The Vaccine Safety Datalink is a collaboration between 9 US integrated health care systems and the Centers for Disease Control and Prevention (CDC).^8,9 Eight data-contributing organizations (Kaiser Permanente: Colorado, Northern California, Northwest, Southern California, and Washington; Marshfield Clinic; HealthPartners; and Denver Health) have access to comprehensive medical records, including vaccinations, for 10 158 003 people aged at least 12 years as of November 10, 2021.

This activity was approved by the institutional review boards with a waiver of informed consent at Kaiser Permanente: Colorado, Northern California, Northwest, Southern California, and Washington; Marshfield Clinic; HealthPartners; and Denver Health. Per 45 CFR part 46.111(f)(3), this study was granted a waiver of informed consent because it posed minimal risk to participants and could not feasibly be conducted without the waiver. The CDC determined that this activity was public health surveillance (45 CFR part 46.102[l][2]) and thus did not require institutional review board review. This report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.¹⁰

Since December 2020, the Vaccine Safety Datalink has been conducting safety surveillance of COVID-19 vaccines, monitoring GBS and 22 other prespecified, serious outcomes after COVID-19 vaccines on a weekly basis. Methods are described in the study protocol and in a prior publication that includes interim results for mRNA vaccines.^7,11 Weekly analyses compared outcome incidence observed during a risk interval after vaccination (eg, 1-21 days after Ad.26.COV2.S) with outcome incidence expected. Analyses were conducted separately by vaccine type. The expected was derived from 1 of 2 types of comparators: (1) similar vaccinated persons who were concurrently (on the same calendar day) in a postvaccination comparison interval following the same vaccine type (eg, 22-42 days after Ad.26.COV2.S) (vaccinated concurrent comparators), and (2) similar persons who were concurrently (on the same calendar day) unvaccinated (ie, had not received any doses of COVID-19 vaccine) (unvaccinated concurrent comparators), where similar means comparators were of the same age group, sex, race or ethnicity, and site as the vaccine recipient. Race and ethnicity data were captured in Vaccine Safety Datalink files in fixed categories based on self-reported data from the participating sites. Race and ethnicity data were used to adjust for confounding that may have arisen if the factor was associated with vaccination dates and outcome events. By protocol, vaccinated concurrent comparators were considered primary and unvaccinated concurrent comparators were considered supplemental owing to concerns about possible bias from unmeasured differences between vaccinated and unvaccinated individuals. This surveillance approach used 2 postvaccination risk intervals, 1 to 21 days and 1 to 42 days, with corresponding postvaccination comparison intervals for vaccinated concurrent comparator analyses, 22 to 42 days and 43 to 84 days, respectively. The 1 to 21-day risk interval allowed for timelier analyses and avoids bias introduced from the short interval between doses for mRNA vaccines. However, a 1 to 42-day risk interval was also used, because this interval is often used in vaccine safety studies of GBS and other outcomes.¹² Individuals who received 2 doses of an mRNA vaccine contributed to analyses of mRNA vaccines when they were 1 to 21 (or 1 to 42) days after dose 1 and again when they were 1 to 21 (or 1 to 42) days after dose 2. However, receipt of dose 2 resulted in censoring of follow-up time after dose 1, therefore most comparison time in the vaccinated concurrent comparator analyses was after dose 2.

We reported on data through November 13, 2021, including descriptive characteristics of COVID-19 vaccine recipients and GBS cases, weekly analyses of GBS with vaccinated and unvaccinated concurrent comparators (previously described) for Ad.26.COV2.S and mRNA vaccines, and additional GBS analyses including comparisons of the incidence of GBS postvaccination by vaccine type to the incidence expected from prior studies, evaluation of the temporal clustering of GBS cases following Ad.26.COV2.S, and head-to-head comparisons of GBS incidence after Ad.26.COV2.S with GBS incidence after mRNA vaccination.

Potential cases of GBS were identified among Vaccine Safety Datalink members aged at least 12 years using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) code G61.0 in the emergency department or inpatient setting, specifically when G61.0 first appeared in an individual’s record in the 1 to 84 days after dose 1 or dose 2 of an mRNA vaccine or after dose 1 of Ad.26.COV2.S. Because disease onset may begin before a diagnosis is recorded in the medical record, potential cases with the ICD-10 code in the 85 to 98 days postvaccination were also reviewed. After review, all cases underwent adjudication according to the Brighton Collaboration criteria.¹³ Briefly, GBS cases meeting Brighton level 1 had the highest level of diagnostic certainty, whereas level 4 included cases of suspected GBS (ie, insufficient information was available in the medical record to meet Brighton level 1-3). In this analysis, we considered Brighton level 1 to level 4 cases as confirmed and sensitivity analyses were conducted excluding level 4 cases. Cases of GBS among unvaccinated concurrent comparators were identified by applying the same algorithm to unvaccinated members of the Vaccine Safety Datalink population during the period of interest. Unvaccinated comparators did not undergo medical record review and adjudication.

As previously described,⁷ primary weekly monitoring of GBS after each type of COVID-19 vaccine used conditional Poisson regression to compare individuals during a postvaccination risk interval with similar individuals who had been vaccinated earlier and were concurrently in a comparison interval. These analyses estimated rate ratios (ie, ratios of incidence rates in a risk interval divided by incidence rates in a comparison interval) that were adjusted for 5-year age group, sex, race and ethnicity, site, and calendar day by conditioning the Poisson regression on strata (risk sets) defined by these factors. Similarly, conditional Poisson regression was used in supplementary weekly analyses to compare individuals in a postvaccination risk interval with similar individuals who were concurrently unvaccinated. One-sided sequential testing was conducted weekly for vaccinated concurrent comparator analyses, with a signaling threshold of P < .0048 (prespecified to keep the overall chance of making a type I error below .05 during 2 years of weekly analyses).

Although our primary weekly analyses of GBS have not met the criterion for a safety signal, concerns about GBS after Ad.26.COV2.S arose from supplementary analyses with unvaccinated comparators and from reports to VAERS. To further investigate we conducted additional analyses. We computed unadjusted incidence rates and 95% CIs of confirmed GBS in the 1 to 21 days and 1 to 42 days following COVID-19 vaccination per 100 000 person-years (PY) by vaccine type and used exact Poisson regression to compare observed incidence rates to the historical background rate of GBS (2 per 100 000 PY).^14,15 We also examined the temporal clustering of the GBS cases within 56 days of Ad.26.COV2.S by day of symptom onset, using a scan statistic.¹⁶ Lastly, we conducted head-to-head comparisons of the 21-day and 42-day risk intervals after Ad.26.COV2.S vs the 21-day and 42-day risk intervals after mRNA vaccination, using conditional Poisson regression to estimate rate ratios (ie, the ratio of GBS incidence after Ad.26.COV2.S divided by GBS incidence after mRNA vaccination), adjusted for age group, sex, race and ethnicity, site, and calendar day as described above. All analyses of mRNA vaccines combined both products (BNT162b2 and mRNA-1273) and primary series doses (1 and 2) to increase statistical power. Analyses were 2-sided with a .05 level of significance unless otherwise specified and were conducted using SAS version 9.4 (SAS Institute) from November 2021 to February 2022.

From December 13, 2020, through November 13, 2021, 15 120 073 COVID-19 vaccines were administered to 7 894 989 individuals who were at least 12 years of age (mean age [SE] age, 46.5 [0.02] years; 8 138 318 doses received [53.8%] by female individuals; 3 671 199 doses received [24.3%] by Hispanic or Latino individuals, 2 215 064 doses received [14.7%] by Asian individuals, 6 266 424 doses received [41.4%] by White individuals), including 483 053 first doses of Ad.26.COV2.S, 8 806 595 first and second doses of BNT162b2, and 5 830 425 first and second doses of mRNA-1273; mRNA vaccines accounted for 96.8% of all doses (14 637 020 of 15 120 073). Compared with mRNA vaccine recipients, a higher proportion of Ad.26.COV2.S vaccine recipients were male (6 721 529 [45.9%] vs 260 226 [53.9%]) and a smaller proportion were at least 65 years of age (3 139 195 [21.5%] vs 55 901 [11.6%]) (Table 1). Ad.26.COV2.S vaccination was concentrated during the 7-week period that corresponded with the period between initial authorization and the pause in use due to concerns about TTS, with 63.5% of Ad.26.COV2.S doses (306 590 of 483 053) administered from February 28, 2021, through April 17, 2021.

During the 1 to 84 days after individuals received Ad.26.COV2.S, 22 potential cases of GBS were identified: all were reviewed and adjudicated, and 11 (50%) were confirmed. Symptom onset ranged from 1 to 77 days after Ad.26.COV2.S; 9 cases (82%) had symptom onset in the 1 to 21 days postvaccination, 1 (9%) had symptom onset in the 22 to 42 days postvaccination, and 1 (9%) had symptom onset in the 43 to 84 days postvaccination (Figure A). Scan statistics identified days 1 to 14 after vaccination as a statistically significant cluster (P = .003). Ten patients were hospitalized (91%) and all were treated with intravenous immune globulin. Patients with confirmed GBS after Ad.26.COV2.S had a mean (range) age of 50 (32-63) years, 9 (82%) were male, 10 (91%) were non-Hispanic White, and 10 (91%) had facial weakness or paralysis in addition to bilateral weakness or paralysis of the limbs (Table 2).

The unadjusted incidence rate of confirmed cases of GBS per 100 000 PY was 32.4 (95% CI, 14.8-61.5) during the 1 to 21 days after Ad.26.COV2.S and 18.0 (95% CI, 8.6-33.1) duringthe 1 to 42 days after Ad.26.COV2.S; both estimates were significantly higher than the background rate of GBS (P < .001) (Table 3). When excluding Brighton level 4 cases, incidence rates were lower (28.8 [95% CI, 12.4-56.8] and 16.2 [95% CI, 7.4-30.8] per 100 000 PY, respectively) but still significantly higher than the background rate (P < .001).

In weekly surveillance using vaccinated concurrent comparators, as of November 16, 2021, the RR of confirmed GBS, adjusted for age, sex, race and ethnicity, site, and calendar day in the 1 to 21 vs 22 to 42 days following Ad.26.COV2.S was 6.03 (95% CI, 0.79-147.79; 2-sided P = .09; 1-sided P = .08). The adjusted RR of confirmed GBS in the 1 to 42 vs 43 to 84 days following Ad.26.COV2.S was 8.64 (95% CI, 1.18-207.32; 2-sided P = .03; 1-sided P = .03). Neither result met the prespecified signaling criteria of a 1-sided P < .0048. In supplemental weekly analyses with unvaccinated concurrent comparators, the adjusted RR of GBS was 10.57 (95% CI, 5.15-20.16; P < .001) in the 1 to 21 days following Ad.26.COV2.S and 10.05 (95% CI, 5.75-16.96; P < .001) in the 1 to 42 days following Ad.26.COV2.S.

During the 1 to 84 days after first and second doses of mRNA vaccines, 78 potential cases of GBS were identified: all were reviewed and adjudicated, and 36 (46%) were confirmed, including 23 after BNT162b2 and 13 after mRNA-1273. Patients with confirmed GBS after mRNA vaccines had a mean (range) age of 53 (14-92) years (Table 2) and symptom onset ranged from 0 to 84 days after vaccination (Figure B). Eleven cases (31%) had symptom onset in the 1 to 21 days postvaccination, 15 cases (42%) had symptom onset in the 22 to 42 days postvaccination, and 9 cases (25%) had symptom onset in the 43 to 84 days postvaccination. One case had symptom onset on the same day as mRNA vaccination (day 0) and was excluded from analyses.

The unadjusted incidence rate of confirmed cases of GBS per 100 000 PY in the 21 days after mRNA vaccines was 1.3 (95% CI, 0.7-2.4) and did not differ from the background rate of GBS (P = .20) (Table 3). Results were similar when excluding Brighton level 4 cases and when using a 42-day risk window, with incidence rates ranging from 1.1 to 2.0 cases of GBS per 100 000 PY.

In weekly surveillance conducted on November 16, 2021, the adjusted RR of confirmed GBS in the 1 to 21 vs 22 to 42 days following mRNA vaccination was 0.56 (95% CI, 0.21-1.48; 2-sided P = .25, 1-sided P = .93). In supplemental weekly analyses with unvaccinated concurrent comparators, the adjusted RRof GBS was 0.83 (95% CI, 0.50-1.33; P = .45) in the 1 to 21 days following mRNA vaccination and 0.85 (95% CI, 0.57-1.27; P = .44) in the 1 to 42 days following mRNA vaccination.

During the 1 to 21 days after vaccination, the adjusted RR of confirmed GBS after Ad.26.COV2.S vs mRNA vaccination was 20.56 (95% CI, 6.94-64.66; P < .001), with 15.5 excess GBS cases in the risk interval per million Ad.26.COV2.S vaccine recipients (Table 4). When using a 1 to 42-day risk interval, the adjusted RRwas lower, 11.46 (95% CI, 4.83-26.16; 2-sided P < .001), otherwise results were similar.

This study had several limitations. First, as mentioned previously, substantially fewer doses of Ad.26.COV2.S were administered relative to mRNA vaccines, resulting in reduced statistical power and wide confidence intervals for some analyses. Second, in this observational study recipients of Ad.26.COV2.S may have differed from recipients of mRNA vaccines in unknown ways that affect GBS risk but were not adjusted for in analyses. Third, the incidence rate of confirmed GBS during the COVID-19 pandemic has not been established, and may differ from prepandemic background rates. Fourth, we could not identify subgroups who may be at greatest risk for GBS after Ad.26.COV2.S given the small number of confirmed GBS cases identified in this surveillance. Fifth, this analysis only included GBS cases after primary COVID-19 vaccination and results may not be generalizable to additional or booster doses.