with time after vaccination. The findings support recommendations for a booster dose after a primary series and consideration of additional booster doses.

Introduction

Randomized trials of BNT162b2 (Pfizer-BioNTech) and mRNA-1273 (Moderna) vaccines showed 94-95% protection against covid-19 among adults and suggested efficacy against covid-19 requiring hospital admission.12 Since the introduction of these vaccines in December 2020, evidence has accumulated that their effectiveness wanes over time since vaccination, especially against milder disease,3456789 they are less effective against omicron than earlier SARS-CoV-2 variants,10 and a third (booster) dose restores high effectiveness against severe disease.10111213 Although protection against severe omicron related disease is believed to be high for several months after a third dose, the durability of protection and how this effect can vary by age group, immunocompromised status, and vaccine product is uncertain. In March 2022, the US Centers for Disease Control and Prevention recommended a second booster dose only for specific subgroups at high risk (such as adults aged 50 and older).14 A more complete understanding of the effectiveness and durability of third and fourth doses of the mRNA vaccines is important to inform policy about booster doses.

The CDC’s VISION network previously examined the effectiveness of mRNA vaccines against admissions to hospital or emergency visits and urgent care visits associated with covid-19, with data from eight healthcare systems.15 In this article, we update VISION’s analyses of mRNA vaccine effectiveness, focusing on the durability of three and four dose protection against severe disease (ie, admission to hospital) during the omicron period. We assess the trajectory of vaccine effectiveness overall and in subgroups defined by age, immunocompromised status, and vaccine product.

Methods

Study design

The VISION network has been described previously.15 We applied a test negative design to estimate vaccine effectiveness of mRNA vaccines using retrospectively collected data. We focused on mRNA vaccines because they comprise more than 95% of covid vaccines administered in the US.16 Separate analyses were done of patients who were admitted to hospital (hospital sample) and patients who received care in an emergency department or urgent care clinic (emergency department or urgent care sample).

Study population and setting

The study population included adults (≥18 years) who received care for covid-like illness at a VISION network hospital or emergency department or urgent care center and had molecular testing for SARS-CoV-2 at least 14 days after vaccines became locally available for their age group (17 January to 3 May 2021). The last contact included in this study period occurred on 12 July 2022. We excluded individuals who received any vaccine other than the BNT162b2 or mRNA-1273 vaccine, individuals who received more than four doses of an mRNA vaccine before the index medical contact, individuals who received only one dose of an mRNA vaccine less than 14 days before the index contact or who had a third or fourth dose less than seven days before the index contact, individuals known to have a positive laboratory test result for a SARS-CoV-2 infection more than 14 days before the index contact, and individuals with a positive SARS-CoV-2 test result but no diagnoses or symptoms suggesting covid-19 illness.

Vaccination status

Vaccination status was categorized by the number of doses received and the number of months between the most recent vaccine dose and the index contact date (referred to as time since vaccination). Patients were considered partially vaccinated if they received only one dose at least 14 days prior to the index contact date or had received a second dose less than 14 days previously. Patients with no record of vaccination before the index contact date were considered unvaccinated. Patients with three doses were those who received a third dose in a primary vaccination series (eg, among immunocompromised individuals) or a booster dose after a primary series of two doses. Aligning with recommendations for receipt of a fourth dose, we examined the effectiveness of four doses among adults aged 50 years or older and among immunocompromised adults of any age. Vaccination status was ascertained from immunization registries, electronic health records, and insurance claims.

Outcomes

The primary outcome was a positive or negative molecular SARS-CoV-2 result for a test done within 14 days before a medical contact to less than 72 h after among patients presenting with covid-like illness, as identified from ICD-9 and ICD-10 (international classification of diseases, ninth and 10th revision, respectively) diagnostic codes (supplemental methods; supplemental table S1). The index date for each contact was the earlier of either the contact date or the date of the closest SARS-CoV-2 molecular assay. An individual could be included as a case once in the emergency department or urgent care sample and once in the hospital sample. Individuals could be included as a control multiple times.

Statistical analysis

We used a test negative case-control design in which cases were patients with covid-like illness with laboratory confirmed covid-19 and controls were patients with covid-like illness and negative SARS-CoV-2 test results (controls could have had positive test results for other respiratory viruses such as influenza). We compared cases with controls in the hospital sample, and separately compared cases with controls in the emergency or urgent care sample. Cases were not individually matched to controls.

Conditional logistic regression was used to examine case-control status in relation to vaccination status categorized as vaccinated with four, three, or two doses, or partially vaccinated; unvaccinated individuals were used as the reference group. To examine waning of vaccine effectiveness, we categorized people who were vaccinated using time specific indicators defined by two month intervals of time since vaccination; unvaccinated individuals were used as the reference group. We exponentiated the regression coefficient of each vaccination status indicator to yield an odds ratio, subtracted the odds ratio from 1 to estimate vaccine effectiveness, and multiplied by 100 to scale vaccine effectiveness as a percentage. In several analyses, a sparse bimonthly interval for which the vaccine effectiveness estimate had a confidence interval wider than 50 percentage points was combined with the previous bimonthly interval to provide a more precise estimate of vaccine effectiveness (see supplemental methods). Vaccine effectiveness estimates (and confidence limits) were scaled to a range of –100% to 100%.17

Logistic regression models were conditioned by calendar week and geographical area such that we compared cases with controls tested during the same week in the same region (supplemental table S2). Covariates included in the models were those determined through bivariate analyses to be statistically significantly associated with both the outcome and vaccination status, as well as those specified a priori as established confounders, including age, race, ethnicity, presence of respiratory and non-respiratory comorbidities, immunocompromised status, and local viral circulation. Cubic splines were used for age, seven day average positivity of SARS-CoV-2 test in the area of the contact, and the propensity to be vaccinated; others were indicator variables. Propensity scores (supplemental methods) predicted vaccination (any versus none) based on demographics, comorbidities (supplemental table S3), and characteristics of the facility (supplemental table S4), and were derived independently for each period of variant dominance (supplemental table S5). Patients who were immunocompromised were identified by ICD-9 and ICD-10 diagnostic codes (supplemental methods).18 We conducted separate analyses for three periods based on when a variant accounted for 50% or more of sequenced isolates in each site: before delta was predominant, when delta was predominant, and when omicron was predominant (supplemental table S6). We assessed the magnitude of waning as the difference in vaccine effectiveness between patients who had recently been vaccinated (defined as less than two months) and patients at a specified level of time since vaccination (eg, four to five months from dose 3), and we examined waning by age (18-44 years, 45-64 years, ≥65 years), vaccine product, and immunocompromised status. Bootstrapping was used to estimate a 95% confidence interval around the difference between vaccine effectiveness at less than two months and vaccine effectiveness at four to five months.

We conducted several sensitivity analyses. First, we added to the study population patients with a known prior infection to assess the sensitivity of results to whether previously infected patients are included or excluded.. Second, we wanted to distinguish results between patients who had been admitted to hospital and patients who had been admitted to an emergency department or to urgent care. Therefore, we examined vaccine effectiveness in the emergency department or urgent care sample and omitted patients admitted to hospital within 30 days. Third, we investigated a negative control exposure19 by examining vaccine effectiveness in patients who received their first dose less than 14 days before the index date of contact. These patients were not expected to have substantial vaccine induced protection, and a vaccine effectiveness estimate substantially more than zero would be evidence of residual confounding.20

Analyses were conducted with SAS version 9.4 and R version 4.1.2. All confidence limits are 95% intervals. Confidence intervals excluding the null value were considered statistically significant.

Patient and public involvement

Study participants contributed in important ways to this research by supplying the underlying data on which the study is based. It was not, however, feasible to involve them in the design, conduct, reporting, or dissemination of this study because the study was conducted under the CDC’s covid-19 incident response structure and limited to analysis of retrospectively collected electronic data only, with no patient interaction.

Results

Study population

From 17 January 2021 to 12 July 2022, 259 006 patients were admitted to 261 hospitals and 634 455 were admitted to 272 emergency departments or to 119 urgent care centers. The hospital sample included 17 446 people with covid-19 during the omicron period, 23 379 during the delta period, and 5078 before delta was dominant. The emergency department or urgent care sample included 57 174 people with covid-19 during the omicron period, 39 909 during the delta period, and 6204 before delta was dominant (table 1; supplementary figs S1-S18).

Table 1

Characteristics of adults with covid-19-like illness who were admitted to hospital or to an emergency department or urgent care, and percentage with laboratory confirmed SARS-CoV-2 infection. Data are number of patients (percentage of column or row) unless stated otherwise

View this table:

In the hospital sample, the median age was 69 years (interquartile range 56-79, 11.2% were black participants, 9.8% were Hispanic, and 23.3% had an immunocompromising condition. In the emergency department or urgent care sample, the median age was 51 years (interquartile range 33-69), 11.0% were black participants, 13.3% were Hispanic, and 4.5% had an immunocompromising condition (table 1). Characteristics by vaccination status are given in supplemental tables S7 and S8. Median times between the last vaccination date and index contact date in the hospital sample were 173 (interquartile range 97-248) days for two doses, 105 (56-156) days for three doses, and 33 (19-50) days for four doses, and in the emergency department or urgent care sample were 179 (110-247) days for two doses, 100 (52-155) days for three doses, and 34 (20-52) days for four doses.

Vaccine effectiveness

Vaccine effectiveness estimates from the hospital and emergency department or urgent care samples are shown in figures 1 and figure 2 and detailed in supplemental tables S9-S14.

Fig 1
Fig 1

Vaccine effectiveness (%) against covid-19-associated hospital admissions by time since vaccination and period of variant predominance. Vaccine effectiveness estimates are adjusted for geographic area, calendar week, age, race, ethnicity, presence of respiratory and non-respiratory comorbidities, immunocompromise status, local viral circulation, and propensity to be vaccinated score. CI=confidence interval; col=column. *Figure 3 presents 4 findings for 4-dose recipients in the subgroups recommended for a fourth dose

Fig 2
Fig 2

Vaccine effectiveness (%) against covid-19-associated emergency department and urgent care visits by time since vaccination and period of variant predominance. Vaccine effectiveness estimates are adjusted for geographic area, calendar week, age, race, ethnicity, presence of respiratory and non-respiratory comorbidities, immunocompromise status, local viral circulation, and propensity to be vaccinated score. CI=confidence interval; col=column. *Supplemental Table 14 presents findings for 4-dose recipients in the subgroups recommended for a fourth dose

Vaccine effectiveness against covid-19 requiring hospital admission was 94% (95% confidence interval 93% to 95%) in the pre-delta period and 96% (95% to 97%) in the delta period, during the initial two months after the second dose. By months four to five after the second dose, vaccine effectiveness against hospital admission decreased to 87% (77% to 93%) in the pre-delta period and 89% (88% to 90%) in the delta period. In the omicron period, two dose vaccine effectiveness against hospital admission was lower than in the earlier periods, both before and when delta was dominant, and waned more, decreasing from 73% (63% to 80%) initially to 57% (51% to 62%) by four to five, and to 40% (32% to 47%) by 12 months after the second dose.

The patterns of vaccine effectiveness estimates from the emergency department or urgent care sample were similar. Vaccine effectiveness of two doses against emergency department or urgent care visits was initially high in the pre-delta period (95%; 94% to 96%) and delta period (93%; 92% to 94%) and then waned. During the omicron period, vaccine effectiveness of two doses against emergency department or urgent care visits was lower initially (63%; 57% to 68%) than in the earlier pre-delta and delta periods and then waned more. From up to one month after the second dose to months four to five, the vaccine effectiveness of a second dose decreased by 9 percentage points (95% confidence interval 4 to 16) during the pre-delta period, by 7 percentage points (7 to 9) during the delta period, and by 26 percentage points (19 to 32) during the omicron period.

A third dose initially restored high levels of protection against both hospital admissions and emergency department or urgent care visits, then began to wane. In the hospital sample, vaccine effectiveness of three doses was initially 96% (95% to 96%) during the delta period and 89% (88% to 90%) during the omicron period. Similarly, in the emergency department or urgent care sample, the vaccine effectiveness of a third dose was initially 96% (95% to 96%) during the delta period and 83% (82% to 84%) during the omicron period. Waning was evident in both samples by four to five months after the third dose during the omicron period, when vaccine effectiveness decreased to 66% (63% to 68%) against hospital admission and to 46% (44% to 49%) against emergency department or urgent care visits.

Vaccine effectiveness against hospital admission after a fourth dose increased to 72% (51% to 83%) in the 50-64 year group and to 76% (71% to 80%) in the 65 years and older age group (fig 3). Similarly, vaccine effectiveness against emergency department or urgent care visits after a fourth dose increased to 57% (47% to 65%) and 73% (69% to 76%) among the 50-64 year and 65 years and older age groups, respectively (supplemental table S14). Vaccine effectiveness of a fourth dose among immunocompromised individuals in the hospital sample was 48% (29% to 62%; fig 4), but we were unable to measure this precisely enough in the emergency department or urgent care sample.

Fig 3
Fig 3

Vaccine effectiveness (%) against covid-19-associated hospital admissions by time since vaccination and age group, restricted to omicron period. Vaccine effectiveness estimates are adjusted for geographic area, calendar week, age, race, ethnicity, presence of respiratory and non-respiratory comorbidities, immunocompromise status, local viral circulation, and propensity to be vaccinated score. CI=confidence interval; col=column. *Patients aged <50 years were excluded from the estimate of fourth dose effectiveness for the subgroup aged 45-64 years.

Fig 4
Fig 4

Vaccine effectiveness (%) against covid-19-associated hospital admissions by time since vaccination and immunocompromise status, restricted to omicron period. Vaccine effectiveness estimates are adjusted for geographic area, calendar week, age, race, ethnicity, presence of respiratory and non-respiratory comorbidities, local viral circulation, and propensity to be vaccinated score. CI=confidence interval; col=column

Vaccine effectiveness in subgroups

In all subgroups examined, vaccine effectiveness waned as time elapsed after the second dose, increased markedly with a third dose, and waned as time elapsed (supplemental tables S9-14). Vaccine effectiveness also substantially improved after a fourth dose among most subgroups for whom this booster dose was recommended. Comparing the initial two months after the third dose with months four to five, vaccine effectiveness against hospital admission during the omicron period decreased by 33 percentage points (95% confidence interval 16 to 56) in the 18-44 years group, 31 (21 to 40) in the 45-64 years group, and 19 (16 to 22) in the 65 years or older group (fig 3,table 2). Results were similar in post hoc analyses that were restricted to individuals without immunocompromising conditions (supplemental table S15).

Table 2

Estimates of mRNA vaccine effectiveness against covid-19 related hospital admissions during omicron period by age group. Data are number of patients (percentage of column or row) unless stated otherwise.

View this table:

Vaccine effectiveness was higher in recipients of the mRNA-1273 than BNT162b2 vaccine in all three variant periods in both the hospital sample and the emergency department or urgent care sample. Vaccine effectiveness waned in recipients of both vaccine products. In the hospital sample during the omicron period, vaccine effectiveness of mRNA-1273 waned from 91% (89% to 92%) to 65% (60% to 70%) by four to five months after three doses whereas vaccine effectiveness of BNT162b2 waned from 88% (86% to 90%) to 66% (63% to 70%) after three doses (table 3).

Table 3

Estimates of vaccine effectiveness against covid-19 related hospital admissions during omicron period by mRNA vaccine product. Data are number of patients (percentage of column or row) unless stated otherwise

View this table:

Vaccine effectiveness after two and three doses was generally lower among individuals who were immunocompromised, in both the hospital and the emergency department or urgent care samples, in each period and at all times since vaccination (fig 4, table 4, supplemental tables S9-S14). In the omicron period, vaccine effectiveness of three doses against hospital admission waned from 78% (73% to 82%) to 48% (40% to 55%) by months four to five in the immunocompromised subgroup compared with 91% (90% to 92%) to 71% (68% to 74%) in the subgroup without immunocompromise (table 4).

Table 4

Estimates of mRNA vaccine effectiveness against covid-19 related hospital admissions during omicron period by immunocompromised status. Data are number of patients (percentage of column or row) unless stated otherwise.

View this table:

Sensitivity analyses

In the first sensitivity analysis, vaccine effectiveness estimates in both samples were similar but slightly lower if patients with previous SARS-CoV-2 infection were included (supplemental tables S16 and S17). In the second sensitivity analysis, vaccine effectiveness estimates were similar but lower if the emergency department or urgent care sample excluded patients who were later admitted to hospital. In the third sensitivity analysis, vaccine effectiveness ranged from –5% to 24% among patients whose index date for medical contact was less than 14 days after the first dose, consistent with the little protection induced by the vaccine during this two week period.

Discussion

Principal findings

Protection against severe omicron related covid-19 was high after three doses of an mRNA vaccine but began to wane less than six months after the third dose. In the hospital sample, vaccine effectiveness after a third doses was 89% among individuals within two months but decreased to 66% among individuals at four to five months. In the emergency department or urgent care sample, vaccine effectiveness of a third dose was 83% within two months but decreased to 46% at four to five months. In all subgroups defined by age, immunocompromised status, and vaccine product, the third dose was initially associated with markedly increased protection, but vaccine effectiveness was lower by four to five months. Vaccine effectiveness increased after a fourth dose for most subgroups for whom this booster dose is recommended in the US. Although we have not yet observed events more than four months from a fourth dose, our results suggest that protection after the fourth dose begins to wane after a few months.

Comparison with other studies

Our vaccine effectiveness estimates for mRNA vaccines are broadly consistent with those in other reports: vaccine effectiveness was lower against the omicron variant than earlier variants,102122 vaccine effectiveness waned after a second dose,3456789 and a third dose restored high levels of protection against severe covid-19 during the omicron and delta periods.10111213 Our results are also consistent with other reports of waning protection after three mRNA doses.232425 As with others, we noted less waning against more severe outcomes,326 lower vaccine effectiveness among individuals who were immunocompromised,1727 and higher vaccine effectiveness among recipients of mRNA-1273 compared with recipients of BNT162b2.102324 We also observed improvement in vaccine effectiveness after a fourth dose.28

Strengths and limitations of this study

One strength of our study is the number and diversity of sites and inclusion of outcomes of varying severity. Additionally, our sample size was large enough to detect modest waning of vaccine protection and to allow stratification of vaccine effectiveness estimates by immunocompromise status. We rigorously controlled for calendar time and geography such that cases were compared with controls tested during the same week in the same geographical area. This comparison allowed us to distinguish differences in vaccine effectiveness attributable to the waning of vaccine induced immunity from those attributable to the change in dominance of SARS-CoV-2 variants.

Our study has limitations. First there is residual confounding if the timing of primary vaccination or booster doses was related to covid-19 risk in unmeasured ways (eg, mask use or occupation). However, we did not observe substantial vaccine protection in the two weeks after a first dose, which provides reassurance that residual confounding is limited. Second, although our test negative design is intended to avoid selection bias from healthcare seeking behavior, the design could induce selection bias arising from factors associated with a covid-like illness but not with covid-19. For example, inclusion of individuals who had influenza as controls could underestimate vaccine effectiveness due to the correlation between covid-19 vaccination and influenza vaccination. Because fewer than 5% of people in the control group in our study were positive for influenza, we expect this bias to be minimal. Also, we cannot rule out selection bias arising from reliance on clinician directed testing, although we note that almost all the patients admitted to hospital with covid-like illness were tested for SARS-CoV-2. Third, immunocompromised status was ascertained only from diagnostic codes at the time of medical contact (without data on prescriptions or laboratory tests), and we could not distinguish whether a third dose was in a primary series for people who were immunocompromised or was a booster dose. Insufficient adjustment for immunocompromised status might have biased vaccine effectiveness estimates downward, especially for those who were vaccinated and received a booster dose relatively early. However, we found waning protection in stratified analyses among both individuals who were immunocompromised and individuals who were not immunocompromised. Fourth, we did not have viral genomic sequence data. Fifth, although we excluded individuals with documented previous SARS-CoV-2 infection, our data might have missed many past infections. Sensitivity analyses that included people with known previous infections suggest that our vaccine effectiveness estimates would be higher if we could have ascertained and excluded everyone with protection induced by infection. Sixth. although we interpret our analyses of the hospital sample as pertaining to severe covid-19, some patients admitted to hospital could have tested positive for other reasons while being in hospital, especially during the omicron period.29 To address this, patients were not eligible for inclusion if they had a positive SARS-CoV-2 test result but no diagnoses suggesting a covid-19 infection. Seventh, although our sample includes enough outcome events to yield precise estimates of vaccine effectiveness for the overall adult population, estimates of vaccine effectiveness against admissions to hospital for covid-19 were less precise for younger adults and individuals who were immunocompromised owing to smaller sample sizes. Finally, we pooled data from heterogeneous populations in 10 US states; however, our findings might not be generalizable to other populations.

Policy implications

To evaluate the clinical significance of waning vaccine effectiveness, consideration of the absolute number of people admitted to hospital that would have been prevented had no waning occurred is helpful. However, this number depends on the background rate of severe covid-19, which sometimes varied 10-fold or more over several weeks. In this context, hospital admissions that would be prevented during an anticipated surge are an appropriate alternative. For example, the rate of hospital admissions related to covid-19 reached about 1500 per million unvaccinated adults each week in January 2022 in the US30; if incidence surges that high again, then for every million adults who lose 20 percentage points of vaccine protection, about 300 additional people each week (1500×0.20) will be admitted to hospital owing to covid-19 compared with no waning effect. During the omicron period, vaccine effectiveness waned within six months of the third dose by about 20 percentage points among those without immunocompromising conditions and by more than 40 percentage points among those with immunocompromising conditions. This amount of waning is enough to be relevant for clinical and policy considerations about the need for boosters or other protective measures. Combined with evidence of the safety and immunogenicity of an additional vaccine dose,313233 our findings lend support for consideration of additional doses beyond the primary series.

Conclusions

Protection conferred by mRNA vaccines against moderate (emergency department or urgent care) and severe (hospital admission) covid-19 waned during the months after primary vaccination, increased substantially after the third dose, and waned again by four to five months. A fourth dose improved vaccine effectiveness among those for whom this booster dose was recommended. Vaccine effectiveness waned less against severe disease than against moderate disease. Vaccine effectiveness of either mRNA vaccine waned among adults of all ages. Among immunocompromised individuals, vaccine effectiveness was lower and waning was more noticeable. These findings support recommendations for a third vaccine dose and consideration of additional booster doses.

What is already known on this topic

  • Studies of the BNT162b2 (Pfizer-BioNTech) and mRNA-1273 (Moderna) covid-19 vaccines suggest that their effectiveness decreases over time and increases with an additional dose

  • How this pattern has varied with the dominant variant and number of vaccine doses, or by age group, immunocompromise status, and vaccine product is, however, not known

What this study adds

  • Among US adults of all ages, protection provided by either mRNA vaccine against moderate and severe covid-19 waned after primary vaccination, increased markedly after a third dose, and then waned again by four to five months after a third dose

  • Vaccine effectiveness diminished less against severe disease than against moderate disease

  • A fourth dose improved vaccine effectiveness among most subgroups for whom it was recommended; overall, our findings support recommendations for broad use of booster doses

Ethics statements

Ethical approval

This study was approved by the institutional review board of Westat.

Data availability statement

No additional data available.

Footnotes

  • Contributors: All authors contributed to the design of the study. PKM, SER, RB, and DY performed the statistical analysis. SR, BD, MBD, SAI, NL, KN, ED, SJG, JH, CM, TCO, ALN, PJE, KD, NPK, IL, WFF, NG, KG, KP, NRV, JA, OZ, CR, MB, MG, and BF were involved in data collection and study coordination at partner sites. EPG, PP, MD, JW, CHB, LB, and RL provided data collection and central study coordination at US Centers for Disease Control and Prevention, supervised by MT. JMF and BF produced the first draft of this manuscript and all authors reviewed, edited, and approved the final version. JMF is the guarantor. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

  • Funding: This study was funded by the Centers for Disease Control and Prevention through contract 75D30120C07986 to Westat and contract 75D30120C07765 to Kaiser Foundation Hospitals.

  • Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/disclosure-of-interest/ and declare: NPK reports institutional support from Pfizer, Merck, GlaxoSmithKline, Sanofi Pasteur, and Protein Sciences (now Sanofi Pasteur) for unrelated studies and institutional support from Pfizer for a covid-19 vaccine trial. CM received institutional support from AstraZeneca for a covid-19 vaccine trial. ALN received institutional support from Pfizer for an unrelated study of meningococcal B vaccine safety during pregnancy. SR received grant funding from GlaxoSmithKline and Biofire Diagnostics. Authors declare no financial relationships with any organizations that might have an interest in the submitted work in the previous three years, and no other relationships or activities that could appear to have influenced the submitted work.

  • The lead author (JMF) affirms that this manuscript is an accurate and transparent account of the study being reported and that no important aspects of the study have been omitted.

  • Dissemination to participants and related patient and public communities: The individual level dataset from this study is held securely in limited deidentified form at the US Centers for Disease Control and Prevention. Data sharing agreements between CDC and data providers prohibit CDC from making this dataset publicly available. CDC will share aggregate study data once study objectives are complete, consistent with data use agreements with partner institutions.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

http://creativecommons.org/licenses/by-nc/4.0/

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

References