Determinants of SARS-CoV-2 IgG response and decay in Canadian healthcare workers: a prospective cohort study

2752 of 4567 vaccinated cohort par ticipa nts agree d to don ate a t leas t one bl ood s ample. Mod elling of anti-RBD IgG tit er from 8903 samples showed an i ncrease in Ig G with each vaccine dose an d with first infection. A decreas e in IgG ti te r was found with the number of months since vaccinatio n or i nfection, with t he sharp est decline afte r th e third d ose. A n immunization r egime tha t included mRN A1273 (Modern a) result ed in higher anti-RBD IgG . Participan ts repo rting mult iple sclerosis, rh eumat oid ar thritis o r taki ng selective immunosuppressan ts, tumor n ecrosis fac tor inhibi tors, calcin eurin inhibi tors an d antine oplastic age nts had lower anti-RBD Ig G. Suppl ement ary anal yses showed higher anti-RBD IgG in thos e repo rting side-effects of vaccination, no r elati on of anti-RBD IgG t o obesity and lowe r tit ers in women immunized ea rly in pregnancy. Sensitivity analysis resul ts suggested no i mportan t bias in th e self-repor t dat a.

Creation of a pr ospective coho rt was cen tral t o the cr edibility of resul ts pres ent ed here. S erial se rology assessments, with longi tudinal an alysis, provided effect es timat es with enhanc ed accuracy and a cleare r underst anding of medical and ot her fact ors affecting respons e to vaccinatio n. Key words: COVID-19; SARS-CoV-2; anti-RBD IgG; cohor t; he althca re worke rs; vaccination . 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under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 3

In the e arly weeks of the COVID-19 pand emic, a cohort of he althca re worke rs (HC Ws) was established in four Canadian provinces t o det ermine modifi able workplace facto rs associat ed with in creased risk of infection a nd of mental dist ress when working thr ough t he pandemic. Although this project envisaged using serology samples to identify or confirm SARS-CoV-2 infection, the ne ed for serial sampl es of asymptomatic HCWs only became appare nt with th e sta rt of SARS-CoV-2 immunization in Decembe r 2020. In r esp onse to a call from th e Canadian Immunology Task Force for data t o bet te r underst and th e immunological resp ons e to SARS-CoV-2 vaccines, we establish ed a serol ogy sub-cohort to coll ect serial bl ood samples for de te rminati o n of SARS-CoV-2 antibodies. At that poi nt in th e pand emic ther e was littl e knowledge abou t the effectiveness of vaccinations or the dec ay in immunological response , alth ough this r apidly became availabl e from clinical tria ls [1][2] [3] and observational studies [4][5]6] [7]. Much useful data has come from cohort studi es of HCWs [8] [ 9] [10] [11] [12 ] that were se t up close to th e star t of vaccination again st COVID-19, but long-term follow-up with repea ted me asures b eyond the first mon ths of vaccination ar e less in evidence [12] [13]. Using data from a serology sub-cohort of Canadian HCWs, we sought to est ablish an over all model of the r elati on betw een an ti-rece ptor binding domain (an ti-RBD) IgG to t he spike pro tein of th e SARS-CoV-2 virus (anti-RBD IgG tite r) and the pa tt er n of SARS-CoV-2 vac cinations and infections over time. As the se rology sub-study was embedded in an on-going cohort, we wer e able to also consider wider qu estions r aised els ewhe re, including th e impact of immune-rela t ed medical conditi on [14] [15] and medicatio ns [16] [17], the role, if an y, of obesity [18] and pregnancy [19] and the rel ation of r epor ted side- effects of vaccination to th e str ength of t he anti-RBD IgG respons e [20]. Since the start of ser ology sampling in this cohort , knowledge on th e immunolo gical response t o vaccines and infection and the d ecay over time has increased expon enti ally [21]. The results from the sta tistical mod els develop ed he re to add ress th ese ques tions can be inte rpre ted in this vastly bet ter-in formed conte xt .

The initial coho rt rec ruit ed HCWs from four Canadian pr ovinces (Albert a, Bri tish Columbia, Onta rio and Qu ebec) in the ea rly months (April-Oct ober) of th e 2020 COVID-19 pandemic. This is described more fully elsewher e [22] but, in brief, HCWs wer e appr oached thr ough their p rovincial profession al organi zations and invit ed to con tact the res earch team if they wer e inte rest e d to take p art . Physicians were recrui ted from all four provinces, registe red and lice nsed prac tical nurs es and health ca re aid es only from Alber ta a nd personal supp ort work ers only from Ontario . All consen ting par tici pants comple ted an onlin e baseli ne/recr uitment qu estio nnair e in Spring/Summer 2020 and were cont acte d to complet e furth er onlin e questi onnai res in lat e fall 2020, in April 2021 and April 2022. The second qu estio nnaire asked if th e HCW would donat e a blood sample to l ook for undiagnosed COVID-19 via detecti on of a nti-nucleocapsid Ig G, emphasi zed a r equ est to no tify the res earch team of all positive tes ts and requ est ed conse nt to ob tain r ecords of positive SA RS-CoV-2 diagnostic nucleic acid tests (NAT) including polymerase chain reac tio n (PCR ) tests from the public heal th au th orities of th eir home p rovince. The 3 rd and 4 th questi onnair es asked for details of all vaccinatio ns received a nd, i n the 4 th qu estionn aire , consent to obt ain immunizatio n recor ds from the province. Shor tly after the first immuniz ations (19 th December 2020) participan ts were told abo ut th e plan t o collect seri al serology blood sampl es fro m those who were willing and who felt it was feasible. Sampl es were co llected 4, 7, 10 and 13 mon ths after the f irst vaccine dose. At th e time of each sample, th e par ticipant was aske d to complet e a very brief ques tionnai re (online or by phone) to upda te . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 4 their r ecord of immuniza tions (date a nd manufacture r) and infection hist ory, wit h details of how this was confirmed (by PCR (NAT) or rapid antige n testing). Blood samples wer e collect ed thr ough commercial or provincial blo od collectio n facilities. Thes e covered the whole province of Albe rta bu t only parts of other pr ovinces. In B ritish Columbia (B C), with the support of th e BC Centre for Diseas e Control , those o utside the ar ea covere d by the commercial se r vice could give samples through th e BC public clinical labora tori e s. In Ont ario an d Queb ec, those n ot cove red by the commercia l services could only give samples through ad hoc a rrangemen ts. Som e did so, but for many it was not feasible to give repea ted sampl es. All sampl es were forw arded t o the Alber ta Bio Sample R eposito ry at the Unive rsity of Alber ta and stor ed at -20C for late r analysis. This was carried out using th e Abbo tt A rchit e ct platform (Abbo tt Diagnostics, Chicago, IL, USA) with the S ARS-CoV-2 IgG II QUANT assay quan tit ati vely detecting an tibodi es direct ed against the RBD region of t he sp ike prot ein. The r esult of this assay, r efer red to as the an ti-RBD IgG tite r, is the o utcome me asure us ed thr ou ghout this ar ticle . The quantifiabl e range was 6.8-40,000 AU/ml: a value of ≥50 AU/ml was interpr ete d as a positive resul t. Sampl es were an alyzed a s they were rec eived, a t eith er of two labora tory cent ers, t o give timely feedback to pa rticipan ts. Any pa rticipan t whose sample result was <50AU/ml was contacte d person ally by the Principal Investiga tor to discuss likely causes and implications . To minimize the possibili ty of changes over time (in labora tory re agents, machin es o r procedur es) that coul d lead to biased r esults in da ta seri ally analyzed , all samples with sufficient residu al seru m were rean alyzed for an ti- RBD IgG under ca refully standa rdized co nditions in a single labo rat ory during su mmer of 2022. The results of th e standar dized r eanalysis have be en used f or the p resen t analysis, when availabl e. Wher e no serum r emained, the original resul t was ret ained . The Abbo tt Architec t SARS-CoV-2 IgG antib ody assay was also used to det ect antibodi es against t he viral nucle ocapsid prote in, reflec ting infection with the SA R S-CoV-2 virus. In line with th e manufacture r’s recommen dati on, a nucl eocapsid resul t was consider ed positive with a value of 1.4 AU/ml or great er. The time be tween each vaccine dose and th e dat e of sample collecti on was calculated for each of the serology samples. Vaccinati on occurring l ess than seven days prio r to sample coll e ction was not conside red for that sampl e. The da te of infectio n with COVID-19 was taken as the r epor ted da te of a positive PCR (NAT) or rapid antig en tes t. Immunological conditio ns were asce rtain ed from a checklist of six conditions plus ‘othe r’ and ‘n one’ o n the third and fourth qu estionn aire (Supplem ent ar y materials 11). Medica tions assess ed were thos e repo rt ed by participan ts as likely or possibly affecting response t o the vaccine (Supplem enta ry materials 1.2). As the re wer e few changes to medication over th e peri od of vaccinations, thos e star ted b efore t he first vaccine were used in the analysis. R esponses to th e checklist o f autoimmune conditi ons (rheumat oid ar thritis , ankylosing spondylitis, psoriatic a rthri tis, psori asis, inflammator y bowel disease, and lupus) were acc ept ed without r eview. Thre e physician members of the t eam reviewe d all free-te xt entri es ent er ed as “othe r chronic autoimmun e inflammatory diseases” or volun tee red i n the medica tion secti on: multipl e sclero sis was identified from these responses . Medica tions wer e coded in to seven categori es based o n the in ter natio nal classification of diseases (ICD) 11 th revision [23] (World Health Org anizati on, 2020): metho tre xa te, tumor n ecrosis factor (TNF) inhibitors , glucocorticoids, in terl eukin inhibit ors, cal cineurin inhibi tors, sel ective immunosup pressants , and an tineopl astic agents (Supplemen tary mat erials 1 .3). Coding was done indepe nden tly with disagreemen ts solved by consensus. Gend er as self-repo rt ed on th e recrui tm ent ques tionnai re and age at e ach sampl e, calculat ed from dat e of birth, were included in e ach analysis. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 5 Variables e xplor ed in th e supplemen tary analyses were defin ed as follows. Body Mass Inde x (BMI) was calculated from self-rep ort of heigh t and weight as kg/m 3 , collected at th e third co ntact . For pr egnancy, dat e of each concepti on was collecte d at th e 4 th contact . Pregnancy gesta tion at the tim e of vaccination could not be calculated exac tly, but was estima ted fro m the month of concep tion re por ted by t he particip ant , with ‘ea rly’ gestati on including those with the vaccin e dose in th e same month as conce ption or in the two following months, ‘mid’ gest atio n in months 4 or 5 and ‘lat e’ gesta tion in mon ths 6-10 of a continuing pregna ncy. Adverse side effects at first an d second vaccinatio n dose were collec ted from a list of thi rt een possible effects at e ach dose (Supplemen tary mat erial 1.4). The participan t was asked to ma rk on a visual analogue scale (from ‘not a t all’ to ‘very’) how much they had be en b other ed by each symptom. Two provinces (Alberta an d British Colu mbia) were able to p rovide public he alth records of vaccination an d infection for par ticipan ts who consent ed . The self-repor t and public he alth r ecord s were reconcile d by adding unrepo rte d vaccinations and cor recti ng dates and types of vaccine. Any PCR(NAT) confirmed cases of COVID-19 from public health r ecords wer e added a nd existing da tes cor rect ed. The r econcil ed records we re used for the analysis repor ted h ere , with a sensitivity analysis comparing models from the o riginal and reconcil ed dat a. Statis tical meth ods. To reduce skewness, th e na tural loga rith m of anti-RBD IgG concentr atio n was used in the analysis. M ean and standar d deviatio n (SD) of the log anti-RBD IgG were calcula ted for e ach catego ry of factors and tabula ted , along with the cor responding geom etric me an of the original an ti-RBD IgG concen tra tio n. Mean tim e in months and its SD from each vaccine dose, or confir med case of COVID-19, to the collec tion of blood sample was calculated and tabul ated for each of the 4 tim e poin ts of blood sample collecti on and over all. We used a line ar mixe d model with a Gaussi an random-in terce pt for each HCW to describ e the vari ation of log anti-RBD IgG values across the HCWs by age and gend er and how the log IgG levels chang ed over tim e according to th e following factors: th e pres ence/absenc e of the firs t and second confirmed COVID-19 infecti ons; time from each infec tion to the IgG measu remen t; nucle ocapsid p ositivity status (positive , negative , missing); reception of up t o four SARS-CoV-2 vaccinations prior to the sa mple; time from each vaccina tion to coll ection of th e sample; typ e of the vaccines (only P fizer, any Modern a, and any non-mRNA vaccines). The times fro m the 1 st vaccination and the 2 nd vaccination were almos t red undan t (i.e., Pearson corr elati on = 0.95) and thus only the lat te r was ente red t o the models. The value ascrib ed to vaccine 1 (present for all pa rticip ants) is the model intercep t. To t his initial mod el, we added th e seven indica tors of the me dical conditions (multiple scle rosis, rh eu matoid ar thri tis, ankylosing spondylitis, psoria tic ar thritis , psoriasis, i nflammatory bowel diseas e, and lupus) a nd seven indicato rs of medication use (met hot rex at e, TNF inhib itors, glucocor ticoids, int erl eukin inhibito rs, calcineurin inhibi tors , selective immunosuppr essants , and an ti neoplast ic agents) to evalu ate the associ ation of th ese clinical factors with log anti-RBD IgG levels, with an in di cator of missing medical data . Thre e supplement ary analyses were conducted t o assess th e associatio ns of the severity of advers e effects (on the 2 nd dose of vaccination), pregnancy at th e time of vaccinati on, an d BMI with th e log anti-RBD IgG l evels: th is was accomplished by adding each of the th re e variables to th e second model above. For the an alysis of adverse react ions to vaccinati on a separa te princip al componen t analysis of log-transformed visual analogu e scales was carried out for adverse responses ascrib ed to the 1 st an d 2 nd dos es. Unro ta ted compon ents with eig en values >1.0 wer e ex trac ted and scores re tained for analysis against an ti-RBD IgG. As a sensi tivity analysis, the ini t ial model was re-fit using only 2031 participan ts with 6630 samples (74% of the total) in Alb ert a and Bri tish Columbia who had been match ed . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 6 to provincial r ecords of PCR (NAT) confir med SARS-CoV-2 infection and vaccinatio n. The Akaike Info rmation Criterion (A IC) was calculated for each da tase t to de termin e wheth er th e fit impro ved when public health d ata were incorpo rat ed All statis tical analyses wer e conduct ed b y Stata version 18 and S AS 9.4. Prob abilit ies quote d were two-side d.

Of the 4964 HCWs consenting t o join the prospective coho rt, 2752 gave at l east o ne post vaccine ser ology sample. This repr esen ted 60% (2752/456 7) of those eligible, e xcluding 397 HCW not known to have be en vaccinated (Figure 1). Of the 2752 giving samples, 292 gave 1, 323 two, 583 thre e and 1554 four, for a to tal of 8903 samples. Of thes e, 96.9% (8629/89 03) had IgG values from the cont rolled re analysis. The geome tric mean (GM) of anti-RBD IgG decre ased from th e 4-month sample (2689 AU/ml) to 1190 AU/ml at 7 months th en increased to 3091AU/ml at th e 10-mont h sample and again to 10,282 AU/ml at 13 months. The GMs of an ti-RBD IgG by month from March/Ap ril 2021, th e first dat e for a 4-month sample since fi rst vaccine, to July 2022, the end of sample collectio n, ar e shown in Supplemen tary mate rials 2. The lowes t G M was in November 2021. The distribu tion of factors conside red in the analysis is given in Table 1, toge the r with the un transforme d GM and mean log anti-RBD Ig G unadjuste d for othe r factors. Th e mean time b etwe en vaccines and infection prio r to each sample is shown in Table 2. All bu t 46 of the 2216 participan ts giving a 4-mo nth sample had r eceived bo th first and second doses b efore t his first sa mple. The initi al model including gende r, age and history of vaccination and infection is shown in Table 3. Gend e r was unrelat ed to the me asured immun e respons e and a decr easing response with ag e was only marked in th ose aged 65 or old er. A p ositive effect on anti-RBD IgG was seen for each of four vaccine doses. Figure 2 sho ws the predict ed values and slop es from the initi al model (Table 3) for those who had r eceived e ach of the four vaccines hypothetically a t its mean time point of the coh ort memb ers (red line), toge ther with smoo the d obser ved values for those with (solid black) and without (dashed black) the third vaccine dose by th e 10-month sam ple. The shar p increas es in the pr edict ed line reflect the es timat ed effects of the 3 rd and 4 th vaccine doses . The predic ted slop e post vaccine 2 is shall ower (estimat ed -0.156) than that from vaccine 3 (-0.156 + -0.105 =-0. 261) or vaccine 4 ( -0.156 + -0.105 + 0.012= -0.249). The smoothed observed values follow th e predic ted line closely to the 7-month sampl e, diverge f or the 10-month sampl e and meet again a t th e 13-month sample The predict ed line in Figur e 3 does not ta ke account of infection or type of vaccine. A to tal of 678 of 2752 HCW in the ser ology sub-study had test ed posi tive for SARS-CoV-2 by July 2022, of who m 19 had two test-posit ive infections (Table 1). The effect estima tes for infection were posi tive, with th e add itional effect for the second case much smaller than the first (Table 3 ). The decay slope after infec tion was sha llower than af ter vaccina tion. Having adjusted for known infection a po sitive nucleocapsid an tibody resul t was r elat ed to a highe r log anti-RBD IgG (Table 3). Those who had receive d o nly BNT162b2 (Pfizer-BioNTech) vaccines had lower anti-R BD IgG tit ers than th ose who had rec eived at l east on e dose of mRNA-1273 (Modern a), while t hose whose doses include d any non-mRNA vaccine had lowe r anti-RBD I gG (Table 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 7 The model was then e xt ended to assess t he effects of medical conditions a nd med ications (Table 4). Among the medical conditions consid ered only HCWs with multiple sclerosis and r heuma toid arthri tis had a lowe r than expec ted immune r esponse , as indicat ed by anti-RBD IgG. Most of th e medica tio ns considered wer e nega tively relat ed to the immune r esponse bu t only tumor necro tic factor (TNF) inhibito rs, calcineurin inhibi tors , selective immunosuppressan ts and anti neopl astic agents re duced th e anti-R BD IgG concent ration to a significant degr ee . Three supplem enta ry analyses consider e d: whethe r severi ty of side effects to vaccination, ges tati on in women pregnan t at th e time of vaccinatio n or bo dy mass index (BMI), had any impact on t he immune respons e, when these fact ors were a dded t o the mod el shown in Table 4. To explo re t he rel atio n of report ed side effects to immune response (refl ected in an ti-RBD IgG concent ra tion) thre e principal comp onents an alyses were e xt ract ed from responses t o the 13 symptoms ra te d by responden ts as thei r exp erie nce aft er the firs t and second vaccine dose (Table 5). Only the first compon ent s, interp re ted as str ength of syst emic rea ctogenicity, r elat ed to a n increased a nti-RBD IgG r esponse in bivari ate an alysis. The first componen ts for th e first and second vaccine doses were cor rela ted (Pearson r = 0.53, p<0.001) but only the s treng th of respon se to th e 2 nd dos e added to th e final model (Table 6). Of the 139 women in the ser ology sub-cohort who rep ort ed a pregnancy since the s tar t of the pand emic, 88 received a t leas t one v accine dose during a t leas t one pr egnanc y. These women appea red to have a somewhat lowe r respons e to vacc ination if the vaccine dos e was given in early or mid-pregnancy, when variables reflec ting stage of pr egnancy were add ed to the pr evious model (Table 6). With BMI in the model , ther e was no suggestion th at pa rticipan t s with a higher BMI ha d a worse immune response (Table 6). In the sensi tivity analysis (Table 7) estimates for 2031 par ticipants with 6630 sam ples (74% of the total) in Alber ta and B ritish Columbia who had be en matched to provincial public he alth r e cords were very similar whether s elf-repor t alon e or da ta reco nciling official and self-report wer e used. T he self-repor t dat a account ed for 75.8% of the variance in log anti-R BD IgG and th e dat a corr ected from public h ealth r ecords 79.8%. The AIC for the model with a djustment for pu blic health da ta was smaller (19208) than with just the self-re por t (18220) suggesting the inclusion of public heal th records had impr oved th e fit. Inspec tion of Table 7 indicates a grea te r (negative) effect of non-mRNA vaccines when self-repor ted vaccine type was cor recte d from public health records.

The serial collec tion of samples for sero l ogy testing demonst rat ed th e anticip ate d increase of ant i-RBD IgG with each vaccine dose and t est-confirmed inf ection and the dec reas e in anti-RBD IgG t iters with t he passage of tim e. Furthe r, it showe d a grea ter incr ease in a nti-RBD IgG with use of th e mRNA-1273 vaccine, that nucl eocapsid positivity was an independ ent p redict or of anti-RBD IgG, th e nega tive impact on anti-RBD IgG of cer tain medical conditions and tre atmen ts, a mut ed vaccine respons e in early pr egnancy and an e nhanced resp onse in th ose repor ting grea ter s everity of common side effects. The strengt hs of the study include the r e peat ed measur es design in a large coh ort , with high participan t ret ention , resul ting in par t from rapid fe edback of serology results . A furth er st re ngth is the very high propo rtion of samples included in the system atic re analysis designed to minimize the impact of variability betwe en labora tori es and over tim e. The coho rt d esign provided informa tion on pe rsonal f actors including medical conditions, adve rse vaccine re actions, B MI and pr egnancy. Alth ough the info rma tion on vaccinatio ns and infection was self-repor ted , the pa rticip a nts were all HCWs who may be more acc urat e in thei r repo rting th an . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 8 the gene ral popul ation . Previous studi es have shown that , in pati ents, self-re por t of vaccination status was very largely concordan t with vaccination r eco rds but tha t self-repo rt ed dat es and type of vaccine were less accurate [24] [25]. In the current s tudy public heal th recor ds were accessed for a majority of the serol ogy sub-cohort and the congru ence in estima tes using pur ely self-report e pisodes and episodes cor rec ted from ex te rnal r ecords gives some confidence to th e findings for the serol ogy cohort over all. Weakness es include th e exclusion of a sizeable numb er of the main coh ort for w hom it was not feasible to obtain se rial blood sampl es. Fact ors asso ciated with joining th e serol ogy cohort h ave been descri bed els ewhere [22]. In addition to thos e exclud ed th rou gh lack of local blood collection facilities , younger age, having a child at home and smoking were associa ted with a lower particip atio n. The r estric ted ser ology analysis was a major

. Only serum was re tain ed for a nalysis and it was not possible t o infer an ything about cellul ar immunity. While t he sample siz e was adequat e to confirm th e nega tive impact of certain condi tions and medications, i t did not all ow confident as sertion of novel findings on th e impact of pre-exis ting conditions . Moreove r, this an alysis does not consid e r the r elati on betw een vaccine resp onse and break through infect ion or severity of disease . The use of 40,000 AU/ml as the upper li mit of quantification p roduced the ceilin g effect evident in Figure 2, a nd will have reduced th e accuracy of estima tes based on sampl es at 10 mont hs (with 221 samples (9.8%) recorded at ceiling) and 13 months (425 samples, 18.9%). This will have biased the estima t ed slopes towa rds the null , as values poten tially higher than 40,000 AU /ml were recorde d as 40,000 AU/ml, res ulting in estima tes of decay slopes after the 3 rd and 4 th vaccine doses that a re less st eep t han would have be en expect ed from dat a withou t the 40000 AU/ml ceiling. The observed incr ease in an ti-RBD IgG fo llowing each vaccine dose and decay wit h time are as r epo rted elsewher e [21] [26], as is the slower decay after infection in this cohor t of vaccinated HCWs [21]. The smaller increase aft er t he 4 th vaccine is consiste n t with oth er studi es [13] but is based her e on only 64 participan ts. Similarly, the small incr ease following a second infection was base d on a confirme d second case in only 19 participan ts. The sha rpe r decline aft er t h e 3 rd and 4 th vaccine dose th an afte r the 2 nd is not consist ent with som e earlie r rep orts which suggested a slow er decline afte r th e third d ose [10] [12]. Th e reasons for this difference are unce rtain . All th re e studies wer e of HCWs and all used the Abbot t ARCHITECT system to obt ain quant ita tive estimat es of anti-RBD IgG . The curre nt st udy was much larger and the follow-up a fter th e third vaccine dos e somewhat longer (mean 3 .3 months) tha n in the previous st udies cit ed her e. Ike z aki et al [10] collected sampl es from 52 HCWs 2 months after the third v accine dose and Dodge e t al [12] from 212 HCWs 2-3 months after the third dose . The sta tistical a pproach differ ed betwe en th e studies . The analysis by Dodge et al [12] suggesting a shallower decay slope after the t hird vaccine considere d only a subgroup in a cros s-sectional analysis ra ther than a longitudinal an alysis of IgG values which would have controll ed for be tween-par ticipant fact ors. Ik ezaki et al [10] reporte d a perce ntage d ecline aft er the 2 nd and 3 rd vaccines without giving more de tail. It should b e not ed that the effect of values above the upp er limit of quantification , observed af ter th e third bu t not seco nd vaccine dose in all thr ee stu dies, will have be en d ifferential , biasing estima tes of decay aft er the third dos e, but n ot th e second, towa rds the null . The finding of a faster declin e afte r the 3 rd vaccine dos e in the curre nt study is not easily dismissed: it is str engthen ed by th e rean alysis of stored serum , ensuring r e sults after the 2 nd and 3 rd doses were direc tly comparabl e. In their compr ehensive review Se tt e and Crot ty [21] commented th at ‘conclusi ons about dur ability of [antibo dies] afte r 3-dose mRNA vaccinati on remain unc ert ain’ . . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 9 The rela tion of a positive nucle ocapsid t e st, assayed from the sam e sample as th e anti-RBD IgG, was r elat ed to a higher IgG conce ntr ation , wheth er or no t there was a r epor ted infec tion (data n ot shown). Without a repo rte d case, this may be assumed t o reflec t an a symptomatic (or unrecogniz ed) infection . In those wit h a recogniz ed infection, this may again r eflect a furt her unrecognized infec tion, o r perh aps different r at es of decay of anti- RBD IgG and nucleoca psid post infecti on [27]. The greate r respons e in thos e with at leas t one dos e of mRNA- 1273 vaccine compared to th ose r eceivin g only BNT162b2 is consistent with th e li tera tur e [4] [12] [13 ] as is the lower IgG r esponse with n on-mRNA vaccines [7] [28] [29]. Our results show th a t HCW s with inflamma tory medical c ondi tions had lower levels of an ti-RBD IgG , as r ep or t ed elsewhere [30] [31]. Not ably thos e partic ipan ts with eith er multipl e scler osis or r h eumat oid a rthr itis had a signific an tly low er c once n tra ti on of an ti- RBD IgG. These findings ma y be r el at ed to the c on dition i tself af f ecting the immune response t o v accine o r t o th e e f f ects of medic al tr e atment f or the co ndition. When we e xamined the immune supp r essa n t medications th a t pa rticipa n ts r epo rted, t he low e s t immune r esp onse w as see n in those r eceiving one of th e select iv e immunosu ppr essa n ts (Suppleme nt a ry ma terials 1.3 ). It has bee n shown that pa ti ents with multiple scle r osis (MS) t aki ng these molecul es ha v e a dec r eas ed ant i-spik e pr otein IgG r espons e to mRNA v accines while untreat e d patie n ts with MS had a r esp onse c omp ar able t o h ealth y c o n trols [16]. Apost olidis et a l [32] similarly demons tr at ed r educe d anti-RBD IgG response t o m RNA v accines in a small gr oup of pa tie n ts with MS on anti-CD20 ther apy . A r ece n t regis t ry s tudy of patie n ts with MS show ed those treated with an ti-CD20 ther api es at the time of v accina tio n had “ Abnormally Low ” v alues [33]. In the cur r e n t study , participa n ts t aking tumor n ecr osis f act o r-alpha inhibi t or s (TNFi) had signific antly low er anti-RBD IgG as did t he thr e e par ticipa n ts prescribed c alcineu rin inhibit o r s. K ashiw ado e t al [17] r epor ted that patie n ts with au t o- immune rheumatoid diseas e t aking TNFi had low er a n ti-RBD IgG than h ealth y c o ntr ols, with similar e f f ects in pa ti ents t aking TNFi f or inflammat ory bo w el disease [34]. Only f our par ticipa n ts repor t ed t aking antine oplastic ag ents at th e time of th eir fir s t v accine d ose. Their lower a n ti-RBD IgG r espons e w as c onsist e n t with th at pr eviously r epo rted [35] but number s w e r e t oo low f or analy sis by ag ent. Levy et al [20] r eported on 831 HCW s f or whom they c ollected adv e r se re actions t o BNT162b2 mRNA vaccinati on through a qu estionn aire se nt seven days after e ach v accine dose. They e xamined t he r el a ti on of r epo r t ed adv er se r eac tions t o anti-RBD IgG in subs eque n t blood sampl es. As in t he current s tudy , they f ound that higher ov er all sc ores of s y s temic r eac tions to th e sec ond v accine r el a ted t o higher level o f an ti-RBD IgG. T wo smaller s tudies of HCW fr om Japa n [36] and Sout h K or ea [37] f ound no clear r el ation between a n ti-spik e pr o tein IgG an d r epo rted adv e r se ev e n ts, while tw o fr om the Uni t ed S ta tes [38] and Gr e ece [9] ag ain f ound those with c ommon adv er se r eac tions had greater anti-RBD I gG lev els. In the cur r e n t study the rel a tio n betw e en adv er s e r e actions aft er th e second dose and anti-RBD IgG p er sist ed after adjustment f or man y of the f act or s (including ag e, g e nder , v accine type) that migh t act as c onf ounder s. Levy and c olle agues [20] c oncluded t ha t the mechanism of this associa ti on between immunological r esp onse and adv e r se e f f ec ts w as y et to be u nder st ood and th a t i t w as unclear wheth er th ere is c ausali ty in eith er direction : does a str onger immunologi c al r esp onse pr o duce more inflamma t o ry mediat o r s and pr oduc ts of inflamma tion th a t spill fr om th e injection sit e t o th e g en er al cir cul a tio n, r esulti ng in s y s t emic s ympt oms [39] or does a s y s temic r espons e trig g e r a mor e mark ed immunologic al on e. Such ques tio ns ha v e be en r aised a lso f or pertussis [40] and adjuv an ts f or th e hum an papilloma virus [41]. The mark ed similarity of r esul ts between the Levy r eport and the cur r e n t study sug g ests that th ere is indeed a phenomeno n t o b e explaine d. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 10 In the p r ese n t study , in which the majorit y of participants w ere f emal e, pregnanci es during the pa ndemic w e r e r epo rted by man y HCW s. While some chose t o d ela y v accina ti on because of thei r pr egnancy , suf ficien t h ad r eceived at least one v accine dos e during pr egnancy t o allow us t o explo r e ea rlier r epo rts of low er a n ti-spik e pr otein IgG [19] or a n ti-RBD IgG [42] aft e r a v accina tion d ose receiv ed du ring pr eg nancy than in non-pregnant r e f erents. G r a y et al [43], look ed at th e same question and r epo rted no e f f ec t alth ough, gr aphic ally , anti-RBD IgG appea r ed lower in pregnant w omen post -v accina tion. A tyeo et al [44] c ompa r ed an ti-spik e protein Ig G in 158 w omen r ec eiving v accines during pr egna ncy (11 fir s t, 88 sec ond and 52 t hir d t rim es ter) and f ound the high es t (log) median IgG in the thi r d trim est er a n d low es t in th e fir s t trimester: t his w as r e ported t o be ‘n on-signific ant ’ without furth er de t ail . The analy sis repor t ed h ere c ompa r ed v accine response in p r egna n t w omen t o that of all health c are w ork er s. Al though th e inf orma tio n on g esta tio n will ha v e bee n less e x a ct than in th e pr evious studies fr om obs t et ric cent e r s, th e s t r eng th of t h e s ta ti s tical modelling giv es cr ed ence t o t he obser v a ti on th a t r educt ion in an tib ody pr oducti on (an ti-RBD IgG) f ollowing v accina tion w as limited to early an d mid-pr egnancy . Ther e is limited evidence that obesi ty is a risk f act or f or a low IgG f ollowing v accin a tio n. W a tanab e et al [18] f ound tha t high er BM I in a sample of 86 HCW s r elat ed t o lower a n ti-spik e protein IgG in biv ari a te but no t multiv ariate models. W ais t cir cumf e r enc e w as mor e str ongly r el at ed than BM I. H erzbe r g et al [45] r epo rted th a t those with a B MI >25 had a lower a n ti-sp ik e pr otein Ig G r espo nse th an oth er s in t heir G erman HCW c oh ort . In a Gr e ek c oho rt of HCW s, P apaioannidou e t al [11] f ound, as in the pr ese n t study , no clear evidence th at BMI w as a risk f act or f or a low an ti-RBD IgG r espons e. The inter e s ting obser v ation t hat neu tralising an tib odies w e r e less e f f ec tiv e and d eclined f ast e r in those clas sified as sev er ely obes e [46], with no dif f er enc e fr om c ontr ols in a n ti- RBD IgG [47], c annot b e in v estig at ed fur t her he r e. A s y s tematic r eview [48] of the r el a tion be tw e en humor al immunity and obesity repor t e d a r ed ucti on in an tib odies in pe ople with ob esity b ut did not dif f erentiate betw e en the types of antibodies me asured. The s tudy r ep or t ed h er e us ed seri al bloo d samples, spaced ov er the fir st 18 mon t hs aft er th e intr oduc tion of v accines ag ains t SAR S-CoV-2 in December 2020 and the end of the s t udy in July 2022. B y analy sing th ese in a longitudinal mod el w e ha ve minimised t he v ariance du e to unc o n trolled dif f eren ces betw e en par ticipants. The r esulti ng model c onfirms the now expec ted incr eas e in anti-RBD IgG with successi v e v accina tion and inf ectio ns but ide n tifies a f ast er d ec a y aft e r the third and f ourth doses th an aft e r th e sec ond . Ha ving es t a blished th is model, w e used i t to determin e whether there w e r e o ther f ac t or s influencing this r espons e. The results demons trat e the impor t anc e of r ecrui tin g and r et ai ning c ohor ts with a rich body of obser v a tional d at a th r ough which t o bett e r unde r s tand response t o v accina tion an d t o h elp ide n tify areas in which ques tions still need t o b e answ er e d. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 11 Criteria for authorship All autho rs at test they mee t the ICMJE cr iteria for autho rship. Conflicts The autho r s declare no c onflict of int e r ests. Funding The Colleg e of Ph y sicians and Sur geons o f Alberta gave s eed funding f or the est abl ishmen t of th e HCW c oho rt. Gr ant funding w as obt ain ed fr om th e Canadian Institu tes of Health Resea r ch (Funding R e f e r enc e number 173209). This funding w as e xt end ed by a gr an t fr om the Can adian Immunol ogy T ask F or ce. Ethics approval Approval for each elemen t of the st udy was given by the University of Alber ta He ath Ethics Boa rd (Pr000099700). The study was also reviewed and approved by Uni ty Health Tor o nto Rese arch Ethics Bo ard (REB# 20-298) for those elements coo rdi nated loc ally for the On tari o par ticipants . All par ticipants gave onlin e written info rmed consen t afte r the n atur e and possible conse quences of th e stud y had been fully explain ed. The resea rch was carried ou t in accorda nce with the Declar ation of Helsinki . Author co ntributions AA, IB , QDM, FL, SR and NC devised th e p roject to establish the he althca re worke r cohort and all ac tively collabora ted to ensur e th e re tent ion of participan ts and th e seri al collectio n of ser ology blood samples. LAT and CC took responsibility for the analysis of samples and resolu tion of dispari ties . TZ and YC curated th e longitudin al datase t. YY, YC and NC carried ou t th e da ta analysis. All a utho rs contribu ted to dr afting the pap er, cri tically reviewed th e final draft and approved th e version submitt ed. Data availability s tatement The data on which this manuscrip t is bas ed will be made availabl e on reas onable reques t to t he corr esponding author . The da ta ar e subject to a data sh aring agreem ent with the major funde r, t he Canadian Immunol ogy Task Force (CITF).. Through the d atab ank curr ently being built individu al anonymized d ata will be made avail able t o resea rchers in tern ation ally on requ est to the CITF

The analysis of serum samples was carrie d out by the Alb er ta Precision Labor ato ri es and we are gra teful t o the Medical Direct or, G raham Tipples a nd to Paul Jacquier , Director , for th eir conside r able help in ensu ring the successful completion of the logistical ly complex analy tical tasks. The d ata from p ublic health r ecords suppli ed through th e Alb ert a and Bri tish Columbia health se rvices were invaluabl e and mad e possible th e sensitivity analysis. The inte rpr eta tion and conclusi ons containe d her ein are those of th e res earche rs and do no t necessarily rep rese nt th e views of the G overnments of Alb ert a or Bri tish Columbi a. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 12 Referenc es 1) Polack FP, Thomas SJ, Kitchin N, Absalon J, Gur tman A , Lockhart S, Per ez JL, Pére z Marc G, M orei ra ED, Zerbini C, Bailey R, Swanson KA, Roychou dhury S, Koury K, Li P, Kalina WV, Cooper D, Frenck RW Jr, Hammit t LL, Türeci Ö, Nell H, Scha efer A, Ü nal S, T resnan DB, Ma the r S, Dormitze r PR, Şahi n U, Jansen KU, Grub er WC; C4591001 Clinical Trial Group. Safe t y and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N Engl J Med. 2020 Dec 31;383(27):2603-2615. doi: 10.1056/NEJMo a2034577. 2) Baden LR, El Sahly HM, Essink B, Kotloff K, Frey S, Novak R, Diemer t D, Specto r SA, Rouphael N, Creech CB, McGet tigan J , Kheta n S, Segall N , Solis J, Brosz A, Fi erro C, Schwar tz H, Ne uzil K, Corey L, Gilbe rt P, Jan es H, Follmann D, Marovich M, Mascol a J, Pola kowski L, Ledgerwood J, Grah am BS, Ben nett H , Pajon R, Knightly C, Leav B, Deng W, Zhou H, Han S, Ivarss on M, Miller J, Zaks T; COVE Study Group . Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine. N Engl J Med. 2021 Fe b 4;384(5):403-416. doi: 10.1056/NEJMoa2035389 . 3) Thomas SJ, Mor eira ED Jr, Kitchin N, Abs alon J, Gurtma n A, Lockhart S , Perez JL, Pérez M arc G, Polack FP, Zerbini C, Bailey R, Swanson KA, Xu X, Ro ychoudhury S, Koury K, Bouguermouh S, Kalina WV, Cooper D, Frenck RW Jr , Hammitt LL, Türeci Ö , Nell H, Schaefer A , Ünal S, Yang Q, Liber ato r P, Tresnan DB, Ma ther S , Dormitzer PR, Şahin U , Gru ber WC, Janse n KU; C4591001 Clinical Trial Group. Safe ty and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine throu gh 6 Months. N Engl J Me d. 2021 Nov 4;3 85(19):1761-1773. doi: 10.1056/NEJMoa2110345 . 4) Dickerman BA, G erlovin H, Mad enci AL, Kurgansky KE, Ferolito BR , Figueroa Muñiz MJ, Gagno n DR, Gazia no JM, Cho K, Casas JP, Hernán M A. Compar ative Effectiveness of BNT162b2 and mR NA-1273 Vaccines in U.S. Veter ans. N Engl J Me d. 2022;386(2):105-115. doi:10.1056/NEJM oa2115463 5) Chemaitelly H, Tang P, Hasan MR, Al Mukdad S, Yassine HM, B enslimane FM , Al Khatib HA, Coyle P, Ayoub HH, Al Kanaani Z, Al Kuwari E, J eremijenk o A, Kaleeckal AH , Latif AN, Sh aik RM, Ab dul Rahim HF, Nasr allah GK, Al Kuwari MG , Al Romaihi HE, Bu tt A A, Al-Thani MH, Al Khal A , Ber tollini R , Ab u-Raddad LJ. W aning of BNT162b2 Vaccine Protecti on against S A RS-CoV-2 Infection in Qatar . N Engl J Me d. 2021 Dec 9;385(24):e83. doi:10.1056/NEJ Moa2114114. 6) Tartof SY, Slezak J M, Fischer H, H ong V, Ackerson BK, Ranasinghe ON, F rankland T B, Ogun OA , Zamparo J M, Gray S. Effectiveness of mRNA B NT162b2 COVID-19 vac cine up to 6 months in a la rge integr ate d heal th system in the USA : A re trosp ective coho r t study. Lance t. 2021 Oct 16;398(10309):1407–1416. doi:10.1016/S0140-6736(21)02183-8. 7) Fedele G, Tren tini F, Schiavoni I , Abrigna ni S, Anton elli G, B aldo V, Baldovin T, Ban dera A, Bonur a F, Clerici P, De Paschale M, Fortu nat o F, Go ri A, G rifa ntini R, Ica rdi G, Lazz aro tto T, Lodi V, Ma stroianni CM, Orsi A, Pra to R, Restivo V, Carset ti R, Piano Mo rt ari E, Leone P, Olivet ta E, Fior e S, Di Martin o A, Brusaferro S , Merl er S, Palamara AT, St efanelli P. Evaluati on of humoral and cellul ar resp onse to fou r vaccines against COVID-19 in different age gro ups: A longit udinal stud y. Front Immunol . 2022 Oct 31;13:10213 96. doi:10.3389/fimmu.2022.1021396. 8) Lustig Y, Sapir E, Regev-Yochay G, Cohen C, Fluss R, Olmer L, Indenb aum V, Mande lboim M, Doolman R, Amit S, Mend elson E, Ziv A, Hupper t A, R ubin C, Freedman L, Kreiss Y. B NT162b2 COVID-19 vaccine and correla tes of humoral immun e respons es and dynamics: a prospective , single-cent re, longitu dinal cohor t study in health-car e workers . Lancet R es pir Med. 2021 Sep ;9(9):999-1009. doi:10. 1016/S2213- 2600(21)00220 -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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 13 9) Hatzakis A, Kar abinis A, Ro ussos S, Panta zis N, Degiannis D, Chaidaroglou A, Pe tsi os K, Pavlopoulou I, Tsiodras S, Paraskevis D, Sypsa V, Psicho giou M. Modelli ng SARS-CoV-2 Binding Antibody Waning 8 Mon ths after B NT162b2 Vaccination. Vaccines (B asel). 2022 Feb 13;10(2):285. doi: 10.339 0/vaccines10020285. 10) Ikezaki H, Nomu ra H, Shimono N. Dynamics of anti-spike IgG antib ody after a thir d BNT162b2 COVID-19 vaccination in Jap anese h ealt h care work ers. Heliyon . 2022 Dec;8(12):e12125. doi : 10.1016/j.heliyon.2022. e12125. 11) Papaioannidou P, Skoumpa K, Bos tanitis C, Michailidou M, St ergiopoul ou T, Bosta nitis I, Tsalidou M . Age, Sex and B MI Rel ations with Anti-SAR S-CoV-2-Spike IgG Antibodi es after B NT162b2 COVID-19 Vaccine in Health Car e Work ers in Nor the rn G reec e . Microorga nisms. 2023 May 13;11(5):1279. doi: 10.3390/microorganisms11051279. 12) Dodge MC, Ye L, Duf fy ER, Cole M, Gawel SH, Werl er MM, Daghfal D, Andry C, Kat aria Y. Kinetics of SARS- CoV-2 Serum Antibodies Through the Alp ha, Delta, and Omicron Su rges Among Vaccinated He alth Car e Workers a t a Bos ton Hospit al. O pen For u m Infect Dis. 2023 May 17;10(7):ofad266 . doi:10.1093/ofid/ofad266. 13) Canetti M , Bard a N, Gi lboa M, Indenb au m V, Mandelboim M, G onen T, Asr af K, Weiss-Ot tolenghi Y, Ami t S, Doolman R, Mend elson E, Hara ts D, Free dman LS, Kreiss Y, Lustig Y, Regev-Yochay G. Immunogenici ty and efficacy of fourth BNT162b2 and mRN A1 273 COVID-19 vaccine doses; thr ee mon t hs follow-up. Nat Commun. 2022 Dec 13;13(1 ):7711. doi:1 0.1038/s41467-022-35480 -2. 14) Lee ARYB, Wo ng SY, Chai LYA, Lee SC, Lee MX, Muthia h MD, Tay SH, Teo CB, Tan BKJ, Chan YH, Sundar R, Soon YY. Efficacy of covid-19 vaccines in i mmunocompromised pati ents : systemati c review and meta- analysis. BMJ . 2022 Mar 2;376:e068632. doi: 10.1136/bmj-2021-068632 . 15) Al-Haideri M , Mohammad TAM , Darvishzadehdel dari S, Karb asi Z, Alimohammadi M, Farama rzi F, Khoras ani S, Rasouli A, Tahmas ebi S, Darvishi M, Ak havan-Sigari R. Immunogenici ty of COVID-19 vaccines in adult patien ts with aut oimmune inflammato ry rheumatic dise ases: A system atic review and meta-analysis. Int J Rheum Dis. 2023 Jul;26(7):1227-1234. doi: 10.1111/1756-185X.14713. 16) Achiron A, M andel M , Dreyer-Als ter S, H arari G, Magalashvili D, Sonis P, Dolev M, Menascu S, Flech ter S, Falb R, Gur evich M. Humora l immune res ponse to COVID-19 mRNA vaccine in pati ents with multipl e sclerosis tre at ed with high-efficacy disease-modifying therapies . Ther Adv Ne urol Disord. 2021 Apr 22;14:17562864211012835. doi: 10.1177 /17562864211012835. 17) Kashiwado Y, Kimoto Y, Ohshima S, Sawa be T, Irino K, N akano S, Hiur a J, Yonekaw a A, Wang Q, Doi G, Ayano M, Mitoma H, Ono N , Arinob u Y, Niiro H, Hotta T, Kang D, Shimono N , Akashi K, Takeuchi T, Horiuchi T. Immunosuppressive t her apy and humor al response to thi rd mRNA COVID-19 vaccination with a six-mon th interval in rh eumatic dise ase pa tien ts. R heumatol ogy (Oxford). 2023 Jun 8:kead2 75. doi: 10.1093/rheumat ology/kead275. 18) Watan abe M , Balen a A, Tuccinardi D, Toz zi R, Risi R, Masi D, Caputi A, Ross et ti R, S poltor e ME, Filippi V, Gangita no E, Manfrini S, M ariani S, Lubr a no C, Lenzi A, Mast roianni C, Gn essi L. Central ob esity, smoking habit, and hyp ert ension ar e associa ted w ith lower an tibody ti tres in r esponse to COVID-19 mRNA vaccine. Diabetes M etab R es Rev. 2022 Jan ;38(1):e3465. doi:10 .1002/dmrr.3465. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 14 19) Blakeway H, Amin-Chowdhury Z, Prasad S, Kalafat E, Ismail M, A bdallah F N, Rezva ni A, Amirth alingam G, Brown K, Le Doare K, Hea th PT, Ladhani SN, Khalil A. Evalua tion of immunogenicity and react ogenicity of COVID-19 vaccines in pregnant women. Ultrasou nd Obst et Gynecol. 2022 Nov;60 (5):673-680. doi: 10.1002/uog.26050. 20) Levy I, Levin EG, Olmer L, Regev-Yochay G, Agmon-Levin N, Wie der-Fin esod A, I n denbaum V, Herz og K, Doolman R, Asraf K, Halpe rin R, Lustig Y, Rahav G. Corr ela tion be tween A dverse E vents and Ant ibody Titers among Healthca re W orkers Vaccinat ed with BNT162b2 mRNA COVID-19 Vaccine. Vaccines (Basel). 2022 Jul 30;10(8):1220. doi:10.3390/vaccines10081220. 21) Sett e A, Cro tty S. Immunol ogical memor y to SARS-CoV-2 infection and COVID-19 vaccines. Immunol Rev. 2022 Sep;310(1):27-46. doi: 10.1111/imr.13089. 22) Cherry N, Adisesh A, Burs tyn I, Durand-M oreau Q, Gala rnea u J-M, Labr èche F, Ruz ycki S, Zadunayski T. Cohort profile : Recrui tmen t and re ten tio n in a prospective coho rt of Canadian h e alth care work ers during the COVID-19 pandemic. Und er review b y BMJ Open an d available as a p repri nt a t https://doi. org/10.1101/2023.04.14.232 88575 Accessed 26 April 2023 . 23) World He alth O rganiza tion (2020). Inte rn ational Classification of Diseases 11 th Re vision. Available on: ht tps://icd.who.int/e n (accessed on 11 August 2023). 24) Archambaul t PM, Rosychuk RJ, Aud et M, Bola R, Vata npour S, B rooks SC, Daoust R , Clark G, Gr ant L, Vaillancourt S, W elsford M, M orrison L J, Hohl CM, Canadian COVID-19 Emergen cy Department Ra pid Response N etwo rk (CCE DRRN) investigators, Ne twork of Canadian Emergency Res earche rs, Canadian Critical Care Trials G roup. Accuracy of Self-Reporte d COVID-19 Vaccination Status Compared Wit h a Public Health Vaccina tion Regist ry in Québec : O bservation al Diagnostic Study. JMIR Publi c Health Surveill . 2023;9:e44465. doi :10.2196/44465. 25) Steph enson M, Ols on SM, Self WH , Gind e AA, Mo hr NM , Gaglani M , Shapir o NI, G ibbs KW, Hager DN, Prekker ME, Gong M N, St eingrub JS , Pelt an ID, Mar tin ET, Reddy R, Busse LW, Du ggal A, Wilson J G, Qadir N, Mallow C, Kwon JH, Exline MC, Chappell JD, Lauring AS, Baughman A , Lindsell CJ, Hart KW, Lewis NM , Patel MM, Tenforde MW, N etwork I nvestigat o rs IV. Ascer tainmen t of vaccination sta tu s by self-report versus source documen tati on: impact o n measu ring COVID-19 vaccine effectiveness. Influenza O the r Respir Viruses . 2022 Nov;16(6):1101–1111. doi: 10.1111/irv.13023. 26) Israel A , Shenh ar Y, Gr een I , Mer zon E, G olan-Cohen A, Schäffer A A, Ruppin E, Vin ker S, Magen E. Larg e- Scale Study of Antib ody Titer Decay follo wing BNT162b2 mRNA Vaccine or SARS- CoV-2 Infection. Vaccines (Basel). 2021 Dec 31;10(1):64. doi: 10.3390/vaccines10010064. 27) Navara tnam AMD, Shr otri M , Nguyen V, Braithwai te I , Beal e S, Byrne TE, Fong WL E, Fragaszy E, Geismar C, Hoskins S, Kovar J, Patel P, Yavlinsky A, Aryee A, Rodg er A, H ayward AC, Aldridge RW; Virus Wa tch Collaborative . Nucl eocapsid and spike an tibody respons es following virologically confirmed SARS-CoV-2 infection: an obs ervati onal an alysis in the Virus Watch community cohor t. I nt J In fect Dis. 2022 Oct;123:104- 111. doi:10.1016/j.ijid.2022.07 .053. 28) Steens els D, Pierlet N, Pende rs J, Mes ott en D, Heylen L. Comparison of SARS-CoV-2 Antibody Respo nse Following Vaccination Wit h BNT162b2 a nd mRNA-1273. JA MA. 2021 Oc t 19;326(15):1533-1535. doi:10.1001/jama.2021.15125. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 15 29) Wei J, Pouwels KB, St oesse r N, Ma tth ews PC, Diamond I, Studley R, Rourke E, Coo k D, Bell JI, N ewton J N, Farra r J, Howar th A, M arsde n BD, Hoosd ally S, Jones EY, Stua rt DI, Crook DW, Pet o TEA, Walker AS , Eyre DW; COVID-19 Infection Survey team . An tibody respons es and corr ela tes of prot e ction in the gen eral populati on afte r two doses of th e ChAdO x1 or BNT162b2 vaccines. N at M ed. 2022 May;28(5):1072-1082. doi: 10.1038/s41591-022-01721-6. 30) Marra AR, Kobayashi T, Suzuki H, Alsuh ai bani M, Tofanet o BM, B ariani LM, A uler MA, Salinas JL, Edmond MB, Doll M, Kutne r JM, Pinho JRR, Rizz o LV, Miraglia JL, Schweizer ML. Sh ort- ter m effectiveness of COVID- 19 vaccines in immunocompromised pati ents: A syst ematic lit era ture review and meta-analysis. J Infect . 2022 Mar;84(3):297-310. doi:10.1016/j.jinf.2021.12.035. 31) Embi PJ, Levy ME, Naleway AL, Patel P, Gaglani M, Na tar ajan K, Dascomb K, Ong T C, Klein NP, Liao IC, Grannis S J, Han J , Ste nehjem E, Dunne M M, Lewis N, Irving SA, Rao S , McEvoy C, B ozio CH, Murthy K, Dixon BE, Grisel N, Yang DH, Godda rd K, Kharba nda AB, R eynolds S, Raiyani C, Fadel W F, Arndorfe r J, Rowl ey EA, Fireman B, F erdinan ds J, Valvi NR, Ball S W, Zerbo O , Griggs EP, Mitchell PK, Porte r RM, Kiduko SA, Blan ton L, Zhuang Y, Steffens A, Rees e SE, Olson N , Williams J, Dickerson M, McM orrow M, S chrag SJ, Verani JR, Fry AM, Azziz-B aumgartn er E, Ba rron M A, Th ompson MG, DeSilva MB . Effectiveness o f 2-Dose Vaccination with mRNA COVID-19 Vaccines Against COVID-19-Associated Hospit aliza tions Among I mmunocompromised Adults - Nine S tat es, Ja nuary-Sep tembe r 2021. MMWR Mo rb Mor tal Wkly Rep . 2021 Nov 5;70(44):1553- 1559. doi: 10.15585/mmwr.mm7044e3. 32) Apostolidis S A, Kakara M , Painter MM, G oel RR, Ma thew D, Lenzi K, Rezk A, Pat te r son KR, Espinoza DA, Kadri JC, Markowitz DM, E Ma rkowitz C, Mexhi taj I, J acobs D, Babb A, B et ts MR, Prak ETL, Weiskopf D, Grifoni A, Lundgr een KA, Gouma S, Se tt e A, Bat es P, Hensley SE, Gr eenpl ate AR, W herry EJ, Li R, Ba r-Or A . Cellular and humor al immune respo nses following SARS-CoV-2 mRNA vaccination in patien ts with multipl e sclerosis on anti-CD20 thera py. Na t Med . 2021 Nov;27(11):1990-2001. doi:10.1038/s41591-021-015 07-2. 33) Klineova S, Farbe r RS, DeAng elis T, Leung T, Smith T, Blanck R, Zhovtis-Ryerson L, Harel A . Vaccine- breakth rough SARS-CoV-2 infections in p eople with multi ple sclerosis an d rela ted conditions: An observati onal study by the New York COVID-19 Neuro-Immunology Consortium (NYCNIC-2). Mult Scler. 2023 Jul;29(8):990-1000. doi:10.1177/13524585231185246. 34) Martins Figuei red o L, Carvalho E Branco J , Carvalho Lourenço L, San tos L, Oliveira AM. An ti-tumor n ecrosis factor th erapy is associat ed with at tenu a ted humor al respons e to SA RS-COV-2 vaccines in patien ts with inflammatory bowel diseas e. Vaccine. 20 23 Jun 13;41(26):3862-3871. doi:10.1016/j.vaccine.2023.05.012. 35) Ruggeri EM, Nelli F , Fabbri A , Ono rato A, Gianna relli D, Gir on Ber rios JR , Virtuoso A, Mar rucci E, Mazzo tta M, Schirripa M , Panichi V, Pessina G, Signorelli C, Chilelli MG, Primi F, N atoni F , Faz io S, Silvestri MA . Antine oplastic tre atmen t class modulat e s COVID-19 mRNA-BNT162b2 vaccine im munogenicity in cancer patien ts: a secon dary analysis of the pr o spective Vax-On stu dy. ESMO Op en. 202 2 Feb;7(1):100350. doi:10.1016/j.esmoop .2021.100350. 36) Takeuchi M, Higa Y, Esaki A, Nabeshim a Y, Nakazono A . Does re actogenici ty after a second injection of th e BNT162b2 vaccine predict spike IgG an ti body levels in healthy J apanes e subjects ? PLoS One. 2021 Sep 20;16(9):e0257668. doi:10.1371/journal . pone.0257668. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 16 37) Hwang YH, Song KH, Choi Y, Go S, Choi SJ, Jung J, Kang CK, Choe PG, Kim NJ, Park WB, Oh MD. Can reactog enicity pred ict immunogenicity af ter COVID-19 vaccination? Kor ean J Int er n Med. 2021 Nov;36(6):1486-1491. doi:10.3904/kjim.2021.210. 38) Oyebanji OA, Wilson B, Ker eszt esy D, Carias L, Wilk D, Payne M, Aung H, Denis KS, Lam EC, Rowley CF, Berry SD, Cameron CM, Cameron MJ , Schmade r KE, Balazs AB, King CL, Canaday DH, Gra venstein S. Does a lack of vaccine side effects correl ate wit h reduc ed BNT162b2 mRNA vaccine r esponse a mong healthcar e workers and nursing home resid ents ? Aging Clin Exp Res. 2021 Nov;33(11):3151-3160. doi: 10.1007/s40520-021- 01987-9. 39) Hervé C, Laupèze B , Del Giudice G, Didier lauren t AM, Tavar es Da Silva F. The how' s and what's of vaccine reactog enicity. NPJ Vaccines . 2019 Sep 2 4;4:39. doi :10.1038/s41541-019-0132-6. 40) Mitchell TC, Casella CR. No pai n no gain? Adjuvant effects of alum and monophos phoryl lipid A in per tussis and HPV vaccines. Curr Opin Immunol. 2 017 Aug;47:17-25. doi:10.1016/j.coi.201 7.06.009. 41) Burny W, Callegaro A, Bech told V, Cleme nt F, Delhaye S, Fisse tt e L, Janssens M, Le roux-Ro els G, Ma rchan t A, van den Berg RA, Garço n N, van der Mos t R, Didierlau ren t AM; ECR-002 Study Gr oup. Different A djuvants Induce Common Inna te Pathways That A re Associat ed with Enhanc ed Adap tive R esponses agains t a Model Antigen in Humans . Fron t Immunol. 201 7 Aug 14;8:943. doi:10 .3389/fimmu.2017.00943. 42) Collier AY, McMah an K, Yu J, Tostanoski LH, Aguayo R, Ansel J , Chandrashek ar A, Patel S, Apr aku Bondzi e E, Sellers D, Bar re tt J , Sanbor n O, W an H, Chang A, Anioke T, Nkol ola J, Bradshaw C, Jacob-Dolan C, Feldman J , Gebr e M, Bo rducchi EN, Liu J, Schmidt A G, Suscovich T, Linde C, Alter G, Hacker MR, Bar ouch DH. Immunogenicity of COVID-19 mRNA Vaccines in Pregnant and Lac tating W omen. J AMA. 2021 Ju n 15;325(23):2370-2380. doi:10.1001/jam a.2021.7563. 43) Gray KJ, Bo rdt EA, Atyeo C, Deriso E, Akin wunmi B, Young N, Baez A M, Shook LL, Cvrk D, James K, De Guzman R, B rigida S, Diouf K, Goldfarb I , Bebell LM, Yonke r LM, Fasano A , Rabi SA, Elovitz MA, Alt er G , Edlow AG. Coron avirus disease 2019 vaccine respons e in pregnan t and lact ating women: a cohor t study. Am J Obst et Gyn ecol. 2021 Sep ;225(3):303.e 1-303.e17. doi:10 .1016/j.ajog.2021.03.0 23. 44) Atyeo CG, Sho ok LL, Brigida S, De Guzma n RM, Demidkin S, Muir C, Akinwunmi B, Baez AM , Shee han ML, McSweeney E, Burns MD, N ayak R, Kumar MK, Patel CD, Fialkowski A, Cvrk D, Gol dfarb IT, Yonker LM, Fasano A, B alazs AB , Elovitz MA, Gray KJ, Alte r G, Edlow A G. Ma ter nal immune res ponse and place ntal antibody t ransfer aft er COVID-19 vaccination across t rimest er and pla tforms. N at Commun. 2022 Jun 28;13(1):3571. doi:10.1038/s41467-022- 31169-8. 45) Herzbe rg J, Fischer B, Lindenkamp C, Bec her H, Becke r AK, Hona rpisheh H, Guraya SY, Strat e T, Knabbe C. Persistence of Immune R esponse in H eal th Care Wo rkers Aft er Two Doses BNT16 2b2 in a Longitudinal Observat ional St udy. Fron t Immunol. 20 22 Mar 4;13:839922. doi :10.3389/fimmu.2022.839922. 46) Malavazos AE, Basilico S, Iacobe llis G, Mi lani V, Cardani R, Bonia rdi F, Dubini C, Prandoni I, Capi tanio G , Renna LV, Boveri S, Rigolini R , Carrar a M, Spuria G, Cuppon e T, D'acquisto A, Carpi nelli L, Sacchi M, Morricon e L, Secchi F, Costa E, Menica nti L, Nisoli E, Carruba M, Amb rogi F, Corsi Romanelli MM . An tibody responses to BNT162b2 mRN A vaccine: I nfection-naïve individuals with abd ominal obesity warr ant att entio n. Ob esity (Silver Spring). 2022 Mar;30(3):606-613. doi:10.1002/oby.233 53. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 17 47) van der Klaauw AA, Ho rner EC, Pereyr a-Gerb er P, Agrawal U, F oste r WS, Spe ncer S, Vergese B, Smith M , Henning E, Ramsay ID, Smith JA , Guillau me SM, Sharp e HJ, Hay I M, Thompson S, Innocen tin S, Boo th LH, Rober tson C, McCowan C, Kerr S, Mulron ey TE, O'Reilly MJ , Gurugam a TP, Guruga ma LP, Rust MA, Ferr eira A, Ebrahimi S, Ceron- Guti err ez L, Scotucc i J, Kronst einer B , Dunachie SJ, Kle nerma n P; PITCH Consortium; Park AJ, Rubino F, Lamikan ra AA , Stark H, Kingston N, Estcour t L, Harvala H, Rob er ts DJ, Doffinger R, Linterman M A, Ma theson NJ, Sh eikh A, F arooqi IS , Thaventhi ran JED. Accel era ted waning of the humoral response to COVID-19 vaccines in obesit y. Nat M ed. 2023 May;29(5):1146-1154. doi:10.1038/s41591-023- 02343-2. 48) Ou X, Jiang J, Lin B, Liu Q, Lin W, Chen G , Wen J . Antib ody respons es to COVID-19 vaccination in peopl e with obesity: A systema tic review and me ta-a nalysis. Influenza Oth er Respi r Viruses. 2 023 Jan;17(1):e13078. doi: 10.1111/irv.13078. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 18 Figure 1 Flow chart of recruitment to the serology sub-study . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 19 Figure 2 Anti-rec epto r binding domain (RBD) IgG t iter t o the spike p rot ein of the S ARS-CoV-2 virus by months from the 1 st vaccine The red line shows th e model pr edicte d anti-RBD IgG level for a p erson who is younger than 35 yea rs old and received 4 vaccinations a t th e resp ective cohort-aver age time p oints of the 4 vacci nations . The vertical jumps around 10 and 12 months indica te the 3 rd and 4 th vaccine boos ts of the IgG l evel, r espectively. The solid an d dott ed black lines show the LO WESS (LOcally Weighted Sca tt erplo t Smoothi ng) lines of observed log anti-R BD IgG levels for heal thcar e workers who ha d and had not , respec tively, receive d the 3 rd vaccination before t heir 10-month samples. No te t hat the LOWE SS lines do not t ake int o account t he lon gitudinal na ture of th e anti-R BD IgG levels within e ach heal thcare work er across time poin ts and just smooth the o bserved values of the respective gr oups. 2468 1 0 1 2 1 4 1 6 Months from the 1st Vaccine Anti-RBD IgG titer [AU/ml] 10 100 1,000 10,000 100,0000 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 20 Table 1 Geom etric mea n (GM) anti-RBD I gG and mean log anti-R BD IgG by person al charact eristics, infection and vaccinati on at t he tim e of sample collectio n. Factor Anti-RBD IgG (GM)AU/ml Log anti-RBD IgG Number of Mean SD Observat ions Participants Gend er Female 3249 8.09 1.56 7424 2292 Othe r 2977 8.00 1.54 1479 460 Age (years) < 35 3032 8.02 1.52 1291 439 35 < 50 3183 8.07 1.56 3472 1125 50 <65 3317 8.11 1.56 3365 1065 65 or older 3094 8.04 1.62 775 243 COVID-19 Confirmed 1 st case 12198 9.41 1.31 1127 678 Confirmed 2 nd case 16248 9.70 0.78 26 19 Nucleocapsid Positive 19583 9.88 1.11 668 544 Missing 2072 7.64 1.05 137 137 Vaccination Any 1 st vaccine 3202 8.07 1.56 8903 2752 A n y 2 nd vaccine 3246 8.09 1.55 8837 2737 A n y 3 rd vaccine 11175 9.32 1.10 3330 2222 A n y 4 th vaccine 9204 9.13 1.40 75 64 Type of vaccine Only Pfizer vaccine 2740 7.92 1.56 7085 2263 Any Modern a vaccine 6323 8.75 1.39 1575 604 Any non-mRNA vaccine 3653 8.20 1.34 243 81 Medical Conditio n Multiple scle rosis 279 5.63 2.84 43 12 Rheumat oid ar thritis 1904 7.55 1.57 145 42 Ankylosing spondylitis 3513 8.16 1.56 48 16 Psoriatic art hritis 3368 8.12 1.65 54 18 Psoriasis 3026 8.02 1.65 271 88 Inflammatory bowel disease 3249 8.09 1.63 243 77 Lupus 2060 7.63 2.02 49 16 Medicati on Metho tr exa te 1679 7.43 1.82 84 26 Tumor Necrosis Fac tor inhibitor 2098 7.65 1.67 61 21 Glucocor ticoids 2163 7.68 1.69 25 8 Inte rleukin inhibi tor 5056 8.53 1.20 23 7 Calcineurin inhibit or 702 6.55 2.45 11 3 Selective immunosuppressan ts 221 5.40 2.31 57 17 Antine oplastic ag ents 906 6.81 3.18 10 4 Medical/medicati on Data missing 3709 8.22 1.44 49 18 Body Mass Inde x <25 3169 8.06 1.53 3992 1243 25<30 3162 8.06 1.55 2832 851 30<35 3161 8.06 1.57 1310 405 35<40 4042 8.30 1.73 388 121 ≥40 3422 8.14 1.71 313 104 Missing 2510 7.83 1.49 68 28 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 21 Vaccine dose during pregnancy No 3172 8.06 1.56 8689 2696 Yes – early stage 4429 8.40 1.67 97 42 Yes – middle stage 3840 8.25 1.55 82 35 Yes – late st age 6156 8.73 1.32 88 35 Overall 3202 8.07 1.56 8903 2752 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 22 Table 2. Mea n time (months) betwe en e ach vaccine dose or case of COVID-19 prior to collec tion of each bloo d sample. Nominal month of sample (since first vac cination). 4-month sample 7-month sample 10-month sample 13-month sample Overall M e a n months SD N Mean months SD N Mean months SD N Mean months SD N Mean months SD N Vaccine 1 4.54 0.51 2216 7.87 0.53 2179 11.25 0.57 2258 14.51 0.68 2250 9.58 3.77 8903 2 2.90 0.87 2170 6.16 0.93 2170 9.54 1.02 2252 12.79 1.08 2245 7.91 3.82 8837 3 2.00 1.03 9 1.70 1.31 93 1.97 1.29 1146 4.17 1.66 2082 3.34 1.87 3330 4 /i1 /i1 /i1 1.89 /i1 1 2.25 1.26 13 2.57 1.89 61 2.50 1.78 75 Case of COVID-19 1 7.93 5.27 9 8.79 5.90 15 6.17 6.57 50 5.15 6.28 604 5.34 6.30 678 2 /i1 /i1 /i1 /i1 /i1 /i1 /i1 /i1 /i1 3.35 6.00 19 3.35 6.00 19 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 23 Table 3. Linear mix ed-effects model of lo g anti-RBD IgG with vaccination an d COVID-19 case history. (N=8903 samples from 2752 participan ts ) Factor Effect estimat e 95% CI P = Lower Upper Gend er: Female 0.003 -0.082 0.089 0.940 Age (years) <35 0.000 35<50 -0.054 -0.146 0.038 0.250 50<65 -0.074 -0.170 0.021 0.130 65 or great er -0.234 -0.366 -0.102 <0.001 COVID-19 Confirmed 1 st case of COVID-19 1.388 1.280 1.495 <0.001 Months first case b efore sampl e -0.049 -0.061 -0.037 <0.001 Confirmed 2 nd case of COVID-19 0.084 -0.380 0.547 0.720 Months second cas e befor e sample 0.005 -0.067 0.078 0.890 Nucleocapsid posi tive 0.687 0.589 0.786 <0.001 Nucleocapsid da ta missing -0.034 -0.190 0.122 0.670 Vaccination Any first vaccine 5.791 5.521 6.062 <0.001 Any second vaccine 2.342 2.087 2.597 <0.001 Months second vaccine b efore sample -0.156 -0.164 -0.148 <0.001 Any third vaccine 3.056 2.983 3.128 <0.001 Months thi rd vaccine befor e sample -0.105 -0.123 -0.087 <0.001 Any fourth vaccine 0.789 0.438 1.141 <0.001 Months four th vaccine before sampl e 0.012 -0.106 0.130 0.840 Type of vaccine Only BNT162b2 (Pfizer/BioNTech) 0.000 Any mRNA1273 (Modern a) 0.346 0.272 0.420 <0.001 Any non-mRNA vaccine -0.456 -0.632 -0.281 <0.001 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 24 Table 4. Linear mix ed-effects model of lo g anti-RBD IgG with medical condi tions a nd medication . (N=8903 samples from 2752 participan ts ) Factor Effect estimat e 95% CI P = Lower Upper Gend er: Female 0.031 -0.051 0.113 0.46 Age (years) <35 0.000 35<50 -0.068 -0.157 0.021 0.13 50<65 -0.074 -0.166 0.017 0.11 65 or great er -0.250 -0.377 -0.123 <0.001 Medical conditions Multiple scle rosis -1.619 -2.178 -1.060 <0.001 Rheumat oid ar thritis -0.383 -0.672 -0.093 0.010 Ankylosing spondylitis 0.388 -0.040 0.817 0.076 Psoriatic art hritis -0.347 -0.784 0.089 0.12 Psoriasis -0.102 -0.281 0.076 0.26 Inflammatory bowel dise ase 0.028 -0.164 0.220 0.78 Lupus 0.303 -0.172 0.779 0.21 Medicati on Metho tr exa te -0.025 -0.429 0.378 0.90 Tumor Necrosis Fac tor inhibi tors -0.712 -1.111 -0.313 <0.001 Glucocor ticoid -0.405 -1.021 0.212 0.20 Inte rleukin inhibi tors 0.419 --0.202 1.041 0.19 Calcineurin inhibit ors -1.149 -2.047 -0.251 0.012 Selective immunosupp ressan ts -1.679 -2.170 -1.187 <0.001 Anti-neop lastic agen ts -1.237 -2.199 -0.275 0.012 Data missing on medical conditions/med ication 0.131 -0.258 0.520 0.51 COVID-19 Confirmed 1 st case of COVID-19 1.374 1.267 1.480 <0.001 Months first case b efore sampl e -0.050 -0.062 -0.038 <0.001 Confirmed 2 nd case of COVID-19 0.067 -0.394 0.527 0.78 Months second cas e befor e sample 0.010 -0.061 0.080 0.79 Nucleocapsid posi tive 0.691 0.593 0.789 <0.001 Nucleocapsid da ta missing -0.038 -0.193 0.117 0.63 Vaccination Any first vaccine 5.836 5.570 6.102 <0.001 Any second vaccine 2.319 2.067 2.571 <0.001 Months second vaccine b efore sample -0.157 -0.165 -0.149 <0.001 Any third vaccine 3.053 2.980 3.125 <0.001 Months thi rd vaccine befor e sample -0.101 -0.120 -0.083 <0.001 Any fourth vaccine 0.838 0.487 1.189 <0.001 Months four th vaccine before sample 0.002 -0.115 0.119 0.98 Type of vaccine Only BNT162b2 (Pfizer/BioNTech)) 0.000 Any mRNA1273 (Modern a) 0.348 0.276 0.429 <0.001 Any non-mRNA vaccine -0.487 -0.656 -0.317 <0.001 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 25 Table 5. Side effects r epor ted for firs t an d second vaccine doses. Mean log visual analogue scor es with weights on th e first 3 principal components e xt ract ed. First dose Second dose Mean SD Component mat rix * Mean SD Component mat rix * Fatigue 1.87 1.63 0.691 -0.434 0.154 2.31 1.75 0.743 -0.408 0.061 Myalgia (muscle pain) 1.62 1.60 0.685 -0.335 0.080 1.84 1.73 0.777 -0.345 0.001 Arth ralgia (joint pain) 0.90 1.25 0.788 -0.143 -0.084 1.21 1.54 0.770 -0.166 -0.109 Headache 1.27 1.51 0.729 -0.273 0.023 1.70 1.70 0.742 -0.269 0.017 Malaise 1.19 1.48 0.796 -0.322 0.024 1.76 1.74 0.788 -0.381 -0.040 Feeling feverish 0.80 1.20 0.811 -0.089 -0.168 0.78 1.17 0.517 0.352 -0.224 Chills 0.77 1.19 0.802 -0.048 -0.179 1.28 1.60 0.753 -0.179 -0.105 Diarrhea/loos e stools 0.60 0.96 0.701 0.346 -0.219 0.66 1.02 0.650 0.379 -0.250 Nausea/vomiting 0.62 0.99 0.753 0.252 -0.227 0.77 1.16 0.680 0.247 -0.207 Skin reaction/r ash 0.55 0.87 0.646 0.499 -0.083 0.57 0.89 0.611 0.547 -0.045 Swollen glands 0.61 0.98 0.668 0.348 -0.104 0.84 1.25 0.618 0.276 -0.095 Pain at injection sit e 2.57 1.47 0.413 0.098 0.763 2.39 1.51 0.491 0.069 0.719 Redness a t injection sit e 1.22 1.42 0.495 0.366 0.569 1.05 1.35 0.530 0.374 0.542 N 3989 2868 *In th e componen t analyses th e varianc e accounted for by the firs t 3 componen ts were: Componen t 1 Component 2 Component 3 First vaccine dose 49.0 9.3 8.6 Second vaccine dose 45.5 10.9 7.7 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 26 Table 6. Rela tion of advers e effects to se cond vaccination, pr egnancy and body mass index (BMI) to an ti- RBD IgG adjusted for all fact ors* in Tabl e 4. Factor Effect estimat e 95% CI P = Lower Upper Adverse 2 nd dose effects Severity of adverse effects (continuous) 0.069 0.031 0.107 <0.001 Missing data on advers e effects -0.049 -0.119 0.021 0.170 Stage of pregnancy Any vaccine at pregnancy– early stag e -0.301 -0.554 -0.048 0.020 Any vaccine at pregnancy–middl e stage -0.299 -0.587 -0.011 0.042 Any vaccine at pregnancy–l ate s tage 0.371 0.084 0.659 0.011 Body Mass Index (BM I) <25.0 (ref) 0 Missing 0.053 -0.274 0.380 0.75 25.0<30.0 -0.012 -0.083 0.059 0.74 30.0<35.0 -0.018 -0.109 0.074 0.70 35.0<40.0 0.132 -0.019 0.284 0.087 40.0 or grea ter 0.121 -0.044 0.287 0.15 *Gen der , age, medical condi tions, med ication, cas es of COVID-19, vaccination. . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint 27 Table 7. Sensi tivity analysis. Linear mix ed effects models of log anti-RBD IgG for 2031 participan ts (6630 samples) for whom pu blic health da ta were availabl e. Self-repor t only Self-repor t plus public heal th Factor Effect estimat e 95% CI P= Effect estimat e 95% CI P= Lower Upper Lower Upper Gend er Female -0.029 -0.137 0.080 0.600 -0.035 -0.138 0.069 0.510 Age (years) <35 0 0 35<50 0.019 -0.091 0.129 0.730 0.024 -0.080 0.128 0.650 50<65 0.002 -0.111 0.115 0.970 -0.004 -0.112 0.103 0.940 65 or great er -0.126 -0.282 0.031 0.120 -0.215 -0.364 -0.066 0.005 COVID-19 Confirmed 1 st case of COVID-19 1.362 1.234 1.491 <0.001 1.389 1.269 1.509 <0.001 Months first case b efore sampl e -0.049 -0.063 -0.034 <0.001 -0.051 -0.065 -0.037 <0.001 Confirmed 2 nd case of COVID-19 -0.010 -0.710 0.691 0.980 -0.063 -0.610 0.484 0.820 Months second cas e befor e sample -0.060 -0.298 0.177 0.620 0.023 -0.050 0.096 0.540 Nucleocapsid posi tive 0.736 0.619 0.853 <0.001 0.685 0.576 0.794 <0.001 Nucleocapsid da ta missing 0.090 -0.090 0.271 0.330 0.033 -0.133 0.200 0.700 Vaccination Any first vaccine 6.209 5.890 6.529 <0.001 5.166 4.795 5.538 <0.001 Any second vaccine 1.745 1.446 2.045 <0.001 2.952 2.593 3.312 <0.001 Months second vaccine b efore sample -0.122 -0.132 -0.112 <0.001 -0.157 -0.166 -0.148 <0.001 Any third vaccine 2.883 2.793 2.972 <0.001 3.113 3.031 3.195 <0.001 Months thi rd vaccine befor e sample -0.120 -0.142 -0.098 <0.001 -0.108 -0.129 -0.088 <0.001 Any fourth vaccine 0.722 0.263 1.181 0.002 0.739 0.324 1.154 <0.001 Months four th vaccine before sample -0.043 -0.200 0.114 0.590 0.049 -0.104 0.202 0.530 Type of vaccine Only BNT162b2 (Pfizer) 0 0 Any mRNA1273 (Modern a) 0.308 0.221 0.395 <0.001 0.296 0.215 0.377 <0.001 Any non-mRNA vaccine -0.283 -0.464 -0.102 0.002 -0.611 -0.811 -0.411 <0.001 . 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)preprint The copyright holder for thisthis version posted September 13, 2023. ; https://doi.org/10.1101/2023.09.12.23295445doi: medRxiv preprint