Older adults mount less durable humoral responses to two doses of COVID-19 mRNA vaccine, but strong initial responses to a third dose

24

Third COVID-19 vaccine doses are broadly recommended, but immunogenicity data 25 remain limited, particularly in older adults. 26

We measured circulating antibodies against the SARS-CoV-2 spike protein receptor-binding 27 domain, ACE2 displacement, and virus neutralization against ancestral and Omicron (BA.1) strains 28 from pre-vaccine up to one month following the third dose, in 151 adults aged 24-98 years who 29 received COVID-19 mRNA vaccines. 30

Following two vaccine doses, humoral immunity was weaker, less functional and less durable 31 in older adults, where a higher number of chronic health conditions was a key correlate of weaker 32 responses and poorer durability. Third doses boosted antibody binding and function to higher levels 33 than second-doses, and induced responses in older adults that were comparable in magnitude to those in 34 younger adults. Humoral responses against Omicron were universally weaker than against the ancestral 35 strain after both second and third doses; nevertheless, after three doses, anti-Omicron responses in older 36 adults reached equivalence to those in younger adults. After three vaccine doses, the number of chronic 37 health conditions, but not age per se, was the strongest consistent correlate of weaker humoral 38 responses. 39

Results underscore the immune benefits of third COVID-19 vaccine doses, particularly in 40 older adults. 41 42 Key words: COVID-19, mRNA vaccine, SARS-CoV-2, humoral immunity, older adults, binding 43 antibodies, ACE2 displacement, viral neutralization, Omicron 44 45 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

46 Older adults are at increased risk of lethal COVID-19 following SARS-CoV-2 infection 47 (SARS-CoV-2) [1-3]. While two doses of a COVID-19 mRNA vaccine broadly protects against 48 hospitalization and death [4-6], weaker vaccine-induced immunity observed in the elderly and certain 49 other groups [7-12] has led to their prioritization to receive third doses [13-16]. Vaccine-induced 50 antibodies also decline over time, which can increase the risk of breakthrough infections [17-19], 51 particularly with the more transmissible and immune evasive Omicron variant (B.1.1.529) [20-22]. 52 We and others have shown that older age is associated with weaker antibody responses to 53 COVID-19 mRNA vaccines, Comirnaty (Pfizer/BioNTech) and Spikevax (Moderna) [10-12]. We 54 previously characterized longitudinal humoral responses up to three months after the second vaccine 55 dose in a cohort of 151 adults 24 to 98 years of age that includes COVID-19 naïve and convalescent 56 individuals [12]. Here, we examine binding and neutralizing antibody responses up to six months 57 following the second vaccine dose, as well as one month following the third vaccine dose. We also 58 evaluate binding antibodies, ACE2 displacement, and virus neutralization against Omicron (BA.1) one 59 month following the second and third doses. 60 61 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

62 Study design. We conducted a prospective longitudinal cohort study in British Columbia, Canada, to 63 examine SARS-CoV-2 specific humoral responses following vaccination with Comirnaty or Spikevax. 64 Our cohort of 151 individuals included 81 healthcare workers (HCW) and 56 older adults (including 18 65 residents of long-term care or assisted living facilities) who were COVID-19 naive at study entry, and 66 14 COVID-19 convalescent individuals with anti-SARS-CoV-2 N antibodies at study entry (including 67 8 HCW and 6 older adults) [12]. Serum and plasma were collected prior to vaccination; one month 68 after the first dose; one, three and six months after the second dose; and one month following the third 69 dose. Specimens were processed same-day and frozen until analysis. 70 71 Ethics approval. Written informed consent was obtained from all participants or their authorized 72 decision makers. This study was approved by the University of British Columbia/Providence Health 73 Care and Simon Fraser University Research Ethics Boards. 74 75 Data sources. Sociodemographic, health and vaccine information was collected by self-report and 76 confirmed through medical records where available. Chronic health conditions were defined as 77 hypertension, diabetes, asthma, obesity (body mass index ≥30), chronic diseases of lung, liver, kidney, 78 heart or blood, cancer, and immunosuppression due to chronic conditions or medication, to generate a 79 score ranging from 0-11 per participant [12]. 80 81 Binding antibody assays. We measured total binding antibodies against SARS-CoV-2 nucleocapsid 82 (N) and spike (S) receptor binding domain (RBD) in serum using the Roche Elecsys Anti-SARS-CoV-83 2 and Anti-SARS-CoV-2 S assays, respectively, on a Cobas e601 module analyzer (Roche 84 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Diagnostics). Following SARS-CoV-2 infection, both assays should be positive, whereas post-85 vaccination only the S assay should be positive, allowing identification of convalescent individuals. 86 Both tests are electro-chemiluminescence sandwich immunoassays, and report results in Arbitrary 87 Units (AU)/mL, calibrated against an external standard. For the S assay, the manufacturer indicates that 88 AU values can be considered equivalent to international binding antibody units (BAU) as defined by 89 the World Health Organization [23]. For the S assay, sera were tested undiluted, with samples above 90 the upper limit of quantification (ULOQ) re-tested at 1:100 dilution, allowing a measurement range of 91 0.4 - 25,000 U/mL. We also quantified plasma IgG binding antibodies against RBD using the V-plex 92 SARS-CoV-2 (IgG) Panel 22 ELISA kit (Meso Scale Diagnostics), which features the ancestral 93 (Wuhan) and Omicron RBD antigens, on a Meso QuickPlex SQ120 instrument. Plasma samples were 94 diluted 1:10000 as directed by the manufacturer, with results reported in Arbitrary Units (AU)/mL. 95 96 ACE2 competition assay. We assessed the ability of plasma antibodies to block the RBD-ACE2 97 receptor interaction by competition ELISA (Panel 22 V-plex SARS-CoV-2 [ACE2]; Meso Scale 98 Diagnostics) on a Meso QuickPlex SQ120 instrument. Plasma was diluted 1:20 as directed by the 99 manufacturer and results reported as % ACE2 displacement. 100 101 Live virus neutralization. Neutralizing activity in plasma was examined using a live SARS-CoV-2 102 infectivity assay in a Containment Level 3 facility. Assays were performed using isolate USA-103 WA1/2020 (BEI Resources) and a local Omicron isolate (BA.1 strain; GISAID Accession # 104 EPI_ISL_9805779) on VeroE6-TMPRSS2 (JCRB-1819) target cells. Viral stock was adjusted to 50 105 TCID50/200 µl in Dulbecco’s Modified Eagle Medium in the presence of serial 2-fold dilutions of 106 plasma (from 1/20 to 1/2560), incubated at 4°C for 1 hour and then added to target cells in 96-well 107 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint plates in triplicate. Cultures were maintained at 37°C with 5% CO2 and the appearance of viral 108 cytopathic effect (CPE) was recorded three days post-infection. Neutralizing activity is reported as the 109 highest reciprocal plasma dilution able to prevent CPE in all three replicate wells. Samples exhibiting 110 only partial or no neutralization at the lowest dilution of 1/20 were coded as having a reciprocal 111 dilution of "10", defined as below the limit of quantification (BLOQ) in this assay. 112 113 Statistical analysis. Comparisons of binary variables were performed using Fisher’s exact test. 114 Comparisons of continuous variables were performed using the Mann-Whitney U-test (for unpaired 115 data) or Wilcoxon test (for paired data). Multiple linear regression was used to investigate the 116 relationship between sociodemographic, health and vaccine-related variables and humoral outcomes. 117 Variables included age (per year increment), sex at birth (female as reference group), ethnicity (non-118 white as reference), number of chronic health conditions (per number increment), mRNA vaccine 119 received (Comirnaty as reference), interval between doses (per day increment), sampling date 120 following the most recent dose (per day increment), and convalescent status (COVID-19 naive as 121 reference). Binding antibody half-lives in serum were calculated by fitting exponential decay curves to 122 antibody concentrations at one, three and six months after the second dose. All tests were two-tailed, 123 with p<0.05 considered statistically significant. Analyses were conducted using Microsoft Excel and 124 Prism v9.2.0 (GraphPad). 125 126 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

127 Participant characteristics 128 As described previously [12], the cohort is predominantly female (Table 1). HCW, older adults 129 and COVID-19 convalescent individuals at study entry were a median of 41, 79 and 48 years old, 130 respectively. Older adults were predominantly (77%) of white ethnicity (compared to 46% of HCW) 131 and had a higher burden of chronic health conditions (a median of 1, interquartile range [IQR] 0-2, 132 range 0-5, vs. a median of 1, IQR 0-0, range 0-3 in HCW). All participants received two COVID-19 133 mRNA vaccine doses between December 2020-July 2021, where the dose interval was up to 112 days 134 as per national guidelines to delay second doses due to initially limited vaccine supply. A total of 141 135 (93%) and 138 (91%) of participants received Comirnaty as their first and second dose, respectively. At 136 the time of writing, 114 participants had received a third dose between October-December 2021, on 137 average 7 months following their second dose. For participants whose third dose was Spikevax (53% of 138 the cohort), those aged ≥70 years received a full dose, whereas those <70 years received a half-dose, as 139 per national guidelines. An additional six (7.4%) HCW and two (3.6%) older adults developed anti-N 140 antibodies during follow-up, reflecting breakthrough infections. Three of these infections, all in HCW, 141 occurred between December 2021-Jan 2022 and are likely Omicron. In longitudinal analyses that span 142 the entire study, participants with a post-vaccination SARS-CoV-2 infection are retained in their 143 original "COVID-19 naive at study entry" groups but identified in the Figures, while in analyses that 144 focus on third dose responses, they are grouped in a single "prior COVID-19" group. 145 146 After two-dose vaccination, lower binding antibodies are associated with older age and burden of 147 chronic health conditions, but older adults mount strong responses after a third dose. 148 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint We measured total anti-RBD binding antibody concentrations in serum before and after 149 immunization (Figure 1A). As reported previously [12], antibody concentrations in older adults were 150 significantly lower than those in HCW one month after the first dose (a median of 2.00 [IQR 1.75-2.25] 151 log10 U/mL in HCW versus a median of 1.50 [IQR 1.05-1.99] in older adults), as well as one month 152 after the second dose (a median of 4.02 [IQR 3.88-4.25] in HCW versus a median of 3.74 [IQR 3.49-153 3.91] in older adults) (Mann-Whitney; both p<0.0001). Three months following the second dose, 154 antibody concentrations had declined by ~0.4 log10 on average, to a median of 3.63 [IQR 3.44-3.83] in 155 HCW versus a median 3.32 [IQR 3.04-3.56] in older adults) (Mann-Whitney p<0.0001 for comparison 156 between groups). Six months following the second dose, antibody concentrations had declined by a 157 further ~0.3 log10 on average, to a median of 3.30 [IQR 3.09-3.47] in HCW versus a median 2.96 [IQR 158 2.68-3.20] in older adults (p<0.0001). This confirms that, following two-dose COVID-19 mRNA 159 vaccination, antibody concentrations remain consistently and significantly lower in older compared to 160 younger adults. By contrast, antibody concentrations in COVID-19 convalescent individuals remained 161 consistently higher than COVID-19 naive individuals at all time points after two doses. Six months 162 after the second dose for example, convalescent individuals maintained median responses of 3.50 (IQR 163 3.40-3.71) log10 U/mL (p=0.027 compared to HCW; p<0.0001 compared to older adults). 164 Multivariable analyses of antibody concentrations after two doses, that adjusted for sex , 165 ethnicity, number of chronic health conditions, first-dose vaccine brand, dosing interval and day of 166 specimen collection post-immunization confirmed that older age remained independently associated 167 with lower antibody concentrations at one and three months after the second dose (Table S1). One 168 month following the second dose for example, each decade of older age was associated with an ~0.06 169 log10 lower antibody concentration (p=0.0067). A higher number of chronic conditions was also 170 independently associated with lower antibody concentrations at both these time points. Six months 171 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint following the second dose, a higher number of chronic health conditions remained the strongest 172 independent correlate of lower responses, with each additional condition associated with an 0.14 log10 173 lower antibody concentration (p=0.0001). A longer dose interval was also associated with higher 174 antibody concentrations at all time points after the second dose (all p<0.05), consistent with previous 175 reports [24-26]. COVID-19 convalescent status was also associated with maintaining 0.26 log10 higher 176 antibody concentrations at three and six months following the second dose (both p<0.05), consistent 177 with superior durability of “hybrid” immunity induced by infection followed by vaccination [27-29]. 178 In both HCW and older adults, the third dose boosted antibody concentrations at least ~0.3-0.4 179 log10 higher than peak values observed after two doses (Wilcoxon paired test p<0.0001 for both 180 groups). Binding antibodies in HCW rose to a median of 4.31 (IQR 4.13 to upper limit of quantification 181 [ULOQ]) whereas those in older adults rose to a median of 4.33 (4.14 to ULOQ) (p=0.33), indicating 182 that older and younger adults mounted comparable initial binding antibody responses following a third 183 dose. In multivariable analyses of third-dose responses, a higher number of chronic health conditions 184 was the sole significant correlate of lower antibody concentrations (p=0.0078), while having received 185 Spikevax as the third dose was associated with higher antibody concentrations (p=0.0091) (Table S2). 186 187 After two-dose vaccination, weaker virus neutralizing activity is associated with age and chronic 188 health conditions, but older adults mount strong responses after a third dose. 189 We performed live SARS-CoV-2 neutralization assays to quantify the ability of plasma to block 190 virus infection of target cells (Figure 1B). Neutralizing activity is reported as the highest reciprocal 191 plasma dilution capable of preventing viral cytopathic effects in all wells of a triplicate assay, where a 192 reciprocal dilution of "10" indicates no or limited neutralization. As previously reported [12], one 193 vaccine dose largely failed to induce neutralizing activity in COVID-19 naïve individuals, though two 194 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint doses induced this activity in most participants, albeit at consistently lower levels in older compared to 195 younger adults. One month after the second dose for example, the median reciprocal dilution was 160 196 [IQR 80-160] in HCW versus 40 [IQR 20-80] in older adults (p<0.0001). Three months after the 197 second dose, neutralizing activity had declined by more than two-fold on average, to a median 198 reciprocal dilution of 40 (IQR 20-80) in HCW versus a median of 20 (IQR BLOQ-40) in older adults 199 (p<0.0001). Six months after the second dose, neutralizing activity had declined to below the limit of 200 quantification (BLOQ) in 58% of HCW and 83% of older adults (Mann-Whitney p=0.0048 for 201 comparison between groups). COVID-19 convalescent individuals by contrast maintained significantly 202 higher neutralizing activity compared to naive individuals at all time points following two-dose 203 vaccination. Multivariable analyses confirmed that older age remained significantly associated with 204 weaker neutralizing activity at one and three months after two-dose vaccination, while COVID-19 205 convalescent status was associated with superior neutralizing activity at all time points following two-206 dose vaccination (all p≤0.0002) (Table S1). 207 A third vaccine dose boosted neutralizing activity in both HCW and older adults, achieving 208 responses that were two-fold and eight-fold higher than peak values after two doses, respectively 209 (Wilcoxon paired test p≤0.006 for both groups; Figure 1B). Specifically, the median reciprocal dilution 210 in HCW and older adults rose to 320 [IQR 160-320] and 320 [IQR 80-320], respectively (p=0.6), 211 indicating that older adults mounted comparable neutralizing responses to younger adults after three 212 doses. A multivariable analysis identified prior COVID-19 as the strongest independent predictor of 213 higher neutralizing activity after a third vaccine dose (p=0.0044; Table S2). 214 215 After two-dose vaccination, binding antibody responses decline faster in those with a higher burden 216 of chronic conditions. 217 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint We next assessed temporal reductions in antibody concentrations after two-dose vaccination 218 (Figure 2A). Assuming exponential decay and restricting the analysis to participants with a complete 219 longitudinal data series with no values above the ULOQ, we estimated antibody concentration half-220 lives to be a median of 59 [IQR 52-75] days in HCW versus a median of 52 [IQR 45-65] days in older 221 adults (p=0.016; Figure 2B). This suggests that, in addition to mounting overall weaker responses to 222 two-dose vaccination compared to younger adults, antibody concentrations in older adults also decline 223 more rapidly. In multivariable analyses however, a higher number of chronic health conditions 224 emerged as the sole independent correlate of antibody decline, with each additional condition 225 associated with a 5-day shorter half-life (p=0.017; Table 2). Furthermore, COVID-19 convalescent 226 status was associated with a 14-day longer antibody half-life after adjustment for other factors 227 (p=0.056), consistent with improved durability of hybrid immunity [27-29]. 228 229 Humoral responses against Omicron following two and three vaccine doses 230 Given the rapid rise of the Omicron variant, we compared peak antibody responses against this 231 strain in plasma collected at one month after the second and third vaccine doses. Here, we grouped all 232 participants with prior COVID-19, regardless of infection timing, in the convalescent category. Overall, 233 IgG binding antibodies against the Omicron RBD, measured using the Meso Scale Diagnostics V-Plex 234 assay, were on average 0.4 to 0.5 log10 U/mL lower than those against the wild type (WT; ancestral 235 Wuhan strain) RBD antigen after two and three doses (all within-group comparisons p≤0.0002; Figure 236 3A). Nevertheless, the third dose universally boosted anti-Omicron IgG concentrations to an average of 237 0.5 log10 higher than levels induced by two doses (all within-group comparisons p<0.05). Consistent 238 with total binding antibody concentrations quantified using the Roche assay (Figure 1A), binding IgG 239 concentrations against the WT RBD were significantly higher in HCW compared to older adults after 240 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint two doses (p<0.0001) but reached equivalence after three doses (p=0.4). IgG concentrations capable of 241 binding Omicron followed a similar pattern, with HCW showing marginally higher anti-Omicron IgG 242 levels compared to older adults after two doses (p=0.09), but equivalent levels after three doses 243 (p=0.49). A multivariable analysis of Omicron-specific IgG concentrations after three doses identified a 244 higher number of chronic health conditions as the strongest correlate of poorer responses, with each 245 additional condition associated with a 0.12 log10 reduction in Omicron binding IgG (p=0.0033; Table 246 3). A longer interval between the first and second vaccine doses was marginally associated with a 247 lower third dose response (p=0.02). 248 We also assessed the ability of plasma to block the interaction between WT and Omicron RBD 249 and the cellular ACE2 receptor, which represents a higher throughput approach to estimate potential 250 virus neutralizing activity (also referred to as a surrogate virus neutralization test [30]). This activity 251 was significantly weaker against Omicron compared to WT RBD after both two and three doses in all 252 groups (all within-group comparisons p≤0.0002; Figure 3B), though the discrepancy was most 253 pronounced for older adults after two doses (where median activity against WT was 90% compared to 254 only 23% against Omicron). The third dose universally boosted anti-Omicron activity (all within-group 255 comparisons p<0.05), with, for example, median anti-Omicron activity in older adults rising from 23% 256 after two doses to 66% after three. Consistent with results for binding IgG antibodies, surrogate 257 neutralization of WT RBD was significantly higher in HCW compared to older adults after two doses 258 (p<0.0001), but reached equivalence after three doses (in fact, activities in older adults were slightly 259 higher at this time point; p=0.08). Surrogate neutralization of Omicron RBD followed a similar pattern, 260 with HCW exhibiting significantly higher activity compared to older adults after two doses (p<0.0001), 261 but equivalent levels after three doses (p=0.2). In multivariable analyses, a higher number of chronic 262 health conditions was the strongest correlate of poorer surrogate neutralizing activity against Omicron 263 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint after three vaccine doses, with each additional condition associated with a ~6% reduction in this 264 activity (p=0.0046; Table 3). Male sex, a longer interval between the first and second doses, and the 265 number of days elapsed since the third dose also correlated with weaker responses after three doses (all 266 p<0.05). 267 Finally, we assessed plasma neutralizing activity against WT (ancestral USA-WA1/2020 strain) 268 and Omicron using a live virus assay in a subset of 20 HCW and 21 older adults who remained 269 COVID-19 negative throughout the study (Figure 4). Neutralizing activity against Omicron was 270 significantly weaker compared to WT following two and three doses in both groups (all p<0.0001). The 271 third dose nevertheless boosted anti-Omicron activity in both groups, where the increase in older adults 272 was particularly pronounced (from a median of BLOQ after the second dose to a median reciprocal 273 dilution of 40 after the third; p<0.0001). Consistent with binding IgG and surrogate neutralization 274 results, anti-Omicron neutralizing activity was significantly lower in older adults compared to HCW 275 after two vaccine doses (p=0.0003) but reached equivalence after the third dose (p=0.79). 276 277 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

278 At every time point following two doses of COVID-19 mRNA vaccine, antibody binding and 279 neutralizing activity were significantly weaker in older compared to younger adults. Antibody 280 concentrations were also less durable in older adults , though responses declined substantially in all 281 groups over time (e.g. by six months after the second dose, neutralizing activities had declined to 282 BLOQ in almost 60% of HCW and >80% of older adults). In multivariable analyses adjusting for 283 sociodemographic, health and vaccine-related variables, a higher number of chronic health conditions 284 remained consistently and independently associated with weaker and less durable binding antibody 285 responses, while a longer interval between first and second doses was consistently associated with 286 higher binding antibody responses after the second dose, as previously reported [24-26]. These findings 287 support public health decisions to provide third doses on or before the six-month mark, with older 288 adults receiving priority. 289 Third doses of COVID-19 vaccine increased antibody binding and neutralizing function to 290 levels that were significantly higher than those achieved by two doses, where the magnitude of 291 boosting in older adults was particularly prominent. Indeed, antibody binding, surrogate neutralization 292 and live virus neutralization activities in older adults were equivalent to those observed in younger 293 adults after three doses. Consistent with recent evidence [20, 21, 31-37], antibody responses against 294 Omicron were universally weaker than those against the ancestral strain after both two and three 295 vaccine doses; nevertheless, anti-Omicron responses in older adults reached equivalence to those 296 observed in younger adults after three doses. Notably, the number of chronic health conditions 297 persisted as an independent correlate of weaker anti-Omicron responses, even after three doses. 298 Similar to other reports [27-29], our findings indicate that individuals who have contracted 299 COVID-19 are likely to benefit from vaccination. Compared to naïve participants, convalescent 300 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint individuals displayed a slower rate of antibody decline, and multivariable analyses demonstrated that 301 binding and neutralization activity was higher in this group at six months after the second dose . 302 Our study has several limitations. As the precise immune correlates of protection for SARS-303 CoV-2 transmission and disease severity remain incompletely characterized [38], the implications of 304 our results on individual-level protection from SARS-CoV-2 infection and COVID-19 remain 305 uncertain. We did not investigate T-cell responses, which may play critical roles in protection against 306 severe COVID-19, particularly in the context of variants [39-46]. Our study was not powered to 307 investigate potential differences in immune responses between the two mRNA vaccines [47, 48], nor 308 differences in full vs. half-doses of Spikevax when administered as third doses to individuals ³70 309 versus <70 years old, respectively, in Canada. Third dose responses were measured at a single time 310 point, so durability assessments are needed. Nevertheless, results provide additional insight into 311 COVID-19 mRNA vaccine immunogenicity in the elderly and in the context of an extended interval 312 between first and second doses (of up to 112 days). 313 In conclusion, while the observation of strong binding and neutralizing antibody responses to 314 third COVID-19 vaccine doses in older adults, including to Omicron, are encouraging, it will be 315 important to closely monitor the durability of these responses over time in this population. 316 317 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

318 We thank the leadership and staff of Providence Health Care, including long-term care and 319 assisted living residences, for their support of this study. We thank the phlebotomists and laboratory 320 staff at St. Paul's Hospital, the BC Centre for Excellence in HIV/AIDS and Simon Fraser University for 321 assistance. Above all, we thank the participants, without whom this study would not have been 322 possible. 323 324 FUNDING 325 This work was supported by the Public Health Agency of Canada through a COVID-19 326 Immunology Task Force COVID-19 "Hot Spots" Award (2020-HQ-000120 to MGR, ZLB, MAB). 327 Additional funding was received from the Canadian Institutes for Health Research (GA2-177713 and 328 the Coronavirus Variants Rapid Response Network (FRN-175622) to MAB), the Canada Foundation 329 for Innovation through Exceptional Opportunities Fund – COVID-19 awards (to MAB, MD, MN, RP, 330 ZLB) and the National Institute of Allergy and Infectious Diseases of the National Institutes of Health 331 (R01AI134229 to RP). MLD and ZLB hold Scholar Awards from the Michael Smith Foundation for 332 Health Research. LYL was supported by an SFU Undergraduate Research Award. GU and FHO hold 333 Ph.D. fellowships from the Sub-Saharan African Network for TB/HIV Research Excellence 334 (SANTHE), a DELTAS Africa Initiative [grant # DEL-15-006]. The DELTAS Africa Initiative is an 335 independent funding scheme of the African Academy of Sciences (AAS)’s Alliance for Accelerating 336 Excellence in Science in Africa (AESA) and supported by the New Partnership for Africa’s 337 Development Planning and Coordinating Agency (NEPAD Agency) with funding from the Wellcome 338 Trust [grant # 107752/Z/15/Z] and the UK government. The views expressed in this publication are 339 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint those of the authors and not necessarily those of AAS, NEPAD Agency, Wellcome Trust or the UK 340 government. 341 342 FIGURE LEGENDS 343 Figure 1. Longitudinal antibody binding and neutralization responses to spike RBD following 344 one, two and three COVID-19 vaccine doses. Panel A: Binding antibody responses to the SARS-345 CoV-2 spike RBD in serum, in HCW (blue circles) and older adults (orange circles) who were COVID-346 19 naive at study entry, as well as COVID-19 convalescent individuals (black circles) at six timepoints: 347 prior to vaccination (pre-vax), one month following the first dose, one, three and six months following 348 the second dose, and one month following the third vaccine dose. Individuals with post-vaccination 349 infections are indicated by red dots at their first N seropositive time point. Participant Ns are provided 350 at the bottom of the plot. A thick horizontal red bar represents the median; thinner horizontal red bars 351 represent the IQR. P-values were computed using the Mann-Whitney U-test (for comparisons between 352 groups) or the Wilcoxon matched pairs test (for comparisons across time points within a group) and are 353 uncorrected for multiple comparisons. ULOQ/LLOQ: upper/lower limit of quantification. Panel B: 354 same as A, but for virus neutralization activity, defined as the lowest reciprocal plasma dilution at 355 which neutralization was observed in all wells of a triplicate assay. Plasma samples showing 356 neutralization in fewer than three wells at a 1/20 dilution were coded as having a reciprocal dilution of 357 10, corresponding to the LLOQ in this assay. The highest dilution tested was 1/2560, which 358 corresponds to the ULOQ. Note that only a subset of pre-vaccine plasma samples was assayed for this 359 activity. 360 361 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Figure 2: Decay rates of serum binding antibody responses to spike RBD following two COVID-362 19 vaccine doses. Panel A: Temporal declines in serum binding antibody responses to spike RBD 363 following two vaccine doses in HCW (blue) and older adults (orange) who were COVID-19 naive at 364 study entry, as well as COVID-19 convalescent participants (black circles). ULOQ: upper limit of 365 quantification. Only participants with a complete longitudinal data series with no values above the 366 ULOQ are shown. Panel B: Binding antibody half-lives following two COVID-19 vaccine doses, 367 calculated by fitting an exponential curve to each participant's data shown in panel A. Participant Ns 368 are indicated at the bottom of the plot. Red bars and whiskers represent the median and IQR. P-values 369 were computed using the Mann-Whitney U-test and are uncorrected for multiple comparisons. 370 371 Figure 3: Anti-Omicron IgG binding and ACE2 displacement activities one month after the 372 second and third COVID-19 vaccine doses. Panel A: Binding IgG responses in plasma to the wild-373 type (WT, ancestral Wuhan strain) and Omicron (OM) S-RBD, measured using the Meso Scale 374 Diagnostics (MSD) V-Plex assay, in HCW (blue circles) and older adults (orange circles) who 375 remained COVID-19 naive throughout the study, as well as individuals with prior COVID-19 376 regardless of infection timing (COVID-19 convalescent; black circles) at one month after the second 377 and third COVID-19 vaccine doses. Participant Ns are shown at the bottom of the plot. A thick 378 horizontal red bar represents the median; thinner horizontal red bars represent the IQR. P-values were 379 computed using the Wilcoxon matched pairs test (for all within-group comparisons) or the Mann-380 Whitney U-test (for between-group comparisons) and are uncorrected for multiple comparisons. Panel 381 B: same as A, but for ACE2 displacement activity, measured using the V-plex SARS-CoV-2 (ACE2) 382 assay, where results are reported in terms of % ACE2 displacement. 383 384 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Figure 4: Anti-Omicron neutralization activities one month after the second and third COVID-19 385 vaccine doses. Neutralization activities, reported as the lowest reciprocal plasma dilution at which 386 neutralization was observed in all wells of a triplicate assay, against the wild-type (WT, ancestral 387 WA1/2020 strain) and Omicron (OM) virus isolates a subset of HCW (blue circles) and older adults 388 (orange circles) who remained COVID-19 naive throughout the study. Participant Ns are shown at the 389 bottom of the plot. A thick horizontal red bar represents the median; thinner horizontal red bars 390 represent the IQR. P-values were computed using the Wilcoxon matched pairs test (for within-group 391 comparisons) or the Mann-Whitney U-test (for between-group comparisons) and are uncorrected for 392 multiple comparisons. 393 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint

394 1. Fisman DN, Bogoch I, Lapointe-Shaw L, McCready J, Tuite AR. Risk Factors Associated With 395 Mortality Among Residents With Coronavirus Disease 2019 (COVID-19) in Long-term Care 396 Facilities in Ontario, Canada. JAMA Netw Open 2020; 3:e2015957. 397 2. Chinnadurai R, Ogedengbe O, Agarwal P, et al. Older age and frailty are the chief predictors of 398 mortality in COVID-19 patients admitted to an acute medical unit in a secondary care setting- a 399 cohort study. BMC Geriatr 2020; 20:409. 400 3. Banerjee A, Pasea L, Harris S, et al. Estimating excess 1-year mortality associated with the COVID-401 19 pandemic according to underlying conditions and age: a population-based cohort study. 402 Lancet 2020; 395:1715-25. 403 4. Baden LR, El Sahly HM, Essink B, et al. Efficacy and Safety of the mRNA-1273 SARS-CoV-2 404 Vaccine. N Engl J Med 2021; 384:403-16. 405 5. Polack FP, Thomas SJ, Kitchin N, et al. Safety and Efficacy of the BNT162b2 mRNA Covid-19 406 Vaccine. N Engl J Med 2020; 383:2603-15. 407 6. Dagan N, Barda N, Kepten E, et al. BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass 408 Vaccination Setting. N Engl J Med 2021; 384:1412-23. 409 7. Grupper A, Rabinowich L, Schwartz D, et al. Reduced humoral response to mRNA SARS-CoV-2 410 BNT162b2 vaccine in kidney transplant recipients without prior exposure to the virus. Am J 411 Transplant 2021; 21:2719-26. 412 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 8. Massarweh A, Eliakim-Raz N, Stemmer A, et al. Evaluation of Seropositivity Following BNT162b2 413 Messenger RNA Vaccination for SARS-CoV-2 in Patients Undergoing Treatment for Cancer. 414 JAMA Oncol 2021; 7:1133-40. 415 9. Monin L, Laing AG, Munoz-Ruiz M, et al. Safety and immunogenicity of one versus two doses of 416 the COVID-19 vaccine BNT162b2 for patients with cancer: interim analysis of a prospective 417 observational study. Lancet Oncol 2021; 22:765-78. 418 10. Collier DA, Ferreira I, Kotagiri P, et al. Age-related immune response heterogeneity to SARS-419 CoV-2 vaccine BNT162b2. Nature 2021; 596:417-22. 420 11. Muller L, Andree M, Moskorz W, et al. Age-dependent Immune Response to the Biontech/Pfizer 421 BNT162b2 Coronavirus Disease 2019 Vaccination. Clin Infect Dis 2021; 73:2065-72. 422 12. Brockman MA, Mwimanzi F, Lapointe HR, et al. Reduced magnitude and durability of humoral 423 immune responses to COVID-19 mRNA vaccines among older adults. J Infect Dis 2021; Online 424 ahead of print. Accepted 9 Dec 2021. doi: 10.1093/infdis/jiab592. 425 13. Barda N, Dagan N, Cohen C, et al. Effectiveness of a third dose of the BNT162b2 mRNA COVID-426 19 vaccine for preventing severe outcomes in Israel: an observational study. Lancet 2021; 427 398:2093-100. 428 14. Gilboa M, Mandelboim M, Indenbaum V, et al. Early Immunogenicity and safety of the third dose 429 of BNT162b2 mRNA Covid-19 vaccine among adults older than 60 years; real world experience. 430 J Infect Dis 2021; Online ahead of print. Accepted, 29 Nov 2021. doi: 10.1093/infdis/jiab584. 431 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 15. Bar-On YM, Goldberg Y, Mandel M, et al. Protection of BNT162b2 Vaccine Booster against 432 Covid-19 in Israel. N Engl J Med 2021; 385:1393-400. 433 16. Falsey AR, Frenck RW, Jr., Walsh EE, et al. SARS-CoV-2 Neutralization with BNT162b2 Vaccine 434 Dose 3. N Engl J Med 2021; 385:1627-9. 435 17. Mizrahi B, Lotan R, Kalkstein N, et al. Correlation of SARS-CoV-2-breakthrough infections to 436 time-from-vaccine. Nat Commun 2021; 12:6379. 437 18. Goldberg Y, Mandel M, Bar-On YM, et al. Waning Immunity after the BNT162b2 Vaccine in 438 Israel. N Engl J Med 2021; 385:e85. 439 19. Levin EG, Lustig Y, Cohen C, et al. Waning Immune Humoral Response to BNT162b2 Covid-19 440 Vaccine over 6 Months. N Engl J Med 2021; 385:e84. 441 20. Collie S, Champion J, Moultrie H, Bekker LG, Gray G. Effectiveness of BNT162b2 Vaccine 442 against Omicron Variant in South Africa. N Engl J Med 2022; 386:494-6. 443 21. Cele S, Jackson L, Khoury DS, et al. Omicron extensively but incompletely escapes Pfizer 444 BNT162b2 neutralization. Nature 2021; Online ahead of print. Accepted, 23 Dec 2021. doi: 445 10.1038/s41586-021-04387-1. 446 22. Song JS, Lee J, Kim M, et al. Serial Intervals and Household Transmission of SARS-CoV-2 447 Omicron Variant, South Korea, 2021. Emerg Infect Dis 2022; 28. 448 23. Bentley E, Mee ET, Routley S, et al. Collaborative Study for the Establishment of a WHO 449 International Standard for SARS-CoV-2 RNA. World Health Organization, Expert Committee on 450 Biological Standardization 2020:WHO/BS/2020.402. 451 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 24. Amirthalingam G, Bernal JL, Andrews NJ, et al. Serological responses and vaccine effectiveness 452 for extended COVID-19 vaccine schedules in England. Nat Commun 2021; 12:7217. 453 25. Parry H, Bruton R, Stephens C, et al. Differential immunogenicity of BNT162b2 or ChAdOx1 454 vaccines after extended-interval homologous dual vaccination in older people. Immun Ageing 455 2021; 18:34. 456 26. Payne RP, Longet S, Austin JA, et al. Immunogenicity of standard and extended dosing intervals of 457 BNT162b2 mRNA vaccine. Cell 2021; 184:5699-714 e11. 458 27. Cho A, Muecksch F, Schaefer-Babajew D, et al. Anti-SARS-CoV-2 receptor-binding domain 459 antibody evolution after mRNA vaccination. Nature 2021; 600:517-22. 460 28. Stamatatos L, Czartoski J, Wan YH, et al. mRNA vaccination boosts cross-variant neutralizing 461 antibodies elicited by SARS-CoV-2 infection. Science 2021; 372:1413-8. 462 29. Wang Z, Muecksch F, Schaefer-Babajew D, et al. Naturally enhanced neutralizing breadth against 463 SARS-CoV-2 one year after infection. Nature 2021; 595:426-31. 464 30. Tan CW, Chia WN, Qin X, et al. A SARS-CoV-2 surrogate virus neutralization test based on 465 antibody-mediated blockage of ACE2-spike protein-protein interaction. Nat Biotechnol 2020; 466 38:1073-8. 467 31. Doria-Rose NA, Shen X, Schmidt SD, et al. Booster of mRNA-1273 Strengthens SARS-CoV-2 468 Omicron Neutralization. medRxiv 2021; Posted 20 Dec 2021. doi: 469 10.1101/2021.12.15.21267805. Preprint. 470 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 32. Schmidt F, Muecksch F, Weisblum Y, et al. Plasma neutralization properties of the SARS-CoV-2 471 Omicron variant. medRxiv 2021; Posted 13 Dec 2021. doi: 10.1101/2021.12.12.21267646. 472 Preprint. 473 33. Garcia-Beltran WF, St Denis KJ, Hoelzemer A, et al. mRNA-based COVID-19 vaccine boosters 474 induce neutralizing immunity against SARS-CoV-2 Omicron variant. Cell 2022; 185:457-66 e4. 475 34. Muik A, Lui BG, Wallisch AK, et al. Neutralization of SARS-CoV-2 Omicron by BNT162b2 476 mRNA vaccine-elicited human sera. Science 2022; 375:678-80. 477 35. Debes AK, Xiao S, Egbert ER, et al. Comparison of total and neutralizing SARS-CoV-2 spike 478 antibodies against omicron and other variants in paired samples after two or three doses of 479 mRNA vaccine. medRxiv 2022; Posted 31 Jan 2022. doi: 10.1101/2022.01.26.22269819. 480 Preprint. 481 36. Ai J, Zhang H, Zhang Y, et al. Omicron variant showed lower neutralizing sensitivity than other 482 SARS-CoV-2 variants to immune sera elicited by vaccines after boost. Emerg Microbes Infect 483 2022; 11:337-43. 484 37. Muecksch F, Wang Z, Cho A, et al. Increased Potency and Breadth of SARS-CoV-2 Neutralizing 485 Antibodies After a Third mRNA Vaccine Dose. bioRxiv 2022; Posted 15 Jan 2022. doi: 486 10.1101/2022.02.14.480394. Preprint.:2022.02.14.480394. 487 38. Feng S, Phillips DJ, White T, et al. Correlates of protection against symptomatic and asymptomatic 488 SARS-CoV-2 infection. Nat Med 2021; 27:2032-40. 489 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 39. Redd AD, Nardin A, Kared H, et al. Minimal cross-over between mutations associated with 490 Omicron variant of SARS-CoV-2 and CD8+ T cell epitopes identified in COVID-19 491 convalescent individuals. bioRxiv 2021; Posted 9 Dec 2021. doi: 10.1101/2021.12.06.471446. 492 Preprint. 493 40. Peng Y, Felce SL, Dong D, et al. An immunodominant NP105-113-B*07:02 cytotoxic T cell 494 response controls viral replication and is associated with less severe COVID-19 disease. Nat 495 Immunol 2022; 23:50-61. 496 41. Swadling L, Diniz MO, Schmidt NM, et al. Pre-existing polymerase-specific T cells expand in 497 abortive seronegative SARS-CoV-2. Nature 2022; 601:110-7. 498 42. Grifoni A, Sidney J, Vita R, et al. SARS-CoV-2 human T cell epitopes: Adaptive immune response 499 against COVID-19. Cell Host Microbe 2021; 29:1076-92. 500 43. Dan JM, Mateus J, Kato Y, et al. Immunological memory to SARS-CoV-2 assessed for up to 8 501 months after infection. Science 2021; 371:eabf4063. 502 44. Riou C, Keeton R, Moyo-Gwete T, et al. Escape from recognition of SARS-CoV-2 variant spike 503 epitopes but overall preservation of T cell immunity. Sci Transl Med 2022; 14:eabj6824. 504 45. Keeton R, Tincho MB, Ngomti A, et al. T cell responses to SARS-CoV-2 spike cross-recognize 505 Omicron. Nature 2022; Online ahead of print. Accepted, 31 Jan 2022. doi: 10.1038/s41586-022-506 04460-3. 507 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 46. Liu J, Chandrashekar A, Sellers D, et al. Vaccines Elicit Highly Conserved Cellular Immunity to 508 SARS-CoV-2 Omicron. Nature; Online ahead of print, 31 Jan 2022. doi: 10.1038/s41586-022-509 04465-y. 510 47. Steensels D, Pierlet N, Penders J, Mesotten D, Heylen L. Comparison of SARS-CoV-2 Antibody 511 Response Following Vaccination With BNT162b2 and mRNA-1273. JAMA 2021; 326:1533-5. 512 48. Nanduri S, Pilishvili T, Derado G, et al. Effectiveness of Pfizer-BioNTech and Moderna Vaccines 513 in Preventing SARS-CoV-2 Infection Among Nursing Home Residents Before and During 514 Widespread Circulation of the SARS-CoV-2 B.1.617.2 (Delta) Variant - National Healthcare 515 Safety Network, March 1-August 1, 2021. MMWR Morb Mortal Wkly Rep 2021; 70:1163-6. 516 517 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Table 1: Participant characteristics and sampling information Variable category Characteristic Healthcare Workers (n=81) Older Adults (n=56) COVID-19 Convalescent at study entry (n=14) Sociodemographic/health Age in years, median [IQR]a 41 [35-51] 78 [73-83] 48 [36-87] Female sex, n (%) 61 (75%) 38 (68%) 10 (71%) White/Caucasian ethnicity, n (%) 37 (46%) 43 (77%) 7 (50%) Chronic health or immunosuppressive conditions, median [IQR] 0 [0-0] 1 [0-2] 0 [0-1] Vaccine information Comirnaty, First mRNA Vaccine, n (%) 80 (99%) 48 (86%) 13 (93%) Comirnaty, Second mRNA Vaccine, n (%) 79 (98%) 46 (82%) 13 (93%) Time between first and second doses in days, median [IQR] 97 [91-102] 76 [45-85] 112 [87-118] Comirnaty, Third mRNA Vaccine, n (%)b 32/61 (52%) 19/47 (40%) 3/6 (50%) Time between second and third dose in days, median [IQR] 210 [200-241] 169 [160-231] 189 [170-194] Specimen collection Specimens collected pre-vaccine, n (%) 80 (99%) 49 (88%) 13 (93%) Specimens collected one month after first dose, n (%) 79 (98%) 49 (88%) 13 (93%) Day of specimen collection one month after first dose, median [IQR] days 28 [27-30] 30 [28-32] 31 [28-32] Specimens collected one month after second dose, n (%) 81 (100%) 55 (98%) 14 (100%) Day of specimen collection one month after second dose, median [IQR] days 29 [29-32] 29 [29-31] 32 [30-36] Specimens collected three months after second dose, n (%) 79 (98%) 53 (95%) 13 (93%) Day of specimen collection three months after second dose, median [IQR] days 90 [90-91] 90 [89-92] 90 [87-91] Specimens collected six months after second dose, n (%) 78 (96%) 40 (71%) 10 (71%) Day of specimen collection six months after second dose, median [IQR] days 181 [179-182] 176 [167-182] 180 [179-181] Specimens collected one month after third dose, n (%) 61 (75%) 47 (84%) 6 (38%) Day of specimen collection one month after third dose, median [IQR] days 30 [29-31] 32 [29-33] 30 [29-30] COVID-19 post-vax Anti-N seroconversion during study follow-up 6 (7.4%) 2 (3.6%) - a interquartile range b denominators are the n of specimens collected one month after third dose . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Table 2: Multivariable analysis of the relationship between sociodemographic, health and vaccine-related variables on serum antibody half-life following two-dose COVID-19 mRNA vaccination Outcome measure Variable Estimate 95% CI p-value Ab half-life after two vaccine doses Age (per year) 0.058 -0.17 to 0.29 0.61 Male sex 5.31 -2.57 to 13.18 0.18 White ethnicity 3.11 -4.67 to 10.88 0.43 # chronic conditions (per add'l) -4.62 -8.39 to -0.85 0.017 Spikevax as first dose 3.37 -13.26 to 20.00 0.69 Dose interval (per day) 0.00014 -0.16 to 0.16 0.99 COVID-19 convalescenta 13.78 -0.37 to 27.93 0.056 a participants with positive anti-N serology at study entry . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Table 3: Multivariable analyses of the relationship between sociodemographic, health and vaccine-related variables on Omicron- specific humoral immunogenicity measures following three-dose COVID-19 mRNA vaccination Humoral measure Variable 1 mo after 3rd dose Estimate 95% CI p-value anti-Omicron RBD IgG (log10)a Age (per year) 0.0035 -0.0027 to 0.0097 0.26 Male sex -0.14 -0.34 to 0.054 0.15 White ethnicity -0.018 -0.21 to 0.17 0.85 # chronic conditions (per add'l) -0.12 -0.20 to -0.041 0.0033 Spikevax as third dose (vs. Comirnaty) 0.15 -0.039 to 0.34 0.12 Interval between 1st and 2nd dose (per day) -0.0066 -0.012 to -0.0011 0.020 Interval between 2nd and 3rd dose (per day) 0.00043 -0.0034 to 0.0042 0.83 Days since 3rd vaccine dose -0.0086 -0.038 to 0.021 0.56 Prior COVID-19b 0.1 -0.16 to 0.37 0.43 anti-Omicron ACE2 % displacementa Age (per year) 0.29 -0.046 to 0.63 0.090 Male sex -12.38 -23.16 to -1.60 0.025 White Ethnicity -2.36 -12.74 to 8.01 0.65 # chronic conditions (per add'l) -6.41 -10.80 to -2.03 0.0046 Spikevax as third dose (vs. Comirnaty) 1.69 -8.58 to 11.97 0.74 Interval between 1st and 2nd dose (per day) -0.41 -0.71 to -0.11 0.0079 Interval between 2nd and 3rd dose (per day) -0.038 -0.25 to 0.17 0.72 Days since 3rd vaccine dose -1.82 -3.43 to -0.21 0.027 Prior COVID-19b 12.28 -2.11 to 26.67 0.094 a Measured using the Meso Scale Diagnostics (MSD) V-plex assay system b Includes all participants with positive anti-N serology at any time during the study (i.e. both pre- and post-vaccine COVID-19 cases) . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint -1 0 1 2 3 4 5 Anti-RBD Ab conc (log10 AU/mL): Elecsys assay HCW Older Adults Conv. HCW Older Adults HCW Older Adults HCW Older Adults HCW Older Adults HCW Older Adults pre-vax 1 mo after 1st 1 mo after 2nd 3 mo after 2nd 6 mo after 2nd 1 mo after boost 1 <0.0001 <0.0001 <0.0001 <0.0001 0.3 n=80 49 13 79 49 13 81 55 14 79 53 13 78 42 10 61 47 6 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.3 0.06 0.027 ULOQ LLOQ 0.3 0.6 0.3 Conv. Conv. Conv. Conv. Conv. 4 8 16 32 64 128 256 512 1024 2048 4096neutralization: reciprocal dilution HCW Older Adults HCW Older Adults HCW Older Adults HCW Older Adults HCW Older Adults HCW Older Adults pre-vax 1 mo after 1st 1 mo after 2nd 3 mo after 2nd 6 mo after 2nd 1 mo after boost 1 0.08 <0.0001 <0.0001 0.002 0.6 n=19 12 8 79 49 13 80 52 14 79 53 13 78 42 10 61 47 6 0.006 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 ns 0.009 0.002 0.005 LLOQ ULOQ 0.6 0.5 0.6 Conv. Conv. Conv. Conv. Conv. Conv. a b Figure 1 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint 0 50 100 150 200 0 5000 10000 15000 20000 25000 days after 2nd dose Anti-RBD Ab conc (AU/mL): Elecsys assay HCW Older adults Convalescent ULOQ HCW Older Adults Conv. 0 20 40 60 80 100 120 140 160 antibody half-life in days 0.016 n=67 37 8 0.13 0.016 a b Figure 2 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint WT OM WT OM WT OM WT OM WT OM WT OM 1 2 3 4 5 6 anti-S-RBD IgG (log10 AU/mL): MSD assay 1 mo post 2nd HCW (COVID-19 naive) 1 mo post 3rd 1 mo post 2nd 1 mo post 3rd 1 mo post 2nd 1 mo post 3rd Older Adults (COVID-19 naive) Convalescent (incl. post-vax infection) n=75 55 52 45 22 13 <0.0001 <0.0001 <0.0001 <0.0001 0.0002 0.04<0.0001<0.0001 <0.0001 0.09 0.40 0.49 <0.0001 a b WT OM WT OM WT OM WT OM WT OM WT OM 0 10 20 30 40 50 60 70 80 90 100ACE2 displacement (%) 1 mo post 2nd 1 mo post 3rd 1 mo post 2nd 1 mo post 3rd 1 mo post 2nd 1 mo post 3rd n=75 55 51 45 22 13 <0.0001 <0.0001 <0.0001 <0.0001 0.0002 0.04<0.0001<0.0001 <0.0001 <0.0001 0.08 0.2 <0.0001 HCW (COVID-19 naive) Older Adults (COVID-19 naive) Convalescent (incl. post-vax infection) Figure 3 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint WT Omicron WT Omicron WT Omicron WT Omicron 4 8 16 32 64 128 256 512 1024 2048neutralization: reciprocal dilutionn=20 n=21 p<0.0001 p<0.0001 p<0.0001 p<0.0001 1 mo post 2nd 1 mo post 3rd 1 mo post 2nd 1 mo post 3rd p=0.14 p<0.0001 p=0.0003 p=0.79 LLOQ HCW (COVID-19 naive) Older Adults (COVID-19 naive) Figure 4 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Table S1: Multivariable analyses of the relationship between sociodemographic, health and vaccine-related variables on immunogenicity measures following two-dose COVID-19 mRNA vaccination Humoral measure Variable 1 mo after 2nd dose 3 mo after 2nd dose 6 mo after 2nd dose Estimate 95% CI p-value Estimate 95% CI p-value Estimate 95% CI p-value anti-RBD Age (per year) -0.0061 -0.010 to -0.0017 0.0067 -0.0047 -0.0085 to -0.00077 0.019 -0.0029 -0.0076 to 0.0018 0.22 Abs Male sex 0.012 -0.14 to 0.17 0.88 0.089 -0.052 to 0.23 0.21 0.062 -0.083 to 0.21 0.40 (log10)a White ethnicity 0.098 -0.056 to 0.25 0.21 0.15 0.012 to 0.29 0.033 0.089 -0.055 to 0.23 0.22 # chronic cond. (per add'l) -0.096 -0.17 to -0.022 0.011 -0.11 -0.18 to -0.047 0.001 -0.14 -0.21 to -0.068 0.0001 Spikevax as first dose 0.25 -0.038 to 0.54 0.088 0.32 0.054 to 0.59 0.019 0.26 -0.039 to 0.57 0.087 Dose interval (per day) 0.0034 0.00011 to 0.0067 0.043 0.0054 0.0025 to 0.0083 0.0003 0.0049 0.00194 to 0.0079 0.0014 Days since 2nd dose 0.0039 -0.026 to 0.033 0.80 0.015 -0.0098 to 0.040 0.24 0.00095 -0.0088 to 0.010 0.85 COVID-19 convalescentc 0.16 -0.087 to 0.42 0.20 0.23 0.0081 to 0.46 0.042 0.26 0.00938 to 0.51 0.042 Viral Age (per year) -0.018 -0.033 to -0.0030 0.019 -0.020 -0.034 to -0.0052 0.008 -0.0022 -0.016 to 0.012 0.75 neut Male sex -0.33 -0.85 to 0.18 0.21 0.22 -0.30 to 0.75 0.40 0.14 -0.29 to 0.57 0.53 (log2)b White ethnicity -0.075 -0.59 to 0.43 0.77 0.27 -0.24 to 0.78 0.30 0.067 -0.36 to 0.49 0.76 # chronic cond. (per add'l) -0.10 -0.34 to 0.14 0.42 -0.16 -0.40 to 0.095 0.22 -0.0080 -0.21 to 0.20 0.94 Spikevax as first dose 0.86 -0.098 to 1.82 0.078 0.71 -0.30 to 1.7 0.17 0.81 -0.088 to 1.71 0.077 Dose interval (per day) 0.0066 -0.0044 to 0.018 0.24 -0.00046 -0.011 to 0.010 0.93 0.0074 -0.0015 to 0.016 0.10 Days since 2nd dose 0.0040 -0.094 to 0.10 0.94 -0.066 -0.16 to 0.026 0.16 0.010 -0.019 to 0.039 0.48 COVID-19 convalescent 1.64 0.81 to 2.47 0.0001 1.84 1.0 to 2.7 <0.0001 1.46 0.72 to 2.19 0.0002 a measured using the Elecsys Anti-SARS-CoV-2 S assay b for viral neutralization, reciprocal plasma dilutions were log2 transformed prior to multivariable analysis. c participants with positive anti-N serology at study entry . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint Table S2: Multivariable analyses of the relationship between sociodemographic, health and vaccine-related variables on humoral responses following three-dose COVID-19 mRNA vaccination Humoral measure Variable 1 mo after 3rd dose Estimate 95% CI p-value anti-RBD Abs (log10)a Age (per year) 0.0018 -0.0011 to 0.0048 0.22 Male sex 0.0080 -0.086 to 0.10 0.87 White ethnicity 0.0089 -0.082 to 0.10 0.85 # chronic conditions (per add'l) -0.053 -0.091 to -0.014 0.0078 Spikevax as third dose (vs. Comirnaty) 0.12 0.030 to 0.21 0.0091 Interval between 1st and 2nd dose (per day) -0.00064 -0.0033 to 0.0020 0.63 Interval between 2nd and 3rd dose (per day) 0.00037 -0.0014 to 0.0022 0.69 Days since 3rd vaccine dose 0.0016 -0.013 to 0.016 0.82 Prior COVID-19c 0.070 -0.057 to 0.20 0.28 Viral neut. (log2) b Age (per year) 0.022 0.0025 to 0.042 0.028 Male sex -0.075 -0.70 to 0.55 0.81 White ethnicity -0.28 -0.88 to 0.33 0.37 # chronic conditions (per add'l) -0.17 -0.43 to 0.081 0.18 Spikevax as third dose (vs. Comirnaty) 0.70 0.11 to 1.29 0.021 Interval between 1st and 2nd dose (per day) 0.012 -0.0056 to 0.029 0.18 Interval between 2nd and 3rd dose (per day) 0.013 0.00097 to 0.025 0.035 Days since 3rd vaccine dose -0.028 -0.12 to 0.065 0.55 Prior COVID-19c 1.23 0.39 to 2.06 0.0044 a Measured using the Elecsys Anti-SARS-CoV-2 S assay b Viral neutralization results were log2-transformed prior to multivariable analysis c Includes all participants with positive anti-N serology at any time during the study (i.e. both pre- and post-vaccine COVID-19 cases) . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 21, 2022. ; https://doi.org/10.1101/2022.01.06.22268745doi: medRxiv preprint